Welcome to Ticol Tcl Interpreter (C) M Shaw 2010-2023
_______ __ ___ ___ ____
/_ __(_)______ / / _ __ < / |_ /_ /
/ / / / __/ _ \/ / | |/ / / / / __/ / /
/_/ /_/\__/\___/_/ |___/ /_(_)____//_/
See any associated README.TXT for latest updates and amendments - Index
---------------------------------------------------------------------------
Ticol v1.27 - Copyright and Licence
---------------------------------------------------------------------------
Copyright (C) to M Shaw 2010-2023
All rights reserved.
Portions Copyright (c) 2007-2016, Salvatore Sanfilippo
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
1 Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
2 Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution
3 This product and its associated documentation may not be sold or
resold for profit
4 This licence does not extend to the Ticol Tcl manual
The Ticol Tcl manual, associated documentation and sample scripts
are not included in this licence but are covered separately by the
standard international author copyright for literary works
Copyright (C) M Shaw (2015-202)
All rights are reserved
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
'AS IS' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
TO,THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
Acknowledgements are made for the following code:
ticol_zip plugin, portions
Copyright (c) 1990-1999 Info-ZIP. All rights reserved
The plugin library is not supplied by Info-ZIP
ticol_sqlite plugin, portions
Copyright attributed to SQLite (Public Domain)
MD5 Algorithm: RSA Data Security, Inc. MD5 Message-Digest Algorithm
Copyright (C) 1991-2, RSA Data Security, Inc. 1991.
All rights reserved
Mersenne Twister Random Number Generator
Copyright(C) 1997-2002, Makoto Matsumoto, Takuji Nishimura
TCP/IP specific extensions in Windows Sockets 2
Copyright(C) 1980,1983,1988,1993
The Regents of the University of California
Documentation generated on Sun, 14th May 2023 21:44:36
Ticol Tcl Manual Index
A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z |
<< >> Bit shift (64-bit)
-> => =>> (struct or class dereference)
. dot (or head) command
:: Double-colon (Scope symbol)
$ dereference
$$ double-dereference (multi-dereference)
$dictionary, [set dictionary] (Dictionary Dereference)
$this
addressofb, strptr (Binary Buffer Address)
addslash
after, every (experimental script)
alias
Alphabetic Index of Built-In Ticol Commands
ANSI (American National Standard for Information) - Unicode Support
ANSI Text Display Plugin
any, all
append
Appendix - Windows Version Numbers
apply
args (proc variable or command)
argv and argc
arity
ARP Address Resolution Protocol and RARP (Reverse ARP) Plugin
array - Tcl Array Commands
array append
array blank
array create
array exists
array find (Reverse array lookup)
array foreach
array get
array is_set
array item
array list
array list
array names
Array Notes
array set
array setint
array size
array sort
array statistics
array swap
array trim
array unset
array walk
array_to_list
array_to_string, string_to_array
arrays of struct
asc
assert
Associative arrays
at_exit
atof, ftoa
autoexec
autoexec.tcl
base64
baseconv, convbase
BEDMAS/BODMAS/BOMDAS (Math Operator Precedence)
beep
Big Integer Math Plugin
Big Text Plugin
binary (Experimental binary command)
bintoint, inttobin
bit
Bitmap Encoding Plugin
bool (boolean values)
box
Braces {} (Grouping) and Brace Style
Brainfuck Interpreter Plugin
break
calc - Auto-optimising alternative to [expr]
call
Callbacks
calldll examples
calldllstr, calldllstr_cdecl
Calling methods within arrays of struct
carray plugin
catch
ceil
CGI Application Back-End Configuration
CGI Script Debugging
CGI Web Plugin
chain
Chaining Tcl Commands
charcount
chartok
chdir
chomp
chr
Class Example
clear
clearprocs
clipboard
clock
clock format
clock scan
Clock Scan Workaround
clock split
close, file close
clreol
cls
cmdcount (Command Count)
cmp (Compare)
Code Highlighting Plugin
Coding Standards
Colour Constants
Combination Math Operators
comma
Command Line Arguments
Command Line History
Command Line Interpreter (CLI)
Command Line Interpreter (CLI) Line Editing
Command Line Prompt String
commands
Commands Grouped by Functionality
Comments
Comparing Ticol/Tcl With PHP
Compatibility with Standard Tcl
Complex Nested Dereferences
concat
Conditional Operator - ?:
console
Console Colours
Console Redirection
Console Screen Capture and Restore
const
content
continue
Control+C (break)
Conversion Routines - mkn, cvn
copyfile
Copyright and Licensing
CPU Information Plugin
CRC32 Plugin
Creating Script-Embeddable Content from Files
CScript - VBScript Console Link Plugin
ctime (Unix time)
ctime_to_double, double_to_ctime
cube, cuberoot
Data Structures (Abstract Data Types)
Data Type Ranges
date
date_to_ctime
dateserial
day_of_week
debug
Debugging Ticol
Debugging With option breakpoint, halt, resume
Decimal to Roman Conversion Plugin
default (Placeholder)
defined
DHCP Plugin (Simple DHCP commands)
dict (Dictionaries)
dict append
dict clear
dict create
dict exists
dict for
dict get
dict info
dict item
dict keys, dict values
dict lappend
dict map
dict merge
dict remove
dict set
dict size
dict unset
dict walk
dict with
die
Differences between Ticol and Standard Tcl
dim
Disambiguity: sort
diskfree
diskstat
disktype
dll_close
dll_close
dll_open
DNS Plugin (Simple DNS / Reverse DNS lookups)
do while, do until (Flow Control)
doevents
Domain Error (Math)
Double Quotes
dummy
dump
dumpmem
dumpvariant
echo
elevate (Windows 2000 or Higher)
Elevate Example Script (Windows 2000 or Higher)
Embedded TCX Plugin
Embedding Ticol Tcl in C++ Programs (ticol.dll)
empty
emulating C++ arrays
encrypt
enum
error
Error Codes
escape
eval
Event Log Handling Plugin
every (Script) - Experimental Threaded Ticol Versions Only
Examples
Exception Handling
exec, shell
Executable-Only Code Obfuscation (ticol.exe filename /c)
exit
Expand - Argument Expansion
Experimental @> Struct Command
explain
expr
Expression Functions (Built In Functions)
External DLL Link Plugin - calldll*, calldll*_cdecl
extract
factorial
FAQ - Frequently Asked Questions
fib (Fibonacci)
file
file exists
file gets
file pathcount
file pathstrip
file stat
filesize
filter
find (Help)
floop (floating point loop)
Flow-control: for foreach if else, switch while do goto on loop
flush
fmod
for (Flow Control)
foreach (Flow Control)
Form Input
format
Format String (printf and format)
fpart
fraction and integral
FTP Plugin (Simple FTP Client)
funct
Function Arity
functions
get_fileopen, get_filesave
get_folder
get_win_error
getcwd, pwd
getenv
getkey
gets
getsid
Getting User Input
glob
global
Global Variables
gosub, sub
goto_block, goto
gotoxy
Handling Byte Arrays
Handling filenames and paths
Handling Sparse Integer-Indexed Arrays
Hash Tables
hasp - Ticol Hasp HL Plugin
help topic
Hex Routines
Hints and Tips
html chomp_tag (Plugin)
html header (Plugin)
html header_field (Plugin)
HTML Plugin
html remove_header (Plugin)
html strip (Plugin)
http
ICMP Ping Plugin
if..elseif..else (Flow Control)
in, ni
Include files
index
info
info args
info cmdcount
info exists
info level
info vars
info winver
INI Handling Plugin
inkey
input
input_csv
inputbox
inrange
inspect
instr, rinstr, strstr, strstrr
int, double (Type Casting)
Integer to English Word Conversion Plugin
interp
Introspection and Reflection
inttooct, octtoint
ip_to_long, long_to_ip
is_*
is_active
is_admin
is_array
is_command
is_const
is_date
is_dir
is_dst
is_elevated
is_empty
is_file
is_leapyear
is_list
is_mod
is_null
is_numeric
is_pointer
is_proc
is_set
is_static
item
item (array, stack or struct)
join
json_to_list (JSON Conversion Plugin)
K Combinator
Key Input Routine
Known Bugs and Feature Anomalies
Lambda Example
lappend
lassign
lastchar
lcount
ldepth (proc)
left, string left
len
let (Assignment command)
level
lib (Ticol Plugin Module Extension Loader)
lib interface specifications
lindex
Line Based String Processing Plugin
Line Continuation Character
Line Numbers
link_to_path
Linked List Plugin
list
List Processing Commands
list_to_array
Listing Files
llength
lmap
load (CLI Command - Loads a file into CLI interpreter memory)
logo
loop (Flow Control)
lpush, lpop, ltail, lpeek
lrand
lrange
lremove (command)
lremove (proc)
lreplace
lreverse
ls
lsearch
lset
lsort
Macro #if #exec #echo
Macro Constants
Macro Escape Sequences
Macro Pre-Processor (MPP)
makestr
malloc, free
man - Ticol Manual
map, string map
Math Expressions
Math Operators
max, min (Command form)
md5
MD5 Batch Example
MD5 Verification
md5_file
mean, median, mode
mem_used
memcpy
Memory Allocation and Garbage-Collection
memset
mid
mids (Return a String Midsection)
Minimal Escaping Style
Miscellaneous Routines Plugin
mkdir
mkdirs
mktemp
Modulo Sign (Discussion)
movefile
msgbox
Multi Line Commands
Multiple Commands
Namespaces and Scope
Network Drive Map Plugin
Networking Commands
new
newline
Non-Decimal Numbers
not (Bitwise and Logical)
now
Number Formats
ofaddress, addressof (Tcl Variable Address)
ofaddressb (Binary Buffer from Address)
on (Flow Control)
on_error
open, file open
Optimising [if]...[elseif]...[else] Ladders
option (Runtime Configuration Options)
Passing Arrays to Procedures Using upvar
Passing Structs to a Procedure By Name
Passing Variables to Procedures by Name (by Reference)
Passing Variables to Procedures by Value
pause
Performance and Efficiency
Performance Example
Performance Graph (Profiling)
pid, tid (Process and Thread ID)
pipe
Piping Commands
play_wav, stop_wav
Plugins Extensions for Ticol Tcl
pointers
Potential Plugins and Extensions
precision
Preset Global Variables and Constants
Principle of Least Surprise (POLS)
Print Functions
printf
proc (Tcl Procedures)
Procedure Argument List Variable: 'args' ($args)
procs (info procs)
Pseudo-Assembler Interpreter Plugin (TicASM)
puts
qbf
Quickstart Guide for Ticol Tcl
Quoting Hell
Random File Database Plugin
randomstr
range
read, file read
readfile
Recursion and Issues Arising
regex
Registry Plugin (ANSI)
rename
replacechar
reserved variables
result
return
Returning Struct Variables from Procedures
rewind
rgb
right, string right
RLM Network Licence Query Plugin
rnd, rndseq, rnd64
round
RTF Output Plugin
run - (CLI Command - Run a Ticol Tcl script)
Running Scripts
sbool (Signed Boolean)
sc
scan
Scientific Format Number Plugin
screen
Security
seek
semi-colon
set
setat
setb, struct setb
setc
setenv
setl
setr
setresult
sgn
Shortcut Math Operators
Single Line Commands
Single Quotes
sizeof
sleep
SMTP Mail Plugin
source
spawn
split
SQLite SQL Plugin
sqrt procedure
Square Brackets
stack
stack create
Stack Examples
stack foreach
static (static variables)
stop
store
strchr
string
String Comparison Operators
string equal
String Functions
string index
string is_alpha, string is_alnum
string isempty
string match
string range
string repeat
string replace
string slice
string to_wchar
string toarray
string tolower, string toupper, tolower, toupper
string_to_array string arrayvar Convert a CRLF delmited string to a Tcl array
stripto
striptor
strsplit
strto, strtor
strtok
struct
struct clear
struct copy
struct create
struct dump
struct elements
Struct Implementation Details
struct item
struct set
struct setb
struct size
struct unset
subst
swap
switch (Flow Control)
Syntax Checking (Static)
sysfree
tab
Tcl Commands
TCL History
Tcl Lists
Tcl Programming Idioms
Tcl Semantics
Tcl Strings
Tcl Syntax
tcl2bat (Conversion Batch Script)
TCP Socket Ping Plugin
TCP/IP and Networking Functionality
tell
textcolor
The Tcl Language
Thread Use
TicASM Instruction Set (Plugin)
ticks
Ticol Design and Picol Design Changes
Ticol Design Goals
Ticol Disadvantages
Ticol DLL Stub Program (ticol_stub.exe)
Ticol Editor and Recommended IDE
Ticol Errors
Ticol is a Tcl interpreter, not a compiler!
Ticol Macro PreProcessor Overview (MPP)
Ticol Manual - Copyright (C) M Shaw 2010-2023
Ticol Tcl Architecture
Ticol Tcl Interpreter for Windows
Ticol Uses
Ticol Workarounds
Ticol-Specific, Non Standard Tcl Commands
ticol.ini - Configuration file for Ticol.exe
time
timer
Tips for Writing Big Programs
tk Support
tohex, fromhex
Towers of Hanoi
trace
trim, trimleft, trimright, string trimleft - string trimright
trimprefix
trimsuffix
Troubleshooting and Common Mistakes
trunc
try ... catch (Flow Control)
Tutorial
type
Unary commands: incr, decr, ++, --
UNC Path Handling
undef
undef command (Undefining Commands)
unescape
unknown (proc)
unlink
unset
unwrap (proc)
uplevel
upvar
urlencode, urldecode
Useful Commands
Useful Debug Routines
Using [upvar] with Structs
Using Logical and/or Within Flow-Control Statements
Using Ticol at the Command Line
Using Ticol With Batch Files
variable
Variable Casting
Variables
Variant Array Functions (Plugin)
Variant Handling Plugin
varname
varray get (Plugin)
varray lbound (Plugin)
varray size (Plugin)
varray ubound (Plugin)
vars
VB (Visual BASIC) Time
vfract
vsplit (Plugin)
vwait (Requires threaded compile of Ticol)
waitkey
walk
watch
wchar_length
wchar_to_string
while (Flow Control)
White Space
Why I Wrote Ticol Tcl?
win active
win box
win circle
win close
win cls
win columns. rows
win create
win cursor
win ellipse
win font
win gotoxy
win height, width
win line
win lineto
win moveto
win paint
win para
win preview
win pset
win puts
win rect
win rndcolour
win setpos
win snooze
win textcolor
win wherex, wherey
win xpos, win ypos
win_error
Windows Dialogs
Windows DLL Test Interface (wintest.exe)
Windows GDI Graphics Plugin
Windows GDI Graphics Plugin - Example (LabelMania.scr)
Windows Service Query Plugin
Wordcount Plugin
wrap
wrap example script
writefile
XBase - A Simple Dbase III/IV Plugin
xbase get
ZIP Plugin
Ticol Manual - Copyright (C) M Shaw 2010-2023 |
Index |
_______ __ ___ ___ ____
/_ __(_)______ / / _ __ < / |_ /_ /
/ / / / __/ _ \/ / | |/ / / / / __/ / /
/_/ /_/\__/\___/_/ |___/ /_(_)____//_/
See any associated README.TXT for latest updates and amendments
Use: help <command> or 'exit' to exit the interpreter
Type the following:for help on a particular command, e.g.
help contents help print
help debugging help load
help faq help run
help flow control help tutorial
help licence help tcl
help math help troubleshooting
help syntax help ticol
http://www.kerys.co.uk/ticol
Alphabetic Index of Built-In Ticol Commands |
Index |
Code used to generate this list:
foreach x [commands] { puts $x }
puts "[lcount [commands]] commands"
Count the commands using:
puts [lcount [commands]]
Redirect to a file in Windows using:
ticol.exe ; "foreach x [commands] {puts $x}" /na > cmdlist.txt
.
$
${}
::
!
!=
%
&
&&
&=
*
**
*=
+
++
+=
-
--
-=
/
/=
<
<<
<=
<=>
==
>
>=
>>
?
@
K
^
^=
add
addressof
addressofb
addslash
after
alias
all
any
append
apply
args
arity
array
array_set
asc
assert
at_exit
atof
autoexec
base64
baseconv
beep
big
bigtext
binary
bintoint
bool
box
bit
break
calc
call
carray
catch
cd
cd..
ceil
cgi
chain
char
charcount
chdir
chomp
chr
clear
clearprocs
clipboard
clock
close
clreol
cls
cmdcount
cmp
comma
commands
concat
console
content
const
continue
convbase
copyfile
cpu
cscript
ctime_to_double
cvn
date
date_to_ctime
day_of_week
debug
dec
decr
decrypt
defined
dict
die
dim
diskfree
diskstat
disktype
div
do while
do until
doevents
double
double_to_ctime
dummy
dump
dumpmem
dumpvariant
echo
elevate
empty
encrypt
enum
eof
eq
eqi
eqn
error
escape
eval
event
exec
exit
expand
explain
expr
extract
false
fib
file
filesize
filter
find
float
flush
fmod
for
foreach
format
fpart
fraction
fromhex
funct
functions
ge
gei
get_folder
get_win_error
getcwd
getenv
get_fileopen
get_filesave
getkey
gets
getsid
glob
global
gosub
goto
goto_block
gotoxy
gt
gti
halt
help
hextoint
highlight
hostname
html
http
if
in
incr
index
info
inkey
input
input_csv
inputbox
inrange
inspect
instr
int
integer_to_words (plugin)
integral
inttobin
inttohex
inttooct
ip_to_long
is_admin
is_command
is_const
is_date
is_dir
is_dst
is_elevated
is_file
is_leapyear
is_list
is_mod
is_null
is_numeric
is_pointer
is_proc
is_set
item
join
json_to_list
lappend
lassign
lastchar
lcount
le
left
lei
len
let
level
lib
lindex
link_to_path
linkedlist
list
list_to_array
llength
lmap
load
logo
long
long_to_ip
loop
lrand
lrange
ls
lsearch
lset
lsort
lt
lti
makestr
man
map
max
md
md5
md5_file
mean
median
mem_used
memset
mid
mids
min
mkdir
mkdirs
mkn
mktemp
mod
mode
movefile
msgbox
mul
ne
nei
new
newline
ni
normal_path
now
octtoint
ofaddress
ofaddressb
on
on_error
open
option
pause
pid
pipe
play_wav
precision
printf
proc
procs
put
puts
pwd
randomstr
range
read
readfile
ren
rename
replacechar
resume
return
rewind
right
rinstr
rnd
rot13
round
run
sc
scan
sci
screen
set
setat
setb
setc
setenv
setl
setr
setresult
shell
short
signed
sizeof
sleep
slice
sockping
source
spawn
split
srand
stack
static
stop
stop_wav
store
strchr
string
stripto
striptor
strptr
strsplit
strstr
strstrr
strto
strtok
strtor
struct
sub
subst
swap
switch
sysfree
tab
tell
test
textcolor
ticks
time
timer
tohex
tokenise
trace
trimprefix
trimsuffix
true
trunc
try
type
undef
unescape
unlink
unset
unsigned
unwrap
uplevel
upvar
urldecode
urlencode
username
variable
varname
varray
vars
volume
vsplit
vwait
waitkey
walk
wallpaper
watch
wchar_length
wchar_to_string
while
wide
win_error
wordcount
words_to_integer (plugin)
wrap
writefile
|
|=
||
~
See also: commands by function, Ticol, Tcl, faq
Commands Grouped by Functionality |
Index |
Data Types and Type Manipulation
Database
Date and Time
Debugging and Introspection
DLL and Struct Interface
Encryption, Compression, Randomness and MD5
File and Disk and I/O
Flow Control
Event Handling (Threaded compile of Ticol)
HTML and Internet
Information and Environment
Languages
List Handling
Logic
Macro
Math
Memory
Network
Program Execution
String
User interaction and Display
Windows System Interaction
Data Types and Type Manipulation
Relational: > < >= <= == !=
array
array_to_list
array_set
base64
bit
carray
chain
const
convbase
ctime_to_double
cvn
dict
dim
enum
eq
eqi
eqn
escape
fromhex
ge
gei
global
gt
gti
hextoint
inttobin
inttohex
inttooct
is_const
is_mod
is_null
is_numeric
is_proc
is_set
json_to_list
le
lei
let
lei
let
list_to_array
linkedlist (Linked list plugin)
long
long_to_ip
le
lt
lti
map
mkn
ne
nei
new
octtoint
rtf (RTF plugin)
sc
set
short
signed
sizeof
stack
static
subst
swap
tohex
type
unset
unsigned
variable
varname
varray
vsplit
wchar_length
wchar_to_string
wide
Database
db (Ticol DB plugin)
sql (SQLite plugin)
xbase (DBase 3/4 plugin)
Date and Time
clock
ctime_to_double
date
date_to_ctime
day_of_week
double_to_ctime
is_date
is_dst
is_leapyear
ticks
time
timer
now
Debugging and Introspection
args
array walk
assert
catch
cmdcount
debug
dummy
dump
dumpvariant
empty
halt
info
inspect
mem_used
resume
sc
time
trace
try
type
vars
walk
watch
DLL and Struct Interface
-> Dereference a struct field
=> Set (assign) a struct field value
=>> Dereference then evaluate
addressof
addressofb
class
new
ofaddress
ofaddressb
setb
strptr
struct
Encryption, Compression, Randomness and MD5
crc (CRC plugin)
decrypt
encrypt
file_crc (CRC plugin)
md5
md5_file
randomstr
rle (RLE plugin
rnd
rot13 (Misc plugin
srand
zip (ZIP plugin
File and Disk and I/O
addslash
cd
cd..
chdir
close
copyfile
diskfree
diskstat
disktype
eof
file
filesize
flush
getcwd
get_fileopen
get_filesave
glob
ini (INI Plugin)
input_csv
is_dir
is_file
link_to_path
ls
md
mkdir
mkdirs
mktemp
movefile
normal_path
open
pwd
read
readfile
ren
rename
rewind
tell
unlink
volume
writefile
Flow Control
{*}
after
arity
at_exit
break
call
catch
clear
clearprocs
continue
defined
die
do while
do until
doevents
error
exit
expand
for
foreach
funct
gosub
goto
goto_block
halt
if
is_proc
level
loop
on
on_error
proc
procs
resume
return
sleep
stop
sub
switch
try
undef
uplevel
upvar
vwait
while
Event Handling (Threaded compile of Ticol)
after
vwait
every (script)
at (script)
HTML and Internet
cgi (Plugin)
html (Plugin)
http
urldecode
urlencode
Information and Environment
cpu
info
is_admin
is_command
is_elevated
is_pointer
$tcl_date
$tcl_config
$tcl_cli
$tcl_threaded
$tcl_version
$tcl_precision
$env
$argc
$argv
$win64
Languages
bf (Plugin)
asm (Plugin)
cscript (Plugin)
List Handling
apply
array_to_list
in
is_list
lappend
lassign
lcount
lindex
linkedlist (Plugin)
lappend
lassign
lcount
list
list_to_array
llength
lmap
lrand
lrange
lsearch
lset
lsort
unwrap
ni
Logic
Logical and/or &&, ||
Bitwise and/or &, |
Logical not !
Bitwise not ~
Combined: &=, |=
all
any
bool
false ($::false)
true ($::true)
xor
Macro
#echo
#else
#endif
#exec
#exit
#if
#ifdef
#ifndef
#undef
Math
Bitshift: << >>
Increment: ++
Decrement: --
Combinational: +=, -=, *=, /=, ^=
atof
baseconv
big (big number plugin)
binary
bintoint
calc
ceil
dec
decr
div
double
eval
explain
expr
fib
float
fmod
fpart
fraction
incr
inrange
int
integer_to_words (plugin)
integral
Math operators: + - * / ** ^ %
max
mean
median
min
mod
mode
mul
pipe
precision
round
sci
trunc
unescape
words_to_integer (plugin)
Memory
free
malloc
memcpy
memset
store
Network
arp (ARP plugin)
cidr_match (Misc plugin)
dns (DNS plugin)
ftp (FTP plugin)
hostname
ip_from_host (Ping plugin)
ip_match (Misc plugin)
ip_to_long
is_mac (Misc plugin)
ping (Ping plugin)
rarp (ARP plugin)
sendmail (SMTP mail plugin)
sockping (Socket ping plugin)
Program Execution
alias
args
autoexec
cmdcount
commands
dump
functions
lib
load
option
run
source
String
<=>
add
append
asc
char
chr
chomp
cmp
comma
concat
content
extract
filter
index
instr
item
join
lastchar
left
makestr
memset
mid
mids
range
replacechar
right
rinstr
scan
setat
setc
setl
setr
slice
split
strchr
string
stripto
striptor
strsplit
strstr
strstrr
strto
strtok
strtor
tab
trimprefix
trimsuffix
wordcount
wrap
User interaction and Display
beep
box
bigtext
clreol
cls
echo
find
format
get_folder
get_fileopen
get_filesave
getkey
gets
gotoxy
help
inkey
input
inputbox
logo
man
msgbox
newline
pause
printf
puts
rainbow
screen
textcolor
waitkey
Windows System Interaction
clipboard
console
elevate
event
exec
get_win_error
getenv
getsid
info
mem_used
pid
play_wav
reg (registry plugin)
setenv
shell
spawn
stop_wav
sysfree
username
wallpaper
win_error
See also: command list, Ticol, Tcl, faq
:: Double-colon (Scope symbol) |
Index |
The double-colon symbol operates in a similar way to that found in C/C++
::varname
represents a variable in global (root) scope
$::varname
represents a dereferenced variable in global (root) scope
Ticol Tcl does not support operational namespaces but can emulate them by
allowing for the inclusion of pseudo-namespace prefixes such as...
namespace::varname
or
$namespace::varname
Example:
In Tcl, unlike other programming languages, variables do not "inherit"
into procedures from higher-scope. Globals are visible if addressed using
a double colon prefix...
set s "Hello world"
proc foo {} {
set s "Goodbye cruel" # Declare a conflicting variable 's'
puts $::s # Reference the global variable 's'
}
Result:
Hello world
For partial/embedded substitution, a variable name prefix or any component
of a variable name may be isolated and forced by wrapping in curly braces.
So to replace a variable placeholder which references a pseudo-namespace
you would use, for example (where q holds the pseudo-namespace value)
${q}::varname
This causes $q to be dereferenced in preference to $q::varname
Invalid single colons may currently be used as part of a variable name, but
these require wrapping the whole variable name in braces as follows...
set foo:a 23 # This is not valid as a scope-delimted name
puts ${foo:a} # Single colon requires full bracing
puts $foo:a # This will throw a trappable error
See also: variables, dereference, set, namespace, global
Ticol Tcl Architecture |
Index |
This is a simplified block diagram of the unthreaded Ticol Tcl architecture
from high-level command inputs to lower-level execution
+--+------------------------------------------------------------+
| 1| ticol.exe binary | |
+--+---------------------------+--------------------------------+
| 2| Command Line Interpreter (CLI) or EXE Command Line |
+--+---------------------------+--------------------------------+
| 3| Tcl script | Automatic TCX script decoding |
| | |--------------------------------+
| | | Obfuscation Decoder/Encoder |
+--+---------------------------+--------------------------------+
| 4| Macro Preprocessor (MPP) script preprocessing/optimisation |
| | [calc] expression simplification+optimisation |
| | Substitution of constants |
| | Macro generation+substitution, line numbers, syntax checks |
+--+---------------------------+--------------------------------+
| 5| Internal command registration |
+--+---------------------------+--------------------------------+
| 6| Tcl Parser | Dynamic error checking |
| | | |
+--+---------------------------+--------------------------------+
| 7| [eval] variable substitution, command execution, trace etc.|
+--+---------------------------+--------------------------------+
| 8| [expr] or [calc] natural expression handling/ordering |
+--+------------------------------------------------------------+
| 9| Registered commands, [proc] registration, [undef] |
| | Registered [lib] commands [lib calldll] plugin etc. |
+--+---------------------------+--------------------------------+
|10| [proc] | Registered Tcl procs |
+--+---------------------------+--------------------------------+
|11| Output,trace, errors, debugger |
+--+------------------------------------------------------------+
The CLI or ticol.exe binary command line loads plaintext or
obfuscated scripts with various command arguments. These are automatically
decoded using relevant credentials
Scripts are processed before execution by the Macro Preprocessor (MPP).
The resulting script content may register [proc] instances and interact
with any registered command once executed. The MPP may also execute
Tcl commands as part of the preprocessor operation
The Macro PreProcessor (MPP) performs a number of tasks, including
stripping out comments while preserving line numbers; processing the
internal macro language and optimising natural [calc] expressions into
native Tcl format, as well as performing basic sanity checks on the script
syntax including brace/quote parity checks and substitution of constants
such as \n or 0xfe. The MPP performs a lot of the "heavy lifting" to
speed up interpreter execution.
All commands are evaluated internally by the internal 'eval()' function
which may also be called directly using [eval]. [eval] parses the script
content and breaks it up into commands and arguments, executes commands
and performs variable substitution. This is also where the debugger hooks
and tracing is performed and this is the core of the interpreter
[proc]s are registered or unregistered on-the-fly according to when
commands are encountered. Nested procs are permitted and may be generated
dynamically
The [lib] command can load exported Tcl commands dynamically into the root
namespace where they may be called the same as any other command
(See: lib interface)
See also: Ticol, Tcl, lib, Ticol, lib interface
Expand - Argument Expansion |
Index |
{*} - Argument Expansion Operator
expand - Argument Expansion Command
command {*}{word ?word...?}
string [expand command string...]
Expands a well-formed Tcl list into an argument array for a single
Tcl command. Note that [eval] is required to process multiple commands
The {*} operator operates only on well-formed Tcl lists, not Tcl script
blocks
Inline Tcl scripts are generally not well-suited to calling via this
method unless carefully constructed. Semi-colon separated scripts are
not suitable for use as well-formed Tcl lists with expand. Instead,
put such scripts within a [sub] or [proc] and resolve the expansion
reference to the name of this subroutine.
{*} makes each item in a well-formed Tcl list an individual argument of
the current command. Under normal circumstances, passing a list to a
command passes only one argument (i.e. the complete list). {*} causes
the list to be broken into separate arguments.
Backslash substitutions are performed as is normal for a list and
individual internal words may be surrounded by either braces or
double-quote characters. Its words are then added as individual
arguments to the command being substituted
Variables are not resolved as expansion is resolved before variables
are processed. [subst] may be used to force variable resolution on the
list before passing to {*}
{*} is typically required with chained commands where the outer command
requires separate arguments but the inner one returns a string
The [dummy] command is useful for tracing results of {*} and to test or
debug an argument list.
set s "puts {Hello world}"
# This is will not work as {*} is processed before $s
{*}$s
# This will work as expected. {*} is delayed till after $s
expand $s
Be extremely cautious not to accidentally use an unquoted "{*}" string
within your code where you do not intend it to expand arguments. e.g.
don't use, e.g. ...
puts [unescape "[string repeat 80 {*}]\\highlight0\\cf0\\par "]
as this will attempt to expand within the nested call to [string repeat]
and will lead to near impossible to debug errors
Supplying a whitespace character (tab or space) after {*} is unlawful
For instance:
cmd a {*}{b [c]} d {*}{$e f {g h}}
is equivalent to supplying:
cmd a b {[c]} d {$e} f {g h}
And results in the effective arguments:
a
b
{[c]}
d
{$e}
f
g h
being passed to a command
Example:
set a "puts {Hello world}" # Set var to a command with args
$a # Error: Bad attempt to dereference
{*}$a # Correct method of dereferencing
eval $a # This works just as well, however
eval {$a} # Results in an error
Results
puts' is not recognized as an internal or external command ...
Hello world
Hello world
puts' is not recognized as an internal or external command ...
Example:
# Unset all variables 'a' to 'z' (supplying individual arguments)
unset {*}[split [range a-z]] -nocomplain
# This works just as well
expand "unset [split [range a-z] ""] -nocomplain"
# So does this
eval "unset [split [range a-z]] -nocomplain"
Example:
# Lambda like call with argument passed via [lmap]
set square {lmap {x {expr $x * $x}}}
# {*} will break string into separate arguments
puts [{*}$square 42]
# This also works just as well
puts [eval "$square 42"]
Result:
1764
1764
Example:
set e {info exists}
# {*} will break string into separate arguments
puts [{*}$e a]
assert [{*}$e a] {== $_ 0} -v -line #__LINE__
set a Hello
puts [{*}$e a]
assert [{*}$e a] {== $_ 1} -v -line #__LINE__
# This also works just as well
puts [expand "$e a"]
Results:
0
1
1
Example:
set args [list range foobar 2 4]
puts [string {*}$args]
# This works just as well
puts [expand "string $args"]
# And so does this...
puts [eval "string $args"]
Example:
Set script literal via 'command list' [...list...]
and execute variable contents
set q "puts {1 2 3 4}" # Or set q {puts {1 2 3 4}}
{*}$q
Results:
1 2 3 4
Example:
# [mean] expects a series of arguments not a single Tcl list
# We can expand using {*}
set l "234.34 23.11 37.333 3828.22"
puts [mean {*}$l]
Results:
1030.750749999999900
See:
http://wiki.tcl.tk/17158
http://wiki.tcl.tk/19707
{*} and [expand] Differences
----------------------------
The [expand] command will supply a complete, pre-formed and fully-bounded
list to the interpreter. Thus it is more reliable. The {*} command has to
operate during the parsing of the entire script block so it is more
complex and problematic to detect the full bound of the relevant list
which is to be decoded. The possibility that the presented list may be
less well-formed has to be handled by {*}.
See also: subst, dummy
A semi-colon is used as a punctuator between groups of statements
It marks the absolute end of a command list and acts equivalent to a
newline character.
Example:
set a 0;puts "hello";incr a;puts "a is $a"
# Is equivalent to
set a 0
puts "hello"
incr a
puts "a is $a"
Result:
hello
a is 1
Statements which require to pass multiple commands as a single argument
must be braced, thus:
some-command {arg1; arg2; arg3}
otherwise the command string will be interpreted as:
some-command arg1
arg2
arg3
This should be particularly noted with commands which take a script
element as an argument such as [after]
Command Line Use:
A semi-colon may also be used as the first argument to ticol.exe to specify
a "dummy" script file which is never executed. If specified then any
remaining arguments will be evaluated by the Ticol Command line Interpreter
(CLI) within the same interpreter instance
Example Command Line:
ticol.exe ; "set a 0;puts hello;incr a;puts \"a is $a\""
Result:
hello
a is 1
Example:
ticol.exe ; "puts \"Hello\tworld\""
Result:
Hello world
See also: syntax, Tcl
addslash varname -var -separator <char> -unix
string [addslash string]
Conditionally append a path separation character (slash) to a path string
The backslash character in the string will be escaped
Unix forward slash./ separators may be forced using the -unix argument
This will permit Unix paths and web URLs to be processed
Alternately, any character separator can be specified using the -separator
argument. It is recommended to specify only Windows \\ or Unix /.
Examples:
puts [addslash C:]
puts [addslash .]
puts [addslash \\path\\to\\file]
Results:
C:\\
.\\
\\path\\to\\file\\
A common task is to take user input for an ambiguous path or filespec
and to normalise this into a full path
Example:
# If user enters "." or a path such as c:\foo then we normalise
# Note that we must 'escape' this new path if it contains '\' chars
# We store the normalised path for later use with say, [ls]
if {eq [file extension $filespec] ""} {
set filespec [escape [file normalize $filespec]]
set filespec [addslash $filespec]
set path $filespec
append filespec "*.tcl"
}
See also: file, info
ofaddress, addressof (Tcl Variable Address) |
Index |
integer [addressof variable]
string [ofaddress pointer-variable|address-literal-value -string]
[addressof] returns the address of a simple Tcl variable object (not the
variable's data buffer) and stores it as a string record of that address
Complex abstract data types such as structs or arrays are not supported.
The internal address of an array variable can't be returned as these are
stored as nodes in a hash table.
[ofaddress] performs the inverse operation. The two functions are provides
for use in conjunction with calldll and calldll_cdecl
Both commands are for use with [calldll*] only, and it is strongly advised
that Ticol variables are not manipulated using these commands
[addressof] [addressofb],[set],[$] # Get address
| |
v v
+---------------------+ +---------------------+
| Tcl Variable Object | -> | Variable Data Block |
+---------------------+ +---------------------+
| |
v v
[ofaddress] [ofaddressb] # Dereference
Where a variable is passed to [ofaddress], the address stored in the
internal buffer is dereferenced and the result is returned. Where an
integer literal is passed to [ofaddress], the value is assumed to be
an address and it is dereferenced back to a buffer address
By comparison with C/C++, [addressofb] returns the equivalent of LPSTR*
or LPPSTR (char**)
If a string value is being dereferenced the '-string' argument is
required, otherwise the string may be truncated.
Note that you cannot evaluate [ofaddress "literal-string"]
Such a value must be an integer pointer variable address value returned
by [addressof]
To return a string from pointer value set by [struct setb var] use:
[ofaddress $var]
To return string as a char array use:
[ofaddress var]
When using [calldll*] the address of a Tcl variable buffer/string pointer
by value may be passed by using [addressofb variable] to pass the literal
string address (LPSTR). strptr is an alias for addressofb
For Ticol structs, member values refer to an offset in the struct data
block:
struct s {a 5 b 10 c 20}
Tcl var Address Struct byte offset
-----------------------------------------------------
s => 32542464 (0x1f08f00) 0
s.a => 32542464 (0x1f08f00) 0
s.b => 32542469 (0x1f08f05) 5
s.c => 32542479 (0x1f08f0f) 15
In the above example, both $s and $s.a is 32542464, $s.b is 32542469 etc.
[addressof] and [ofaddress] are useful for assigning values to structs
where the size of the value to be stored is unknown. In such cases we
can store the variable address provided the variable is not reassigned.
See also: setb, ofaddressb, calldll, calldll_cdecl, struct
addressofb, strptr (Binary Buffer Address) |
Index |
integer [addressofb variable]
string [strptr variable]
Inverse of [ofaddressb]
[addressofb] and [ofaddressb] are 'unsafe' and not recommended
Returns the address of the binary string buffer of a variable/string
[addressofb] (or [strptr]) returns the equivalent of LPSTR or (char*)
The command performs a similar function to VarPtr() in Visual BASIC
[addressofb] is for use with [calldll*] only, and it is strongly advised
that Ticol variables are not manipulated using this command
[addressof] [addressofb],[set],[$]
| |
v v
+---------------------+ +-----------------------+
| Tcl Variable Object | -> | Variable's Data Block |
+---------------------+ +-----------------------+
| |
v v
[ofaddress] [ofaddressb]
The [addressof] command, unlike [addressofb], returns the address of a
Ticol variable as a numeric string which can be dereferenced back using
ofaddress
However, this does not point to the variable's string buffer and is not,
therefore suitable to pass via a [calldll*] to a DLL. Instead,
[addressofb] should be used to return the true string buffer address
Example:
set a Hello
[addressof] [addressofb],[set],[$]
| 34602704 | 34600096
v v
+---------------------+ +-----------------------+
| a | -> | Hello |
+---------------------+ +-----------------------+
| |
v v
[ofaddress 34602704] [ofaddressb 34600096]
Hello Hello
[addressofb] may be used to pass the address of a Tcl variable's string
buffer to an external DLL where the contents will be recovered and used
in your program. Care should be taken to avoid buffer overflows. It is
preferable to use a struct when passing data to external DLLs
[strptr] is a friendly alias for [addressofb]
Retrieve 2 byte short value: [ofaddressb var 2]
Retrieve 4 byte long value: [ofaddressb var 4]
Retrieve 8 byte int64 value: [ofaddressb var 8]
Retrieve 256 byte struct 's': [ofaddressb s 256]
Example:
set timeout 0
set datalen 64
set buf [makestr 16] # Create 16 byte buffer as Tcl variable
# Call ping32.dll, passing the binary address of the Tcl buffer
# data block, which will receive the return data
calldll ping32 PingAPI google.com [addressofb buf]:4 datalen:2\
timeout:4 1000
puts $buf
Result:
173.194.78.105
See also ofaddressb
See also: addressof, ofaddress, ofaddressb, set, setb, calldll
ofaddressb (Binary Buffer from Address) |
Index |
string [ofaddressb struct-variable ?byte-width?]
Inverse of [addressofb]
[addressofb] and [ofaddressb] are 'unsafe' and not recommended
Return a binary interpretation of data pointed to in a struct field by
the struct-variable. Normally a string interpretation of the memory
location is returned
[ofaddressb] is for use with [calldll*] only, and it is strongly advised
that Ticol variables are not manipulated using this command
[addressof] [addressofb],[set],[$]
| |
v v
+---------------------+ +-----------------------+
| Tcl Variable Object | -> | Variable's Data Block |
+---------------------+ +-----------------------+
| |
v v
[ofaddress] [ofaddressb]
This will usually be unprintable without using [tohex]. What is often
required is to retrieve a literal numeric value written to memory by
[struct setb] which is the inverse function. [ofaddress] could be used but
that will simply map 'n' bytes from the given memory location
On Windows Intel platforms the data is stored in inverted 'little endian'
format (byte order is reversed)
A byte-width can be specified but this will be rounded to 1,2,4 or 8 bits
You must correctly supply the bit-width interpretation of the data to
match the struct storage field definition. The default bit-width is
4 (a DWORD or 32-bit long)
Example: (where field2 is defined as 8 bytes width)
struct s {field 8} # Single-field struct
struct setb s.field 123 # Set member field to 123 binary (0x7b)
puts [ofaddressb s.field 8] # Retrieve and display
Dump:
008f3440 00 00 00 00 00 00 00 00 - 00 00 00 7B 00 00 00 00 ...........{....
-- -- -- --
00 00 00 7B
Result:
123
Note: Literal integer values greater than the range of DWORD must be
retrieved back using [ofaddressb <varname> 8]
Example:
struct setb foo.a 1234567890123456789 -value
puts [ofaddressb foo.a 8] # Byte width of 8 is required
Result:
1234567890123456789
See also: addressof, ofaddress, struct, struct setb, setb, calldll,
calldll_cdecl
after, every (experimental script) |
Index |
after milliseconds command ?arg? ?arg...?
after milliseconds {script}
after milliseconds {run scriptfile}
after milliseconds {source scriptfile}
Threaded Ticol versions *:
after info
after count
after cancel ?id?
Run a command after a given number of milliseconds
For unthreaded releases of Ticol, processing will halt until the specified
time-interval has elapsed. [after] is implemented as a simple time-delay
(see notes below)
For experimental, threaded releases of Ticol; [after] is event-driven and
will run asynchronously with main scripts with separate [after] scripts
interleaved
The threaded version of Ticol multitasks at script rather than command-
level. [after] and [run] lock mutually so [after] cannot launch either
[after] or [run] whilst a script is in progress from either of these
commands
Interaction with scripts launched using [after] should be controlled using
[vwait]. A continuously running script will typically block an [after]
script. Consider rewriting scripts which need to run continually as an
[every] proc which can be polled at intervals rather than run in a main
script loop (see below)
Extreme care should be taken to avoid scripts clashing, particularly
with respect to common variables or procedures. Pseudo namespace variables
may be used outside procs. Procs will have their own temporary namespace
A proc return value can be passed to any variable which may then be used
as a [vwait] control flag, thus:
after 2500 {set vret [draw_logo yellow $cursor_yloc]}
vwait vret
[vwait] is 'blocking' and will suspend the current script until the named
variable changes
IMPORTANT [after] scripts do not interact with the CLI as these run under
a separate thread. There is no concept of 'input focus'
Note:
Be careful to brace [after] arguments and not use:
after 4000 set vret [foo]
for example, the unbraced [foo] proc will be called before [after]
and thus polled immediately by the interpreter
Threaded versions can be detected via the const '::tcl_threaded' e.g.
if {$tcl_threaded} {
# Call some [after] code
}
Arguments will be concatenated with a space. Formatting and grouping may
be maintained by wrapping arguments in braces
Multiple commands may be specified by using an intervening semi-colon (;)
An [every] command may be defined using a simple 4-line script (see below)
An [at] command can be defined using an 11-line procedure
Unthreaded Compile Example:
# Unthreaded Ticol version example
# Instructions are executed sequentially after a simple delay
after 1000 box 3 3 20 10 - blue
after 1000 load demo.tcl
after 1000 run
after 1000 puts {"hello world"}
after 1000 {puts "hello there...";} {puts "How are you?"}
after 1000 goto 10
after 2000 ping 127.0.0.1
Threaded Version Example (requires threaded compile of Ticol):
proc every {ms code} {
source $code
after $ms every $ms $code # Respawn
}
set done 0
set i 0
proc tick {} {
++ ::i
if {> $::i 4} {set ::done 1}
}
every 1000 tick # Tick once per second
vwait done
puts "Finished ticking at $i"
Results:
Finished ticking at 5
Threaded Version Example (A simple 'DOS TSR-like' clock)
# When run from the CLI the CLI will remain interactive
proc every {ms code} {
source $code
after $ms every $ms $code # Respawn
}
every 1000 {
set x [screen curx]
set y [screen cury]
gotoxy 72 1
textcolor white blue;
puts [time] -nonewline
textcolor
gotoxy $x $y
}
Threaded Compile Example (A useful Tcl [at] scheduler)
# Be careful to avoid calling the Windows 'at' command
if {! [defined at]} {
proc at {time args} {
global at_flag
if {[>= [date $time] [fraction [date now]]] } {
set dt [int [* 86400000 [- [date $time] \
[fraction [date now]]]]]
} else {
set dt [int [* 86400000 [- [+ [date $time] 1] \
[fraction [date now]]]]]
}
after $dt $args
return [/ $dt 1000] # Convert ms to seconds
}
}
at 23:59 {puts "The witching hour is nigh!"}
* NOTE: Threaded Ticol is currently experimental alpha and unlikely
to reach final release quality
See also: vwait, every, time
every (Script) - Experimental Threaded Ticol Versions Only |
Index |
every ms {script}
every cancel
every reset
every trace
[every] is a script which can poll code at regular intervals. Because
each thread is transient (it is launched then exits) it can't be
managed using [after cancel]. Instead, [every cancel] will cancel all
threads. [every reset] must be called before making another script call
to [every]. This is a crude experimental script
Note that the script payload will be run in procedure scope so global
variables must be prefixed with '::'
set _clock_running 0 # Variable return for vwait
set _every::cancel_flag 0 # Module flag variable
set _every::every_trace 0 # Module flag variable
undef every -nocomplain
proc every {ms args} {
if {$::_every::every_trace} {
puts "'[after list]' [after count]"
}
if {eq $ms "cancel"} {
set ::_every::cancel_flag 1
after cancel # Crude: Kills all
} elseif {eq $ms "reset"} {
set ::_every::cancel_flag 0
} elseif {eq $ms "trace"} {
set ::_every::every_trace 1
} elseif {! $::_every::cancel_flag} {
eval $args
after $ms every $ms $args
}
}
Example:
A simple title bar clock:
every 1000 {
console set_title [time]
set _clock_running 1 # Returns for [vwait]
}
See also: after, vwait
Using Logical and/or Within Flow-Control Statements |
Index |
Using with: 'option expression off' (forcing non-expression syntax):
&& || Will perform logical operations on arguments
Arguments will be evaluated as true or false
& | Will perform bitwise operations on arguments,
the result of which may be boolean true/false
0b10 | 0b01 == 3 # i.e. true
0b10 & 0b01 == 0 # i.e. false
Ticol does not perform lazy evaluation with these expression. Tcl always
evaluates every [] statement within a complex expression. The use of
[any] or [all] is recommended as this not only simplifies nested
expressions but performs lazy evaluation where there is more than one
clause. See: any, all
Note that sub-expression which are to be tested using logical and/or
should be bracketed if possible, e.g. ((22/7==3) && (12>1))
Example:
if {[expr ($k != 4) && ($k != 6) && ($k != 9) && ($k != 11)]} { ...
Due to the structure of the Tcl language, compound expressions with [if]
command which use 'and' or 'or' may become complicated. You can either
pass to an [expr] statement or wrap multiple bracketed statements as follows:
if { [&& [statement1] [statement2] } { ...
if { [|| [statement1] [statement2] } { ...
if { [expr [statement1] || [statement2]] } { ...
if { [expr [statement1] && [statement2]] } { ...
Or you can use [option expression on] which will enable expressions for
flow control structures.
Commands presented in [] format are evaluated before the outer expression
statement...
if { [statement1] || [statement2] } { ...
if { [statement1] && [statement2] } { ...
option expression off
for {set i 0} {< $i $adapter_count} {incr i} {
set addr [info ip_address $i]
if { [&& [ne $addr "0.0.0.0"] [ne $addr "127.0,0.1"]] } {
set adapter $i
break
}
}
Example:
# if $op is * and (a xor b) is -ve then option a else option b ...
if {[eq $op *] && [^ [eq $a "-"] [eq $b "-"]] } {
You may also nest 'if' statements to achieve the same effect as 'and'
option expression off
for {set i 0} {< $i $adapter_count} {incr i} {
if { ne [info ip_address $i] "0.0.0.0"} {
if { ne [info ip_address $i] "127.0,0.1"} {
set adapter $i
break
}
}
}
[if.. else] statements may be used to replicate 'or' constructs
--
Using with: 'option expression on' or with [expr]
Logical operators &&, || (and, or) are evaluated left-to-right
Complex expressions using logical combination operators should be properly
bracketed. e.g. 'if {(22/7)==3 && (12>1)} {puts yes} else {puts no}'
Expressions may combine individual values, bitwise to yield a testable
result: expr "1 || 2 || 4"
You may still use bracketed 'eval' type expressions
See also: any, all, if
ANSI Text Display Plugin |
Index |
To load, use: lib ticol_ansi ?-verbose? ?-nocomplain?
Displays ANSI encoded text which would usually require the ANSI.SYS driver
and which is not available in some recent versions of Windows
Syntax:
ansi <text> # Display ANSI or plain text
ansi <text> -slow # Slow down output (useful for animations)
ansi <text> -veryslow # Slowest output (useful for animations)
ansi <text> -mono # Monochrome only (disable colour codes)
ansi <text> -width N # Wrap to console width of N characters
ansi <text> -80 # Force 80 column line-wrap in > 80 col mode
# 80 column displays will wrap at column 80
ansi <text> -step # Single-step output by CSI code
ansi <text> -trace # Show translated CSI code to the Title bar
ansi <text> -noansi # Don't apply the detected ANSI codes
ansi <test> -trans # List translated codes as detected
ansi <test> -codes # Output found CSI codes. 1 per line
# Add -text to include plaintext
ansi <text> -nocrlf # Suppress CRLFs in the source text
ansi <text> -fix # Fix with single line drawing characters
ansi <text> -fix2 # Fix with double line drawing characters
ansi <text> -unescape # Unescape the string if supplied 'escaped'
ansi varname -var ?args? # Supply a variable name instead of string
[ansi] allows rich text display of classic ANSI formatted text of the kind
which was popular in the 1990s. This can be useful for rendering splash
screens or displaying rich content in text mode
[ansi] will not be as fast as a dedicated ANSI.SYS driver written in
assembler but performance is still reasonable
Not all existing ANSI files will display well. Some ANSI drivers will work
by forcing 80 column mode for systems such as DOS which commonly used
80 x 25 resolution. This is incompatible with modern console use
but both 40 and 80 column mode may be forced using the '-40' and '-80'
arguments. Some existing ANSI art files may contain NULL characters and
thus cannot be displayed unless the NULL characters are removed.
Problems may occur displaying old files which relied on console terminal
behaviour or fixed console attributes of a legacy ANSY.SYS driver. Such
files will rely on console line-wrapping for correct display
Windows code pages tend to be missing some line-drawing characters, and
common ones, 221 and 222 may be replaced by 25% block shade (219) using
the argument '-fix'
The 'bell' character (0x07) is filtered out to prevent ANSI-'bombs'
which are comprised of files full of this character
Example:
set s [readfile xmas.ans] # Read a legacy ANSI text file
ansi $s # Display the content in ANSI format
pause -q # Pause with no prompt
Handled CSI Codes
-----------------
A small subset of common and safe ANSI codes are handled, which will
enable most legacy 16 colour content to be displayed: The Windows console
prior to Windows 10 does not support the blink or other attributes.
Keyboard redefinition codes are not supported
The ANSI Control Sequence Introducer (CSI) is comprised of character 27
(ESC) followed by an open square bracket. Various parameters may then
follow, separated by semi-colon characters. A final operation type
character terminates the instruction sequence. These being as follows:
CSI m Screen display (attributes and colour). Intensity (bold) only
CSI A Cursor up (CUP)
CSI B Cursor down (CUD)
CSI C Cursor forward (CUF)
CSI D Cursor backward (CUB)
CSI E Newline (CNL)
CSI F Previous line (CPL)
CSI G Cursor horizontal absolute (CHA)
CSI H Set cursor position (CUP)
CSI f As per: CSI H
CSI 2J Clear screen
CSI K Clear to end of line
CSI U Cursor up (CUU)
CSI s Store cursor position
CSI u Restore cursor position
A Control Sequence Introducer (CSI) string should be defined in Ticol as:
set CSI "[chr 27][chr 91]"
Don't be tempted to use an escaped literal square bracket in place of
[chr 91] or you will need to use [escape] to pass arguments and
the -unescape flag with [ansi]
For more information on CSI visit:
https://en.wikipedia.org/wiki/ANSI_escape_code
Sample ANSI Text
----------------
The following code displays "Hello" in multi-colour and backspaces over
the word 'Coder' to finally print the string '"Hello World'
set CSI "[chr 27][chr 91]"
ansi "${CSI}1;34mH${CSI}33me${CSI}31m${CSI}36ml${CSI}\
0;33ml${CSI}1;32mo ${CSI}1;37m${CSI}sCoder${CSI}uWorld\n"
Creating A Simple ANSI File Viewer
----------------------------------
A simple ANSI file viewer which can be run from the Windows command line
can be created as follows:
option expression off
set filename ""
set arg ""
lib ticol_ansi
if {> argc 1} {
set filename [item argv(1) ""]
}
if {eq [file type $filename] ""} {
append filename ".ans"
}
if {! [file exists $filename]} {
die "File: $filename not found"
}
set s [readfile $filename]
# Using [item] is far easier than checking for valid argc counts
ansi $s $arg [item argv(2)] [item argv(3)] [item argv(4)]
lib ticol_ansi -unload
pause -q
Save the code above as ansi.tcl and use the supplied batch script:
converter, tcl2bat.tcl, to convert using the command:
ticol tcl2bat ansi
You can then run the batch file with optional arguments:
ansi xmas -slow -fix
or run directly as:
ticol.exe ansi xmas -slow -fix
Troubleshooting
---------------
Mangled displays: Check your terminal console is set to 80 columns
Go to the console menu, select properties, then 'Layout' tab.
Ensure you have set BOTH 'Screen Buffer Size' width and 'Window Size'
width to 80 columns or the correct number of columns intended by the ANSI
file
The most useful troubleshooting arguments will probably be '-trace -step'
which will single-step CSI processing and show the codes in the
console title bar
The 'blink' character attribute is not supported in the Windows console
See also: plugins, readfile, item
ANSI (American National Standard for Information) - Unicode Support |
Index |
Ticol is an ANSI only language. That is, extended/international characters,
often called "Unicode", "Wide" or "Multi-byte" characters are not supported
Non-Roman languages such as Chinese are not directly supported but data
might be handled via external DLLs and using the [calldll] plugin
Only the standard 8-bit ANSI character set from 0..255, including many
international accented characters, is supported
This is by design. Ticol is a personal hobby project with no requirement
for non-Roman character sets
Many other applications similarly lack Unicode/MBCS support
e.g. Unix du.exe for Windows cannot handle Unicode path names
Ticol may be used to "drive" Unicode-aware command-line application such as
RoboCopy or XCOPY.
See also: Ticol, why, ANSI plugin
bool [alias alias-name existing-command ?-nocomplain?]
bool [alias alias-name existing-proc ?-nocomplain?]
Create an alias of a proc or command. The command or proc being aliased
must exist and the alias must not exist
An alias can be undefined using [undef] or [rename] as with any command
Take extreme care when creating aliases to procedures and subsequently
renaming or deleting the original reference. The procedure code will
only be removed after the last reference is removed. Also, when testing
for removed references, take care to ensure that proc [unknown] is
not being called as this could confuse your checks.
Example:
alias fred puts
fred "Hello world"
Results:
Hello world
See also: undef, proc, commands
append variable value ?value? ...
string [append variable value ?value? ... -return]
Append a value or series of values to a variable. If the variable does not
exist it will be created and any arguments appended to it
Append provides an efficient means of concatenating values into variables
However, [store add] is much faster
[append] will append to any simple variable or array element only
To concatenate strings without passing to a variable see: [concat]
As the string size expands and to aid speed and efficiency, [append]
returns no string unless the -return argument is used
Example:
option expression off
set var 0
for {set i 1} {<= $i 10} {incr i} {
append var "," $i
}
puts $var
Result:
0,1,2,3,4,5,6,7,8,9,10
See also: concat, store, set, string
string [apply funct ?arg1 arg2 ...?]
Apply an anonymous function
The command [apply] applies the function 'func' to the arguments
arg1 arg2 ... and returns the result.
In mathematics apply is a function that applies functions to arguments.
It is key to programming languages derived from lambda calculus, such as
LISP and also in functional languages
The function 'func' is a two element list {args body} or a three
element list {args body namespace} (as if the list command had been
used). Note that namespaces are only emulated in Ticol
The first element args specifies the formal arguments to func.
The invocation of apply adds a call frame to Ticol's evaluation stack
and acts in equivalent manner to [proc]. So the [info level] value will
be +1 higher than the calling frame
Empty function args and an empty arglist as {} are acceptable
The specification of the formal arguments args is shared with the proc
command. See: proc
Example:
This shows how to make a simple general command that applies a
transformation to each element of a list
proc lmap {lambdaExpression list} {
set result {}
foreach item $list {
lappend result [apply $lambda $item]
}
return $result
}
lmap {x {return [string length $x]:$x}} {a bb ccc dddd}
lmap {x {expr {$x**2 + 3*$x - 2}}} {-4 -3 -2 -1 0 1 2 3 4}
Results:
1:a 2:bb 3:ccc 4:dddd
2 -2 -4 -4 -2 2 8 16 26
Example:
puts [apply {{x y} {expr hypot($x,$y)}} 3 4]
Result:
5.0
Example:
# Empty lambas are allowed
# Show callframe level
puts [apply {{} {return [info level]}} {}]
Result:
1
See:
http://wiki.tcl.tk/4884
http://wiki.tcl.tk/5892
https://en.wikipedia.org/wiki/Apply
https://www.tcl.tk/man/tcl/TclCmd/apply.htm
See also: proc, lassign, K, uplevel, lmap, map, lambda
::argc Is always defined.
::argv Will be defined only if there are command arguments
::argv0 Will always be defined but may be empty
If a script is launched via ticol.exe's command-line then argc will always
be at least 1 and argv(0) will contain the full path and name of the script
file. This applies to both TCL and TCX filetypes. The argv() array starts
at zero and with the second Windows argument (i.e. the Tcl script name)
# Windows/C++ argv[] index
# 0 1 2 3
ticol.exe myscript.tcl foo bar
# Ticol argv index
# N/A 0 1 2
If there are no command-line arguments and [source] has not been used to set
argv(), then argv will not be set, argc will be 0.
Use [is_set], [array item], or ::argc to check if argv exists before
processing $argv()
Examples:
ticol.exe argv.tcl 1 2 3
Result:
--------------------------------------------
hash_table object:0x3c4088 size:250
--------------------------------------------
Bucket(1) Node: 0x20a3f10 chained nodes:1
1: 0x20a3f10 key:'0' => 'argv.tcl'
Bucket(69) Node: 0x20a3fd0 chained nodes:1
1: 0x20a3fd0 key:'3' => '3'
Bucket(94) Node: 0x20a3f50 chained nodes:1
1: 0x20a3f50 key:'1' => '1'
Bucket(135) Node: 0x20a3f90 chained nodes:1
1: 0x20a3f90 key:'2' => '2'
Caution!
--------
If a Ticol filetype (tcl/tcx) is associated with the Windows shell then
this behaviour will differ unless the association definition has been
modified to include command line arguments. Simply associating TCL or TCX
files with Ticol.exe is insufficient to pass command line arguments.
For this reason, take extreme care in launching Ticol scripts via Windows
file associations
Example:
Command: argv.tcx 1 2 3
argv.tcx: array walk argv
Result: (argv higher than 0 is not passed via the command-line)
---------------------------------------------------------------
hash_table object:0x5c4088 size:250
---------------------------------------------------------------
Bucket(1) Node: 0x2113f40 chained nodes:1
1: 0x2113f40 key:'0' => 'C:\\VC5\\MyProjects\\ticol\\Re...'
Example:
if {is_set argv} {
array foreach argv val ss {
puts "argv($ss)='$val'"
}
}
Safe Dereferencing
------------------
[array item] offers the safest means of dereferencing argv() and offers
a default return value
puts [array item argv 0 ""] # Default is ""
puts [array item argv 1 "argv 1 is not set"] # Default is a message
See also: globals, array item
args (proc variable or command) |
Index |
proc <name> {args} {... # Variable
list [args argument-list...] # Command
$args proc variable
------------------
$args is a procedure variable which acts analogously to the C++ ellipsis
pseudo variable placeholder. Where specified,'args' will contain all
subsequent arguments similar to the "C" ellipsis (...)
Example:
proc foo {args} {
puts "\$args=='$args'"
}
foo 1 2 3 4
Result:
$args=='1 2 3 4'
[args] command
--------------
See: proc args for additional information
The [args] command echoes its arguments as a Tcl list.
Example: (Using the {*} expand operator)
puts [args "hello world" 1 2 3]
Result:
{hello world} 1 2 3
See also: proc, expand, debugging, argv
Procedure Argument List Variable: 'args' ($args) |
Index |
proc {argument-list ... args} { ...
The special variable 'args' may be defined as the last argument of a
procedure. When defined this variable will receive a list of zero or more
arguments which are supplied when the proc is called. It is similar in
function to the C/C++ ellipsis function argument "..."
If the special argument, 'args', is not specified as the final argument of
a proc then any excess arguments supplied when the proc is called will be
discarded and ignored
When including several scripts it is recommended to ensure procs from
other scripts are cleaned-up before defining using [undef], e.g.
(See also args for the command [args])
undef foo -nocomplain
Example:
proc foo {a args} { # Will include unlimited trailing args
puts "foo a is '$a'"
puts "foo args is '$args'"
}
proc bar {a} { # Will allow only one arg
puts "bar a is '$a'"
}
foo Hello
foo Hello World
foo Hello World How {Are You} Today
newline
bar Hello
bar Hello World
bar Hello World How {Are You} Today
newline
Results:
foo a is 'Hello'
foo args is ''
foo a is 'Hello'
foo args is 'World'
foo a is 'Hello'
foo args is 'World How {Are You} Today'
bar a is 'Hello'
bar a is 'Hello'
bar a is 'Hello'
Example:
proc foo {x y} { ...
Example:
proc demo {first {second "two"} args} { ...
Results:
See: proc for more information
Procs can also be called in many ways, directly by name or via [call],
[gosub], [@>] or [$]
Example:
proc foo {} {puts Hello}
foo
call foo
gosub foo
@> foo
set p foo
$foo
Results:
Hello
Hello
Hello
Hello
Hello
See also: proc, args, varname, call, gosub
Arity is the number of arguments presented to a given function. There are a
variety of functions available either within expr or funct as follows:
The command [arity] will return the arity of a built-in function
Function Arity 0
----------------
srand Initialise the random number generator
rand Return double random value from 0 to 1
The rand() and srand() functions are not
cryptographically secure, and should not be used
to generate one-time passwords or session keys
Link to the Windows API using [calldll] instead
Function Arity 1
----------------
abs value Return the absolute value of a number
acos value Return the arc cosine of value (arccos)
asin value Return the arc sin of value (arcsin)
atan value Return the arc tangent of value (arctan)
bool value Evaluates a string and returns 1 or 0
ceil value Round to the next-highest integer value
cos/cosh value Return the cosine of value
decr value Decrement value
degrees value Convert radians to degrees
double value Cast value to double
exp value Return the exponent of value
fib value Return the Fibonacci of value
fix value Return integer value rounded up
floor value Round to the next-lowest integer value
fraction value Return the fraction part of a real/double
incr value Increment value
int value Cast value to integer
integral value Return the integer part of a real/double
log value Return the logarithm of value
log10 value Return the log10 of value
round value (0 places) Round value to 0 places using standard rounding
sgn value Return the sign as an integer (1, 0, -1)
sin/sinh value Return the sine of value
sqr value Return the square of value
sqrt value Return the square-root of value
srand seed Seed the random number generator
tan/tanh value Return the tangent of value
wide value Force a value to 64 bit signed integer
Function Arity 2
----------------
div a b Functional division of a / b
fmod a b Return the remainder of a / b
hypot a b Return the hypotenuse of a and b
max a b Return the maximum of a and b
min a b Return the minimum of a and b
mul a b Functional multiplication of a * b
pow a b Return a to the power of b
rnd a b Return a random number between a and b inclusive
round a b (b places) Round a to b decimal places
You can use the [functions] command to get a list of supported [expr] functions
You can use the [funct] command to export and access an expression function
See also: expression functions, arity, functions, funct, expr, eval
integer [arity functname]
Return the arity count of an inbuilt function as an integer value
If the function is not found then -1 is returned
Example:
puts [arity rand] # No arguments
puts [arity abs] # One argument
puts [arity max] # Two arguments (funct)
puts [arity fred] # No such command
Results
0
1
2
-1
For an arity list of each available function. See: function arity
See also: function arity, expression functions, functions, funct,
expr, eval
ARP Address Resolution Protocol and RARP (Reverse ARP) Plugin |
Index |
Load using: lib ticol_arp?.dll?
arp ipv4-address ?-list? # ARP lookup of IPV4 address
arp table # Return the ARP table as a Tcl list
rarp mac-address ?-list? # Reverse ARP lookup. Look up in ARP table
Address Resolution Protocol (ARP) queries will query the local ARP table.
Results will depend on the content of the table
[arp] returns a string value and may return an empty string for interfaces
which return no MAC address. No mac address is returned for addresses which
cross NAT firewalls or for loopback/dummy interfaces. Such addresses must
be on the same LAN
[arp table] returns a Tcl list with one {ip-address mac-address status}
list per interface found
[rarp] can only return results for "known" IPV4 addresses which have been
discovered and placed in the local ARP table cache. Your workstation must
have been in conversation with these addresses, (e.g. via ping)
Lookups which return empty {} string are not fatal errors. The return should
be tested to see if the return value is empty, which indicates the IP or MAC
address was not found
For exceptions such as malformed IP or MAC addresses an error exception is
raised with an internal ARP library error code or a Windows system error code
The -list option will return a list which includes the error code as the
2nd item.
[arp] Error Codes
-----------------------------------------------------------------
0 Success
1 Success but empty MAC address returned
2 Failed to load iphlpapi.dll
3 Failed to link to lib function SendARP
4 MakeMacAddress failed
5 Bad PMAC buffer
6 Bad IP address
7 Unhandled error
* SendArp Windows API error code
31 ERROR_GEN_FAILURE (Network error/disconnected interface)
50 ERROR_NOT_SUPPORTED
67 ERROR_BAD_NET_NAME
87 ERROR_INVALID_PARAMETER
111 ERROR_BUFFER_OVERFLOW
1168 ERROR_NOT_FOUND
1784 ERROR_INVALID_USER_BUFFER
[rarp] Error Codes
-----------------------------------------------------------------
0 Success (match found)
1 Failure (no match)
2 Bad MAC address
3 Bad MAC format
4 Unknown/unhandled error
Example:
puts [arp table]
Result:
{224.0.0.2 00-00-00-00-00-00 Static} {224.0.0.22 00-00-00-00-00-00 St
atic} {224.0.0.252 00-00-00-00-00-00 Static}
Example:
set t [arp table]
foreach l $t { # foreach list item
foreach {i m s} $l { # IP, MAC, Status
puts "$i\t$m\t$s"
}
}
Result:
224.0.0.2 00-00-00-00-00-00 Static
224.0.0.22 00-00-00-00-00-00 Static
224.0.0.252 00-00-00-00-00-00 Static
Examples:
arp 127.0.0.1 # Loopback adapter
arp 8.8.8.8 # Google DNS
arp 192.168.1.1 # Class C local IPV4 address
rarp 3c46d8cc5cce # Must be in the ARP table (ping first)
rarp beefdeadbeef # Must be in the ARP table (ping first)
Results:
{} # Empty string as loopback adapter not in cache
{} # Empty string as 8.8,8.8 is an internet address
3c46d8cc5cce # MAC address of workstation on local LAN
192.168.1.1 # Reverse lookup from MAC to IP
{} # No such MAC address in the ARP table
Examples:
arp 1.2.3.4 -list # No such address?
rarp pqrs -list # Bad MAC address
Result:
{} 31 # Empty return + network error
See also: ping, dns, dhcp, plugins
array - Tcl Array Commands |
Index |
Various commands for manipulating standard Tcl arrays. Arrays are also
addressed and dereferenced using the dollar ($) sign in the same way
as for languages such as PHP. See a topic below for more information:
arrays Information about standard Tcl arrays
array append a b ?n? Append array b to array a for ?n? rows
array blank arrayVar s ?c? ?n? Blank an integer-indexed array
array create arrayVar ?n? Create an empty array
array exists arrayVar Return boolean true if array exists
array find arrayVar string Return index/subscript or empty string
array foreach arrayVar v s code Iterator
array get arrayVar Return an array as a key + value list
array is_set arrayVar Boolean 1 if an array exists else 0
array item arrayVar subscript Array indexing with default option
array list arrayVar Exports an array to an unsorted list
array names arrayVar Returns a list of subscript values
array set arrayVar list Set array using a list of parameters
array setint arrayVar list... Set integer-indexed array using lists
array size arrayVar Return the size of array if it exists
array sort arrayVar Exports the array to a sorted Tcl list
array statistics arrayVar Show statistics for a given array
array swap arrayVar s1 s2 Show swap 2 array elements
array trim arrayVar ?mask? Trim L/H and R/H end of array values
array unset arrayVar Delete array & all allocated elements
array walk arrayVar Walk the array hash-table
string_to_array string arrayvar Convert a CRLF delmited string to a Tcl array |
Index |
Example:
set a(0) Hello
puts $a(0)
Example:
set a(0) Hello
puts [array item a 0]
Example:
set a(0) Hello
puts [array get a 0]
Example:
set a(0) Hello
puts [set a(0)]
See: help array <subcommand>
e.g.: help array list
For "C"-like integer indexed, non-associative arrays see: help carray
See also: arrays, variant arrays, array foreach, passing variables by name
varname, carray, string_to_array
integer [array append array1 array2 ?count?]
Append one integer-indexed array onto the end of another. Optionally, this
can be restricted to one or more elements of the 2nd array. The returned
integer is the new count of the appended-to array
Integer-indexed arrays are returned by a number of Ticol command such as
[split]
Example:
split $qbf " " -array a # Splits into an integer-indexed array
split "Now is the time for all men to come to the party" " " -array b
set r [array append a b] # Append 2nd array on to the 1st
assert [array size a] {== $_ 22} -v -line #__LINE__
puts "New size is $r"
Result:
New size is 22
This is an equivalent [proc] solution
option expression off
proc array_append {_a _q} {
upvar $_a a
upvar $_q b
set q [array size a]
set r [array size b]
for {set i 0} {< $i $r} {++ i} {
set a([+ $q $i]) $b($i)
}
return $_a
}
See also: arrays, array foreach, array size, carray
integer [array blank arrayname size ?blankchar? ?-start N?]
Blank (or create) an integer-indexed array of the given size
[array blank] is useful for quickly creating and initialising very
large arrays
If 'blankchar' is specified then each element will be intialised to this
value. The array index will start at 0 unless otherwise specified by
the '-start N' argument.
An integer count of elements successfully set is returned
Example:
array blank foo 4 0 -start 3
loop i 3 7 {
puts "foo($i) == '$foo($i)'"
}
Result:
foo(3) == '0'
foo(4) == '0'
foo(5) == '0'
foo(6) == '0'
See also: array, array set, array walk, carray
array foreach arrayName iteratorVar ?subscriptVar? {code}
Iterate an array, applying a block of code with each element's value
passed into valVar and, optionally, the array's subscript value passed
into subscriptVar. The array element value precedes subscript due to
subscript being an optional parameter
[array foreach] is functionally equivalent to a combination of [foreach]
and [array get] but is more efficient
# Emulate [array foreach] using [array get] and [foreach]
foreach {x y} [array get env] {
printf "%-20.20s\t=%-45.45s\r\n" $x $y
}
One common alternative is to get the names and then order them
Tcl arrays are associative and intrinsically unordered. Values cannot be
retrieved from an array in the same order that they were set. Arrays will
be retrieved in random (unsorted) order. Integer-indexed arrays should
be used where a sorted order is required
Don't wrap a pair of value and subscript variables in braces
Example:
set i 0
array foreach env val ss {
printf "$i:\t%-20.20s=%-45.45s\n" $ss $val
incr i
}
Result: (prints out environment table)
1: ProgramFiles =C:\Program Files (x86)
2: USERPROFILE =C:\Users\Admin
Example:
ticol.exe ; "set i 0; array foreach env x {puts $i:$x;[incr i]}"
Result:
Prints out environment
See also: foreach, arrays, array, carray
array create arrayName ?size?
Alias for [dim arrayName size]
Creates a Tcl associative array with the default pre-allocation of 1
If 'size' is specified then the array will have memory preallocated to that
size. This will prevent a need to reorganise the storage table on the fly
It is not mandatory to use [array create]. Arrays will be created on first
use as with standard variables, for example, the statement:
set a(0) Hello
will create array 'a' and array element '0'
[array set] will be the most common means of creating arrays with pre-
determined contents, e.g.
array set colours {
red 1
green 2
blue 3
white 4
}
will create array 'colours' with 4 elements
Example:
array create names 10
array walk names
Results:
-----------------------------------------------------------------
hash_table object:0x35ec650 size:10
-----------------------------------------------------------------
Hash Table Statistics:
Root size:10 Occupation:0 (0.00%) Max depth:0 Nodes:0
-----------------------------------------------------------------
See also: array, arrays, dim, carray
array_to_list arrayvar ?listvar? ?-values?
Converts an array to a well-formed but unsorted Tcl list containing pairs
of keys (subscripts) and matching values. Optionally this list can be
passed into an output variable. Both the array and optional list variables
must be passed by name
If the -values argument is used then only values are returned
Use [lsort] to sort the resulting list
Example:
set q [lsort [array_to_list env]]
puts "Returned [lcount q] item(s)"
newline
set i 1
foreach pair $q {
foreach {l r} $pair {
puts "$i\t$l\t=$r"
}
++ i
}
Result:
1 ALLUSERSPROFILE =C:\ProgramData
2 APPDATA =C:\Users\Admin\AppData\Roaming
3 COMPUTERNAME =FIDO
4 ComSpec =C:\Windows\system32\cmd.exe
5 CommonProgramFiles =C:\Program Files (x86)\Common Files
6 CommonProgramFiles(x86) =C:\Program Files (x86)\Common Files
7 CommonProgramW6432 =C:\Program Files\Common Files
...
See also: list_to_array, lsort, array, list, carray
array is_set arrayVar
Returns boolean 1 ($true) if an array exists, otherwise 0 ($false)
Example:
set a(0) Hello
puts [bool [array is_set foo]]
puts [bool [array is_set a]]
Result:
false
true
See also: is_set, carray
Return a boolean indicating if an array exists
array exists arrayName
array exists arrayName(element)
When used with the name of an array variable this tests only if the array
variable name has been set. This does not guarantee any elements in the
array are assigned. To check the size of an array use [array size arrayName]
An array element represents a single variable bound to an array and should
be tested using [array exists arrayName(element)]
You may also use [info exists varname] to test for an array element
In all cases the variable to be tested must NOT be prefixed by a dollar $
sign
Case will be ignored when searching for the special system arrays 'env' and
'argv'
Examples:
array exists argv
array exists env
array exists argv(0)
array exists env(TEMP)
info exists argv(0)
info exists env(TEMP)
See also: is_set, is_empty, arrays, array, array size, info exists, carray
array find (Reverse array lookup) |
Index |
array find arrayName valuestring ?sequence? ?-nocase? ?-length N?
Searches an array by looking up a value to return the index/subscript
of the first matching item with full string match, or an empty string if
the item does not exist. The match is case-sensitive and will match the
entire string. It will not match a substring
[array find] is not a standard Tcl command and is much less efficient
than subscript indexing, but should be adequate for small arrays of less
than a few thousand elements
Example:
split $qbf " " -array a # Split into integer-indexed array
set i [array find a "fox"]
puts "fox is in element '$i'"
Results:
fox is in element '3'
The '-length' argument allows comparison on a given number of characters
Example:
puts [array find env "x86" -nocase -length 3]
Result:
PROCESSOR_ARCHITECTURE
Example:
array set fruits {
# Subscript/key
# Value
blue berry
green apple
orange orange
red cherry
yellow lemon
}
set i [array find fruits cherry]
puts "cherry is in element '$i'"
Results:
cherry is in element 'red'
Name to Integer Lookups
-----------------------
String to integer lookups using an array indexed on integer values can be
done using [array find]
array months {
1 January
2 February
# ... etc.
} -const
# e.g. where argv(1) is a month name we return an integer
set month [array find months $argv(1) -nocase -length 3]
However, consideration should be given to inverting the lookup array so
as to render the use of [array find] unnecessary. Note however that
[array item] will fail "safe" if the item is not found whereas $()
reference does not unless the [array set -nocase] option is used
array months {
Jan 1
Feb 2
# ... etc.
} -const -nocase
# e.g. For "May" we return 5
set month $month([string left $argv(1) 3])
# or
# set month [array item months [string left $argv(1) 3] ""]
Sequential Search
-----------------
Tcl arrays are not ordered vectors and Tcl arrays are stored in a
randomised hash table. This can make sequential searching for multiple
identical items problematic if the array is not integer indexed.
A sequence number may be given to find the Nth matching item in an array
A sequence value < 1 is ignored and returns the first matching item
A sequence value >= 1 returns the Nth matching item
Sequence numbers are inefficient but allow loop searching for N matching
items within an array. Note that arrays are stored unordered and case
significant searches may return items in a different sequence to case
insignificant ones. Case and case-insignificant searches should not be
mixed within the same sequential search.
# findstr is set with some string to search in array 'a'
option expression off
set found ""
set sequence 1
do {
set found [array find a $findstr $sequence]
puts "'$found'\t'[array item a $found {<not found>}]'"
++ sequence
} while {ne $found ""}
See also: arrays, carray
list [array get arrayname]
string [array get arrayName ?mode? ?pattern?]
string [array get arrayName ?pattern? ?mode?]
Return an array as a key + value list. [array get arrayName] will return
the entire array as a list of pairs of keys+values
Where ?mode? is any of:
-glob Wildcard string matching (the default)
-exact Exact string matching (-glob is the default)
-all Return all items
and optionally:
-pair Return array subscript+value pairs
-nocase Ignore case when making comparisons
Note that [array get] is less efficient than direct array addressing as
Tcl associative arrays are hashed and meant to be indexed efficiently by
the precise key value (subscript)
Case will be ignored automatically for arrays set using -nocase or for the
special global arrays env() and argv()
Example:
array get env u*
Result:
USERPROFILE C:\Users\Admin USERNAME Admin USERDOMAIN SNOOPY
Example: (Prints out the environment table)
foreach {x y} [array get env] {puts $x=$y}
Result:
ProgramFiles=C:\Program Files (x86)
USERPROFILE=C:\Users\Admin
OS=Windows_NT
ComSpec=C:\Windows\system32\cmd.exe
PROCESSOR_ARCHITEW6432=AMD64
SystemDrive=C:
Example:
foreach x [lsort [array get env -nocase -pair]] {
puts [lindex $x 0]=[lindex $x 1]
}
Result:
ALLUSERSPROFILE=C:\ProgramData
APPDATA=C:\Users\Admin\AppData\Roaming
HOMEDRIVE=C:
HOMEPATH=\Users\Admin
LOCALAPPDATA=C:\Users\Admin\AppData\Local
USERNAME=Admin
WINDIR=C:\Windows
Example:
foreach {x y} [array get env proc* -nocase] {puts $x=$y}
Result:
PROCESSOR_ARCHITEW6432=AMD64
PROCESSOR_IDENTIFIER=Intel64 Family 6 Model 37 Stepping 5, GenuineIntel
PROCESSOR_REVISION=2505
PROCESSOR_LEVEL=6
PROCESSOR_ARCHITECTURE=x86
Example:
puts [wrap [array get env -glob c*] 60]
Results:
ComSpec C:\\Windows\\system32\\cmd.exe COMPUTERNAME LENOVO C
ommonProgramFiles(x86) {C:\\Program Files (x86)\\Common File
s} CommonProgramFiles {C:\\Program Files (x86)\\Common Files
} CommonProgramW6432 {C:\\Program Files\\Common Files} COMMP
ath {C:\\Program Files\\Lenovo\\Communications Utility}
See also: array names, arrays, array, info exists, carray
array list arrayName
Exports the array values to an unsorted Tcl list
Note that this does not include the array subscripts. Ticol lists are
slow and inefficient, particularly when used to store large numbers of
items (> 1000 or more). For small lists the difference in performance
is insignificant
Example:
# Export and iterate each list value item
foreach {x} [array list env] { puts $x }
Result:
C:\Program Files
C:\Users\Admin\AppData\Local
C:\Users\Public
C:\Users\Admin\AppData\Roaming
C:\ProgramData
$P$G
See also: arrays, array names, array item, carray
array item arrayName subScript ?default value? ?-nocase? ?-nocomplain?
Resolve an array index with optional default return value. This is helpful
since associative arrays may be 'sparse' and there may therefore be no
guarantee that any element exists at all
[array item] is not a standard Tcl command and it offers a safe lookup of
arrays. The standard, 'inline' method of referencing arrays will return a
fatal error where the array element does not exist. [array item] will
instead return an empty string where the array element does not exist
The referenced array variable must exist
[array item] may be combined with [array exists array subscript -nocase] to
safely handle ambiguous/case-insensitive array lookups
If the element is not found and a default value given then the default value
will be returned. This permits safe/simple array lookups with defaults
A fatal error is raised only where the referenced array does not exist
unless '-nocomplain' is passed. In which case the default will be returned
The variable name argument must not be dereferenced using the $ symbol
Correct:
puts [array item env temp -nocase]
Wrong:
puts [array item $env temp -nocase] # $env returns the tag "<array>"
Example:
array set smtp_code { # Array of SMTP error codes
221 "Server ready"
250 "Send success"
252 "Cannot verify user, but it will deliver"
}
puts [array item smtp_code 250 "Unknown error"]
puts [array item smtp_code 251 "Unknown error"]
Result:
Server ready
Unknown error
See also: array, arrays, carray
array names arrayName ?mode? ?pattern?
Where ?mode? is any of:
-glob Wildcard string matching (the default)
-exact Exact string matching (-glob is the default)
-nocase Ignore case when making comparisons
-all Return all items
Returns a list of array subscript values (array element name identifiers)
Case will be ignored automatically for arrays set using -nocase or for the
special global arrays env and argv
Example:
puts [array names env u*]
Result:
USERPROFILE USERNAME USERDOMAIN
Example:
foreach {x} [array names env t* -nocase] {puts "$x\t$env($x)"}
Result:
TEMP C:\Users\Admin\AppData\Local\Temp
TMP C:\Users\Admin\AppData\Local\Temp
Example:
foreach {x} [array names env -exact windir] {puts "$x=$env($x)"}
Result:
windir=C:\Windows
See: array get for similar examples
See also: arrays, carray
integer [array set arrayName list ?-const? ?-nocase?]
Set and initialise an array using a list of parameters. This may be a single
Tcl list or series of arguments. [array set] cannot be used to append items
to an existing array. The array must not already exist
The standard Tcl command is extended to add constant arrays (-const) and
case-insignificance lookup attributes (-nocase). Arrays which are constant
cannot be reassigned-to but can be unset
Arrays which are set using -nocase will be much less efficient but can be
indexed by subscript values in a case-insensitive manner. Scripts should be
designed to use large arrays in a case-sensitive manner. For small arrays
the performance impact will be negligible
A subst is performed on array pairs before assignment. Variables are
substituted and escaped characters are translated. Commands are not
executed
Use: [set array(subscript) value] - to set individual array elements
[array create] can be used to create an array without assigning any
elements and [array create size] can be used to pre-allocate an array of a
fixed size to avoid time-consuming dynamic reallocation
[array set] returns an integer count of the number of array elements
created
Example:
array set colours {
red 1
green 2
blue 3
white 4
}
array set colours {red 1 green 2 blue 3 white 4}
array walk colours
array foreach colours a b { printf "%s\t%s\n" $b $a}
Result:
---------------------------------------------
hash_table::walk() Object:0x1df43f0 size:250
---------------------------------------------
Bucket(138) Node: 0x1e08840 chained nodes:1
1: 0x1e08840 key:'green' value:'2'
Bucket(140) Node: 0x1e08870 chained nodes:1
1: 0x1e08870 key:'blue' value:'3'
Bucket(181) Node: 0x1e088a0 chained nodes:1
1: 0x1e088a0 key:'white' value:'4'
Bucket(238) Node: 0x1e08810 chained nodes:1
1: 0x1e08810 key:'red' value:'1'
green 2
blue 3
white 4
red 1
Example: (Calling commands and passing the result into [array set]
set p [expr 4*atan(1)] # An accurate representation of pi
set bp [expr 22/7.0] # An inaccurate representation of pi
array set mathstuff {
pie $p
badpie $bp
}
array foreach mathstuff v ss { puts "Item: $ss\tValue '$v'" }
Result:
Item: badpie Value '3.142857142857143'
Item: pie Value '3.141592653589792'
See also: set, array, arrays
array setint arrayvar {list} ?{list} ...? ?-base N? ?-const? ?-nocase?
Set an integer-indexed array using a standard Tcl list of values.
More than one list may be passed, in which case, each subsequent list of
items will be appended to the array. If an array does not exist one will
be created.
By specifying the array base, singular or groups of items may also be
replaced within an existing array.
The default array subscript base starts at 0
The '-const' argument allows a const array to be created.
Constant arrays require the use of [array unset varname -const] to unset
The '-nocase' argument allows an array to be created which is case-
insensitive with respect to subscript references (See: arrays)
Example:
array setint a {one two three} {four five six}
loop i 0 [array size a] {
puts "array a: $i -> '$a($i)'"
}
newline
array setint a {THREE FOUR} -base 2
loop i 0 [array size a] {
puts "array a: $i -> '$a($i)'"
}
Results:
array a: 0 -> 'one'
array a: 1 -> 'two'
array a: 2 -> 'three'
array a: 3 -> 'four'
array a: 4 -> 'five'
array a: 5 -> 'six'
array a: 0 -> 'one'
array a: 1 -> 'two'
array a: 2 -> 'THREE'
array a: 3 -> 'FOUR'
array a: 4 -> 'five'
array a: 5 -> 'six'
See also: array, arrays, array set, array unset, is_const, carray
array size arrayName # Return the size of an array if it exists
If the array is not assigned (does not exist) then an empty string is
returned
If the array exists (has been declared) but with no elements assigned then
zero is returned
Example:
# a # Not declared
set b {} # Not an array
dim c 1 # Create empty array using [dim]
set d(1) 0 # Assigned array and element
array set e {1 one 2 two 3 three} # Multi-element set
puts "Undeclared a '[array size a]'"
puts "Non-array b '[array size b]'"
puts "Empty-array c '[array size c]'"
puts "Declared d '[array size d]'"
puts "Declared e '[array size e]'"
Results:
Undeclared a ''
Non-array b ''
Empty-array c '0'
Declared d '1'
Declared e '3'
See also: array, array unset, carray
array list arrayName # Exports the array to an unsorted Tcl list
Note that the list contents will be in random, not original, order
Example:
foreach x $qbf { set q($x) $x }
Result:
the brown jumps lazy fox quick over The dog
Example:
foreach x [array list env] { puts $x }
Result:
D:\WINDOWS
D:\WINDOWS\system32\cmd.exe
See also: arrays, array sort, array, carray
list [array sort arrayname ?newvar? ?-reverse? ?-names? ?-numeric?]
integer [array sort arrayname ?newvar? ?-reverse? ?-array name?\
?-names? ?-numeric?]
Creates a sorted Tcl list of array values. Note that the subscript (key)
values are not returned and the internal array hashtable structure cannot
actually be sorted, only the returned output
The default sorting is to sort alphabetically (A..Z). To sort numeric data
use the -numeric argument. This will sort on the basis of 64 bit integers.
Float sorting is not supported
By default a sorted Tcl list is returned. Alternately, another integer-
indexed array can be created, in which case the new array element count
is returned.
The default sort order is normal (A-Z)
Sorting in Integer Sequence
---------------------------
[array sort] can also output an integer sorted array indexed at base 0 by
specifying the '-array' argument
Example:
option expression on
for {set i 0} {$i < 2000} {incr i} {
set a($i) [randomstr 20]
}
puts [llength [array sort a]]
Result:
2000
Example:
option expression on
for {set i 0} {$i < 2000} {incr i} {
set a($i) [randomstr 20]
}
puts [array sort a b]
Result:
2000
Example:
option expression off
array sort env -array out
set s [array size out]
for {set i 0} {[< $i $s]} {incr i} {
printf "%4i:\t%-40.40s\r\n" $i $out($i)
}
Results
Outputs values in the ENV table
Example:
# Using a temporary and [swap] to swap between sorted and unsorted arrays
# [swap] moves no data, it simply renames the two variables
option expression off
array sort foo -array out
swap out foo # Swap $foo and $out
set s [array size foo]
for {set i 0} {[< $i $s]} {incr i} {
printf "%4i:\t%-40.40s\r\n" $i $foo($i)
}
Sort by Array Names (Subscripts/Keys)
-------------------------------------
If -names is specified together with -array and an array value then the
array will contain an exported, integer-indexed list of array names
(subscript keys) rather than array values. The resulting integer-indexed
array can then be used to index the original array in that sorted order
Example:
array set unsorted {
The orange
quick yellow
brown blue
fox black
jumps magenta
over white
the cyan
lazy green
dog red
}
array sort unsorted -array sorted -names # Output to array 'sorted'
set size [array size sorted] # Grab the array size
option expression off
for {set i 0} {[< $i $size]} {incr i} { # Iterate numerically
printf "%4i:\tref: %s =>\t%-40.40s\r\n" $i $sorted($i) \
$unsorted($sorted($i))
}
Results:
0: ref: brown => blue
1: ref: dog => red
2: ref: fox => black
3: ref: jumps => magenta
4: ref: lazy => green
5: ref: over => white
6: ref: quick => yellow
7: ref: The => orange
8: ref: the => cyan
[array sort] is not a standard Tcl command
See also: arrays, array list, array, swap, Tcl lists, carray
array statistics arrayname
Show statistics for a given array. This shows the total number of root
entries in the array's hash-table as well as the total number of buckets
(nodes) including the root entries.
Additionally, a summary is shown of the first 10 most relevant top-level
bucket allocation counts. Ideally the average search depth for a given root
entry should not be much more than about 2 or 3 but as the allocation is
largely random and this is therefore impossible to achieve completely.
Ticol enhances the usual Tcl statistics table by showing the total bucket
count per-category in brackets.
Example:
option echo on
option expression on
for {set i 0} {$i < 2000} {incr i} {
set a($i) [randomstr 20]
}
array statistics a
Result:
2000 entries in table, 1000 buckets
number of buckets with 0 entries: 0 (0)
number of buckets with 1 entries: 270 (270)
number of buckets with 2 entries: 269 (538)
number of buckets with 3 entries: 173 (519)
number of buckets with 4 entries: 94 (376)
number of buckets with 5 entries: 36 (180)
number of buckets with 6 entries: 16 (96)
number of buckets with 7 entries: 3 (21)
number of buckets with 8 entries: 0 (0)
number of buckets with 9 entries: 0 (0)
number of buckets with 10 or more entries: 0 (0)
average search distance per entry: 2.32
Array hash-table performance can be configured using ticol.ini settings
but it is strongly recommended to keep to the defaults
See: ticol.ini
See also: arrays, array, array walk, carray
bool [array swap arrayVar subscript1 subscript2]
Swap the contents of two array elements
Example:
set a(0) world
set a(1) Hello
puts "$a(0) $a(1)"
array swap a 0 1
puts "$a(0) $a(1)"
Results:
world hello
Hello world
See also: array, array sort, array item
array_to_string, string_to_array |
Index |
string [array_to_string arrayname ?-join string? \
?-start element? ?-stop element? ?-stopstr string? ?-nocase?]
integer [string_to_array string arrayname ?-start elementnum? \
?-chars? ?-asc? ?-append? ?-stopchar char?]
[array_to_string] converts an array which may be, and is most useful with,
a byte array, to a string. See [split] for the inverse command.
[string_to_array] will convert either:
a multi-line string (default) to a series of character array elements
with a base 0 index containing each line.
or:
if '-chars' is used then will split the string into characters with one
character in each array element. When used in character mode with '-chars'
this behaviour can also be modified to store either ASCII characters
(default) or the ASCII integer value by using '-asc' option
If the array exists then unless -append is used an error will be raised to
prevent accidental overwriting. When using -append mode a starting
index value in the existing array may be specified.
[string_to_array] returns a count of the number of lines appended or the
number of characters appended if used in -chars mode.
'-stopchar <char>' will halt processing at the given character. This is
useful for trimming whitespace or other characters from the stored string
[string_to_array] can't have both -start and -append. -append will
override any -start setting and will, instead, append to the end of any
existing array. -start will assign elements from the starting subscript
value onwards
Example:
Byte-array to string conversion
array set a {
0 H
1 e
2 l
3 l
4 o
}
puts [array_to_string a]
Result:
Hello
Example:
String to byte-array conversion with append and stop filter
array set a {
0 H
1 e
2 l
3 l
4 o
5 " "
}
set s "world! 12345"
string_to_array $s a 6 -append -chars -stopchar " "
loop i 0 [array size a] {
puts "$i:\t$a($i)"
}
Result:
0: H
1: e
2: l
3: l
4: o
5:
6: w
7: o
8: r
9: l
10: d
11: !
See also: array, arrays, carray
array trim arrayVar ?maskchars? ?-left|-right?
Trims the left and/or right hand ends of all array values
Example:
# Create a ragged, CRLF-delimited string with leading/trailing spaces
option expression off
set s " The quick brown
fox jumped
over the
lazy dog"
puts $s
puts [string repeat - 30]
split $s "\r\n" -array a
array trim a
for {set i 0} {< $i 4} {++ i} { # Print in array order
puts '$a($i)'
}
Result:
The quick brown
fox jumped
over the
lazy dog
------------------------------
'The quick brown'
'fox jumped'
'over the'
'lazy dog'
See also: array, trim, string repeat, carray
integer [array unset ?::?arrayName ?::??arrayName?... ?-nocomplain? ?-const?]
integer [array unset ?::??arrayElement(X)?... ?-nocomplain?]
Delete an array and all elements or individual elements.
Returns an integer count of items successfully deleted
The dollar $ prefix must not be passed to [array unset]
Const arrays must be deleted using the -const argument
If -nocomplain is used with multiple arguments then execution will continue
with the remainder if an error is encountered
array unset arrayName
If the array is in global scope then use the '::' prefix (omitting the $)
as:
array unset ::arrayName
Array object types may be mixed, either whole array names or individual
array elements.
array unset a b(1) c d e(index) -nocomplain
It is an error to attempt to unset a non-existent or const array
variable but the -nocomplain argument may be used to ignore this behaviour
If -nocomplain is not passed then processing will halt at the first error
See also: unset, arrays, array, clear, vars, carray
array walk arrayName ?count? ? ?-integer? ?-bare?
Walk (iterate) the array hash-table (not a standard Tcl command)
If the 'count' argument is omitted the entire array is iterated
If -integer is used then an array will be output in integer order starting
at array element "0"
If -bare is used then no structural debug information is included and
only the array contents are output
--------------------------------------------------
hash_table::walk() Object:0x973290 size:500
--------------------------------------------------
Table(1) Node: 0x9d8500 chained nodes:1
0x9d8500 depth:1 key:'75' value:'WdqUPVnSLNLTMfeeEqTO'
Table(5) Node: 0x9d8a80 chained nodes:1
0x9d8a80 depth:1 key:'50' value:'gPkhMdxxowGitxhibQgd'
Table(10) Node: 0x9d8d90 chained nodes:1
0x9d8d90 depth:1 key:'76' value:'KMtJygZvZouAvRBUPVqg'
Root size:500 Root occupation:89 (17.80%) Max depth:2 Occupied nodes:100
--------------------------------------------------
[array walk] is not a standard Tcl command
See also: arrays, array, array statistics, carray
Ticol supports subscripted, associative Tcl arrays. This offers arrays with
either traditional numeric subscripts or subscripts indexed by a case-
sensitive string. The subscript may be a string-literal or a Tcl variable
name prefixed by a dollar $ sign or $:: global scope indicator to access
global variables within procedures. Numeric subscripts are still supported
but the subscript will simply be a string value which may be intermixed
with non-integer subscript values within the same array
Some languages such as PERL call associative arrays 'hashes' which
reflects the underlying hash-table data structure used for storage. Ticol
also uses dynamic hash tables to store array data.
Example
-------
set a(0) Hello # Integer subscript
set a(place) world # String subscript
puts "$a(0) $a(place)" # Prints 'Hello world'
Variable Promotion
------------------
Ticol Tcl allows the promotion of simple variables to array type using
[set]. This is not a standard Tcl feature and is provided to simplify
the working of the [static] command. No checks are made when promoting
(overwriting) a variable from a simple type to a more complex array type.
set a Hello # Set 'a' as a standard variable
puts $a
set a(0) Hi # Reset and promote to array type, discarding contents
puts $a # Will show <array> type placeholder
puts $a(0)
Result:
Hello
<array>
Hi
See: array
Multi-Dimensional Arrays (Arrays of Arrays)
-------------------------------------------
There are no multi-dimensional arrays in standard Tcl but they may be
emulated by various means. This usually involves compound subscript values
using embedded commas which, when combined, form a unique hash key e.g.
'arrayvar(x,y)' where the key value is "x,y"
These arrays allow 'tree-like' structures to be created using arrays
Such structures will typically require recursive code to iterate
1 and 2 dimensional C-like arrays are provided via the ticol_carray plugin
See the Tcl websites for a discussion
Example:
# Emulated multi-dimensional array
set a(0,1) "Item one"
set a(0,2) "Item two"
set a(1,1) "Item three"
set a(1,2) "Item four"
Ticol, however, supports enhanced, nested, arrays of arrays via nested
dollar addressing. So, ${$a(1)(0)} will reference array element 1 of a(),
subscript item 0. Array elements must exit. If the array is sparse and
not all array elements in a range exist then use [array item] to address
safely
Example:
# Will iterate any number of child arrays belonging to array 'a'
# First, init the top-level array, a
# This is as simple as naming the array in the element
# If tested, these elements will return "<array>"
# [is_array] may also be used to test, selectively
set a(0) b
set a(1) c
# Init child array, b
set b(0) $qbf
set b(1) "b element 1"
# Init child array, c
set c(0) "Hello world"
set c(1) "c element 1"
loop i 0 [array size a] {
loop j 0 [array size $a($i)] {
puts "a($a($i)($j)) -> '${$a($i)($j)}'"
}
newline
}
Interaction of array 'trees' can also be achieve using recursion as follows:
Note that Ticol uses the '@' character for [uplevel] rather than '#'
const QUOTE [chr 34] # [chr 34] could be used inline
proc delete_array_recursive {_q} {
upvar $_q q
if {is_array q} {
loop i 0 [array size q] { # Check element for child array
if {[uplevel @0 "is_array $::QUOTE$q($i)$::QUOTE"]} {
delete_array_recursive ::$q($i)
}
}
uplevel @0 "array unset $_q" # Use [uplevel 1 array unset]
}
}
# Assumes parent array 'a' in this example
delete_array_recursive a
Case (In)Sensitive Subscripts
-----------------------------
Array subscripts are case-sensitive with the exception of the $argv and
$env system arrays. Subscripts are referenced case-insensitively so that
$env(TEMP) and $env(temp) will both return the correct array element.
Large arrays cannot be efficiently traversed when addressed case-
insensitively. The internal hash table has to be traversed in a linear
fashion in such cases (See: array index)
Some Tcl commands have been extended to include case-insensitive lookups.
These lookups are far less efficient than case sensitive ones
An array can be defined as case-insensitive using [array set] with the
-nocase argument
Example:
array set a {q Hello} -nocase
puts $a(q) # Reference using lower case
puts $a(Q) # Reference using upper case
Result:
Hello # Same result for either upper or lower case
Hello
Spaces in Subscripts
--------------------
Whitespace is allowed in a subscript value but should be wrapped in double
quotes
Example:
set foo("a b c") zzz
set bar("d e f") "Hello world"
Declaration and Use
-------------------
Arrays can be declared using set only one level at a time. Multiple,
simultaneous, nested arrays may not be declared (e.g. set a(b(c(1))) hello)
You can use [array set] to initialise multiple array elements
set a(1) Hello # A traditional literal numeric subscript
puts $a(1) # Example dereference
Hello # Result
set a(one) Hello # An associative array with literal subscript
puts $a(one) # Example dereference
Hello # Result
set i 0 # Create an index variable
set a($i) Hello # An associative array with variable subscript
puts $a($i) # Example dereference
Hello # Result
array set days { # Create a const array of 7 elements
1 Monday
2 Tuesday
3 Wednesday
4 Thursday
5 Friday
6 Saturday
7 Sunday
} -const
Arrays are referenced by dollar prefix, the same as any other Tcl variable
Note that the array syntax prohibits any whitespace between the variable
name and the open parenthesis:
Correct: "$varname($index)"
Incorrect: "$varname ($index)" (looks for non-array variable '$varname')
Arrays as Command Arguments
---------------------------
Not all Ticol commands accept arrays as output arguments. Some do, such as
[setat]. Check with the section for each command
Most commands won't accept nested/complex array expressions as output
arguments (i.e. as the name of target variables to be written).
Most will accept arrays and complex array expressions as input arguments
as long as the array is resolved before the call
Sorting
-------
Associative arrays cannot be sorted internally, they are, however, rapidly
accessible by the hashed key (subscript) value so sorted indexes are
typically unnecessary.. See [array sort] for sorting workarounds
Storage and Efficiency
----------------------
All arrays are stored in a dynamic hash-table as associative arrays. Due to
this internal storage method, arrays are not contiguous in memory and thus
cannot be natively sorted. To sort an array, output it to a sorted list
You can create an array with a fixed table allocation and avoid the
performance costs of resizing by declaring to a high enough initial size.
Use array create arrayName size, e.g. array create big 10000
Arrays indexed by numeric value can be indexed in pseudo-sorted order by
use of a numeric subscript index value and a [for] loop or returned as a
list
Array storage is expanded dynamically from a minimum reserve allocation of
1 element
Any valid object which can be stored in a standard Tcl variable may be
stored in an array element
Storing Types
-------------
The empty string "" is stored in an array element as an empty Tcl list {}
Arrays of lists and other types are possible but it should be understood
that a Tcl array is not a contiguous binary array and is unsuitable for
passing forward to an external C/C++ DLL
Arrays of Lists
---------------
Since Tcl lists are strings, Tcl arrays may have a list as the array value
array set economists {
1 {HH002 "Henry Hazlitt" 002 87}
2 {WB001 "Walter Block" 072 77}
3 {PB001 "Phillip Bagus" 033 22}
4 {MR001 "Murray Rothbard" 023 19}
} -const
And this would be equally valid since the subscript key is unique
array set economists {
HH002 {1 "Henry Hazlitt" 002 87}
WB001 {2 "Walter Block" 072 77}
PB001 {3 "Phillip Bagus" 033 22}
MR001 {4 "Murray Rothbard" 023 19}
} -const
Passing Arrays to Procedures
----------------------------
Arrays can only be passed via a proc body by reference (by name) and using
a subsequent upvar command. To pass an array by value you must convert to
a list
This is by design. Arrays are intended to be passed by reference (by name)
[array item] is the most convenient way to handle the double-dereference of
unknown array names passed to a proc
Example:
array set colours {
1 red
2 green
3 blue
} -const
proc list_by_key {arrayName} {
upvar $arrayName
foreach key [lsort [array names $arrayName]] {
puts "Key: '$key [array item $arrayName $key]'"
}
}
list_by_key colours
Nested Array References
-----------------------
Nested array references are permissible. The outer variable must follow Tcl
syntax and NOT include a dollar prefix when calling [set]
set c 1 # Create a subscript variable
set b($c) hi # Create an array variable
set a($b($c)) Hello # Set a nested associative array
puts $a($b($c)) # Example nested dereference
Hello # Result
Each nested level must first be declared either using set or the [array]
command. Attempting to set an un-dereferenced intermediate array (i.e.
missing the $ symbol) will trigger an invalid argument error.
(e.g. set a(b(c))) hello) -> error in 'b(c)')
Advanced Arrays
---------------
Ticol cannot create complex arrays such as arrays of struct, arrays of
stack etc. The array variable, can however, point to the name of a
complex variable type or a standard Tcl variable which holds a list
Array Debugging
---------------
The statistics for an array are inspected using:
[array statistics arrayname]
and the array can be inspected in detail using [array walk]
See also: array, array create, array exists, array list, array size,
array sort, array statistics, array unset, array walk,
dictionaries, carray, string_to_array
integer [asc character|string]
integer [asc "character"|"string"]
The inverse of chr. Returns the ASCII (ANSI) numeric value of a string
Special chars such as []${} etc. must be escaped (See: escape sequences)
Strings with escaped characters must be wrapped in double-quotes
Examples:
puts [asc "A"] # Double-quoted argument
puts [asc {A}] # Braced argument
puts [asc A] # Raw argument
puts [asc "\["] # Special character [ must be escaped and quoted
Result:
65
65
65
91
A procedural equivalent would be something like:
undef asc -nocomplain # Undefine the built-in command
proc asc { char } {
scan $char %c value
return $value
}
puts [asc A]
puts [asc a]
Result:
65
97
See also: chr, string, left, right, mid, index
assert lvalue ?{rvalue-expression}? ?-line line? ?-v? ?-var name?\
?-crlf? ?-trace? ?-unescape? ?-debug?
Asserts an expression using a substitution variable in the expression
This expression is in the current scope context.Failure raises a
trappable error exception which can be handled by [catch] or [try/catch]
A temporary variable is created in the current scope which receives the
result of the first argument. The default name is $_ . You can change this
to any name using the ?-var name? option as long as the variable does not
already exist in the current scope. The substitution variable represents
the evaluated expression and may be used in the test expression
If you supply a line number this will be displayed in the output
No output is issued if the assertion is true unless the '-v' (verbose)
argument is specified. Failures are always displayed. Thus, [assert]
statements may be left within active code. See: NDEBUG (below)
By default, [assert] will not escape the lvalue input. You can enable
this by using the '-unescape' argument which will unescape backslash
escape sequences. Nor does [assert] subst its lvalue input so variables
which are wrapped in braces and [subst] or [eval] must be called directly
if required
The expression rvalue argument is optional must be supplied in braced
format. The default rvalue is {== $_ 1} which resolves to TRUE
If you want a CRLF emitting after the test expression has been executed
use the -crlf argument. This option should come before '-line #__LINE__'
The -trace argument will allow you to inspect the lvalue and rvalue
being compared. This can be useful when debugging unexpected results from
escaped strings for example.
After debugging is complete all [assert] statements can be disabled
or re-enabled globally by using the following Macro definition.
#define NDEBUG # Disable all [assert] commands
#undef NDEBUG # Re-enable all [assert] commands
Once the macro variable NDEBUG is defined [assert] statements are skipped
and have almost no execution overhead within a script
Float values should incorporate an appropriate level of rounding using
[expr ... round(n)] or [round n]
-debug shows the argument list as it is passed to [assert] to aid
introspection and error checking.
Resetting NDEBUG with #undef NDEBUG will have no effect on the rest of
the script. The entire script will be affected by the first instance of
#define NDEBUG which is found since this has effect only at macro
preprocess time and these events are out-of-order with sequential Tcl
command processing
It may be convenient to write [assert] as a one-line command
For more information, see: macro preprocessor
Examples
--------
# Note that the default placeholder for 'this' is $_
assert 1 {eq $_ 1}
assert 1
Result:
# PASS - Program continues without interruption (no output)
Example:
assert 1 {eq $_ 1} -line 10 -v
assert 1 -line 10 -v
Result:
# Program continues without interruption and displays confirmation
assert: PASS 'eq $_ 1'
Example:
assert 0 {eq $_ 1} -v -line #__LINE__ # Default internal variable
assert 0 -v -line #__LINE__
Result:
assert: Line 10: Assertion failed: '0' != 'eq $_ 1'
# Program execution halts
Example:
assert 1 {eq $x 1} -var x
Result:
# Test passes OK. No output
Example:
assert 1 {eq $x 1} -v -var x -line #__LINE__
assert 0 {eq $x 1} -v -var x -line #__LINE__
Result:
assert: Line 1: PASS 'eq $x 1'
assert: Line 2: Assertion failed: '0' != 'eq $x 1'
# Program execution halts
Example:
#define ndebug
# The following statement would normally fail but is now ignored
assert 1 {eq $_ 0}
Result:
# Assertion is ignored. Execution continues without assert
Example:
assert [round $pi 5] {== [round $_ 5] 3.14159} -v -line #__LINE__
Result:
assert: Line 1: PASS '== [round $_ 5] 3.14159'
See also: debugging
Older languages like C, C++, BASIC, and Java support arrays in which the
array subscript index value is an integer and where that subscript index
represents some offset into a single memory block. Tcl, in common with
many modern scripting languages such as Perl, Python or PHP supports
'associative' arrays in which the index value is an arbitrary string value
or string representation of a number. In associative arrays there is no
counted binary offset from one location in a block of memory to the next.
The underlying data structure in the case of Ticol is the "hash table"
Associative arrays held in a hash-table are inherently unordered
Associative array syntax places the subscript within parentheses as with
other languages. There is no need to declare an array before assigning.
You can assign any valid variable name as an array a long as it is not
already assigned and in use
If you intend to assign a very large array you can use [dim] to pre-
allocate the underlying hash-table to speed-up assignment later on
Example:
set name(first) "Gordon" # 'first' is a string literal
set name(last) "Bennett" # 'last' is a string literal
puts "Full name: $name(first) $name(last)"
Result:
Full name: Gordon Bennett
As with all Tcl variables, you dereference a variable by prefixing with
a dollar symbol $ but cases which refer to a variable name without
dereferencing do not require a dollar prefix e.g. puts $name(first)
Some functions, such as [unset] require the literal name of the array,
e.g. [unset name], [unset name(first)]
The system environment is provided via a pre-allocated const array 'env'
e.g.
puts $env("PATH")
Ticol command line arguments are provided via the const array 'argv' with
integer 'argc' representing the argument count.
e.g. puts $argv(0)
Double-dereferencing such as $$array(subscript) can be handled
For detailed information, see: arrays
See also: array, arrays, array find, env, argv, dim, dump, array unset
double dereference
autoexec ?on|off?
This option interacts with 'proc unknown'. It controls the passing of
unrecognised/unknown commands to the underlying operating system or the
calling of a special procedure. Autoexec is ON by default, allowing Ticol to
be used as a Windows command-shell
If autoexec is 'on' then unknown commands are passed to Windows IF and only
IF a proc called "unknown" isn't defined. C/C++ style escape characters such
as \r or \n are also disabled at the console to ensure correct operation
with Windows
If autoexec is 'off' and a proc called "unknown" is not defined, then the
command will return an error. Turning the option 'off' will re-read any INI
configuration settings for BackslashEscape for the CLI (see: help ini)
The command 'autoexec' with no arguments will return the current status
as 'on' or 'off'
This is also configured in ticol.ini [config] as AutoExecCommand=<boolean>
The autoexec option is Ticol specific and not standard Tcl.
The autoexec option does not relate to the use of an autoexec.tcl file
For more information, see: unknown
See also: autoexec.tcl, option, ini, commands, exec, spawn, cli,
debugging, macro, unknown
autoexec.tcl - The Ticol startup script
If a file called autoexec.tcl is found in the same folder as ticol.exe at
startup then this will be executed automatically unless configured
otherwise or set to do so via the command line. This is the Ticol
equivalent to the DOS/Win AUTOEXEC.BAT file.
Due to inherent risks in running a script automatically, the autoexec.tcl
file may not be obfuscated, thus preventing potentially damaging commands
from being hidden. Any obfuscated autoexec.tcl file will not be processed
autoexec.tcl will only be run from the exe path location
You may disable the execution of autoexec.tcl using either the command line
option /NA or via the ticol.ini file setting:
[config]
AutoExecScript=FALSE
An autoexec.tcl file may be used to initialise common variables or define
a set of common procedures before an interactive CLI session or running
another script
If autoexec.tcl ran successfully and returned code ok then variable
$::autoexec_ran will be set to $::true (1), otherwise $::false (0)
autoexec.tcl is never run in CGI mode. Programmers must rely on ticol.ini
or the active script for configuration
Note that autoexec.tcl may NOT be encrypted/obfuscated. This is by design
autoexec.tcl has nothing to do with the [option autoexec] setting
for the Ticol Command Line Interpreter (CLI)
See also: configuration, obfuscation, autoexec, at_exit
proc at_exit {} {?script?}
sub at_exit {?script?}
[at_exit] is an optional sub or user-defined procedure which is called when
a Tcl script exits. [at_exit] takes no arguments or than the main body
script. Any arguments passed to [at_exit] will be ignored
The script body is optional and may be empty
[at_exit] is called only via running scripts and never from the CLI. It is
not called unless the script exits the Ticol binary via control statements
or by conclusion of a script. Exiting Ticol using CTRL+BREAK or by typing
'exit' at the CLI prompt will bypass any defined [at_exit] script.
The [upvar] command is required within an [at_exit] procedure to access
variables defined elsewhere in a script as with any other procedure or sub
[at_exit] is intended for use as a cleanup routine, to close DLL and open
file handles safely, avoid memory leaks, file or database corruption etc.
[at_exit] is not called when Ticol is exited abruptly using CTRL+BREAK
Where a [stop] command is encountered within [at_exit], execution will be
returned to the CLI. Where an exit command is met within [at_exit], Ticol
will exit
Where neither is given the default behaviour will apply (exit if run from
Windows or stop and return to the CLI if run from the CLI)
Unlike C++, multiple [at_exit] routines are not provided for. There is no
need to register multiple [at_exit] routines. One routine can call all
desired shutdown code, including other procedures
Example:
undef at_exit -nocomplain
proc at_exit {} {
puts "Inside proc 'at_exit'"
upvar handle
upvar pointer
puts "Handle is $handle"
puts "Pointer is $pointer"
puts "Exiting Ticol"
return 0
}
puts "In main()"
set handle 10
set pointer 2
puts "Exiting script"
stop
Result:
In main()
Exiting script
Inside proc 'at_exit'
Handle is 10
Pointer is 2
Exiting Ticol
Example:
proc at_exit {} {
# If undefined then has no effect on the program exit
textcolor lightred
puts "* Stopped *"
textcolor
# exit # Exit to O/S if desired
# stop # Return to CLI if desired
return 23 # Return ERRORLEVEL value to the O/S
# Else exit to original start environment
}
Example:
ticol.exe "proc at_exit {} {puts Done}"; test.tcl
Example:
ticol.exe "sub at_exit {puts Done}"; test.tcl
See also: on_error, upvar, run, cli, exit, stop, return, autoexec.tcl
proc on_error {} {?script?}
sub on_error {?script?}
[on_error] is an optional sub or user-defined procedure which is called when
a Tcl script encounters a non-caught trappable error. [on_error] takes no
arguments or than the main body script. [on_error] is not called where
errors are caught or trapped by [catch] or [try]..[catch].
Any arguments passed to [on_error] will be ignored
The script body is optional and may be empty
[on_error] may be used to cleanup or take remedial action in the event of an
error occurring such as unloading a Ticol plugin module or logging an event.
Errors generates within the [on_error] script are blocked from recursively
generating futher errors. The preferable course of action is to perform any
necessary cleanup then call [die]
Note that variables such as errorMsg are global and should be addressed
within the [on_error] script via the '::' scope prefix.
See also: error, at_exit, die, sub, proc
double [atof <scientific-number-format-string>]
string [ftoa value]
Converts a string containing a number in "scientific notation" format
e.g. "2e+10" into a double value. This will convert only within the range
of an 8 byte double representation
Scientific notation numbers are not implemented within [expr] in order
to increase efficiency of the interpreter and since their use is likely
to be quite low.
This is a temporary/intermediate fix for the lack of "scientific" number
handling within Ticol. [atof] may be embedded within [expr] expressions
The inverse of [atof] is [ftoa] which will convert a floating point number
in non "scientific" notation back to such. [format %e] may also be used to
convert from standard to "scientific notation" format
Be aware that double precision numbers may not precisely represent certain
values
Similar functionality is provided by the ticol_sci plugin via [sci] and [dec]
Examples:
puts [atof 2e10]
puts [atof 2e-5]
puts [atof 32e+4]
puts [ftoa 320000.000000]
puts [atof 32.345e-4]
puts [atof 0.001234e+6]
set q "0.001234e+6"
puts [expr [atof $q]*$pi]
puts [ftoa [expr [atof $q]*$pi]]
# Division and cast using scientific notation
set a 2.2e1
set b 7
puts [/ [double [atof $a]] $b]
Results:
20000000000.000000
0.000020
320000.000000
3.200000e+005
0.003235
1234.000000
3876.725334529804500
3.876725e+003
3.142857142857143
https://en.wikipedia.org/wiki/Scientific_notation
See also: sci, dec, math
string [base64 string -encode | -decode ?-wrap width?]
string [base64 varName -var -encode | -decode ?-wrap width?]
Base64 encode or decode a string or variable according to the argument
parameter given. -var will specify that a variable rather than string
constant has been passed.
Where a -wrap value is given this will control line-wrapping on the
returned base64 code. The default is no line-wrapping (0)
The minimum -wrap width is 4 characters
Example:
set a [base64 "Hello world!" -encode]
puts $a
puts [base64 $a -decode]
Result:
SGVsbG8gd29ybGQh
Hello world!
Example:
puts [base64 $qbf -encode]
newline
puts [base64 $qbf -encode -wrap 10]
Result:
VGhlIHF1aWNrIGJyb3duIGZveCBqdW1wcyBvdmVyIHRoZSBsYXp5IGRvZw==
VGhlIHF1aWNr
IGJyb3duIGZv
eCBqdW1wcyBv
dmVyIHRoZSBs
YXp5IGRvZw==
See also: encrypt, decrypt, mkn, cvn
Tips for Writing Big Programs |
Index |
Ticol Tcl is intended for small scripts relating to administrative use.
It was written by the author to be used for general housekeeping tasks
as a simple alternative to PowerShell and batch script
Here are some tips for developing larger programs:
Make full use of [proc]
Divide large scripts into many small procedures except where
performance is critical and inline code required
Use file-level modularity
Break big scripts into smaller files and call using [source scriptname]
Be aware that each call to [source scriptname] will evoke the macro
preprocessor (MPP) which has a time cost
Arguments to external files are permissible via [source scriptname]
Use the Macro Pre Processor
For conditional execution, where appropriate
Make full use of plugin libraries with [lib]
A range of functionality is available via Ticol plugin DLLs
See also: performance, faq, eval
Currently, only one subcommand implemented [bit invert]
integer [bit invert value ?-8|16|32|64?]
Bitwise invert a value, optionally according to a given mask width
The inversion mirrors or flips the bit pattern left-to-right
0x0001 -> 0x1000
The inversion is performed against the specified mask-width with
a mask of 64 being the default
Example:
puts [inttobin [bit invert 1]]
puts [inttobin [bit invert 1 -8]]
puts [bit invert 0x0180 -16]
puts [inttobin [bit invert 0x0180 -16]]
puts [inttobin [bit invert 0b1100000000 -16]]
Result:
0b1000000000000000000000000000000000000000000000000000000000000000
0b10000000
384
0b110000000 # 384 decimal
0b110000000 # 384 decimal
See also: << >> ~
calc - Auto-optimising alternative to [expr] |
Index |
calc <expression> ?-explain?
calc {<expression>} ?-explain?
calc "<expression>" ?-explain?
calc <arg> <arg...> ?-explain?
Inline Tcl commands are faster and more efficient than using [expr].
The Macro PreProcessor will expand [calc] natural 'infix' expressions
into native Tcl commands using Polish 'prefix'-style notation
[calc] is identical to [expr] and may be used in place of [expr].
It accepts the same expression format argument(s). However, unless the
command-line argument /NEO (No Expression Optimisation) or the
[option preprocessor off] directive is used, then the expression will
be expanded and rewritten by the Macro PreProcessor (MPP) into native Tcl
format commands before running. If /NEO is used then [calc] will be
interpreted exactly as for [expr]
The '-explain' argument is should be used only at the Ticol CLI rather
than load/macro-preprocess runtime. It must also be the last argument in
any test expression. Since constants are optimised-out, the resulting
expression may be a single value. A separate command, [explain] can be
used to view a fully-expanded expression.
Example:
# Example using -explain from the CLI
puts [calc 1+2/3.0-4*4]
puts [calc 1+2/3.0-4*4 -explain]
puts [explain 1+2/3.0-4*4]
Result:
-14.333333333333332
[- [+ 1 [/ 2 3.0]] [* 4 4]]
[- [+ 1 [/ 2 3.0]] [* 4 4]]
[calc] supports the same range of math operators as [expr]
(excludes ++ and --)
[calc] requires a bracketed declaration - Use [calc ...] not calc ...
The translated command strings are typically around 50% faster than [expr]
as [expr] has an error and syntax-checking overhead as well finally calling
[eval] with the results. [calc] hugely speeds up performance by
pre-processing this task before run time
Ticol expressions are BEDMAS-optimised unless brackets are used to enforce
ordering. Function calls are resolved to the [funct] command.
If the command-line argument /NEO is used then [calc] will behave in exactly
the same way as [expr]; This makes [calc] much more flexible as well as
significantly improving perform performance on math tasks. In addition,
this enables the '-explain' argument. If problems are encountered then use
/NEO to disable [calc] translation and it will be interpreted as [expr]
If the result of [calc] is a string value and if [option expression] is
set to OFF and this value is passed to [eval] then the result may be
misinterpreted as a command. The fix is to ensure that you use
[option expression on] when handling string returns when chaining the
result to [eval]
You may examine the translated result using the /ECHO command line argument
if manual checks are required on the translated expression
Example:
set f 3
puts [calc {$f+2-($f==2)}]
Translation (requires ticol.exe /ECHO to output):
puts [- [+ $f 2] [== $f 2]]
Result:
5
Example:
option PreProcessor off # Turn off MPP for the command line
set s [calc 1/2.0+3*4 -explain] # Get the translated Tcl code
puts $s # Display the Tcl code
puts [eval $s] # Evaluate it and get the result
option PreProcessor on # Turn the MPP back on again
Result: (Expanded Tcl commands)
[+ [/ 1 2.0] [* 3 4]]
12.5
[calc] will also attempt to simplify constant-only expressions by
evaluation and then substitution. Disable using the /NEO command argument
Example:
puts [calc {[rnd 2 4] > 3 ? {return Higher} : {return "Same/lower"}}]
Interpreted as:
puts [? [> [rnd 2 4] 3] {return Higher} {return "Same or Lower"}]
Result:
Either Higher or lower
Example: (save as script.tcl)
# script.tcl
set s [calc "1 / 2.0 + 3 * 4"]
puts $s
Run as: ticol.exe script.tcl /echo /neo
set s [calc "1 / 2.0 + 3 * 4"]
puts $s
Run as: ticol.exe script.tcl /echo
set s [12.5]
puts $s
ticol.exe script.tcl
Note: (/NEO command-line argument)
The /NEO (No Expression Optimisation) argument also applies when creating
protected TCX files. Any TCX sub-file which needs to be run using the
/NEO argument must be obfuscated using /NEO
Note: (Excess {} braces)
Do not wrap the entire expression to be passed to [calc] in braces or
[calc] will be unable to interpret it. Braces prevent evaluation.
Spurious error messages will result.
Example: (WRONG - expression is 'masked' by braces)
puts "a and b are: '[calc {{$a eq $b} ? {return equal}\
: {return different}}]"
Example: (CORRECT)
puts "a and b are: '[calc {$a eq $b} ? {return equal}\
: {return different}]"
Note: (/NES argument)
The /NES (No Expression Simplification) will disable automatic
simplification and optimisation of expressions generated by [calc]
Note that where [calc] const expressions are pre-evaluated and generate
an error then this error will always be fatal and the preprocessing phase
will halt. This feature can be disabled using /NES on the command-line
Note also, that constants or variables with embedded # characters must
be enclosed in double-quotes
Example:
puts [calc acos(10)] # Will evaluate to -1.#IND00000000000 (invalid)
Note that the literal hash (#) comment character can be used safely
within a quoted string
See also: explain, expr, Macro PreProcessor, /NEO
Handling Byte Arrays |
Index |
Integer-indexed arrays may be used to store byte arrays as integer values
These may be converted to string using the following procedure:
Buffered [store] is used for rapid I/O. No pad characters are added.
Any zero (NULL) character encountered will truncate the string
proc array_to_string_asc {arr} {
#################################################
# Convert an integer byte-array to ASCII string
#################################################
upvar $arr a
set j [array size a]
for {set i 0} {< $i $j} {++ i} {
store add [chr $a($i)]
}
return [store get]
}
set a(0) 72
set a(1) 101
set a(2) 108
set a(3) 108
set a(4) 111
puts [array_to_string_asc a]
Result:
Hello
See also: arrays, array_to_string, string_to_array
Ticol Tcl procs can perform 'callbacks' to other Tcl procs where the
name of a callback proc is passes via the initial command reference
Example:
option expression off
proc test_callback {cbp} {
puts "Enter test_callback -> '$cbp'"
for {set j 0} {[< $j 10000]} {++ j} {
if {[&& [> $j 0] [is_mod $j 1000]]} {
$cbp $j # Call the callback proc
}
sleep 1
}
}
proc callback {x} {
puts "In callback $x"
}
test_callback callback
Results:
Enter test_callback -> 'callback'
In callback 1000
In callback 2000
In callback 3000
In callback 4000
...
See also: proc
signed char [char value] # 8 bit signed
unsigned char [uchar value] # 8 bit unsigned
signed short [short value] # 16 bit signed
unsigned ushort [ushort value] # 16 bit unsigned
signed long [long value] # 32 bit signed
unsigned long [ulong value] # 32 bit unsigned
signed integer [int value] # 64 bit signed
signed integer [signed value] # 64 bit signed
unsigned integer [uint value] # 64 bit unsigned
unsigned integer [unsigned value] # 64 bit unsigned
double [double value] # 8 byte FP
Tcl is effectively a typeless language. It handles all variables internally
as string values and, in most cases these are converted by Ticol to 64-bit
integers or 8-byte floating-point doubles when necessary for arithmetic
purposes
All Ticol integers are stored as 64-bit signed integers, internally
(See: 'data type ranges' for data range values)
Conversion and type detection is automatic. There may be cases, such as
marshalling data for DLL interfaces, or certain math routines, where you
may wish to coerce the type manually. The built-in Ticol casts are more
efficient than casts written in Tcl script
Non-decimal constants such as 0x1234 or 0o1234 are usually translated into
decimal format before execution by the Macro PreProcessor (MPP). If the
MPP is disabled using the /NP command-line option then translation is done
on-the-fly.
Binary conversion [bintoint], [inttobin] converts to unsigned integer
The work of unnecessary runtime conversion is therefore avoided by the MPP,
but this might cause unexpected side-effects if a script relies on the
original notation. If in any doubt, wrap such values in double-quotes,
this will prevent MPP translation
You may use the functions above to cast the string values to an appropriate
type. This may involve the truncation of the value with loss of precision.
[char] differs from [chr] in that [char] truncates any value to a valid
char integer range whereas [chr] accepts only a fixed range of argument
values (0..255) outside of which an error will be generated
[signed] and [unsigned] will be coerced from 64-bit signed or unsigned
integers with consequent 'wrap around'. Overflows will encounter truncation
e.g. unsigned short values over 65535 will be truncated at 0xFFFF and the
remainder will 'wrap around'
Appending ".0" to an integer will have the same math effect as a [double]
cast
Example:
# Same effect from both concatenation of ".0" and [double] cast
[12345].0 -> 12345.0
[double 12345] -> 12345.000000000000000
Numeric type conversions are built-in to Ticol. A cast will accept any kind
of numeric input, decimal integer, decimal real, hexadecimal, octal or
binary
You will still have to specify byte widths when marshalling data for export
via a DLL interface via [calldll*]
Note again, that non-decimal number types such as hex (0xNN) which are not
protected within quoted strings will be converted to decimal by the Macro
Pre-Processor (MPP) before any script is run. This is designed to speed up
execution
Example: Result Comments
--------------------------------------------------------------------------
puts [int $pi] 3 Cast a constant
puts [char 240] -16 Cast a char
puts [uchar 12345] 57 Masked to 0xFF
puts [chr 12345] Overflow Invalid for [chr]
puts [char 0b1111000] 120 Cast binary (0bX)
puts [uchar 0b1111000] 120 Cast binary (0bX)
puts [double 0b111100011011] 3867.000000000000000 Cast binary (0bX)
puts [short 0o123] 83 Cast octal (0oX)
puts [short 0x123ffff] -1 Truncated at 0x7FFF
puts [ushort 0x123ffff] 65535 Truncated at 0xFFFF
puts [long 44073709551615] -1244839937 Mask 0x7FFFFFFF
puts [ulong 4294967297] 3050127359 Mask 0xFFFFFFFF
puts [int 0x123ffff] 19136511 Valid integer
puts [double 0x123ffff] 19136511.000000000000 Valid double
puts -1 -1 64-bit integer
puts [unsigned -1] 18446744073709551615 64-bit integer
puts [signed 18446744073709551615] -1 64-bit integer
puts [signed 0xFFFFFFFFFFFFFFFF] -1 64-bit integer
Code Example:
option expression off
for {set i 0} {< $i 64} {incr i} {
puts "1 << $i == [unsigned [<< 1 $i]]"
}
See also: data type ranges, fpart, integral, asc, calldll, chr, error
# With option expression off
errorcode [catch {sourcecode} resultVariable ?-quiet?]
# With option expression on
errorcode [catch [sourcecode] resultVariable ?-quiet?]
[catch] intercepts error returns from the source code expression which is
passed to it and prevents the script from halting, instead returning a
boolean value. If there is an error condition then 'variable' is set to the
error message. [catch] returns a success code and the intercepted error
code as a return value.
The '-quiet' argument will suppress local error messages and the caller
must rely on using other methods such as $errorMsg to interpret failures
The [catch] command operates at the level of Tcl error return values and
does not provide any additional O/S CPU exception handling.
[catch] returns 0 on SUCCESS and a non-zero value on error
The non-zero value is a Tcl error code. See: error code
The [catch] expression must be wrapped in braces and, if a command, the
command must not be wrapped in square brackets or catch will evaluate the
command return rather than the command itself and may fail. This behaviour
is dependent on the [option expression] setting
If the expression is successful then 'resultVariable' is set to the return
value of the expression. It is important that the catch expression is
wrapped in curved braces not square brackets as square bracketed
expressions will be evaluated before the catch test
Don't wrap an assignment within a [catch] statement unless you want to
test the assignment itself. For example...
# Where 'nosuchvar' does not exist, s is not set
set s Hello
catch {set s [string left $nosuchvar 1]}
assert $s {ne $_ Hello} -v # This will fail to clear s
# Use instead...
option expression off
catch {string left $nosuchvar 1} s
Effective use of [catch] requires 'option echo off' as [catch] will handle
the display of error messages rather than direct console output
Within flow-control statements with: [option expression on] catch must be
wrapped in square brackets to be evaluated otherwise it will be treated as
an expression. Catch is most useful when combined with [if] (see below)
See: error values for a list of returned error codes
Example:
option expression off
if {[catch {set h1 [xbase open user.dbf]}]} {
die "Failed to open database address.dbf"
}
Example:
option echo off
set file "nosuchfile.txt"
# Note [catch] is wrapped in square brackets for [if]
if { [catch { open $file r } err] } { # Any value > 0
puts stderr "Could not open $file for reading\n$err"
exit 1
}
Result:
Could not open nosuch.txt for reading
Exception raised from catch. Error code 1
Example:
option echo off
stack create s 1000
stack unset s
catch {puts "Stack empty? [bool [stack is_empty s]]" } err
puts stderr "Stack error\n$err"
exit 1
Result:
Stack error
stack is_empty: Stack 's' not initialised
Example:
set line #__LINE__
if {[! [catch
{puts 'pwd: [pwd]'} # Must be wrapped in braces
]]} {
die "Failed to trap bad puts command at line: $line"
}
Result:
# No error occurs
Global variables are also set by internal error-handlers as follows:
errorCode Tcl error code returned to the CLI
errorLine May be set by the script using #__LINE__
errorMsg Console message generated by an internal error-handler
See also: try, cmdcount, trace, error, run, time, debugging, flow control
stop, exit, errorCode, errorLine, errorMsg, errorInfo,
error values
string [chain varname ?defaultvalue? ?-depth N?]
[chain] will follow a chained set of standard variable references until
the terminating case is met. It performs the equivalent of multiple $$
or [set] dereferences and follows a chain until the terminal case is found
[chain] handles any $-resolvable variable including arrays and struct
elements
The command is useful if the dereference depth is unknown, thus precluding
the use of a fixed number of $ or [set] commands
The command accepts an optional default value. If specified, this will be
returned if the chain of references fails.
The optional '-depth N' argument can be used to restrict the search depth
A value of -depth 0 will perform a full search, 1 will search to the 1st
level below, 2 the 2nd and so on...
[chain] won't know if a terminal case string is a reference to a broken
variable name or a string literal
Example:
set a Hello # Terminal node
set b(23) a # b(23) points to a
set c b(23) # c points to b(23)
puts [chain c] # Chain through: c->b(23)->a->Hello
unset a # Break the chained reference
puts [chain c]
puts [chain foobar {Not found}] # No such reference, 'foobar'
catch {
puts [chain foobar] # Will raise a catchable error
}
Results:
Hello # The chain is fully-dereferenced
a # Chain search halts at 'a'
Not found # A default value is returned
<Error is raised> # A trappable error
See also: double dereference, dereference, item, try, catch
integer [ceil double-value]
Returns a double, rounded up to the nearest integer value
Example:
puts [ceil 7.6]
puts [ceil 3.2]
Results:
8
4
See also: floor, math, integer, double
integer [clear ?-list?]
Clear (unset) all user-level variables and consts
System-level consts such as $env are unaffected.
Although consts cannot be changed during the life of a script, the entire
variable table may be cleared before running a new script. This prevents
conflicts where scripts are loaded from the CLI.
By default, [clear] returns the count of cleared variables
Note that certain variables such as error* are recreated each
time the Command Line Interface reloads
Adding the '-list' argument will return a list of cleared variables
instead of a count of variables cleared
Ticol also maintains a small number of reserved system variables which
are protected from deletion by [clear] or [run]
The [run] command will also clear the user-level variable table
The loaded script may be cleared using the command [load -clear]
An alternative is to use:
unset * -glob -nocomplain # -const is required to delete consts
Example:
puts [clear -list]
Result:
errorMsg errorLine errorInfo errorCode
See also: clearprocs, reserved variables, load, unset, set, vars
integer [charcount string mask-chars ?-nocase?]
Count the number of instances of each character in the mask in the
input string.
Case may be ignored using the -nocase argument
Where $qbf is a standard test variable:
"The quick brown fox jumps over the lazy dog" ...
Each successive mask character is counted only once. i.e. the string is
iterated against the mask and each string char is matched only once
Thus [charcount $qbf Tt] and [charcount $qbf TTtt] both return 2
This applies also to -nocase where both [charcount $qbf Tt -nocase]
and [charcount $qbf t -nocase] return 2
[charcount] is useful for checking the number of backslash escapes and
estimating a shrinkage factor when a string is unescaped
Escaped backslash characters - "\\" are counted as one character,
otherwise each backslash is counted as one character and any non-backslash
mask character which follows is counted as one character
Example Result
puts [charcount $qbf a] # 1
puts [charcount $qbf r] # 2
puts [charcount $qbf t] # 1
puts [charcount $qbf tT] # 2
puts [charcount $qbf t -nocase] # 2
puts [charcount $qbf tT -nocase] # 2
puts [charcount $qbf aeiou] # 11
puts [charcount $qbf fox] # 6 x o and f
puts [charcount "\[puts \"hello\"\]" "\\"] # 4 x '\\'
puts [charcount "c:\\windows\\system32\\" "\\"] # 3 x '\\'
puts [charcount "c:\\windows\\system32\\" "ws\\"] # 8 x w, s and \\
See also: chartok, char, string
integer [chartok string charmask array ?-nocase?]
[chartok] 'tokenises' (breaks apart) a string as delimited by characters
specified in the 'charmask' argument and returns the individual components
in a base 0 integer-indexed array. A count of the number of tokenised items
in the array is returned
The mask argument may comprise of one or more characters by which the
string will be broken-up
The mask characters are not included in the output. Typically a string will
be split by spaces, tabs or CRLF pairs
If the 'array' argument exists it will be overwritten without warning
If the -nocase option is used then case will be ignored when comparing the
'charmask' argument.
Example:
option expression off
set count [chartok $qbf " " a]
for {set i 0} {< $i $count} {++ i} {puts "$i $a($i)"}
Result:
0 The
1 quick
2 brown
3 fox
4 jumps
5 over
6 the
7 lazy
8 dog
Example:
set s "Live-long and+prosper"
set count [chartok $s "+- " a]
for {set i 0} {< $i $count} {++ i} {puts $a($i)}
Result:
Live
long
and
prosper
See also: array, arrays, charcount, split, strto, strtor
integer [clearprocs ?filter? ?-nocomplain?]
Clear all user-defined procedures from the current workspace
If a 'glob' name filter is not given then * is assumed (all procs)
User procedures are all defined at scope level 0 (root/global)
[clearprocs] returns a count of the number of procs which have been
successfully erased. This may be used to cross-check with a list produced
by [procs]
[clearprocs] will terminate any [after] script which is proc dependent
Example:
option echo on
procs
puts [clearprocs]
Result:
cube rainbow square testmean
4
See also: clear, glob
Command Line Arguments |
Index |
Syntax: ticol.exe <scriptname>?.tcl? ?/argument...?
Syntax: ticol.exe ; "tcl script commands"
The following command line switch options are accepted:
/ECHO Echo preprocessed source code and exit
Invalid for encrypted files
/EXPR:<expression> Evaluate a Tcl expression (calls [expr expression])
Expression results can be piped back to scripts etc.
/CRLF Issue a CRLF after /EXPR (default is OFF)
/LIB:libname?.dll? Load a Ticol plugin library DLL at startup
/C?:OPT? ?outfile? Protect (encrypt) a Ticol source file
where OPT is one of the following options:
<No option> Automatic encryption/decryption
USER Lock to a Windows user name
VERSION Lock to a Ticol version
MAC Lock to a network card address
WINVER Lock to a Windows version x.xx
HOST Lock to a Windows host name
IP Lock to a valid IP address
PASSWORD Require a user-specified password
Decryption is automatic except for /C:password
If <outfile> is omitted then infile.tcx is created
/64 Base64 encode TCX output produced with /C (above)
/PW:<password> A password phrase relevant to the above /C cases
This may include any field including host or IP address
which will be used to encrypt the script
The MD5 value is displayed on successful encryption
/I Info. Show encryption information for a TCX file
/LOG Enable startup logging to ticol.log
/MD5 filename Display the MD5 value of a file (or: /MD5:filename)
/NA No autoexec.tcl Don't load/execute autoexec.tcl
/NB No console I/O buffering
/NE Disable the exit command, disable CTRL+BREAK
/NP No Macro PreProcessor
/NOESC No MPP processing of escape sequences
/NEO No Expression Optimisation of [calc] during preproc
/NES No Expression Simplification of [calc] during preproc
/NX No CLI autoexecute unknown commands (autoexec off)
/NMX No MPP #exec allowed (see also INI->PreprocessorExec)
/SC Enforce Tcl standard comments
/G Show performance analysis graph at end of run
/PAUSE Pause after execution (keep console window open)
/BP Enable breakpoints and single-step debugging
This also enables the [halt] debugging command and
stack trace
/ST Enable Stack Trace only. Will be displayed only if
an exception occurs. Not valid for plugin DLL libs
/DEBUG Enter immediate single-step debugging mode
Enables breakpoints (/BP), stacktraces and issues a
[halt] instruction to start debugging immediately on
running a script
/Q Quiet mode. Don't show load/run/PreProcessor errors
/VER Echo the version
Use: ticol.exe /? to get details of command line arguments for the current
version of Ticol
Note that under some versions of Windows, the maximum command line length
is 8191 characters (8Kb-1)
Expressions are accepted if the first argument is specified as a semi-
colon. Where this format is used, the entire expression must be wrapped
at the command-line level, in double-quotes otherwise the following error
may be encountered:
Error: 1 too many [ brackets found
Braces may be used within Windows CMD line arguments to delineate Tcl
strings
Examples:
ticol.exe ; "puts [{hello world}]"
ticol.exe ; "puts [cd {\\\\server\\share\\path}]"
See also: argc, argv, command line interpreter, CLI
Command Line Interpreter (CLI) |
Index |
void [cli]
This is an interactive shell within which Tcl commands may be directly
executed. From here you can also run programs or interact with the O/S
shell. The command [cli] launches another instance in the same thread
and is launched automatically on running ticol.exe without a script
name. The [exit] command will exit any [cli] instance.
The CLI is a single-line editor with a maximum line length of 4096
characters. Multi-line freeform editing is not supported but multiple
commands may be separated by a semi-colon.
Command history is retrieved using the up/down cursor keys control the
command history both at the CLI and within most text input routines
Macro-preprocessing such as #ifdef...#endif won't apply to
commands entered directly within the CLI. The [at_exit] command also not
enabled from the CLI
Scripts may be run using: run script?.tcl? or evaluated using: eval script
The CLI returns the status error code in square brackets, followed by any
applicable return value if 'option echo' is set to ON
[x:status] <return-value>
The primary Ticol status codes are:
0 OK/Success
1 Error condition. The [error] command may give more information
2 [return] statement encountered
3 [break] statement encountered
4 [continue] statement encountered
5 [exit] statement encountered
6 [stop] or abort condition (stop command, wrong arguments etc.)
7 [goto] command encountered
8 A command or proc is already defined
9 [gosub] statement encountered
Commands results are not echoed by default. To enable command console echo
use the command: option echo on
- or configure the ticol.ini file and add an entry:
[Config]
ConsoleEcho=TRUE
Example:
set a "Hello"
set $a
Result (at the CLI):
[0:ok] Hello
Security
--------
For obvious security reasons, 3rd party users should not be given access to
the Ticol CLI e.g. via CGI, remote command scripting or other mechanisms.
Direct CLI scripting (direct argument interpretation) is blocked in CGI
mode; e.g. /EXPR:<string> and the [cli] command which provides a command-
line interpreter...
http://127.0.0.1:8800/cgi-bin/ticol.exe?/expr%2022/7.0
The [cli] command should be undefined in all secure scenarios using the
ticol.ini -> CommandUndefList="command..."
See also: security considerations, ticol.ini, command line arguments,
command line editing, command line history, exec, autoexec,
debugging, macro
Command Line History |
Index |
The Ticol CLI stores input history. You can retrieve this using the up
and down cursor arrow keys.
See also: command line editing, command line editing
clipboard get
clipboard set string ?-append?
clipboard clear
clipboard isempty
clipboard printscreen
clipboard gettext
Copy text to or from the clipboard. Optionally, clear the clipboard or check
to see if it is empty. clipboard printscreen will capture the current window
to the clipboard as an image
[clipboard get] will retrieve ANSI text from a selected area in the console
window. You can append to the clipboard using the -append argument
[clipboard set] will place ANSI text into the clipboard
[clipboard clear] empty the clipboard.
It is recommended that you test for a non-empty clipboard and clear it prior
to copying text to it. Note that the clipboard supports multiple data formats
[clipboard printscreen] will grab the current visible console window as a
graphic image equivalent to a PrtScrn operation
[clipboard gettext] will retrieve 80 x 25 characters of ANSI text from
cursor location 1,1 onwards in the current console. If the screen is scrolled
after a [cls] command then only the first 2,000 ANSI characters will be
returned. Select-zones are not supported by this command
Other than clipboard printscreen, only text mode operations are supported
See also: Single line commands, Ticol
Command Line Prompt String |
Index |
void [prompt {braced-argument-list} ?-reset?]
Configures the Ticol CLI prompt. The option '-reset' may be used to reset
to the internal default of "$u@$h$s$k$g". The argument list must be
enclosed in braces. Any tokens which are not recognised as prompt symbols
will be treated as string literals.
Example:
prompt {Hello $u it is $d at $t $g}
prompt -reset
Result:
Hello Admin it is 19-11-2021 at 12:49:38 >
Admin@Snoopy>
The command line prompt can also be configured via the TICOL.INI file using
a subset of the DOS/Windows PROMPT characters as follows
[Config]
Prompt=<string>
Where the default internal value is Prompt=$u@$h$g
Which will return: user@host>
Example:
[Config]
Prompt=$u@$h$s$k$g ; Show user@hostname,space,command-count
Result:
Admin@Snoopy 132>
Where: the string contains literal text and any of the following tokens:
$c An opening curved parenthesis (
$g A greater-than symbol >
$h The current hostname (if any)
$f A closing curved parenthesis )
$k Command count (See: info cmdcount)
$l A less-than symbol <
$p The current path
$n The current drive letter only, e.g. C
$s A single space character
$u The current username (if any)
$_ A newline
See also: cli, ini file, info cmdcount
The following colour constants are defined. You may also use numeric values:
0 0x00 black
1 0x01 darkblue
2 0x02 darkgreen
3 0x03 darkcyan
4 0x04 darkred
5 0x05 darkmagenta
6 0x06 darkyellow (orange)
7 0x07 grey (gray)
8 0x08 darkgrey (darkgray)
9 0x09 blue
10 0x0a green
11 0x0b cyan
12 0x0c red
13 0x0d magenta
14 0x0e yellow
15 0x0f white
Both 'grey' (UK) and 'gray' (US) spellings are accepted
See the example file: colors.tcl
Most modern versions of Windows will display "darkyellow" as a brown/khaki
colour where orange would be more useful. You can change darkyellow to
orange using the following REG file. Save as "console-orange.reg" and
merge into the current session then log out and back in again to enable
More recent versions of Windows will reset this colour at each boot
---------------------------------------------------------------------
Windows Registry Editor Version 5.00
; Colour table with DarkYellow == Orange 0x00c0c0c0 (BGR)
[HKEY_CURRENT_USER\Console]
"ColorTable06"=dword:00508dfc
---------------------------------------------------------------------
See also: textcolor, textbackground
Conditional Operator - ?: |
Index |
Command syntax:
bool [? {test-expression} {true-expression} ?:? {false-expression}]
or: (using expr syntax):
bool ( test-expression ? true-expression : false-expression )
Ternary (3 argument) or 'functional' if
The first operand is evaluated and all side effects such as expression calls
are completed before continuing to evaluate
If the first operand evaluates to true (a nonzero value), the second operand
is evaluated. If the first operand evaluates to false (0), the third operand
is evaluated. The result of the conditional operator is the result of
whichever operand is evaluated the second or the third.
Whichever of the operand expressions applies will be evaluated using 'expr'
and this may include other Tcl commands. To return literal strings use
'return string'
Only one of the last two operands is evaluated in a conditional expression
Note: Nested functional if commands should be avoided as these can lead
to difficult-to-debug problems, bracket parity issues and possible
stack overflow conditions in recursive code. Where a nested functional if
must be used it is strongly recommended that each level is bracketed
Arguments for the command version should be properly braced, particularly
if the expressions contain commands wrapped in square brackets
The functional if may be faster than an if command for simple expressions
To return a string or other literal from the command form [?], use the
[return] command. This is not necessary in the expression form
The expression form may be deprecated in future versions of Ticol as the
command form is easy to include within expressions
Example:
puts [expr "rnd(1,2) ==1 ? One : Two"]
Example:
set i 0; while {[< $i 5]} {puts [? [rnd 0 1] {return true}\
{return false}]; incr i}
Result:
false
true
true
false
true
Example:
# Command form [?] requires []
puts "[? [rnd 0 1] {return true} {return false}]"
Result:
(true or false)
Command Version of [?]
----------------------
Infix functional [expr ?] if expressions may be rewritten using the
command version [? arg true-expr | false-expr]
proc factorial {n} {? $n<2 1 : {? $n>=20 0 : {expr $n * \
[factorial [-- n]]}}}
[?] Sub Expressions and Nesting Caveats
---------------------------------------
Infix [expr ?] sub expression should call [expr] separately to evaluate
sub expression. Calls to [expr] must have braced arguments for [expr]
# Contains 3 levels of call to [expr], each with braced arguments
# restrict to factorial 20 to avoid 64-bit integer overflow
proc factorial n {expr {$n<2 ? 1: expr {$n>20 ? $n*($n-1) :\
expr {$n * [factorial [-- n]]}}}}
See also: flow control, if, else, while, for, foreach
Conversion Routines - mkn, cvn |
Index |
string [mkn <numeric-value>] # Make a compressed number
integer [cvn <mkn-string>] # Reconvert a compressed number
These commands are similar to Visual BASIC's Mki/Cvi, Mkf, Cvf functions
[mkn] converts a decimal numeric string of any length into a compressed
4-bit encoding format which is universal to all decimal-based number types
Large amounts of numeric data may be compressed or numeric data may be
efficiently compressed for storage in say a database field
This format does not embed null characters within the string and encodes a
precise string representation rather than a binary translation. It encodes
to precisely half the original string length
The size of the numeric string is limited only by available memory
The encoded character ranges is: "0-9 + - . <space>"
Encoding pattern is as follows:
------------------------------------------------------------------
ASCII Value Input
Encoding Symbol
Encodes
Comment
------------------------------------------------------------------
48..57 0x1..0xa '0..9' Decimal digit
43 0xb 11 '+' Positive sign
44 0xc 12 ',' Comma (thousands separator)
45 0xd 13 '-' Negative sign
46 0xe 14 '.' Decimal point
32 0xf 15 ' ' Space
0 0x0 '\0' Reserved for C/C++ NULL terminator
------------------------------------------------------------------
Examples:
puts [mkn 616]
puts [mkn $pi]
puts [mkn +3.14159]
puts [cvn [mkn 3452345]]
puts [cvn [mkn +3.14159]]
Results (may include extended characters not reproducible here):
rp
N%&úvFÜèC@
¦ÔRj
3452345
+3.14159
Using [mkn], [cvn] to store float/double values in struct fields
----------------------------------------------------------------
Since [struct setb] stores integers only you must either store
floats in string format or, if space is at a premium, you can encode
them using [mkn] to create a compressed floating point number
representation which can be stored and then retrieved using [cvn]
Example:
struct s {a 16 b 16}
struct setb s.a 1234.5678] # Will truncate and store integer
struct set s.b [mkn 1234.5678] # Will store full float value
puts "s.a is: '[ofaddressb s.a]'"
puts "s.b is: '$s.b' -> '[cvn $s.b]'"
Results:
1234
1234.5678
Struct dump:
struct 's' level 0 (0x1f389f0) data @32737744 (0x1f389d0)
| 32 byte(s), 2 member(s)
+ 1:'->s.a' => 32737744 (0x1f389d0) 16 byte(s)
+ 2:'->s.b' => 32737760 (0x1f389e0) 16 byte(s)
Address 0 1 2 3 4 5 6 7 8 9 A B C D E F ASCII
-------- ----------------------- ----------------------- ----------------
01f389d0 D2 04 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
01f389e0 23 45 E6 78 90 00 00 00 00 00 00 00 00 00 00 00 #E.x............
See also: base64, fromhex, tohex, inttohex, md5, crc, encrypt, decrypt
integer [baseconv number base ?-ucase?]
integer [convbase number base]
[baseconv] can convert from base 10 to any base (2..16), [convbase] can
convert positive integers from an arbitrary base (2..16) to base 10.
No type prefix or zero padding is added to the result. If you wish to use
the resulting value via casting or other commands you may need to prefix
with a type identifier such as "0x", "0b" or "0o".
[baseconv] output may include alpha characters depending on the selected
base. By default these will be lower-case unless the '-ucase' argument is
passed
CAUTION: Do not prefix convbase inputs with a number-type prefix such as 0x
or the Macro PreProcessor will pre-convert to decimal before it is passed to
the command
Example:
convbase 499602D2 16
convbase 1010 2
Result
1234567890 (Base 10)
10 (Base 10)
Example:
baseconv 10 2
Result
1010
The resulting numeric values are not prefixed by a type flag e.g. 0x, 0o or 0b
See also: inttobin, inttohex, inttooct
beep ?frequency? ?duration?
Play a note. The default frequency is 700Hz and the default duration is
200ms
Example:
option expression off
for {set i 300} {< $i 900} {incr i 100} {beep $i 400}
Examples:
beep 600
beep - 2000
Note that best performance is on systems which support the 8254 timer chip.
This is emulated on Windows 7 and later via the sound card, and not at all
on some versions of Windows. Performance may be very with such emulated
sound output.
Support for the Beep API was dropped in Windows Vista and Windows XP 64-Bit
See:
https://msdn.microsoft.com/en-gb/library/windows/desktop/ms679277%28v=vs.85%29.aspx
See also: playwav
Using Ticol With Batch Files |
Index |
It is possible to emulate the Linux 'shebang' system and create a batch-
script which can automatically execute Ticol Tcl code within the same file
The following batch script can be used as a stub-wrapper for Tcl code
::set - {
@echo off
REM ##################################################################
REM Name: magic.bat
REM Purpose: Command-line chaining test for Ticol Tcl
REM Notes: Run as: magic.bat
REM The initial enclosing [set] command is ignored by Tcl
REM ##################################################################
REM Check filetype using BATCH script
if "%~x0" equ "" (
set SCRIPT=%0.bat
)
if "%~x0" neq "" (
set SCRIPT=%0%~x0
)
ticol.exe %SCRIPT% %1 %2 %3 %4 %5 %6 %7 %8 %9 /na
exit /b
}
# start Tcl Code ####################################################
textcolor magenta
puts "Hello from Ticol Tcl"
textcolor
newline
See also: tcl2bat, Ticol, md5 batch example, autoexec.tcl, autoexec
exec, spawn, shell
bool [in list value ?-glob|-nocase?]
bool [ni list value ?-glob|-nocase?]
Return a boolean indicating whether an element is in or not in a well-
formed single-level Tcl list. Multi-level lists are unsupported.
[in] and [ni] are list rather than string matching commands
[lsearch] is more suited to complex list-searching
A 'glob' style wildcard or case-insensitive search may be used.
These options cannot be combined and -glob is case-significant
Example Result
# qbf: The quick brown fox jumps over the lazy dog
puts [in $qbf fox] # 1
puts [in $qbf FOX -nocase] # 1
puts [in $qbf f?x -glob] # 1
puts [in $qbf f*x -glob] # 1
puts [in $qbf fred] # 0
puts [ni $qbf fox] # 0
puts [ni $qbf fred] # 1
See also: list, lsearch, lindex, llength, instr, string substr
Ticol is a Tcl interpreter, not a compiler! |
Index |
Ticol is a Tcl interpreter. This means that Tcl scripts are evaluated
on-the-fly and are not translated into machine instructions for the CPU
Interpreters are necessarily slower in terms of execution than any compiled
language. Some other languages compile to an intermediate form (p-code)
which is designed to run on a virtual machine (virtual CPU) and these can
perform very fast and in some cases out-perform native compiled object code.
However these languages are not interpreters in the truest sense. This means
Ticol will be relatively slow and should not be used for tasks where
performance is required
For this reason it is vital to optimise Tcl scripts. One of the ways this is
achieved is to use a Macro PreProcessor (MPP) to pre-process scripts and
perform any necessary static translations, e.g. on non-decimal const numbers
such as 0x123 and convert these to decimal values. C-like escape sequences
such as \r or \n are also converted to literal characters.
Ticol includes a number of non-standard Tcl commands which are designed to
facilitate optimal performance. See: performance
The command [calc] is an optimisable alternative to [expr]. This command
will be optimised into simpler Tcl commands by the Macro PreProcessor unless
disabled. If disabled it will be interpreted identically to [expr]
See: Macro PreProcessor for more information
See:
https://en.wikipedia.org/wiki/Tcl
See also: calc, Tcl, Ticol
string [interp ?script?]
Launches a new instance of the Ticol interpreter with a separate CLI
and preserving the original interpreter. At present only a rudimentary
implementation of [interp] is offered.
Currently this runs in the same thread as the original intepreter and
the original interpreter is 'suspended' during execution. The child
interpreter will inherit the configuration of the parent. In a future
version of Ticol, [interp] will run in a separate thread.
To exit this interpreter instance and return to the original interpreter
instance, use the [exit] command. If [exit] or [return] is not present
in the script body then the child CLI will be entered
Example:
set a 23
assert $a {== $_ 23} -v -line #__LINE__
interp {
puts "Enter child interpreter @"#__LINE__
set a 24
assert $a {== $_ 24} -v -line #__LINE__
exit
}
assert $a {== $_ 23} -v -line #__LINE__
[interp] may be used to return a result as with [proc]
Example:
set r [interp {
return [* 4 [funct atan 1]]
}]
puts "interp result was $r"
Result:
interp result was 3.141592653589792
See also: cli, spawn, exec, commands, interpreter
binary (Experimental binary command) |
Index |
bool [binary set varname escapedString ?-length x?]
string [binary set varname ?-length?]
string [binary tohex ?-var? varname]
bool [binary fromhex string varname]
string [binary base64 string|varname -var ?-encode? | ?-decode var?]
integer [binary length varname]
binary memcpy tcl-var source-address length
bool [binary update varname]
Ticol uses uncounted ASCIIZ string which limits flexibility for use with
binary strings which contain embedded NULL characters
Ticol supports a few rudimentary binary operations. Note that where
'varname' is specified as an argument this should NOT be dereferenced as
dereferenced strings are not handled as binary and will be truncated at
the first NULL character.
Binary strings cannot be returned from commands directly, so binary
commands which return binary strings cannot be chained together (nested)
If an argument variable does not exist then it will be created. If it
exists and has content it will be overwitten. Useful for handling binary
registry values. [binary] is not a full implementation of the Tcl [binary]
command
[binary] commands can not be used on non-binary variables
binary set varname escapedString ?-length x?
Set a standard Tcl variable with binary content
binary tohex string|varname -var
Return a hex interpretation of a binary variable or string
If a variable is passed, use the -var argument
binary fromhex string varname
Convert a hex string to a Ticol binary variable
The contents of the variable will be overwritten
If the variable does not exist it will be created
binary base64 string|varname -?var? ?-encode? | ?-decode varname?
Returns the contents of a binary variable or string, base64-
encoded or decoded. -encode is the default mode for this command
If a variable is passed, use the -var argument
If -decode is specified then the argument must be
followed by the name of a new output variable. The output
from -decode will/may be a binary string with embedded NULLs
Where a variable name is specified the return value is a boolean
Otherwise (for -encode with no varname), the return value is
the encoded ASCIIZ string in base64 format
binary length varname
Return an integer value representing the stored length
The variable must be a binary one which has been set using
the [binary] command. Standard variables will return a binary
length of 0
binary memcpy varname sourceAddress length
Copies data from memory address to a Tcl variable
The Tcl variable will have memory allocated to hold the source
The variable is marked as 'binary' and will have a length value
The resulting binary string is "safe" and is null-terminated
It may safely be printed using [puts]
binary update varname
Updates a Ticol variable and forces one with an existing value
to be recognised as a binary type
Notes:
The Ticol Macro PreProcessor (MPP) will not allow \x00 or \x22 entries
as these would cause strings to malfunction, but these are internally
handled by [binary]. Remaining values are unescaped by the MPP before
the script is loaded
Example:
binary set a \xfe\xffhello\x21\x00\x22world\x02\xff\xfe -length 18
puts "'$a!'"
puts "normal: length=[string length $a]"
puts "binary: length=[binary length a]"
puts "tohex: '[binary tohex a]'"
puts "base64: '[binary base64 a]'"
Results:
'. Hello'
normal: length=8
binary: length=18
tohex: 'feff68656c6c6f210022776f726c6402fffe'
base64: '/v9oZWxsbyEAIndvcmxkAv/+'
Example:
set a 23
binary memcpy b [addressofb a] 2
puts "\$b=='$b'"
puts "binary length=[binary length b]"
puts "string length=[string length b]"
Result:
$b=='23'
binary length=2
string length=2
See also: set, tohex, encrypt, base64, tohex, fromhex
integer [bintoint binary-value]
string [inttobin value ?-full? ?-noprefix?]
[bintoint] converts a binary value to a decimal
[inttobin] converts a decimal value to a binary string with the prefix
"0b". Conversion of binary to decimal integer is automatic in Ticol
Values are calculated as 64-bit integers
Values given to inttobin may be decimal integers, hexadecimal, octal or binary
The optional ?-full? arg includes leading left-hand zeroes for the full,
64-bit internal representation
[format] and [printf] may be used to force full width formatting
CAUTION
-------
The Macro PreProcessor (MPP) will automatically convert octal values in
the form 0bNN into decimal. Be cautious when passing unquoted binary
values with "0b" prefix to [bintoint]
Examples:
option echo on
inttobin 9223372036854775808 -full
inttobin 666 -full
inttobin 666 -noprefix
inttobin 666
inttobin 0
printf "0b%08s\r\n" [inttobin 0 -noprefix]
Results:
0b1000000000000000000000000000000000000000000000000000000000000000
0b0000000000000000000000000000000000000000000000000000001010011010
0000000000000000000000000000000000000000000000000000001010011010
0b1010011010
0b0
0b00000000
See also: big number, inttohex, inttoct
bool [any bool ?bool...?]
bool [all bool ?bool...?]
The [any] and [all] commands offer efficient and compact methods of testing
multiple boolean or pseudo-boolean results without requiring either the
use of [expr] or deep command nesting necessary with multiple [||] commands
Both forms perform lazy evaluation. [expr] with natural expressions may be
used but this is less efficient than using the native Tcl command form.
[any] will return $true (1) if any clause is true
[all] will return $true (1) if and only if all clauses are true
To test say four different variables, a, b, c, d and would require the
following using Tcl format expressions:
Example:
option expression off
# We assume a b c d and e are already set
if { || $a [|| $b [|| $c [|| $d $e]]] } {puts "At least one variable is set"}
if { && $a [&& $b [&& $c [&& $d $e]]] } {puts "All variables are set"}
Can be replaced with the following, which is easier to read and less prone
to hard-to-trace bracket parity errors:
option expression off
# We assume a b c d and e are already set
if {any $a $b $c $d $e} {puts "At least one variable is set"}
if {all $a $b $c $d $e} {puts "All variables are set"}
if {all [any $a $b $c] [any $d $e $f]} {puts "Pass"}
if {any [all $a $b $c] [all $d $e $f]} {puts "a,b,c XOR d,e,f"}
Example:
puts [any 0 1 0]
puts [any 0 0 0]
puts [any 1 1 1]
newline
puts [all 1 1 1]
puts [all 1 0 1]
puts [all 0 0 0]
newline
set a [rnd 0 1]
set b [rnd 0 1]
if {all $a $b} {puts "True"} else {puts "False"}
Results:
1
0
1
1
0
0
<Prints True or False depending on [rnd]>
See also: bool, &&, ||, if, flow control
bool (boolean values) |
Index |
string [bool value] # Command
$true # Global constant ($::true in proc scope)
$false # Global constant ($::false in proc scope)
[bool] returns the string "true" or "false" depending on the input numeric
value. A non-zero value returns true, a zero value, false
If the string "true" or "false" is passed to [bool] then 1 or 0 is
returned with the string being evaluated for a match
Constants are defined as $true and $false, but these reside in global
scope and must be referred within procs using the scope prefix $::true and
$::false
Raw boolean words 'true' or 'false' may be defined by the Macro
PreProcessor (MPP) as follows. These substitutions are unaffected by
scoping issues
# Define macro symbols for 'true' and 'false'
#define true 1
#define false 0
assert true {== $_ 1}
assert false {== $_ 0}
# String comparisons for macro values *must* be dquoted and not braced
# this prevents the MPP performing a substitution, braces do not
assert [bool true] {eq $_ "true"} -v -line #__LINE__
# Otherwise ...
assert [bool true] {eq $_ true} -v -line #__LINE__
# Would translate to...
assert [bool 1] {eq $_ 1} -v -line #__LINE__
# In which case, [bool 1] would return the string "true"
Example input mappings for [bool]:
Input value Result
--------------------
1 True
3 True
-3 True
0 False
x False
True 1
TRUE 1
False 0
FALSE 0
true and false are defined as constants 1 and 0 respectively
Examples:
puts [bool 1]
puts [bool true]
puts [bool "true"]
puts [bool {true}]
puts [bool false]
puts [bool [> [funct rnd 1 10] 5]]
Results:
true
1
1
1
0
The string "true" or "false" depending on [rnd] output
See also: math, string
sbool (Signed Boolean) |
Index |
signed-boolean-integer [sbool value]
Returns -1 (true) or 0 (false)
Converting BASIC code to Tcl can be problematic as expressions may include
boolean results which behave differently to Tcl. This can be because BASIC
usually uses the signed integer -1 (0xFFFF) as a 'true' value whereas Tcl
uses the positive integer, 1.
[sbool] takes a value, not an expression. Use [expr] or [eval] to evaluate
If combined with bitwise operations translated from BASIC, unexpected
results may occur.
Example:
# BASIC code example: (7 AND -8 < -5)
# WRONG because [< -8 -5] returns 1 not -1, which breaks [&]
puts [& 7 [< -8 -5]]
# CORRECT because [< -8 -5] will now be cast to -1 (0xFFFFFFFFFFFFFFFF)
puts [& 7 [sbool [< -8 -5]]]
Result:
1 # Wrong
7 # Correct
See also: math, bool
Draw a graphic box
box x y w h ?forecolour | -? ?backcolour? ?-single? ?-solid? ?-nofill?
Color may be a symbolic name such as 'blue' or a numeric value which
corresponds with standard Windows console values (e.g. blue, cyan and
these may optionally be prefixed with 'light', e.g. 'lightred')
Both 'grey' and 'gray' are accepted.
-single will draw a single-line border (the default is double line)
-solid will remove the border pattern and draw a solid box of 'backcolour'
-nofill will draw only the box outline in 'forecolour'
'backcolour' is the box 'fill' colour
'forecolour' may be ignored when specifying 'backcolour' by using a -
Displaying the box will change the cursor position. Use gotoxy to
reset. gotoxy may also be used to draw within the box area
Screen colours are reset back to defaults after a call to [box]
If width w or height h is zero then a vertical bar or horizontal line
will be drawn
Examples:
box 3 4 30 10 - lightcyan -solid
box [rnd 1 20] [rnd 1 20] [rnd 5 30] [rnd 1 20] - lightcyan
See also: msgbox, cls, gotoxy
Braces {} (Grouping) and Brace Style |
Index |
Braces group Tcl commands including string characters. Braces may be nested
Typically each call to a Tcl command unwraps one level of braces
In some cases, double-quotes may be used to group words together. Note that
single-quotes do not group, so ...
puts 'Hello world'
is not a valid group/string, whereas ...
puts "Hello world"
or ...
puts {Hello world}
... is valid, as braces may also delimit a string and can be useful where
there would otherwise be nested double-quotes.
Braces delay evaluation and may be used strategically to prevent
evaluation at a specific level. Evaluation may be forced using [subst]
Each command-call removes one set of braces and evaluates the result
Because braces delay evaluation for one level they will prevent variables
from being replaced unless [subst] is called on the group. Thus later
evaluations which are dependent on such variables will fail
Nested levels of braces, used outside of quoted strings may be used to
prevent interpretation within the current context, i.e. to delay
interpretation to a later evaluation in a sequence
Braced Variables
----------------
Braces may also be used to isolate concurrent variables within strings and
to prevent misinterpretation of variables adjacent to literal string
content e.g. in "$variablestring" as "${variable}string" we want to isolate
'variable' from 'string' and dereference $variable but not 'string'.
This is easier in languages which have a dollar suffix but Tcl has a dollar
prefix which requires a more complex solution to strings 'butting up'
against each other. When used in this context, whitespace is significant
and may be interpreted as part of a variable name. Therefore care should
be taken to avoid including unnecessary whitespace in braced variable
references.
Braces which fully-enclose variables name within strings do not delay
evaluation and may also be used to distinguish variable names from string
content which follows immediately-after
e.g.
set will_be_evaluated_ok "Hello "
puts ${will_be_evaluated_ok}string # Normal dereference
set q will_be_evaluated_ok
puts ${$q}string # Double dereference
puts $[$q }string # Will look for var 'q '
Result:
Hello string
Hello string
# Variable 'q ' not found
Braces may be used within lists to delineate groupings
Example:
set a "Goodbye "
set b "cruel "
puts "$a${b}world"
Result:
Goodbye cruel world
Brace Style
-----------
Tcl requires K&R style brace placement as the line end forms part of the
command syntax. Failure to adhere to this style will cause fatal errors
proc foo x { # Correct K&R Tcl style
...
}
proc foo x # Incorrect Allman/BSD C/C++ style
{
...
}
See:
https://en.wikipedia.org/wiki/Indent_style
http://wiki.tcl.tk/708
See also: Tcl commands, double dereference, dereference, flow control,
subst, eval, math, Tcl
call proc ?argument...?
call command ?argument...?
call "<tcl script>"
Allows the calling of a procedure, Ticol command or scriptlet
Similar to the function of CALL in legacy Visual BASIC. Call allows a
proc to be called directly bypassing expression-evaluation, [trace]
and /G graph statistics.
A command can be called. In this case the normal evaluation chain is
followed including [trace] and /G graphing
Finally, a text script containing Ticol Tcl commands may be called
Note that Ticol treats numeric labels within square brackets as
literal numbers e.g. [123]. This is by design. [call] may be used
to call procs with all numeric names.
Proc Example:
proc 123 {} {puts "I am in bad proc 123"}
[123] # Not called as [123] is evaluated as a number
call 123 # Proc [123] is called successfully
Result:
I am in bad proc 123
Command Examples:
option echo on
call expr "sqrt(99)"
Result: (Submitted via the CLI prompt)
[0] 9.949874371066199
Script Example:
call "textcolor magenta; puts {Hello } -nonewline ;\
textcolor red; puts world; textcolor"
Result: (Submitted via the CLI prompt)
Hello world
See also: puts, eval, expr, foreach, sub, gosub, goto, dummy
External DLL Link Plugin - calldll*, calldll*_cdecl |
Index |
To load, use: lib ticol_calldll ?-verbose? ?-nocomplain?
Syntax:
calldll* dllname?.dll? function-name|@ordinal arg arg...
calldll* handle function-name|@ordinal arg arg...
See also: dll_open
See also: dll_close
[calldll*], where * is one of the suffix types listed below, can safely
call an external ANSI DLL routine.
This command is still experimental and calling-conventions may change
Both __stdcall and __cdecl non-COM DLLs are supported thus offering
connectivity with a wide range of other applications including the
Microsoft Visual C++ runtime (msvcrt40.dll), SQLite3, AutoIT and
other popular DLLs which use standard interface methods
Calling DLLs is not a trivial task an the marshalling of variables both
outward and return can be quite complex and difficult to debug. The correct
call-type (standard __stdcall or __Cdecl) must be used to match the DLL
The DLL may be called either by name or by a handle returned from dll_open
Additionally, an exported DLL function may be called by ordinal number by
prefixing the ordinal with a @ symbol. e.g.
# Must be run from an elevated console prompt or session
puts "Is user an admin?: "[bool [calldll "Shell32" "@680"]]
Handling Partial and Default DLL Arguments
------------------------------------------
All arguments must be supplied, even if optional for the DLL interface
Pass default NULL arguments as 0. Failure to supply sufficient arguments,
even if defaults exist, will result in a [calldll] exception
e.g. for C++ export prototype in foo.dll
void __stdcall bar(LPSTR, long* = NULL);
Pass:
calldll foo.dll bar "Hello" 0 # Equivalent to passing NULL
set r "000000000000000000000" # Ensure buffer is sufficient
calldll foo.dll bar "Hello" r # And pass 'r; by reference
struct r {s 21} # Create a small struct
calldll foo.dll bar "Hello" r # And pass by reference
Quoting
-------
Quotes around the dll and function name are optional, as is the DLL suffix
It is usually recommended that calldll calls be wrapped inside a procedure
Byte Widths
-----------
Byte widths may be specified in arguments as 'var:<width>' e.g. 'data:4'
If a variable or string contains a ":" character (e.g. in a Windows path),
a byte-width MUST be specified. This feature is not ideal and will be
improved in a future version. At present [calldll] is fairly primitive
and satisfied only the basic design requirements.
Passing by Value or by Reference
--------------------------------
Variables are passed by reference if a variable name is used and the
variable exists, if prefixed by a $ then they are passed by value.
If you pass a value which could be interpreted as a variable name then
ensure you [unset] that name before calling [calldll] as detection is
automatic.
Passing a Tcl variable by reference does not guarantee sufficient
allocated memory space to hold any return. Either the called DLL must
allocate using malloc() or you should allocate a sufficient buffer
by using [makestr] or assigning a suitably large value using [set]
Alternatively, you may pass the address of a single-member struct
which has sufficient static buffer space in the member field.
Literals, e.g. "Hello" or 23, are passed by value for example
Passing by Name or Reference
----------------------------
Where a variable is passed by name rather than value, the argument is
passed to the DLL by reference. String variables will be passed to the
DLL as char* which will allow the original string data block to be modified
This should be used with extreme caution and the variable must be created
with sufficient space allocated to handle any return. Use [makestr] for
example, to do this.
# 'funct()' below is an exported __stdcall function in 'dllname.dll'
set var 10
# Passed as char* but DLL should not modify the string contents
calldll dllname funct $var # Pass a variable by value
# Passed as char* and DLL may modify carefully
calldll dllname funct var # Pass by Tcl name (reference)
# Passed as char* and DLL may modify carefully
calldll dllname funct [addressofb var] # Pass by reference
# Passed as char* and DLL may modify carefully
calldll dllname funct [strptr var] # Pass by reference
Note: The Microsoft Visual C++ runtime (msvcrt40.dll) is compiled as a
__cdecl DLL and calldll*_cdecl must be used to call it.
Passing Binary Addresses
------------------------
Passing the binary address of a variable buffer/string pointer by value may
be forced by using [addressofb variable] to pass the literal buffer address
(LPSTR in C++). [strptr] is an alias for [addressofb]. This is most useful
for struct field address offsets. In most cases [calldll*] will auto-
detect the correct type of cast to apply.
Passing C/C++ Pointer Arrays (String Arrays)
--------------------------------------------
Pointer array such as char** are not directly supported by can be emulated
using [struct] to store an array of 4-byte addresses. An example is given
elsewhere. See: calldll examples
Return types and calls
----------------------
Return Type Use
----------------------------------------------------------------------
long, short, void: calldll dllname|handle funct ?var...?
You may also recast a long value to string pointer
<type*> You may also cast any accessible type
double, float: calldlldbl dllname|handle funct ?var...?
string, bstr: calldllstr dllname|handle funct ?var...?
variant calldllvariant dllname|handle funct ?var...?
Return variant must be released using [free]
long, short, void: calldll_cdecl dllname|handle funct ?var...?
double, float: calldlldbl_cdecl dllname|handle funct ?var...?
string, bstr: calldllstr_cdecl dllname|handle funct ?var...?
variant calldllvariant_cdecl dllname|handle funct ?var...?
Return variant must be released using [free]
Return Error Codes
-------------------
0 No error
DLL Name only (file exists)
DLL Name + Funct Name (file and funct exist)
DLL Name + all parameters (success)
1 No DLL filename was given
2 Failed to get DLL handle (DLL not found/doesn't exist)
3 Can't get proc address to given function in existing DLL
4 Unhandled data type was passed via the VARIANT to CallDLL*
5 Invalid ordinal number was passed for given DLL
6 An invalid handle was passed
Examples
--------
Example:
calldll_cdecl msvcrt40.dll printf "Hi from printf %i %f\n" 123 4.567
Result:
Hi from printf 123 4.567000
Example:
# Must be run from an elevated console prompt or session
puts "Shell32 is user an admin: [bool [calldll shell32 @680]]"
Result:
true
Example:
# Must be run from an elevated console prompt or session
puts "AutoIt is_admin: [bool [calldll AutoItX3.dll AU3_is_admin]]"
Result:
true
Notes on Calling MSVC Print Functions From ticol.dll
----------------------------------------------------
[calldll] cannot call MSVC (msvcrt40.dll) console print functions
After 256 calls they will return -1 (EOF)
Console output requires Windows API functions and which are
exported as lib functions WriteF() and WriteLn()
Notes on DLLs Returning Static Structs
--------------------------------------
If a DLL returns a pointer to a static struct or object which requires
persistent memory read, for example, msvcrt40.dll function asctime()
will return a pointer to 'static struct tm', it is vital that
[dll_load] is used to load and keep the DLL in memory while that data
structure is referenced or copied. You may call the DLL using the handle
returned by [dll_open]
Example:
lib ticol_calldll
set handle [dll_open msvcrt40.dll]
set t 1470404087
set tm_ptr [calldll_cdecl $handle localtime t]
set s [calldllstr_cdecl $handle asctime tm_ptr]
dll_close $handle
puts $s
Result:
Fri Aug 05 14:34:47 2016
Data Structures
---------------
External API routines can often use structs. Ticol provides binary-
compatible structs which can interface with both Ticol and external DLLs
The return from a C++ struct* may be passed to a local Tcl struct using
[struct copy]
Variant returns, including 1D and 2D string arrays are handled
See: structs
See also: calldllstr, calldll examples, dll_open, dll_close, struct,
addressof, addressofb, strptr, ofaddress, struct copy
Note: All examples require the statement: lib ticol_calldll
--
set t 1470404087 # Fri Aug 05 14:34:47 2016
set ret [calldllstr_cdecl msvcrt40.dll ctime t] # Convert time_t to string
puts "Result is ctime(FIXED-DATE): -"[string trim ret [chr 10]]-
--
calldll user32 SetWindowTextA [calldll Kernel32 GetConsoleWindow] \
"Ticol Tcl"
--
puts "Is user an admin?: [calldll Shell32 @680]"
--
puts "is_admin: [bool [calldll AutoItX3.dll AU3_is_admin]]"
true
--
calldll_cdecl msvcrt40.dll printf "Hi from printf %i %f\n" 123 4.567
Hi from printf 123 4.567000
--
calldll mslib145 MapNetworkDrive "\\srv01\c[chr 36]" "q:" "" "" errcode 1 0
--
puts "proc mem: "[comma [calldll "mslib145.dll" "GetProcessMemoryUsed" \
[pid]]]
--
proc ipfromhost {hostname} {
return [calldllstr ping32 GetIPFromHostName $hostname]
}
--
puts "calldlldbl_cdecl->"[calldlldbl_cdecl msvcrt40.dll sqrt $pi]
--
Example passing the address of a buffer by value to call a DLL:
set timeout 0
set datalen 64
set buf [makestr 16]
calldll ping32 PingAPI google.com [addressofb buf]:4 datalen:2 timeout:4\
1000
Returns the string: 173.194.78.105 in variable 'buf'
--
Example querying Windows OS version
struct OsVersionInfo {
{OSVSize 4} # long
{dwVerMajor 4} # long
{dwVerMinor 4} # long
{dwBuildNumber 4} # long
{PlatformID 4} # long
{szCSDVersion 128} # char* x 128 bytes
}
setb OSVSize [struct size OsVersionInfo] 4 -value
puts "GetVersionEx=[calldll kernel32 GetVersionExA OsVersionInfo]"
puts "OSVSize: [ofaddressb OSVSize 4]"
puts "dwVerMajor: [ofaddressb dwVerMajor 4]"
puts "dwVerMinor: [ofaddressb dwVerMinor 4]"
puts "dwBuildNumber: [ofaddressb dwBuildNumber 4]"
puts "PlatformID: [ofaddressb PlatformID 4]"
puts "szCSDVersion: [ofaddress szCSDVersion]"
dump OsVersionInfo
--
# Requires the author's vbsqlite3.dll port
set QueryString "SELECT t1.ArtistName,CDs.Title,CDs.date, CDs.CDID\
FROM Artists t1, CDs\
Where t1.ArtistID = CDs.ArtistID\
ORDER BY ArtistName, CDs.Date ASC "
set dbfile ".\\test.db"
set h [dll_open "vbsqlite3.dll"]
set r [calldll $h sqlite3_open [strptr dbfile] db]
set vr [calldllvariant $h sqlite_get_table_basic $db:4 $QueryString]
set rows [varray size $vr]
set cols [varray size $vr 2]
set dims [calldll $h "GetArrayDimensions" $vr]
puts "Table has $rows rows and $cols columns and $dims dimensions"
calldll $h sqlite3_close $db
dll_close h
--
# Call dll by handle
# Requires the author's VB Toolbox DLL
option expression off
set handle [dll_open mslib145.dll]
calldll $handle LoadIPHLPAPI
set adapters [calldll $handle GetAdapterCount]
puts "Querying $adapters adapter(s)..."
option expression off
for {set i 0} {< $i $adapters} {++ i} {
set addr [calldllstr $handle GetIPAddress $i]
set name [calldllstr $handle GetAdapterDescription $i]
puts "$i: [left $name 40] -> $addr"
}
dll_close $handle
--
Calling a DLL or EXE with exports which requires a pointer array
Pointer arrays (char*[]) are not supported by Ticol but can be
emulated using [struct] as follows
N.B. Calling ticol.exe internal exports is not recommended!
# Run under ticol.exe or ticol.dll
if {[info is_dll]} {
set filename ticol.dll
} else {
set filename ticol.exe
}
lib ticol_calldll
# Open a binary which contains valid exports...
set h [dll_open $filename]
# Get the global interpreter handle (internal workings)
set i [calldll_cdecl $h picolThis]
# Create a pesudo-array of 2 x 4-byte "pointers" in a struct
# as successive memory addresses
struct foo {
a1 4
a2 4
}
set arg1 "eval"
# A simple test which is interpreted via [eval]
set arg2 {textcolor magenta;puts [expr 22/7.0];textcolor}
# Map the var's binary string into the struct address space
struct setb foo.a1 [strptr arg1]
struct setb foo.a2 [strptr arg2]
# Call the DLL (must be __cdecl)
set s [calldll_cdecl $h picolCommandEval $i 2 foo 0]
puts "Tcl result code is $s"
--
ReactOS
-------
[calldll] may not work reliably with the ReactOS operating system
See also: calldllstr, calldll, calldll_cdecl, dll_open, dll_close,
single line commands
calldllstr, calldllstr_cdecl |
Index |
To load, use: lib ticol_calldll ?-verbose? ?-nocomplain?
Syntax:
calldllstr dllname?.dll? function-name|@ordinal arg arg...
calldllstr_cdecl dllname?.dll? function-name|@ordinal arg arg...
Call a DLL which returns a string result, exported as a Tcl string
If there is a need to dereference or address a returned pointer, e.g. to
release memory allocated by the DLL call then you must use [calldll] instead
and use the returned integer pointer address value. When a Tcl string is
returned then the original address will be lost
Static references to strings returned via [calldll] will return a reference
to some part of an external variable. When [calldllstr] is used the original
string (pointer address) reference is lost. No attempt should be made
therefore, to release values returned by [calldllstr] other than using
standard Tcl methods such as [unset]
[calldllstr_cdecl] is used to call DLLs compiled with the __cdecl DLL
calling convention. Use [calldll_cdecl] to return an address value
If memory wasn't allocated by the DLL call then any return address
returned by [calldll] when returning a string may be discarded.
A null address will return 0
Example:
# Uses shell32 StrChrA to locate a substring by character match
# Shows differences between addressof and addressofb
# and use of ofaddressb to dereference
set s "Hello World"
puts "addressof s [addressof s]" # Note how these addresses differ
puts "addressofb s [addressofb s]" # This is the var string address
puts "calldll: H is '[calldllstr shell32 StrChrA [addressofb s] [asc H]]'"
set ptr [calldllstr shell32 StrChrA [addressofb s] [asc W]]
puts "StrChrA($s,W) offset should be 'World' -> '[ofaddressb $ptr']"
See also: calldll
To load, use: lib ticol_carray ?-verbose? ?-nocomplain?
Syntax:
handle [carray create ?size_x? ?size_y? ?size_z?]
integer [carray add handle lval ?rval?]
carray delete handle index
integer [carray count handle]
string [carray item handle index ?default-value? ?-rvalue? ?-list?]
string [carray get handle index ?-rvalue?]
string [carray get_2d handle x y ?-rvalue? ?-tuple?]
carray foreach handle lvalue ?rvalue? {code}
string [carray set handle index]
carray set handle index lvalue ?rvalue?
carray set_2d handle x y lvalue ?rvalue?
carray sort handle ?-reverse? ?-rvalue? ?-numeric?
carray walk handle
bool [carray unset handle ?-nocomplain?]
Emulates "C"-like fixed arrays rather than associative arrays. Fixed
arrays are indexed using 'base zero' integer index values and may be
sorted, Carrays are never sparse. Deletions are infilled by rearranging
the array. Sorting will rearrange the element order
[carray] can optionally act as an integer-indexed tuple and store an
"RValue" as well as the required "LValue".
IndexValue -> 'LValue' + ?RValue?
Multi-dimensional arrays are not supported. As with early versions of
FORTRAN, they can be emulated using integer arithmetic to calculate an
offset within the 1-dimensional array.
create # Can pre-allocate memory to a given number of elements
# [carray create] Create a dynamic array
# [carray create x] Create a 1 dimensional fixed array
# [carray create x y] Create a 2 dimensional fixed array
# [carray create x y z] Create a 3 dimensional fixed array
# Address 2D arrays using set_2d and get_2d
add # Adds an lvalue and, optionally, a corresponding rvalue
# An Rvalue only may be added by specifying the Lvalue as {}
# The updated record count is return on success, otherwise 0
delete # Removes an item by index (both Lvalue and any Rvalue)
# The array is re-packed and indexed items will be re-ordered
count # Returns a count of items in the array
foreach # Offers a means of iterating each element of the array in
# ascending integer storage order. [foreach] is affected by
# the internally-sorted order of the array
# The variables are placeholders for lvalue and rvalues
# with rvalue being optional
# [foreach] works only with 1 dimensional arrays
# use iteration and direct addressing for 2d and 3d arrays
get # Retrieve an array item by integer subscript. The element must
# exist, otherwise an error will be returned. Optionally, an
# RValue may be returned instead. [carray get] fails on error
get_2d # Get an lvalue and optionally an rvalue in a 2D array
get_3d # Get an lvalue and optionally an rvalue in a 3D array
item # Index the array by integer subscript, returning an LValue and
# optional Rvalue. A default value may be specified
# If the element does not exist either "" or an optional default
# value will be returned. [carray item] returns default on error
set # Either set an element's lvalue (optionally also rvalue) at a
# specific index position. Called with no value arguments this
# command will echo the lvalue at the index position
set_2d # Set an lvalue and optionally an rvalue in a 2D array
set_3d # Set an lvalue and optionally an rvalue in a 3D array
sort # Can sort the array in normal or reverse order. Note that fixed
# arrays are not as efficient at sorting as associative arrays
# and it may be impractical and slow to sort very large arrays
# The default sort is 'string', a numeric sort can be forced by
# using the -numeric argument. 64-bit Ticol integers are correctly
# handled by this option
# Record numbers may no longer match after a sort is performed
walk # Is useful for debugging. It will iterate the array elements
unset # Destroys the array and renders the handle unusable. Using
# the handle after a call to [unset] will result in an error
Example:
set h [carray create]
carray add $h Quick Brown
carray add $h Fox Jumps
carray add $h Lazy Dog
carray add $h LValue
carray add $h {} RValue
carray sort $h
carray foreach $h lv rv { # foreach iterator example
puts "Lvalue:'$lv'\tRvalue:'$rv'"
}
carray sort $h -reverse
set count [carray count $h]
puts "We have $count item(s)"
for {set i 0} {[< $i $count]} {++ i} { # Loop example
puts "Item $i: '[carray item $h $i]'\t'\
[carray item $h $i -rvalue]'"
}
Results:
Lvalue:'' Rvalue:'RVOnly'
Lvalue:'Fox' Rvalue:'Jumps'
Lvalue:'Hello' Rvalue:'World'
Lvalue:'LVonly' Rvalue:''
Lvalue:'Lazy' Rvalue:'Dog'
Lvalue:'Quick' Rvalue:'Brown'
Item 0: 'Quick' 'Brown'
Item 1: 'Lazy' 'Dog'
Item 2: 'LVonly' ''
Item 3: 'Hello' 'World'
Item 4: 'Fox' 'Jumps'
Item 5: '' 'RVOnly'
Advantages and Disadvantages
----------------------------
Unlike associative arrays, "C" arrays can have an inherently ordered
(potentially sorted) state can can only be indexed by integer value
rather than by a string. They are, on average, slower to insert/add new
records and sorting a large array may take a long time and searching a
record on an unsorted array will may be much longer than a hashtable-
based associative array. A Tcl array could be used to index a carray.
Deleting All Nodes
------------------
carrays are not 'sparse' so they will reorganise and resize downwards on
deletion, and any intermediate records will be removed. Thus record
ordering will change on the deletion of an intermediate node
As the count and record numbers may change after each call to
[carray delete] the following method should be used to iterate through and
delete all nodes:
set i 0
while {[> [carray count $h] 0]} {
carray delete $h 0 # Delete the first element
}
Deleting in reverse order will also work
for {set i [- [carray count $h] 1]} {>= $i 0} {-- i} {
carray delete $h $i
}
A naive [for] loop which iterates the count as follows won't work as the
array will be resized downwards after each deletion:
# Won't work as [carray count] changes after each [carray delete]
for {set i 0} {[< $i [carray count $h]]} {++ i} {
carray delete $h $i
}
This method also will not work...
set count [carray count $h]
for {set i 0} {[< $i $count]} {++ i} {
carray delete $h $i
}
Emulating Multi-Dimensional Arrays
----------------------------------
A linear array can be manipulated as a multi-dimensional
(N-dimensional) array with some trivial mathematics
A proc may be defined to handle N-dimensional indices and resolve these
into a 1-dimensional index value. [carray create] must be called with
fixed array dimensions
2D and 3D addressing commands are built-in from build date 19/Dec/2021
lib ticol_carray
set dim_x 5
set dim_y 3
# Proc [in] x and y should be array-base 0 values
proc array_set {handle x y value {rvalue}} {
carray set $handle [+ [* $y $::dim_x] $x] $value $rvalue
}
proc array_get {handle x y} {
return [carray get $handle [+ [* $y $::dim_x] $x]]
}
set h [carray create [* $dim_x $dim_y]]
array_set $h 0 0 "0,0" Alpha element
array_set $h 1 0 "4,2" Omega element
puts "1 0 -> '[array_get $h 1 0]'"
lib ticol_carray -unload
See the 2d_carray.tcl and 3d_carray.tcl examples
https://www.geeksforgeeks.org/emulating-a-2-d-array-using-1-d-array/
See also: plugins, arrays, dictionaries
Handling Sparse Integer-Indexed Arrays |
Index |
Tcl associative arrays are naturally "sparse" and have no innate connected
ordering. Where ordering is required, some sort of ordering array subscript
such as an integer subscript is required.
# The phrase "Hello world" is produced by implicit ordering by subcript
set a(1) "Hello"
set a(2) "world"
Where more control is required over ordering then the "carray" plugin
should be used. This allows "C"-like arrays which are indexed by an
integer address value and where both Lvalue and Rvalues may be stored.
A command based indexing syntax is provided rather than bracketed array
references. One, two and three dimensional carrays may be sparse in that
unoccupied elements which have been declared are set to NULL.
Iterating large sparse arrays may be problematic, performance wise. It may
be necessary to use a Tcl array to index the sparse array in order to avoid
say iterating a sparse array of say a million elements
To load, use: lib ticol_cgi ?varname?
A rudimentary command to process CGI server POST input as a CGI
(Common Gateway Interface) application and splits into an array variable
The default variable name is _POST but this can be changed by passing an
alternate name as an argument
Ticol can be used to generate a modest CGI 'back-end' system including
database using SQLite or XBase. The language is not, however, recommended
for commercial or industrial CGI use
[get] will set and populate an array variable called _POST()
Syntax:
integer [cgi] # Get/set the '_POST()' array
integer [cgi get ?-var name?] # Get a CGI variable
string [cgi arg name ?-nocase?] # Match QUERY_STRING
bool [cgi argfound name ?-nocase?] # Test QUERY_STRING
string [cgi post name ?-nocase?] # Match _POST value
bool [cgi postfound name ?-nocase?] # Test _POST value
Where 'name' is an optional alternative to the default var, '_POST'
If 'var' is omitted then $_POST() will be used from the last [get] call
Errors:
To avoid fatal errors within a CGI script, all errors other than a fatal
exception are non-fatal and will return code 'OK' an error code as follows
>0 SUCCESS Count of items returned
0 No items returned from POST
-1 Failed to initialise library
-2 Incorrect argument count
-3 CONTENT_LENGTH was not set
-4 CONTENT_LENGTH <= 0
-5 CONTENT_TYPE was not set
-6 REQUEST_METHOD not set to POST
-7 Unknown/unhandled error
Example:
Front-end form
<!-- Web form -->
<html>
<head>
<title>Test form</title>
</head>
<body><pre>
<form name="A" method="POST" action="/cgi-bin/ticol.exe?cgi_post.tcl"
enctype="multipart/form-data" encoding="multipart/form-data"><p>
Enter your name: <input type="text" name="name" size="30"><p>
Enter some text: <input type="text" name="text" size="40"><p>
<input type="submit" value="Submit" name="button1">
<input type="reset" value="Reset" name="button2"><p>
</form></pre></body>
</html>
Tcl back-end script
option expression off
puts "Content-type: text/html\r\n\r\n" # Send HTTP header
lib ticol_cgi -nocomplain # Load the plugin dll
set s [cgi] # Call & set $_POST() array var
if {> $s 0} { # Test for -ve error return
puts "<font color=\"blue\"><strong>"
array foreach _POST val subscript { # Iterate array $_POST()
puts "<br>Result \$_POST($subscript)='$val'<br>"
}
puts "</strong></font>"
}
flush # Flush stdout
Result:
<Depends on data typed into the input form>
See example file: lib_cgi_post.tcl and cgi_post.tcl (non lib version)
See also: security, cgi setup, plugins, dll interface, flush, plugins
CPU Information Plugin |
Index |
To load, use: lib ticol_cpu ?-verbose? ?-nocomplain?
Shows various low-level CPU information by directly accessing the CPU
Syntax:
string [cpu name] # The full, descriptive name of the CPU
string [cpu vendor] # The CPU vendor, e.g. "GenuineIntel"
integer [cpu bitness] # The underlying O/S bitness, (32 or 64)
integer [cpu cores] # Estimate the number of available cores
bool [cpu hyperthreading] # Hyperthreading supported. Returns 1 or 0
bool [cpu hypervisor] # Running on hypervisor. Returns 1 or 0
integer [cpu eax] # eax register return as integer
integer [cpu ebx] # ebx register return as integer
integer [cpu ecx] # ecx register return as integer
integer [cpu edx] # edx register return as integer
Not all VMs will show hypervisor enabled. Some VMs will emulate the CPU
entirely, others will use the underlying CPU but may disable some features
The number of cores is an estimate and may be incorrect. The operating
system may disable cores as it wishes
Examples:
lib ticol_cpu
puts [cpu name]
puts [cpu vendor]
puts [cpu bitness]
puts [cpu cores]
puts [cpu hypervisor]
puts "Virtual Machine Extensions " -nonewline
if {& [cpu edx] 0x2} { # Test edx bit 2
puts "present"
} else {
puts "not present"
}
Results:
Intel(R) Core(TM) i5 CPU M 540 @ 2.53GHz
GenuineIntel
64
4
FALSE
Virtual Machine Extensions present
See also: plugins
string [concat string ?string ...?]
Efficiently concatenate one or more strings and return the result
[concat] treats each argument as a list and concatenates them into a single
list. It permits any number of arguments. The arguments are unchanged
[concat] is much faster at accumulating strings than using chained [set]
operations as the original source string is not repeatedly passed via the
Ticol interpreter and parsed
To concatenate items to a variable see [append] or [store]
Example:
puts [concat The quick brown fox jumps over the lazy dog]
Result
The quick brown fox jumps over the lazy dog
Example:
concat a b {c d e} {f {g h}}
Result:
a b c d e f {g h}
See also: append, store, setat, cmp, set, variable, performance
integer [setc varName string ?length?]
Insert a source string into a variable's contents, centrally padded,
optionally for 'length' characters. The target variable is not truncated
Any value for 'length' < 0 will be ignored. Any value for 'length' greater
than the source string length or variable length will be ignored
The number of characters copied from the source string will be returned
If varName is empty then the request will be ignored and 0 will be returned
Use [setl] and [setr] to align left and right
Example:
set s [makestr 30]
set r [setc s Hello]
puts '$s'
puts $r
Result:
' Hello '
5
Example:
set s ""
set r [setc s Hello]
puts '$s'
puts $r
Result:
''
0
See also: setl, setr, string
cd path
chdir path
Change the current directory. Does not support wildcard * chdir prefix
Both forward and backslash characters are accepted
Where cd is called from the command-line as a script expression, remember
to escape all backslash characters and any reserved Tcl character such as
the dollar ($) sign
UNC paths are accepted but can have no action relative to a current, fixed
drive-letter:
Examples:
cd..
cd ..
cd \\rootdir
cd \\rootdir\\subdir
cd \\\\server\\share\\path # Will have no action relative to path
cd \\\\server\\c\$\\path # Note that $ must be escaped
cd /rootdir
cd /rootdir/subdir
cd //server/share/path # Will have no action relative to path
cd //server/c\$/path # Note that $ must be escaped
See also: mkdir, mkdirs, getcwd, file, pwd
string [chr n ?added-offset-value?]
The inverse of asc. Returns character number 'n' as a string
The argument 'n' must be an integer value, not a character
[chr] also permits a value to be added. This is useful where an integer is
converted into an ASCII letter e.g. 0 -> 'a', 1 -> 'b' and avoids an
additional and time-consuming call to [+] to calculate the ASCII value
Example:
puts [chr 65]
Result:
A
Example:
set i 2
puts [chr $i 97] # Add offset: character 2 + 97
Result:
c
A procedural equivalent might look something like this:
undef chr -nocomplain # Undefine the built-in command
proc chr {i} {
return [format "%c" $i]
}
puts [chr 65]
puts [chr 97]
Result
A
a
See also: asc, string, left, right, mid, index
string [chomp var ?-leave?]
string [chomp var N ?-leave?]
[chomp] takes only a variable name as its argument and is a non-standard
command to efficiently slice-up large strings.
If an argument length is not specified, then [chomp] removes and returns
characters up to and excluding the next whitespace character as one of
"\r\n\t ".
If an argument length, 'N' is specified, [chomp] removes and returns 'N'
characters from the left-hand side of a string. The remainder of the
characters not removed are left in the variable
The source variable will have the returned characters removed and they
are thus 'eaten' or 'chomped' away unless the '-leave' argument is used
[chomp] is faster and more memory-efficient with very large strings than
compounding [set] and [mids] or [set] and [string left] etc.
If argument N is larger than the string length then an whole string is
returned. If the source variable is empty then an empty string is
returned
chomp var N # Removing N characters is equivalent to
mids $var [+ N 1] # this command using [mids]
Example:
set s $qbf
chomp s 1 # Discards 1 character and retains 2nd char onwards
puts $s
set s $qbf
puts [mids $s 2] # Returns 2nd character onwards
Results:
he quick brown fox jumps over the lazy dog
he quick brown fox jumps over the lazy dog
See also: mids, left, string, string left
Standard Tcl time/date routines
Returns dates or times in the Unix epoch (01/01/1970)
integer [clock clicks|seconds|milliseconds]
string [clock format seconds|now ?-format "%Y-%m-%dT%H:%M:%S"?]
Where seconds must be a positive integer
See also: clock scan, clock split
The -format option controls what format the return will be in
The contents of the string argument to format has similar contents as the
format statement Additionally, there are several more %* descriptors which
can be used to describe the output.
%a Abbreviated weekday name (Mon, Tue, etc.)
%A Full weekday name (Monday, Tuesday, etc.)
%b Abbreviated month name (Jan, Feb, etc.)
%B Full month name (January, February, etc.)
%d Day of month
%j Julian day of year
%m Month number (01-12)
%y Year in century
%Y Year with 4 digits
%H Hour (00-23)
%I Hour (00-12)
%M Minutes (00-59)
%S Seconds(00-59)
%p PM or AM
%D Date as %m/%d/%y
%r Time as %I:%M:%S %p
%R Time as %H:%M
%T Time as %H:%M:%S
%Z Time Zone Name
The default -format mask is "%a %b %d %H:%M:%S %Y"
Examples:
Command: Result
clock clicks # 80579
clock seconds # 1425214130
clock format now -format "%Y %m %d" # 2015 03 01
clock format 9999 -format "%Y %m %d %H:%M:%S" # 1970 01 01 02:46:39
Example:
puts [clock format [double_to_ctime [now]] -format "%Y/%m/%d %H:%M:%S"
Result:
2015/04/08 12:15:41
Example:
puts "[clock milliseconds] v's [clock seconds]"
Result:
1512702231579 v's 1512702231
See: clock, clock scan, clock scan workaround
See also: clock format, clock scan, clock scan workaround, clock split,
time, date, ticks, double_to_ctime, ctime_to_double,
date_to_ctime, now, vars, is_dst
string [clock format seconds|now ?-format "%Y-%m-%dT%H:%M:%S"?]
Format codes for the clock -format string are as follows. These are
standard C/C++ formatting codes with a couple of extra Tcl codes
%a Abbreviated weekday name
%A Full weekday name
%b Abbreviated month name
%B Full month name
%c Date and time representation appropriate for locale
%D Day of month as a decimal number with no 0 prefix (1..31)
%d Day of month as decimal number (01..31)
%H Hour in 24-hour format (00 .. 23)
%I Hour in 12-hour format (01 .. 12)
%j Day of year as decimal number (001..366)
%m Month as decimal number (01 .. 12)
%M Minute as decimal number (00 .. 59)
%N Month of the year as a 1 or 2 digit number (1..12)
%p Current locale's A.M./P.M. indicator for 12-hour clock
%o Day of month ordinal. E.g. "st", "nd", "rd", "th"
%S Second as decimal number (00 .. 59)
%U Week of year as decimal number, Sunday as first day of week (00..51)
%w Weekday as decimal number (0 .. 6; Sunday is 0)
%W Week of year as decimal number, Monday as first day of week (00..51)
%x Date representation for current locale
%X Time representation for current locale
%y Year without century, as decimal number (00 .. 99)
%Y Year with century, as decimal number
%z Time-zone name or abbreviation; no characters if time zone is unknown
%Z Ditto
%% Percent sign
Example:
puts [clock format 0 -format "%A, %D%o %B %Y"]
Result:
Thursday, 1st January 1970
Example:
puts [clock format now -format "%A, %D%o %B %Y %H:%I:%S"]
Result
Wednesday, 4th October 2017 20:04:37
Example:
puts [clock format [date_to_ctime 1972 04 04] -format %D%N%y]
Result:
4472
See also: clock, clock split, clock scan, clock scan workaround,
time, date
clock scan <date time>
[clock scan] parses the date/time argument and returns the number of
seconds since the start of the Unix epoch which may be converted back
to a date using various methods such as [date]
[clock scan] accepts input in the following common formats
US date format - 'MMM DD ?HH:MM:SS?'
US date format - 'MMM DD YY ?HH:MM:SS?'
US date format - 'MMM DD YYYY ?HH:MM:SS?'
US date format - 'MMM DD, YY ?HH:MM:SS?'
US date format - 'MMM DD, YYYY ?HH:MM:SS?'
US date format - 'MMMM DD ?HH:MM:SS?'
US date format - 'MMMM DD YY ?HH:MM:SS?'
US date format - 'MMMM DD YYYY ?HH:MM:SS?'
US date format - 'MMMM DD, YY ?HH:MM:SS?'
US date format - 'MMMM DD, YYYY ?HH:MM:SS?'
US date format - 'DD MMM ?HH:MM:SS?'
US date format - 'DD MMM YY?HH:MM:SS?'
US date format - 'DD MMM YYYY?HH:MM:SS?'
US date format - 'DD MMMM ?HH:MM:SS?'
US date format - 'DD MMMM YY ?HH:MM:SS?'
US date format - 'DD MMMM YYYY ?HH:MM:SS?'
US date format - 'MM/DD ?HH:MM:SS?'
US date format - 'MM/DD/YY ?HH:MM:SS?'
US date format - 'MM/DD/YYYY ?HH:MM:SS?'
US date format - 'DD/MM ?HH:MM:SS?'
US date format - 'DD/MM/YY ?HH:MM:SS?'
US date format - 'DD/MM/YYYY ?HH:MM:SS?'
Where:
MM is a 2-digit month value
MMM is a 3 character month name
MMMM is a full month name
DD is a 1 or 2 digit day
YY is a 2 digit year > 100 AD, where > 50 -> 19xx and <= 50 -> 20xx
YYYY is a 4 digit year
The output is always limited to the start of the Unix time epoch
- 1970/Jan/01 00:00:00 (0)
[clock scan] can be emulated for various date formats using [scan] and
[date],
Example:
# Scan an ISO date into [date] format
scan 20161122 %4i%2i%2i y m d
puts "$y/$m/$d"
Result:
2016/11/22
Checks are made only on each individual parameter. No check is made for
the validity of the entire date string.
See: clock scan workaround
Example:
set a [clock scan "Jan 12, 1992"]
set b [clock scan "Jan 12, 1997"]
puts $a
puts $b
puts "Difference is [- $b $a]"
Results:
695174400
853027200
Difference is 157852800 second(s)
The result of [clock scan] can be converted from a "C" integer time value
to double using [ctime_to_double]. This value can then be manipulated as
an ISO date string using [date]
Example:
set b [clock scan "Jan 12, 1997, 23:59:12"]
puts [date [ctime_to_double $b]]
Result:
1997/01/12 23:59:12
set q [yyyy-mm-dd "2018-01-05 23:59:01"] # "2018-01-05 00:00:00"
puts $q
puts [date [ctime_to_double $q]]
1515196741
Error Codes
-----------
1 Bad slash separator count
2 Day value < 1 || > 31
3 Month value < 1 || > 12
4 Missing value field
5 Resulting date is invalid
6 Unhandled error
See also: clock scan workaround, clock, clock_split, date, time, scan, is_dst
Clock Scan Workaround |
Index |
[clock scan] is not fully-implemented as the date inputs are too widely
varying to be practical. Using the popular Tcl example code we can
demonstrate an easy workaround using [scan] and [date]
# We replace: set bookSeconds [clock scan $halBirthBook]
set h 00
set i 00
set s 00
set halBirthBook "Jan 12, 1997 01:02:03"
set count [scan $halBirthBook "%s %s %s %02d:%02d:%02d" m d y h i s]
puts "Scanned $count item(s)"
set d [string trim d ,]
puts "date string $y/$m/$d $h:$i:$s"
set bookSeconds [date $y/$m/$d $h:$i:$s -ctime]
puts "bookSeconds=$bookSeconds"
newline
# We replace: set movieSeconds [clock scan $halBirthMovie]
set h 00
set i 00
set s 00
set halBirthMovie "Jan 12, 1992 23:59:01"
set count [scan $halBirthMovie "%3s %s %s %02d:%02d:%02d" m d y h i s]
puts "Scanned $count item(s)"
set d [string trim d ,]
puts "date string $y/$m/$d $h:$i:$s"
set movieSeconds [date $y/$m/$d $h:$i:$s -ctime]
puts "movieSeconds=$movieSeconds"
newline
set diffseconds [expr $bookSeconds - $movieSeconds]
puts "The book and movie versions of '2001, A Space Odyssey' had a"
puts "discrepancy of $diffseconds second(s) [/ $diffseconds 86400] day(s)\
in how"
puts "soon we would have sentient computers like the HAL 9000"
Based on example from:
https://www.tcl.tk/man/tcl8.5/tutorial/Tcl41.html
See also: clock, clock scan, clock split, date, time, scan
[clock split seconds ?days|-? ?hours|-? ?minutes|-? ?seconds|-?]
Splits a number of seconds into variables containing an absolute number of
days, hours, minutes or seconds
Each variable argument is optional but at least one variable argument must
be present. Arguments must be in order of days, hours, minutes and seconds.
Arguments may be skipped by using a "-" placeholder
Variable arguments must not exist
Example:
puts [clock split 3601 d h m s]
puts "$d day(s) $h hour(s) $m minute(s) $s second(s)"
Result:
4
0 day(s) 1 hour(s) 0 minute(s) 1 second(s)
Example:
puts [clock split 86461 d h m s]
puts "$d day(s) $h hour(s) $m minute(s) $s second(s)"
Result:
4
1 day(s) 0 hour(s) 1 minute(s) 1 second(s)
Example:
puts [clock split [now -ctime] d h m s]
puts "$d day(s) $h hour(s) $m minute(s) $s second(s)"
Result:
4
17522 day(s) 12 hour(s) 42 minute(s) 49 second(s)
See also: clock, clock scan, time, date
close handle
file close handle
Close a file-handle previously returned from the open command
Handle must be > 0 and a valid stream handle
[puts] may be used to write to the file
Example:
set fp [open "input.txt" w+]
puts $fp "test"
close $fp
See also: open, gets, file, readfile, writefile
Clear text from the current cursor position to the end of the line
Example
cls
gotoxy 1 1
puts $qbf # Output some text which we will clear
gotoxy 1 1
puts "Hello" -nonewline
clreol # Clear the previous text
newline
cls ?character|{}? ?background-colour?
Clears the console window, optionally to a given character or colour
The default screen-clear character is a space. A character value is
required in order for the background colour to have effect
Example:
cls [chr 176] darkcyan
See also: console
cmdcount (Command Count) |
Index |
integer [cmdcount ?-reset?]
integer [cmdcount script|command ?var?]
Similar to [catch], [cmdcount] will either show the current count of
commands executed or will execute a command or script and return a count
of the number of commands which were executed. Along with the [time]
command this may be useful for profiling and optimising Tcl scripts
The '-reset' argument can be used to reset cmdcount
Example:
puts [cmdcount] # Show the current command count
Result:
1
Example:
puts [cmdcount {puts "Hello"}] # Command cost of running this command
Result:
1
Example:
puts [cmdcount {run test.tcl}]
Result:
2786
See also: debugging, time, trace, catch
bool [cmp string1 string2 ?start-pos? ?len? ?-nocase?]
Efficient compare function. Avoids the need to extract string segments for
comparison by comparing directly within the string.
Locates string2 within string1 dependent on string1 start position and
comparison length
The start position is indexed at a base 1 offset in string 'string1'.
i.e. 1 indexes the first character, 2, the second etc.
A case-insensitive comparison can be performed using -nocase
[sc] (string compare) is a similar command but which returns an index value
Example: (Detect C style opening comment)
option expression off
# Variable 'data' is read from a file of 'count' characters in length
for {set i 1} {< $i $count} {incr i} {
if { cmp data "/*" $i 2 } {
incr i 2
set in_comment $true
}
}
See also: cat, eq, ne, efficiency
The following coding standards have been used when developing Ticol Tcl
Methodology
-----------
A 'test-driven' methodology has been used. Code modules are developed
against small test routines, not all of which form the release package
test scripts
Language
--------
The following C/C++ language standard library functions have been avoided
at all cost and alternative 'safe' versions written:
Avoided Alternative used Explanation
---------------------------------------------------------------
gets getstr Buffer size specification
sprintf _snprintf Buffer size specification
strcpy strcpyn Null termination guarantee
strncpy strcpyn Null termination guarantee
strcat strcatn Null termination guarantee
strncat strcatn Null termination guarantee
Leak tests have been performed on modules where feasible
See also: ticol, why
string [comma string ?separator-char? ?-round N?]
Comma-formats a numeric value. Optionally, the separator character may be
customised and a rounding value may be specified.
Example:
puts [comma 1234567890]
puts [comma 1234567890 +]
puts [comma 123456.7890 / -round 2]
puts [comma 123.4567890 -round 2]
Result:
1,234,567,890
1+234+567+890
123/456.7890
123.45
Example:
puts [comma 1234567890 " "]
Result:
1 234 567 890
Example:
puts "[comma [/ [filesize "ticol.exe"] 1024]] Kb"
Result:
214 Kb
See also: format
list [commands ?filter?]
Returns a list of all known/registered commands and procs
A filter may be used to select by filter prefix match
Example:
commands sc
Result:
scan screen
[is_command] may be used to test for the presence of a defined command or
proc. A count of commands registered may be obtained using:
puts [lcount [commands]]
Use: option echo on to display directly to the console
See also: Tcl commands, commands by function, info, is_command,
vars, help, exit
Using Ticol at the Command Line |
Index |
ticol.exe filename[.tcl] [arg ...] [/switches]
ticol.exe /expr:<expression> [/switches]
ticol.exe ; [statement ...] [/switches]
Running Tcl scripts from the Windows command-line:
If the first argument to ticol.exe is a Tcl script then this will be loaded
and executed after running any autoexec.tcl file (unless /NA is specified)
Any subsequent arguments can be accessed from the argv() array
See: argv
Using Ticol as a command-line expression-evaluator from Windows:
You can call Ticol with Tcl commands if you prefix them with a semi-colon
character as the first command-line argument and quote the remainder when
necessary. This avoids Ticol attempting to run a Tcl script file
ticol.exe ; "puts [+ 1 1]" ;
Embedded double quotes may be escaped in Windows using \" e.g.
ticol.exe ; "puts [expr \"22/7.0\"]" ;
or you may use the /EXPR command-line argument. This will evaluate
and return the result by calling [expr]
ticol.exe ; "/expr:22/7.0"
The semi-colon variant can handle multiple arguments in the same
interpreter context:
Example:
ticol.exe ; "set a 4" "printf \"Pi is %3.2f\r\n\" \
[calc round(4*atan(1),2)]" "puts [expr $a*3]" /na
Result:
Pi is 3.14
12
Examples:
ticol ; "load test.tcl; run"
ticol.exe ; "puts [expr \"22/7.0\"]" ; "puts \"hello world\"" ; "for \
{set i 0} {< $i 3} {++ i} {puts $i}"
Result:
<runs script test.tcl and then returns to the O/S>
3.142857142857143
hello world
0
1
2
Expressions with embedded spaces, commands etc. must be wrapped in
double-quotes. Embedded double-quotes may be escaped
ticol "/expr:[cube 3]" /na # Quoted works OK
ticol /expr:[cube 3] /na # Unquoted does not
ticol "/expr:\"Hello world\"" /na # puts not required here
Calling from Windows Batch Script FOR Command
---------------------------------------------
The Windows FOR command will require usebackq and backtick wrappers to use
double-quotes around any Tcl /EXPR command which contains spaces
for /F "tokens=* usebackq" %%i in (`ticol.exe "/expr:round(22/7.0,4)"`) do (
set RESULT=%%i
)
echo Ticol expression result for 22/7.0 is %RESULT%
for /F "tokens=* usebackq" %%i in (`ticol.exe "/expr:4*atan(1)"`) do (
set RESULT=%%i
)
echo Ticol expression result for 4*atan(1) is %RESULT%
/EXPR may also be used to call external __stdcall or __cdecl DLLs ...
for /F "tokens=* usebackq" %%i in (`ticol /expr:"[calldlldbl_cdecl \
msvcrt40.dll sqrt 10.0]" /na`) do (
set RESULT=%%i
)
echo Result for [calldlldbl_cdecl msvcrt40.dll sqrt 10.0] is %RESULT%
The Windows FOR command requires no CRLF to be issued. You can add a CRLF
to the output by adding the /CRLF argument to ticol.exe
See also: command line, command line arguments, Ticol, commands
string [command ?arg...?]
A Tcl command is some alpha string which is most often contained within
square brackets
If a known/registered command string is met outside of a quoted
string it will be executed without the need to be wrapped in square
brackets. Thus, in
puts "Hello world"
the command, [puts] does not need to be wrapped in []
Unless evaluation is delayed by wrapping in braces, commands are
evaluated recursively within strings before the string or command-
sequence is evaluated
Thus...
puts "Hello [chr [+ 80 7]][chr 111][chr 114][chr 108][chr 100]"
will display
Hello world
The innermost [+ 80 7] command is resolved first, then the result,
[chr 87], followed by [chr 111] ...
The command may have optional arguments. Most Tcl commands take arguments,
whereas a few do not. Thus [stop] takes no arguments whereas [puts] does.
Arguments act as modifiers of a command similar to 'attributes' in
certain languages such as C#
Tcl commands use "Polish" type prefix notation where an operator or
command appears first, followed by any arguments. Flow-control commands
may accept either Tcl command or natural expression syntax. This behaviour
can be controlled using the [option expression] command. Explicitly
called commands which are wrapped in [] do not require the use of
[option expression off]
The command name can be supplied via a variable which will be resolved
before the command is executed (if braces are not used to delay
evaluation)
set p puts
set q Hello
set r world
$p "$q $r"
Result:
Hello world
Example:
set p puts
set q randomstr
set r 10
$p [$q $r]
Result:
UaMLNJYcxl
See also: braces, option expression
Standard Tcl single-line comments are prefixed by the # character
Example:
# This is a comment
Standard Tcl single-line comments are both supported and are enhanced,
since a semi-colon is not required before a # comment. This is because
single-line comments are implemented by the macro preprocesor and are
removed from the script before being interpreted by the Tcl interpreter
You can add semi-colons before a # comment if you wish, these will be
ignored:
Example:
puts "A standard Tcl comment with ; follows..."; # ...here
puts "An enhanced Ticol Tcl comment follows..." # ...here
A # comment start symbol is valid only outside of quoted strings. Hash
symbols within quoted strings, even multiline-ones, are not treated as
comment start markers and may be used as literal characters.
Outside of quoted strings a literal # character may be represented by
being 'escaped' as \#.
Unquoted escape sequences should not be used for [puts] or other print
commands.
Example:
puts "This string contains a literal # character"
A literal has character can be also represented using [asc] and [chr]
as follows:
Example:
puts [asc \#] # Prints 35 (unquoted not recommended)
puts [asc "#"] # Safely quoted (recommended)
puts [chr 35]
Result:
35
35
#
It is strongly-recommended that, as a matter of style, a # comment
character is followed by a space. This will avoid accidental and
difficult to trace clashes with the Ticol macro commands
(e.g. #exec, #if, #exit)
The Tcl semi-colon+hash end of line syntax is supported but not required
# This is a comment
puts "Hello World" # This is a valid Ticol comment
puts "Goodbye" ;# This also a valid comment
puts "Quoted # char" ;# This is a valid Tcl comment too
Any information after a comment character is completely removed by the
Ticol Macro PreProcessor. To disable this use the /NP (no preprocesor)
command-line argument. Note that this also disables #ifdef etc. macros
and these must therefore be removed from the source file
A # comment should not appear after a line continuation character '\'
Example:
"The quick brown\ # This comment is illegal because of \
fox jumps over\ # line-continuation. So is this ...
the lazy dog"
Long Comments
-------------
Long comments are supported in two forms via the Macro PreProcessor.
#ifdef... #endif statements and C/C++ style /* ... */ long comment tags.
C/C++ style comments are the most efficient and both styles are processed
by the Macro PreProcessor (MPP) before the script is executed, both from
the command-line or CLI during [load] or [run] operations.
Long comments are valid within braces and outside of quoted strings.
If it is desired to place literal comments within strings, such strings
must be wrapped in double quotes rather than braces.
Example:
# Comment sequence visible as a string literal...
puts "This string contains a /* comment */ sequence"
# Comment within braced string acts as a comment...
puts {This string contains/* comment */ characters}
Results:
This string contains a /* comment */ sequence
This string contains characters
Example Tcl comments using macro variables and #ifdef commands
#ifdef comment
set a 22
set b 7.0
puts [/ $a $b]
#endif
To revert a long macro comment either delete or prefix again with a 2nd
# character ...
##ifdef comment # This ifdef statement is commented-out
#ifdef comment # This ifdef statement is active
And (perhaps heinously, for some Tcl enthusiasts) C/C++ style comments
are supported ...
/*
This is a comment, this code is ignored
So is this invalid expression [expr 1/0]
Some may not like this, but it is convenient and useful
*/
C/C++ long comments are not valid inside quoted strings and are
treated as string literals within them. Thus...
set a "Brave /* new */ world"
will result in the string ...
Brave /* new */ world
C/C++ long comments may be ignored by prefixing with a standard Tcl #
comment ...
# /*
puts "This code will be executed as it is not commented"
# */
Selective enabling of comment blocks during debugging can be facilitated
as in C++
/*
puts "This code will not be executed"
# */
C++ single-line style comments may be defined using this dummy proc:
proc // args {
# C++ style comment - do nothing. Discard all args on this line
}
Be careful of Unix-style wildcard filename/path values within unquoted
strings. It is good practise to always wrap strings in double quotes
puts [ftp $hostname ls /tools/ * -u admin -p xxxxx]
Should be:
puts "[ftp $hostname ls /tools/*.zip -u admin -p xxxxx]"
Since the unquoted '/*' in '/*.zip' will be interpreted as a long comment
start. Quoting will prevent this.
Mixed Long and Short Comments
-----------------------------
Mixed long and short comments should appear on separate lines and
apart from other statements, particularly quoted strings.
# /*
puts "This code uncommented until the leading # is removed"
# */
Dangerous Comment Examples
--------------------------
These are examples of a dangerous comments because '#if' and '#else' are
reserved Macro Preprocessor keywords. Commenting out a Tcl [if] or [else]
statement will inadvertently create a valid Macro directive
#if This line is present the MPP will execute the '#if' clause
#else This comment is also dangerous for the same reason
#exit This comment will cause the Macro Preprocessor to exit
#echo "--x--" This is bad if you intended to comment out [echo]
See also: command line, cli, Ticol
Compatibility with Standard Tcl |
Index |
Ticol Tcl was designed to be flexible and have a reasonable degree of
compatibility with standard Tcl. Standard Tcl has changed over time and has
incorporated many new features. Ticol was written for personal use by the
author and incorporates a number of changes which were desired by the author
Some of these changes reduce compatibility with standard Tcl.
One of the main design issues was a desire to improve performance by use of
a Macro PreProcessor (MPP) to optimise source code before execution. It was
decided that the # character would be an absolute comment delimiter and
therefore cannot be uses in commands such as [uplevel]. C/C++ like comments
are also available as these suited the author's preference
The Macro PreProcessor is non-Tcl-standard and is more C-like. Although
basic it can offer some advantages when developing Tcl code
Tests showed that standard Tcl command such as [+] or [-] were much faster
than the expression-handler. Perhaps because Ticol doesn't make use of any
compiler-like optimisations when interpreting the script and it leans on
the basic Picol kernel interpreter design. For this reason, a new command
[calc] was added which enables the MPP to expand expressions into inline
Tcl code. If the MPP is disabled then [calc] will be interpreted as [expr]
Some advanced Tcl features are not implemented either fully or at all.
Arrays are preferred by the author to lists and dictionaries and have been
implemented as hashtables. Performance of arrays is quite fast with Ticol
being able to handle a million array items with no particular problems and
with excellent access speed
Ticol has its own (DLL) plugin library support which is not supported by
standard Tcl. One of these plugins offers the ability to interface into
external DLL libraries (either __stdcall or __Cdecl). To support this
interface Ticol also offers a simple binary struct which enables
API/external DLL calls and struct returns. Variants are also supported
The plugin interface C++ API code is available on request so you can
extend Ticol by adding new commands
Ticol also offers many non-standard commands which can be used to improve
performance and flexibility, e.g. [funct] will allow you to call any [expr]
function directly. [funct] is used by the MPP when it expands [calc]
expressions
Other commands include: [at_exit], [base64], [cat], [chr]. [encrypt],
[ctime_to_double], [decimal_to_roman], [dump], [elevate], [inputbox]
[json_to_list], [msgbox], [shell], [spawn], [win_error], [write]
Standard Tcl may require some modifications before it will run on Ticol
Tk is not supported
See also: Non Standard Tcl Commands, Ticol, FAQ, Performance, Plugins, Tcl
Ticol DLL Stub Program (ticol_stub.exe) |
Index |
This is a Windows console program used to test interactivity with the
DLL version of the Ticol Tcl interpreter (ticol.dll)
The stub program is a very simple console app which dynamically links
to ticol.dll and runs Tcl commands. These commands may be used to
run a script file via either [run] or [source]
The syntax is as follows:
ticol_stub.exe <options> tcl commands...
where <options> are one or more of
/Q or -Q Quiet mode
/BP or -BP Enable breakpoints
/ST or -ST Enable stack trace
/RA or -RA Enable run autoexec.tcl
/DEBUG or -DEBUG Run in debug mode
Examples:
ticol_stub.exe ?-ra? ?-bp? ?-st?
ticol_stub "puts \"hello world\""
ticol_stub puts [expr 22/7.0]
ticol_stub -q run dbview locos
ticol_stub -q run tests.tcl
ticol_stub -q run ansishow ansi -fix
ticol_stub -q source hanoi.tcl 17
ticol_stub -q source run_asm.tcl maze
ticol_stub -q run test_lib_asm
ticol_stub help puts
ticol_stub source dbview.bat albums.dbf
Note that this stub application is for testing only and should not
be used for high security purposes or in a production environment
The test stub exe must not be exposed as a CGI back end application as it
accepts direct Tcl commands and could compromise a system.
See also: ticol.dll, ticol.ini
console attribute # Return the attribute at the cursor pos
console borders on|off # Enable the console window frame border
console center # Centre the console on the screen
console close # Closes the console and exits Ticol
console curx # Return the cursor x position in columns
console cury # Return the cursor y position in rows
console get_attributes # Get attrbutes as a base64 block
console get_text # Grab the current console as plain text
console get_title # Retrieve the console title (ANSI only)
console height # Return the console height in rows
console hidden # Returns a boolean 1 if hidden, 0 if not
console hide # Hides the console
console hidecursor # Hide the console cursor
console maximize # Maximize the console window
console minimize # Minimize the console window
console move x y ?w? ?h? # Move or resize the console window (pix)
console noborders # Disable the console window border
console scroll x y w h ?n? # Scroll a console area by 'n' lines
console set_attributes string # Set attributes saved by get_attributes
console set_title string # Set the console title (ANSI only)
console setfocus # Set focus on the console window
console show # Shows the console
console showcursor # Show the console cursor
console size w h # Resize the console columns and rows
console width # Return the console width in columns
Show or hide the console window or check if the console is already hidden.
Note: if you hide the console window you will lose control of the Ticol
CLI
The command 'console hide' should be run only from a running script or you
may lose control of the console. Hiding the console may be useful for CGI
web scripts
Setfocus has limited use and may be blocked by recent versions of Windows
[console scroll] accepts an optional lines argument. Currently 1 (default)
scrolls the contents of the area up and -1 scrolls the window content
downwards
[console set_title] will only have effect during script operation. The CLI
will reset the console title on script exit. The console may only be set
to ANSI (ASCII) string values. Paths should be unescaped for set_title
Both [console get_title] and [console set_title] will return a boolean
indicating success or failure. [console set_title] does not unescape
input strings, you must use [unescape]
Usage:
console hide
console show
console setfocus
Example:
console set_title "Ticol Demo"
puts "Hiding the console..."
console hide
puts "Console hidden? [bool [console hidden]]" # Not visible until show
sleep 3000
console show
puts "I should be back now"
Result:
The console window will disappear for 3 seconds then reappear
Example:
# Smooth scroll an area down by 1 line, 5 times
for {set i 0} {< $i 5} {++ i} {
console scroll 26 3 50 4
sleep 80
}
# Smooth scroll an area up by 1 line, 5 times
for {set i 0} {< $i 5} {++ i} {
console scroll 26 3 50 4 -1
sleep 80
}
See also: screen, screen capture, CLI, box, gotoxy, console colours
Console Screen Capture and Restore |
Index |
A console screen can be saved and restored as follows
Note that the console must not be scrolled by restoring text before
any attributes are restored
# Save the text
writefile console.txt [base64 [escape [console get_text]] \
-encode -wrap 100]
# Save the attributes
writefile console_attribs.txt [console get_attributes \
-wrap 100]
cls
pause
# Restore
set console_text [readfile console.txt]
set console_attribs [readfile console_attribs.txt]
string_to_array [base64 $console_text -decode] a
# Restore the text
loop i 0 [array size a] {
printf $a($i) # printf will not add CRLFs
}
# Restore the attributes
console set_attributes $console_attribs
See also: console, console get_text, console get_attributes
The standard console colours can be customised. This may be necessary as
some colours in the standard Windows console palette are not represented
well. These also may not match those interpretations on very old PC system
For example, orange appears as khaki brown for some reason on some
versions of Windows.
Example: (Set console "DarkYellow" to be orange)
Save the following file as: console-orange.reg
To load it, double click and accept the prompts
---------------------------------------------------------------
Windows Registry Editor Version 5.00
; Colour table with DarkYellow == Orange 0x00c0c0c0 (BGR format)
[HKEY_CURRENT_USER\Console]
"ColorTable06"=dword:00508dfc
---------------------------------------------------------------
Note that the file may only be valid for the current session and unless the
defaults are changed the colours will revert after the next reboot
Colour changes made via the registry have no effect until a new console
window is opened
See also: console, big graph
const varName value
Define a simple constant variable. You cannot assign to a constant
The use of [const] is preferable to macro #define
Const values will typically be created in global scope and can be accessed
from within procedures using the $::constname scope prefix. Or they may be
declared using upvar to create a local const instance
Constants can be deleted using [unset]
Const arrays can be set using [array set arrayname -const]
Example:
const ONE_GIGABYTE 1073741824.0
puts $ONE_GIGABYTE
puts [bool [is_const ONE_GIGABYTE]]
proc foo {} {
puts "foo thinks 1Gb is: $::ONE_GIGABYTE" # :: scope operator used
}
foo
unset ONE_GIGABYTE -const
puts [bool [is_set ONE_GIGABYTE]]
Results:
1073741824.0
True
foo thinks 1Gb is: 1073741824.0
False
See also: enum, set, array set
string [content string left-anchor right-anchor]
Extracts and returns the portion of a string bounded by the left and right
anchor characters, not including the anchor character.
Example:
puts [content "abc(12345)def" ( )] # Returns '12345'
Note that to pass braces or whitespace characters as anchor characters you
must either enclose these in double quotes, or specify as [chr] values
puts [content "abc{12345}def" "{" "}"] # Returns '12345'
puts [content "abc{12345}def" [chr 123] [chr 125]] # Returns '12345'
See also: string
integer [copyfile sourcespec target ?-y?]
Copy a file from sourcefile to targetfile. Wildcards are supported
Path targets or specified filename targets are supported
Paths accept \ or / separators. Where a trailing -y is used from the CLI
then a trailing / separator must be used to signify a path as the backslash
will be interpreted as a line-continuation escape character (\).
Where backslashes are used then take note that the Macro PreProcessor (MPP)
will require these to be escaped as double backslashes (\\) e.g. "c:\\path"
Where wildcard+forward-slash combinations are present, be sure to wrap
any filename in double-quotes to avoid /* being interpreted as a comment.
The count of successful copies is returned else 0 on failure
CAUTION: Supports only ANSI filenames, not Unicode/MBCS
Example:
copyfile ticol.exe c:\temp\ticol.exe -y
copyfile ticol.exe c:/temp/new Copy to target file /temp/new
copyfile c:\\tcl\\*.tcl c:\\temp Existing directory
copyfile c:\\tcl\\*.tcl c:/temp/new Will fail as '/new' is a file
copyfile c:\\tcl\\*.tcl c:/temp/new/ Force create '/new/' directory
See also: movefile, file
Copyright and Licensing |
Index |
---------------------------------------------------------------------------
Ticol Copyright and Licence
---------------------------------------------------------------------------
Copyright (C) to M Shaw 2010-2023
All rights reserved.
Portions Copyright (c) 2007-2016, Salvatore Sanfilippo
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
1 Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
2 Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution
3 This product and its associated documentation may not be sold or
resold for profit
4 This licence does not extend to the Ticol Tcl manual
The Ticol Tcl manual, associated documentation and sample scripts
are not included in this licence but are covered separately by the
standard international author copyright for literary works
Copyright (C) M Shaw (2015-2023)
All rights are reserved
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
TO,THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
---------------------------------------------------------------------------
Licence Summary
---------------
As with the Info-ZIP licence, this modified BSD licence has additional
clauses added to protect the author:
Clause 3 is to avoid the author having to encounter the downstream
cost of commercial support
Clause 4 is to ensure that the literary rights to the documentation
are fully protected to ensure that it is not edited, repackaged or
sold by a 3rd party
A modified licence may be issues on request and only by agreement with
the author
This project was initially derived from a micro-project called "Picol",
by Salvatore Antirez Sanfilippo, which implemented a Tcl interpreter
in 500 lines.
This project is closed source and the source code is not distributed
There is no obligation under the BSD licence to distribute source code
The original 500-line Picol source is freely available from
http://oldblog.antirez.com/post/picol.html
This documentation (manual), in whatever format distributed, is an
original creative work by the author, It is licensed separately from
the program binaries and protected by standard international literary
copyright. It may be distributed with the program free of charge,
but may not be sold, resold, modified or incorporated into other works
Copyright (C) to M Shaw (2015-2023)
All rights reserved
To the best of the author's knowledge, no GPL code is used in this project
Portions of certain plugin code may contain code attributable to other
authors including but not limited to the following code:
For more information, see any accompanying readme.txt for a full list
ticol_zip plugin, portions
Copyright (c) 1990-1999 Info-ZIP. All rights reserved
The plugin library is not supplied by Info-ZIP
ticol_sqlite plugin, portions
Copyright attributed to SQLite (Public Domain)
Acknowledgements are made for the following code
MD5 Algorithm: RSA Data Security, Inc. MD5 Message-Digest Algorithm
Copyright (C) 1991-2, RSA Data Security, Inc. 1991.
All rights reserved
Mersenne Twister Random Number Generator
Copyright(C) 1997-2002, Makoto Matsumoto, Takuji Nishimura
TCP/IP specific extensions in Windows Sockets 2
Portions Copyright(C) 1980,1983,1988,1993
The Regents of the University of California
The Picol project source code has been replaced with around 100,000
lines of C++ code. This source code is not available but the program
binaries will be made available for both private and commercial use as
freeware. Should any disputes arise the author reserves the right to
withdraw the Ticol program from further public distribution
Please read the LICENCE.TXT or README.TXT file for full licence terms
In summary:
No warranty of fitness for use or any other type of warranty is offered.
By using Ticol you agree to use it at your own risk and indemnify the
author against all liabilities arising as per the licence agreement.
Ticol must not be used in a commercial environment without adequate and
thorough testing
You may freely use Ticol whether for private or commercial use
You may give individual copies of Ticol to friends or colleagues
You may not sell Ticol or any accompanying documentation
You have no right whatsoever to technical support or fixes
You agree to satisfy yourself of any fitness for use regarding Ticol
You accept this as software offered free of charge and in good faith
You agree that this software is not warrantied in terms of fitness for
any purpose whatsoever
You accept that this is a personal/hobby/educational project and you hold
no commercial or professional expectations of it either in terms of
reliability, fitness for use, performance or levels of support
You agree to use Ticol entirely at your own risk and indemnify the author
from any and all liabilities arising from its use
You agree not to use this product in any mission critical systems where
life or financial loss is at risk
You must include the BSD copyright notice/licence terms with any copies
you pass on to a 3rd party
See: Ticol, why, frequently asked questions
C or "Unix" time is defined as a signed integer representing the number
of seconds elapsed since 00:00:00 hours on January the 1st 1970
Dates prior to this cannot be represented
Dates after 03:14:07 UTC on 19 January 2038 cannot be represented
The Visual BASIC method is to use a double value to represent time
Example:
puts [clock seconds]
Result:
1449684929
See:
https://en.wikipedia.org/wiki/Year_2038_problem
Convert to a string format using the [clock format] command
See also: vbtime, date_to_ctime, double_to_ctime, ctime_to_double,
clock, time, now
ctime_to_double, double_to_ctime |
Index |
double [ctime_to_double seconds]
integer [double_to_ctime double]
[ctime_to_double] converts the standard internal Tcl "C" time format from
seconds in the Unix epoch of 01/Jan/1970 to an industry-standard or Visual
BASIC double value. [double_to_ctime] performs the inverse operation
Example:
puts [clock format [double_to_ctime [now]] -format "%Y/%m/%d %H:%M:"
Example:
puts [double_to_ctime [now]]
puts [clock seconds]
Result:
1449684929
1449684930
See also: date, time, is_date ,date_to_ctime, clock, clock format, now,
ctime
[cube] and [cuberoot] are not implmented internally but can easly be
implemented as Tcl procedures. The following examples may be used.
Note that when dealing with double/float precision numbers, accuracy
can be an issue and rounding may be required (due to binary to decimal)
conversion.
Prefix notation is slightly faster than using [expr] for math.
proc cube {x} {
return [round [* $x [* $x $x]] 5]
}
proc cuberoot {x} {
set start 0
set end $x
set mid [/ [+ $start $end] 2.0]
while {!= [round [* $mid [* $mid $mid]] 4] $x} {
set mid [/ [+ $start $end] 2.0]
if {< [* $mid [* $mid $mid]] $x} {
set start $mid
} elseif {> [* $mid [* $mid $mid]] $x} {
set end $mid
} else {
return [int [round $mid 5]]
}
}
return [round $mid 5]
}
See also: math functions
VB (Visual BASIC) Time |
Index |
Visual BASIC stores Date values as an IEEE 64-bit (8-byte) floating point
number, identical to a VB Double. Digits to the left of the decimal point,
when converted to decimal, are interpreted as a date between 1 January 100
and 31 December 9999.
Values to the right of the decimal indicate a time between 0:00:00 and
23:59:59 and is significant to 5 decimal places at 1 second resolution
The date section of the Date data type, is a count of the number of days
that have passed since 1 Jan 100, offset by 657434. That is, 1 Jan 100 is
denoted by a value of -657434; 2 Jan 100 is denoted -657433; 31 Dec 1899 is
1; 1 Jan 2000 is 36526 etc.
The time section of the date (after the decimal point) is the fraction of a
day, expressed as a time. For example, 1.5 indicates the date 31 Dec 1899
and half a day, i.e. 12:00:00. So, an hour is denoted by an additional
4.16666666666667E-02, a minute by 6.94444444444444E-04, and a second
by 1.15740740740741E-05
The Ticol [date], [dateserial] and [time] commands understand Visual BASIC
format doubles representing date values
See also: date, dateserial, time, ctime, now, date_to_ctime
See:
is_active procname # Test if a proc is currently active
is_admin # Test if user has administrative rights
is_array # Test if a variable name is an array
is_command string # Is a string a registered command
is_const string # Is a string a constant variable
is_date string # Is a string a valid formatted date or time
is_dir path # Is a string a valid directory
is_dst # Test if we are in Daylight Savings Time
is_elevated # Test if we are running with elevated rights
is_empty var # Test if a variable is both set and empty
is_file name # Test a file
is_leapyear string # Is a date value a leap year
is_list string # Test if a string is a well-formed Tcl list
is_mod x # Optimised modular division test
is_numeric x # Is a string a valid number
is_pointer x # Is a variable a pointer (struct member)
is_proc name # Test if a string is a defined proc
is_set var # Test for a valid/existing variable
array exists a # Tests if an array exists
array is_set a # Tests if an array element is set
# C++ Ctype Emulation (Applies to whole strings)
is_alpha # Test if all A..Z or a..z
is_alnum # Test if all A..Z or a..z or 0..9
is_digit # Test if all 0..9
is_lower # Test if all chars are lower case alpha
is_punct # Test if all chars are punctuators
is_space # Test if all chars are space or tab
is_upper # Test if all chars are upper case alpha
is_xdigit # Test if all chars are A..F, a..f or 0..9
All the above commands return type "bool" (0 | 1)
See also: array exists, array is_set
bool [is_array varname]
Test if a variable name is an array variable
dim a 10 # Empty array
set b Hello # Simple variable
array c {a 10} # Array
struct set d {a 10} # Struct
puts [is_array a]
puts [is_array b]
puts [is_array c]
puts [is_array d]
puts [is_array e] # e does not exist
Results:
1
0
1
0
0
See also: is_*, is_set, array exists, varname
integer [is_active procname]
Test to see if a proc or a proc's parent proc is active
[is_active] returns the following values
0 If neither are active
1 If the current proc is active
2 If the current proc and a proc's parent is active
The integer return from [is_active] may be interpreted as a boolean in
most situations. [is_active] may be used to test before calling [undef]
or [rename] safely
Example:
proc foo {} {
puts [is_active foo]
}
foo
Result:
1
See also: proc, rename, undef, is_*
bool [is_dst ?-bias? ?-offset? ?-standardname? ?-daylightname?]
[is_dst] returns a boolean (1 or 0) indicating whether we are currently
in Daylight Savings Time. The actual time offset is not returned
The '-bias' argument will return the local geographic time zone bias from
UTC (GMT) in minutes.
The '-offset' argument will return the local Daylight-Savings-Time (DST)
offset in minutes
The '-name' argument will return the name of the time zone
Example:
# US Central time (-6hrs) will return -360 (minutes)
puts [is_dst -offset]
Result:
-360
Example:
# UK summer time (+1hr) will return 60 (minutes)
puts [is_dst -offset]
Result:
60
Example:
puts [is_dst -standardname]
Result:
GMT Standard Time
Example:
puts [is_dst -daylightname]
Result:
GMT Daylight Time
See also: is_*, time, date, clock, ctime_to_double, double_to_ctime
bool [is_date double]
bool [is_date date|time]
bool [is_date date]
bool [is_date time]
bool [is_date "date time"]
bool [is_date "time date"]
Performs basic sanity checks in a date-interpreted double value or string(s)
Examples:
puts [is_date 0] # True
puts [is_date -1] # True
puts [is_date 42657.0430902778] # True
puts [is_date 2016/Oct/14] # True
puts [is_date 2016/10/14] # True
puts [is_date 00:00:59] # True
puts [is_date 2016/10/14 01:02:03] # True
puts [is_date 01:02:03 2016/10/14] # True
puts [is_date "2016/10/14 01:02:03"] # True
puts [is_date {2016/10/14 01:02:03}] # True
puts [is_date 9999999.0430902778] # False (out of range)
puts [is_date "00:00:59 2016/oct/39"] # False (invalid day)
See also: is_*, date, dateserial, vbtime
bool [is_file filename]
Currently checks for the existence of a file and returns a boolean value
indicating whether the file exists or not
Examples:
puts [is_file ticol.exe]
puts [is_file nosuch.txt]
Results:
1
0
See also: file, is_*
bool [is_leapyear year]
Confirms whether a year value is a leapyear or not.
Valid from 1 AD to 9999 AD
Returns a boolean (integer) 1 or 0
The date of the 1582 Pope Gregory declaration is used as the calculation
threshold
Common algorithms which use leap years to calculate are not valid for
years in the Julian calendar before 1752 in the British Empire and US.
The year 1700 was a leap year in the Julian calendar, but not in the
Gregorian calendar.
Example:
puts [is_leapyear 0001]
puts [is_leapyear 0004] # Is a leap year
puts [is_leapyear 4] # Is a leap year
puts [is_leapyear 1896] # Is a leap year
puts [is_leapyear 2001]
puts [is_leapyear 2012] # Is a leap year
puts [is_leapyear 9996] # Is a leap year
puts [is_leapyear 9999]
Results:
0
1
1
1
0
1
1
0
See also: is_date, is_*
bool [is_list string]
Test if a string is a valid Tcl list
Example:
puts [is_list {}]
puts [is_list ""]
puts [is_list a]
puts [is_list {a b]
Results:
0
0
1
1
See also: is_*, list, lindex, lcount
bool [is_static varname]
Tests if a given variable is a protected static variable
Static variables are protected from deletion using [unset] and require the
use of [static varname -unset] to remove
Static variables are intended to persist for the life of a script but there
may be certain circumstances where a variable needs to be tested for and
selectively removed
See also: static, is_*
date ?-separator char? # Return the current date as YYYY-MM-DD
# Use [date -separator {}] for YYYYMMDD
date -year # Return the current year as YYYY
date -month # Return the current month as M
date -day # Return the current day as D
date -yday # Return the current day of the year as D
date double ?-separator char? # Returns YYYY-MM-DD + optional separator
date double ?-list? # Return each field as a Tcl list
date double ?-notime? # Suppress display of time value
time double # Use [time] to suppress date value
date double ?-longmonth? # Display 3-char month names e.g. "Jan"
date double -format <string> # Return specific date using PHP format
date now # Returns current date as a double value
date now -format <string> # Return current date using PHP format
date YYYY/MM/DD ?hh:mm??:ss? # Returns a double representing the date
date YYYY-MM-DD ?hh:mm??:ss? # Do
date YYYY/MMM/DD ?hh:mm??:ss? # Do
date YYYY-MMM-DD ?hh:mm??:ss? # Do
date DD/MM/YYYY ?hh:mm??:ss? # Do
date DD-MM-YYYY ?hh:mm??:ss? # Do
date DD-MMM-YYYY ?hh:mm??:ss? # Do
date DD-MMM-YYYY ?hh:mm??:ss? # Do
date <date-string> -ctime # Return the time as a C-long integer
dateserial YYYY MM DD hh mm ss # Use [dateserial] for dates as Tcl lists
Shows the current system date in ISO YYYY-MM-DD format
ISO date format is preferred
Returns a VB6-compatible date value as double
Or, returns a formatted date/time string
Date Ranges
Numeric inputs range from -693958.000000 (0000/01/01 AD) to
2958465.000000 (9999/12/31 AD). Units are in days and fractions of
a day as per Visual BASIC 6.0
US Date Formats
The US date formats of MM-DD-YYYY and MM-DDD-YYYY are not supported
Comma Date Formats
The date format Jan, 1, 2016 and similar formats are not supported
Input Locales
Automatic/Windows system input locales are not supported
The epoch starts the same as for VB6 at 30/Dec/1899 as value 0.0
Negative values extend as far back as January, 1st 0 AD
1/1/70 00:00:00 (Unix epoch in VB) is 25569.0 as a double
February 5, 2036 (Unix epoch end) is 49711.00001157407
(00:00:01 6th Feb 2036)
Valid Range
Dates are valid in the range from 0 AD to 9999/12/31 AD
Dates older than the positive VB6 date epoch are negative
Year values must be 4-digit. Leading zeroes are required if a date is
< 4 digits. e.g. 325 AD will be year "0325" not "325"
Values outside the of range 0/00/00 to 9999/12/31 (-693959.0 to 2958465.0)
will raise an error
Similarly: puts [date [time [date now] -ctime]] is invalid because
[date] cannot take a C time value as an argument. In such cases we would
use [ctime_to_double] to convert the result from [time]:
puts [date [ctime_to_double [time [date now] -ctime]]]
Short Year Values
Short year values such as 99 for 1999 are not supported
4-digit years are mandatory otherwise 0.0 will be returned
Date Arithmetic
Since date values are "real" (floating point double) values they may be
added, subtracted, multiplied or divided with caution and providing the
number range is adhered to when reconverting back to date format using
[date]
Year/month/day values can be compared with now by using [int] to
strip off the fractional (time) part of the number
Example:
if {[eq [int [date now]] [date $year-$month-$day]]} {...}
Exercise caution when adding dates on or before the zero epoch date of
30/dec/1899 as time values will wrap-around. You may need to use [fpart]
or [fraction] to split-off the fractional part of a date, save it and
then re-add after arithmetic. This is because dates < 30/dec/1899 are
represented negative floating point and changing the sign may invert
the fractional part
When manipulating for whole days you will usually need to strip off the
fraction parts of double values to avoid rounding up with 1 day error
You may use [fpart] to strip-off and save the decimal fraction part
Example:
# Increment year from 325 AD to 1776 AD, keeping the time
# An alternate method is to return a list and increment item 0
set d [date 0325/jan/01 09:30:00] # Save full date
set t [fpart $d] # Split off and save time
set i [integral $d] # Split and save the date
puts "Stored: $d [date $d] and time as $t" # Display saved values
set r [+ [* 1452 365] $i] # Add 1452 365 day years
puts "Result: '${r}.$t' -> [date ${r}.$t]" # Concatenate and display
Result:
Stored: -575255.3958333334 0325/01/01 09:30:00 and time as 3958333333
Result: '-45275.3958333333' -> 1776/01/01 09:30:00
Example:
# Date subtraction in epoch range > 0.0 (30/Dec/1899)
# Days to 2019/01/10
puts "[int [- [date 2019/01/10] [date now]]] days"
# Days to Xmas 2016
puts "[int [- [date 2016/12/25] [date now]]] days"
Results:
813 days
67 days
Example:
# Add 8.5 years to today and display in YYYY/MMM/DD format
puts [date [+ [date now] [* 8.5 365.25]] -longmonth]
Time Arithmetic
---------------
Time values should be first converted using [date]. Next, any arithmetic
should be performed, and the result converted back using [time]
puts [date 11:58]
puts [date 5:30]
# Add 5.5 hours to 11:58
puts [time [+ [date 11:58] [date 5:30]]]
# Same action using [now]
puts [time [+ [now] [date 5:30]]]
# Same action, giving a full date
puts [date [+ [now] [date 5:30]]]
Results:
0.4986111111
0.2291666667
17:28:00
17:29:58
2017/04/14 17:30:58
PHP Date Formats
----------------
The [date now -format <string>] and [date <value> -format <string] options
permit the construction of date strings using PHP style character tags
such as "r" or "H:i:s". There is a maximum output string length of
64 characters, any excess will be truncated. Any unhandled characters will
be passed as literals e.g. the colon ":" or hyphen "-" characters
All tags other than U and b are implemented
Example:
puts [date now -format l]
puts [date [+ [date now] 2.5] -format r]
puts [date $dateval -format "D, jS F Y H:i:s"]
Results
Thursday
Wed, 20 Oct 2021 03:37:04 +0100
Sun, 17 Oct 2021 15:37:24 +0100
Exception Errors
----------------
Accessing a date value outside the valid range will trigger an exception
error. Other date errors will return 0.0
Notes - Converting from 1 second interval cdates to decimal day fractions
Multiply each second by 0.000011574074074 or [/ 1 86400.0]
Example:
# 1 second past the Unix epoch end
puts [+ 49711 [/ 1 86400.0]]
# 1 second past the Unix epoch end
puts [date [+ 49711 [/ 1 86400.0]]]
Result:
49711.000011574077
2036/02/08 00:00:01
Visual BASIC Anomalies
----------------------
In order to accommodate two digit year values, VB5 interprets
Cdbl(Fix(Cdate("0000/jan/1"))) as 36892.0; which is 01/01/2000
Dates with a year of < 100 AD are interpreted as being in the year 2000
Ticol does no such 2-digit year interpretation adjustment. Post Y2K there
is now no valid reason to use two-digit year dates within scripts.
4-digit years are mandatory in Ticol for all dates after 999 AD
More Examples Result
puts [date] 2016-10-18 # ISO date
puts [date -year] 2016 # Current year
puts [date -month] 10 # Current month
puts [date -day] 18 # Current day
puts [date -yday] ? # Day of year
puts [date now] 42661.785891 # "now" keyword
puts [date [date now]] 2016/10/18 18:51:41
puts [date 0] 1899/12/30 # Same as VB6
puts [date 1] 1899/12/31 # Same as VB6
puts [date 2] 1900/01/01 # Same as VB6
puts [date 0000/1/1] -693958 # VB6=01/01/2000
puts [date -693958.000000] 0000/01/01 # Jan, 1st 0 AD
puts [date 0001/1/1] -693593 # VB6=01/01/2001
puts [date -693593.000000] 0001/01/01 # Jan, 1st 1 AD
puts [date 42661.785891] 2016/10/18 18:51:40
puts [date 42661.785891 -separator -] 2016-10-18 18:51:40
puts [date 42661.785891 -separator " "] 2016 10 18 18:51:40
puts [date 42661.785891 -separator _] 2016_10_18 18:51:40
puts [date 42661.785891 -longmonth] 2016/Oct/18 18:51:41
puts [date 2016-10-18] 42661 # YYYY-MM-DD
puts [date 2016/10/18] 42661 # YYYY/MM/DD
puts [date 18-10-2016] 42661 # DD-MM-YYYY
puts [date 18/10/2016] 42661 # DD/MM/YYYY
puts [date 0325/01/01] -575255 # -ve (325 AD)
puts [date 1400/01/10] -182610 # Same as VB6
puts [date 1700/01/10] -73037 # Same as VB6
puts [date 1970/01/01] 25569 # Unix epoch
puts [date 2036/02/06] 49712 # Unix epoch end
puts [date 9999/02/06] 2958101 # Same as VB6
puts [date 2016-10/18] 0.000000 # ERROR
puts [dateserial 2016 10 18 18 51 40] 42661.785880 # For lists
puts [date 2016/10/18 18:51:40] 42661.785880 # 2 arguments
puts [date 18:51:40 2016/10/18] 42661.785880 # Reverse is OK
puts [date "2016/10/18 18:51:40"] 42661.785880 # Quoted is OK
puts [date {2016/10/18 18:51:40}] 42661.785880 # Braced is OK
puts [date {18:51:40 2016/10/18}] 42661.785880 # Reverse is OK
See also: time, dateserial, is_date, ticks, clock, ctime, vbtime,
date_to_ctime, fpart, clock scan workaround, clock scan,
day_of_week, format
double [dateserial ?year|-? ?month|-? ?day|-? ?hour|-? ?minute|-?\
?second|-?]
double [dateserial ?tcl-list?]
Calculates a date value and returns a double from individual arguments
Each arguments is optional with the defaults being 1 1 1 0 0 0
(1st January, 1 AD 00:00:00). Optional arguments are omitted by specifying
a hyphen character '-', using empty "" or {} arguments
[dateserial] may also be called with these items as a space-separated
Tcl list, as produced by [date double -list]
All arguments must be integer values. Values must be given in order but
may be omitted with a '-' used as a placeholder
Example:
puts [dateserial - - - 00 05 36]
puts [dateserial "" "" "" 00 05 36]
puts [dateserial {} {} {} 00 05 36]
puts [date [dateserial - - - 00 05 36]]
Result:
-693958.00388888887
-693958.00388888887
-693958.00388888887
0000/01/01 00:05:36
Example:
puts [dateserial 1899 12 30 0 0 0]
puts [date [dateserial 1899 12 30 0 0 0]]
Result:
0.000000000000000
1899/12/30 00:00:00
Example:
puts [date now] # Returns a double
puts [date [date now] -list] # Returns date as list
puts [dateserial [date [date now] -list]] # Returns a double
puts [date [dateserial [date [date now] -list]]] # Returns full date
Result:
42663.971261574072
2016 10 20 23 18 37
42663.971261574072
2016/10/20 23:18:37
Example:
puts [fraction [dateserial - - - 2 4 6]] # Extracts .nnnn
puts [time [fraction [dateserial - - - 2 4 6]]] # Extracts .nnnn
Result:
-0.086180555517785
02:04:06
See also: date, time, vbtime, format
integer [date_to_ctime year ?month? ?day? ?hour? ?minute? ?second?]
Convert string-formatted date values into a C/Unix time value as a number
of seconds since January 1st 1970. Passing a date value earlier than this
will raise an error.
The inverse command is [ctime_to_double]
Example:
puts [date_to_ctime 1972 apr 04] # Flying Scotsman Day
Result:
71193600
Example:
puts [date [ctime_to_double [date_to_ctime 1972 04 04]]]
puts [clock format [date_to_ctime 1972 04 04] -format %D%N%y]
Result:
1972/04/04 00:00:00
4472
See also: ctime_to_double, ctime, time, date, dateserial, vbtime
The following data type ranges may be coerced using data type casts
The internal storage format for numbers is string. Calculations are
performed on integers by converting to 64-bit values and real numbers by
converting to 64-bit (8 byte) doubles. However, not all routines will
handle such values. Some which link to the underlying C runtime
may be limited to 32 bit signed long values.
By default, signed values are used internally. Some commands such as
bitwise operations [bit] may use unsigned values. See the documentation
for each command
Type casts may be forced using various cast commands (See: variable
casting)
Type Bits Sign Example Comma-Range
--------------------------------------------------------------------------
char 8 signed -128 to 127
char 8 unsigned 0 to 255
short 16 signed -32,768 to 32,767
short 16 unsigned 0 to 65535
long 32 signed -2,147,483,648 to 2,147,483,647
long 32 unsigned 0 to 4294967295
int 64 signed -9,223,372,036,854,775,808 to 9,223,372,036,854,775,807
int 64 unsigned 0 to 18,446,744,073,709,551,615
double 64 signed 1.7E +/- 308 (15 digits of precision)
Example Integer Range
-------------------------------------------------------
char -128 to 127
unsigned char 0 to 255
short -32768 to 32767
unsigned short 0 to 65535
long -2147483648 to 2147483647
unsigned long 0 to 4294967295
int -9223372036854775808 to 9223372036854775807
int unsigned 0 to 18446744073709551615
double 1.7E +/- 308 (15 digits of precision)
Note that [scan] with %d, %ld etc. will not truncate excess range values
down to handled ranges
Command and Function Argument Limits
------------------------------------
Although the default internal integer range is 64-bit signed integer, most
commands are based on code exported from the underlying C library and are
thus limited to signed 32 bit values unless otherwise stated
For example, the [rnd] command takes two signed long values. Calling with a
Ticol 64-bit integer outside this range will return an invalid value.
Useful Definitions
------------------
# Refer within source without a $ prefix e.g. CHAR_MIN
# The MPP will detect within substrings e.g. MYCHAR_MIN
# const or enum is preferable
#define CHAR_MIN -128
#define CHAR_MAX 127
#define UCHAR_MIN 0
#define UCHAR_MAX 255
#define SHORT_MIN -32768
#define SHORT_MAX 32767
#define USHORT_MIN 0
#define USHORT_MAX 65535
#define LONG_MIN -2147483648
#define LONG_MAX 2147483647
#define ULONG_MIN 0
#define ULONG_MAX 4294967295
#define INT_MIN -9223372036854775808
#define INT_MAX 9223372036854775807
#define UNSIGNED_MIN 0
#define UNSIGNED_MAX 18446744073709551615
or
# Refer within source code using a $ prefix e.g. $CHAR_MIN
# Detection and parsing is more precise but slower
# const or enum is preferable to #define
const CHAR_MIN -128
const CHAR_MAX 127
const UCHAR_MIN 0
const UCHAR_MAX 255
const SHORT_MIN -32768
const SHORT_MAX 32767
const USHORT_MIN 0
const USHORT_MAX 65535
const LONG_MIN -2147483648
const LONG_MAX 2147483647
const ULONG_MIN 0
const ULONG_MAX 4294967295
const INT_MIN -9223372036854775808
const INT_MAX 9223372036854775807
const UNSIGNED_MIN 0
const UNSIGNED_MAX 18446744073709551615
Within the C language these values are defined within file "limits.h"
See also: variable casting, cast, numeric types, vbtime
Data Structures (Abstract Data Types) |
Index |
Ticol offers the following data structures and ADTs
strings natively
buffered strings natively
integers as string or 64-bit internal representation
double/floats/real as string or internal double representation
binary values as string
hexadecimal values as string
octal values as string
big integers plugin
Roman numbers plugin
compressed numbers natively via mkn, cvn
Scientific numbers native and plugin
associative arrays natively
lists as string
binary lists plugin
stacks natively
structs natively
dictionaries <feature incomplete>
line addressed strings plugin
date/time natively including VB-like dates and C dates
dynamic memory blocks natively
xbase/dbase databases plugin
Ticol databases plugin and similar to DBF format
sqlite databases plugin
C-like arrays (1, 2 or 3D) plugin with command (v's $) addressing
Variant array plugin for DLL interfacing
Memory block access is available via malloc/free which (in theory at least)
could offer the ability to create other complex ADTs.
Ticol does not offer ...
bit-limited types - char, short etc. (all integers are 64-bit)
complex or nested binary structs
classes (other than very simple emulation)
objects/inheritance
trees
counted/binary strings (other than in a very limited manner)
See also: data type ranges
integer [day_of_week day month year]
Returns the day of the week for a given day in a given month and for a
given year. All arguments are required. Valid for range 1000 AD to 9999 AD
The values are all 'base 1' e.g. day 1..31, month 1..12, year 1..2017+
The return value is 'base 0' index where 0==Sunday, 1==Monday etc...
[day_of_week] can be used to construct a Gregorian calendar
Calculations are based on the introduction of the Gregorian calendar in
1582 not 1752
Common algorithms which use leap years to calculate are not valid for
years in the Julian calendar before 1752 in the British Empire and US.
The year 1700 was a leap year in the Julian calendar, but not in the
Gregorian calendar.
Example:
# 23rd December 2017 is a Saturday (day 6)
puts [day_of_week 23 12 2017]
puts [day_of_week 4 4 1972]
puts [day_of_week 21 2 1804]
Result:
6 # Saturday
2 # Flying Scotsman Day was a Tuesday
2 # The first steam locomotive ran on a Tuesday
[day_of_week] can be used to create an extended-range Gregorian
calendar. See calendar.tcl example file
Example:
JANUARY 1172
Sun Mon Tue Wed Thu Fri Sat
1
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31
# 1st January is a Saturday (Gregorian)
# Validation Reference:
# https://moodle.lse.ac.uk/calendar/view.php?\
view=month&course=1&time=-25182489600
JANUARY 14001
Sun Mon Tue Wed Thu Fri Sat
1 2 3 4 5 6
7 8 9 10 11 12 13
14 15 16 17 18 19 20
21 22 23 24 25 26 27
28 29 30 31
# Validation reference:
# http://elearning2.msbm-uwi.org/calendar/view.php?\
view=month&course=1&time=379661749200
References:
https://calendar.zoznam.sk/nameday-enathor.php
See also: date, time, clock
debug ?on|off|print? ?-quiet?
Turn on/off debug tracing or view the current setting or print to the
console.
[debug] with no arguments issues a verbose status unless -quiet is used.
In all cases a boolean 1 or 0 is returned indicating the current debug
status
[debug -quiet] can be used to programmatically detect debugging status
and act accordingly
if { [debug -quiet] } {
debug off
}
[debug print] or [debug print string] is a command which prints to the
console unless a ...
#define ndebug
statement is issued. In which case all [debug print] statements
are disabled in the script, regardless of where the #define occurs
Because this is an interpreter, not a compiler and because the MPP runs
before runtime interpretation the '#define ndebug' statement can appear
anywhere and will have an effect on the whole file
If selective output is required then do not define 'ndebug', instead
define a different macro variable and wrap statements in #ifdef
Example:
debug print "Hello from debug"
Result:
Hello from debug
Example:
debug print "This statement will not appear"
#define ndebug
debug print "Nor will this"
Result:
<no output>
Example:
#define debug_print
#ifdef debug_print
debug print "I will appear"
#else
debug print "I won't appear"
#endif
Result:
I will appear
See also: trace, debugging, breakpoints, halt, time, debugging, commands
Useful Debug Routines |
Index |
Two useful routines are offered to aid testing and debugging, [check] and
[checkex]. [checkex] is an extended form of [check] which allows for more
complex expressions in the test rather than just simple equivalence
[assert] offers a built in alternative
# Use: set line #__LINE__ at parent level
if {! [defined checkex] } {
proc checkex {x expr} {
upvar line
set NAN "-1.[calc [chr 35]]IND00000000000"
set q {$x}
puts "Line $line: " -nonewline
puts "'$q'=='$x': '$expr'"
if { [eval $expr] } {
# puts "Pass"
} else {
textcolor red
puts "checkex assertion failed: '$q' -> '$x' at line $line"
textcolor
stop
}
}
}
if {! [defined check] } {
proc check {x y} {
if {!= $x $y} {
puts "Error in test at after line $::line"
puts "'$x' != '$y'"
stop
}
}
}
Examples:
check [calc "ceil(2.8)"] 3
checkex [comma 1234567890] {eq $x "1,234,567,890"}
Results:
line 3348: x:[3.000000000000000] == y:[3]
Line 3348: '$x'=='1,234,567,890': 'eq $x "1,234,567,890"'
See also: assert, debug, faq, debugging
default (Placeholder) |
Index |
See: proc for default proc arguments
See: switch for default switch clause
bool [defined command|proc|function ?-command? ?-proc? ?-funct?]
Returns a boolean 1 or 0 indicating whether the given command or procedure
is defined. The [defined] command is useful in allowing scripts to be re-
run after a stop event without encountering a 'procedure already defined'
error.
Defined procs and built-in commands (which have not been undefined) share
the same calling space.
Valid flag arguments:
<none> Test for either a command or defined proc
-command Test only for a built-in command (which has not been undefined)
-proc Test only for a registered proc
-funct Test only for a math function which can be called by [funct]
Example:
puts [defined puts]
puts [defined foo]
Result:
1
0
The default is to detect any command or proc but not a function as
functions cannot be called directly and present a minimal need to be
detected
Example:
if {! [defined ipfromhost]} {
proc ipfromhost {hostname} {
return [calldllstr ping32 GetIPFromHostName $hostname]
}
}
Example:
puts [bool [defined puts]]
proc bar {} {}
puts [bool [defined bar -proc]]
puts [bool [defined foo]]
puts [bool [defined abs -funct]]
puts [bool [defined quux -command]]
Result:
true
true
false
true
false
See also: undef, commands, is_set, funct, proc
dict (Dictionaries) |
Index |
Simplified Tcl Dictionaries. (Incomplete Feature)
Note: Ticol dictionaries are not currently 'order-preserving'
dict append var key ?string? Append (concatenate) string values
dict clear var Empty a dictionary but retain type
dict create ?key value?... Create dict, Reference is returned
dict exists value key ?key? Test if a key exists in a dict arg
dict for {key val} dictValue body Dictionary iterator
dict get dictValue key ?key...? Get a dictionary item(s)
dict info Show dictionary information
dict item dictValue key ?default? Return a single key match or default
dict keys dictValue ?glob? Return a list of dictionary keys
dict set var key ?key...? value Add/update a dict item by key value
dict lappend var key ?value? Append items as well-formed Tcl list
dict map Iterate a dictionary against a script
dict merge ?dictValue? Merge dictionary items
dict remove dictValue ?key ...? Remove items from a dictionary arg
dict replace dictValue ?key value? Replace items by key search
dict set var key ?key...? value Set key values in a dictionary
dict size Return the number of element pairs
dict unset Unset individual key entries
dict values dictValue ?glob? Return a list of dictionary values
dict with var {script} Set variables to dict key values
dict walk Diagnostic iteration of a dictionary
set dictVar Return a dictionary as a Tcl list
$dictVar Ditto
unset dictVar Delete a dictionary variable
A dictionary is a series of pairs of keys and values which are may be
stored either in a Tcl list or in a dictionary object. Nested dictionaries
are permissible in the same manner as nested Tcl lists. Each nested
dictionary level will consist of a matched key+value pair.
Dictionaries may be populated from arrays, text files or data files for
example
Dictionary string formatting follows the same Tcl list rules in that empty
items are represented as {} and strings containing whitespace are wrapped
in braces.
Dictionary data is stored internally as a hash table for fast retrieval
and can handle very large dictionaries. Dictionaries may be exchanged
between Tcl list strings and back using standard dereference commands $
and [set] with conversion back to dictionary via [dict create] and
[dict set]
Dereferencing a Ticol dictionary with $ returns a Tcl list string, not
another dictionary object. This means dereferencing the contents of a
dictionary can be used with other related commands such as list-processing
It also means that dereferencing is inefficient with large dictionaries
and dereferencing an entire dictionary should be avoided
If dictionary variable assigned using [set] is dereferenced using $ or
[set], the output will be a well-formed Tcl list.
Dictionary variables are deleted using [unset], not [dict unset]
Example:
set mydict [dict create 4470 {Great Northern} 4471 "Sir Frederic Banbury"]
puts "Type='[type mydict]'"
dict lappend mydict 4472 "Flying Scotsman"
dict lappend mydict 4473 "Solario"
dict lappend mydict 4474 "Victor Wild"
dict lappend mydict 4475 "Flying Fox"
dict lappend mydict 4476 "Royal Lancer"
set k [dict keys $mydict]
foreach item [lsort $k] {
puts [format "%-10s\t%s" $item [string trim [dict get $mydict $item] \
[chr 34]]]
}
unset mydict
Results:
Type='dictionary'
4470 Great Northern
4471 Sir Frederic Banbury
4472 Flying Scotsman
4473 Solario
4474 Victor Wild
4475 Flying Fox
4476 Royal Lancer
See also: dict append, dict create, dict exists, dict info, dict map,
dict remove, dict rename, dict set, dict size, dict walk, list,
arrays
$dictionary, [set dictionary] (Dictionary Dereference) |
Index |
string $dictionaryVariable
string [set dictionaryVariable]
Dictionary data is stored internally as a hash table for fast retrieval
and can handle very large dictionaries. Dictionaries may be exchanged
between Tcl list strings and back using standard dereference commands $
and [set] with conversion back to dictionary via [dict create] and
[dict set]
Dereferencing a Ticol dictionary with $ returns a Tcl list string, not
another dictionary object. This means dereferencing the contents of a
dictionary can be used with other related commands such as list-processing
It also means that dereferencing is inefficient with large dictionaries
and dereferencing an entire dictionary should be avoided
If dictionary variable assigned using [set] is dereferenced using $ or
[set], the output will be a well-formed Tcl list.
Double-dereferencing is not yet implemented for dictionaries. Note that
the dictionary variable is simply a handle to a dictionary object.
Example:
set dict [dict create 4470 {Flying Fox} 4472 {Flying Scotsman}]
puts "Type dict '[type dict]'"
puts "Type \$dict '[type $dict]'"
puts $dict
set a dict
puts [set $a] # [set] will double-dereference ok
puts $$a # Double-dereferencing is not implemented
Result:
dictionary
string
4470 {Flying Fox} 4472 {Flying Scotsman}
4470 {Flying Fox} 4472 {Flying Scotsman}
<dict>
See also: set, dict
Note: Ticol dictionaries are not currently 'order-preserving'
dictionary [dict append dictionaryVariable key ?string ...?]
Append new key+value pairs to an existing dictionary as specified by
the dictionaryVariable argument. The return value is a dictionary.
[dict append] appends the values to an existing key by string concatenation
rather than using a list operation (see [dict lappend])
[dict append] takes a dictionary variable as its argument. This
variable need not exist. If it does not exist it will be created.
A dictionary value is returned which may be echoed using $ or [set]
Example:
set mydict [dict create 1 The]
dict append mydict 1 " quick brown" # 1 {The quick brown}
dict append mydict 1 " " # 1 {The quick brown }
dict append mydict 1 fox # 1 {The quick brown fox}
puts $mydict # 1 {The quick brown fox}
Result:
1 {The quick brown fox}
See also: dict, dict lappend
bool [dict clear dictionaryVar ?dictionaryVar? ... ?-nocomplain?]
Clears one or more dictionaries of all keys and values but retains the
data structure and type information ready for repopulation
Errors are fatal and will halt the script unless the -nocomplain argument
is used
This is not a standard Tcl command
Example:
set mydict [dict create a b c d]
puts $mydict
assert [type mydict] {eq $_ dictionary} -v
dict clear mydict
assert [type mydict] {eq $_ dictionary} -v
assert $mydict {eq $_ {}} -v
Result:
assert: PASS 'eq $_ dictionary'
a b c d
assert: PASS 'eq $_ dictionary'
assert: PASS 'eq $_ {}'
See also: dict, dict unset, unset
Note: Ticol dictionaries are not currently 'order-preserving'
dictionary [dict create ?key value ...?]
[dict create] takes a series of key+value pairs, supplied as a well-formed
Tcl list and returns a dictionary value. The returned dictionary may be
assigned to a variable using [set]. One or more of the value arguments may
be null '{}' if desired but null should be avoided for key values
If the set dictionary variable is dereferenced using $ the output will be
a well-formed Tcl list. Although this may show no transformation in the
data, the dictionary data is stored internally as a hash table for
efficient retrieval
[dict create] does not take a variable name as an argument. The return
value is a dictionary list which must be set into a variable. The
resulting variable created by [set] will be of type 'dictionary'
Example:
# This example uses the expand {*} operator to expand [split] return
set s {4470 {Flying Fox} 2553 Manna 2555 Centenary 60100 Spearmint}
set d [dict create {*}[split $s]]
puts "Type d is: [type d]"
# Dereference dictionary and return a list
puts $d
Result:
Type of d is: dictionary
4470 {Flying Fox} 2555 Centenary 2553 Manna 60100 Spearmint
See also: dict, dict exists, dict map
bool [dict exists dictionaryValue key ?key ...?]
[dict exists] returns a boolean value (1 or 0) indicating whether the key or
series of keys within a set of nested dictionaries) exists in the given
dictionary value. This returns a true (1) value if a [dict get] on that
series will succeed.
Example:
dict set a the quick brown fox jumps over the lazy dog
bool [dict exists $a the]
bool [dict exists $a The]
bool [dict exists $a the quick]
bool [dict exists $a the quick brown fox jumps over the lazy]
bool [dict exists $a the quick brown fox jumps over the lazy dog]
Results
true
false
true
true
false
See also: dict
<null string> [dict for {keyVar valueVar} dictValue script]
Iterates each key+value pair within a dictionary value string, mapping
each key+value pair to the specified variable names and applying
the script with those variables for each iteration. The result of the
command is an empty (null) string.
[dict for] allows the contents of the keys and values to be changed
However, it is not recommended that keys be changed
The script cannot unset the original source dictionary which it is
currently iterating as this is passed by value, it can, however,
change the variables in the local scope of the script and in
the dictionary which is returned
The keyVar and valueVar variables are created in the current scope and
will remain after [dict for] has exited
Example:
dict for {key val} {a b c d e f} {
puts "$key $val"
}
Result;
a b
c d
e f
See also: dict, dict set, for, foreach
void [dict info dictionaryVariable]
Displays diagnostic information about the underlying dictionary hash
table
Dictionary 'mydict' at address 0x2056e20 7 item(s)
7 entries in table, 100 bucket(s) (7.00% use)
Buckets with 0 or more entries: 93 (Total: 0)
Buckets with 1 or more entries: 7 (Total: 7)
Buckets with 2 or more entries: 0 (Total: 0)
Buckets with 3 or more entries: 0 (Total: 0)
Buckets with 4 or more entries: 0 (Total: 0)
Buckets with 5 or more entries: 0 (Total: 0)
Buckets with 6 or more entries: 0 (Total: 0)
Buckets with 7 or more entries: 0 (Total: 0)
Buckets with 8 or more entries: 0 (Total: 0)
Buckets with 9 or more entries: 0 (Total: 0)
Buckets with 10 or more entries: 0 (Total: 0)
Mean search distance per entry: 1.00
See also: dict, dict walk, array walk
dictionary [dict get dictionaryValue key ?key ...?]
Returns a value matching a given key or keys
Multiple keys may be supplied to search within a nested dictionary
A failure to find a match to a given key will raise an error
Use [dict exists] to check for the presence of keys before retrieving
Example (non-nested):
# Non-nested dictionary
set mydict [dict create 4470 {Great Northern} 4473 Solario]
puts [dict get $mydict 4470]
puts [dict get $mydict 4473]
Result:
Great Northern
Solario
Example (nested):
# Nested dictionary
# Locomotive, BR number, LNER number, Name
dict set a4 60014 2509 Silver Link
dict set a4 60015 2510 Quicksilver
puts [dict get $a4 60014 2509]
puts [dict get $a4 60015 2510]
puts [dict get $a4 60103 4472]
Result:
Silver Link
Quicksilver
<error>
See also: dict, dict item
string [dict item dictValue key ?default? ?-nocase? ?-glob?]
Retrieve a single dictionary item from a dictionary value which matches a
single key value.
If the key is not found then an empty string will be returned unless an
alternate default value is specified, in which case this default will be
returned instead.
The default comparison will be made by case-sensitive, full-string match
unless the -nocase and/or -glob arguments are used.
-glob will treat the key value as a glob wildcard match template
-nocase will perform either a standard or glob search as case-insensitive
If a wildcard search using -glob is performed then there may be other
matching keys present in the dictionary. [item] will return the first
matching key only. Use [dict keys] or [dict values] to collect multiple
matching values
[item] is a non-standard Tcl command and it matches the behaviour of
[item], [array item] etc. as found elsewhere in Ticol
Example:
set mydict [dict create The quick brown fox]
puts [dict item $mydict the] # No match
puts [dict item $mydict The] # Direct match
puts [dict item $mydict the -nocase] # Case-insignificant match
puts [dict item $mydict t?e] # No match
puts [dict item $mydict T?e -glob] # Wildcard match
puts [dict item $mydict t?e -glob -nocase] # Wildcard match, nocase
unset mydict
Result:
''
'quick'
'quick'
''
'quick'
'quick'
See also: dict, dict get, dict keys, dict values
dict keys, dict values |
Index |
Note: Ticol dictionaries are not currently 'order-preserving'
dictionary [dict keys dictionaryValue ?globPattern?]
dictionary [dict values dictionaryValue ?globPattern?]
Return either a list of all or selected keys in a dictionary value or a
list of all values stored in a dictionary value. Note that [dict keys] and
[dict values] both take a dictionary value not a variable
Example:
set mydict [dict create foo 18 bar 19 moo 20 par 20]
puts [dict keys $mydict]
puts [dict keys $mydict ?oo]
puts [dict keys $mydict *ar]
puts [dict values $mydict]
puts [dict values $mydict 1?]
Result:
moo foo par bar
moo foo
par bar
18 19
18 19 20 20
See also: dict, dict get
Note: Ticol dictionaries are not currently 'order-preserving'
dictionary [dict lappend dictionaryVariable key ?value ...?]
Appends one or more values to a dictionary key value in Tcl list format
[dict lappend] takes a dictionary variable as its argument. This
variable need not exist. If it does not exist it will be created.
A dictionary value is returned which may be echoed using $ or [set]
Example:
set mydict [dict create keyVal foo]
dict lappend mydict keyVal {some marmalade} {I like marmalade}
dict lappend mydict keyVal bar
puts $mydict
Result:
keyVal foo {some marmalade} {I like marmalade} bar
See also: dict, dict map, lappend, append
Note: Ticol dictionaries are not currently 'order-preserving'
dictionary [dict map {keyVariable valueVariable} dictionaryValue body]
[dict map] transforms a literal dictionary string via a pair of variables
which represent the key/value pairs within the passed dictionary string
and which will (possibly) be transformed by the body script.
For each iteration of a key value in the dictionary, 'body' is called
with the specified variables substituted to the values of the dictionary
key pairs. The script may modify the key pair values as desired and the
changes are written back to the returned dictionary.
The returned dictionary represents the transformation applied to each
key+value pair by the script argument. The return value is cast to a
dictionary type
It is not recommended to change the key value within the script as this
will rename the original key within the dictionary
ActiveState Tcl's behaviour is to clear all values to null '{}' unless
explicitly set by the script argument. Ticol leaves values unchanged
unless specifically set or cleared by the script as this makes more sense
If you wish a value to be cleared you must do it explicitly by using
set valueVariable "" or valueVariable val {}.
A key value which is changed will replace the original dictionary key
rather than creating a new key
Example:
set i 1
set dict [dict map {key var} {1 {Blink Bonny} 2 {Sir Nigel Gresley}\
3 {Flying Scotsman}} {
if {[expr $key == 1]} {
set var "Solario"
}
}]
puts "The dict is now: '$dict'"
Results:
The dict is now: '3 {Flying Scotsman} 1 Solario 2 {Sir Nigel Gresley}'
See also: dict, map
dictionary [dict merge ?dictionaryValue ...?]
[dict merge] merges together one or more dictionary values to return a new
dictionary variable. Where key collisions occur the result will be based
on the last key+value pair presented
Example:
# Create and merge three dictionaries
set dict1 [dict create 2550 *Unknown* 4498 {Sir Nigel Gresley}]
set dict2 [dict create 4470 {Flying Fox} 4472 {Flying Scotsman}]
set dict3 [dict create 2553 Manna 2555 Centenary 2550 {Blink Bonny}]
set dict [dict merge $dict1 $dict2 $dict3]
dict for {key val} $dict {
puts "$key $val"
}
unset dict1 dict2 dict3 dict
Result:
2553 Manna
2550 Blink Bonny
4470 Flying Fox
4472 Flying Scotsman
4498 Sir Nigel Gresley
2555 Centenary
See also: dict, dict set
dictionary [dict remove dictionaryValue ?key ...?]
Removes a key+value pair from a dictionary by matching a key entry to
return a new dictionary value
Be careful to supply a dereferenced dictionary variable or literal
dictionary list and not a variable name. Also note that the original
dictionary will remain unchanged
Example:
set dict1 [dict create 4470 {Flying Fox} 4472 {Flying Scotsman}]
set dict2 [dict remove $dict1 4470]
puts $dict1
puts $dict2
Result:
4470 {Flying Fox} 4472 {Flying Scotsman}
4472 {Flying Scotsman}
See also: dict replace, dict unset
dictionary [dict replace dictionaryValue ?key value ...?]
Replaces one or more key+value pairs in a specified dictionary value and
returns a new dictionary with the amended key+pair combinatins. Where
collisions exist between successive keys, the result will be that produced
by the last key entry supplied.
Be careful to supply a dereferenced dictionary variable or literal
dictionary list and not a variable name. Also note that the original
dictionary will remain unchanged
Example:
set dict1 [dict create 4470 FOOBAR 4472 {Flying Scotsman}]
set dict2 [dict replace $dict1 4470 {Flying Fox}]
puts $dict1
puts $dict2
Result:
4470 FOOBAR 4472 {Flying Scotsman}
4470 {Flying Fox} 4472 {Flying Scotsman}
See also: dict replace, dict unset
Note: Ticol dictionaries are not currently 'order-preserving'
dictionary [dict set dictionaryVariable key ?key ...? value]
[dict set] creates a new dictionary by assigning one or more key arguments
to a single value. Multiple key arguments will create a nested dictionary
A dictionary variable argument must be specified and this variable must not
already exist.
Example:
# Non-nested example
dict set mydict 4470 {Great Northern}
dict set mydict 4471 "Sir Frederic Banbury"
puts $mydict
unset mydict
Result:
4470 {Great Northern} 4471 {Sir Frederic Banbury}
Nested dictionary keys must access the keys by nesting the first
keys innermost within any nested sequence
Example:
# Multiple, nested keys mapping to a singular value, 'f'
dict set mydict a b c d e f
puts "Dictionary is: '$mydict'"
puts "Nested retrieval: '[dict get [dict get [dict get \
[dict get $mydict a] b] c] d]"
unset mydict
Dictionary is: 'a {b {c {d {e f}}}}'
Nested retrieval: 'e f
See also: dict, dict create, dict exists, dict map
integer [dict size dictionaryValue]
Return the size (count) of key+value pairs in a dictionary
Note that this command takes a VALUE (not a dictionary variable as its
argument
Example:
puts [dict size {a b c d}]
Result:
2
See also: dict
CAUTION: Don't use [dict unset] to unset (delete) a dictionary variable!
Use [unset dictvar] instead
dictionary [dict unset dictionaryVariable key ?key ...?]
Remove one or more dictionary items by matching to a given key value.
A (possibly) updated dictionary is returned. [dict unset] does not unset a
dictionary variable.
Example:
set mydict [dict create 4470 {Flying Fox} 4472 {Flying Scotsman}\
2553 Manna 2555 Centenary]
puts [dict size mydict]
dict unset mydict 2553 2555
puts [dict size mydict]
puts [dict get $mydict]
unset mydict
Result:
4
2
4470 {Flying Fox} 4472 {Flying Scotsman}
See also: dict, unset, dict rename, dict remove
void [dict walk dictionaryVar ?-bare?]
Iterates the underlying hash table object for a specified dictionary
variable. Useful for debugging dictionaries.
The '-bare' argument changes output to a key+values only output with no
structural debug information
------------------------------------------------------
hash_table object:0x2216100 size:100
------------------------------------------------------
Bucket(6) Node: 0x2217b00 chained nodes:1
1: 0x2217b00 key:'4475' => 'Flying Fox'
Bucket(26) Node: 0x2217810 chained nodes:1
1: 0x2217810 key:'4476' => 'Royal Lancer'
Bucket(32) Node: 0x2216500 chained nodes:1
1: 0x2216500 key:'4470' => 'Great Northern'
Bucket(37) Node: 0x22177d0 chained nodes:1
1: 0x22177d0 key:'4472' => 'Flying Scotsman'
Bucket(53) Node: 0x2217160 chained nodes:1
1: 0x2217160 key:'4474' => 'Victor Wild'
Bucket(64) Node: 0x2217990 chained nodes:1
1: 0x2217990 key:'4473' => 'Solario'
Bucket(87) Node: 0x2216cb0 chained nodes:1
1: 0x2216cb0 key:'4471' => 'Sir Frederic Banbury'
Hash Table Statistics:
Root size:100 Occupation:7 (7.00%) Max depth:0 Nodes:7
------------------------------------------------------
-bare style output...
60014 2509 {Silver Link}
60015 2510 Quicksilver
60016 2511 {Silver King}
60017 2512 {Silver Fox}
See also: dict, array walk, dict info
<result> [dict with dictionaryVariable ?key ...? body -unset]
[dict with] sets a series of variables within the current scope to one
or more of the key values found in the supplied dictionary variable.
The dictionary variable will be a well-formed Tcl list
Once these are set the script body will be called and the result will be
the result of that script evaluation
This could be say the 'value' variable of [dict for]
If -unset is used then [dict with] will unset the variables it set
Specifying keys is not yet implemented. Nested dictionary handling is
not yet implemented
Example:
# List Railway Locomotive Numbers
# Set fields individually to create a single-level nested list
# Dictionary name
# | Unique key
# | | Key+value pair
# | | | |
# v v v v
dict set a4 1 name Silver Link
dict set a4 1 br 60014
dict set a4 1 lner 2509
dict set a4 2 name Quicksilver
dict set a4 2 br 60015
dict set a4 2 lner 2510
dict for {key val} $a4 {
dict with val {
printf "%2s-> Name: %-10s\tBR Number: %5s \
LNER Number: %4i\n" $key $name $br $lner
}
}
Result:
1-> Name: Silver Link BR Number: 60014 LNER Number: 2509
2-> Name: Quicksilver BR Number: 60015 LNER Number: 2510
See also: dict, dict for, dict set
The following Windows dialogs are available:
inputbox Accept user input, optionally with password masking
get_fileopen Browse for a file to open
get_filesave Browse for a file to save
get_folder Browse for a folder name
msgbox Get user input or display a message from a popup dialog
There is no Tk GUI support in Ticol, unless an external DLL is called or the
Win32 API called via the calldll* plugin
See also: inputbox, get_fileopen, get_filesave, get_folder, msgbox, gets
die ?message? ?-code return-code?
A PHP-like command to exit the current program, optionally displaying a
message or return code. If running in the CLI, control is returned to the
CLI, otherwise the program will exit to the operating system.
Example:
die "Something went wrong"
die "Exiting to system" -code 10
See also: exit, stop, at_exit, return
Differences between Ticol and Standard Tcl |
Index |
Ticol is intended to be a subset of standard Tcl with a few differences
The following aspects differ from ActiveState Tcl:
$$ Double/multiple $ dereferencing is supported
${...} Complex, nested dereferencing with braces is supported
Thus, braces have extra meaning when associated with $
$this Ticol offers $this in [switch] and [goto_block]
$env $env (and other) consts are provided. $env has case-
insensitive subscript reference
any A simplified method of handling multiple || clauses
all A simplified method of handling multiple && clauses
arrays Variant arrays are available for use with calldll
A standard variable may be promoted to an array by
assignment but not vice-versa. Arrays may also be const or
case-insensitive. Arrays use efficient hashtables and
are associative, also supporting integer subscripts
"C" like arrays are available as a plugin
autoexec.tcl An optional startup, autoexec.tcl script is supported
big math Big number integer support available as a plugin and with
some floating point support (fdiv, fmod)
binary Binary values are handled as 0bNNNNN... with options
calc [calc] interacts with the Macro Preprocesor (MPP) which will
transform [calc] commands to more-efficient raw Tcl. If the
MPP is disabled then [calc] will be interpreted as [expr]
calldll An external plugin lib which offers DLL interfacing similar
to Visual BASIC, dotNet etc.
chain Follows a chain of multiple nested dereferences (see $$)
commands Many additional commands which are implemented at binary
level for speed. This includes some useful BASIC-like
commands. To facilitate [calc] macro optimisation a command
will accept a reduced calc expression as [number]. e.g. [10]
comments Comments require no ; prefix as ;#
Comments also include "C" style /* ... */ long-comments
console echo To make the CLI more useful, Ticol turns console echo off
by default. This may be enabled using: option echo on
date A wide variety of date commands are available
Debugger A single-step debugger and error stack trace
do Ticol implements a do..while loop as well as [loop]
encryption Source code obfuscation with various locking schemes are
available in Ticol Tcl
eval External scripts are linked and evaluated using [eval]
expand [expand] is an alternative to {*} in some cases
explain Can 'explain' or expand natural expression as Tcl braced
commands
expr Expressions can sometimes be handled differently, due to
less flexibility in the Ticol expression handler
e.g. Spaces required between commands when evaluating
strings in [expr]
ActiveState Tcl: if {""==""} {puts PASS}
Ticol Tcl: option expression on
if {"" == ""} {puts PASS}
Sometimes Ticol is more flexible:
ActiveState Tcl:
set op ==
set foo 1
if "$foo $op 1" {puts YES} # can't read "foo": no such var
Ticol Tcl:
option expression on
set op ==
set foo 1
if "$foo $op 1" {puts YES} # Passes (echoes YES)
if {"$foo$op 1"} {puts YES} # Passes (echoes YES)
if {"$foo${op}1"} {puts YES}# Passes (echoes YES)
gosub [proc] without arguments or call a [goto_block] label
goto Block-limited gotos (purely for entertainment purposes)
with localised scope and jump out of flow-control direct
to a label
if The keyword 'then' is not supported. It serves no useful
purpose and processing it slows down the interpreter
item The [item] keyword allows safe dereference calls to objects
which may not exist and offers default return values
Supported for all object types
lib External plugin modules (as DLLs) are loaded using [lib]
The standard Tcl plugin method is different and incompatible
Linked exports are automatically unregistered on unloading
a lib. Aliases are also checked.
loop A simple and efficient equivalent to [for]
namespaces Real namespaces are not supported. They are emulated using
name-scope prefixes such as 'fake_scope::var'
macros A full "C"-like Macro PreProcessor with load-time
optimisation and extended commands
Nested #ifdef is not currently supported
math All integer math is calculated using 64-bit integers
Hex, octal and binary consts are handled transparently in
math.:e.g. expr 0b1010*3 => 30, expr 0x0f*10 => 150
The new Tcl 9.x 0oNN octal format specifier is supported
Math expressions are optimised by the Macro PreProcessor
unless /NEO and /NES command line arguments are used
memory Memory allocation and free for use with [calldll] which
exposes the standard "C" malloc, free and memcpy
obfuscation Script encryption (obfuscation). Ticol scripts may be
obfuscated using symmetric encryption and may also be
execution-locked by various means to user, workstation
or by password etc.
on Similar to BASIC's ON GOSUB
option Certain functionality is controlled via the [option] command
particularly, the 'option expression' option
Performance Performance analysis, profiling and graphing
plugins Around 50 plugins are available to extend the language
proc Ticol procs may not be integers or real numbers
A proc name may, however, contain numeric digits
ActiveState Tcl allows numeric values as proc names
A proc name may also consist of only one or more spaces
Procs defined with numeric prefixes can be called only
directly using [call procname]
Proc default args cannot contain math or other symbols
unless wrapped in double-quotes. It is good practice to wrap
strings within double-quotes in any case
set A standard variable may be promoted to an array by
assignment using [set] and no warning is issued, This helps
facilitate [static] variables in procedures. e.g.
set a Hello
set a(10) 23 # Variable a is now type <array>
puts $a # Returns "<array>"
speed Ticol Tcl is purely an interpreter. Even though the MPP will
do some optimisation it will always be slower than
ActiveState Tcl or compiled code
stacks Stacks are built-in to the main interpreter core
static Ticol offers native static variables for procs, and
this can be extended to include structs
structs Structs are built-in. Ticol provides for binary struct
emulation in order to interface with external DLLs. Structs
may also be used with care with native Tcl code
sub [sub] offers argument-less procedures
time Optionally returns the current time as [time]
Ticol also has an extensive [date] command
uplevel The absolute level parameter is specified by a @ symbol
not # as # is reserved by the MPP for comments
e.g. uplevel @1 command # Executes at absolute level 0
upvar The absolute level parameter is specified by a @ symbol
not # as # is reserved by the MPP for comments
[upvar] also assumes the local variable name (optional)
e.g. upvar @1 s # Executes at absolute level 0
Also, [upvar] optionally assumes the alias, which makes
far more sense than making it mandatory
e.g. upvar s # Not upvar s s
variants Some variant handling for interfacing with external DLLs
These are not recommended for use with native Tcl code
variable [variable] is implemented an alias for [set] because Ticol
does not support real namespaces
Variable names allow underscores, dots and colon characters
Variable names can embed "::" as a pseudo namespace
e.g. variable foo::bar
variable More restrictive than standard. Allows only the following
characters names "a-zA-Z0-9_.:& \t"
Windows Basic Windows dialogs including file open, file save,
folder-browse, input box and message box
xbase/dbase xbase database support as a plugin as well as an extended
database support which offers large fields/records
The Macro-PreProcessor (MPP) extends the standard Tcl comment character.
This adds trailing comments but reserves the # symbol. The MPP s designed
to optimise and speed up interpretive scripts but can be disabled using
the /NP command argument (not recommended). The MPP consumes no run-time
resources as pre-processing is performed at load time
Macro #__LINE__ , #__DATE__ #__TIME__, #__FILE__, #__USER__ and #__HOST__
debugging constants are available. Macro PreProcessor #define, #ifdef,
#ifndef, #undef processing controls
Ticol has no Tk/GUI support unless you call an external dll via the calldll
plugin to facilitate graphics output via the Windows API
See also: non standard
See also: ticol, gui, arrays, calc, calldll, do, encryption, global, goto,
lib, math, option, struct, time, uplevel, upvar, variant
dim arrayname size
Dimensions an array according to size. The underlying hash table is
created and required memory is allocated. No array elements are assigned
and the element count is set to zero. See also [array create]
A size value of 0 is acceptable
[is_set] will return $true (1)
[array exists] will return $true (1)
[array count] will return 0
[dim] is not necessary to create array elements as array allocation is
dynamic, but it may speed up execution when allocating a very large number
of elements within a loop as the array hash table is expanded dynamically
and geometrically as elements are added. This introduces a time cost
overhead to critical loops
Example:
dim names 100
puts [array count names]
array walk names
Results:
0
-----------------------------------------------------------------
hash_table object:0x35ec650 size:10
-----------------------------------------------------------------
Hash Table Statistics:
Root size:10 Occupation:0 (0.00%) Max depth:0 Nodes:0
-----------------------------------------------------------------
See also: array, arrays, array set
integer [disktype x:]
Returns a numeric value indicating the disk type as follows:
0 DRIVE_UNKNOWN
1 DRIVE_NO_ROOT_DIR
2 DRIVE_REMOVABLE
3 DRIVE_FIXED
4 DRIVE_REMOTE
5 DRIVE_CDROM
6 DRIVE_RAMDISK
Example:
array set status {
0 "Unknown"
1 "No root directory"
2 "Removable"
3 "Fixed"
4 "Remote"
5 "CDROM"
6 "RAMDisk"
} -const
puts "[array item status [disktype c:] Unknown"
Result: (where c: is a hard drive)
Fixed
Note that a trailing path is allowed. e.g. [disktype "c:\\temp"]
See also: diskfree, diskstat, info
dll_close dllname handle
Release a DLL previously opened by [dll_open]. Accepts either an integer
literal or a variable. Returns a boolean success flag.
See: calldll
Example:
set handle [dll_open "vbsqlite3.dll"]
dll_close $handle
Result:
1
See also: dll_open, calldll
integer [dll_open dllname]
Load a DLL into persistent memory. This may be necessary to process DLL
calls where there is a need for a persistent handle such as SQLite
queries
[calldll] will accept a handle to an open DLL in place of a DLL name
See: calldll
Example:
set handle [dll_open "vbsqlite3.dll"]
Result:
13631488
See also: dll_close, calldll
Unary commands: incr, decr, ++, -- |
Index |
integer [incr value|variable ?increment? ?-unsigned? ?-width 8|16|32?]
integer [decr value|variable ?increment? ?-unsigned? ?-width 8|16|32?]
integer [++ value|variable ?increment? ?-unsigned? ?-width 8|16|32?]
integer [-- value|variable ?increment? ?-unsigned? ?-width 8|16|32?]
The above commands increase or decrease a value, simple or array variable
by a specified amount. No other variable type is accepted.
The increment or decrement is by 1 unless increment is specified
If specified the increment may be +ve or -ve and greater than 1
The optional increment argument may be any signed 64-bit integer
Real numbers (1.2345) will be rounded to integer before applying the
operator so: [++ -1.23] gives 0, and [++ 1.23] gives 2
The command is extended with the addition of the following arguments:
-unsigned Return an unsigned 64-bit integer
-width N Return the value (truncated if nesc.) to 8,16 or 32 bits
-width 8 Unsigned char (8 bits)
-width 16 Unsigned word (16 bits)
-width 32 Unsigned long (32 bits)
Example:
set line #__LINE__
++ line 2
puts "Line is $line"
Example:
set a(23)
puts [++ a(23)]
puts $a(23)
Result:
24
24
Unary increment/decrement is not implemented within [expr]. This is due to
potential conflicts with expressions such as "3 - -7"
Embed [++] [incr] or [--] [decr] commands within the expression as follows...
Example:
set a 1;puts [expr "[++ $a]*2"] # Result 4
Example:
set a 1;puts [expr "[++ $a 5]*2"] # Result 12
Where the incrementing operator is used in an expression as an infix
operator it should be wrapped in parentheses if an increment value is given
Example:
set a 1;puts [expr "($a++ 10)*2"] # Result 22
See also: math, expr, eval
integer [diskfree x:]
Shows the free disk space for a given drive/path. [diskfree] works with
UNC path. You can use comma to format the result
Example:
puts "[comma [diskfree d:]] byte(s)"
Result:
135,071,793,152 byte(s)
Example:
Show free space on UNC path for workstation, "snoopy" C:
puts "[format %.3f [/ [diskfree \\snoopy\c$] [* [* 1024.0 1024] 1024]]] Gb"
Result:
7.89 Gb
See also: disktype, diskstat, info, file
integer [diskstat x:]
string [diskstat x: -name]
Check the status of a disk drive, including network drives and return a
status code
The -name argument will cause either a drive letter or network share name
to be returned depending on the connection type instead of a status code
Useful in conjunction with [disktype]
The following codes are returned where -name is omitted:
0 Absent/error (invalid spec/removable drive absent/no such drive)
1 Drive is present and available
2 Networked drive present but disconnected state
3 Networked drive is absent
See also: diskfree, disktype, info, file
do while, do until (Flow Control) |
Index |
do while {condition-statement} {code-block}
do {code-block} while {condition-statement}
do {code-block} {condition-statement}
do until {condition-statement} {code-block}
do {code-block} until {condition-statement}
A flexible-syntax [do] loop
The following additional commands interact with [do]
break # Is passed back to any enclosing structure
stop # Halt operation, return to the CLI if loaded
exit # Exit the script and return to Windows
goto # Where used with an enclosing [goto_block]
continue # Resume execution at the head of the loop
Ticol offers both standard Tcl while loops as well as do...while loops
Well-formed single statements need not be braced but where braces are
present and over more than one line the open brace must be at the end of the
line not the start of a line (see flow control). Statements may be combined
on a line using the semi-colon separator
Examples:
do {puts [rnd 1 10] } while 1
do {puts [rnd 1 10] } while { 1 }
do {puts [rnd 1 10] } while {
1
}
set i 0
do { # Correct brace layout (Tcl)
puts "i is $i"
if {$i == 2} { return }
incr i
} while { $i < 10}
do while {1} {puts [rnd 1 10] }
do {puts [rnd 1 10] } {1}
set i 0 ; do {puts "i is $i"; if {== $i 2} { return } ; incr i } \
while { $i < 10}
do while {[< $i 10]} {
puts "do-while i is $i"
++ i
}
set i 0
do {
puts "do {} i is $i"
++ i
} {[< $i 10]}
set i 0
do {
puts "do-while i is $i"
++ i
} while {[< $i 10]}
See also: flow control, while, for, loop, if, switch, goto, time
$varname
[set varname]
Dereferences a variable to return the contents (see: double dereference)
Very early versions of Tcl lacked the '$' syntax and, instead, used the
[set] command to dereference variables. In some cases [set] may still be
useful. Variable names may be braced as ${var} to isolate from strings
Because braces delay evaluation, the use of braces to isolate variable
names can be problematic, especially with nested sub-variables:
e.g. ${name_$var}. Such cases will require the use of intermediate
variables since translation would require the order of nested brace
evaluation to be reversed for a $ operation. Nested, braced dereferences
are allowed (e.g. ${var1${var2}} )
Strings which contain non-variable embedded dollars must be escaped.
If it is not possible to manually escape then such strings may be
escaped using the [escape] command
Example:
set var {"\\path\\to\\badfile.$$$"} # Simulate read from [ls]
puts [escape $var] # Prevent problems
Unlike standard Tcl, Ticol supports multi-level dereferences ($$var)
as well as complex, nested dereferences using braces.
See: nested dereferences
Note that you cannot dereference a struct member using $ or [set] as
this merely returns the member offset address within the struct
Example:
set s 23
puts "a is '$s'"
puts "a is '[set s]'"
Result:
s is 23
s is 23
Example:
set a(1) 23
puts "a(1) is '$a(1)'"
puts "a(1) is '[set a(1)]'"
Result:
a(1) is '23'
a(1) is '23'
Example:
set a "Hello "
puts ${a}world
Result:
Hello world
Double-Dereferencing
--------------------
Double/multiple level dereferences are allowed in Ticol Tcl using either
nested [set] or multiple $ prefix statements. (See: double dereference)
Examples:
set a 23
set b a
set c b
set d c
puts $$$$d # 4 levels of dereference
puts [set $$$d] # 4 levels of dereference
puts [set [set $$d]] # 4 levels of dereference
puts [set [set [set $d]]] # 4 levels of dereference
puts [set [set [set [set d]]]] # 4 levels of dereference
Result:
23
23
23
23
23
See also: set, double dereference, nested dereference, variables, arrays
Domain Error (Math) |
Index |
Some math functions have a restricted set of valid inputs or 'domain'
submitting values outside these limits will result in a 'Domain Error'
The following functions have limited domains
Function Range Limit
--------------------------------------
acos -1 to +1
asin -1 to +1
log Positive numbers only
log10 Positive numbers only
sqrt Positive numbers only
Example:
puts [expr acos(23)]
puts [funct acos 23]
Result:
acos(23): Domain error
acos(23): Domain error
http://www.cplusplus.com/reference/stdexcept/domain_error/
See also: math functions
In common with most programming languages, double quotes in Tcl are used
to delineate and group words together into strings. They are not mandatory
unless an intended string value contains whitespace. Although braces are
preferable as delimiters, double quotes strongly recommended to be used in
all cases where a string is intended and braces can't be used.
Braces may also be used to perform this task since braces can group
characters into a single argument for a command. Since nested double-
quotes as not permitted, braces may be used to group string sequences
within double-quoted strings
Example:
set "I am a {string within} a string"
set "I am a {delayed $var within} a string"
Example:
puts "Hello world"
puts {Hello world}
It should also be noted that while unescaped double-quotes are not
allowed within a string, adjacent strings will concatenate, thus in the
following example \\r\\n isn't embedded but is, instead, concatenated
Example:
store add s "[unescape "\\r\\n"]\r\n"
Single-quote characters (') do not group and therefore cannot be used to
group characters into a string argument
The empty string or empty list may be represented by pairs of double
quotes or braces
Example:
set s ""
set s {}
Numeric hex or other literals for which it is desired that the MPP not
translate into decimal should be quoted. This will affect the interaction
with string comparison commands such as eq, ne. Numeric comparison
commands will perform automatic decimal conversion
Examples:
set a 0xABCDEF
set b "0xABCDEF"
set c 0o12345
set d "0o12345"
set e 0b1010101
set f "0b1010101"
puts $a
puts $b
puts [int $b]
puts $c
puts $d
puts [int $d]
puts $e
puts $f
puts [int $f]
Result:
11259375
0xABCDEF
11259375
5349
0o12345
5349
85
0b1010101
85
It is not allowed to embed unescaped double-quotes within a quoted-string
even if quote parity is maintained. In standard Tcl this is an error
Example:
option echo on
set q Hello"world" # This is actually two adjacent strings
Result
Hello"world # But the embedded quote is visible
However, a workaround is to wrap the unescaped string in spaces or to
escape the string and then use [unescape]. For example, when addressing
an array with a subscript which has spaces...
set q {a("with spaces")}
vars q*
assert $q {eq $_ {a("with spaces")}} -v
assert $q {eq $_ [unescape "a(\"with spaces\")"]} -v
See also: quoting hell, braces, single quotes, dereference,
minimal escaping style
Tcl requires careful attention to using quoted versus braced strings.
Usually the braced form is preferable but there may be a few cases where
quoted strings are preferable. In Ticol Tcl double-quoted strings (") are
preferable where a string contains long-comment /* */ tags.
Tcl may look like C/C++ but it is not and the handling of quoted strings
differs since, in essence, all but special types are handled as strings
The need for nested quotes may require a careful rethink of your source
code since nested double quotes are not allowed. Quotes are permitted only
at one level of nesting or the macro preprocessor will complain!
Quotes are often wholly unnecessary. A string is still a string even if
not enclosed in either quotes or braces. A means of enclosure becomes
necessary when whitespace is involved, however, even in this case, if
the this is an escaped space character then delimiters are unnecessary
The single quote - ' - is not a string delimiter
Example:
puts Hello # Hello is a string (undelimited)
puts "Hello" # Hello is a string
puts {Hello} # Hello is a string
puts "Hello world" # Hello world is a string
puts {Hello world} # Hello world is a string
puts Hello\ world # Hello world is a string (undelimited)
puts "Hello\ world" # Not valid at the CLI, valid only in scripts
# unless option escape is set to on
set a(Hello\ world) 23 # Array subscript: Hello world is a string
set a("Hello\ world") 23 # Array subscript: Hello world is a string
puts 'Hello' # 'Hello' is a string, but not delimited by '
puts 'Hello world' # Not a properly formed string
puts [Hello world] # Not a properly formed string
Links
-----
Recommended reading:
https://wiki.tcl-lang.org/page/Quoting+hell
https://wiki.tcl-lang.org/page/Tcl+Quoting
See also: quoting, minimal escaping style
Minimal Escaping Style |
Index |
This is a code style which avoids unnecessary Tcl escaping syntax wherever
possible and is highly recommended. For example, while it is strongly
recommended to properly brace arguments for flow-control commands such
as [if] or [while], these braces are optional and the recommendation is
merely to reduce mistakes.
In Ticol, much of the work of translating backslash-escaped characters
is performed during macro preprocessing which means its behaviour
differs slightly
Curly braces prevent proper and timely variable substitution
"In some languages such as C, Javascript, Lisp, and Python, double
quotes indicate a certain type of value: a string. In Tcl all values
are already strings, so double quotes do not have that function.
Instead, they provide an environment in which whitespace characters
are ordinary characters rather than word delimiters. Braces do the
same, but additionally inhibit the standard substitutions"
(https://wiki.tcl-lang.org)
General Rules
-------------
Don't use quotes where braces could be used.
Don't use braces where no backslash substitutions are needed
Examples
--------
puts Hello\ world # Valid (escaped spaces are ok)
puts Hello\t world # Not valid, will raise an error
puts {Hello\tworld} # Valid grouping using braces
puts "Hello\tworld" # Valid grouping using dquotes
Links
-----
https://wiki.tcl-lang.org/page/Tcl+Minimal+Escaping+Style
https://web.archive.org/web/20081009203437/http://www.tclscripting.com/articles/jan06/article3.html
See also: quoting hell
$$ double-dereference (multi-dereference) |
Index |
$$var
set $var
Ticol allows for a non-standard double-dereference similar to PERL on
simple variables and arrays elements. This avoids the need for clumsy
[set] statements and is much faster, avoiding a slow command-call
Multiple levels of dollar-dereferencing are supported e.g. $$$$var
Deeply-nested levels can also be handled using [set] in the approved
manner for example:
puts [set $var]
puts [set [set [set [set $var]]]]
Complex, braced variable expressions such as '${var}::name' are
supported by multiple dereferences.
Example:
# Prints out '23'
set prefix::name 23 # set 'var prefix::name' to value '23'
set p prefix # p points to 'prefix'
puts ${${p}::name} # Double-dereference via $p
puts [set ${p}::name] # Same functionality
puts [set [set p]::name] # Same functionality
[set] can also be used to handle very complex nested dereferences and
can be combined with the use of one or more dollar $ dereferences
Note that struct member fields must be dereferenced using [struct item]
and not the dereference dollar prefix $
An alternative to double/multiple dereferences is to use [chain] which will
automatically resolve down to a root case.
More examples...
Example:
set a 23
set myvar a
puts "Double dereference using \$\$ is: '$$myvar'"
puts "Double dereference using \[set\] is: '[set $myvar]'"
Result:
Double dereference using $$ is: '23'
Double dereference using [set] is: '23'
Example:
# d -> c -> b -> a == 23
set a 23; set b a; set c b; set d c
puts $$$$d # x4 levels of dereference
puts [set $$$d] # x4 levels of dereference
puts [set [set $$d]] # x4 levels of dereference
puts [set [set [set $d]]] # x4 levels of dereference
puts [set [set [set [set d]]]] # x4 levels of dereference
Result:
23
23
23
23
23
Example:
set a(1) one
set b a(1)
set c(1) b
set d c(1)
puts "\$\$\$\$d resolves to '$$$$d'"
puts "\$\$\$\$d resolves to '[set [set [set [set d]]]]'"
Results:
$$$$d resolves to 'one'
$$$$d resolves to 'one'
Array double-dereferences, where the array name prefix is contained
in a variable must be resolved as follows. This method is useful when
addressing array names returned from procedures etc.
set a(0) "Hello"
set q a
set i 0
puts "Array element $i: '${${q}($i)}'" # Correct syntax
# ^
# Resolve first ${q}
# Resolve second ${____($i)}
Result:
Hello
This breaks down as follows...
a) ${${q}($i)}
b) Resolve innermost reference: ${q} -> which gives 'a'
c) Resolve outermost array: ${a($i)} -> Hello
Complex array elements containing whitespace can be referenced using
quoted-escaped strings and the [unescape] command as follows: Whilst
variables can be declared which have complex whitespace, literal strings
require escaping.
Example:
set a("with spaces") "Hello world!"
# ^ Array definition (escapes not required)
set b [unescape "a(\"with spaces\")"] # Escaped string
# ^ Literal string (must be escaped)
set c b
puts "Triple dereference: '$$$c'"
Result:
Triple dereference: Hello world!
Structs
-------
Structs can slightly more complex to multi-level dereference, but this
depends on the circumstances:
Simple example:
struct create st {a 10} # Create struct s with 1 field
set s Hello # Create a separate variable
struct set st.a s # Point s.a to variable s
puts $st.a # Single-dereference
puts ${$st.a} # Double-dereference
# ^ # Dereferences to 's'
# ${s} # Dereferences to 'Hello'
Result:
s
Hello
Alternate Simple Example:
struct create st {a 10} # Create struct s with 1 field
set s hello # Create a separate variable
struct set st.a s # Point s.a to variable s
puts $st.a # Single-dereference gives s
puts $$st.a # Double-dereference gives $s
Result:
s
hello
Complex example:
struct create s {a 10 b 20} # Create struct s with 2 fields
struct set s.a Hi # Set field a to "Hi"
struct set s.b There # Set field b to "there"
set q s # Create 1 level of indirection
set r q # Create another level in r
puts "s.a '${${$r}.a}' " # Dereference 3 times
puts "s.a '${${$r}.b}' " # (2 indirection and 1 final)
Result:
Hi
there
How this resolves (from the innermost braces outwards) ...
# Expression: ${${$r}.a}
# ^
# Resolve first $r -> q
# Resolve second ${q} -> s
# Resolve third $s.a -> "Hi"
#
# Note that $s.a is equivalent to ${s.a}
Dictionaries
------------
Double/indirect dereferencing for dictionary types is not yet implemented
For such variables, use [set $varname]
Example:
set dict [dict create 4470 {Flying Fox}]
set a dict
puts $$a # Will return "<dict>"
puts [set $a] # Will correctly dereference dict ok
Other Data Types
----------------
All other data types such as linkedlist, stack, variant etc. use simple
handles or variables which can be multi-dereferenced as such
Useful resources:
http://phaseit.net/claird/comp.lang.tcl/tcl_deref.html
https://psg.com/~joem/tcl/faq.html#DoubleIndirect
See also: chain, dereference, nested dereference, set, variables
Complex Nested Dereferences |
Index |
Complex nested dereferences are supported for arrays and standard variables
These may also be braced to isolate variable names from strings. This is an
enhancement on standard Tcl
Braces are resolved within variable expressions in the opposite order to
braces within standard Tcl expressions. That is, they are evaluated
recursively, from the innermost set of braces outwards. Where braces are
used within a variable expression they must follow immediately after the
dollar sign. Braces must wrap the whole variable expression
Examples:
$s # A simple variable dereference
$a(text) # Array with subscript literal value 'text'
$a($ss) # Array with subscript variable '$ss'
${s}text # Variable 's' isolated from remainder of text
${${s}text} # Double dereference of 's' including 'text'
${a($ss)}text # Array 'a' with subscript '$ss', isolated from text
abc${a(1)}def # Array 'a(1)' embedded within the string "abcdef"
$a(${i}wo) # Where $i == 't', this references subscript 'two'
$$$c # Triple dereference (c->b->a->result)
$$${c} # Ditto (equivalent)
$${${c}} # Ditto (equivalent)
${${${c}}} # Ditto (equivalent)
${$p$q$r} # Where p, q, r == v, a, r and var == "x" then "x"
${${p}${q}${r}} # Ditto (equivalent)
${q}(${i}wo) # Where q == a and i == t then "a(two)"
${${q}(${i}wo)} # Where q == a and i == t and a(two) == "x" then x
${${$q}.a} # Where q => r => struct s with member field a
${::x}abc # A global variable called x embedded against 'abc'
${::{$x}.a} # Where x == s; refer to global struct member s.a
${::foo::{$x}.a} # Where x == s; refer to namespace struct ::foo::s.a
Invalid Expressions:
${q}(text) # Where q == a. Array 'a' will return "<array>text"
# Should be ${${q}(text)}
{$a(1)} # Will not resolve due to protective braces/missing $
# Should be ${$a(1)}
{$p$q$r} # Will not resolve due to protective braces/missing $
# Should be ${$p$q$r}, however ...
# puts [eval $[subst {$p$q$r}]] ... will resolve
$::{x}abc # A global variable requires :: inside the braces
Dereferencing [struct] Inside [proc]
------------------------------------
Dereferencing an [upvar] struct inside [proc] can be problematic due to
indirection in the struct name which requires a complex, nested
dereference.
If desired, [set] or [->] can be used as an alternative to resolve the
final dereference where ${${argvar}.field} would be required
Example:
struct s {a 10}
proc foo {x} {
upvar $x # Dereferencing, x gives the struct name 's'
puts "-> ${x}.a" # This requires x.a to be braced
puts "-> ${${x}.a}" # This requires $s.a to be braced
}
struct set s.a Hello
foo s
Results:
s.a
Hello
See also: chain, dereference
Dump loaded source-code or known variables
dump ?-echo? ?-crlf? # Display syntax-highlighted source code
dump variable # Analyse a variable
Dump with no arguments will display any preprocessed source code loaded
within the CLI by either [run] or [load] as long as it is not encrypted.
Code which has been encrypted (obfuscated) using ticol.exe /C cannot be
listed
Source code is highlighted using the following convention
blue Inbuilt Tcl commands
magenta Procs defined within the script
yellow Strings (embedded commands not highlighted)
green Dereferenced variables
darkyellow Numeric values
grey Variables, arguments and lib commands
white Square, curved and curly braces
- Comments are removed by the MPP
Example: (procs will only be highlighted once known)
proc foo {x} {
puts "Hello from [info procname] "$x # Alert user
++ x
}
puts [foo 12] # Prints 13
Line numbers will no longer match the original source-code but #__LINE__
tags will correctly reference the original source. Trailing whitespace is
removed but indenting is preserved.
See: load
You can return the source code as a single string to be captured by a Tcl
variable by using the -echo argument. Unless the Macro PreProcessor has
been disabled the returned code will be preprocessed and will have had
empty lines removed with comments and macro #ifdef..#else..#endif blocks
stripped
Example:
load msgbox # Load msgbox.tcl to the CLI
set a [dump -echo] # Echo the source into a variable
puts $a # Display the source
Result:
Loaded msgbox.tcl OK 105 byte(s)
set r [msgbox [join [split $qbf " "] "\n" 768] "Choose" 36]
msgbox "You chose [map {6 Yes 7 No} $r]"
--
dump ?variable?
[dump] with a variable name displays an analysis of a variable. When used with
structs full member details will be shown, this includes the offset into the
struct and the member (field) size in bytes
struct 's' - at 9384672 (0x8f32e0), 146 byte(s), 5 member(s)
+ member 1: 's->s_a' => 0x8f32e0, 4 byte(s)
+ member 2: 's->s_b' => 0x8f32e4, 30 byte(s)
+ member 3: 's->s_c' => 0x8f3302, 100 byte(s)
+ member 4: 's->s_d' => 0x8f3366, 4 byte(s)
+ member 5: 's->s_e' => 0x8f336a, 8 byte(s)
-------- ----------------------- ----------------------- -- ASCII -------
008f32e0 48 65 6C 6C F0 FF 00 00 - 6F 00 33 32 31 41 42 43 Hell....o.321ABC
008f32f0 44 45 46 47 48 49 4A 4B - 4C 4D 4E 4F 50 51 52 53 DEFGHIJKLMNOPQRS
008f3300 54 55 54 68 65 20 71 75 - 69 63 6B 20 62 72 6F 77 TUThe quick brow
008f3310 6E 20 66 6F 78 20 6A 75 - 6D 70 73 20 6F 76 65 72 n fox jumps over
008f3320 20 74 68 65 20 6C 61 7A - 79 20 64 6F 67 00 00 00 the lazy dog...
008f3330 00 00 00 00 00 00 00 00 - 00 00 00 00 00 00 00 00 ................
[dump], when used with upvar arrays or standard variables will also show the
parent linked-list chain. This may be useful to understand how variables work
array 'b' level 3 (0x200bde0) data @33603120 (0x200be30) 7 byte(s)
(Array data not shown. Use array walk arrayname)
Address 0 1 2 3 4 5 6 7 8 9 A B C D E F ASCII
-------- ----------------------- ----------------------- ----------------
020cbe30 3C 61 72 72 61 79 3E 00 20 27 00 78 69 74 20 6C <array>. '.xit l
+ Parents of 3::'b' (0x20cbde0 / 34389472):
+ Value: '<array>'
|
| Lvl::Variable Name This -> Parent Type Const
| ------------------------------------------------------------------
+ 2::b 0x020cbe40 -> 0x020cb890 array False
+ Value: '<array>'
+ 1::b 0x020cb890 -> 0x020c9540 array False
+ Value: '<array>'
+ 0::b 0x020c9540 -> 0x00000000 array False
+ Value: '<array>'
See also: inspect, vars, walk, calldll, dumpvariant, array walk, load, run
dumpvariant address
Defined in lib plugin: ticol_variant.dll
Displays information relating to a variant object which has been returned
from a calldll*variant operation. This operation requires a valid variant
address
Example:
set v [vsplit $qbf] # Split to Variant (SAFEARRAY) array
dumpvariant v
varray unset v
Result:
DUMP: Variant 'v' at 36012656 (0x2258270)
Type is 8204 (0x200c) BYVAL
Variant is an array of type 0xc
Type is 12 byte Variant VT_VARIANT
See also: dump, inspect, walk, vars, varray get, array walk
void [dumpmem address ?count?]
Debugging routine to dump memory contents at the specified location
Attempting to access a forbidden location will result in a trappable error
Example:
lib ticol_calldll # Load plugin
set h [dll_open msvcrt40.dll] # Load the MSVC runtime
set t 0 # Set a value (integer)
set ret [calldll_cdecl $h localtime t] # Call 'localtime'
dumpmem $ret 32 # Dump 32 bytes of the result
dll_close $h # Release the MSVC runtime
To dump the binary contents of a Tcl variable use [addressofb] as follows
set s "Hello world"
dumpmem [addressofb s] 16
See also: dump, dumpvariant, walk, array walk, vars, info, introspection
Ticol Editor and Recommended IDE |
Index |
Notepad++ is the recommended editor:
https://notepad-plus-plus.org
You can add the TCL syntax by copying the output of: puts [commands]
Next, open Notepad++ and go to:
Settings->Style Configuration->Tcl (from list)
->Instruction Word (from list)
and then paste the list of functions into the
"User defined keywords" area
Executing Ticol Scripts from Notepad++
--------------------------------------
* Click the "Run" menu option on the Notepad++ main menu
* Select "Run..."
* Click the elipsis button (...) to browse to the location of ticol.exe
and select it
* Add a space to the resulting command line and append the following,
wrapped in double-quotes: "$(FULL_CURRENT_PATH)"
* You should now have a command something like
C:\path\to\ticol.exe "$(FULL_CURRENT_PATH)"
* Click "Save..."
* Enter an appropriate name, e.g. "Ticol Tcl"
* If desired select a shortcut such as CTRL+F1
* Click "OK" to save
* You should now be able to run the current script using your shortcut or
the menu
* You may need to add /NA to the command string to prevent autoexec.tcl
from running, e.g.
C:\path\to\ticol.exe "$(FULL_CURRENT_PATH)" /na
* You may also need to add /PAUSE to halt the screen before closing
C:\path\to\ticol.exe "$(FULL_CURRENT_PATH)" /na /pause
* Note that command-line editing is non-existent via the menus, if you
wish to change the command line you will need to delete the shortcut and
recreate from scratch
See also: Ticol, ticol.ini
echo value|varname ?-var? ?-hex?
Print a string without backslash escape translation other than performed by
the Macro PreProcessor when the script is loaded. Echo is marginally more
efficient than either [puts] or [printf] and is useful for debugging string
translation, escaping etc.
The -hex argument will output hex character pairs
Additionally, [echo] with -var can output a string which contains Tcl code
without the $ dereference causing the contained code to be executed.
[echo] is equivalent to: puts [set value] except that a string value, not a
variable is required as the only argument
[echo] returns the untranslated string intact, and also does not append an
additional carriage return/linefeed pair at the end. Use [newline] or
puts "" after a call to [echo] if you wish to generate a new line
Examples:
echo $env(programfiles)
puts [set env(programfiles)]
puts $env(programfiles)
Results:
C:\Program Files (x86)
C:\Program Files (x86)
See also: puts, printf
elevate (Windows 2000 or Higher) |
Index |
elevate ?scriptname? ?command-line-arg...?
Raise the current Ticol interpreter elevation level if running on Windows
XP, 7 or higher by spawning a new instance of the Ticol interpreter
Elevating executables (other than ticol.exe) is not supported
You will be prompted for UAC if necessary. If the console or launching
process is already elevated then the [elevate] command will be ignored
Examples:
# With no script loaded ########################
# Elevate the current session
elevate
# Elevate session and does not run autoexec.tcl
elevate /na
# With script loaded ###########################
# Elevate the loaded script and run it
elevate
# Elevate and run the loaded script, doesn't run autoexec
elevate /na
# Elevate the script 'foo.tcl'
elevate foo.tcl
# Elevate the script 'foo.tcl' (filetype is optional)
elevate foo
# Elevate the script 'foo.tcl' with no autoexec.tcl
elevate foo /na
# Elevate the script 'foo.tcl' with 3 arguments
elevate foo arg1 arg2 arg3
# Elevate the script 'foo.tcl' with 3 arguments, no autoexec
elevate foo arg1 arg2 arg3 /na
[elevate] can elevate any script which has been pre-loaded using [load]
or will elevate specified scripts together with their optional arguments.
[elevate] can also be used to run a script in elevated mode from the
Ticol CLI
Prompts depend on the current UAC setting to determine wither or not
user-interaction may be required to grant elevated permission. Ticol
cannot override this,
[elevate] returns a boolean indicating whether the session is/was elevated
successfully
To exit an elevated Ticol CLI with no active script:
type exit and press ENTER
Elevation will work differently between Windows XP and 7 and this also
depends on the service pack level for Windows XP. To kill an elevated
instance of Ticol you will need to run pskill.exe from an elevated command
prompt. Elevation may block 64-bit console applications from running
properly on a 64-bit system.
Use [is_elevated] to test whether or not a script is elevated
The following snippet of code is usually all that is required to elevate
and continue running a script...
if {! [is_elevated]} {
elevate # Elevate and re-call with any args
exit # We continue, when elevated (not here)
}
As the spawned, elevated instance of Ticol will close automatically on
exit you may wish to use this snippet to prevent the script from closing
option expression off
if {is_elevated} {
pause
}
Example:
puts [elevate]
Result:
1
Example:
if( ! [is_elevated] } {
elevate
} else {
puts "[escape [info script] is running elevated"
# ...
pause
}
Result:
# If UAC is accepted then will display
<path-to-script> is running elevated
Example:
load test # Contains [elevate] command as shown above
run
Result:
# The script is run in elevated mode
Example:
load tests # No [elevate] command present
elevate # Elevate and run
Result:
# The script is run in elevated mode
Chaining $argv to an elevated script
------------------------------------
The expansion operator can be used to pass a concatenated $argv() array
if {! [is_elevated] } {
set cmdline ""
if { [array is_set argv] } {
for {set i 0} {< $i $argc} {++ i} {
append cmdline "$argv($i) "
}
}
elevate {*}$cmdline
} else {
newline
textcolor green
puts "* [escape [info script]] is now elevated!\r\n"
textcolor
# ... further commands ...
}
See also: is_elevated, is_admin, spawn, exec, shell
Elevate Example Script (Windows 2000 or Higher) |
Index |
The following script will create a combined Tcl and Windows batch script
which, when run will elevate a Windows executable (with optional arguments)
A script created by tcl2bat.bat is insufficient to perform this task
Use as (example):
elevate.bat "explorer.exe /e /select,c:\windows"
elevate.bat "notepad %temp%\hello.txt"
elevate.bat "ticol.exe /na"
Script:
::set - {
@echo off
REM ####################################################################
REM File: elevate.bat
REM Purpose: Command-line self-chaining batch script for Ticol Tcl
REM Language: Windows BATCH script and Ticol Tcl v1.26+
REM ####################################################################
set SCRIPT=%~dp0%~nx0
ticol.exe %SCRIPT% /na %1 %2 %3 %4 %5 %6 %7 %8 %9
exit /b 2> nul
REM Windows NT4.0 does not support exit /b so we must goto end
goto end
REM End Batch ##########################################################
}
option expression off # Use non-expr style
if {! [is_elevated] } {
set cmd ""
if { [array is_set argv] } {
for {set i 0} {< $i $argc} {++ i} {
append cmd "$argv($i) "
}
}
elevate {*}$cmd
} else {
set cmd ""
if { [array is_set argv] } { # Array to list
for {set i 2} {< $i $argc} {++ i} {
set arg_v [array item argv $i {}]
if {! [eq $arg_v ""]} {
append cmd " $arg_v"
}
}
}
# We MUST unescape any outgoing path otherwise path supplied
# to Windows will be incorrect
exec {*}[unescape $cmd]
}
/* # End Tcl / EOF #
:END
REM */
See also: elevate, is_elevated, escape, exec, spawn, array, append,
expand, {*}
Executable-Only Code Obfuscation (ticol.exe filename /c) |
Index |
* For the ticol_tcx.dll plugin see: tcx plugin
Code obfuscation/encryption or code-protection is available. This encodes
a Ticol source file to prevent inspection and tampering. You should only
run protected and unprotected Ticol scripts from a source you trust.
Code obfuscation is designed to deter casual tampering, not a determined
or professional hacker. The encryption mechanism used is not industrial
strength but nor is it trivial. However, Tcl naturally exposes portions
of any script so parts of a loaded script may still be visible. TCX
encoding merely protects the source script from casual tampering.
This option can apply only to a single Ticol file module. TCX files can be
run either from the command-line or from the Ticol CLI console
Where an output isn't specified when encoding, the filename will be used
with the filetype-suffix ".tcx". When running .tcx files the filetype must
be given
The code is obfuscated and randomised. e.g. (hex dump)
000000 2C 15 72 09 1F 46 1B 40 - 78 77 66 0C 0F 7F 7F 0A ,.r..F.@xwf..+
000010 29 7A 3F 4E 62 6A 2A 33 - 7B 05 6E 0F 41 77 16 3A )z?Nbj*3{.n.Aw.:
000020 7C 2D 6E 13 68 6F 4C 15 - 7B 77 5F 40 58 28 13 7D |-n.hoL.{w_@X(.}
000030 29 21 47 60 72 22 76 71 - 1A 1E 75 2A 39 34 47 4A )!G`r"vq..u*94GJ
Advantages of Obfuscation
-------------------------
Scripts may be emailed or transmitted securely. The contents are tamper-
proofed which means their safety can be verified if the source is known
and trusted and if checked using MD5 signatures. Source code and IP is
not divulged to third-parties. MD5 signatures should always be used to
check
Macro PreProcessing, other than processing backslash escape sequences is
not performed in advance. These are not processed as the MPP will be run
a 2nd time on the TCX file and strings with resolved \n sequences would be
invalid. Advanced processing enables advance checking of basic syntax
before runtime.
Comments are removed in TCX files, leading to smaller files
Disadvantages of Obfuscation
----------------------------
Obfuscated scripts could hide malicious code if run from untrusted sources
You should never run a tcx script from an untrusted source nor if the MD5
is not verified. Preferably reserve protected code for internal use only.
Treat protected Ticol scripts the same as unverified EXE files
Use the sample MD5 check BAT file as a specimen for checking against a
list of known scripts
Caution: The macro, "#exec" can cause malicious scripts to be executed not
only when scripts are run but when they are encoded using /C since the
Macro PreProcessor will be invoked when encoding unless /NP is used
You should only run (including obfuscate) scripts you trust and preferably
only ones you generate and lock-down yourself
[info body] can reveal [proc] source information at runtime
Caveats on Code Protection
--------------------------
Although the encryption type can be divulged, protected scripts cannot
be decrypted back to the full, original source. The /ECHO command is
disabled. This is by design. You MUST retain unencrypted copied of any
encrypted source code
When running in breakpoint/debug mode the individual commands are shown
at each breakpoint but the script summary is unavailable for TCX files.
Likewise, the commands [commands], [functions] and [procs] will reveal
the names of loaded code but not the code itself
As the command [info body] is necessary for some scripts to run it is
not blocked and will reveal code snippets for procedures only. The main
code and its full structure are not revealed
Security Verification
---------------------
The validity of any decoded TCX block or file should be tested by checking
the code's MD5 value to ensure it matches what is expected.Use an external
MD5 checking application or Ticol's /MD5 argument to generate a checkable
MD5 signature.
Example:
ticol.exe /md5 test.tcx # Generate an MD5 for an existing file
ticol.exe foo.tcl /C # Generate foo.tcx encoded file
ticol.exe foo.tcl /C /64 # Generate TCX block file (foo.t64) and MD5
Result:
AC6960741A575B1C0928EEE817D72272
* Encrypt: Successfully created obfuscated output file
C:\tcl\foo.tcx
MD5 is: 0823aceda778c586e1ee072c2f7d9919
* Encrypt: Successfully created obfuscated output file
C:\tcl\foo.t64
MD5 is: 14701fd2324ee68888b1238808ad3ae8
See TESTMD5.BAT for a specimen script to automatically check a TCX file
before executing
Encryption Syntax
-----------------
ticol.exe tclfile.tcl ?output-file? /C?:<option>? ?/PW:<password>?
The command-line option /C takes the following arguments
Case is ignored for all options other than /C:PASSWORD
/C:PASSWORD Requires /PW:<password> to be successfully encrypted
/PW Values are optional for the remainder and if omitted,
Local workstation/user details, Tcl version will be used
/C Automatic encryption. No run restrictions. Can be run on
any workstation and by any user. Source cannot be viewed
or edited
/C:VERSION Encrypt to be run only with a specific version of Ticol
/C:PASSWORD Uses a password specified using /PW:<password>
Will not execute unless password is specified
/C:MAC Encrypt on the basis of a MAC address which must be
present on the executing system
/C:IP Lock to an IP address specified by /PW
All valid addresses are checked on decryption
/C:USER Encrypt to be run by a current Windows username
/C:HOST Encrypt to be run on a specific workstation name
/C:WINVER Encrypt to be run on a specific windows version (x.xx)
Controlling WinVer execution is better done in real time
/64 Base64 encode TCX output when used with /C (above)
and save as a text file with a '.t64' filetype
/I Will display the encryption mode info for a protected file
/MD5:<filename> Produce an MD5 string for any file. This may be used by
an external script to verify a Ticol file which is about
to be executed
Operation Syntax
----------------
No arguments are required to run a protected file except /C:password
ticol.exe filename.tcx # Run a default-encrypted file
ticol.exe filename.tcx /pw:secret # Run a password protected file
ticol.exe filename.tcx /i # Show obfuscation info
ticol.exe /md5:filename.tcx # Show a tcx file's MD5
The encryption type is automatically detected. The only mode for which a
password can be specified is that encrypted using /C:PASSWORD mode
tcx files may be associated with ticol.exe to run as executable files. e.g.
C:\>hello.tcx
Hello World
in such cases command-line arguments my not be passed properly
Code Protection Examples
------------------------
ticol.exe foo.tcl /c # Obfuscated (no run restrictions)
ticol.exe foo.tcl /c:mac /pw:01020365A690 # Run only on MAC 01020365A690
ticol.exe foo.tcl /c:user /pw:admin # Run only for user 'admin'
ticol.exe foo.tcl /c:host /pw:server1 # Run only on host PC server1
ticol.exe foo.tcl /c:version /pw:1.27 # Run only on Ticol v1.27+
ticol.exe foo.tcl /c:password /pw:secret # Run only with this password
Result:
Creates 'foo.tcx' which can be run using the command:
ticol.exe foo.tcx
The MD5 signature of the new file will be displayed if successful. This
should be stored if necessary for use with script validation
Creating Base64 Blocks for the [tcx] Command
--------------------------------------------
Adding the /64 argument to /C will output to a file in base 64-encoded
format suitable for use with the [tcx] command provided by the
ticol_tcx.dll plugin. See also: tcx plugin
CGI Mode
--------
TCX scripts may be used as web-server CGI applications. The embedded
TCX block may be wrapped to a fixed line length and delimited with
newline characters if desired. These will be ignored
TCX blocks embedded in a TCL script may be executed using [tcx run]
or decoded using [tcx decode] and then run using [eval]
Example:
set s "\
bOevxqW17U\
eCvogqO3Ju\
4c5Asi/FpO\
UdXWNsHPtH\
QfHX6HdMut\
Fr4JT82PtF\
oOA+Ax53YI\
wheQsssDIq\
A3g="
Security
--------
Cryptographic features in this program are not cryptographically secure
and should not be relied on to secure confidential information
Obfuscation or Encryption?
--------------------------
You can decide which term is more appropriate
http://csharptest.net/865/obfuscation-is-not-encryption/index.html
See also: tcx plugin, run, load, Ticol, md5, md5_file
Pseudo-Assembler Interpreter Plugin (TicASM) |
Index |
Load using: lib ticol_asm?.dll?
string [asm string ?-debug?]
This plugin allows you to interpret 'assembler-like' ASM scripts
The assembler interpreter is reasonably fast. There is a Tcl script
version of this interpreter (asm.tcl) which is about an order of
magnitude or two slower than the Ticol plugin DLL version
The language is a pseudo-assembler it has an assembler-like language
and works very much like a normal assembler but is not bound by CPU
limitations, and registers are able to operate using float/double
values as well as integer. It has less functionality than a real
assembler and it is important to note that some of the constructs
and paradigms differ from those of a real assembler. In particular,
dereferencing [VAR] or [REG] differs
Performance
-----------
A minimalist language can perform quite efficiently. A benchmark fractal
script takes 320ms for a compiled EXE, 1605ms for a Ticol interpreted
script and only 400ms for a TicASM interpreted script. Which shows that
the processing overhead is minimal compared to native machine code
Advantages
----------
* A useful prototyping and experimental tool
* May prove useful as a learning tool
* ASM code can run much faster than Ticol TCL
* Errors trace opcodes and line numbers for debugging
* Would probably not be too difficult to write a compiler for TicASM
* External DLLs may be linked directly from ASM
* The Ticol Tcl interpreter can be called from ASM
* A small number of CRT functions are emulated in CALL
* External __cdecl DLLs may be called e.g. msvcrt40.dll
Disadvantages
-------------
* Experimental, not for production use
* Assembler is hard, code is bulky and extremely verbose
* There is no sophisticated debugger for this version
* Can't compile this language to machine code
* Difficult to debug and write error-free code
* Lack of full API/library support although Ticol libs can be called
* An 'assembler-like' language, not a true assembler
* Opcodes differ from those of a real physical CPU
* Paradigms differ in some subtle ways from 'real' ASM
The following opcodes are supported:
ADD AND CALL CMP DB DEC DUMP END FADD FCMP FDIV FMOD FMUL
FPOP FPUSH FSUB HALT INC INT JG JGE JL JLE JMP JNZ JZ MOV
MUL NOP NOT OR ORG POP PUSH RET SHL SHR SUB TRNC XOR
The language can be extended in the future with no runtime overhead
There are 4 64-bit/float/real registers: AX, BX, CX, DX,
a data index pointer EDP, which references the memory block, and
instruction pointer EIP. EIP can't be modified but EIP can be
addressed the same as other registers
Memory is allocated via a 'tape' of 64Kb similar to 'Brainfuck'
and is indexed by EDP. A separate execution stack of 32kb is kept
for PUSH, POP, CALL and RET (CALL pushes the return address to
the stack, RET will POP it off and resume at that address)
The data stack works in reverse order to normal. Additional memory
is used by adding to EDP rather than subtracting
Comments are supported, prefixed by a semi-colon and C++ style long
/*...*/ comments are supported
A __LINE__ macro will substituted into the source code line as an integer
ORG is supported and it understands ORG <hex>h, decimal and 0xNN
The J* instructions currently understand only labels so ORG is ignored
There are no cast modifiers to opcodes (qword, dword, byte ptr) etc.
Float values may be truncated (rounded), to integer if necessary using
the TRNC instruction. Otherwise a register will store an 8 byte double
value where a floating point number is detected. Future versions of
the language may have more sophisticated number type casting. Double/
float values can be stored only in literals, registers or as BCD strings
mov ax, 123456 ; AX stores 64-bit integer
mov ax, 123.456 ; AX stores 8-byte double
Literal integer values may be presented as Hex (0xNN),
Octal (0oNN) or binary (0bNN), e.g.
mov ax, 23
mov ax, 0x17
mov ax, 0o27
mov ax, 0b10111
Three standard Tcl/C++ escape characters are understood in literal
strings used by DB, these are \t (TAB), \r (0x0d), \n (0x0a)
e.g. db ? "Hello\tworld!\r\n"
Loading
-------
The source is 'compiled' to a bytecode memory image at load time
Basic checks are made on syntax etc. No further interpretation is
performed at runtime other than variable lookup
Returning Values to Tcl
-----------------------
Returns can be made to Ticol Tcl via a BCD string indexed via the AX
register. To avoid 'accidental' returns It is suggested that AX is cleared
on exit using xor ax,ax, and the first byte of the data table is set to
NULL using: db ? "" as the first entry
lib ticol_asm
buf: ; Declare variable buf
mov buf, edp ; Set buf address
db ?, "12345" ; Store a value at buf address
mov ax, buf ; Return address of buf in AX
Calling Tcl Commands
--------------------
Tcl routines can be called from TicASM using 'call eval'
A buffer of sufficient size to handle the result must be declared and
the buffer plus length arguments pushed in right-to-left order on the
stack before calling
lib ticol_asm
asm {
expr: ; Define a variable
buf: ; Define a variable
mov expr, edp
db ? "puts [/ 22 7.0]"
mov buf, edp
db 256, ""
push 256 ; Push the buffer size
push buf ; Push the buffer address
push expr ; Push the expression address
call eval ; Call the Ticol Tcl [eval] command
mov ax, buf ; Print out the result: '3.142857142857143'
call 9
}
lib ticol_asm -unload
Calling External DLLs
---------------------
External DLLs may be called. Subject to the limits of the interface
e.g. MSVCRT40.DLL has been tested. (See: asm instruction set)
Precision
---------
The underlying data type is IEEE double. Integer precision is limited by
double (float) representation to 52 bits or 16 digits. This is less than
that offered by 64-bit integers but represents a compromise. No exceptions
are generated where precision is lost. Code should check/guard against
Performance
-----------
For comparison. A standard Mandelbrot fractal test which takes around 250
to 300 milliseconds as a compiled EXE takes about 420 milliseconds to run
in TicASM as an ASM script. The ASM.TCL script will run the same ASM
script in about 13 minutes The corresponding Tcl version of the same
code runs in 1760 ms which means that for many tasks ticol_asm.dll could
be much faster
Standard Mandelbrot Test
Approx Implementation Type
Time/ms
300 Compiled EXE
420 Ticol ASM Plugin running fractal.asm
1,760 Ticol interpreting fractal.tcl
780,000 Ticol asm.tcl interpreting fractal.asm
The specification is experimental and subject to revision and improvement
If curly-braces are avoided in the source then the entire source, including
double-quotes may be braced and set within a Tcl variable and included
within a Tcl script, the return from which can be returned to Tcl
Comments and whitespace affect only load-time not run-time performance
Example Calls:
lib ticol_asm # Load the interpreter plugin
set script [readfile fractal.asm]
asm $script
lib ticol_asm -unload
lib ticol_asm # Load the interpreter plugin
asm { ; We are now using ASM syntax
msg:
mov msg, edp ; Point var 'msg' to current EDP location
db ? "Hello world\n" ; Initialise buffer space
mov ax, msg ; Point AX to the address in 'msg'
call 9 ; Call print function
}
lib ticol_asm -unload # Unload the interpreter plugin
Example Code:
lib ticol_asm # Load the ASM interpreter plugin
set r [ # Note square bracket here
asm { ; Approximation to PI
buf: ; Define a variable
mov buf, edp ; Set buffer address
db 21, "" ; Define buffer storage
mov ax, 22 ; AX = 22
mov bx 7 ; BX = 7
fdiv ax, bx ; Float divide AX / BX
mov [buf], ax ; Copy AX to string buffer
mov ax, buf ; Index AX to buffer
end ; and return
}
] # Note square bracket here
assert $r {== $_ 3.142857} -v -line #__LINE__
lib ticol_asm -unload ; Unload the interpreter plugin
Example Code:
lib ticol_asm # Load the interpreter plugin
asm {
org 100h ; Example only, not required
; as we won't JMP < 0x100
msg: ; Declare a variable/label
buf: ; Declare a variable/label
db ? "" ; Ensure AX is not printable on return
mov msg, edp ; msg points to this string (EDP)
db ? "Loop: " ; Define a string (6 bytes+null)
mov buf, edp ; buf points to EDP
db 21, 0 ; 21 byte buffer to store number
mov ax, 4 ; Initialise AX
mov bx, 0 ; Initialise BX to counter
L1: ; Declare a label
push ax ; Save AX
mov ax, msg ; Point AX to our message
call 9 ; Print string
inc bx ; Increment BX
mov [buf], bx ; Store BX value in our buffer
mov ax, buf ; Point AX to our printable buffer
call 9 ; Call print address at AX
mov ax, 10 ; CR (newline)
call 2 ; Call print char
pop ax ; Restore AX
dec ax ; Decrement AX
jnz L1 ; If not zero then loop L1
xor ax, ax ; Clear AX
end ; Finish
}
lib ticol_asm -unload # Unload the interpreter plugin
Result:
Loop: 1
Loop: 2
Loop: 3
Loop: 4
Execution took 1 ms
Development
-----------
TicASM is in a relatively early state of development
There is potential to write a Tcl to ASM transcoder
There is also potential to write a true assembler for the language
For more information, see the separate documentation
See also: asm instruction set, plugins, brainfuck
TicASM Instruction Set (Plugin) |
Index |
See: Ticol Assembler
The opcodes are similar to those in a normal CPU based assembler but the
goal is not the compilation of object code but fast execution of an
interpreted script language at as close a speed as possible to that of a
compiled executable
Things to note:
* This is not a real assembler. Syntax will differ
* Registers may contain either float (real) numbers or 64-bit integers
* Variables and labels are interchangeable (same object type)
* Variables and labels must be declared and are declared by first
reference in the source with a colon suffix. e.g. var:
* Variables are integer-only and cannot contain float (real) numbers
* Registers and memory storage may contain strings and real numbers
* A variable may point to a location with allocated memory which can be
dereferenced to retrieve or set the contents of that location
* Variables should be initialised to zero using, say, MOV VAR, 0
* An allocated memory location may contain either a integer or float
number, or an ASCIIZ string
* Dereferencing a register will retrieve or set a single byte at
the address location referenced by the register. Dereferencing a
variable will fetch a BCD-translated number (integer or float)
[REG] Fetch a single char
[VAR] Fetch an integer or float value (if translatable)
* Both source and target operands determine whether the result will be
restricted to integer. e.g. MOV VAR, [VAR] is restricted by target type,
MOV [VAR], VAR is restricted by source type
* Memory is allocated using DB
As with C++, a 1 byte is added to the requested storage size, which will
contain a NULL byte terminator character (0x0). This should be allowed-
for if incrementing memory address pointer locations
* A variable may be incremented, decremented or set to point to any valid
memory location. You are responsible for overflow checking
* Fixed, automatic and fill modes are supported by DB
DB 100, "x" ; Allocate 100 bytes, repeat fill with char "x"
DB 100, "foo" ; Allocate 100 bytes, repeat fill with string "foo"
DB 3, "foo" ; Allocate 3 bytes, fill exactly with string "foo"
DB 100, "" ; Allocate 100 bytes, repeat fill with NULL chars
DB ?, "Hello" ; Allocate 6 bytes and fill exactly with "Hello\0"
DB AX, "x" ; As for DB 100,"x"
A buffer can't be allocated to a size larger than a fill string with
zero fill repeat. In such cases, allocate to NULL ("") and copy later
* A DB fill of NULL bytes is achieved using a null string "" argument
* After memory is allocated using DB, the EDP register will be
incremented to the next location after the NULL terminator character
* Items are stored in memory locations as ASCIIZ strings
* CALL can only call a label value or a preset library function at present,
not an arbitrary integer address within the source
* ORG may be set but currently has no usage
* OP [VAR], [VAR] performing location-to-location byte transfers is not
supported. Use intermediate registers to facilitate memory content
transfers. Character strings must be copied one character at a time using
[REG] to dereference the character. (Alternately use: call strcpy)
mov ax, [var1] ; Copy a BCD double or integer
mov [var2], ax
; Copy x1 char from var1 to var2
mov ax, var1 ; Point AX to source location
mov bx, var2 ; Point bx to target location
mov cx, [ax] ; Fetch x1 byte from source to CX
mov [bx], cx ; Write x1 byte from CX to target
[VAR] Dereference the integer variable address VAR to access
the contents in allocated memory
[REG] Retrieve a single byte from the address location indexed
by REG
* Literal values may be decimal, hexadecimal (0xNN), octal (0oNN), or
binary (0bNN). The ORG instruction understands <HEX>h
* TRNC (truncate) is an approximate rounding which will truncate a float
value to N decimal places. TRNC is useful with FCMP to bound a value to
some known, comparable range
* Comma separators between operands are not mandatory, 'MUL AX 12' is ok
* The underlying stack type is a 64-bit integer stack
* Float values can be stored on the stack using FPUSH, FPOP
* FPUSH *must* be followed by FPOP unless the result is to be discarded
in which case you can POP into a discarded register
The following opcodes are supported:
Instruction Description
-----------------------------------------------------------------------
; Comment delimiter
Add: Integer addition of two values (See FADD for float)
ADD REG1, REG2 Integer add register 2 to register 1
ADD REG, N Add integer value N to register
And: Bitwise AND two integer values
AND REG1, REG2 Bitwise AND register 1 with register 2
AND REG, N Bitwise AND register with value N
Call Label Execution jumps to the specified label
The return address is pushed onto the stack
Calling functions should preserve the return address
CALL <function> Call an emulated "C" runtime function
CALL dll_open Open (load) a standard __cdecl DLL
CALL dll_run Execute a function call in a DLL loaded by dll_open
CALL dll_close Close a DLL loaded by dll_open
CALL N Call emulated O/S function
Db: Define bytes (allocate memory) as follows:
DB ? <string> Allocate string length bytes and copy string
DN N <string> Allocate N bytes, copy as much of string as will fit
If string is shorter then the string is repeated
DB REG, <string> Allocate REG bytes. Behaviour as for DB, N
DEC REG Integer decrement the contents of register REG
DUMP Display register and variable contents etc.
END Immediately end script processing
Fadd: Add two float values
FADD REG1, REG2 Float add registers 2 to register 1
FADD REG, N Add value N to register contents
FADD REG, VAR Add the integer contents of VAR to register REG
Fcmp: Compare two float values. FCMP is an alias for FSUB
FCMP REG1, REG2 Compare register 1 with register 1 (corrupting 1)
Fdiv: Float divide A by B, giving A
FDIV REG1, REG2 Float divide REG1 by REG2, giving REG1
FDIV REG, N Float divide REG by N, giving REG
FDIV REG, VAR Divide register REG by the integer contents of VAR
Fmod: Modular float divide two values and return the result
FMOD REG1, REG2 Modular float divide register 1 by register 2
FMOD REG, N Modular float divide register by value N
FMOD REG, VAR Modular float divide register by integer variable
Fmul: Float multiply A by B, giving A
FMUL REG1, REG2 Float multiply register 1 by register 2
FMUL REG, N Float multiply register by value N giving register
FMUL REG, VAR Multiply the integer contents of VAR to register REG
FPOP REG POP stack to REG as float. Must be pushed with FPUSH
FPOP VAR POP stack to VAR as float. Must be pushed with FPUSH
FPUSH REG PUSH float REG to stack. Must be popped with FPOP
FPUSH N PUSH float N to stack. Must be popped with FPOP
FPUSH VAR PUSH float VAR to stack. Must be popped with FPOP
Fsub: Subtract two values
FSUB REG1, REG2 Subtract value of register 2 from register 1
FSUB REG, N Subtract value N from register
FSUB REG, VAR Subtract an integer variable from register
HALT Pause execution
INC REG Integer increment a register value
INT N Emulated/simplified O/S interrupts (See below)
J* label: Jump operations can jump only to a label not an integer
JG LABEL Jump to label if last op was > than zero
JGE LABEL Jump to label if last op was >= than zero
JL LABEL Jump to label if last op was < than zero
JLE LABEL Jump to label if last op was <= than zero
JMP LABEL Unconditionally jump to label
JNZ LABEL Jump to label if last op was == zero (Z flag set)
JZ LABEL Jump to label if last op was != zero (Z flag not set)
Mov: Move data from one location to another
MOV REG, REG Move contents of the right hand register to the left
MOV REG, N Move the float or integer value N to register
MOV REG, VAR Move the integer contents of variable to register
MOV REG, [VAR] Move float or integer value from variable to register
MOV VAR, REG Move the contents of register to variable as integer
MOV VAR, N Move value N to variable as integer
MOV VAR1, VAR2 Move value from variable 1 to variable 2
MOV VAR1, [VAR2] Move value at variable 2 to variable 1
MOV [VAR], REG Move value of register to variable at location VAR
MOV [VAR], N Move value N to location at VAR
MOV [VAR1], VAR2 Move value from variable VAR2 to location at VAR1
MOV [VAR1], [VAR2] Not supported. See mov [var], var; or mov var, [var]
May be implemented later as a byte copy
MUL A, B Integer multiply A by B
MUL REG1, REG2 Integer multiply register 1 by register 2
NOP No operation (ignored)
NOT REG Bitwise invert integer in register
NOT VAR Bitwise invert integer value
Or: Bitwise OR two integers
OR REG1, REG2 Bitwise OR integer register 1 with register 2
OR REG, N Bitwise OR integer register with value N
OR REG, VAR Bitwise OR integer register with integer value
ORG N Set the code origin (once only per script)
POP REG Retrieve value from stack and assign to register
PUSH REG Push register value to the stack. Leave REG unchanged
RET Return from CALL. Pop the return address from the stack
Shl: Integer bit shift left two integers
SHL REG1, REG2 Integer bit shift left register 1 by register 2
SHL REG, N Integer bit shift left register by integer value N
SHL REG, VAR Integer bit shift left register by integer variable
Shr: Integer bit shift right two integers
SHR REG1, REG2 Integer bit shift right register 1 by register 2
SHR REG, VAR Integer bit shift right register by integer variable
Trnc: Truncate value A at B digits after the decimal point
TRNC REG1, REG2 Truncate float value in register 1 by register 2
TRNC REG, N Truncate float value in register by value N
TRNC REG, VAR Truncate float value in register by integer variable
Sub: Integer subtract 2 items
SUB REG1, REG2 Integer subtract register 2 from register 1
SUB REG, N Integer subtract value N from register
SUB REG, VAR Integer subtract value in variable from register
Xor: Integer XOR two items
XOR REG1, REG2 Integer xor register 2 with register 1
XOR REG, N Integer register 1 with value N
XOR REG, VAR Integer register with value in variable
Math opcodes currently operate only on registers not variables
The emulated O/S layer is in very early stages of development and may be
changed substantially to bring it more into line with real API code
Abstract Data Types (ADTs)
--------------------------
Arrays, structs, stacks, heaps, queues etc. must be emulated in the same
way as they would using any assembler. It is not anticipated that TicASM
will be used for particularly complex programs which feature such ADTs
Emulated "C" (CRT/C Runtime) Functions
--------------------------------------
A small number of essential C runtime functions are emulated and can be
called directly using CALL. The function prototypes are the same as
for the C functions except that strchr, and strstr return an integer
character position not a char* offset. For features not linked directly
it is possible to call MSVCRT.DLL via calldll_*
Unless stated, any required argument(s) should be pushed onto the stack in
__stdcall reverse order (right to left). These functions will clean up the
stack. A return value (if any) will be in register AX
dtoa Convert double (float/real) to string
Buffer should be at least 21 bytes
dtoa(integer buffer buflen)
fpush: buflen, push: buffer double
flush Flush stdout, stdin and stderr
No arguments, no return
free Free memory
AX: Address value returned by malloc
No return
eval Call Tcl [eval] to run a Tcl command / script
Returns string result in a buffer
Defaults of 0 (NULL) may be passed for buffer/length
integer eval(argument-string buffer buflen)
push: argument-string buffer buffer-length
AX: Returns Tcl error code
getvar Get a simple Tcl variable (arrays not supported)
string getvar(varname buffer length)
push: length buffer varname
AX: Returns buffer address (NULL on error)
gotoxy Locate the cursor at X, Y
void gotoxy(x y)
push: y x
No return
itoa Convert integer to string
Buffer should be at least 21 bytes
itoa(integer buffer buflen)
push: buflen buffer integer
malloc Allocate memory
AX: Bytes to allocate
AX: Returns memory handle/address (0 on failure)
memset Erase (set) a block of memory
void memset(address char length)
AX: Memory address to write
BX: Char value to set (integer)
CX: Length of block to write
AX: Return from malloc
printf Print an ASCIIZ string
integer printf(string ...)
AX: Returns index to string
setvar Set a Tcl variable
varname must be a declared memory location
integer setvar(varname value)
push: value varname
AX: Returns Tcl error code
strchr Index presence of a character in a string
String must be a declared memory location
integer strchr(string char)
push: char string
AX: Returns base 1 integer index
strcmp Compare one string to another (case-sensitive)
strcmpi Compare one string to another (case-insensitive)
Strings must be a declared memory location
integer strcmp(string compare)
push: compare string
Returns standard "C" integer
0 equal, >0 greater-than, <0 less-than
strcpy Safe copy bytes from one location to another
Source and target must be declared memory locations
The new target string length is returned
integer strcpy(target source length)
push: length source target
strlen Return the length of an ASCIIZ string
string must be a declared memory location
integer strlen(string)
push: string
AX: Returns integer string length
strstr Index presence of a substring in a string
string must be a declared memory location
integer strstr(string substring)
push: substring string
AX: Returns base 1 integer index
unlink Delete a file
string must be a declared memory location
filename must be a declared memory location
integer unlink(filename)
push: filename
AX: Returns boolean success
External DLL Linkage
--------------------
Primary arguments are passed in registers (AX..DX),
Optional arguments are pushed onto the stack
The functions perform internal cleanup of the stack (__stdcall)
call dll_open ; Load an external __cdecl DLL
; DLL name in AX, handle returned in AX
call dll_close ; Close the handle of an open DLL
; DLL handle in AX, boolean success in AX
call dll_run ; Execute a DLL function which returns a long/int
; Result may be a string address
; AX Handle
; BX Function name
; CX Arg count (not including fn name)
; <stack> args in reverse order (R to L)
; AX integer return
call dll_run_double ; Execute a DLL function which returns a double
; As for dll_run: AX double return
call dll_run_string ; Execute a DLL function which returns a string
Emulated O/S Functions
----------------------
These require the argument to passed in register AX
CALL 2 ; Print single byte
CALL 3 ; Set the console colour (0..15)
CALL 8 ; Print numeric literal
CALL 9 ; Print ASCIIZ string
Emulated MS/DOS / Windows Interrupts
------------------------------------
These expose "C" interfaces and require the argument to passed in
register AX. Unless specified, any return is in register AX
More features may be added in due course. Some functions are non-standard
int 0x21, <ax> ; (Hex 21, Decimal 33)
int 0x21, 0x2 ; Print single byte indexed by AX
int 0x21, 0x9 ; Print ASCIIZ string indexed by AX
int 0x21, 0x0f ; Open file for sequential read or write
; Filename indexed by BX, mode string CX
; Standard "C" file modes ("rwa+")
; Returns a "C" FILE* handle in AX cast to integer
int 0x21, 0x10 ; Close file opened with 0xf.
; Pass file handle in BX
; Returns boolean in AX
int 0x21, 0x14 ; Sequential read file
; Handle in BX, buffer address CX, length DX
; String buffer will not have NULL termination added
; Byte count returned in AX
int 0x21, 0x15 ; Sequential write file
; Handle in BX, string address CX, length DX
; If DX == 0 then whole string will be written
; Byte count returned in AX
int 0x21, 0x2a ; Get current date as an ASCIIZ string (return in AX)
int 0x21, 0x2b ; Get current time as an ASCIIZ string (return in AX)
int 0x21, 0x2c ; Get the clock time in seconds (return in AX)
int 0x21, 0x30 ; Get clock time in milliseconds (return in AX)
int 0x21, 0x31 ; Get pseudo-random number (PRNG) (return in AX)
The TicASM spec is experimental and subject to revision
See: ticol assembler
Embedding Ticol Tcl in C++ Programs (ticol.dll) |
Index |
There are 2 potential methods of embedding/linking Ticol Tcl into other
language.
1) The use of a C++ static LIB file
2) Using a DLL file
At some future point a mechanism for producing compilable C++ source
code for embedded Ticol Tcl projects may be released, until then it is
possible to call via the available 'ticol.dll' version of the interpreter.
The interface prototypes are not yet finalised and may be subject to
revision.
The static LIB version is working but not yet publicly released since it
requires extensive documentation.
Example Calling ticol.dll
-------------------------
How to call the DLL version of the Ticol interpreter from C++ code...
// Prototype signature
LPSTR __stdcall ticol_command(LPSTR s,
const long opt_breakpoint=FALSE,
const long opt_stacktrace=FALSE,
const long opt_run_autoexec=FALSE,
const long opt_debug_mode=FALSE,
long* result=NULL)
// Example C++ calling ticol.dll (A __stdcall DLL)
// Ensure stack is sufficient for recursion (try 30,000-100,000)
// Loading the DLL dynamically is straightforward
#pragma comment(linker, "/stack:0x300000")
#include <windows.h>
#include <stdio.h>
int main(int argc, char** argv)
{
LPSTR ptr=NULL;
long breakpoint=0;
long stacktrace=0;
long autoexec=0;
long debug=0;
long err=0;
// Embedded Tcl source code:
// DO NOT indent this multi-line literal!
// Some compilers may also restrict the size of string literals
// e.g. MSVC 5.0
LPSTR script="\
textcolor magenta\r\n\
puts \"Hello world\"\r\n\
textcolor\r\n\
";
// Declare funct*
LPSTR (__stdcall* ticol_command)(LPSTR s,long,long,long,long,long*);
HINSTANCE hAPI=LoadLibraryA("ticol.dll"); // Load the DLL
if(hAPI) // Check handle
{
ticol_command=(LPSTR (__stdcall*)(LPSTR,long,long,long,long,long*)) \
GetProcAddress(hAPI,"ticol_command"); // Load the export
if(ticol_command) // If successful
{
printf("test: Loaded DLL ticol.dll->ticol_command()\r\n");
// Call Tcl ...
ptr=ticol_command(script,breakpoint,stacktrace,autoexec,debug,&err);
printf("test: ticol_command: Returned result:'%s' errorcode:%li\r\n"\
,(ptr?ptr:"NULL"),err);
}
else
{
printf("test: Failed to ticol_command in ticol.dll\r\n");
return 1;
}
}
else
{
printf("test: Failed to load ticol.dll\r\n");
return 2;
}
return 0;
}
A Windows test interface is also provided to test ticol.dll
(See dll interface)
See also: dll interface, plugins, Ticol
emulating C++ arrays |
Index |
Emulating C/C++ Arrays When Calling External DLLs
When using [calldll] to link to external DLLs it may be necessary to pass
a C/C++ "pointer array", i.e. an array of pointers rather than a fixed
size array of strings. This sort of array is not natively supported by
Ticol Tcl. However, it can be emulated using Ticol structs providing that
the called DLL treats the passed array as 'readonly'
# Create an array of 4-byte "pointers" in a struct as successive
# addresses with individual elements in, say,, e1..e4
struct a {
e1 4
e2 4
e3 4
e4 4
}
# Create string variables which will be mapped-to by the pointer array
set a1 "The"
set a2 "quick"
set a3 "brown"
set a4 "fox"
# Connect the pointer array elements to the string variables by
# storing the binary representation of the internal pointer address
# using [strptr]
[struct setb] will set a binary representation rather than ASCII
struct setb a.e1 [strptr a1]
struct setb a.e2 [strptr a2]
struct setb a.e3 [strptr a3]
struct setb a.e4 [strptr a4]
# Call 'someFunct' in the DLL, passing the array size and the array
# N.B. the DLL must NOT reallocate the array or expand string contents
set r [calldll my.dll someFunct 4 a]
See also: calldll, arrays, carray, struct, struct setb, strptr
string [encrypt string password]
string [encrypt var password -var]
string [decrypt string password]
string [decrypt var password -var]
Encrypts either a string or variable (if the -var argument is used)
Encrypt a string or variable to a base64 string using moderate-strength
symmetric encryption. Encryption will result in string size increase due
to base64 encoding
[encrypt] should not be used to store highly-vulnerable information such
as credit cards, personal or bank details etc. For high-strength secure
encryption call an external DLL or API which offers AES or similar
asymmetric encryption
[decrypt] decrypts a string or variable from a base64 string which was
encoded using [encrypt]. Decoding will result in a string size reduction
No error is raised where the password is incorrect. The caller should
check any decrypted string and test the result as appropriate.
If you require stronger encryption use [calldll*] to call an external
encryption library
Example:
puts [encrypt "Hello world!" secret]
puts [decrypt [encrypt "Hello world!" secret] secret]
set s "Hello world!"
puts [encrypt s secret -var]
puts [decrypt [encrypt s secret -var] secret]
Result
N4w8n7XnePSvob9G
Hello world!
N4w8n7XnePSvob9G
Hello world!
Security
--------
Cryptographic features in this program are not cryptographically secure
External routines should be called for security cryptography
Properties
----------
The following techincal information may be relevant to anyone writing C
code to link the library
ticol.dll is compiled with the following properties:(* is default)
Stack 0x30000
Processor: Blend*
User lib: single threaded*
Calling convention: __stdcall
Struct byte alignment: 8 bytes*
See also: tcx, base64, mkn, cvn
Embedded TCX Plugin |
Index |
Load using: lib ticol_tcx?.dll?
string [tcx encode <script> ?-password password?]
string [tcx decode "base64-block" -password password]
string [tcx run "base64-block" ?-password password?]
integer [tcx writefile filename <script> ?-64?]
The ticol_tcx plugin allows blocks of obfuscated Ticol Tcl code to be
embedded within windows batch or other Ticol Tcl scripts
Generate encoded blocks using: ticol.exe <filename> /c /64 or by using
[tcx encode] to generate a base64 encoded block of data.
Extreme caution should be used when executing such code from unknown or
untrusted sources. The use of [md5] is recommended to check the code has
not been modified. (See: md5)
[tcx run] can run any base64 encoded TCX block which has been created
using 'ticol.exe /c /64' or using [tcx encode]
Arguments are not directly currently supported via [tcx run]
but you can set argc and argv consts as follows:
# Example setting argc and argv to pass to a [tcx] block
# This example is for Towers of Hanoi (hanoi.tcl) with 15 hoops
unset argc -const -nocomplain
unset argv -const -nocomplain
set argc 2 # 2 arguments to be passed
set argv(0) "" # 1st argument will be the script name (not required)
set argv(1) 15 # 2nd argument will be the ring count
Decrypt failure will return an error code (2) and should be handled using
[try] or [catch]
A companion utility called base64.exe can be used to convert TCX binary
code into base 64 format text for use with [tcx] or the /64 argument can
be used with the /C argument. Base64 encoded TCX code can only be executed
by the [tcx] command, and not by ticol.exe directly unless re-saved as
a binary file.
Example:
# Will read 'filename.tcl' and write 'filename.t64'
ticol.exe filename /c /64
Example:
# Uses default encryption
# Encode using ticol.exe filename.tcl /c /64
# Or using base64.exe filename.tcx to encode
set code "ibuhXiuUJYyD6Uubn7G4L/wRkK/BtOkaHi5IFuULSaPU/3YskM139se
IuesC7KEhQjprZY83PyZAqzgwFmlfnIkpMMmSMvs="
# Load the plugin
lib ticol_tcx
# Execute the code (call [eval] internally)
tcx run $code
Decoded script:
textcolor yellow
puts "Hello World"
textcolor
;
Results:
Hello world
Note that Ticol's string processor doesn't care how the quoted base64
string is formatted. Line-endings are ignored and this 'ragged-line' example
would work just as well.
set code "ibuhXiuUJ
YyD6Uubn
7G4L/wRkK/BtOkaHi5IFuULSaPU/3Y
skM139se
Iues
C7KEhQjprZY83PyZAqzgw
FmlfnIkpMMmSMvs="
Code validation can be performed using [md5]
Example:
# Encode using ticol.exe filename.tcl /c which will also generate the
# MD5 value, and then: base64.exe filename.tcx
set code "<some-block-of-tcx-code-generated-via-the-above-method>"
if {ne [md5 $code] 2e65ca46a5e2e605e5e675d935033c1a} {
textcolor red
die "Error: MD5 validation failed at line "#__LINE__
textcolor
}
# Otherwise, continue and evaluate
tcx run $code
Error Handling
--------------
The -nocomplain argument can suppress error messages. To control program
flow during errors, use [try..catch] as follows
try {
lib foo.dll
} catch {
puts "Can't load plugin foo.dll"
}
Security
--------
Cryptographic features in this program are not cryptographically secure
TCX encoding is intended for obfuscation only
See also: tcx, base64
enum var ?var? ...
enum var = integer ?var = integer? ...
enum {var|var = integer ?var|var = integer? ...}
A simple implementation of enumerated integer constants. You cannot assign
to an enum. Enums will have visibility scope at the level where they are
defined and each member will be treated as a separate const variable
Enums are assigned values in sequence starting at zero to guarantee they
are unique. Names may not have already been assigned to a const or variable
Member names must be unique
Assignment is optionally permitted in any number format which resolves to a
signed integer. The enum increment sequence will continue to increment by
+1 from that number in the next constant. The assignment operator '=' must
be separated by a space or tab character
[enum] avoids more convoluted methods of declaring a unique series of
variables
[enum] const values will typically be created in global scope and can be
accessed from within procedures using the $::constname scope prefix
Enumerated values cannot be assigned but they can be [unset] and re-
declared. You must call as [unset varname -const]
Example:
enum zero one two three
Result:
integer three 3
integer two 2
integer one 1
integer zero 0
Example:
enum zero one six = 6 seven
Result:
integer seven 7
integer six 6
integer one 1
integer zero 0
Example:
enum {quick brown = 3 fox}
Result:
integer quick 0
integer brown 3
integer fox 4
Example: (For Windows MessageBox API)
enum {
MB_OK
MB_OKCANCEL
MB_ABORTRETRYIGNORE
MB_YESNOCANCEL
MB_YESNO
MB_RETRYCANCEL
MB_CANCELTRYCONTINUE
MB_ICONSTOP = 0x10
MB_ICONERROR = 0x10
MB_ICONHAND = 0x10
MB_ICONQUESTION = 0x20
MB_ICONWARNING = 0x30
MB_ICONINFORMATION = 0x40
MB_ICONASTERISK = 0x40
MB_DEFBUTTON1 = 0
MB_DEFBUTTON2 = 0x100
MB_DEFBUTTON3 = 0x200
MB_DEFBUTTON4 = 0x300
MB_APPLMODAL = 0
MB_SYSTEMMODAL = 0x1000
MB_TASKMODAL = 0x2000
MB_HELP = 0x4000
}
Example:
enum zero one two
unset zero -const
unset one -const
unset tw* -const -nocomplain
See also: const, set, unset, vars, walk
error "message" ?info? ?returnCode? ?line?
Raises an error condition
The [error] command may be used in conjunction with catch to display/set
error strings message is an informative error message. info gives more
information about the error. code is intended to be a numeric error code.
Line is optional and non-standard and may report the line number by using
the #__LINE__ macro together with a variable
Example:
option expression off
proc Div {a b} {
set line #__LINE__
if {== $b 0} {
error "Error generated" "Info for error" 1 [- $line 1]
} else {
return [/ $a $b]
}
}
Div 3 0
See also: on_error, preset globals, error codes, try, catch, return
Some common Tcl/Ticol errors;
Example:
my_proc [[string left [getenv PROMPT] 4]]
Result:
eval: No such command or procedure '$P$G'
Fix:
Remove the outer layer of square brackets from the expression, thus:
checkex [string left [getenv PROMPT] 4]
Reason:
The outer brackets will try to evaluate the string which is
returned and it will be executed as a Ticol or Windows command
See also: error, debugging
Ticol error code return values are as follows
These are returned by [catch] but may be passed as option arguments to
other commands such as [return]
0 Success
1 Error
2 return statement met
3 break statement met
4 continue statement met
5 exit statement met
6 stop statement met
7 goto statement met
8 value already defined
See also: error, catch, return
Chaining Tcl Commands |
Index |
Many commands are able to be chained together to form command sequences
Example:
set r [msgbox [join [split $qbf " "] "\n" 768] "Choose" 36]
msgbox "You chose [map {6 Yes 7 No} $r]"
See also: K combinator
string [source filename.tcl|filename.tcx ?arg ...? ?-password <xxx>?]
Evaluate a Ticol tcl or obfuscated tcx script file, retaining the currently
loaded script and active variables. Use [eval] to evaluate an expression
[source scriptname] will execute the named script.
Note: [source] will perform a time-consuming Macro PreProcess on the
script each time it loads. Do use [source] within fast loops
The filetype for normal TCL files is not enforced for [source] but TCX
files will be detected by the filetype 'TCX' and is thus mandatory for
obfuscated TCX files
TCX files which do not require a password may be used if the password is
passed using the -password argument. Care should be taken when placing
high security passwords into Ticol scripts. Such passwords could be
stored in encrypted form and locked to a local machine etc. Alternatively
use a scriptable external app such as pwsave.exe which can retrieve
a stored password locked to a local machine and user account.
[run] will force [option escape] to ON, [source] leaves this setting
unchanged. Entering the CLI will force this option OFF so using [source]
from the CLI will require correctly setting [option escape] in a script
All arguments passed to [source] are passed to the referenced script except
the -password argument will be replaced with "******"
Example:
source hanoi.tcx 17
source pwprotected.tcx -password foobar
Result
Loads, decrypts and executes hanoi.tcx setting argv(1) to 17
Tcl script files may be called/chained from a script by using [source],
with execution returning to the original script. Note that variables for a
chained script exist in the same namespace (true namespaces are not
supported) so care should be taken to avoid clashes in variable names in
the chained script.
[source filename] will not reset the interpreter command count
Example
puts "Calling child script"
source child.tcl
puts "Parent script is still here"
Note that where a script which calls other TCL files is encrypted as a TCX
file, this does not automatically alter those TCL links to TCX format.
If you wish to 'chain' only to TCX files you must write the script to call
TCX versions of a file before encoding
Note that chaining scripts may accumulate uncleared variables. A script
should clean-up its own global variables before exit if it is likely to
be called by another script. Local proc variables are cleared
automatically. (See: clear, unset)
A script file may also be executed by reading the contents to a variable
and then calling [eval] with the loaded variable contents
Exiting a Child Script
----------------------
The spawned child script may be exited prematurely by [return -code exit]
[return] or [return -code ok] at root level will not cause a child script
to exit and return to the parent
Caution: Mutual Script Loading Via [source]
-------------------------------------------
If two files mutually call each other via [source] then this may lead to
rapid memory consumption and memory exhaustion since memory from each
script instance will be preserved upon loading the next
See also: eval, load, run, dump. spawn, exec, tcx
string [eval math-expression]
Evaluate a Tcl expression
For simple commands it is more efficient to call directly than via [eval]
(e.g. [+ $x $y] [min $x $y] v's [eval "+ $x $y"] [eval "min $x $y])
Standard Tcl expressions are in 'prefix' format whereas expr expressions are
in 'infix' format
Example:
puts [eval "> 10 2"]
Result:
true
Freeform/unquoted number formats understood by eval are as follows:
decimal values NNNN
hexadecimal prefixed by 0xNNNN
octal prefixed by 0oNNNN
binary prefixed by 0bNNNN
See also: expr, bool, math, ticol, clear, unset, tcx
Event Log Handling Plugin |
Index |
To load, use: lib ticol_event ?-verbose? ?-nocomplain?
bool [event is_source name]
bool [event add name ?0..4?]
bool [event remove name]
bool [event log name string type ?postscript?]
[event] offers basic Windows event logging commands.
[event add] requires Windows UAC elevation and adds an SZ_EXPAND registry
event entry under:
"SYSTEM\\CurrentControlSet\\Services\\EventLog\\Application\\%s"
and this registers ticol_event.dll as the string table source for event
messages. Only 4 basic message types are supported
The following Macro defines are useful. Note that these can only be used
outside quoted strings
#define EVENTLOG_SUCCESS 0
#define EVENTLOG_ERROR_TYPE 1
#define EVENTLOG_WARNING_TYPE 2
#define EVENTLOG_INFORMATION_TYPE 4
Note
----
If an event source is removed then the Windows Event Log application will
no longer be able to query the string table of the linked Ticol plugin DLL.
You will see an error:
The description for Event ID N from source XXXX cannot be found.
The event source should be left registered as long as ticol_event.dll is
installed on the system and you need to query related events
Example:
#define EVENTLOG_INFORMATION_TYPE 4
lib ticol_event
if {[is_elevated]} {
if {event add ticol EVENTLOG_INFORMATION_TYPE} {
puts "Added an event to the event log"
} else {
puts "Failed to add an event"
}
puts [bool [event is_source ticol]
}
Results: (If elevated)
True
See also: is_elevated, elevate, plugins
The following commands are useful in handling exceptions:
catch
try / catch
error
The following variables are pre-defined and set on an error condition
arising
Name Type Comments
::errorCode Var Tcl error code sent to a the console
::errorLine Var May be preset using the #__LINE__ macro
::errorMsg Var Error message generated by a Ticol error
Fatal exceptions generate no errorMsg variable
::errorInfo Var Additional information
See also; try, catch, error
exec program ?arg? ?arg...?
@program ?arg? ?arg...?
@ program ?arg? ?arg...?
shell
Run an external program from the Ticol shell. Ticol can be used as a
Windows command shell.
If autoexec is "on" then unrecognised commands will be shell-executed
Prefixing a command with '@' can be useful to override conflicts with
built-in commands.
[shell] launches an instance of CMD.EXE. Type exit to exit and return
to Ticol
To grab the output text from a spawned application, see the [spawn]
command
CTRL+C is available for external applications, if quit using CTRL+C then
execution will be returned to the Ticol CLI shell
Examples:
@type ticol.man # Call the Windows command shell
exec dir *.tcl /s # Use [exec] to spawn the command
ftp localhost # Call the program directly
xcopy *.* \\server\share\folder # Call xcopy with arguments
See also: autoexec, spawn, exit
Various example scripts are included with the standard Ticol distribution
This includes a series of test files
See also: Tcl, Ticol, faq
exit ?errorlevel-value?
Exits a Ticol Tcl program. If the program was launched from a Windows
console then execution will return to the console. If launched from the
Ticol CLI then execution will return to the CLI
The errorlevel value will be returned to the operating system
CTRL+C is disabled in Ticol. If you wish to quit the CLI prompt, either
press CTRL+BREAK or type 'exit' at the CLI prompt
See also: stop, at_exit, return
string [explain expression]
Expand or 'explain' a natural expression in terms of the Polish Notation
(PN) interpretation of an expression as used with [calc]. The expression
handler, [expr] does not translate into intermediate Polish notation but
[calc] does. [explain] may be used to check the internal interpretation
of any expression
The result of [explain] may be evaluated using [eval]
Example:
puts [explain 1+2*3+4]
puts [eval [explain 1+2*3+4]]
Result:
[+ [+ 1 [* 2 3]] 4]
11
See also: calc, expr
string [expr expression items]
string [expr {expression-group}]
string [expr "expression-string"]
string [expr arg arg...]
Evaluates a mathematical expression held in a Tcl string or variable
Expressions within 'expr' allow 'infix' format whereas native Tcl 'eval'
commands are in 'prefix' notation as [operator var1 ?var2?],
Individual [expr] arguments may be grouped by double-quotes or braces but
will otherwise be concatenated into a single argument. It is strongly
recommended that the argument to [expr] be wrapped in braces or double-
quotes or complex expressions, (particularly multi-line) will be mis-
evaluated
[calc] is recommended as an alternative to expr for expressions not
generated at run-time. [calc] will reduce the natural-expression to
'Polish notation' format which runs faster than iterating over the
expression interpreter
Examples:
puts [expr 2*3+2] # Internal precedence
puts [expr $b*(3+2)] # Natural expression with brackets
puts [expr 2*[+ 3 2]] # Infix sub-command
puts [expr "2*[+ 3 2]"] # Infix sub-command (quoted)
puts [expr "2*[- $a 2]"] # Infix sub-command (quoted)
puts [expr "2*[decr 3 2]"] # Infix sub-command (quoted)
puts [expr {2*[+ 3 2]}] # Infix sub-command (braced)
puts [expr "2*[incr 3 2]"] # Infix sub-command (quoted)
Result:
8
10
10
10
2
2
10
10
Expression operators include the following:
+ - * / ^ # Add, sub, mul, div, power
== != > < >= <= # Relational math operators
&= |= ^= += -= *= /= # Combined math operators
<< >> # Bit-shifting operators
** ? () # Exponent, long-if, brackets
& | # Bitwise and/or
&&, || # Logical and/or
The following are not supported. Use embedded [command] versions
++ -- # Increment, decrement
and or # Keywords (use symbols instead)
mul div add sub # Keywords (use symbols instead)
Note particularly that:
++ and -- are not supported other than as separate [++], [--] commands
Unary increment/decrement is not implemented within [expr]. This is due to
potential conflicts with expressions such as "3 - -7"
Embed [++] [incr] or [--] [decr] commands within the expression.
You may intermingle expression functions, procedures, bracketed commands,
constants, variable names etc. within the expression string. Standard
BEDMAS operator precedence applies unless curved parentheses are used to
group
A wide range of arithmetic functions are supported. See "functions" for
more information. Curved brackets may be used to prioritise sub-expressions
Parity checks are made on bracket pairing outside of quoted strings.
Number evaluation follows Tcl rules. All numbers are evaluated as integer
unless a decimal point is specified. Division will therefore be truncated
to integer unless a decimal point is given for the divisor
Freeform expressions are also permitted but it is recommended that
expression-strings are braced or enclosed in double-quotes
Performance
-----------
Because it is a generalised expression-handler with significant syntax-
checking overheads and which calls the underlying Tcl routines, [expr] may
be far slower than decomposing an expression into component operations with
Tcl 'prefix' syntax. See [calc] for an optimised equivalent to [expr]
Example:
set src [expr ($x & $y) % 3] # Freeform expression syntax
set src [expr {22/7.0}] # Braced expression syntax
set src [expr "22/7.0"] # Quoted expression syntax
set src [% [& $x $y] 3] # Decomposed commands (~50% faster)
The Macro PreProcessor (MPP) may be used to optimise expressions into native
Tcl commands (see [calc] which is a substitute command which will auto-
optimise expressions into simpler Tcl commands)
Embedded Commands
------------------
Tcl commands may be embedded within expressions. These will be evaluated
first but do not affect overall operator precedence - e.g.
Use: puts [expr 22/7.0+[incr 2]] -> 6.142857142857142
Not: puts [expr 22/7.0+incr 2]
Use: puts [expr [incr 10]]
Not: puts [expr "incr 10"]
Whitespace
----------
Whitespace is handled between bracketed expression calls but the expression
should ideally be wrapped in double quotes or curly braces to ensure safe
grouping
Examples:
expr 22 / 7.0
expr 22/7.0
expr "22/7.0"
Returns:
3.142857
Whitespace is permitted it is recommended that it is avoided
Number Formats
--------------
The following 64-bit encoded number formats are understood by [expr]:
decimal any base 10 number 12345
hexadecimal prefixed by 0xNNNN 0x3039
octal prefixed by 0oNNNN 0o30071
binary prefixed by 0bNNNN 0b11000000111001
Transiting Array Brackets
-------------------------
An [expr] statement does not transit expression array () brackets.
While array indices may contain expressions those expressions are not
resolved unless they are in Tcl "command" format e.g. [* 2 3] or if the
[expr] command is used
Thus: [expr $a(22/7*2)] would be one of:
[expr $a([expr (22/7)*2])]
[expr $a([* [22/7] 2])]
Note that [calc] will attempt to resolve all sub-expressions, including
array parentheses expressions
Incorrect Use
-------------
Tcl commands which are not part of the [expr] syntax cannot be embedded
without square brackets. Commands [++] and [--] are not allowed as these
conflict with arithmetic addition/subtraction followed by sign value
Example:
set a 1;puts [expr "set b; ++ $a"] # ;separators are not valid
set a 1;set b [expr "[++ $a]"]; puts '$b' # Valid
Caution
-------
Because Tcl does not distinguish between data types you should avoid
passing unchecked string data into [expr], particularly string data which
may contain valid Tcl commands. This is because the arguments will be
evaluated
Example:
set s "@format c: /y"
expr "$s == {}" # This could be bad!
instead, embed string comparison commands such as [eq] or [string equal]
as follows...
set s "@format c: /y"
expr [eq $s {}]
https://en.wikipedia.org/wiki/Polish_notation
Stored, Nested [expr]
---------------------
If [expr] is both stored within a variable which is resolved later using
[expr] *and* the stored string contains [expr], then you must not only
brace the [expr] argument but include square brackets as follows:
set a {[expr {1 + 2 * 3}]} # Correct
# set a {expr {1 + 2 * 3}} # Wrong
puts "Stored a is: '$a'"
puts [expr {$a}]
Result:
Stored a is: '[expr 1 +( 2 * 3)]'
7
See also: calc, eval, bool, functions, math, arrays, inttohex, inttooct,
inttobin
This is example code which calculates a factorial
This sample may overflow into negative or zero values without further
error checking.
option expression off
proc factorial {val} {
puts "Current level: [info level] - val: $val"
set lvl [info level]
if {== $lvl $val} {
return $val
}
return [expr {($val-$lvl) * [factorial $val]}]
}
puts [factorial 20]
Results:
2432902008176640000
See also: fib, math
integer [fib n]
Compute Fibonacci series value from 0 to 93
Accessing a value over 93 will return an overflow error
[fib] uses a rapid, iterative method of calculation internally
Example:
option expression on
for {set j 0} {$j < 81} {incr j} {puts "$j [fib $j]"}
Result:
0 0
1 1
2 1
3 2
4 3
5 5
6 8
7 13
8 21
9 34
10 55
11 89
12 144
13 233
14 377
. . .
See also: math, big
Enhanced Tcl 'file' command
Offers a variety of stream-based file manipulation commands as well as
O/S-like commands. These are similar to those offered by the "C" language
Many of the sub-commands can be used independently without the 'file' prefix
For I/O see commands such as [puts], [format], [read]
file attributes filename # Return a Tcl list of common attributes
file close fp # Close an open file
file delete filename # Delete a file (alias for unlink)
file dirname name # Return the directory name of a filename/path
file eof fp # Tests file-read for end of file condition
file executable filename # Tests if a file is executable, return 1 or 0
file exepath # Return the path of the Ticol executable
file exists filename # Test if a file exists
file extension filename # Get the filename-extension/filetype as ".typ"
file flush fp # Flush file output to a stream
file gets fp ?var? ?bytes? # Read a line of text data from an open file
# Returns string or count of bytes if var given
file isdirectory filename # Test if a directory exists (returns a boolean)
file isfile filename # Test if a file exists (returns a boolean)
file isroot path # Test if an unescaped path is root, X:\ or \
file md5 filename # Generate an MD5 value for a file
file mkdir filename # Create a directory
file name filename # Return the filename+ext only (non standard)
# "C:\path\to\my.txt" gives "my.txt"
file normalize filename # Get the full, normalised path of a filename
# e.g. win.ini in c:\windows\system32 gives
# c:\windows\system32\win.ini
file open filename mode # Open a file for reading or writing
# Uses C++ fopen() modes e.g. "rb", wb", "rt"
# Returns a file handle
file pathcount filename # Return count of path segments (\\ or /)
file pathstrip filename n # Strips N path segments from the given path
file readable filename # Check a file for read access & return 1 or 0
# Does not check file read/write locking
file rename oldname newname # Rename a file. *CAUTION* - Ticol is ANSI only
# and does not support Unicode filename
file resolve path # Resolve an ambiguous path
file rewind fp # Rewind the file pointer of an open-file
file runtype filename ?-v? # Tests executable run type
# Returns code: Optional -v verbose flag
# -1 unknown Unknown type/non-executable
# 0 32bit 32-bit Windows exe
# 1 dos DOS executable
# 2 wow Windows on Windows64 exe (wow)
# 3 pif PIF file
# 4 posix Posix binary
# 5 os216 OS/2 16-bit exe
# 6 64bit 64-bit Windows exe
file seek fp offset # Set/place the file pointer of an open-file
file size filename # Get the size of a filename (64-bit)
file split filename # Split a path into components in list format
# {drive: path filename}
file stat filename var # Retrieve 'stat' data. See: file stat
file tell fp # Get the file-pointer location of an open-file
file type filename # Return file, directory, link or unknown
file unlink filename # Delete a file (alias for delete)
file volumes # Return a Tcl list of valid volumes
file writeable filename # Check file/dir for write access. Returns bool
Examples:
--------
All examples are with: option echo on
Example:
file attributes autoexec.tcl
Result:
-archive 1 -hidden 0 -system 0 -readonly 0 -longname autoexec.tcl \
-shortname {}
Example:
file readable autoexec.tcl
Result:
1
Example:
file type autoexec.tcl
Result:
file
Example:
file volumes
Result:
C:/ D:/ E:/ F:/ G:/ I:/ J:/
See also: info, exists, filesize, mkdir, mkdirs, unlink, dir, input,
readfile, writefile, normal_path, unc paths
bool [file exists filespec]
Check if a file exists or not. [file exists] also supports wildcards
Example:
puts [file exists ticol.exe]
puts [file exists *.tcl]
Result:
1
1
See also: file, readfile, writefile
integer [file pathcount filename]
Returns a count of the individual path segments (including any filename)
[pathcount] accepts backslash or Unix-style forward slash characters
If using backslash characters then [pathcount] requires an 'escaped' path
Use [escape] to escape a path string delimited by single backslash
characters
Examples:
puts [file pathcount c:\\]
puts [file pathcount c:/]
puts [file pathcount c:\\file.txt]
puts [file pathcount c:/file.txt]
puts [file pathcount c:\\path\\to\\some\\file.txt]
puts [file pathcount [escape c:\path\to\some\file.txt]]
puts [file pathcount c:/path/to/some/file.txt]
Results:
1
1
2
2
5
5
5
See also: file
string [file pathstrip pathstring n ?-unix?]
Strips 'n' path segments from the specified path.
If 'n' is zero then no segments are removed
If 'n' is greater than the number of path segments then an empty string is
returned
Both Windows and Unix style path separators are accepted as input, but if
Windows backslash characters are used they must be 'escaped' (use [escape])
A Windows style double backslash-escaped path will be returned unless the
-unix argument is used. Any trailing path separators are removed.
Note that path strings should always be wrapped in double quotes
Examples:
puts [file pathstrip "c:\\windows\\system32\\drivers\\etc\\" 2]
puts [file pathstrip "c:\\windows\\system32\\drivers\\etc\\" 2 -unix]
puts [file pathstrip "c:/program files/microsoft shared/" 1]
Results:
c:\\windows\\system32
c:/windows/system32
c:\\program files
See also: file, file pathcount, escape, unescape
bool [file stat filename arrayvar]
Fills an array with _stat information using the C++ '_stat()' function
The array variable referenced in the argument must not exist. A boolean
True (1) is returned if the file exists, False (0) if it doesn't
Example:
file stat ticol.exe a
array foreach a val ss {
puts "\$a($ss)\t$val"
}
Results:
$a(mode) 33279 # Protection
$a(ino) 0 # inode number (always 0)
$a(dev) 2 # ID of device containing file (A==0)
$a(atime) 1481501018 # C time of last access
$a(mtime) 1481501108 # C time of last modification
$a(gid) 0 # Group ID of owner (always 0)
$a(size) 190976 # Total size, in bytes
$a(nlink) 1 # Number of links (always 1)
$a(rdev) 2l # device ID (if special file)
$a(ctime) 1481501018 # time of last status change
$a(drive) C # Drive letter of 'dev' (upper case)
$a(uid) 0 # User ID of owner (always 0)
References:
https://msdn.microsoft.com/en-us/library/14h5k7ff.aspx
https://stackoverflow.com/questions/6094665/how-does-stat-under-windows-exactly-works
See also: file
string [file gets ?fp? ?varname? ?-length bytes?]
Read a line of text data from an open file
Returns a string value if no variable specified, else returns the count of
bytes successfully read if a var argument was given. If an EOL byte sequence
is encountered before 'bytes' have been read then a short count will be
returned. If 'var' does not exist it will be created in the current scope
[file gets var] will always clear var to {} if it exists
Example:
set fp [file open [unescape $argv(1)]]
set r [file gets $fp s -length 100]
puts "s is '$s"
file close $fp
Result:
<Text from input file up to 100 chars or first CRLF sequence>
See also: readfile, file
string [extract string maskchars ?default? ?-nocase? ?-offset N?]
string [extract varname maskchars ?default? -var ?-nocase? ?-offset N?]
Extract strings matching 'maskchars' from the left hand end of a string
Can be useful for stripping and parsing number sequences from large
strings. The first 4 arguments are positional
Very large strings in a variable may be passed by name using the -var arg
To simplify use with math commands, a default value may be returned
Using -nocase will match on characters regardless of case
If -offset is used with an integer value then the search will begin at 'n'
characters into the string where 0 is the first character 1 is the 2nd etc.
The offset value shifts to the starting point to the right by N chars
Note that the mask string must include a whitespace character if multiple
words are to be found in any search string. Omitting a space can be useful
in using [extract] to search and extract single words. Using '-offset N'
with a loop variable it is possible to iterate through all word tokens in
a string
If the search mask is braced it will be interpreted as a literal string
and can therefore not contain embedded commands. Any complex search
mask relying on commands should be generated and passed via a variable
Example:
set s "123.45Hello World"
puts [* [extract s "0123456789." "0" -var] $pi]
set s "No numbers in here"
puts [* [extract s "0123456789." "1" -var] $pi]
puts [extract qbf [range a-z] -var -nocase -offset 4]
puts [extract qbf "[range a-z] " -var -nocase -offset 35]
Result:
387.829613085659960
3.141592653589793
quick
lazy dog
See also: string, filter
string [filter string charmask ?-exclude? ?-nocase?]
string [filter varname charmask ?-exclude? ?-var? ?-nocase?]
Filter characters in a string, either including (default) or excluding
characters from string 'charmask'. The first 3 arguments are positional
Large variables which have a heavy copy cost may be passed by name using
the -var argument
Using -nocase will match on characters regardless of case
Example:
# Snippet of Brainfuck code
set s "+++++ +++++ Initialize counter (cell #0) to 10"
set s [filter $s "+"]
puts '$s'
Result:
'++++++++++'
See also: string, extract
integer [filesize filename]
Returns the size of a given file in bytes as a 64-bit value
See also: file, unlink
find string
find "string with spaces"
Search the Ticol manual for topics which contain a given search string
Since the macro preprocessor (MPP) processes all Ticol command-line
arguments including both 'help' and 'find', this will affect the ability
to search for the # character (which is interpreted as a comment)
See: help option PreProcessor
Example:
user@host> find upvar
Result:
Help topics containing keyword 'upvar'
arrays subsequent upvar command. To pass an array by value
at_exit The upvar command is required to access variables d
combinatory logic upvar 1 $_stack stack
comparing ticol with php Standard (downwards) variable inheritance No (upva
const declared using upvar to create a local const instan
contents upvar Reference a variable in a higher
differences global Global is an alias for upvar
dump dump, when used with upvar arrays or standard varia
find user@host> find upvar
See also: help
Provides a for iterator similar to that of C/C++
With: option expression on (Standard Tcl):
Default:
for {init-eval} {condition-expr} {post-eval} {eval-code}
With INI ForInitAllowsExpression=TRUE
for {init-expr} {condition-expr} {post-eval} {eval-code}
for {[init-eval]} {condition-expr} {post-eval} {eval-code}
This uses an [expr] statement as the 3rd (condition) argument and
if configured, can optionally use an [expr] statement for the 2nd (init)
argument. Always ensure you have the correct [option expression] setting
With: option expression off (Non-standard):
for {init-eval} {condition-eval} {post-eval} {eval-code}
This uses an [eval] statement for the 3rd (condition) argument
The following additional commands are available for use with [for]
return # Will exit the script when used in root-level [if]
break # Will be passed back to any enclosing structure
stop # Halt operation, return to the CLI if loaded
exit # Exit the script and return to Windows
goto # Where used with an enclosing [goto_block]
continue # Resume execution at the head of the [for] loop
Flow-control for loop
Example with single loop variable
option expression off
for {set i 0} {< $i 10} {incr i} {
puts "i is $i"
++ j
}
option expression on
for {[set i 0]} {$i < 10} {incr i} {
puts "i is $i"
++ j
}
Example - Handling multiple loop variables:
[for] will quite happily accept multiple or complex arguments since each
braced argument is merely a Tcl expression which will be interpreted
using either [expr] or [eval] depending on the [option expression]
setting.
option expression on
for {[set i 0]; [set j 0]} {($i < 10) && ($j < 10)} {incr i; incr j} {
puts "i is $i, j is $j"
++ j
}
option expression on
for {[set i 0; set j 0]} {$i < 6} { ++ i; ++ j 2} {
puts "i=$i; j=$j"
}
option expression off
for {[set i 0]; [set j 0]} {[expr ($i < 10) && ($j < 10)]} {[incr i]; \
[incr j]} {
puts "i is $i, j is $j"
++ j
}
Emulating C++ for(;;)
---------------------
The [for] command with null arguments will result in an infinite loop
See also: while
for {} {} {} {
puts "Infinite loop"
}
See also: loop, while, do, time, flow control
The standard interrupt handler CTRL+C is disabled to prevent accidental
or undesired exit from a Ticol script. Script execution can be terminated
using CTRL+Break instead
See also: break, exit, stop
Flow-control: for foreach if else, switch while do goto on loop |
Index |
The following construct flow-control constructs are supported:
array foreach # Iterator
carray foreach # Iterator (Plugin)
catch # Error handler
do # Iterator
floop # Floating point loop
for # Iterator
foreach # Iterator
gosub # Subroutine execute
goto / goto_block # Selection
if # Selection
linkedlist foreach # Iterator (Plugin)
loop # Integer loop iterator
on # Selection
proc # Procedure execute
stack foreach # Iterator
switch # Selection
time # Iterator
try/catch # Error handler
varray foreach # Iterator (Plugin)
while # Iterator
Details
-------
array foreach array element ?subscript? {code-body}
carray foreach handle lvalue ?rvalue? {code-body}
catch {code-body}
do {code-body} while {...}
floop variable float-start float-limit ?float-increment? {code}
for {initiator} {condition-test} {post-action} {code-body}
foreach variable list {code-body}
foreach varlist1 list1 ?varlistN listN ...? {code-body}
gosub function
goto_block {?goto? statement} {...}
if {condition-test} else {...}
if {condition-test} elseif {...} ... else {...}
linkedlist foreach handle lvalue ?rvalue? {code-body}
loop variable start limit ?increment? {code-body}
on integer {...} ?{...}...? ?-default {...}?
proc name {?arg...?} {code-body}
stack foreach varName indexVar {code-body} ?-r?
switch ?options? {arg} {{case script} ... | {default script}}
time ?expression? ?iterations? ?-info | -quiet?
try {code-body} catch {code-body}
varray foreach handle variable {code-body}
while {condition-test} {code-body}
Additionally: the following control keywords are available:
break # Exit the loop
continue # Continue execution from the loop head
default # (For use with [switch] only)
die # Exit to windows with message
exit # Halt processing and exit to Windows
goto # Only valid within a [goto_block]
return # Return from the loop with optional code etc.
stop # Halt processing and return to the CLI
Behaviour is modified by: option expression on|off
if/else/elseif keywords must be separated by whitespace. The Tcl syntax
requires that the opening brace following if, else, elseif, be on the same
line as the command. e.g. (See: brace style)
if { condition } { # Correct layout (K&R/Tcl format)
statement
} elseif { condition } {
statement
} else {
statement
}
if { condition } # Incorrect layout (Allman/BSD format)
{
statement
}
else if { condition } {
statement
}
else
{
statement
}
while { condition } { # Correct layout (Tcl)
statement
}
while condition { statement } # Correct layout (Tcl)
do { # Correct layout (Tcl)
statement
} ?while? {condition}
do {statement} {condition} # Correct layout (Tcl)
set i 0
do { # Correct layout (Tcl)
puts "i is $i"
if {== $i 2} { return }
incr i
} while { $i < 10}
while 1 {puts [rnd 1 10] }
do {puts [rnd 1 10] } while 1
All of the above cases support the break or continue keywords
break exits the block loop structure such as while, for, foreach
continue jumps to the start of any block structure and repeats
Return values set using 'return' can be passed from while or do-while by
chaining from 'set' - e.g.
set r [do { ...
if { $i == 2 } {return string}
set r [while { ...
if { $i == 2 } {return string}
If 'option expression' is set to ON then expr with infix notation will be
used for tests instead of the default of eval with prefix notation
A limited-scope form of goto is supported within a goto_block
(see goto_block)
See also: for, foreach, array foreach, if, elseif, else, switch,
while, do, loop
See also: option expression, foreach, array foreach, if, proc, return,
at_exit, linkedlist, carray, stack, loop, try, catch
break
Halt execution within a flow-control block such as [for], [while], [do] or
[goto_block]
If executed outside a flow control block and at level 0 (root) [break]
will exit the current script and have the same effect as [exit]
[break] is not valid within other flow-control structures
Example:
option expression off
for {set i 0} {< $i 10} {incr i} {
if {== $i 5} {break}
}
Example:
break
Result
# Script exits
See also: flow control, continue, stop, exit, return
continue
Continue execution of a loop structure from the head of the loop
[continue] skips any body code and executes the increment argument of a
[for] loop. Relevant to [for] [foreach] [while] [do]
Example:
# print a line for each of the integers from 0 to 10 except 5:
option expression on
for {set x 0} {$x < 10} {incr x} {
if {$x == 5} {
continue
}
puts "x is $x"
}
Result:
x is 0
x is 1
x is 2
x is 3
x is 4
x is 6
x is 7
x is 8
x is 9
Example:
option expression off
set sum 0
set x 0
while {< $x 10} {
incr x # Pre-increment before [continue]
if {== $x 5} {
continue
}
puts "x is $x"
+= sum $x
}
Result:
x is 1
x is 2
x is 3
x is 4
x is 6
x is 7
x is 8
x is 9
x is 10
See also: flow control, break, stop, exit, return
Handling filenames and paths |
Index |
Wherever possible, Ticol handles both / and \ path separators in filenames.
Within complex strings and expressions it may be preferable to use / as
this does not need to be escaped and avoids problems with 'double-
unescaping'
Literal strings must have the \ character 'escaped' as \\
Example:
set filename "c:/path/to/file.txt" # Correct
set filename "c:\\path\\to\\file.txt" # Correct
set filename "c:\path\to\file.txt" # Incorrect (\t is a TAB)
Command Line Use
----------------
When using filenames from the CLI this behaviour is modified to allow
unescaped paths. See: option escape for more information
See also: escape sequences, escape, unescape, option escape
integer [fpart double-value ?precision?]
Return the fractional part of a double value as an integer
The precision part must be >= 1 and <= 10
[fpart] preserves any zeroes after the decimal, stripping the sign, leading
zero (if any) before the decimal point and the decimal point itself
Example:
puts [fpart $pi]
puts [fpart $pi 2]
puts [fpart $pi 6]
Results:
1415926535
14
141592
Example:
puts [fpart 0.12]
Result:
1200000000
Example:
# Splitting a double value for processing with reassembly
puts "[integral $pi].[fpart $pi 7]" # Recommended
puts "[integral $pi][fraction $pi]" # Bad idea!
puts "[integral $pi].[fraction $pi]" # Bad idea!
Result:
3.1415926 # Correct
30.141592653589793 # Wrong!
3.0.141592653589793 # Wrong!
Example:
puts [fpart [dateserial - - - 2 4 6]]
puts [time .[fpart [dateserial - - - 2 4 6]]]
Results:
0861805555
02:04:06
See also: fraction, integral, vfract
fraction and integral |
Index |
double [fraction realValue]
double [integral realValue]
Returns either the fractional or integer part of a real number (float)
Example:
fraction [expr atan(1)*4]
Result:
0.141593
Example:
integral [expr atan(1)*4]
Result:
3.00000
Note: Use [fpart] to retrieve the decimal fraction as an integer value
See also: fpart, math, vfract
Introspection and Reflection |
Index |
Ticol has modest commands for type reflection and introspection,
as follows:
args # C++ like vararg placeholder
array walk # Dump a variable's contents
assert # Debug assertion
catch # Trap errors (see also; try..catch)
cmdcount # Count of the # of executed commands
dump # Dump details of a standard variable or struct
dump # Show syntax-highlighted source code
dumpvariant # Dump the contents of a variant
halt # Enter the debugger when debugging active
info # Various introspective commands
inspect # Inspect raw Tcl variable information
mem_used # Returns the current memory used by Ticol
resume # Resume execution after debugging halt
time # Time execution calls
trace # Display calls and shows results
try .. catch # C++ like try...catch blocks
type # Display variable type names
vars # List variables for a given context
walk # Dump the variable table at any callframe level
watch # Set a variable watch for use with breakpoints
/echo # Arg. Show Macro PreProcessed source code
/bp # Arg. Breakpoint execution stacktrace
/debug # Arg. Enable debugging options including /bp
/g # Arg. Performance graph
Variables also exist which reveal information about the system
These can be inspected using the command [vars -const]
string exe_name
string tcl_date
integer tcl_cli
string tcl_config
integer tcl_threaded
float tcl_version
integer tcl_precision
array env
empty argv0
integer argc
See also: debugging
doevents ?count? ?-sleep?
This is analogous to Visual BASIC's DoEvents command.
[doevents] yields thread control to Windows and allows processing of the
Windows message queue. doevents may be combined with sleep to minimise CPU
use during tight loops. [doevents] will poll the Windows message queue
An optional count argument may be added which will cause [doevents]
to poll a given number of times. Additionally, a 'sleep' of 10ms
may be performed during each loop.
You can use [sleep 0] to achieve a similar effect
Example:
set i 0
while {$i < 1000000 } {
# A tight, CPU-intensive loop
doevents
incr i
sleep 0
}
See also: sleep, performance
flush ?stream?
Flushes output to an open file stream. This may include the symbolic strings
stdout and stderr as well as open file handles
See also: open, close, file, puts, printf, echo, ticol_cgi
. dot (or head) command |
Index |
char [. string|var ?-var?]
The 'dot' command returns the first (head) character of a string or
variable. This is far more efficient than [string left $string 1] with
a timing of about 1 or 2 microseconds per call on an i5 upwards
It is equivalent to C/C++ '*string'
The command [asc] will return the first character as an ASCII integer
Example:
puts [. $qbf]
puts [. qbf -var]
puts [. qbf]
Result
T
T
q
See also: pipe, asc, chr, string
foreach (Flow Control) |
Index |
foreach var list ?var list...? {script}
foreach {varlist} list ?{varlist} list...? {script}
foreach var|{var...} list|{vars...} {script}
Iterates the RValue (right hand value) list using LValue (left hand value)
variable(s) as placeholder(s) for the list item(s)
Can be considered equivalent to the pseudocode ('in' is not supported):
foreach variable in list {script}
If a literal list of items is given then it should not be wrapped in
braces to ensure grouping. If a braced list of variables is given then
don't enclose in quotes or braces. Rvalue variables should not be braced
as this will prevent evaluation. LValue variables may be braced
The opening brace of the body statement should be on the same line as
foreach. See the rules about grouping in Tcl in the Tcl documentation
The iterator variable(s) may exist in the current scope, but will be
reused and if they exist must be of standard type. [foreach] will not
check/reject existing variables before using them as iterators
The iterator variable(s) will remain in the current scope after [foreach]
exits and must be unset. [foreach] will not unset iterators on exit
If a break statement is encountered before the list is exhausted, the tail
of the list will be returned.
foreach f [ls "c:\\windows\\*.*" -af -list] {
puts "File: $f"
}
unset f
foreach {f g} {one {two three} four five six {seven eight} nine ten} {
puts "Items f g: $f, $g"
}
unset f g
ls *.* a -la -list
foreach {i j k l} $a { # Lvalue vars may be braced to group
puts "$i, $j, $k, $l"
}
foreach f {one two three} {puts "$f,"}
unset f
set args {one two three four}
set arg [foreach {a b c} $args break]
foreach s [commands] {puts $s}
unset s
set l [array list argv]
foreach s $l {puts $s}
unset s
Example:
set a "1 2"
set b "3 4"
foreach x $a y $b {puts "$x $y"}
unset x
Results:
1 3
2 4
When using with arrays take care to wrap multiple Lvalue variable arguments in
braces
WRONG: foreach x y [array get env] {puts $x=$y}
RIGHT: foreach {x y} [array get env] {puts $x=$y}
foreach may also be used to split lists into variables using an empty body
set s "123 456"
foreach {p q} $s {}
Admin@Lenovo 134> vars
Defined variables at level 0
----------------------------------------
integer p 123
integer q 456
See also: for, while, time, switch, array foreach, list, ls, ticol
The [input] and [gotoxy] commands can be used to create simple form-based
input screens to capture user-input. There is no support for Tk and
Windows GUI based forms
Example:
# Prompt for 3 basic input fields
cls
textcolor white darkmagenta
puts " * Ticol Demo Form Input * "
textcolor
if {< $tcl_version 1.27} {
puts "Error: Requires Tcl v1.27 or higher. Found v$tcl_version"
die
}
proc at_exit {} {
gotoxy 1 15
}
box 4 4 74 7
gotoxy 6 6
puts "Enter your name: ________________________________________"
gotoxy 6 7
puts "Enter your age: ___"
gotoxy 6 8
puts "Enter your address: ________________________________________"
gotoxy 6 6
puts "Enter your name: " -nonewline
gotoxy 26 6
set name [input 40 -f white -b red]
textcolor
if {eq $name ""} return
gotoxy 6 7
puts "Enter your age: " -nonewline
gotoxy 26 7
set age [input 3 -f white -b blue -password]
textcolor
if {eq $age ""} return
gotoxy 6 8
puts "Enter your address: " -nonewline
gotoxy 26 8
set address [input 50 -f black -b yellow]
textcolor
if {eq $address ""} return
# Display our results ...
box 4 11 74 5
gotoxy 6 12
puts "Your name is: $name"
gotoxy 6 13
puts "Your age is: $age"
gotoxy 6 14
puts "Your address is: $address"
gotoxy 1 17
Result:
+--------------------------------------------------------------------+
| Your name is: Me |
| Your age is: 23 |
| Your address is: me@example.com |
+--------------------------------------------------------------------+
See also: input, inputwin, msgbox, getkey
string [format format-string|literal ?arg? ...]
Standard Tcl format command. The syntax is similar to the C printf command.
format prints no output. You must call puts or printf to display the result.
Note that format and printf do not automatically issue a newline character
Examples:
const euler 2.71828182845904523
format %x 123456
puts [format "%.10f" [/ $euler $pi]]
puts [format "%.0i" [expr 1/20.0*100]]%
Results:
1e240
0.8652559794
5
[format] and [printf] both include 64-bit number formatting as %I64i and
%I64u (case-sensitive). Overflows occur at about 19 digits.
Format field numeric specifiers may be passed using a * token
Example: (interprets as '%-20.2s')
Result: printf "'%-*.*s'\n" 20 2 Hello => 'He '
See also: format string, set, puts, printf, comma
Format String (printf and format) |
Index |
The [format] and [printf] commands generate a formatted string in a fashion
similar to the ANSI C sprintf function. A format string indicates how to
format the result, using % conversion specifiers as in C sprintf, and the
additional arguments, if any, provide values to be substituted into the
result. The return value from format is a formatted string
A format token sequence is introduced via a percent % character
The following formatting tokens may be used with [format] and [printf]
%<flags><field-width><precision><size-modifier>conversion-type
%<+|-| |0><N|*><.N|*><ll|i64>b|c|d|e|E|f|g|i|o|s|u|x|X|%
Positional Specifier
--------------------
In standard Tcl the first position is an optional positional specifier
This is unsupported in Ticol Tcl
Optional Flags
--------------
The section portion of a format string consists of an optional flag
character
- Left-justify the converted argument
+ Print number with a sign, even if positive.
space A space should be added to the beginning of the number
if the first character is not a sign
0 Specifies that the number should be padded on the left
with zeroes instead of spaces
0n Specify 'n' number of leading zeroes (e.g. %05i, see below)
Optional Field Width
--------------------
The third portion of a conversion specifier is a decimal number giving a
minimum field width for this conversion. It is typically used to make
columns line up in tabular printouts. If the converted argument contains
fewer characters than the minimum field width then it will be padded so
that it is as wide as the minimum field width.
Padding normally occurs by adding extra spaces on the left of the converted
argument, but the 0 and - flags may be used to specify padding with zeroes
on the left or with spaces on the right, respectively. If the minimum field
width is specified as * rather than a number, then the next argument to the
format command determines the minimum field width; it must be an integer
value
printf "'%*s'\n" 20 Hello => ' Hello'
printf "'%-*s'\n" 20 Hello => 'Hello '
Optional Precision Specifier
----------------------------
The fourth portion of a conversion specifier is a precision, which consists
of a period followed by a number
If the precision is specified with * rather than a number then the next
argument to the format command determines the precision; it must be a
numeric string.
Example: (interpreted as '%-20.2s')
printf "'%-*.*s'\n" 20 2 Hello => 'He '
Optional Size Modifier
----------------------
The fifth part of a conversion specifier is a size modifier, which must be
ll, h, or l. If it is ll or I64 then the value will be cast to 'wide'
format
lli 64-bit signed integer
llu 64-bit unsigned integer
I64i 64-bit signed integer
I64u 64-bit unsigned integer
Mandatory Conversion Types
--------------------------
The sixth part of a format string is a mandatory conversion type specifier
b Binary integer
c Character type (Non-Tcl standard)
d Signed decimal
e or E Scientific notation
f Floating point/double
g General format
i Signed integer
o Octal
s String
u Unsigned integer
x or X Hexadecimal (lower, upper case)
% Literal % symbol (%%)
Differences from ANSI printf
----------------------------
* Added Tcl support
%n Unsupported as per standard Tcl
%p Pointer format. (Unsupported as per standard Tcl, use %x)
Differences from Standard Tcl
-----------------------------
%N$ Prefix $ unsupported
See also: format, printf, puts
Memory Allocation and Garbage-Collection |
Index |
Ticol allocates and frees memory for variables automatically. There is
no need to perform any specific 'garbage collection' operations
Array variables are assigned from a self-managing hashtable which expands
dynamically and is recovered automatically as scope exits
Uncontrolled recursion can also cause a 'stack overflow' memory error
where the application will generate a fatal exception error. Recursive
procedures should be carefully-designed to avoid such situations arising
Memory returned by external DLL calls or assigned by [malloc] for an
external DLL call should be carefully managed
See also: set, proc, malloc, free, memory
integer [get_win_error]
Retrieve the last registered Windows error. This should be retrieved as soon
as possible before a new event resets the error value to zero
Use [win_error] to format the error message
See also: win_error
list [glob ?-file d|f? ?-directory name? ?-nocase? pattern]
A limited implementation of the Tcl [glob] command. The [ls] command
offers more flexibility but [glob] is useful for generating lists of
files/folders or taking file counts
The glob pattern MUST be wrapped in double-quotes if it contains '/*'
[glob] performs a case-sensitive comparison on search results, this can be
overridden using -nocase
The search pattern may specify only * and ?; character matches [chars],
\x and {a,b,...} are unsupported. Only one search pattern is supported
The options currently support only -type d (directory) or -type f (file)
arguments b,c,l,p,s are unsupported. (df is the default)
The -directory name may be either a path name or a filespec
The types may be combined using multiple -type arguments
A Tcl list is returned with the path separator format will always be in
Unix style "/"
Examples:
puts [glob -type f] # Find all files
puts [glob -type fd] # Find all files and dirs
puts [glob -type f -type d] # Find all files and dirs
puts [glob -type f d*] # Find all files matching d*
puts [glob -directory c:\tools -type f a*] # Find files a* in c:\tools
puts [glob "/windows/*.dll"] # Dquotes required for '/*'
puts [lcount [glob -nocase "/windows/system32/r*"]]
Example:
puts [glob -type d "\windows\system32\r*"] # From the CLI
Result:
/windows/system32/ras /windows/system32/Recovery
/windows/system32/restore /windows/system32/ro-RO
/windows/system32/RTCOM /windows/system32/ru-RU
Example:
foreach x [glob -type f "/tools/*"] {puts [file normalize $x]
Result:
C:/tools/ask.exe
C:/tools/md5.exe
C:/tools/printf.exe
C:/tools/ticol.exe
Glob is also a term which refers to wildcard string matching present in
the following commands:
array get
array names
lsearch
string match
switch
unset
vars
See also: ls
string [get_folder ?-initdir initdir? ?-flags flags? ?-title string?]
Creates a browse dialog with a user-input edit box
Uses the Windows shell GUI to browse for a folder, and, if required create
a new foldername. If the interface is cancelled then an empty string is
returned. Unicode is not supported. No error codes are returned
[get_folder] accepts Unix style forward-slash '/' path separators but
requires an unescaped string for '-initdir' if Windows backslash path
separators are used. Use [unescape]
A path containing Windows '\' path separators will be returned unless
the user cancels input
Some useful flag values (may have no effect in some Windows releases):
#define BIF_RETURNONLYFSDIRS 0x00000001
#define BIF_EDITBOX 0x00000010
#define BIF_NEWDIALOGSTYLE 0x00000040
See the Windows API guide for full details of folder browse flags
Examples:
#define BIF_RETURNONLYFSDIRS 0x00000001
set s [get_folder -flags BIF_RETURNONLYFSDIRS]
set s [get-folder -initdir [unescape "c:\\temp"]
set s [get-folder -initdir [unescape "c:/windows"]
set s [get-folder -initdir [unescape "c:\\"] -title "Select txt files"
See also: get_fileopen, get_filesave
string [getkey ?-integer? ?-lower? ?-upper?]
Wait for a keypress and return the character. If the -integer argument is
given then an integer value representing the key input is returned
Upper or lower case can be forced using -lower or -upper
Note that PC extended key handling for keys such as F1 will return a
sequence of two integer values. For more information, search for PC keyboard
handling
Use [inkey] to handle keyboard polling loops
Example:
A useful Yes/No/Quit key-input procedure
if {! [defined get_key_input]} {
proc get_key_input {} {
while {1} {
set k [getkey -lower] # Returns a lower-case keypress
if {eq $k "y"} {
puts $k
return "-r"
} elseif {eq $k "n"} {
puts $k
return {}
} elseif {eq $k "q"} {
puts $k
return q
}
sleep 20
}
}
}
Example:
Prompt user to copy text to clipboard
textcolor yellow
puts "Copy screen to the clipboard ? Y/N" -nonewline
set k [getkey -lower]
if {eq $k "y"} {
clipboard gettext # Grab the screen as text
puts "\\nCaptured screen to clipboard"
} else {
newline
}
textcolor
See also: key input, inkey, gets, input, inputbox
integer [getsid <username>]
integer [getsid <domain\\username>]
Retrieve a Windows SID string for a given username. The domain prefix is
optional. If omitted the current workstation/workgroup will be used
Example:
puts [getsid administrator]
puts [getsid snoopy\\administrator]
puts [getsid [username]]
Result:
S-1-5-21-460261754-543217201-1416368962-500
S-1-5-21-460261754-543217201-1416368962-500
S-1-5-21-460261754-123456789-9876543210-500
See also: pid, tid, username
global varName ?varName? ... ?-value value?
Link to an existing variable in global scope or, if this does not exist,
create a new, simple variable (non struct/array) in global scope with an
alias of the same name in the local scope
[global] is valid only within procedures
[global] is used within procedures to create a global in the root (::global)
namespace and a local (upvar) copy of the same variable. The value argument
is optional and if omitted the new variable will be set to empty ""
When the local copy of the global variable is updated the global copy will
also be updated. The local copy is destroyed on exit of the procedure but
the global copy will remain with values set by the procedure
Ticol global differs from standard Tcl in supporting a single 'value' argument
which may be used to initialise all of the specified values to, say, zero
Where a scope prefix is given (e.g. a in 'a::b', this will be emulated
and a local variable created which uses the tail value (e.g. 'b' in this
example)
Example:
proc create_global {} {
# Create a global var with local upvar copy, set it then change it
puts "Calling \[global global_var -value 12345\]"
# Creates var in global scope
global global_var -value 12345
puts "Local: \[bool \[is_set global_var\]\]=\
[bool [is_set global_var]]"
puts "Global:\[bool \[is_set ::global_var\]\]=\
[bool [is_set ::global_var]]"
puts "Check: \$::global_var=$::global_var"
# Update the local copy
set global_var 11111
# Then check the global copy
puts "Recheck:\$::global_var=$::global_var"
}
create_global
vars gl*
Result:
Calling [global global_var -value 12345]
Local: [bool [is_set global_var]]=true
Global:[bool [is_set ::global_var]]=true
Check: $::global_var=12345
Recheck:$::global_var=11111
Defined variables matching 'gl' at level 0
------------------------------------------
integer global_var 11111
Example:
proc create_global {} {
# Create a global var with an emulated scope prefix
global a::b # Create '::a::b' and local 'b'
set b 12345 # Local 'b' and '::a::b' are set
}
create_global
vars gl*
Result:
Lists variable 'a::b' as 12345
See also: global variables, preset globals
Variables in global scope, including arrays, may be accessed by the global
scope prefix '::'. This must appear immediately after any dollar sign, e.g.
$::globalVar
Global constants, when accessed within procedures, must be either referred
to using $:: or declared using upvar which will create a local const instance
See: preset globals - for a discussion of pre-set Ticol global variables
Nested scope levels exist within procedures. This is referred to as the
given 'level'. This value increases from 0 at the top (global) level to
higher values as procedure call-frames are added.
Variable scope within flow-control structures is the same as at the current
procedure level
Example:
proc getvar {} {
set a "I am here"
puts "proc local a is: '$a'"
puts "proc global a is: '$::a'"
puts "proc const b is $::b"
}
set a "Hello"
const b 1234
puts "a in global scope is $::a"
getvar
Result:
a in global scope is Hello
proc local a is: 'I am here'
proc global a is: 'Hello'
proc const b is 1234
Example of declaring a global variable and setting as an array/element
set a ""
proc set_global_array {} {
# Set (cast) global non-array variable as array with one element
set ::a(1) "One"
}
set_global_array # Call the proc
puts "\$a(1) set by proc is '$a(1)'" # Print the results
Result:
$a(1) set by proc is 'One'
See also: preset globals, global, set, let
goto_block, goto, gosub
goto_block {
goto local_label
gosub local_label
label {statement}
...
}
[goto] is implemented purely for the entertainment of the author
[goto] statements are valid only within the scope of a [goto_block]
statement
The [goto] paradigm is unpopular and, if not used correctly, can lead to
messy and difficult to debug source code; but it may be of interest to
computer science students
A [goto_block] is a simple set of locally-constrained labels and matched,
braced script pairs which works in a manner similar to PHP's goto statement
but which is confined in scope to one or more [goto_block] statements.
If a label's script argument consists of only a single-command there is no
need for braces. In addition to pairs of labels and scripts, a [goto label]
or [gosub label] entry with matching label argument is allowed
It is not possible (nor recommended) to jump into any structured flow
control using [goto]. A [goto_block] cannot span or overlap a flow control
command such as [for] or [while]
No freeform Tcl code is allowed in the [goto_block] outside a properly-
defined block. All Tcl script code must be declared under a block
associated with a valid label - e.g.
goto_block {
puts "Hello" # Wrong (freeform code with no label)
}
goto_block {
init {puts "Hello"} # Correct (label + code block pair)
}
[goto_block] labels may be any string (including Tcl command names) but
they must be unique within the context of each [goto_block]
Unless a [goto] or [gosub] command is encountered, a [goto_block] will
sequentially evaluate each label's script argument in turn, one after the
other. A label's script is executed immediately when encountered with
execution starting at the first valid label in the block. Each successive
label will be evaluated unless otherwise directed by a [goto] or [gosub]
statement
A [break] or [return] within a label's script will exit the [goto_block]
and ignore all remaining labels and code within that block. e.g.
goto_block { # Sequential labels + statements ...
label1 { puts Hello } # 1)
label2 {} # 2) Empty arguments are permitted
label3 { break } # 3) Will cause the goto_block to exit
label4 { puts world } # 4) Never executed
label5 { return } # 5) [goto_block] would exit
}
Result:
Hello
Each label entry within a [goto_block] must consist of a label followed by
a code block. The code block must be wrapped in braces if there is more
than one command present. Labels need not be all-numeric and can be any
valid string. Empty label script argument parentheses are allowed
Any required block initialisation code for the [goto_block] MUST appear
within a dummy label entry and not 'inline'. It is suggested that an
'init' label is used as the first label entry and any setup code executed
within the 'init' code-block.
Label statements are executed sequentially until a [goto] statement is
encountered or after one or more goto statements has been executed
[goto_blocks] may be nested but must appear within the scope of a properly
constructed block statement. However, labels are unique and local to each
[goto_block] at each level. Nested goto_blocks are probably a bad idea and
could lead to very messy and hard to debug code
Variable $this is set to the label entry on entering a code block
Example:
goto_block {
label1 {
puts "In parent goto_block label1"
goto_block {
label1 {
puts "In child goto_block label1"
goto_block {
label1 {
puts "In grandchild goto_block label1"
goto label2
}
label2 {
puts "In grandchild goto_block label2"
}
}
}
label2 {
puts "In child goto_block label2"
}
}
}
}
Result:
In parent goto_block label1
In child goto_block label1
In grandchild goto_block label1
In grandchild goto_block label2
In child goto_block label2
[gosub] may be used to call a local label and resume execution at the next
statement after that original [gosub]
[gosub], when used within a label's argument script, behaves normally and
will only call a registered [sub]. This [sub] may be declared within the
[goto_block] or externally (e.g. globally)
Example:
sub external {
puts "In external sub"
}
goto_block {
gosub label_two # Call local block
gosub label_one # Call local block
label_one {
puts "In yr label1"
gosub external # Call external [sub]
}
label_two {
puts "In yr label2"
}
}
Results:
In yr label2 # [gosub] calls directly
In yr label1 # [gosub] calls directly
In external sub # Called by non-[goto_block] [gosub]
In yr label1 # 'Fall-through' to label1
In external sub # Called by non-[goto_block] [gosub]
In yr label2 # 'Fall-through' to label2
A [return] within a label's script argument will behave as follows:
* If called from [goto] or sequential execution then [return] will
always exit the [goto_block]
* If called from a [gosub] a will return from the [goto_block] unless
[return -code ok] is used (default is to return -code return)
If [return -code ok] is used then execution will continue with the
next statement in the [goto_block]
goto_block {
gosub label_one
label_one {
puts "Hello user"
return # Use 'return -code ok' to continue execution
}
label_two {
puts "I am never executed"
}
}
If a [break] statement is encountered this will cause the [goto_block] to
exit. A [goto_block] cannot be exited using return, other than using
[return -code break]
Example:
goto_block {
10 { puts "Going to 30"
goto 30
}
20 { puts "Going to end"
goto end
}
30 { puts "Going to 20"
goto 20
}
end { puts "Exiting" }
}
Example:
set a 1; goto_block { 10 {;} 11 {;} 12 {puts $a} 12 {incr a; goto 10} }
Example:
goto_block {
# goto_block used to comment out a block of code
goto end
puts "This statement won't be executed"
end
}
A BASIC-like ON GOTO can be simulated using the following. This behaves the
same as C++ switch()
goto_block {
init { # Initialisation must be within a dummy block
set var [rnd 1 3]
goto $var # Jump to a [goto] entry
}
1 {
puts "One"
break
}
2 {
puts "Two"
break
}
3 {
puts "Three"
break
}
}
Braces are required for the rvalue of each goto sub-block when more than
one command is present and items are separated by whitespace. The following
[goto_block] will jump to the selected label similar to ON GOTO and will
then execute all subsequent blocks sequentially
A [break] may be introduced after each proc call if each jump case needs
to be 'atomic' ...
Example:
proc one {} {
puts "One"
}
proc two {} {
puts "Two"
}
proc three {} {
puts "Three"
}
proc four {} {
puts "Four"
}
goto_block {
init { # Initialisation is in a dummy block
set var [rnd 1 4] # Initialise 'var'
goto $var # Jump to a [goto] entry, below ...
} # .. then continue, sequentially
1 one # Call proc [one]
2 two # Call proc [two]
3 three;break # Call proc [three] (with break example)
4 four # Call proc [four]
}
Example emulated loop with conditional exit:
option expression off
set s 0
goto_block {
start { # Loop start
++ s # Increment
if {> $s 9} { # Test
goto end # Exit
}
goto start # Loop end
}
end {
textcolor magenta
puts "Puts exited OK at label: end"
textcolor
}
}
Example:
'Flip-flops' are memory-stable and can be run infinitely
set sleep_delay 2000
goto_block {
10 {puts "in 10";sleep $sleep_delay;goto 20}
20 {puts "in 20";sleep $sleep_delay;goto 10}
}
Variable $this is set by [goto_block]
Example:
proc foo {} {
puts "Called by $::this"
}
goto_block {
one {
foo
}
goto three
two {
foo
}
three {
foo
}
}
Result:
Called by one
Called by three
Used outside of a [goto_block], [goto] will echo its first argument if
[option echo] is on
Infinite Loops
It is possible to make coding 'mistakes' whereby jump errors within a
[goto_block] may result in infinite loops. Here is a clumsy exmaple...
option expression off
goto_block {
# This block never exits...
init {
set i 0
}
10 {
while {< $i 10} {
puts "Loop $i of 10"
if {> $i 5} {goto init}
++ i
}
}
}
Dynamically-Defined Block Labels
Variables may be used to define [goto_block] labels but these are evaluated
only once and on entry into the block. Once control has passed into the
[goto_block] these can no longer be changed and changes in a label's
variable will have no effect.
set a 1
set b 2
set z 0
goto_block {
init {
puts "In $this"
if {rnd 0 1} {
set a 2 # Reset the ordering
set b 3 # Has no effect on labels
}
goto $b$z
}
$a$z { # Label is fixed at 10
puts "In $this"
goto $a$z
}
$b$z { # Label is fixed at 20
puts "In $this"
}
}
Implementing [break]
--------------------
A root-level immediate break within a [goto_block] can be implemented as
follows
goto_block {
label1 {code}
{} break
label2 {code}
}
or
goto_block {
label1 {code}
label2 break
label3 {code}
}
Background Reading
------------------
https://www.php.net/manual/en/control-structures.goto.php
https://www.mcherm.com/in-defense-of-the-much-maligned-goto.html
http://www.davar.net/PROGRAM/EXTRACTS/GO-TO.HTM
See also: gosub, sub, call, do, while, for, foreach, flow control
The [sub] and [gosub] commands are relatively useless and has been
implemented for entertainment value only. [proc] should be used instead
The [gosub] command is 'syntactic-sugar' and is an optional part of the
Ticol Tcl language. For a discussion of [goto_block] see: goto_block
sub label {code}
gosub label # When used within goto_block as an lvalue
gosub proc # When used within a code block as a command
A [sub] subroutine is a procedure ([proc]) which takes no arguments and
(usually) returns no value. [return] commands are ignored and return
values are discarded. Any arguments given to a [gosub] command are ignored.
[gosub] may also be combined with [goto] within a [goto_block]
For use of [gosub] within a [goto_block],
[sub] labels are case-sensitive. [gosub] may also call any proc but will
pass no arguments to it.
The commands [static], [upvar], [uplevel] etc. may be used within a
subroutine exactly as with [proc]. Local variables in [sub] work in exactly
the same way as with [proc]
[gosub] has two different behaviours. Within a script, including script
blocks of a [goto_block] it will [gosub] (call) a defined subroutine.
At 'root' level within a [goto_block] it will [gosub] (or, call) a block
label, See [goto_block] and [sub]
When calling a label within a [goto_block] a [return] is optional
[gosub] can also call any defined [proc] but will pass no arguments to it
[gosub] will return the return value of any [sub] or [proc] which is
called although in most cases this would be discarded
Example:
sub sub1 {
static counter
++ counter
puts "In sub1. Counter is $counter"
}
gosub sub1 # Used here as an lvalue
gosub sub1 # Used here as an lvalue
Result:
In sub1. Counter is 1
In sub1. Counter is 2
Example:
sub sub1 {
static counter1
++ counter1
puts "In sub 1. counter1 is $counter1"
return # optional, return values are discarded
}
sub sub2 {
static counter2
++ counter2
puts "In sub 2. counter2 is $counter2"
gosub sub1 # Call another sub. Scope prefix required
return # optional, return values are discarded
}
gosub sub1
gosub sub2
gosub sub1
vars
Result:
In sub 1. counter1 is 1
In sub 2. counter2 is 1
In sub 1. counter1 is 2
In sub 1. counter1 is 3
Defined variables at level 0
--------------------------------
integer sub2::counter 1
integer sub1::counter 3
CAUTION: [gosub] subroutines may not call each other or call themselves
recursively and block execution or a stack-overflow error may result.
Example:
Stack overflow conditions can be triggered by the following abuses...
sub sub1 { # Problem 1
gosub sub1 # Self-recursion will cause a stack-overflow
}
gosub sub1
sub sub2 { # Problem 2
gosub sub3 # Mutual calls will cause a stack-overflow
}
sub sub3 {
gosub sub2
}
gosub sub2
The [gosub] prefix is optional outside of [goto_block]. A defined
subroutine may be called directly by referencing its name
Example:
sub sub1 {
puts "In sub1"
}
sub1 # Call the subroutine without [gosub] prefix
Result:
In sub1
Using [gosub] within [goto_block] (Block labels as subroutines)
---------------------------------------------------------------
[gosub] is also allowed within a [goto_block], but the [gosub] call must
be at the same scope level as the labels (i.e. not within a label's code
block scope). Execution of defined subroutine blocks may be skipped by
using a dummy (empty) skip label as in the example below.
The label's code-block must return code 'ok' using 'return -code ok'
Note that [gosub], where used in a block's script, will call an
externally defined [sub], not a block label.
Example:
goto_block {
goto skip # Skip execution of the blocks below...
one {
textcolor darkyellow
puts "* In $this"
textcolor
return -code ok
}
two {
textcolor magenta
puts "* In $this"
textcolor
return -code ok
}
skip {} # Empty/dummy label placeholder
gosub two # Call block 'two' (gosub keyword required)
gosub one # Call block 'one' (gosub keyword required)
}
Results:
* In two
* In one
See also: goto_block, goto, proc, static, upvar, uplevel, eval, dummy,
hasp - Ticol Hasp HL Plugin |
Index |
A HASP HL plugin to query local and remote network HASP licence servers
To load, use: lib ticol_hasp ?-verbose? ?-nocomplain?
Prerequisite:
The HASP HL driver must be installed and running on the target host
Command and Syntax:
list-of-dongle-ids [hasp keys hostname|ip ?-port value?]
list-of-fids [hasp modules hostname|ip keyid ?-port value?]
date-string [hasp expiry hostname|ip keyid fid ?-port value?]
string [hasp product hostname|ip keyid fid ?-port value?]
string [hasp type hostname|ip keyid ?-port value?]
list [hasp data hostname|ip ?-port value?]
list [hasp keydata hostname|ip ?-port value?]
Note that the HASP HL series driver must be installed to run this plugin
Note also that 'fid' is Feature ID code, 'keyid' must be a single key ID,
not a list of key IDs
Non networked (HASP HL Time) can only be discovered on the local machine
[hasp keys] Returns a list of discovered HASP key IDs
[hasp keydata] Returns a list of basic information on a unique HASP key
basis for all known keys. 5 fields are returned per key
Key#, IP, Hostname, Type, Module FID, Module Expiry
[hasp modules] Returns a list of known module FIDs by key ID
[hasp expiry] Returns the expiry of a given module
[hasp product] Returns the product name for a given FID
[hasp type] Returns they key type of a given HASP key ID
[hasp data] Returns all of the known HASP key data from a host
Example:
# List basic key information (per key)
set i 1
foreach x [hasp keydata localhost] {
puts "$i '[lindex $x 0]' '[lindex $x 1]' '[lindex $x 2]'"
++ i
}
Example:
# Do not use
puts [hasp modules localhost [hasp keys localhost]]
Examples:
lib ticol_hasp.dll
# Show all connected+known keys on server foo
puts [hasp keys foo]
# List known module FIDs on the first known key
lib ticol_hasp.dll
puts [hasp modules localhost [lindex [hasp keys localhost] 0]]
Useful references
-----------------
https://sentinel.gemalto.com/software-monetization/sentinel-hasp-hl/
See also: plugins, ticol_rlm
source filename
There is no #include macro for Ticol. Includes are processed in real-time
by using the [eval] command. As the 'included' file will undergo Macro
Preprocessing it should be arranged to include the file only once near the
start of any script.
Avoid using [source filename] inside loops unless careful thought is put
into the script design, including handling variable conflicts and chained
error conditions
Example:
source defs.tcl
puts $LONG_MAX
You can use [eval] to include data to set arrays etc. which has been
defined in a separate file
Example:
# File. include_data.tcl
return {
1 {foo}
2 {bar}
3 {baz}
4 {quux}
}
# Main file which will include the data file
puts "Including include_data.tcl"
array set data [source include_data.tcl]
for {set i 1} {<= $i [array size data]} {++ i} {
puts "$data($i)"
}
Results:
foo
bar
baz
quux
See also: Macro, eval
string [index string N]
Rapid indexing of strings to return a single character. [index] uses base 0
indexing values where 0 represents the first character of a string. Out of
bounds index values will return the null string {}
Example:
puts -[index $qbf 0]-
puts -[index $qbf 1]-
puts -[index $qbf 2]-
puts -[index $qbf 42]-
Result
-T-
-h-
-e-
-g-
See also: strstr, instr, string
Tcl Programming Idioms |
Index |
Idioms are code fragments with a common theme and frequent usage
"A programming idiom expresses a recurring construct in one or more
programming languages. Developers recognize programming idioms by
associating and giving meaning to one or more code fragments" (Wikipedia)
Idioms are often implemented 'longhand' and inefficiently using a number
of separate commands (words). Such idioms may be compressed into a single
command. In the case of Ticol, which is implemented more efficiently as
a distinct command
In essence, they represent one or more ways of doing a common task
Some non-standard idiomatic commands have been included with Ticol.
These can be written by combining several other commands but these
inbuilt idioms improve speed and efficiency by compressing into a
single command
Removing a matching sequence of characters from the start of a string
See [chomp], [html chomp_tag]
Extract complete HTML tags from a text string
See [html chomp_tag]
Align one string into another
See [setl], [setc], [setr]
Wrap text into a fixed width boundary
See [wrap]
Convert JSON data to a Tcl list
See [json_to_list]
Efficiently concatenate strings
See [concat], [append], [store]
Match one of a series of characters to a location in a string
See:[instr], [strstr]
Create a repeated list/string sequence
See [string repeat], [range]
Return the last character of a string
See [lastchar], [mids]
Split a string along fixed width boundaries
See [slice]
Return a single character within a string
See [index], [mids]
Return part of a string
See [mids]
Return all of a string after a given index position
See [mids]
Format a number as a comma-separated sequence
See [comma]
Replace part of a string with a sequence of characters
See [setat], [mid]
Declare and initialise
See [static], [let], [dim]
Show Disk Free Space (UNC or local)
See [diskfree]
References:
https://en.wikipedia.org/wiki/Idiom
if..elseif..else (Flow Control) |
Index |
if {condition} {statement}
if {condition} {statement} ?else {statement}?
if {condition} {statement} ?elseif {statement}? ?else {statement}?
The if statement controls conditional branching
The body of an if statement is executed if the value of the expression is
evaluated as nonzero. The number of elseif statements is unlimited but
execution will slow down as more cases are added
The opening brace for the statement following if or else MUST be on the same
line as the if or else verb.
The standard Tcl keyword 'then' is wholly superfluous and not supported in
order to speed-up execution of the interpreter
Ensure arguments are enclosed in braces
if {condition} {
statements...
}
if {condition} {
statements...
} elseif { condition } {
statements...
} else {
statements...
}
Testing Numeric Conditions
--------------------------
The 'condition' clause may be an expression, numeric literal or [eval]
return. If 'condition' is a numeric expression then test commands are
not required
Example:
set s 23
if {$s} {puts Yes} else {puts No}
set s 0
if {$s} {puts Yes} else {puts No}
if {string length hi} {puts Yes} else {puts No}
if {string length ""} {puts Yes} else {puts No}
Results:
Yes
No
Yes
No
Compact if..elseif Form
-----------------------
Line-continuation characters may be used to produce a compact elseif
ladder without braces. Note that the \ backslash is not required on the
last elseif line:
set failed $false
if {eqi $tag area} {set failed $true}\
elseif {eqi $tag base} {set failed $true}\
elseif {eqi $tag br} {set failed $true}\
elseif {eqi $tag col} {set failed $true} # No backslash on last clause
if {$failed} {
...
}
Subcommands
-----------
The following additional commands are available with [if-elseif-else]
return # Will exit the script when used in root-level [if]
break # Will be passed back to any enclosing structure
stop # Halt operation, return to the CLI if loaded
exit # Exit the script and return to Windows
goto # Where used with an enclosing [goto_block]
Optimising Very Large if...elseif Ladders
-----------------------------------------
See: optimising if
Expression v's Eval
-------------------
Flow control structures use the settings of 'option expression' to decide
whether to test statements using [expr] or [eval]. When using expression on
bear in mind that the argument must be a valid math expression and NOT a
command
If 'option expression' is ON, then you MUST use valid math expressions. Math
expressions must use direct or implied logical tests on command results.
Command calls require square brackets with expression ON in order to return
a boolean value to test
Caution: Don't do this ...
option expression on
if {[$argc > 1]} { ... # [] transforms into a command call
As Ticol will try to dereference argc first and then try to run a command
with that name. Use this form...
option expression on
if {$argc > 1} { ... # A braced expression is required here
Example Expression:
option expression on
if {is_set w} {puts yes} else {puts no} # Not a math expression
if {[is_set w]!=0} {puts yes} else {puts no} # Correct math expression
if {[is_set w]} {puts yes} else {puts no} # Correct math expression
Example Eval:
option expression on
if {[is_set w]!=0} {puts yes} else {puts no} # Wrong for use with eval
if {[is_set w]} {puts yes} else {puts no} # OK for eval
if {is_set w} {puts yes} else {puts no} # OK for eval, [] optional
More complex, combined conditional expressions may be formed using [expr]
Example:
option expression off
if {expr [ne $var_1 ""] && [ne $var_2 ""]} { ...
option expression on
if {[ne $var_1 ""] && [ne $var_2 ""]} { ...
If Tests for Non-Zero Results
-----------------------------
[if] tests for a non-zero expression or command result so expressions may
sometimes return unusual or unexpected results
option expression on
set i 9
if {$i - 10} {puts yes} else {puts no}
if {$i - 9} {puts yes} else {puts no}
if {$i - 8} {puts yes} else {puts no}
Results
yes
no
yes
See also: else, elseif, on, switch, flow control
* Note that adapter and TCP/IP information is unsupported in Windows NT
info adapter_count # Return the count of known network adapters
info adapter_id n # Return adapter GUID
info adapter_name n # Return the adapter name
info adapter_type n # Return the adapter type value
info argcount a # Return the number of non-switch (/) args
info args procname # Return the argument list of a procedure
# Cannot be blocked by TCX encryption
info body procname # Return the script body of a procedure
# Cannot be blocked by TCX encryption
info cmdcount # Return a count of total command calls
info commands ?pattern? # List available commands
info domain ?n? # Show the DNS domain name bound to adapter
info dhcp_enabled n # Return boolean 1 if DHCP is enabled else 0
info dhcp_end n # Show the DHCP lease expiry date
info dhcp_server n # Show any currently configured DHCP server
info dhcp_start n # Show the DHCP lease start date
info disks # Return a Tcl list containing active disks
info disksize x: # Show size of the given disk
info disktype x: # Show the disk type for a given drive/path
# 0 Unknown type
# 1 No root directory/Unavailable/Error
# 2 Removable
# 3 Fixed
# 4 Remote
# 5 CDROM/DVD
# 6 RAMDISK
info encrypted # Returns 1 if script was obfuscated else 0
info executablename # Returns the command line EXE filename
info exists varname # Check if a variable exists, returns 1 or 0
info fmac_address n # Return a formatted adapter MAC address
info frame ?number? # Show the frame number
info functions # Return a list of functions
info gateway n # Return the configured IP gateway address
info globals ?filter? # Returns a Tcl list of variables at level 0
info guid # Return a new, unique GUID string
info hostname # Return the configured PC host name
info ip_address n # Return the IP address for an adapter
info is_dll # Returns true (1) if running in ticol.dll
info is_elevated # Show whether running elevated (Win7+)
info is_networked # Show whether networked or not
info level ?n? # Show the call frame level (top level is 0)
info lib # The library path name
info locals ?filter? # Tcl list of variables at the current level
info mac_address n # The adapter's MAC address
info memtotal # The total physical memory in bytes
info memavail # The available physical memory in bytes
info memtotalpage # The total swapfile memory in bytes
info memavailpage # The available swapfile memory in bytes
info memtotalvirtual # The total virtual memory in bytes
info memavailvirtual # The available virtual memory in bytes
info mempercentfree # Show the % free to 2 decimal places
info monitorcount # Show the number of connected monitors
info nameofexecutable # Ticol's EXE name and path
info netmask n # The adapter's IP subnet mask
info primary_dns n # The adapter's primary DNS server
info primary_wins n # The adapter's primary WINS server
info procname # The currently active proc else ""
info procs ?filter? # A list of user-defined procedures
info script # The name of the active script (if any)
# Is (re)set by [source],[load],[run]
info scriptpath # The escaped, qualified path of the script
# Is (re)set by [source],[load],[run]
info secondary_dns n # The adapter's secondary DNS server
info secondary_wins n # The adapter's secondary WINS server
info tclversion # The Ticol version. Same as $::tcl_version
info username # Return the currently logged-on user
# List variables. See also: info vars
info vars ?pattern? ?-const? ?-nocase?
info winver ?-list? # Windows version information (as a list)
# See: info winver
Where a network adapter is given this is a base 0 value
Multiple network information calls are necessarily slow as an API call has
to be repeatedly made to retrieve each attribute.
# Example Result Comment
# ----------------------------------------------------------------------------
puts [info argcount argv] # 1 Count non-switch '/' args
puts [info args square] # x List procedure arguments
puts [info body square] # * $x $x List procedure body
puts [bool [info exists pi]] # true Check if var pi exists
puts [info exists commands c] # call calldll... List all starting with 'c'
puts [info executablename] # c:\tcl\ticol.exe Shows the EXE name and path
puts [info ip_address 0] # 0.0.0.0 Show IP for adapter #1
puts [info ip_address 2] # 192.168.1.23 Show IP for adapter #3
puts [info hostname] # WORKSTATION23 Show the host name
puts [info tclversion] # 1.06 Ticol-specific version no
puts [info vars pi -const] # pi 3.141592... List vars starting with pi
puts [info procs f*] # foo List procs starting with f
puts [info script] # test.tcl Must be loaded
puts [info level -1] # foo 10 Where called from proc foo
Example:
option expression on
for {set i 0} {$i < [info adapter_count]} {incr i} {
puts "Adapter $i: [info ip_address $i]"
}
Example:
# This script file extension
set filetype [file extension [info script]]
puts "File ext is: $filetype"
See also: mem_used, commands, dump, set, vars
string [info args procname]
Returns the argument parameter of a defined proc. The proc must exist
Example:
proc foo {x {y NULL}} {
puts "In proc [info procname] [info args [info procname]]"
}
foo
Result:
In proc foo x {y NULL}
See: info, mem_used
integer [info cmdcount]
Return a count of total Ticol command calls as an integer
Example:
option expression off
proc factorial {val} {
puts "Current level: [info level] - val: $val"
set lvl [info level]
if {== $lvl $val} {
return $val
}
return [expr {($val-$lvl) * [factorial $val]}]
}
puts "This is how many commands have been executed: [info cmdcount]"
puts "Now this is how many commands have been executed: [info cmdcount]"
puts "\nThis interpreter is revision level: [info tclversion]"
puts "The process id for this program is [pid]"
set count1 [info cmdcount]
set fact [factorial 3]
set count2 [info cmdcount]
puts "The factorial of 3 is $fact"
puts "Before calling the factorial proc, $count1 commands executed"
puts "After calling the factorial proc, $count2 commands executed"
puts "It took [expr $count2-$count1] commands to calculate a factorial"
Results:
This is how many commands have been executed: 133
Now this is how many commands have been executed: 135
This interpreter is revision level: 1.14
The process id for this program is 4352
Current level: 1 - val: 3
Current level: 2 - val: 3
Current level: 3 - val: 3
The factorial of 3 is 18,3,3
Before calling the factorial proc, 142 commands executed
After calling the factorial proc, 176 commands executed
It took 34 commands to calculate a factorial
See also: info
bool [info exists variableName|arrayName]
The argument for variable or array name must be a name-literal. i.e. it
cannot be prefixed with a dollar sign as this will cause automatic
dereference to the variable value. Dollar-subscript values may be required
for arrays.
Note that the array syntax prohibits whitespace between the variable name
and the open parenthesis:
Correct: "$varname($index)"
Incorrect: "$varname ($index)" (looks for non-array variable $varname)
Use [array exists arrayname] to test for the presence of an array by name
info exists argv0
info exists argv(0)
set i 0
info exists argv($i)
info exists env(TEMP)
array exists argv
array exists env
See also: info, array exists
info level ?number?
Displays information about the calling frame at a given level. Level 1 in
a proc will be the current level, level 0 the calling frame for that proc.
If 'number' is not specified, info level returns a number indicating the
current calling level, i.e. the level from which info level was called with
0 being the top level, and each subsequent level incremented by 1
If 'number' is specified, info levels returns a Tcl list containing words
of the command for the level indicated by number. A number value greater
than 0 indicates the level identified by that number. 0 indicates the
current level, -1 indicates one level prior to the current level,
-2 indicates the level prior to that one etc.
Example:
proc foo {bar {baz NULL}} {
puts "info level 0 is \[[info level 0]\]"
}
foo
foo abc
foo abc def
Results:
info level 0 is [foo]
info level 0 is [foo abc]
info level 0 is [foo abc def]
See:
http://wiki.tcl.tk/1720
See also: info
list [info vars ?pattern? ?-const? ?-nocase?]
Return a Tcl list of matching variables. A glob pattern may be supplied
and the search will be case-sensitive unless the -nocase argument is
included. A list of constants may be returned by using the -const
argument
set foo 1 # Create a standard var
const Example 1 # Create a const
puts [info vars] # All variables regardless of case
puts [info vars e*] # All variables precisely matching e*
puts [info vars f??] # All variables precisely matching e*
puts [info vars F?? -nocase] # All variables matching f?? or F??
puts [info vars N* -nocase] # All variables matching n* or N*
puts [info vars -const] # All constants
puts [info vars e* -const] # All constants matching e*
puts [info vars e* -const -nocase] # All constants matching e* or E*
Results:
argc argv argv0 env eof errorCode errorInfo errorLine errorMsg exe _name
false logo NAN pi qbf tcl_cli tcl_config tcl_date tcl_precision tcl_thre
aded tcl_version true win64 env eof errorCode errorInfo errorLine errorM
sg exe_name
foo
foo
NAN
argc argv argv0 env eof exe_name false logo NAN pi qbf tcl_cli tcl _conf
ig tcl_date tcl_precision tcl_threaded tcl_version true win64 env eof eu
ler exe_name env eof Example exe_name
The [vars] command gives a diagnostic snapshot of current variables and
an abridged view of their type and contents. See: vars
See also: vars, info
info winver ?-list? ?-vernum? ?-truever?
Show information about the current Windows version
Argument -list will return full Windows information in Tcl list format as
{Version name} major.minor.build {Service pack} bitness
Argument -vernum will return the full version number as a double value
Argument -truever will show the true version number on later Windows
builds
Example:
puts [info winver -list]
Result:
{Windows XP} 5.01.2600 {Service Pack 3} 32
Example:
puts [info winver]
Result:
Windows XP [Version 5.01.2600] (Service Pack 3) 32 bit
Example:
# Show whether Windows is 64 or 32 bit
lindex [info winver -list] 3
Result:
64
Example:
puts [info winver -vernum]
Result
5.1
See also: info
string [input ?length? ?default-string? ?-f forecolour? \
?-b backcolour? ?-password? ?-crlf? ?-var varname?]
Input a fixed length string from the console. This is a simple input
editor with backspace delete only. Full line-editing is not supported.
WARNING: DO NOT EVALUATE USER INPUT WITHOUT INPUT SANITISATION
User input should be protected by being wrapped in {} braces
as the user might input malicious Tcl code. Avoid using [eval]
or [expr] on the result of user-inputs without checks.
The default input length is 30 characters
Optionally, the input area foreground and background colour can be set
and input characters can be masked with '*' if the -password argument
is used. Backspace editing/deleting is supported. Insert mode is not
supported. The input length cannot be exceeded or a bell alert will
sound
Input is confirmed by pressing the RETURN, ENTER or TAB keys
If ESC is pressed then any input text is discarded and input exits
If '-var varname' is passed then an integer keycode value will be
returned in that variable
Example:
# Input 20 characters using white on blue text
puts "Enter 20 chars: "
set s [input 20 -f white -b blue]
newline
puts "You input '$s'"
Example:
# Input a password, hiding the input
puts "Enter password of 20 chars: "
set s [input 20 -password]
newline
puts "You input '$s'"
See also: inkey, getkey, gets, input, input_csv, gotoxy, box, inputbox
integer [inkey]
Check if there is a key input event pending and retrieve the keypress code
If a key-event is pending the key value as an integer representing the ASCII
code is returned, otherwise 0 is returned
[inkey] does not halt for keyboard input. Use [getkey] to wait for input
Example:
puts "Press ESC to halt"
while {1} {
puts "Looping"
sleep 1000
if {== [inkey] 27} { # ESC is 27
stop
}
}
See also: key input, getkey, gets, input, input_csv, gotoxy, box
input-csv filename ?varname | -? ?-max items-max? ?-crlf? ?-array name?\
?-r? ?-retries retry-count?
Read lines from a standard-format CSV file and translate to a properly-
formatted Tcl list. The CSV input-file may be contain double-quote wrapped
strings but all fields must be comma separated
If no variable name is given then [input_csv] returns the list. If a
variable name is passed then the variable is assigned (even if this is an
empty list) and a count of the number of list elements is returned
The new variable name is optional but if lines-max is specified and varname
is omitted then supply a hyphen (-) character instead
The optional '-crlf' argument will retain newline pairs in the resulting
output
'-max' is the maximum number of array items to read from the file
'-retries' is the maximum number of 1000ms file open retries to attempt
The minimum is zero retries and the maximum is 200 (200 x 1000ms)
CSV read and conversion may become slow for files over about 5Mb and
excessively so for files above 10 or 20Mb+. Consider sequential. line-based
input into smaller Tcl lists instead
Each line (row) of the file can be read to a Tcl array using the -array
argument with a non-existent variable name. Subscript->value assignment
can be reversed or 'inverted' using the -r argument
No unicode/wide character conversion is performed
Example:
CSV file:
"Hello world",1234,"ABC",5.678
"The quick brown fox",0123,"CDE",23.23
puts [input_csv csvfile.csv - -max 4]
Result:
{Hello world} 1234 Tcl 5.678
See also: json_to_list, file read
[input] offers console line-input with text input length limits, backspace
editing and text input area colourisation as well as password character
hiding
[inkey] provides polled single key input to enable you to build your own
keyboard-interaction routines
[gets stdin] is a stream-based method of retrieving user input from stdin
It offers limited line-editing facilities and no option to fix the length
of the user-input or prevent line-wrap. It is not recommended for general
user interaction. Use [input] instead
Example:
puts "Enter 20 chars: "
set s [input 20 -f white -b blue]
newline
puts "You input '$s'"
Example:
puts -nonewline "Throw a die. What value did you get? "
flush stdout
set top [gets stdin]
switch -- $top {
1 {set end "st"}
2 {set end "nd"}
3 {set end "rd"}
default {set end "th"}
}
puts "You selected the $top$end face"
See also: gets, flush, switch
string [inttooct integer-value]
integer [octtoint octal-value ?-unsigned?]
Converts to and from from decimal integer and octal format
Conversion from octal to decimal is automatic in all routines
Octal format is defined by an "0o" prefix. e.g. "0o12"
Values given to [inttooct] may be decimal integers, hexadecimal, octal or
binary values
Tcl v9 octal number identifiers are implemented, which require 0oNN format
CAUTION
-------
The Macro PreProcessor (MPP) will automatically convert octal values in
the form 0oNN into decimal. Be cautious when passing unquoted octal
values with "0o" prefix to [octtoint]
Examples:
puts [inttooct [expr "0o123+1"]]
puts [inttooct 0b1101001]
puts [inttooct 0x68]
puts [octtoint 0o124] # Unquoted octal will be converted by MPP
puts [octtoint "0o124"] # Quoted octal will be retained
Results:
0o124
0o151
0o150
0
84
See:
http://wiki.tcl.tk/498
See also: intotohex, inttobin, tohex, fromhex, cast
integer [strchr string charstring ?-nocase?]
C++ like string search command. Finds the location of the first character
from 'charstring' in 'string' If found a base 1 value is returned,
otherwise zero is returned.
Example:
puts [strchr $qbf "k"]
puts [strchr "Hello" "x"]
Result:
9
0
See also: strstr, instr
bool [inrange integer_value reference_target ?percent?]
[inrange] tests a signed integer value to see if it is within
+ or - X percent of the given reference range integer value.
Integers are standard 64 bit range for Ticol integers
The default percentage is +/- 10%. Alternate ranges may be specified as
an integer value in the range 0% upwards
[inrange] is useful for testing integer values which may deviate by small
amounts within a required reference range. e.g. timing keyboard inputs
in milliseconds
Example:
set r [rnd 9000 1000]
puts [inrange $r 1000]
Result:
1
See also: data type ranges, inkey, pause, gets, input
void [inspect varname ?tag-string?]
[inspect] allows raw variable information to be inspected. Optionally
a tag string may be included, which may be a comment string or line number
Non-printing characters are not displayed
Example:
set a(0) 23
inspect a(0) Line:#__LINE__
See also: walk, dump, debugging, troubleshooting
instr, rinstr, strstr, strstrr |
Index |
integer [instr haystack needle ?startpos? ?-nocase? ?-var?]
integer [rinstr haystack needle ?startpos? ?-nocase? ?-var?]
integer [strstr haystack needle ?startpos? ?-nocase? ?-var?]
integer [strstrr haystack needle ?startpos? ?-nocase? ?-var?]
VB and C++ - like string search commands. Use either the VB or C++ alias
[instr] and [strstr] search forward from the start of a string
[rinstr] and [strstrr] search backward from the end of the string
Where 'haystack' may be a variable or literal string and where 'needle'
may also be a variable or literal string
The 'startpos' argument is an index base 1 value indicating which character
to start the comparison from. e.g. 1 is the 1st character in the string
Returns a 'base 1' index value of 'needle' into 'haystack'
If 'needle' is not found in 'haystack' then 0 will be returned.
If 'haystack' is a null "" string then 0 is returned
If 'needle' is a null "" string then 0 is returned
The result may be treated as a boolean - not found being 0 and found non-zero
Example:
puts [instr $qbf "fox"]
puts [instr $qbf "QUICK" -nocase]
if {[instr $qbf "lazy"]} { puts "True" }
Results:
17
5
True
Example:
puts [strstr "quick brown fox" "brown"]
puts [strstr "quick brown fox" "BROWN" -nocase]
Result:
7
7
Example:
option expression on
if {instr $qbf fox} {
puts "I found the fox!"
}
Result:
I found the fox!
See also: setl, setr, mid, string
inputbox ?prompt? ?title? ?default? ?-password? ?-x X? ?-y Y?\
?-w W? ?-h H?
Creates a windows which can receive user input, optionally with input masked
by a password-masking character. Inputs are confirmed by ENTER/OK click or
cancelled using ESC/CANCEL click
Options
-------
-password Mask input characters with an asterisk (*)
-x X Locate at cursor X
-x Y Locate at cursor Y
-w W Make dialog W wide
-h H Make dialog H high
If -x and -y are omitted the dialogue will centre on the screen
Examples:
puts [inputbox "Enter your name:"]
puts [inputbox "Enter password:" "secret" -password]
puts -[inputbox "Enter your name" "Rupert" -w 500 -password]-
puts [inputbox "Enter your name:" -x 200 -y 300 -w 500]
See also: msgbox, get_filesave, get_filesave, get_folder, input, gets
ip_to_long, long_to_ip |
Index |
integer [ip_to_long value]
string [long_to_ip value]
Convert an IPV4 dotted-quad address to an unsigned long value or convert a
long value to an IPV4 dotted-quad address. The address can be partial, in
which case it will be evaluated from left to right (e.g. "192.168.0")
The caller is responsible for any 32-bit integer bounds checking
Example:
puts [inttohex [ip_to_long 255.255.255.255]]
Result:
0xffffffff
Example:
# Partial address
puts [inttohex [ip_to_long 192.168.0]]
puts [ip_to_long 192.168.0]
Result:
0xc0a80000
3232235520
Example:
puts [long_to_ip 0xffffffff]
Result:
255.255.255.255
See also: info
item (array, stack or struct) |
Index |
string [array item arrayName subScript ?default? ?-nocase?]
string [stack item stackName index ?default?]
string [struct item structName.fieldname ?default?]
The item argument returns a value if it exists. If the referenced object
does not exist then either "" or a default value is returned
See also: item, array, stack, struct
string [tohex string]
string [fromhex string]
Converts a string to and from a hex-encoded version
Examples:
tohex "hello"
fromhex 68656C6C6F
puts [fromhex [tohex "Hello world"]]
Result:
68656C6C6F
hello
Hello world
See also: inttohex, inttooct, cast
number [funct function-name arg ?arg? ...]
Call an [expr] function directly and separate from the [expr] command
[funct] is more efficient than calling expr where access to a single
function is required. Complex expressions are not supported within the
arguments unless returned by [expr]
Examples:
puts [funct log10 2]
puts [funct round $pi 2]
puts [funct rnd 1 10]
puts [* [funct atan 1] 4]
Result:
0.301029995663981
3.140000
6
3.141592653589792
See: 'functions' and 'expression functions' for the full list of functions
exported from [expr]
See also: functions, expr, eval, function arity
list [functions ?glob-pattern?]
Return a Tcl list of math functions defined for use with the expr command
The search may be constrained by a glob search value. The list may be
iterated with [foreach]. The search case is not relevant as all functions
are pre-defined internally as lower-case.
See: 'function arity' and 'expression functions' for more details about
particular functions and their arguments
Example:
# Full list of all [expr] and [funct] functions in this version
puts [functions]
Result:
abs acos asin atan bool ceil decr div cos cosh degrees double exp fib
fix floor fmod fraction hypot incr int integral log log10 max min mul
pow rand rnd round sgn sin sinh sqr sqrt srand tan tanh wide
Example:
# List all functions starting with the letter 'm'
puts [functions m*]
Result:
max min mul
Use: option echo on to display directly to the console
See also: expression, procs, funct
BEDMAS/BODMAS/BOMDAS (Math Operator Precedence) |
Index |
Ticol implements BEDMAS to automatically prioritise math expressions:
BEDMAS: Brackets, Exponent, Division, Multiplication, Addition, Subtraction
This is also known by other acronyms such as PEDMAS, BOMDAS etc.
The order of operations used throughout mathematics, science, technology and
many computer programming languages is expressed as:
Exponents and roots
Multiplication and division (of equal preference, thus left-to-right)
Addition and subtraction (of equal preference, thus left-to-right)
Where an operator has equal precedence, (e.g. * and /) then a left to
right evaluation is performed
This aspect of mathematics appears to date from well before the 17th
century and frequent internet postings suggest that this comes as a
surprise to many
See:
http://en.wikipedia.org/wiki/Order_of_operations
See also: math operators, math expressions, expression functions
Tcl does not natively understand math, although individual commands are
able to process math commands and expressions
The built-in [expr] command can handle complex expressions and applies
standard BEDMAS rules by reprioritising expressions before evaluation.
During the checking phase any non-decimal consts such as 0xFF or 0o777 are
automatically converted to their decimal equivalents. Multiplication and
division are given priority over addition and subtraction
Main flow-control commands may take either an [expr] or [eval] argument
which is controlled by the 'option expression' setting. With 'option
expression on' commands such as [if] and [while] will take an [expr]
argument. With 'option expression off' you will need to pass eval results
of individual Tcl commands
e.g.
option expression on
if {30 / 10 == 3} { ...
option expression on
if {== [/ 30 10]} { ...
Exponentiation is supported via the operator '**' and pow() expression
function
Example:
puts [expr 2**8]
puts [expr pow(2,8)]
Results:
256
256.000000000000000
The contents of pre-bracketed expressions and function brackets are also
evaluated for correct ordering and precedence.
Value interpretation follows Tcl rules. Integers are subjected to integer
division. If you wish float division then specify a decimal point.
e.g. - 22/7 will return 3, 22/7.0 will return 3.142857142857143
For a full list of supported math functions see Expression Functions
Note that commas ',' are undefined within math expressions other than as
function parameter separators and their inclusion elsewhere may cause
unpredictable results.
You may also apply your own expression parentheses.
Example:
# Call [expr] then external BAT file which calls cscript->eval
# abs(1+11*2-1/7.0)/sqrt(2/1.6) is reinterpreted as:
# abs(1+(11*2)-(1/7.0))/sqrt(2/1.6)
puts [expr abs(1+11*2-1/7.0)/sqrt(2/1.6)]
20.444050079998078
puts [eval.bat abs(1+11*2-1/7.0)/sqr(2/1.6)]
20.44405007999810
Note: The big math plugin module can be used to extend the number range
See also: math operators, BEDMAS, expression functions
float [fmod float float]
Return the remainder after float division
Example:
puts [fmod 1.34 0.45]
Result:
0.440000
See also: %, mod, math
Combination Math Operators |
Index |
Combined operation and assignment commands
value [+= var|value value] # Combined addition and assignment
value [-= var|value value] # Combined subtraction and assignment
value [*= var|value value] # Combined multiplication and assignment
value [/= var|value value] # Combined division and assignment
value [|= var|value value] # Combined or + assignment
value [&= var|value value] # Combined and + assignment
value [^= var|value value] # Combined xor + assignment
These operators can take either a variable name or a value (dereferenced
from a variable or as a literal numeric constant)
Unsigned 64 bit operations can be forced using the '-unsigned' argument
They are more efficient than using a longhand form with [set] using
set var [+ var value]
Examples:
set a 23
set b 12
puts [+= 23 12]
puts [+= $a 12]
puts [+= a 12]
puts [+= 23 $b]
puts [+= $a $b]
puts [+= a $b]
Results:
35
35
35
35
35
35
See also: math operators, optimisation
The following math operators are supported.
CAUTION: Operations are treated as integer or float unless mixed types are
encountered, in which case float is used. Take care with division; if you
want a float result then one of the operands must be a float value or your
result will be truncated to an integer value: e.g. Use: [expr "22/7.0"]
not [expr "22/7"]
+ Addition
- Subtraction
* Multiplication
/ Division. At least one argument must be float to return a float
% Modular integer division
** Power. Returns x raised to the power of y (See: pow)
Integer overflow is fatal
Double overflow returns 1.#INF00000000000
fmod Floating point Modular integer division as a command
fmod() Floating point modular division (as an [expr] function)
Combination Operators (See also: Combination operators)
+= Combination add and assign
-= Combination sub and assign
*= Combination mul and assign
/= Combination div and assign
|= Combination or and assign
&= Combination and and assign
^= Combination xor and assign
Integer Only:
^ Bitwise xor
<< Bitwise shift left (64-bit)
>> Bitwise shift right (64-bit)
| Bitwise or
|| Logical or
& Bitwise and
&& Logical and
The above may be applied as combined operator= symbols, e.g. ^=, &=
Relational Operators:
All relational operators may be qualified using an '-unsigned' option
which will force an unsigned 64 bit integer comparison
== Equivalence
!= Inequality
> Greater than
>= Greater than or equal to
< Less than
<= Less than or equal to
<=> Relational test (same as PERL operator)
Returns -1 for <, - for equal, +1 for > as per C++ strcmp
is_mod Safe modular division test
() Precedence bracketing
Division by zero will return an error as either:
Integer division by zero (10 / 0)
Float division by zero (10 / 0)
When used as a command, the math operators such as [add] [sub] etc. can be
used with a variable name as the following method of shortcut assignment...
set a 1
add a [expr 3*3]
add a [* 3 3]
Result:
10
19
Note: The big math plugin module can be used to extend the number range
See also: incr, decr, ++, --, fmod, is_div, shortcut operators, power,
BEDMAS, math expressions, is_mod, big math, combination operators
Shortcut Math Operators |
Index |
Syntax: <operator> <variable> <value>
Operator Mnemonic Domain
------------------------------------------
+= Add Integer, float
-= Sub Integer, float
*= Mul Integer, float
/= Div Integer, float
&= BitAnd Integer
|= BitOr Integer
^= BitXor Integer
The shortcut math operators help simplify and speed up math expressions.
For example, the longhand: set a [* a 3] may be condensed to: *= a 3
Examples:
Set a 1 # 1
+= a 10 # 11
-= a 2 # 9
*= a 3 # 27
/= a 9 # 3
|= a 4 # 7
&= a 6 # 6
^= a 2 # 4
^= a 2 # 6
See also: math operators
<< >> Bit shift (64-bit) |
Index |
integer [<< value bits ?bit-width?] # Bitwise shift left
integer [>> value bits ?bit-width?] # Bitwise shift right
Either left or right-shifts a value by N bits. This is a valid operation
for integers only. Negative shift values will raise a catch exception
The expression forms of << and >> do not accept bit-width arguments.
If you need to specify a bit width then insert the command in either
expression or eval pass as [<< value shift bits] [>> value shift bits]
Examples:
puts [expr [<< [expr 20+3] [expr 2^4] 32] - 472]
puts [expr 1*[<< 1 1]*3/4.0*5]
Result:
1000
7.5
Note that all Ticol integers are 64-bit and by default these bit operations
work on 64-bit integers. This can be modified by adding an output bit-width
argument of either 8,16 or 32. Any other values will evaluate as 64-bit
Shift does not wrap or suffer from an underlying 32-bit bias for other
bit-widths. All shifts equal-to or in excess of the specified bit width
will return 0
Wrapping does not occur (32-bit wrapping is blocked)
Shift [<<] and [>>] fills vacated bits with 0. All values are unsigned.
Sign-bits do not therefore, cause 'backfill' with [>>]
Examples:
puts [inttohex [<< 0xff 32]]
puts [inttohex [<< 0xff 63]]
puts [inttohex [<< 0xff 64]]
loop i 0 40 8 {
printf "\[inttohex \[<< 0xff %2i 32\]\]='%20s'\r\n" $i\
[inttohex [<< 0xff $i 32]]
}
Result:
0xff00000000
0x8000000000000000
0x0
Examples:
puts [inttohex [<< 0xff 0 32]] # 0xff
puts [inttohex [<< 0xff 8 32]] # 0xff00
puts [inttohex [<< 0xff 16 32]] # 0xff0000
puts [inttohex [<< 0xff 24 32]] # 0xff000000
puts [inttohex [<< 0xff 32 32]] # 0x0
See also: math operators, functions
Expression Functions (Built In Functions) |
Index |
The following standard C/C++ compatible functions are available within Tcl
expressions using the [expr] command.
See: function arity for more information
abs(x) # Return the absolute value of a number
acos(x) # Return the arc cosine of value (arccos)
asin(x) # Return the arc sin of value (arcsin)
atan(x) # Return the arc tangent of value (arctan)
bool(x) # Evaluates a string and returns 1 or 0
ceil(x) # Round to the next-highest integer value
cosh(x) # Return the cosine of value
decr(x) # Decrement value
div(x,y) # Functional division of x / y
cos(x) # Return the cosine of value (in radians)
cosh(x) # Return the cosine of value (in radians)
degrees(x) # Convert radians to degrees
double(x) # Cast value to double
exp(x) # Return the exponent of value
fib(x) # Return the Fibonacci of value
fix(x) # Return integer value rounded up
floor(x) # Round to the next-lowest integer value
fmod(x,y) # Return the remainder of x / y
fraction(x) # Return the fraction part of a real/double
hypot(x,y) # Return the hypotenuse of x and y
incr(x) # Increment value
int(x) # Cast value to integer
integral(x) # Return the integer part of a real/double
log(x) # Return the logarithm of value
log10(x) # Return the log10 of value
max(x,y) # Return the maximum of x and y
min(x,y) # Return the minimum of x and y
mul(x,y) # Functional multiplication of x * y
pow(x,y) # Return x to the power of y
rand() # Return double random value from 0 to 1
rnd(x,y) # Return a random number between a and b inclusive
# The rand() and srand() functions are not
# cryptographically secure, and should not be used
# to generate one-time passwords or session keys
round(x,y) # Round value x to y places using standard rounding
sgn(x) # Return the sign as an integer (1, 0, -1)
sin(x) # Return the sine of value (in radians)
sinh(x) # Return the sine of value (in radians)
sqr(x) # Return the square of value
sqrt(x) # Return the square-root of value
srand() # Seed the random number generator
tan(x) # Return the tangent of value (in radians)
tanh(x) # Return the tangent of value (in radians)
wide(x) # Force a value to 64 bit signed integer
?x:y # Functional [if] statement (deprecated)
Available functions may be listed using the 'functions' command
See: function arity for more details about functionality
The functional int differs from the int conversion command. int(x) will
round to the lowest integer whereas [int x] will simply cast and convert
Int returns the first -ve integer less than or equal to value, whereas Fix
returns the first -ve integer greater than or equal to value.
Functions may be nested within expressions passed to the 'expr' command
and may also be called using the [funct] command using [funct <name>]
Where radians are required (cos, sin etc.) radians = degrees * (pi/180)
or: degrees * 0.017453292519943.
The 'sgn' function returns:
Greater than zero 1
Equal to zero 0
Less than zero -1
Example:
expr "int(22/7.0)+int(5.2)"
Result:
8
Example:
expr "round(exp(log(10)),0)"
Result:
10
Example:
expr ([rnd 0 1]?yes:no)
Result:
Randomly yes or no
Examples using [funct]
puts [funct max 2 10]
puts [funct sqrt $pi]
Result:
10
1.772453850905516
See also: function arity, math operators, math expressions, arity, expr,
sqrt proc, eval, math, funct, functions
A square root command is not defined but is available as a math function
Math functions can be called via the [funct] command or a useful code
snippet can be wrapped-up as shown below
This allows a default precision of 3 and the option of passing a precision
parameter to suit
proc sqrt {x {precision 3}} {
return [format %.${precision}f [funct sqrt $x]]
}
puts [sqrt 169]
puts [sqrt $pi]
puts [sqrt $pi 5]
puts [sqrt $pi 7]
Results:
13.000
1.772
1.77245
1.7724539
See also: sqrt function, functions, math
getcwd ?-unix?
pwd ?-unix?
Returns the current working directory
[getcwd] will return the path as a backslash-escaped string for Tcl
processing [pwd] will print out the current working directory for display
purposes. The -unix argument will force Unix path separators to be returned
Example:
pwd
getcwd
getcwd -unix
Results:
C:\windows\system32
C:\\windows\\system32
See also: mkdir, mkdirs, file, dir
get_fileopen, get_filesave |
Index |
get_fileopen filter title initdir ?flags?
get_filesave filter title initdir ?flags? ?defaultname?
Launches a Windows file select (open) dialogue or a file select (save)
dialogue For get_filesave a default filename may be specified after the
flag value. Unwanted flag values should be specified as 0
Any option which is specified as "-" will be ignored
Example:
set f [get_filesave *.txt "Save Txt File" "d:\temp"]
set f [get_filesave *.txt "Open Txt File" "d:\temp" 0 fred.txt]
See also: msgbox, inputbox, get_folder
string [getenv]
Return an environment variable from Windows
Example:
puts [getenv ProgramFiles]
Result:
C:\Program Files (x86)
See also: setenv
string [gets ?stdin|filenum|-? ?var|-? ?-length count?]
[gets] has an enhanced syntax compared to standard Tcl. The first two
arguments are optional and each of those may have a default specified
using "-". The default stream is stdin and the default variable is none.
Read data from an open file-stream. This may either have been opened by the
open command or the stdin read stream may be used. Optionally, a variable
may be specified to receive the data. This need not be declared in advance.
The default file-stream is 'stdin'. A numeric value is assumed to represent
a length value
Additionally an optional number of bytes to read from the stream may be
specified
If 'count' is specified then CTRL+Z (^Z or EOF) terminates input before
the required number of characters has been read. If count is not specified
then input terminates when ENTER is met
When using [gets] to read from the stdin console it supports up/down cursor
history, home/end and full inset-mode line editing
Example:
puts "Enter your name: " -nonewline
set a [gets]
newline
Example:
puts "Enter your name: " -nonewline
gets a
newline
Example:
puts "Enter your name up to 20 characters: " -nonewline
gets a 20
newline
Example:
puts "Enter your name up to 20 characters: " -nonewline
gets stdin a 20
newline
Example:
set fp [open "input.txt" r]
gets $fp a 100 # Read 100 chars
close $fp
See also: input, inputbox, file, open, close
gotoxy x y
Locate the screen cursor at an x/y location. The values are base 1
If using [puts] with absolute screen coordinates via [gotoxy] you should
use the -nonewline argument if possible to avoid screen scrolling issues
[screen curx] and [screen cury] give the current x/y locations
Example:
option expression off
for {set i 0} {< $i 10} {++ i} {
gotoxy 5 5
puts "I am at 5,5" -nonewline
}
See also: screen, cls, box
The Towers of Hanoi script is a useful benchmark for speed. A compiled
C++ version of the Tcl algorithm is a couple of orders of magnitude
faster, but this is to be expected as Ticol Tcl is interpreted and
outside the plugin module level, is not intended to be coded for
performance
The following example code, 'hanoi 17' runs in around 4 seconds on a
low-powered i5 under Windows 7.
option expression off
cls
set n 0
set yloc 1
if {! [defined draw_status]} {
proc draw_status {} {
upvar yloc
gotoxy 1 $yloc
upvar a
upvar b
upvar c
upvar n
upvar moves
upvar x
textcolor white
puts "Towers of Hanoi ($n) - Pole Status ([comma $x] moves)\n"
textcolor red
printf "Pole A(%2i) %-67s\r\n" [stack count a] [stack list a]
textcolor yellow
printf "Pole B(%2i) %-67s\r\n" [stack count b] [stack list b]
textcolor green
printf "Pole C(%2i) %-67s\r\n" [stack count c] [stack list c]
textcolor
}
}
halt
textcolor white blue
puts " Towers of Hanoi "
textcolor
newline
if {[< $argc 2]} {
puts "How many rings (2..25)? " -nonewline
gets stdin n
if {|| [== $n ""] [== $n 0]} {stop}
} else {
set n $argv(1)
}
set n [min $n 25]
stack create a $n # Init 3 stacks to given size
stack create b $n
stack create c $n
for {set x 0} {[< $x $n ]} {++ x} { # Create initial stack of hoops
stack push a [+ $x 1] # 0..n-1
}
set x 1
set shln [<< 1 $n] # Precalc (1 << $n), << for n
draw_status
set start_secs [clock seconds]
set start [timer]
while {< $x $shln} {
set xminus1 [- $x 1]
stack push [chr [calc ((($x|$xminus1) + 1) % 3 )] 97] \
[stack pop [chr [calc ($x & $xminus1) % 3] 97]]
is_mod $x 50000 draw_status
++ x
}
set end [timer]
set end_secs [clock seconds]
draw_status
newline
puts "Took [- $end $start] millisecond(s) and [comma $x] move(s)"
puts "Took [- $end_secs $start_secs] second(s) and [comma $x] move(s)"
newline
puts "Done."
stack unset a
stack unset b
stack unset c
Results:
Took 3884 millisecond(s) and 131,072 move(s)
Took 4 second(s) and 131,072 move(s)
For comparison. The compiled C++ hanoi.exe 17 gives the following
and can run 'hanoi 23' in about the same time with 8,388,607 moves
Took 47 ms and 131071 move(s)
string [tohex string]
string [fromhex value]
string [inttohex decimal-variable ?-noprefix? ?-full?]
integer [hextoint string]
Conversion routines to convert from text and integers, to and from
hexadecimal
Note that because the Macro PreProcessor will translate hex values if
prefixed with '0x', you should wrap any such values in double quotes
before calling [hextoint], or you should strip any prefixing '0x',
otherwise inaccurate results will be returned due to double, interpretation
of the hex value
[tohex] Converts a string into hex-encoding, the inverse is [fromhex]
[inttohex] converts a decimal value to base 16, the inverse is [hextoint]
Conversion from hex to decimal integer is automatic in Ticol
The -full argument presents all prefixing zero values.
Normally these leading zeroes are omitted
If the -noprefix argument is given then the 0x... prefix will be omitted
[printf] and [format] can also be used to generate more control over
hex formatted output
Example:
puts [inttohex 12345]
Result:
0x3039
Examples:
puts [inttohex 12345]
puts [inttohex 12345 -noprefix]
puts [hextoint "0x3039"]
puts [tohex "Hello world"]
puts [fromhex 48656C6C6F20776F726C64]
Results:
0x3039
3039
48656C6C6F20776F726C64
Hello world
See also: int, double, data types
Code Highlighting Plugin |
Index |
Load using: lib ticol_highlight?.dll?
string [highlight string ?-f bat|asm|html? ?-rtf?]
string [highlight {} -reset]
[highlight] was created to output the Ticol man file as both HTML and RTF
format files and is designed for highlighting on a black background but
could be used with grey background.
It can create a HTML or RTF-highlighted block of text from Tcl, Windows
Batch, Ticol Assembler or HTML script code using the Ticol man file
markup syntax.
Arguments
---------
-html Generate HTML output (default, -html is not required)
-rtf Generate RTF output
-f asm Ticol Assembler (TicASM) format
-f bat Windows Batch Script format
-f html HTML code format
-f tcl Tcl format (default, -tcl is not required)
-reset Reset multiline string highlighting
When using in -html output mode, any < and > characters will be replaced
with < and >
Rich Text (RTF) output can also be generated using a custom {colortbl}
entry. The default output format is HTML.
Call as [highlight {} -reset] to reset multi-line string text highlighting
if necessary
Unescaping Inputs
-----------------
Escaped Tcl source should be passed in 'unescaped' form using [unescape]
RTF Colour Table
----------------
Highlighted code will refer to \cf1 to \cf16 and the following RTF colour
table header can be used and which match the standard DOS/Windows console
colours: This will allow embedded colour codes '\cf1' to \cf16' to map
into this console colour table. Note that "darkyellow" is represented
as 'orange' not 'khaki'.
# RTF emulation of classic Windows console colours
array set rgb_colour { # Used to set individual RGB colortbl elements
# array {R G B} cfN hex binary name \cfN
0 {0x00 0x00 0x00} # 01 0x00 0b0 black cf1
1 {0x48 0x3D 0x8B} # 02 0x01 0b1 darkblue cf2
2 {0x00 0xC6 0x00} # 03 0x02 0b10 darkgreen cf3
3 {0x00 0x8B 0x8B} # 04 0x03 0b11 darkcyan cf4
4 {0x80 0x00 0x00} # 05 0x04 0b100 darkred cf5
5 {0x8B 0x00 0x8B} # 06 0x05 0b101 darkmagenta cf6
6 {0xD2 0x69 0x1E} # 07 0x06 0b110 darkyellow cf7
7 {0x69 0x69 0x69} # 08 0x07 0b111 darkgrey cf8
8 {0x80 0x80 0x80} # 09 0x08 0b1000 grey cf9
9 {0x00 0x00 0xFF} # 10 0x09 0b1001 blue cf10
10 {0x7C 0xFC 0x00} # 11 0x0a 0b1010 green cf11
11 {0x00 0xFF 0xFF} # 12 0x0b 0b1011 cyan cf12
12 {0xFF 0x00 0x00} # 13 0x0c 0b1100 red cf13
13 {0xFF 0x33 0xFF} # 14 0x0d 0b1101 magenta cf14
14 {0xFF 0xFF 0x00} # 15 0x0e 0b1110 yellow cf15
15 {0xff 0xff 0xff} # 16 0x0f 0b1111 white cf16
}
# Build RGB RTF colour table header #
option expression off
set ct "{\\colortbl ;"
for {set i 0} {< $i 16} {++ i} {
append ct "\\red[lindex $rgb_colour($i) 0]"
append ct "\\green[lindex $rgb_colour($i) 1]"
append ct "\\blue[lindex $rgb_colour($i) 2];"
}
append ct "}"
Example files are supplied, man_to_rtf.tcl and man_to_html.tcl
Example:
lib ticol_highlight
set s "<font color=\"blue\">"
puts [highlight [unescape $s] -html]
See also: Plugins, lib, rtf
Various HTML manipulation commands
To load, use: lib ticol_html ?-verbose? ?-nocomplain?
See separate help topics for:
html chomp_tag Parse an HTML file
html header Return the entire HTTP header
html header_field Return an HTTP header field
html remove_header Strip off an HTTP/HTML header
html strip Remove all HTML tags from a string
See also: plugins, http
html chomp_tag (Plugin) |
Index |
To load, use: lib ticol_html ?-verbose? ?-nocomplain?
string [html chomp_tag varName ?varOut ?-nobrackets?]
Remove the next HTML tag from a variable, return it and remove that
part of the string (including <> brackets)
An optional array or simple variable (varOut) can receive any pre-tag
prefix text, thus enabling combined HTML and plaintext to be parsed.
Optionally the enclosing brackets can be omitted from any return
Example:
lib ticol_html
set s [http http://localhost:8800]
puts [html chomp_tag s]
puts [html chomp_tag s]
puts [html chomp_tag s]
Result:
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
<html>
<head>
Example:
lib ticol_html
set s " <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">\
<html><head><title>Hello world</title></head><body>\
Hello World
</body></html>"
puts [html chomp_tag s] # Tag:doctype
html chomp_tag s # Tag:html
html chomp_tag s # Tag:head
html chomp_tag s # Tag:title
html chomp_tag s t # Tag:/title. Return the title text value in 't'
puts $t
Example:
lib ticol_html
option expression off
set s [http http://localhost:8800]
do {
set t [html chomp_tag s]
puts $t
} while {ne $t ""}
Results:
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
<html>
<head>
<title>
</title>
</head>
<body>
...
See also: plugins, html strip, html header_field, html remove_header
html header (Plugin) |
Index |
To load, use: lib ticol_html ?-verbose? ?-nocomplain?
string [html header string]
string [html header varName -var]
Returns the entire header from an HTTP 1.1 block of HTML code
Example:
set s [http http://localhost:8800]
puts [html header $s]
Result:
Header: HTTP/1.1 200 OK
Date: Mon, 31 Jan 2022 17:01:42 GMT
Server: Apache
Last-Modified: Thu, 25 Feb 2021 01:58:24 GMT
ETag: "40f48-a5e-7cb6c213"
Accept-Ranges: bytes
Content-Length: 2654
Vary: Accept-Encoding
Content-Type: text/html
See also: html, html header_field
html header_field (Plugin) |
Index |
To load, use: lib ticol_html ?-verbose? ?-nocomplain?
string [html header_field string field-name]
string [html header_field varName field-name -var]
Return a header field from an HTTP 1.1 block of HTML code. Case is ignored
during comparison and trailing colons may be omitted for the field name.
Whitespace is trimmed from the return string
The -var argument allows a variable name to be passed instead of a string
literal
If either the string argument or field-name is null then an empty string
is returned
Example:
# Emulate an HTTP header
set x "HTTP/1.1 200 OK\r\n\
Date: Fri, 12 Jun 2017 13:26:15 GMT\r\n\
Server: Apache/2.4.10 (Ubuntu)\r\n\
X-Powered-By: PHP/5.5.12-2ubuntu4.3\r\n\
Redirect: http://localhost.com/index.php\r\n\r\n"
puts "Date: \[[html header_field $x Date:]\]"
puts "Redirect: \[[html header_field $x rEdIrEcT:]\]"
puts "Server: \[[html header_field x server -var]\]"
puts "Foo: \[[html header_field $x foo]\]"
Result:
Date: [Fri, 12 Jun 2017 13:26:15 GMT]
Redirect: [http://localhost.com]
Server: [Apache/2.4.10 (Ubuntu)]
Foo []
See also: plugins, html strip, html header_field, html remove_header
html remove_header (Plugin) |
Index |
To load, use: lib ticol_html ?-verbose? ?-nocomplain?
string [html remove_header string]
Servers may return a 'header' section as an HTML page prefix. This
command offers an easy way to remove this
[html remove_header] strips off and removes the HTTP header from an
HTML body return string stripto may be used to remove other junk up to
the first "<" bracket
This command is also useful for removing any HTML prefix from JSON data
which is prefixed by a header. See: json_to_list
See also: plugins, http, html strip, html header_field, html chomp_tag
html strip (Plugin) |
Index |
To load, use: lib ticol_html ?-verbose? ?-nocomplain?
string [html strip string]
Removes all HTML tags from a string
tags are replaced with a single space
Example:
set s "<b><em>Hello world</em></b>"
puts [html strip s]
Result
Hello world
See also: plugins, html remove_header, html header_field, html chomp_tag
http ?http://?domain?:port??url? ?port | -? ?var? ?-u username?
?-p password??-1? ?-nocomplain? ?-head? ?-t N?
Retrieve a page of HTML text from a remote website via a HTTP 1.1 request
HTTP is supported on any port but HTTPS and FTP etc. are unsupported
For FTP, see ticol_ftp
This is currently a minimal HTTP 1.1 implementation with automatic chunked
download support which should be sufficient for rudimentary data download
from a server.
HTTP 1.0 using the ?-1? argument is available, otherwise HTTP 1.1 is used
Binary downloads are currently unsupported but you can easily [spawn]
external programs such as wget.exe for binary content
The default port value is 80. This may be specified either in the URL or via
the port argument. Extra headers cannot be specified.
A URL which contains whitespace must be enclosed in braces or double-
quotes
Optional Arguments:
-1 Use HTTP 1.0 instead of HTTP 1.1
-u Specify a username for BASIC Authentication (plaintext)
-p Specify a password for BASIC Authentication (plaintext)
-t N Specify an optional connect timeout in seconds
-head Issue an HTTP HEAD command instead of HTTP GET
-nocomplain Don't issue verbose output on errors
Positional Arguments:
Example:
http "http://en.wikipedia.org/wiki/Main_Page" - x
Result:
puts "HTML is: $x"
Example:
puts [http http://localhost:8800/cgi-bin/ticol.exe?cgi.tcl]
Result:
<displays Tcl script page contents>
See also: urlencode, urldecode, html remove_header, html header_field
int, double (Type Casting) |
Index |
Type-casting operators:
integer [int variable|number]
double [double variable|number]
Converts a numeric character to the requested type, by truncation of
necessary. Input can either be a number or a variable name
Bear in mind that Tcl is a "typeless" language where "everything is a
string" However strings still have an apparent type, for example the
string:
3.141592653589792
has "real" number characteristics, whereas ...
3
does not. Consequently one can transpose real numbers to integers. Also,
Ticol assumes type and will assume integer values in some cases where this
is undesired. This may affect certain math operations such as division
forcing a truncated integer result. You can ensure a real number (double)
operation by casting integer values to double
Example:
expr 22/7
Result:
3
Examples:
puts [expr 22/double(7)] # Expression syntax
puts [/ 22 [funct double 7]] # Using [funct] to call
puts [/ 22 [double 7]] # [double] direct command
puts [/ 22 7.0] # Command syntax with double value
Result:
3.142857142857143
[int] and [double] are available as both expression functions and Ticol
commands
See also: vars, type
bool [is_admin]
Return $::true (1) if the current Windows user is an administrator
See also: true, false, elevate, is_elevated, is_*
bool [is_command command|procname]
Returns either $::true (1) or $::false (0) indicating whether a command or
procedure is defined. [is_defined] may also be used to test for definition
Example Result
is_command puts 1
is_command put 1
is_command puttt 0
See also: is_defined, commands, if, while. for. foreach, flow control, is_*
true, false
bool [is_const varName]
Returns a boolean $::true (1) or $::false (0) indicating whether a
variable is a constant
Example:
set sid Hello
enum {
fred
}
puts "Is sid const? [bool [is_const sid]]"
puts "Is fred const? [bool [is_const fred]]"
Results:
False
True
See also: bool, true, false, is_pointer, const, set, is_*
bool [is_elevated]
Returns $::true (1) if Ticol is running with elevated privileges
(Windows 7+) or $::false (0) if not
[elevate] will self-call, passing the original command-line arguments
Example:
if {! [is_elevated]} {
elevate
} else {
# Run the script
}
See also: true, false, elevate, is_admin, is_*
bool [is_empty varname]
Test if a variable is both set and empty. If a variable of either array or
simple type is both set and empty it returns 1. If the variable does not
exist 1 is also returned.
Be careful to pass the name of the variable rather than its value.
Empty strings also return 1. [is_set] may be used in conjunction with
[is_empty].
Use [is_null] to test strings for 'emptiness'
Type Empty Not-empty
---------------------------------
unset var 1 1
set var 1 0
string literal 1 0
---------------------------------
Example:
set s {}
if {is_empty s} {
puts "Empty"
} else {
puts "Not empty"
}
puts [is_empty {}]
puts [is_empty nosuchvar]
puts [is_set nosuchvar]
Result:
Empty
1
1
0
See also: is_set, is_null, is_*
bool [is_dir path]
Tests if the given path string is a path or not, returning a boolean
$::true (1) or $::false (0)
See also: true, false, mkdir, mkdirs, unlink, file, is_*
bool [is_pointer varName]
Returns a boolean $::true (1) or $::false (0) indicating whether a
variable is a pointer variable or not
See also: true, false, is_const, setb, addressofb, ofaddressb, struct,
is_*
bool [is_proc name]
Test if a command is a user-defined procedure. Returns a boolean
$::true (1) or $::false (0)
Example:
if {is_proc foo} {
undef foo
}
See also: true, false, proc, undef, is_*
bool [is_null string]
Test if a string is null (empty, "" or {})
Use [is_empty] to test a variable name and [is_set] to test if a variable's
status
Example:
puts [is_null ""]
puts [is_null {}]
puts [is_null Hello]
puts [is_null $qbf]
Results:
1
1
0
0
See also: is_set, is_empty
bool [is_mod value divisor ?{command}?]
Fast, division-safe modular division check. If the divisor is zero then
$false (0) is returned. This avoids extra code required to prevent division
by zero
In loops this can lead to improved performance. A boolean $::true (1) or
$::false (0) is returned. Performance may be enhanced by avoiding [if]
and executing an optional command to evaluate if [is_mod] is true
[is_mod] replaces the longhand expression:
if {[== [% $x 1000] 0]} {call_proc}
with:
if {[is_mod $x 5000]} {call_proc}
Calculations are performed on 64-bit integer values
Example:
[is_mod $x 5000 call_proc]
Avoid using this...
if {is_mod $i 1000} {
printf "\r %2.f %% " [* [/ ${i}.0 $size] 100.0]
}
Use this instead, as it is more efficient and avoids calling [if]
is_mod $i 1000 {
printf "\r %2.f %% " [* [/ ${i}.0 $size] 100.0]
}
See also: is_, is_number, math, is_*
bool [is_numeric string ?-comma?]
Tests if a string is composed entirely of the following characters and
returns boolean 1 if so, otherwise 0
+-0123456789,.
This may include comma-formatted numbers where -comma is used
Examples:
puts [is_numeric hello]
puts [is_numeric 10/01/2017]
puts [is_numeric 23:59]
puts [is_numeric 12345]
puts [is_numeric +123.45]
puts [is_numeric -12,345.67]
Results:
0
0
0
1
1
1
See also: is_*
bool [is_set variableName ?-nocase?]
Tests to see if a variable exists (is set). Similar to php's 'isset'
This command is useful for testing potentially unset variables with [if]
You cannot do the following with an array variable because the variable
is resolved before [if] is called ...
if {$_array(nosuch_element)} { # Will raise an unset variable error
Instead use [is_set]:
if {is_set _array(nosuch_element)} {
You should pass a variable name not the dereferenced contents of a variable
[is_set] returns a boolean value (1 or 0)
[is_set] may be used on standard Tcl variables such as string variables,
arrays, lists, stacks, structs. For arrays you may also use [array is_set]
If using option expression on then the 1st argument will not be interpreted
as a command. Thus care should be taken with flow-control statement
arguments
Correct use of [is_set]:
option expression on
if { [is_set argv] } {...}
option expression off
if { is_set argv } {...}
Incorrect:
if { is_set $argv } {...}
if { [is_set $argv] } {...}
Array existence should be tested using [array exists]
See also: is_empty, is_null, array exists, is_admin, is_dir, is_elevated,
set, vars, is_*
string [item varname ?default?]
Safe reference of any variable with optional default value which is returned
if the specified variable is not found. The default default return is ""
[item] avoids fatal errors where a variable does not exist and will,
instead, return a default value without complaint
Any type of variable may be referenced including simple variable, array,
struct member etc. The variable name only should be given and the name
must not be prefixed with a dollar sign
Examples:
# Where simple variable q does not exist ...
puts [item q]
puts [item q "q was not found"]
set q foo
puts [item q]
# Where array item a(1) does not exist ...
puts [item a(1)]
puts [item a(1) "Array element a(1) was not found"]
set a(1) bar
puts [item a(1)]
set q 1
puts [item a($q)]
Results:
""
q was not found
foo
""
Array element a(1) was not found
bar
See also: array item, stack item, struct item
string [join list ?separator?]
Join elements within a space-separated Tcl list and return a string. The
spaces are removed during the process. The list elements may be separated
by an optional separator character
Example:
set s "The quick brown fox"
set r [join $s "-"]
puts $r
Result:
The-quick-brown-fox
[join] can also be used to 'flatten' a list by one level. The outer-
braces are always removed when an argument is passed to a command but
an additional +1 level of braces is removed
Example:
set data {1 {2 3} 4 {5 {6 7} 8}}
puts [join $data]
Result:
1 2 3 4 5 {6 7} 8
See also: split, list
json_to_list (JSON Conversion Plugin) |
Index |
This is a very simple JSON conversion command
To load, use: lib ticol_json?.dll? ?-verbose? ?-nocomplain?
list [json_to_list json_string|var ?-var? ?variable-out?]
Convert a sequence of JSON format data into Tcl list format
Converting JSON using native Tcl code can be quite slow. This function
speeds up conversion significantly compared to a Tcl code implementation
The command was written for a specific data-processing task and may not,
therefore, be feature-complete as far as the JSON specification is
concerned. It contains a subset of the JSON feature set
If -var argument is passed then the first input argument is interpreted as
a variable name. An optional, trailing output variable is supported.
This may be either a simple variable or an array element
Example:
lib ticol_json -verbose
set json { "name":"John", "age":30, "car":null }
set l [json_to_list $json]
puts "Name: [lindex $l 0 1]"
puts "Age: [lindex $l 1 1]"
puts "Car: [lindex $l 2 1]"
lib ticol_json -unload
Results:
Name: John
Age: 30
Car: null
See also: Plugins, list, list_to_array, unwrap
Line Based String Processing Plugin |
Index |
Efficiently process lines from CRLF-delimited text. Allows a block of
CRLF-delimited text to be addressed as conveniently as if it was an array
To load, use: lib ticol_line?.dll? ?-verbose? ?-nocomplain?
integer [line count string|?varname -var?]
string [line <number> string|?varname -var?]
integer [line <number> string|?varname -var? -array name]
[line count] will return a count of CRLF-delimited lines in a string
[line <number>] will return a particular line number from a CRLF-
delimited string
If the -var qualifier argument is appended then the passed string argument
will be assumed to be the name of a variable. This avoids having to make
copies of large strings internally and should improve performance in loops
If '-array name' is appended then an array will be returned containing
the requested lines of text. The command will return the count of lines
returned
Example:
lib ticol_line
set s [readfile foo.txt]
puts [line count $s]
Example:
lib ticol_line
set s [readfile foo.txt]
puts [line 237 $s]
Example:
lib ticol_line
set s [readfile ovid.txt]
puts [line 237 $s 4 -array a]
Result:
# Returns 4 lines from 237 on in array 'a'
See also: plugins
string [K x ?y? ...]
Functional programming K-combinator. Allows chaining of Tcl expressions
Functionally equivalent to:
proc K {a b} {return $a}
It can be used in all situations where you want to deliver a result that is
not the last. i.e. the first item is returned but the last discarded
The K is a capital letter K (not lower-case k)
For instance, reading a file in one go:
proc readfile filename {
set f [open $filename]
set data [read $f]
close $f
return $data
}
can be simplified, without need for the data variable, to:
proc readfile filename {
K [read [set f [open $filename]]] [close $f]
}
Swap two variables, say a and b, without any temporary variables:
set a [K $b [set b $a]]
Another example, popping a stack:
proc pop _stack {
upvar 1 $_stack stack
K [lindex $stack end] [set stack [lrange $stack 0 end-1]]
}
We can also achieve an efficient implementation of lambda expressions in
Tcl using K:
proc lambda {arg body} { K [set n [info $body 0]] [proc $n $arg1 $body] }
For more information see:
http://wiki.tcl.tk/1923/
See also: list functions
Here is an example input procedure for single key input with echo
This could also be implemented as a [sub]
if {! [defined get_key_input]} {
option expression off
proc get_key_input {} {
# Returns a lower-case keypress
while {1} {
set k [getkey -lower]
if {eq $k "y"} {
puts $k
return "-r"
} elseif {eq $k "n"} {
puts $k
return {}
} elseif {eq $k "q"} {
puts $k
return q
}
sleep 20
}
}
}
Example usage
option expression off
set r [get_key_input]
if {eq $r "q"} {
textcolor red
puts "\n * Quit *"
textcolor
stop
}
See also: getkey, input, proc, sub
Here is an example Lambda function from the http://tcl.tk website
When called from a script the lambda function can be called using
puts [$inv x]. This script cannot handle lists of arguments
When called from the Ticol CLI you can use just $inv x - e.g. $inv 5
proc lambda {args body} {
puts "in lambda args:$args body:$body"
set name [info level 0] # Should return lambda x {expr 1./$x}
puts "Name:$name"
commands $name
proc $name $args $body
set name
}
set inv [lambda x {expr 1./$x}] # Calls proc lambda
puts [$inv 5] # Call with $inv 5 from the Ticol CLI
Result:
0.2
The [lmap] command can also be used to construct lambda expressions such
as:
puts [lmap {x {expr {$x * $x}}} {1 2 3 4 5}]
For more on lambda expressions see: lmap and apply
See:
http://wiki.tcl.tk/519
See also: K, apply, lassign, lmap, map
list [lappend listVar string ?string?...]
Append list elements onto a variable or creates a list from a series of
strings. Strings with spaces are wrapped in double quotes. The variable
referenced in the argument will be created if it does not exist
Where the string to be appended contains reserved Tcl characters such as
{}{}$\
these will be escaped in order to print safely
Creating Nested Lists
---------------------
Don't use [lappend] to add enclosing { and } braces to create sub lists as
this will add superfluous spaces and break parsing. Either append sub
items to a sub-list and [lappend] that or use string routines to wrap in {}
Examples
--------
Example:
puts [lappend var [list one two] three]
puts $var
Result:
{one two} three
Example:
set s ""
echo [lappend s "{escaped"] # [echo] is required to see the escapes
Result:
\{escaped
Example:
# Will create a new variable
lappend nosuchvar 1 2 3 # Variable 'nosuchvar' does not exit
puts $nosuchvar
Result:
1 2 3
See also: lindex, list
list [lassign list ?varName ...?]
This command treats the value list as a list and assigns successive
elements from that list to the variables given by the varName arguments
in order. If there are more variable names than list elements, the
remaining variables are set to the empty string.
If there are more list elements than variables, a list of unassigned
elements is returned, otherwise an empty list {} is returned
Example:
lassign {a b c} x y z
Result:
$x =>'a'
$y =>'b'
$z =>'c'
Example:
lassign {a b c} v w x y z
Result:
$v =>'a'
$w =>'b'
$x =>'c'
$y =>'{}'
$z =>'{}'
Example:
lassign {a b c "The quick" "brown fox" "jumps"} x y z
Result
$x =>'a'
$y =>'b'
$z =>'c'
{The quick} {brown fox} jumps
See also: lappend, list, apply, lmap, map, apply, lambda
integer [lcount list]
integer [lcount varname -var]
Returns an integer count of the number of items in a list.
Non-standard Tcl. See [llength]
Using -var and a variable name avoids a time-consuming variable copy
which may improve performance in tight loops or with very large lists
The argument list must be either a variable or, if a literal string,
enclosed in braces or double-quotes
Example:
puts [lcount $qbf] # Count the number of words
puts [llength $qbf] # Count the number of words (standard)
puts [lcount qbf -var] # Count the number of words
puts [lcount {{a b} c}]
Result:
9
9
9
2
Example:
See also: list, lindex, lappend, lassign
integer [ldepth list]
Returns the depth of a list or 0 for an empty or flat list
Procedure definition:
proc ldepth {list} {
if {! [is_list $list]} {
return 0
}
set flatList [join $list]
if {[string equal $flatList $list]} {
return 1
}
return [expr {[ldepth $flatList] + 1}]
}
Example:
set l {}
puts [ldepth $l]
set l a
puts [ldepth $l]
set l {a b}
puts [ldepth $l]
set l {a {b {c d}}}
puts [ldepth $l]
Results:
0
1
1
3
See also: llength, unwrap, list, lappend, lassign, lcount, ldepth, lindex,
list, list_to_array, llength, lmap, lrand, lrange, lreverse,
lsearch, lset, lsort
string [lastchar string]
Returns the last character of a string (or list interpreted as a string)
If the string is empty then an empty string "" is returned
[lastchar] is more efficient than convoluted combined commands such as:
puts [mids $qbf [string length $qbf] 1]
Examples:
puts [mids $qbf [string length $qbf] 1]
puts [lastchar $qbf]
puts [lastchar Hello]
puts [lastchar ""]
Results:
g
g
o
""
See also: string, left, right, mids, mid
string [left <string> n]
string [left string var -var]
string [string left <string> n]
string [string left var n -var]
Return the leftmost n characters of a string
A common task is to look at the first character of a string. The most
effective way to do this is to use [index] rather than [string left]
If the string contains special Tcl characters []{}$ then call [unescape]
on the string before calling [string left]
Example:
puts [string left $qbf 1]
puts [index $qbf 1]
Results:
T
T
See also: index, right, mid, lastchar, setc, setl, setr, strto, strtor
filter
integer [len string]
Return the length of a string
This is a convenient alias for [string length]
Example:
puts [len $qbf]
Result
43
See also: string, string length, lastchar
let (Assignment command) |
Index |
let Lvalue = Rvalue Evaluate Rvalue(s) as an expression then assign
let Lvalue := Rvalue Assign rvalue(s) without evaluating (assignment)
The MPP may simplify/evaluate expressions
Rvalue of [expr] or [command] will still evaluate
If Rvalue is a string then a well-formed Tcl list
will be returned (this may be brace-wrapped)
let Lvalue ++ Increment lvalue(s) - equivalent to [++ var]
let Lvalue -- Increment lvalue(s) - equivalent to [-- var]
let Lvalue += Rvalue Assign to existing value. Rvalue may be expression
Will cast to double before addition
let Lvalue -= Rvalue Assign to existing value. Rvalue may be expression
Will cast to double before subtraction
let variable|array = expression|variable|array|[command] # Eval with expr
let variable|array := list|variable|array|[command] # Assignment only
let variable|array = {variable|array} # Evaluate contents with [expr]
let variable|array := {variable|array} # Assignment of values only
[let] implements a proposed extension to the Tcl language. It acts similarly
to [set] except that it accepts a Tcl expression as the Rvalue
The expression may be another Tcl variable or a complex statement
[let] returns the resulting Lvalue variable's contents if there is only one
lvalue. If several variables are [let] then the last variable's contents
is returned
Whitespace is allowed between individual Rvalue expression arguments where
'=' is used to separate. This will be removed before the result is
evaluated using [expr]
To assign a literal string rather than evaluate it, use the ':=' operator
instead of '='. This also allows the result of a bracketed [eval] operation
or commands to be assigned. For numeric expressions there will be no
difference between the two methods since the numeric result will be
correctly checked by [expr]
Example:
let x y z := "06/Jun/2015" # Note: just assign, don't [eval]
Note that expressions contained within square [] brackets will have already
been evaluated before being presented to [let]. Therefore the result will
be re-evaluated again when using '='. Where this is not desired use direct
assignment ':='
For example let a b c = [info commands] return an invalid result since the
resulting command string will be re-evaluated as an expression. The correct
command would be: let a b c := [info commands]
Example:
set x "Hello world"
let a b c := $x
puts "$a, $b, $c"
Result:
Hello world, Hello world, Hello world,
Example:
let x := [/ 22 7.0] # Call assign instead of expr
Result:
3.142857142857143 # [] are always resolved
Example
let x := [/ 22 7.0]*2 # Call assign instead of expr
Result:
3.142857142857143*2 # [] are always resolved
Example
let x = [/ 22 7.0]*2 # Call expr and assign the result
Result:
6.285714285714286
Example:
let x := "Hello world" # Will return a Tcl list
Result:
{Hello world}
Example:
let x := one "two three" four # Will return a well-formed Tcl list
Result:
one {two three} four
Example:
set q 0;let q += [rnd 1 10]; puts $q
Result (some value between 1 and 10):
2.000000
Let is an efficient way of setting variables to a complex expression which
avoids intermediate steps but will be slower and less-efficient for simple
assignments of const values
When using the '=' operator, variables, including array variables will be
evaluated before the call to let, and the contents of the variable will
then be evaluated by let as an expression, thus if you wish to assign the
value of a variable without interpretation you must wrap the rvalue variable
in braces thus:
Example:
let path = {$env(TEMP)} # Prevent [expr] evaluation, pass as a literal
let path := $env(TEMP) # Alternate method, direct assignment
Result:
x is correctly set to 'c:\temp'
Example:
let path = $env(TEMP)
Result:
x is incorrectly set to 'temp' after evaluation of c:\temp
Expression examples:
Example:
let q = 4 * atan(1)
puts "q is $q"
Result:
q is 3.141592653589792
Example:
let q = [fib 12]
puts "q is $q"
Result:
q is 233
Example:
let x = {rnd(1,5)>2?return Hi:return Lo} # rvalue must be wrapped in {}
Result:
Lo
See: Tcl language extension proposal:
http://wiki.tcl.tk/3880
See also: set, unset, vars
integer [level]
Returns the current callframe level
See also: info, proc
lib (Ticol Plugin Module Extension Loader) |
Index |
Load or reference a library using the following syntax:
result [lib <libraryname>?.dll? ?-option?]
integer [lib dllname ?filter|modulename? ?-list? ?-nocomplain? \
?-verbose? ?-replace?] # Load a library plugin
integer [lib dllname -unload] # Unload a module
string [lib -modules] # List loaded modules
string [lib dllname -commands] # List commands for a
# loaded plugin
string [lib dllname -info] # List information for a
# a loaded plugin
bool [lib -loaded module_name] # Test if a module is loaded
bool [lib module_name -loaded] # Test if a module is loaded
ticol.exe /lib:dllname ?/lib:dllname...? # Load at startup
(For a list of available plugins see: help plugins)
Load a Ticol plugin extension DLL. The DLL must have been written for Ticol
using the Ticol interface standards/ It must provide query interfaces
according to the supplied API definitions. (See the accompanying API
documentation)
If no path is given, [lib] will attempt to load the plugin library from the
current ticol.exe path location. The command returns a count of the number
of modules loaded or unloaded
[lib] will load commands into the current interpreter instance. Once loaded
these cannot be reloaded by [lib] until they have been undefined using
[undef]. [lib -nocomplain] may be used to ignore any reload errors.
No load status will be printed to the console unless the -verbose arg is
used
Conflicting existing commands can be automatically unregistered using the
-replace argument which will cause existing commands to be unlinked
Once loaded, plugin library functions may be used the same as any other
internal Ticol command and variable memory assigned is automatically
garbage-collected
[lib -info] Calling separately with the -info arg will return the version
number as x.xx and the build date in ISO format as a Tcl list. This command
does not load the library
To unload a plugin and remove all registered commands use the -unload
argument. -unload will return a count of modules successful de-registered
An -unload call to a lib which is not loaded will raise an error unless
the -nocomplain option is used
Unloading a [lib] will destroy any aliases to exported [lib] commands
A library may be loaded at startup using: ticol.exe /lib:dllname
Any number of modules may be loaded at startup
[lib -modules] will return a Tcl list of all registered Ticol module DLLs
Examples:
lib ticol_zip # Load in silent mode
lib ticol_mail.dll -verbose # Load in verbose mode
lib ticol.dll d* -list # Load and return modules as a list
lib ticol.dll -unload # Unload ticol.dll
lib ticol_bignum.dll -replace # Reload
puts [lib ticol_ping -info] # Returns "1.02 20160331"
puts [lib -modules] # "ticol_bignum ticol_zip ticol_reg"
To unload all loaded plugins use the following code snippet
foreach x [lib -modules] {lib $x -unload}
An detailed audit of all modules can be performed using the following code
set file [escape [ls "ticol_*.dll" -list]]
set l ""
foreach x $file {
set cmds [lib $x -list] # Load with option '-list'
set info [lib $x -info] # Module info
textcolor yellow
puts $x # Lib filename
textcolor
puts "\t$info"
puts "\t$cmds"
lib $x -unload # Unload the module
}
newline
puts "[lcount $file] module(s)"
Example:
# Find the first loaded module and confirm it is loaded
# Yes, this is not particularly useful!
lib ticol_cpu
puts [bool [lib [lindex [lib -modules 0]] -loaded]]
lib ticol_cpu -unload
puts [bool [lib [lindex [lib -modules 0]] -loaded]]
Result:
true
false
See: plugins for a list of available Ticol plugin modules
See also: plugins, lib interface, undef, commands
lib interface specifications |
Index |
Plug-in DLLs can be produced for Ticol which adhere to the following
specifications:
All functions must be exported as __cdecl
The following exports must be offered:
long __cdecl _get_dll_version()
{
return _DLL_VER // C++ macro definition
}
LPCSTR __cdecl _get_dll_date()
{
return _DLL_DATE; // C++ macro definition
}
LPCSTR* __cdecl _get_function_list()
{
return _dll_function; // Returns a C++ array
}
LPCSTR* __cdecl _get_command_list()
{
return _dll_command; // Returns a C++ array
}
long __cdecl _get_export_count()
{
return _export_count; // Defined from arrays
}
Interface functions (as defined above) must be exported by name as __cdecl
with the following Ticol command signature:
int (__cdecl* CmdFunc)(struct Interp* i, int argc, char** argv, void* pd)
Your functions will use i as the interpreter instance and use the argc/argv
arguments to receive arguments. These arguments may be altered but not
deleted or reassigned. Full details of available library functions are given
in the plugin documentation. For more information see the example C++ source
code
ticol_init(); // Call before calling any Ticol library functions
Only locally-allocated memory may be freed. The C++ lib header has
prototypes for a number of functions exported back from Ticol to the DLL
for memory housekeeping etc.
Commands are specified as case-sensitive strings which must not contain spaces
Example skeleton stub code:
#define _DLL_VER 100 // Required (v1.00)
#define _DLL_DATE "20211221" // Required (21/Dec/2021)
#define _DLL_COMMENT "A Demo Plugin" // Required (Caption)
LPCSTR _dll_command[]= {"mytest" }; // Registered command name
LPCSTR _dll_function[]= {"CommandTest"}; // The C++ function here
long _export_count=howbig(_dll_command); // howbig() is a C++ macro
long __cdecl lib_cleanup(long) // Any cleanup goes in here
{
// Empty (no cleanup required)
}
int __cdecl CommandTest(struct Interp *i, int argc, char **argv, void *pd)
{
ticol_init(); // MUST be called
print_error("%s: Command not implemented\n",argv[0]);
_ticol_set_result(i,"0"); // Return "0" to Tcl
return TICOL_OK; // Return code SUCCESS
}
The associated project DEF file...
LIBRARY ticol_my.dll
DESCRIPTION "Ticol Demo Plugin"
EXPORTS
CommandTest
See also: plugins, commands, faq, calldll
string [lindex list index ?index|end-N?... -blanks]
Index an item from a list. Multiple index selectors may be given, in which
case, each will be selected in turn from left to right. Brace binding is
honoured and quoted strings are treated as distinct objects outside of
braces. Index values are array base indexed at zero
Items are separated by whitespace which is removed unless braced
Indexing very large lists (megabytes and above) can be quite slow. Arrays
should be used in preference
For single indices, the mnemonic, "end" may be used either with or without
an adjustment value, e.g. end-4
If -blanks is used then empty items will be returned
To return a range of elements from a list, see [lrange]
Example:
set l [list colour "icon cornflower blue"]
puts $l
Result:
colour {icon cornflower blue}
Example:
puts [lindex $l 0]
Result:
icon
Example:
puts [lindex $l 1 0]
Result:
icon
Example:
puts [lindex $l 1 1]
Result:
cornflower
See also: list, lrange, llength, lappend
Line Continuation Character |
Index |
A backslash character can be appended to the end of a Tcl source file line
This will cause the line to continue onto the next and the backslash to be
ignored.
Care should be taken to include formatting spaces when breaking up a string
The next line should start at the beginning of the line and not be indented
within the source-code unless you wish this to be included in the resulting
string
The line-continuation character must not be followed by any other character
or comment such as # on that line or it will be interpreted as a literal
backslash rather than end of line marker
Ticol also handles multi-line strings within a source file.
Example: (First char of each next line is will become a space)
.........1.........2.........3...
puts "The quick brown \
fox jumps over \
the lazy dog"
Result:
The quick brown fox jumps over the lazy dog
Example: (Multi-line string. The text format is retained)
.........1.........2.........3...
puts "The quick brown
fox jumps
over
the lazy
dog"
Result:
.........1.........2.........3...
The quick brown
fox jumps
over
the lazy
dog
Example: (Results in extra spaces)
.........1.........2.........3...
puts "The quick brown \
fox jumps over \
the lazy dog"
Result:
The quick brown fox jumps over the lazy dog
Example:
.........1.........2.........3...
puts \
"Hello world"
Result:
Hello world
Where if statements are used with [option expression on] and the expression
is on multiple lines then line break continuation characters must be used,
e.g.
set r [rnd 1 10]
if { $r \
> \
5
} {
puts Yes
} else {
puts No
}
Numbers may be declared using line-continuation providing they are
wrapped in double quotes, thus:
set n "12345\
6789"
See also: escape sequences, macro, #ifdef, #define
Command Line Interpreter (CLI) Line Editing |
Index |
The CLI implements standard 'insert-mode' line-editing with the following
keystrokes. The command history is reloaded on next running Ticol
UP CURSOR Previous command in history
DOWN CURSOR Next command in history
HOME Move cursor to the start of the line
END Move cursor to the end of the line
ENTER/RETURN Submit the edited line to the interpreter (command)
BACKSPACE DELETE Delete backwards (right to left)
DEL Delete forwards (text to right of cursor is deleted)
CTRL+C Cancel the current edit and start a new line
CTRL+M Same as ENTER/RETURN
CTRL+S Halt the interpreter
CTRL+BREAK Immediately quit the interpreter (even while running)
The command [exit] will exit the CLI whereas [stop] will not
See also: command line history, CLI
Absolute line numbers exist only in an original Ticol source script. In Tcl
there is no concept of absolute line numbers within operational code since
code may be broken up into many separate fragments (e.g. procs)
The Macro PreProcessor (MPP) will completely remove comment lines and thus
the original line numbers cannot be traced. Also, procedures use copied
segments of code which have no absolute line numbering.
Line numbers referring back to the original source code can be matched
during debugging by the use of a variable set within the source code at
strategic points using the #__LINE__ macro
Example:
set line #__LINE__
if {! [defined check] } { # [assert] is the native alternative
proc check {x y} {
if {!= $x $y} {
puts "Error in test at after line $::line"
puts "$x != $y"
stop
}
}
}
set line #__LINE__
check [expr "abs(-100)"] 100 # Should pass
set line #__LINE__
check [expr "abs(-100)"] 101 # Should fail
Result:
Error in test at after line 15
Example:
The [assert] command using #__LINE__ to add a trace output
assert [/ 22 7.0] {> $_ 3} -v -line #__LINE__
See also: assert, line continuation character, debugging, performance
integer [llength list]
Return the count of items in a Tcl list or 0 if the list is empty
llength takes a dereferenced list literal as argument, not a variable name
Example:
set a [list one two three]
puts [llength $a]
Result:
3
See also: list, lappend
string [link_to_path filename]
Resolve a Windows LNK shortcut file to the target path for that link file
The result is an unescaped string. [escape] may be required to use the
return with other commands
Example: (This uses an external app: makelink.exe to create the LNK file)
makelink.exe ticol.exe .
puts [link_to_path ticol.exe.lnk]
Result:
C:\Tcl\ticol.exe
See also: info file, escape
To load use: lib ticol_list ?-verbose? ?-nocomplain?
handle [linkedlist create]
bool [linkedlist add handle lvalue ?rvalue?]
string [linkedlist get handle lvalue]
string [linkedlist item handle integer]
code [linkedlist foreach $handle lvalue ?rvalue? {command}]
bool [linkedlist in_list handle key]
bool [linkedlist insert handle key lvalue ?rvalue?]
long [linkedlist count handle]
void [linkedlist walk handle]
bool [linkedlist unset handle key]
bool [linkedlist unset handle]
[linkedlist] offers standard low-level, linked lists. These are not the
same as Ticol built-in Tcl lists (which are represented as ordered strings)
[linkedlist] enables a programmer to create rapid, robust linked lists in
the same manner as offered by many other high-level programming languages
External data structures such as [linkedlist] require a valid handle
variable to access
A list has a start (often called a 'head') and an end (often called a
'tail'). Items are usually added to the start of the list and the 'tail'
value represents the most recently added item. Lists are iterated by
starting at the 'start' node and reading each successive node in the chain
until the 'tail' or last node is met
the quick brown
+------+ +------+ +------+ +-----------+
| head |->|node 1|->|node 2|->|tail (NULL)|
+------+ +------+ +------+ +-----------+
|lvalue| |lvalue| |lvalue| | N/A |
+------+ +------+ +------+ +-----------+
|rvalue| |rvalue| |rvalue| | N/A |
+------+ +------+ +------+ +-----------+
Each node in the linked list can store two variables. Each will store an
lvalue (left hand value) and may optionally store an rvalue (right hand
value)
[linkedlist create] Creates a new list and returns a handle to the created
list. A list is identified and manipulated by its handle. It is a fatal
error to issue a list command with an invalid, freed or absent handle
[linkedlist add] will add a new item to the linked list. Each successive
new item is added at the head of the list. An "rvalue" is optional
Example:
linkedlist add $handle hello
linkedlist add $handle pi 3.14159265358979323
[linkedlist get] retrieves an 'rvalue' (if any) from a list item. A list
item may either simply be present in the list (as an lvalue) or the
lvalue may have an associated rvalue. (lvalue and rvalue stand for left
hand value and right-hand value respectively)
Example:
# The lvalue is 'pi', the rvalue is '3.14159265358979323'
linkedlist add $handle pi 3.14159265358979323
puts [linkedlist get $handle pi]
[linkedlist item] Allows for random access of the linked list by integer
value. The first item in the list will be '1' (i.e. a base 1 index)
[linkedlist count] should be used to obtain a count of list items
Example:
option expression off
set handle [linkedlist create]
set i 1
foreach {x} $qbf {
linkedlist insert $handle {} $x $i
++ i
}
for {set i 1} {< $i [linkedlist count $handle]} {++ i} {
set lv [linkedlist item $handle $i] # LValue
set rv [linkedlist item $handle $i -rvalue] # RValue
puts "$i:\t$lv\t$rv"
}
Result:
1: lazy 8
2: the 7
3: over 6
4: jumps 5
5: fox 4
6: brown 3
7: quick 2
8: The 1
[linkedlist foreach] allows iteration over the list from start to head,
calling some body of code. An lvalue variable must be supplied. An rvalue
variable is optional. Standard flow-control commands such as [break],
[return] and [stop] are supported and these will return an appropriate
error code (See: error codes, stop, break, return)
Example:
set handle [linkedlist create]
set i 1
foreach {x} $qbf {
linkedlist add $handle $x $i
++ i
}
linkedlist foreach $handle lv rv {
puts "foreach is:\t'$lv'\t'$rv'"
if {[eq $lv fox]} {
puts "* Break *"
break
}
}
Results:
foreach is: 'The' '1'
foreach is: 'quick' '2'
foreach is: 'brown' '3'
foreach is: 'fox' '4'
* Break *
[linkedlist in_list] can be used to test if a given lvalue (key) is present
in a given linked-list structure
Example:
set handle [linkedlist create]
set i 1
foreach {x} $qbf {
linkedlist insert $handle {} $x $i
++ i
}
puts "dog present?: [bool [linkedlist in_list $handle dog]]"
puts "cat present?: [bool [linkedlist in_list $handle cat]]"
Results:
true
false
[linkedlist insert] inserts a new entry AFTER any key value which is found
or at the start of the list if the key value was not found or was supplied
as an empty/null value
Example:
linkedlist insert $handle quick blue
linkedlist unset $handle fox
Repeated calls to [linkedlist insert handle {}] can be used to create a
list in reversed-order from the default. Each successive call will add
the new value to the beginning of the list
Example:
set handle [linkedlist create]
set i 1
foreach {x} $qbf {
linkedlist insert $handle {} $x $i
++ i
}
linkedlist walk $handle
Results:
List 0x2d60cb0 0 item(s)
List::walk 0x2d60f70 VT_001e [dog]=>'9'
List::walk 0x2d60f20 VT_001e [lazy]=>'8'
List::walk 0x2d60ed0 VT_001e [the]=>'7'
List::walk 0x2d60e80 VT_001e [over]=>'6'
List::walk 0x2d60e30 VT_001e [jumps]=>'5'
List::walk 0x2d60de0 VT_001e [fox]=>'4'
List::walk 0x2d60d90 VT_001e [brown]=>'3'
List::walk 0x2d60d40 VT_001e [quick]=>'2'
List::walk 0x2d60cf0 VT_001e [The]=>'1'
[linkedlist count] will return a count of the number of items in a list
[linkedlist walk] will perform a debug 'walk' of the linked list, starting
at the beginning and iterating all nodes until the head is met
Cleanup
-------
It is not recommended to [return] from a linked script with ticol_list
loaded and objects created unless you first clean-up using
[lib ticol_list -unload] and/or specifically call [linkedlist unset]
[linkedlist unset $handle] Destroys the linked list and renders the
handle invalid. Attempts to use a handle after a call to unset will raise
an exception. [linkedlist] objects are not subject to automatic garbage-
collection on proc-exit
[lib unload] destroys all created lists and reclaims their memory
Example:
set handle [linkedlist create]
puts "Handle to list is $handle"
set i 0
foreach {x} $qbf {
linkedlist add $handle $x $i
++ i
}
linkedlist insert $handle jumps "rapidly" # Insert after jumps
linkedlist insert $handle {} *start* # Insert at start
linkedlist insert $handle fox dog # Add dog after fox
linkedlist unset $handle fox # Remove fox
linkedlist walk $handle
linkedlist unset $handle
Results:
List 0x2d20fe0 10 item(s)
List::walk 0x2d20cd0 VT_001e [The]=>'0'
List::walk 0x2d20d20 VT_001e [quick]=>'1'
List::walk 0x2d20d70 VT_001e [brown]=>'2'
List::walk 0x2d21150 VT_001e [dog]=>'10'
List::walk 0x2d20e10 VT_001e [jumps]=>'4'
List::walk 0x2d21100 VT_001e [rapidly]=>'9'
List::walk 0x2d20e60 VT_001e [over]=>'5'
List::walk 0x2d20eb0 VT_001e [the]=>'6'
List::walk 0x2d20f00 VT_001e [lazy]=>'7'
List::walk 0x2d20f50 VT_001e [dog]=>'8'
See also: list, array, stack, struct, error codes
list {vars...}
Return a Tcl list from a series of variables. List creates an atomic
parameter which can be supplied/chained into other parameters
Example:
call [dict create d colours [list orange red blue green purple]]
Ticol supports a range of standard Tcl list commands
Note that a plugin also offers standard linked lists: (See: linked list)
See: list commands for a full list of Tcl list commands
See also: lappend, lreverse, lcount, lindex, llength, lassign,
list_to_array, array_to_list, string, linked list, apply
List Processing Commands |
Index |
Ticol lists are simple formatted strings, and are quite inefficient.
List processing operations should be kept to relatively short lists
For greater efficiency, use arrays or stacks
See the following separate help topics:
array_to_list Convert an array to a list
in Test if an item is in a list
lappend Append an item to a list
lassign Anonymous functions. Apply a command to a list
lcount Return the count of elements in a list
ldepth Test the depth of a list (proc)
lindex Return an item from a list by integer index
list Return a well-formed Tcl list
list_to_array Convert a list to an array
llength Return the length of a list
lmap Iteratively assign elements of lists to variables
lpop Pop an item from head of a list, optionally splitting the list
lpush Push item to head of a list
lrand Return a random element from a list
lrange Return a range of items from a list
lreverse Reverse the order of a list
lsearch Search a list for an item
lset Set a list element
lsort Sort a list
ltail Returns the tail of a list held in a strong constant
ni Test if an item is not in a list
See also: Tcl lists
integer [list_to_array <tcl-list> arrayname ?array-base?]
Converts a well-formed Tcl list into an array indexed by integer from
element base 0.
Example:
set r [list_to_array $qbf out]
puts "list_to_array returned $r item(s)"
set s [array size out]
for {set i 0} {[< $i $s]} {incr i} {
printf "%i:\t%-40.40s\r\n" $i $out($i)
}
Result:
list_to_array returned 9 item(s)
0: The
1: quick
2: brown
3: fox
4: jumps
5: over
6: the
7: lazy
8: dog
See also: list
lpush, lpop, ltail, lpeek |
Index |
list [lpush list value]
void [lpush variable value -var]
list [lpop list ?-n X?]
list [lpop variable -var ?-n X?]
list [ltail list]
list [lpeek list]
list [lpeek variable -var]
These commands offer 'stack-like' push and pop operations which will
operate either on variables or lists held in strings.
Where lists are handled as strings the [lpush] and [lpop] oeprations
may be 'chained' together. This is not recommended when using variables
via the -var argument.
[lpop] will return a list containing one item. This may be a a string of
one or more words, depending whether it was braced or not within the
input list.
When using string lists, [lpop] simply returns the head of the list and
leaves the input list unchanged. i.e. the input string is a constant
The default behaviour for [lpop] is to pop on item from the list, this
can be changed using '-n' to specify how many items to return.
[lpush] will return an updated list when used with a string based list
input. When used with a variable, the command returns nothing and
the updated list appears within the given variable.
[ltail] is provided so that lists held in string constnats can be
split in conjunction with [lpop]. [ltail] operations may be chained.
[lpeek] will return the head item from a list without removing it
Example:
# List push sing string literals
puts [lpush [lpush "" Hello] world]
# List push using a variable
set s ""
lpush s Hello -var
lpush s world -var
puts $s
# Decompose the list into head and tail using a variable
set head [lpop s -var]
puts $head
puts $s
# Build multi-word list items...
set s ""
lpush s {Now is the winter} -var
lpush s {of our discontent} -var
puts $s
# Return the tail of a list held in a string constant
puts [ltail $qbf]
Results:
Hello world
Hello world
Hello
world
{Now is the winter} {of our discontent}
quick brown fox jumps over the lazy dog
See also: list commands, list, lindex, lmap, lreverse
string [lrand <list> ?-norepeat?]
Returns a random element from a well-formed Tcl list
If -norepeat is used then the same indexed element will not be returned
more than once in succession. i.e. unless the list comprises a single
element then a different element will be returned at each call.
[lrand] is concise and avoids the use of a complex and syntactically
error-prone compound statements
Example:
puts [lrand $qbf]
Result:
fox # (probably)
See also: list, lrange, rnd, rndseq, randomstr
list [lrange <list> start finish|end|end-N]
[lrange] returns a range of list elements from the given list.
The index values are 'base 0' values where 0 is the first element
The end value may be specified using either 'end' for the last item
or 'end-N' for the Nth item before the last
To select single elements from a list or sub-list see [lindex]
Example:
puts [lrange $qbf 1 3] # 2nd to 4th items
Results:
quick brown fox
See also: list, lindex, list commands, lpush, lpop
list [lremove list item ?-nocase? ?-all?]
list [lremove varname item -var ?-nocase? ?-all?]
Removes one or more instance(s) of a single item from a well-formed Tcl
list. The list may be passed directly via a literal or dereferenced
variable or passed as a variable name using the '-var' option
Only complete list items may be removed. [lremove] does not remove
substring values and is not a string-processing command
Unless '-all' is used, only the first matching reference to 'item' will
be removed. If '-all' is used then all matching instances will be removed
The default is to perform a case-sensitive match. The '-nocase' argument
may be used to perform case-insensitive matching
Example:
set s $qbf
puts [lremove $s fox]
puts [lremove $s the -nocase]
Result:
The quick brown jumps over the lazy dog
quick brown fox jumps over lazy dog
A simple verison of [lremove] can be implemeted using [lreplace] as
follows ...
proc lremove {listVariable value {case {-exact}}} {
upvar 1 $listVariable var
set idx [lsearch $case $var $value]
set var [lreplace $var $idx $idx]
}
See also: lremove (proc), lreplace, list, lpush, lpop
list [lreplace list <start> <end> ??list-item? ?list-item?...?]
Replace items within a list over a range where <start> and <end> are
integer index values and where 0 refers to the first item of the list
(base 0 index)
[lreplace] returns a new list with zero or more items replaced by those
specified in the argument(s). The supplied list argument is not modified
It is important to note that <start> and <end> refer to a range
Successive optional elements are inserted into the specified range
Example:
puts [lreplace {a b c d e} 1 1 foo]
puts [lreplace {a b c d e} 1 3 foo]
puts [lreplace {a b c d e} 1 1 foo bar moo par]
puts [lreplace {a b c d e} 1 3 foo bar moo par]
Result
a foo c d e
a foo e
a foo bar moo par c d e
a foo bar moo par e
See also: lremove, list, lappend, lassign, lmap, lreverse, lsearch, lrange,
lpush, lpop
Removes one unique item from a well-formed Tcl list
[lremove] can be implemented as the following proc and which includes the
ability to use '-nocase' to perform a case-insensitive replacement
proc lremove {listVariable value {case {-exact}}} {
upvar 1 $listVariable var
set idx [lsearch $case $var $value]
set var [lreplace $var $idx $idx]
}
Note that [lremove] does not yet support negative index values. The 'value'
argument should be tested to ensure that it is non NULL
A proc would be required to remove all instances of a particular item. This
could be achieved by looping and testing the [lsearch] return as follows...
proc lremove_all {listVariable value {case {-exact}}} {
upvar 1 $listVariable var
while {>= [set idx [lsearch $case $var $value]] 0} {
set var [lreplace $var $idx $idx]
}
return $var
}
The functionality of both could be combined with an '-all' argument
See also: lsearch, lreplace, list, lpush, lpop
lreverse tcl-list
Reverse a Tcl list. The list must be a well-formed Tcl list or simple
space/tab-separated string
Example:
puts [lreverse $qbf]
Result:
dog lazy the over jumps fox brown quick The
Example:
set q "Now is the {time for all good men} to come to the aid of \
{the party}"
puts [lreverse $q]
Result:
{the party} of aid the to come to {time for all good men} the is the is Now
See also: list, lindex, lpush, lpop
A Tcl list is a sequence of characters, optionally separated by whitespace
and grouped, where necessary by braces. To all intents and purposes a Tcl
list is a human-readable string. Lists get complex when there are nested
lists and nesting where braces protect space-separated lists. Sub-lists
are likewise braced
"This is a list"
Placing this as a sub-list we get:
"This is a list {This is a list}"
As with Lisp, complex "tree-like" hierarchies can be represented using
lists
See also: list, lindex, lassign
load (CLI Command - Loads a file into CLI interpreter memory) |
Index |
# Unencrypted file
load filename?.tcl? ?targetVarName? ?-password password?
# Encrypted file
load filename.tcx ?-password password?
# Clear a loaded file
load -clear
Load to the Console and Run
---------------------------
Loads a Tcl file from the Command Line Interpreter (CLI) shell
[load] with no arguments will display information about a loaded script
Use [run] to run the file once loaded
Use [dump] to display non-encrypted (obfuscated) loaded source code to the
console and optionally to a Tcl variable
Use [load -clear] to erase the script buffer (See also: clear)
The [load] command does not clear the user-level variable table or non-
system level consts (See: run, clear)
[load] with a .tcx file and, if necessary a password via -password will
load and decrypt a protected file.
Load to a Variable
------------------
Optionally, a file may be loaded to a variable instead of the CLI. In this
case any loaded program will be cleared and the preprocessed source loaded
to the given variable. If the variable exists it will be overwritten. You
can then run this source using [source $varname]. The source is not
returned or echoed by the [load] command
Example:
load test.tcl code # Load 'test.tcl to variable 'code'
> Loaded test.tcl OK 1420 byte(s) # CLI acknowledgement
eval $code # Runs the script at the CLI
See also: run, dump, exit, clear, source, clearprocs, call
logo fore ?back? ?x? ?y?
Draw the Ticol logo
See also: info, console, screen, puts
lmap {lambda-expression} list
Iteratively assigns elements of one or more lists to variables,
executes the lambda-expression argument, and collects the results into
a new list by calling [apply]. The lambda-expression will be an expression
such as:
{x {expr {$x * $x}}} # Given x return x * x
Which is equivalent to f(x) -> (x * x)
[lmap] applied to an empty list will return an empty list
[lmap] is defined as a native command in Ticol. This can also be emulated
using the following Tcl code and using [apply]. The native command is more
efficient, performance-wise
proc lmap {lambdaExpression list} {
set result {}
foreach item $list {
lappend result [apply $lambdaExpression $item]
}
return $result
}
Example:
puts [lmap {x {expr {$x * $x}}} {1 2 3 4 5}]
Result:
1 4 9 16 25
Example:
puts [lmap {x {return [string length $x]:$x}} {a bb ccc dddd}]
Result:
1:a 2:bb 3:ccc 4:dddd
Example:
lmap {x {expr {$x**2 + 3*$x - 2}}} {-4 -3 -2 -1 0 1 2 3 4}
Result:
2 -2 -4 -4 -2 2 8 16 26
See:
https://wiki.tcl-lang.org/13920
http://www.tcl.tk/man/tcl/TclCmd/lmap.htm
See also: lambda, list, apply
loop (Flow Control) |
Index |
loop limit ?increment? {code} # Short form
loop variable start limit ?increment? {code} # Long form
A flexible and efficient [for] loop. [loop] similar in implementation to
the TclX [loop] construct.
A short form simply iterates for 'limit' times using an optional increment
value. This increment may be negative but negative increments will work
only with negative limit values.
loop 10 {puts "Hello"} # Iterates 10 times
loop 10 2 {puts "Hello"} # Iterates 5 times
loop -10 -2 {puts "Hello"} # Iterates 5 times
It is important to understand that the long form of [loop] iterates from
'start' to 'limit-1' and for a total of 'limit' times (depending in any
increment value)
e.g. [loop i 0 5] will iterate 5 times through '0 1 2 3 4'
loop i 0 10 2 {puts "Hello $i"}
is equivalent to:
for {set i 0} {< $i 10} {incr i 2} { # Loop exits when i == 10
puts "Hello $i"
}
[loop] optionally takes 5 argument with an optional 6th argument to
specify an increment. There is no need to brace all but the last argument
[loop] takes a loop variable name as its 2nd argument, which will be
initiated to the value of the 3rd argument. The final 'code' argument
will be evaluated, incrementing the loop variable by 1 unless an increment
value is specified as the optional 5th argument
The increment argument may be positive or negative but not zero. If
argument 3 is higher than argument 4 and no increment argument is given
then an increment of -1 will be used. An increment value of zero will
raise an error.as this would lead to infinite loops
The [loop] variable will always be the same as the limit variable on exit
but the loop will execute for 'limit-start' times. So 'loop i 0 10'
will execute 10 times, looping from 0 to 9 inclusive the same as [for]
if 'start' is equal to 'limit' then the loop code body is never executed
and 'limit' is returned in the loop variable
[loop] supports [continue], [break] and [return] which behave the same as
with [while] and [for] constructs
[loop] uses signed 64-bit integers as its loop arguments and can thus
loop to INT_MAX (9223372036854775807)
Example:
# Automatic use of +1 increment
loop i 0 5 {
puts "i is $i"
}
puts "Final i is $i"
# Automatic use of -1 increment
loop i 5 0 {
puts "i is $i"
}
puts "Final i is $i"
Results:
i is 0
i is 1
i is 2
i is 3
i is 4
Final i is 5
and
i is 5
i is 4
i is 3
i is 2
i is 1
Final i is 0
Example:
option expression off
set i 0
proc foo {} {
upvar i
loop i 0 10 1 {
puts "i is $i"
if {== $i 5} {return $i}
}
}
foo
puts "i is $i"
Result:
i is 5
Example:
option expression off
loop i 0 10 1 {
if {== $i 5} {
# Skip 5 and increment past 6
++ i
continue
}
}
puts "i is $i"
Result:
i is 10
See:
https://osr507doc.xinuos.com/en/man/html.TCL/loop.TCL.html
See also: floop, for, while, do, flow control
floop (floating point loop) |
Index |
floop variable float-start float-limit ?float-increment? {code}
Implements simple loop constructs as with [loop] but using floating point
values. [loop] is 64-bit integer only while [floop], which is slower, allows
double value floating point increments and counters.
See [loop] for more details on structure and control.
Note that there is no short-form variant of [floop] as offered by [loop]
See also loop, for, while, do, flow control
See: ls
See also: file
ls <filespec> ?-r? ?-ad | -af | -all? ?-la? ?-list ?var?? ?-array var?
List files This is a partial implementation
-r Recursive search (includes subfolders)
-ad | -af Attributes: directory or file
-all List all files
-la List all file details
-lc Force lower case path names
-path Return full path (default is filename only)
Paths are returned with Unix style '/' separators
-list ?var? Return in list format (variable not mandatory)
-array var Return the list as an array (variable is mandatory)
var Storage variable is optional (for either -array or -list)
If var is given then a filespec must also be given
Ordering/sorting is not supported. You can sort the result from -list
Perform a Windows-style directory listing. The result can optionally be
fed into a variable and optionally in Tcl list format using the -list arg
If [option echo] is set to OFF then [puts] or [printf] must be used to
display the results
Forward-slash and backslash path separators are supported
CAUTION: The default is not to return a Tcl list. Use -list to return one
CAUTION: Large folders or whole disk searches have the potential to create
huge lists which may exhaust available memory. Lists are slow and
inefficient. Use arrays for potentially large searches
Example:
option echo off
ls *.txt -la -list a # Return list in var 'a'
foreach {q} [lsort $a] { # For each file row entry
foreach {i j k l} $q { # Iterators must match return fields
puts "'$i' '$j' '$k' '$l'"
}
}
Example:
option echo on
ls c:/temp/newdir
ls "c:/temp/newdir/*.tcl"
Example:
option echo on
ls *.tcl -array a
loop i 0 [array size a] {
puts $a($i)
}
Note: where the external Windows DIR command is used this requires backslash
syntax and does not support Unix/Tcl style forward slash path separators
See also: unlink, mkdir, mkdirs, chdir, getcwd
integer [lsearch ?option ...? list pattern]
Where ?options? are implemented as follows:
-glob Default. Use a glob style match v's -exact (see: glob)
-exact Make an exact match (case-sensitive) v's -glob matching
-nocase Make a case-insensitive comparison
-inline Return the value rather than the list index position
-start N Start processing at list index item offset N
An array base 0 value corresponding to [string index]
where 0 is the first item in the list, 1 the 2nd etc...
This offset will apply to the pre-sorted input list
Values <0 are ignored and 0 assumed
-not Invert the search. Return items which don't match 'list'
Typically this will return a well-formed Tcl list
-all Return a list of all items matching the search pattern
of the first item. Output may be modified by -inline
-sort Sort the input list before comparing (non-standard)
-sorted Ignored. Ticol does not optimise sorted list searches
-increasing Ignored. Used with -sort only - otherwise ignored
-decreasing Ignored. Used with sort only - otherwise ignored
-ascii Ignored. Ticol is ANSI-only
[lsearch] matches the pattern argument to the list to find a match. If no
match is found then -1 is returned. Nested lists are not supported.
If all matching style options are omitted then the default matching style is
-glob. If more than one matching style is specified, the last matching style
given takes precedence
Sub-list matching is not yet supported
Examples:
lsearch {ba ze fi eo cu pa le ri} fred # => -1
lsearch {ba ze fi eo cu pa le ri} * # => ba
lsearch -inline {ba ze fi eo cu pa le ri} * # => 0
lsearch -not {ba ze fi eo cu pa le ri} * # => -1
lsearch -all -not {ba ze fi eo cu pa le ri} * # => {}
lsearch -inline -all -not {ba ze fi eo cu pa le ri} * # => {}
lsearch {ba ze fi eo cu pa le ri} ?o # => 3
lsearch -inline {ba ze fi eo cu pa le ri} ?o # => eo
lsearch -inline -all {ba ze fi eo cu pa le ri} ?e # => ze le
lsearch -inline -glob -all {ba ze fi eo cu pa le ri} ?e # => ze le
lsearch -inline -all -not {ba ze fi eo cu pa le ri} ?e # => ba fi eo cu pa ri
lsearch -all -not {ba ze fi eo cu pa le ri} ?e # => 0 2 3 4 5 7
lsearch -glob -all -not {ba ze fi eo cu pa le ri} ?e # => 0 2 3 4 5 7
lsearch -exact -nocase {{a b c} {d e f}} {D E F} # => 1
lsearch -exact -nocase {{a b c} {d e f}} {p q r} # => -1 # Not found/-1
lsearch -exact -sorted -inline -start 4 {b z f e c p l r} {p} # => p
lsearch -exact -sorted -start 4 {b z f e c p l r} {p} # => 2
lsearch -inline -list -exact\
-not -start 0 {a "Hi there" x y z} {x} # => "a {Hi there} y z"
lsearch -exact -inline -all -not {a c b c a b c} c # => a b a b
lsearch -exact -all -not {a c b c a b c} c # => 0 2 4 5
lsearch -exact -inline -all {a c b c a b c} c # => c c c
lsearch -exact -all {a c b c a b c} c # => 1 3 6
lsearch -exact -all -start 2 {a c b c a b c} c # => 3 6
Note that [lsearch] does not 'eval' its arguments...
lsearch -all -inline {$qbf} * # => $qbf (correct)
lsearch -all -inline {list $qbf} * # => list $qbf (correct)
lsearch -all -inline "$qbf" * # => quick brown ...
lsearch -all -nocase -inline $qbf t* # => The the
See also: list, lappend, lindex
list [lset varname list]
list [lset varname {} list]
list [lset varname ?index|end?-N?|{}? replacelist]
Set a list element in a list variable. [lset] takes a string variable
name and assigns the Nth element of that list with a value passed in the
'replacelist' argument. A list is returned
It is permissible for the string variable to be empty or not exist and
the variable will be created if it does not already exist
A replacement list is mandatory even if this is a single empty element
If the variable exists and it is of string type it will be overwritten
If the variable exists and it is not of string type (e.g. struct or stack)
then an error will occur and execution will halt unless [try/catch] is used
The element index list (3rd+ argument) need not be braced
Each list element must be >= 0 and < the number of elements in each
addressed sublist. e.g. if a sublist has 3 elements {a b c} then
valid element addresses range from 0 to 2.
[lset] takes either 3 or 4 or more arguments
Nested lists of unlimited depth (other than memory) are supported
lset x {j k l} # Set x with list value {j k l}
lset x {} {j k l} # Set x with list value {j k l}
lset x 0 j # Set x list element 0 with value j
lset x 2 j # Set x list element 2 with value j
lset x end j # Set x last list element with value j
lset x end-1 j # Set x last but one list element with value j
lset x {2 1} j # Set x nested list element 2,1 with value j
lset x 2 1 j # As above but with unbraced element list
lset x {2 3} j # Set x nested list element 2,3 with value j
..code
Examples:
..code
set x [list [list a b c] [list d e f] [list g h i]]
puts [lset x end-1 j]
set x [list [list a b c] [list d e f] [list g h i]]
puts [lset x 2 1 j]
set x {0 {0 {0 {0 1 2 3 4 5}} 1 2 3} 1 2 3}
puts [lset x {1 1 1 3} 23]
Results:
{a b c} j {g h i}
{a b c} {d e f} {g j i}
0 {0 {0 {0 1 2 23 4 5}} 1 2 3} 1 2 3
See also: list, lappend, lindex, lsearch, lmap
string [setl target-variable source-variable|source-string]
A Procrustean command. Inserts a source string or variable into the
target, left-aligned. [setl] is based on the BASIC command LSET
Any excess length in the source is stripped to fit.
The standard Tcl [setl list] is not yet implemented
Example:
set s [makestr 20 *]
setl s "hello"
puts $s
Result:
hello***************
Example: (Blank out part of a string)
set s "0849 666 616"
setl s " "
puts "'$s'"
Result:
' 666 616'
See also: setc, setr, mid, strstr, string
list [lsort list ?-ascending | -descending? ?-nocase? ?-integer? ?-real?]
Sorts a Tcl list in ascending or descending order, optionally ignoring case,
comparing as integer, or comparing as a double value
lsort does not support -step. -index, -stride etc.
Braced list sub-items are supported
lsort will perform a simple sort on a list. e.g.
foreach x [lsort [array get env]] {puts $x}
Example:
set ls [lsort {
{{d e m o} 34512}
{{c o d e} 54321}
{{v e r y} 13254}
{{b i g} 12345}
{}
} ]
puts "=$ls="
Result:
={{b i g} 12345} {{c o d e} 54321} {{d e m o} 34512} {{v e r y} 13254}=
See also: list
string [makestr length ?char? ?var?]
string [makestr length -- var]
Creates and allocated a string of given length, optionally filled with a
specified character. If more than one fill character is passed then only
the first character of the fill-string will be used. The default fill
character is a space. Arguments are positional
Don't mix [makestr] with [malloc] for standard Tcl variables. Variables
created with [makestr] are standard Ticol variables and will be
automatically released
Literal numeric fill characters will be used if passed:
e.g. [makestr 5 0] => "00000"
Low and high order ASCII characters may be passed using [chr N] as an
argument
Examples:
makestr 5 [chr 0x21] # => "!!!!!"
makestr 5 [chr 42] # => "*****"
makestr 5 [char 0b101011] # => "+++++"
[makestr] does not accept the value [chr 0]
Ticol Tcl does not like NULL (0x0) characters within strings.
Use [memset] to zeroise the contents of a Tcl string buffer if this
is required.
If variable is specified and the default filler-character of a single space
is required either the dummy argument '--' should be used to skip at this
position or a space " " should be explicitly specified
Example:
set buf [makestr 10]
puts "-'$buf'-"
Result:
-' '-
Example:
# Use of -- to ensure default single space is used
makestr 10 -- buf
puts "-'$buf'-"
# Alternative method, using an explicit space
makestr 10 " " buf
puts "-'$buf'-"
Result:
-' '-
-' '-
Example:
set buf [makestr 1024] # Allocate a string buffer of 1024 spaces
memset buf [chr 0] 1024 # Set all characters to [chr 0]
See also: chr, asc, memset, calldll
address [malloc byte-count]
free address
free variable -var
[malloc] is provides for integration with external DLLs using [calldll],
in particular, with [calldllvariant]. The command [calldllvariant] within
plugin ticol_calldll returns a VARIANT which must be releasd using [free]
[malloc] allocates memory using the C malloc() function from the heap
memory manager. An address to a heap-allocated block of memory is returned
This address is freed by a call to [free] using the same address.
The address argument or variable value must be > 0
[malloc] and [free] should be used with extreme care and are intended for
use only with Ticol structs and for interfacing with [calldll]
CAUTION
-------
You are responsible for the safe management of allocated memory blocks
Don't use [malloc] for standard Tcl variables. Used correctly, [malloc]
will be stable and not leak memory. Memory blocks allocated with [malloc]
are NOT automatically freed. You must call [free] on the address returned
by [malloc]. Do not attempt to [free] an address which has not been
allocated via [malloc]
Value Zeroed After Free
-----------------------
If [free] is passed the name of a variable then the variable will be reset
to "0" (zero) on successful free to prevent the address being re-used.
In other cases you must take care to set malloced addresses to zero to
as appropriate in order to prevent re-use after [free]
Example:
set ptr [malloc 10]
puts "malloc ptr is $ptr"
set source Hello
# MUST copy length of string + 1 to copy the NULL char
memcpy $ptr [addressofb source] 6
puts "ofaddressb ptr =='[ofaddressb $ptr]'"
free $ptr
Result:
malloc ptr is 33394064
ofaddressb ptr =='Hello'
See also: memcpy, memset, calldll, addressofb, ofaddressb
.key
..control-code
.. comment
The Ticol manual is a simple text file which is formatted using a method
which can break the file into easily-accessible sections. Topic key markers
are phrases which must be prefixed by a dot character and the dot must be
the first character on a line. Control codes and comments are prefixed
with two dots.
The format is as follows
---------------------------------------------
.. comment and end of section marker
.topic index key
.another optional key for this topic
.yet another optional key for the same tipic
topic title
Manual text...
..code
<some Tcl code>
..end
..colour
<some text
..end
..end (of topic section)
---------------------------------------------
Any text following ".." will be ignored
See also: ticol
string [map {token-pair-list} variable|string-literal ?-nocase?]
string [string map {token-pair-list} variable|string-literal ?-nocase?]
Remaps strings using a paired list. The list-pair must have even parity.
If a variable is given for the last argument then the variable contents
will be reassigned. In either case the remapped string is returned.
If there are no valid mappings then the original input string is returned
Mapping is expensive on long strings where no multi-char remaps are found
and where a large input-map is given
Also implemented as [string map]
Examples:
puts "Result is: [string map {abc 1 ab 2 a 3 1 0} 1abcaababcabababc]"
puts "Should be: 01321221"
puts "Result is: [string map {1 0 ab 2 a 3 abc 1} 1abcaababcabababc]"
puts "Should be: 02c322c222c"
puts "Result is: [string map {1 0 ab 2 a 3 abc 1} frederick]"
puts "Should be: frederick" # No valid mapping returns original string
set s "1<2 2>1 2≥2 2≤2"
puts "Result is: "[string map -nocase {
"<" "<"
">" ">"
"≤" "<="
"≥" ">="
} $s]
puts "Should be: 1<2 2>1 2>=2 2<=2"
See also: map, apply, string, filter
max, min (Command form) |
Index |
number [max x y ...]
number [min x y ...]
Return the maxima or minima of two numeric values, either double or
integer. String arguments are not accepted. Either a double or integer
value is returned, depending on the inputs
max() and min() are also available as expression functions. You can also
access these in command-style using [funct max x y] and [funct min x y]
Example:
puts [max 3 7]
puts [max 3.0 7.0]
puts [max foo bar] # Invalid string literals
puts [max $pi 2.0]
Results:
7
7.000000000000000
0
3.141592653589793
See also: max function, min function, functions, math
integer [mid varName string|var startpos ?length? ?-var?]
Set the middle of a variable to a portion of a 2nd variable or string literal
If the length parameter is omitted then the entire source string will be
copied to the target (or as much as will fit)
[mid] is not a standard Tcl command and only operates on variables, not strings
Example:
set q $qbf
puts [mid q cat 17 3]
puts $q
Results:
3
The quick brown cat jumps over the lazy dog
See also: mids, left, right, setr, setl, strstr, string
mids (Return a String Midsection) |
Index |
string [mids string start ?length?]
string [mids var start ?length? -var]
string [string mids string start ?length?]
string [string mids var start ?length? -var]
Return a mid-section of a string
The 'start' position argument is an array 'base 1' index value
The length argument is optional and if omitted then the entire string
from 'start' will be returned
If the 'start' argument is <= 0 then a start position of 1 is assumed and
the whole string will be processed
If the 'start' argument is greater than the string then an empty string,
as "" will be returned.
If -var is passed as the last argument then the value argument will be
interpreted as a variable name instead of a string. The variable must exist
Use [index <string> n] instead of [mids <string> n 1]
[mids] is not a standard Tcl command
Example:
if {eq [mids $data $i 2] "/*" } {
incr i 2
set in_comment $true
}
Example:
puts [mids "hello world" 7]
Result:
world
Example:
puts [mids "hello world" 2 4]
Result:
ello
Example:
set s "Hello"
puts [mids $s 3 2]
Result:
ll
See also: mid, left, right, setc, setr, setl, strstr, string, index,
lastchar, strto, strtor, filter
bool [mkdir path]
bool [md path]
Create a single directory (folder)
Example:
mkdir foldername
See: [mkdirs] to create a series of folder names in one command
See also: unlink, file, ls
bool [mkdirs path]
Create a series of directories (folders)
[mkdirs] is not a standard Tcl command
Example:
mkdirs \\one\two\\three
See: [mkdir] to create a single folder
See also: unlink, file, ls
string [mktemp ?typ? ?path?]
Returns a temporary filename, optionally specifying the filetype (default
is TMP) and the path (default is the current path)
If the path does not exist an attempt will be made to create it
Example:
puts [mktemp]
puts [mktemp {} c:/temp]
string [md5 string|var ?-var?]
Return the 32-character MD5 value of a string
[md5] is not a standard Tcl command
The -var option is binary-compatible
Example:
puts [md5 $qbf]
puts [md5 qbf -var]
Result:
9e107d9d372bb6826bd81d3542a419d6
9e107d9d372bb6826bd81d3542a419d6
Example: MD5 hash-based random number generator
set seed $pi # Could instead use [clock]
set counter 0 # Global/static counter
option expression on # Use [expr] for flow-control
proc hashrnd {lo hi} { # Hash-generated random number
upvar seed # Not recursive. No level given
upvar counter
incr counter
set r [md5 [+ $counter $seed]] # Get an MD5 value
set q "0x" # Literal 0x prefix
append q [mids r [% $counter 24] 8] # Slice a section of the MD5
return [expr "$q % (($hi-$lo)+1)+$lo"] # Evaluate as hexadecimal
}
option expression on
for {set i 0} { $i < 10000} {incr i} {
puts "hashrnd 10..10000:\t[hashrnd 10 10000]"
}
See also: md5 batch example, md5_file, protection, MD5 Verification,
upvar, append
This is a simple batch file which will run a Ticol script depending on a
successful MD5 check. You will need to know the MD5 signature in advance and
if the signature changed then you would need to update the script
@echo OFF
set FILE=hello.tcx
set MD5=C7A37A87B4E6B7D311C0BF430844B94A
for /F "tokens=*" %%i in ('ticol.exe /md5 %FILE%') do (
set RESULT=%%i
)
REM ECHO DEBUG: MD5 result is: %RESULT%
if /I %MD5% equ %RESULT% (
hello.tcx
) else (
echo MD5 failure. Cannot execute %FILE%
echo MD5 is:
ticol.exe /md5:%FILE% /crlf
echo Should be:
echo %MD5%
)
See also: md5
string [md5_file filename]
Return a 32-bit MD5 hex-value for a given filename. Performance is
reasonable even on large ISO images.
This should match the MD5 of any external MD5 EXE program
Example:
puts [md5_file ticol.exe]
Result:
82bef32153657dc25dc439a3b500a614
See also: md5, protection, MD5 Verification
The following skeleton batch file may be used to generate a mechanism to
test encrypted Ticol TCX files before execution. This method requires that
the MD5 value be entered into the batch script
REM ----------------------------------------------------------
@echo OFF
set FILE=hello.tcx
for /F "tokens=*" %%i in ('ticol.exe /md5 %FILE%') do (
set RESULT=%%i
)
REM ECHO DEBUG: MD5 result is: %RESULT%
if /I 314ebb9f13a413ab4419fc9d5691011d equ %RESULT% (
ticol.exe %FILE%
) else (
echo MD5 failure. Cannot execute %FILE%
)
REM ----------------------------------------------------------
See also: md5, md5_file, security issues
double [mean value ?value? ...]
double [median value ?value?...]
double [mode value ?value?...]
[mean] calculates the arithmetic mean of a series of values and return the
result as a floating point double value. Individual values and the total are
evaluated as floating point doubles with a double return
[median] returns the 'middle' value of the sorted list of numbers after
sorting into ascending order. If 1 value is supplied, this is returned, if
2 are supplied then the lower (floor) number is returned.
[mode] returns the most frequent number. If no number is repeated more
than once then no mode value exists and zero is returned.
(double 64 bit signed 1.7E +/- 308 (15 digits of precision))
Example:
puts [mean 1.2 19.8 22.7 4343 40.5 19.4 222]
puts [median 3.4 5.4 1.2 2.2 5.1]
Result:
666.942857142857060
3.4
[mean] expects a series of arguments not a single Tcl list but e can expand
using the {*} operator to enable list handling
Example:
set l "234.34 23.11 37.333 3828.22"
puts [mean {*}$l]
Results:
1030.750749999999900
[mean] is not a standard Tcl command
See also: math functions, big mean
integer [mem_used ?pid?]
Show the memory used by the current Ticol process or any other process in
bytes. This corresponds to the data 'working set' in Windows Task Manager
not to 'memory (private)'.
This will be the same amount of memory as displayed in Task manager for
"mem usage". This can be used to track application memory use for fault-
finding etc. There should be initial memory allocation as routines are
set-up, but this should not continually increase. The underlying memory-
manager may also fluctuate but the usage trend should remain broadly
stable
Error returns:
On error mem_used returns a negative value indicating one of the following:
-1 Can't load PSPAPI.DLL
-2 Can't get procaddress for "GetProcessMemoryInfo"
-3 Can't open given Process ID (usually insufficient permissions)
Example:
set memstart [mem_used]
set last 0
set this 0
set iters 0
while {1} {
# command to check ########################
some-valid-tcl-function
###########################################
set this [mem_used]
if {!= $last $this} {
# Next line ends with a line-continuation character...
puts "Leaked [/ [- $this $memstart] 1024] Kb\tTotal used: \
[comma [/ $this 1024]] Kb\t$iters iteration(s)"
}
set last $this
incr iters
}
[mem_used] is not a standard Tcl command
See also: Ticol, debugging
memcpy target-address source-address byte-count ?-t?
[memcpy] calls the standard "C" memcpy() routine to copy a block of
memory from one location to another.
[memcpy] can take a dereferenced struct member name as the target, or
an address of a block of memory allocated by [malloc]. The copied block
may not be null-terminated and therefore may not be safely printable.
[memcpy], [malloc] and [free] should be used with extreme care and are
intended for use only with Ticol structs and for interfacing with [calldll]
[addressofb] can be used to return the binary/byte address of the allocated
buffer in a Ticol variable (similar to StrPtr() in Visual BASIC)
[ofaddresb] is used to dereference a Ticol variable which stores a memory
address
Example:
set ptr [malloc 10]
puts "malloc ptr is $ptr"
set source Hello
memcpy $ptr [addressofb source] 6
# MUST copy length of string + 1 to copy the NULL char
puts "ofaddressb ptr =='[ofaddressb $ptr]'"
free $ptr
Result:
malloc ptr is 33394064
ofaddressb ptr =='Hello'
If the -t argument is used then a null terminator will be substituted
for the last byte copied. Thus x1 byte of the source is discarded
Example:
set a [malloc 256] # Allocate a large buffer
set s "hello" # Set a standard Tcl variable
memcpy $a [addressofb s] 5 -t # Copy 5 bytes. Make the last byte NULL
puts "a is '[ofaddressb $a]'" # Print the binary data at address a
Result:
a is 'hell'
[memcpy] is not a standard Tcl command and should be used with extreme
care
See also: malloc, free, memset
memset varName character length
[memset] may be used to format or zeroise buffers such as those created
using [makestr] when interfacing with external DLLs using calldll*
[set] will copy only the usable part of a string, not the full, allocated
buffer size. Thus using [set] with [makestr] and [chr 0] could create an
under-sized target buffer string and cause instability with [calldll]
[memset] will clear memory only up to the allocated length of the string
it will not clear/set the internal/terminating null character of any string
Don't use [memset] with [makestr] as the latter is a Tcl command, not
a heap memory command.
It is allowable to call memset with a calculated address offset
Avoid using memset with Tcl variables other than those used via calldll*
A record should be kept of the return address and that address manipulated
rather than passing data into a Ticol variable object.
Example:
set buf [makestr 1024 " "] # Creates a buffer of 1024 spaces
memset buf [chr 0] 1024
puts "'$buf'"
Result
'' # Buffer has been allocated to 1024 chars but set to [chr 0]
Example:
set ptr [malloc 200] # [malloc] returns an address
memset $ptr x 10 # Write 10 char "x"
memset [+ $ptr 5] "q" 2 # Write "qq" to offset ptr + 5
puts "Address is: $ptr"
puts "Contents: '[ofaddressb $ptr]'"
Result:
Address is: 34791712
Contents: 'xxxxxqqxxx'
[memset] is not a standard Tcl command and should be used with extreme
care
See also: calldll, malloc, free, memcpy, makestr
Troubleshooting and Common Mistakes |
Index |
Here are some of the most common mistakes when first using Tcl.
Particularly when coming from a C/C++ background:
No Text Output
Command Switch Arguments are Case Sensitive
Using Double Quotes Where Braces Are Preferable
Mismatched Square Brackets in Quoted Strings
Failure to Spot an Error Condition During Debugging
Line Continuation \ Is Not The Same As A Multiline String
Accidental Misuse of # (Hash) Comments With Multiline Strings
Accidental Misuse of # (Hash) Comments With Macros
Very Long Strings in Notepad++
Using [unset] instead of [undef]
Failure to Use [option expression off]
Using 'VB' Style Variable Syntax
Failure to Use [upvar] for [proc] Arguments
Invalid Variable Names
Variable Name Clash Within Procedures when Using [upvar]
Typo Leads to Infinite Loop
Storing [expr] Commands in Strings
Lapsing into [expr] Format Within Flow Control Statements
Failure to clear/delete variables or procs
Failure to Evaluate [proc] Returns
Using [static] where [set] is required
Handling Sparse Integer Indexed Arrays
Incorrect Options Setting with [for] etc.
Can't Use Tcl Variables for [proc] Default Arg Values
Erroneously Declaring a Global as a Local Var
Incorrect [option expression] Setting
Mis-handling of Number Range
Misplaced End of Line Brace Characters
Failure to Separate Commands by Whitespace
Failure to Protect Strings With Double Quotes
Misplaced Double Quotes
Incorrect Delimiting of Command Arguments
Using [else if] Instead of [elseif]
Using [then] Within [if]
Finding Bad PreProcessor Expressions
Malformed CRLFs in Files
Junk Characters
Windows Command Line Breaks Direct Commands
Memory Use
Big Number Library Returns Unexpected Results
No Text Output
--------------
Problem: No output or activity past a certain point in the script
Fix: Check for a malformed string and malformed escape sequences with
an accidental string overrun to the end of the file
Find the halt point in the script using line traces to locate the
statement where the problem occurs and isolate the malformed statement
Try to simplify the string and avoid an excess of \ escape sequences
or embedded double-quotes as it is easy to lose track. Use minimal quoting
style and use braces instead where possible. Break down complex strings
into simpler tasks with smaller strings then reassemble.
Don't embed '#' comment characters within braced strings, and if embedded
within double-quoted strings, be sure to escape using '\#'
Command Switch Arguments are Case Sensitive
-------------------------------------------
Where command options are available, unless otherwise specified, these
must be passed in lower case. e.g. pause Hello -quiet
Using Double Quotes Where Braces Are Preferable
-----------------------------------------------
Tcl does not support nested quoted strings (nested double quotes)
and since braces perform the same task it is often preferable for
braces to be used unless string substitution or immediate command
execution is required
Note, however, that some escapes such as \# are valid only within quoted
strings. These should be avoided if possible
Forgetting to Include Required Whitespace Between Words
-------------------------------------------------------
Because Tcl consists of a series of commands which are delimited by
<whitespace>[command-name]<whitespace>, if this whitespace is
inadvertently omitted then a syntax error will result. What would
be distinct operators in other languages are commands in Tcl...
Example:
# [!] is a command not an operator
if{![expr 2+2]} {... # C++ like syntax is wrong
if{! [expr 2+2]} {... # Whitespace is required after '!'
Mismatched Square Brackets in Quoted Strings
--------------------------------------------
Square bracket parity is checked outside of quoted strings by the
macro preprocessor (MPP) but not within quoted strings. Since Tcl
allows quoted strings to contain commands which will be evaluated
it is important to check for mismatched [] brackets if anomalous
behaviour occurs.
Failure to Spot an Error Condition During Debugging
---------------------------------------------------
You may not want console error messages displayed and may prefer to
handle these yourself and display your own. However, if you disable
console echo of fatal errors using [option echo off] you risk failing
to spot the cause of a fatal error and this could be difficult to
debug.
If in doubt, always ensure that [option echo on] overrides as close
to the location of any potential fault during debugging. This is
particularly the case if you spawn external script files using [source]
and which may disable console echo.
Ideally, child scripts should save the current echo setting using
[option push] and restore it on exit using [option pop]
e.g. [option echo push on], [option echo pop]
Example:
# Here a fatal error triggers the premature exit of [time] with no
# indication of why. 'child_script.tcl' inadvertently unsets
# variable 'i', and also turns echo off, thus '++ i' fails
# The only indication is given by the error code from [time -info]
set i 0
time [
puts "Loop $i"
source child_script.tcl
# option echo on # This is required to spot the problem
++ i
] 100000 -info
Line Continuation \ Is Not The Same As A Multiline String
---------------------------------------------------------
Line-continuation characters at the end of a quoted string will cause the
lines to merge and behave as a single line. Where a quoted string is
defined over several lines without continuation characters, this string
will retain CRLF line endings after each line.
# String will be "The quick brown fox"
set s "The \
quick \
brown \
fox"
Is not the same as
# String will be "The \r\nquick\r\n brown\r\n fox"
set s "The
quick
brown
fox"
The first example will result in a single line, the second, multiple lines.
Accidental Misuse of # (Hash) Comments With Multiline Strings
-------------------------------------------------------------
Hash characters are not interpreted within quoted strings. Thus a #
comment can't be included at the end of any line within a multiline
string definition. For example, the comments below are invalid:
set s "The
quick brown # Foxes aren't that quick
fox jumps
over the
lazy dog"
set s "The\
quick brown\ # Foxes aren't that quick
fox jumps\
over the\
lazy dog"
Accidental Misuse of # (Hash) Comments With Macros
--------------------------------------------------
It is good practice to insert a space after a # character. This prevents
inadvertently calling a macro command. For example. If one intended to
comment-out the 'echo' command by prefixing with a # then the macro
preprocessor version would be called. With CGI applications this would
trigger text output before headers were sent and generate a 500 server
error.
Examples:
#echo "Some text" # This calls the macro #echo command
# echo "Some text" # Inserting a space is recommended, thus
#if # This will execute the MPP '#if' clause
#else # Dangerous for the same reason
#exit # Will cause the MPP to exit immediately
These are examples of a dangerous comments because '#if' and '#else' are
reserved Macro Preprocessor keywords. Commenting out a Tcl [if] or [else]
statement will create a valid Macro directive
Very Long Strings in Notepad++
------------------------------
It has been noted that some versions of NPP may automatically wrap and
insert CRLFs into very long strings, e.g. very large numbers of say
1000+ digits
The solution to this is to break the string into smaller pieces, wrap in
double quotes and use line-continuation backslash characters at the end
of each intermediate line. These will then be correctly handled by the
Macro PreProcessor
Example:
assert [big fact 200] {== $_ "7886578673647905035523632139321850622951\
3597768717326329474253324435944996340334292030428401198462390417721213\
8919638830257642790242637105061926624952829931113462857270763317237396\
9889439224456214516642402540332918641312274282948532775242424075739032\
4032125740557956866022603190417032406235170085879617892222278962370389\
7374720000000000000000000000000000000000000000000000000"} -v
Using [unset] instead of [undef]
--------------------------------
Intending to clear a proc (e.g. foo) but using 'unset foo' where
'undef foo' is required, or intending to clear var foo using
'undef foo' where 'unset foo' is required
Failure to Use [option expression off]
--------------------------------------
When using non-expression (Tcl command format) arguments to flow-control
The preference can be set up in ticol.ini
Using 'VB' Style Variable Syntax
--------------------------------
Using var$ instead of $var. Tcl requires the latter
Example:
set i 0
for {[> i$ 10]} { ... # Wrong
set i 0
for {[> $i 10]} { ... # Correct
Failure to Use [upvar] for [proc] Arguments
-------------------------------------------
When passing by name (by reference) rather than by value, you must use
[upvar] to create a reference to the variable. If a [proc] is called by
another [proc] then [upvar] giving no level reference is sufficient as
each [proc] call is one level higher than the next.
Example:
proc foo x { # Wrong. Variable x won't be referenced
return [* $x 1]
}
proc foo x { # Correct
upvar x # [upvar] creates a linkage to the source var
return [* $x 1]
}
proc foo x { # Correct (using an alternate local variable)
upvar x y # $y is an alias for $x
return [* $y 1]
}
Invalid Variable Names
----------------------
Variable names in Ticol are more restricted than in standard Tcl
The reason for this is to speed-up the interpreter core.
It is is possible to create variable names which can't be dereferenced
using the dollar '$' symbol. In such cases it is usually possible to
dereference using [set] instead.
The dollar parse routine differs from the string parse routine and
commands which accept strings will tolerate a wider range of characters.
Under normal circumstances it is strongly recommended to adhere to the
restricted range of characters available when choosing variable names.
Example:
set [chr 0x1e] Hello
puts [set [chr 0x1e]] # Will dereference okay
puts ${[chr 0x1e]} # Will raise an error
Results:
Hello
# (12400) eval: No variable reference to '{[chr 30]}' at level 0
Variable Name Clash Within Procedures when Using [upvar]
--------------------------------------------------------
Consider the following proc which will pass its argument by name
(reference)
proc foo {a} { # Variable 'a' is created in [proc] scope
upvar $a # Attempts to declare a second local var 'a'
puts $a
}
set a Hello
foo a # Calling with 'a' means [proc]->'a' resolves to 'a'
An error will result, pointing out that variable 'a' already exists in
the current scope This occurs because we passed the name 'a' via argument
'a' which then resolves to 'a' when dereferenced
upvar: 'a' already exists at level 1
which it does, since it was passed via the proc body. Because you can't be
sure what variable name will be passed into the proc later on you must
avoid such clashes where the inbound name is the same as the local one
You must either explicitly declare the local variable in the [upvar]
statement, e.g. in this case: upvar $a b or 'mangle' the names
using underscores e.g.
proc foo {_a} {
upvar $_a a # Declares a local var 'a'
puts $a
}
Alternatively, just ensure the variable names for the local reference
are different from the arguments when calling by name
proc foo {a} {
upvar $a b # Declares a local var 'b'
puts $b
}
Typo Leads to Infinite Loop
---------------------------
option expression off
for {set i 0} {< $i count} {++ i} {
This should be '$count' not 'count' (a difficult bug to spot)
option expression off
for {set i 0} {< $i $count} {++ i} {
Storing [expr] Commands in Strings
----------------------------------
If storing [expr] or [calc] command sequences, do not also store the
command [expr] within the string. Store only the argument tail unless
you wish to use the {*} expansion operator to resolve:
# Correct
set a {1+2*3}
puts [expr $a]
# Wrong
set a {expr {1+2*3}}
puts [$a]
# Correct
set a {expr {1+2*3}}
puts [{*}$a]
# Wrong. [expr] args should be braced but '{*}a$' is a command
# The subcommand should be wrapped in square brackets
set a {expr {1+2*3}}
puts [expr {{*}$a}]
# Correct
set a {expr {1+2*3}}
puts [expr [{*}$a]]
This can be debugged in single-step detail by using [option breakpoint on]
and issuing a [halt] instruction
Lapsing into [expr] Format Within Flow Control Statements
---------------------------------------------------------
# Wrong '!' is a Tcl command
if {![info exists env(QUERY_STRING)]} {
set env(QUERY_STRING) ""
}
# Correct (a space separator is inserted)
if {! [info exists env(QUERY_STRING)]} {
set env(QUERY_STRING) ""
}
Failure to clear/delete variables or procs
------------------------------------------
A conflicting proc or variable exists somewhere else in the code, possibly
within an external file. It is good practise to clear procs and variables
as soon as you have finished within them, particularly when working in the
global (root) scope. This can be done safely using:
undef procname -nocomplain
unset variablename -nocomplain
[clear] will clear all locally created variables
A useful shortcut scriptlet is as follows: This will remove all single
letter variables 'a' to 'z'. The {*} will cause each list element to be
passed to [unset] as a separate argument:
unset {*}[split [range a-z]] -nocomplain
Failure to Evaluate [proc] Returns
----------------------------------
Example:
return * $i $i
Should be:
return [* $i $i]
Reason:
[*] is a command which must be evaluated. The first example, instead,
supplied 3 invalid arguments to the [return] command.
i.e. [return * $i $i].
Square brackets cause an advance recursive evaluation, thus, if i was
set to 3 then [return [* $i $i]] would be interpreted as [return 9]
Using [static] where [set] is required
----------------------------------------
Static is used for declaration with an optional initialiser
It should not be used for assignment after declaration
# In root-level scope
static foo 23
assert $foo {== $_ 23} -v # PASS
static foo 11
assert $foo {== $_ 11} -v # FAIL
set foo 11
assert $foo {== $_ 11} -v # PASS
Handling Sparse Integer Indexed Arrays
--------------------------------------
Tcl arrays are associative rather than sequential binary blocks. They have
no implicit order to them, so iterating over a 'sparse' integer sequence
can be problematic. Use of [foreach] may not deliver the required numeric
sequence or ordering.
[array item] can be used to safely iterate over a desired sequence but
there may be a computational cost to doing this.
This solution is far from idea. [carray] in the ticol_carray plugin offers
an alternative to Tcl arrays. It may also be useful to built up an index
to large sparse arrays. "C" arrays may be non-sparse, expanded and kept in
explicit, sorted order. They may therefore be traversed at very low cost.
Example:
# Sparse array example
# Assuming we don't know the contents of array 'a' in advance and that
# the array size could be up to 100,000 elements
set a(1) "Hello "
set a(999) "Fine "
set a(10000) "World "
loop i 0 100000 {
set q [array item a $i]
if {ne $q ""} {
puts $q -nonewline
}
Incorrect Options Setting with [for] etc.
-----------------------------------------
Example:
variable not found after [for] loop initiation
Should be:
for {[set i 0]} ...
or
option expression off
for {set i 0} ...
or:
TICOL.INI -> ForInitAllowsExpression=FALSE
for {set i 0} ...
Reason:
Standard Tcl forces the first (initiator) statement of a [for] loop
to be a Tcl command (via [eval]). Ticol Tcl allows the first command
to be an expression if configured in TICOL.INI using:
ForInitAllowsExpression=<boolean>
Can't Use Tcl Variables for [proc] Default Arg Values
-----------------------------------------------------
String or numeric literals should be used for default values in [proc]
default argument lists. Such default values should be considered as
string or numeric constants.
Braces will precent these from being resolved when the proc is defined
Example:
Here the global '$::true' will not be resolved to '1'
proc foo {{x $::true}} {...
Instead use:
proc foo {{x 1}} {...
A future version may offer a workaround for this problem
Erroneously Declaring a Global as a Local Var
---------------------------------------------
The intent is to declare a global variable or global reference but instead a
local is declared due to the inadvertent omission of the '::' global scope
prefix. Often this will be where the global is 'masked' by a local with the
same name.
Example:
# Local 'masks' intended global reference in for loop
option expression off
set maximum 10
proc foo {} {
set maximum 0
for {set i 0} {< $i $maximum} {++ i} {
puts $i
}
}
Incorrect [option expression] Setting
-------------------------------------
Example:
option expression on
if {file exists test.tcl} { ... }
Should be:
if {[file exists test_enum.tcx]} {
or
option expression off
if {file exists test.tcl} { ... }
Reason:
If option expression is on, either by command or INI config file, then
the arguments to flow-control commands such as [if], [for] etc. will be
interpreted as math expressions. If this is the case any command must be
evaluate (called) by wrapping in square brackets.
--
Example:
option expression off
for {set i 0} {$i < 10} {incr i} { ...}
Should be:
option expression off
for {set i 0} {< $i 10} {incr i} { ...}
or:
option expression on
for {[set i 0]} {< $i 10} {incr i} { ...}
Reason:
With option expression off, arguments to flow-control commands such as
[for], [while] etc. will not be evaluated in "expression format" but in
[eval] format. The command expression in the first argument will have no
effect
With 'option expression on' '{set i 0}' is not a valid expression
Mis-handling of Number Range
----------------------------
Ticol natively works with 64-bit signed integers but many commands accept
and return only 32-bit signed integers. Care should be taken to appreciate
the limits of specific commands as stated in their definitions here.
Misplaced End of Line Brace Characters
--------------------------------------
C/C++ and C# programmers may inadvertently assume that the following brace
style can be used with Tcl. It cannot. Opening braces mark delayed
expansion wrappers for commands. CRLFs are valid within delayed brace
wrappers but not within argument tails outside curly braces
Example: (wrong - opening brace must be on same line)
if {== $i 10}
{
puts "i is 10"
}
else
{
puts "i is not 10"
}
Should be: (correct - opening brace on same line)
if {== $i 10} {
puts "i is 10"
} else {
puts "i is not 10"
}
or
if {== $i 10} {puts "i is 10"} else {puts "i is not 10"}
Reason. The first expression is "C++"-like but illegal. The CRLF marks the
termination of the command [for] and it is prematurely truncated
Failure to Separate Commands by Whitespace
------------------------------------------
Example:
option expression on
for{set i 0}{$i < 10}{incr i}{ ...}
Should be:
option expression on
for {set i 0} {$i < 10} {incr i} { ...}
Reason:
Tcl is a discrete, command-based language. Spaces are required between
commands and grouped arguments (those wrapped in braces)
Failure to Protect Strings With Double Quotes
---------------------------------------------
Strings need not be wrapped in double-quotes but failure to do so leaves
a path open for frustrating errors. e.g.
puts [ftp $hostname ls /tools/* -u admin -p xxxxx]
Should be:
puts "[ftp $hostname ls /tools/*.zip -u admin -p xxxxx]"
Since the unquoted '/*' in '/*.zip' will be interpreted as a long comment
start. Quoting will prevent this.
Misplaced Double Quotes
-----------------------
If an extra, unescaped double quote is embedded in a properly quoted string
then this error may be issued:
"Premature end preprocessing source file"
Possibly with a preprocessor line number. Work backwards from the line
number to check for strings with embedded/unescaped double quotes
Incorrect Delimiting of Command Arguments
-----------------------------------------
Bad example:
puts The quick brown fox jumps over the lazy dog
Should be:
puts "The quick brown fox jumps over the lazy dog"
or, (preferably):
puts {The quick brown fox jumps over the lazy dog}
or:
puts Thequickbrownfoxjumpsoverthelazydog
Reason:
Arguments to Tcl commands such as [puts] must be either wrapped in double-
quotes (where appropriate) or (preferably) grouped using curly-braces.
The first example instead supplies 9 independent and invalid arguments to
[puts]. Arguments are also implicitly grouped if there is no intervening
whitespace, thus an argument which contains no whitespace need neither be
braced or quoted
Using [else if] Instead of [elseif]
-----------------------------------
Example: (wrong)
if {[rnd 0 1]} {
puts "True"
} else if {[== [rnd 2 3] 3] } # 'else if' is not a valid keyword
puts "Three"
}
Example: (correct with nested construct)
if {[rnd 0 1]} {
puts "True"
} else {
if {[== [rnd 2 3] 3] }
puts "Three"
}
}
Example: (correct using 'elseif')
if {[rnd 0 1]} {
puts "True"
} elseif {[== [rnd 2 3] 3] }
puts "Three"
}
Reason:
'elseif' is not a command but an argument of the command [if]
Using [then] Within [if]
------------------------
Example: (wrong)
if {[rnd 0 1]} then {
puts "True"
}
Example: (correct)
if {[rnd 0 1] {
puts "True"
}
Reason:
There is no [then] argument to the Ticol [if] command
Finding Bad PreProcessor Expressions
------------------------------------
Problem: Preprocessor Bad expression at line xxx, near
and the script fragment can't be distinguished
Solution: Place #exit statements at strategic locations and run
ticol.exe scriptname /echo /na
until the line is located. Use binary search strategy in
placing the #exit statements
Malformed CRLFs in Files
------------------------
Problem: Files loaded into an editor have double-line spacing
Evidence: A hex dump of the file shows line endings to be a sequence of
0x0d 0x0d 0x0a ("\r\r\n")
Reason: File I/O is translating "\n" (0x0d) characters
to "\r\n" (0x0d 0x0a) pairs
Thus, passing \r\n will keep the "\r" and substitutes "\r\n"
for "\n" and generating "\r\r\n"
Fix: Emit a "\n" character in the appropriate sections of code
instead of "\r\n"
Junk Characters
---------------
Be careful when pasting-in code from websites and non-text editors.
Occasionally, non-visible "junk" characters may find their way into source
code and result in difficult to debug issues. These characters may not be
visible unless certain options are enabled in your editor
Take particular care when pasting-in Unicode/UTF-8 source
Windows Command Line Breaks Direct Commands
-------------------------------------------
Passing a direct command via the Windows command line requires that the
command is wrapped in double-quotes rather than braces. If the command
is braced then Windows will break apart the command string and supply
only the first part. The result will be a brace parity error.
Example:
ticol.exe ; {puts Hello} # Too many { characters
# Windows will supply '{ puts'
Instead, direct commands passed from the command-line should be as
follows and should use bracing for embedded strings...
ticol.exe ; "puts {Hello World}"
Memory Use
----------
Various structures and objects within Ticol may progressively consume
amounts of memory, the allocation of which may be visible within
Windows Task Manager or another process monitor. Array hash tables
perform an incremental "smart resize" periodically and are not reduced
until the array is unset. Various aspects of memory may become fragmented
during long and complex loops. Unless memory use is continually escalating
there should be no issues and this is normal behaviour.
Common memory load on first loading Ticol is usually just under 2Mb
During operation this may rise to 10 or 20 Mb particularly if very large
data structures are created
See:
http://phaseit.net/claird/comp.lang.tcl/fmm.html Frequently Made Mistakes
Big Number Library Returns Unexpected Results
---------------------------------------------
The big number library is an integer only library with a few exceptions
Attempts to supply real numbers is not error-checked. The caller must
'sanity check' all inputs. Calculations may return zero or unexpected
results
puts [big pow 2 10] # Returns 1024
puts [big pow 2.6 10] # Also returns 1024 (accepts only integer)
See also: optimisation, performance
bool [movefile sourcefile targetfolder ?-y?]
If target is specified as a foldername, ensure it ends with a trailing '\'
character or it will be assumed to be a new filename
Wildcards are not supported
[movefile] can rename files as well as move them
Forward-slash characters '/' may be used as path separators to avoid
false escape issues with '\' characters
Where wildcard+forward-slash combinations are present, be sure to wrap
any filename in double-quotes to avoid /* being interpreted as a comment.
CAUTION: Supports only ANSI filenames, not Unicode/MBCS
Examples:
movefile test.txt test\\ # Move into subfolder 'test'
movefile test.txt test/ # Move into subfolder 'test'
movefile test.txt test.dat # Rename file as test.dat
See also: copyfile, file
integer [msgbox message ?title | -? ?flags?]
Displays a standard Windows messagebox dialogue
A title is optional and may be omitted or specified as blank using the '-"
placeholder in order to supply a flag value with no title.
msgbox returns Windows-standard numeric return-codes
Return values:
yes 6
no 7
Flag values:
0x0 MB_OK
The message box contains one push button: OK
This is the default
0x1 MB_OKCANCEL
The message box contains two push buttons: OK and Cancel
0x2 MB_ABORTRETRYIGNORE
The message box contains three push buttons: Abort, Retry, and Ignore
0x3 MB_YESNOCANCEL
The message box contains three push buttons: Yes, No, and Cancel
0x4 MB_YESNO
The message box contains two push buttons: Yes and No
0x5 MB_RETRYCANCEL
The message box contains two push buttons: Retry and Cancel
0x6 MB_CANCELTRYCONTINUE
The message box contains three push buttons: Cancel, Try Again, Continue
Use this message box type instead of MB_ABORTRETRYIGNORE
0x10 MB_ICONSTOP
A stop-sign icon appears in the message box
0x10 MB_ICONERROR
A stop-sign icon appears in the message box.
0x10 MB_ICONHAND
A stop-sign icon appears in the message box.
To indicate the default button, specify one of the following values
0x20 MB_ICONQUESTION
A question-mark icon appears in the message box.
The question-mark message icon is no longer recommended because it
does not clearly represent a specific type of message and because the
phrasing of a message as a question could apply to any message type.
In addition, users can confuse the message symbol question mark with
Help information. Therefore, do not use this question mark message
symbol in your message boxes. The system continues to support its
inclusion only for backward compatibility.
0x30 MB_ICONWARNING
An exclamation-point icon appears in the message box.
0x40 MB_ICONINFORMATION
An icon consisting of a lowercase letter i in a circle appears in the
message box.
0x40 MB_ICONASTERISK
An icon consisting of a lowercase letter i in a circle appears in the
message box.
0x0 MB_DEFBUTTON1
The first button is the default button.
MB_DEFBUTTON1 is the default
unless MB_DEFBUTTON2, MB_DEFBUTTON3, or MB_DEFBUTTON4 is specified.
0x100 MB_DEFBUTTON2
The second button is the default button.
0x200 MB_DEFBUTTON3
The third button is the default button.
0x300 MB_DEFBUTTON4
The fourth button is the default button.
To indicate the modality of the dialog box, specify one of the
following values.
0x0 MB_APPLMODAL
The user must respond to the message box before continuing work in the
window identified by the hWnd parameter. However, the user can move to
the windows of other threads and work in those windows
Depending on the hierarchy of windows in the application, the user may
be able to move to other windows within the thread. All child windows
of the parent of the message box are automatically disabled, but pop-up
windows are not
MB_APPLMODAL is the default if neither MB_SYSTEMMODAL nor MB_TASKMODAL
is specified
0x1000 MB_SYSTEMMODAL
Same as MB_APPLMODAL except that the message box has the WS_EX_TOPMOST
style. Use system-modal message boxes to notify the user of serious,
potentially damaging errors that require immediate attention (for
example, running out of memory). This flag has no effect on the user's
ability to interact with windows other than those associated with hWnd
0x2000 MB_TASKMODAL
Same as MB_APPLMODAL except that all the top-level windows belonging to
the current thread are disabled if the hWnd parameter is NULL. Use this
flag when the calling application or library does not have a window
handle available but still needs to prevent input to other windows in
the calling thread without suspending other threads.
0x4000 MB_HELP (Not supported)
Adds a Help button to the message box. When the user clicks the Help
button or presses F1, the system sends a WM_HELP message to the owner
You can use [enum] to set these const values as follows:
enum {
MB_OK
MB_OKCANCEL
MB_ABORTRETRYIGNORE
MB_YESNOCANCEL
MB_YESNO
MB_RETRYCANCEL
MB_CANCELTRYCONTINUE
MB_DEFBUTTON1 = 0
MB_ICONSTOP = 0x10
MB_ICONERROR = 0x10
MB_ICONHAND = 0x10
MB_ICONQUESTION = 0x20
MB_ICONWARNING = 0x30
MB_ICONINFORMATION = 0x40
MB_ICONASTERISK = 0x40
MB_DEFBUTTON1 = 0
MB_DEFBUTTON2 = 0x100
MB_DEFBUTTON3 = 0x200
MB_DEFBUTTON4 = 0x300
MB_APPLMODAL = 0
MB_SYSTEMMODAL = 0x1000
MB_TASKMODAL = 0x2000
MB_HELP = 0x4000
}
Example:
msgbox "Yes or no" "Question" [| 4 32]
Result:
7
Example:
option autoexec off
option escape on
msgbox [join [split $qbf " "] "\n"] "Choose" 36
Result:
Messagebox which displays each word on new line with Yes, No buttons
See also: inputbox, box, get_fileopen, get_filesave, get_folder, enum
As with C/C++, multiple commands may be separated using an intervening
semi-colon (;)
This is particularly useful when calling Ticol from the Windows console
Example:
set a 22;set b 7.0;set p [expr $a/$b];puts "Pi is $p"
Command line example:
# Run with no autoexec
ticol.exe ; "set a 22;set b 7.0;set p [expr $a/$b];puts {Pi is $p}" /na
See also: syntax, Tcl, Ticol
Namespaces and Scope |
Index |
Ticol, as a minimal Tcl implementation, does not support full Tcl namespaces,
however they may be emulated by prefixing procedure and variable names with
namespace tag value and double-colon. You will need to use the full,
emulated namespace name of the procedure when calling as a namespace context
cannot be set
Namespace-prefixed variables are useful to help avoid variable name
collisions where the same name may be inadvertently be chosen in more than
one Tcl script
Care should be taken to avoid misinterpretation of namespaces by adjacent
variables or punctuation in strings. This may require enclosing the tag
value in curly-braces.
Note that the double-colon 'namespace' scope prefix should not be confused
with global scope use (e.g. '::var' and '$::var')
If you wish to use colons adjacent to variables you must enclose the
variable name in braces ...
Example:
set s "I am a pretty string"
puts "${s}::Punctuated::By colons" # Correct
puts "$s::Punctuated::By colons" # Incorrect
Result:
I am a pretty string::Punctuated::By colons
eval: No variable 's::Punctuated::By' in the current scope
Here is a demonstration of procedure and variable namespace emulation:
Example:
set local::a 12345 # We will overwrite this using the proc
proc math::div {x y} { # Example divide procedure
upvar local::a a
set a "div was called"
if {!= $y 0} {
[/ $x $y]
} else {
return 0 -code error
}
}
set a [math::div 22 7.0]
puts $a
puts "\$local::a=$local::a (Should say \"div was called\")"
Result:
3.142857142857143
$local::a=div was called (Should say "div was called")
Example:
An example of scoped variable use
set scope fred
set ${scope}::j 0
++ ${scope}::j
puts "fred is $fred::j"
Result:
fred is 1
This example of scoped variable use using pseudo-namespaces is more
complex in that the nested-resolution of an indirect scope variable (fred)
and subsequent dereference is problematic without the use if [set] to
dereference since the outer '$' reference will be evaluated before the
inner reference unless a [command] is used
[set] can echo or 'dereference' a variable as well as set it
Example:
option expression on
set scope fred
for {set ${scope}::j 0} {[set ${scope}::j] < 500} {++ ${scope}::j} {
set q [set ${scope}::j]
if {[is_mod $q 100]} {
puts "${scope}::j is $q"
}
}
Results:
fred::j is 0
fred::j is 100
fred::j is 200
fred::j is 300
fred::j is 400
See also: proc
Multi Line Commands |
Index |
Providing that a newline is not encountered before an opening brace or
bracket, the contents of a command or expression can be spread across
multiple lines. This is normal for braced multiline expressions but is
also true for commands which are enclosed in square brackets
Example:
# Single-line command
puts [eval "/ 22.0 7.0"]
# Multi-line command
puts [
eval "/ 22.0 7.0"
]
See also: command separator
Networking Commands |
Index |
See the [info] command for various commands for querying the local network
External DLLs such as ping32.dll may be dynamically linked-to and used
See also: info, ping, ftp, dns, arp, dhcp, calldll
new sourceobject newobject
Create a new and empty instance of an object. Source/template objects have
no special designation. Any handled object can be used as the source
template. Currently, [new] detects and handles only structs
For structs, the total binary struct size, field count, field names
and field widths are copied. Field names will be distinguished (decorated)
with the name prefix of the struct parent variable name
Structs are binary-compatible and may be used with [calldll_*] and passed
to the Windows API and other DLLs. Each Tcl field variable is mapped to
a binary-offset in the struct's data-block
[new] is not implemented for arrays, standard variables or other data
types as it would serve no useful purpose
Manual deletion of objects created using [new] is not required. Garbage
collection is automatic
Example:
struct s {{field1 10} {field2 20}} # Create a source struct
struct set s.field1 Hello # Assign some data
struct set s.field2 World
new s s_new # Copy s as s_new
dump s_new # Display debug info
Results: (Displays a blank struct)
var type is 3 v->val is 36811792
struct 's_new' level 0 (0x231b430) data @36811792 (0x231b410)
¦ 30 byte(s), 2 member(s)
+ 1:'s_new->s_new.field1' => 36811792 (0x231b410) 10 byte(s)
+ 2:'s_new->s_new.field2' => 36811802 (0x231b41a) 20 byte(s)
Address 0 1 2 3 4 5 6 7 8 9 A B C D E F ASCII
-------- ----------------------- ----------------------- ----------------
0231b410 00 00 00 00 00 00 00 00 - 00 00 00 00 00 00 00 00 ................
0231b420 00 00 00 00 00 00 00 00 - 00 00 00 00 00 00 00 00 ................
* Var 0::'s_new' has no parents
See also: struct, set
newline ?count?
Prints a new line to the console.
Equivalent to [puts ""] or [printf "\r\n"]
The default is to issue one newline (CRLF pair). An optional argument may
specify the number of new lines to print
[newline] offers no particular performance advantage but may make code
easier to read
See also: puts
not (Bitwise and Logical) |
Index |
integer [! expression] # Logical not
integer [! string] # Logical not
integer [~ expression] # Bitwise not
The '!' operator returns a logical true/false as 1 or 0. [!] will also
operate on non-numeric strings as a test for an empty string. An empty
string or empty list will return true (1), otherwise false. Standard
number detection rules apply. Anything which is not a valid number will
be interpreted as a string
The bitwise not '~' operator performs a bit-inversion on the expression
and returns the result
Example Returns
puts [expr "!(-12 -3)" ] 0
puts [expr "! $true" ] 0
puts [expr "! $false" ] 1
puts [expr "! -12 -3" ] -3
puts [expr "~(-12 -3)" ] 14
puts [~ 0] -1
puts [~ 1] -2
puts [! Hello] 0
puts [! 1.2.3.4] 0
puts [! ""] 1
puts [! {}] 1
Example:
set s {}
if {! $s} {
puts "s is empty"
}
Result:
s is empty
Example:
set s Hello
if {! $s} {
puts "s is empty"
} else {
puts "s is not empty"
}
Result:
s is not empty
Example:
set ip_address 1.2.3.4
if {! $ip_address} {
puts "ip_address is empty"
} else {
puts "\$ip_address is not empty"
}
Result:
$ip_address is not empty
See also: expr
Non-Decimal Numbers |
Index |
Ticol supports several non-decimal const number formats as follows:
Unquoted values (freeform integer):
Type Prefix Range Example
---- ------ ----- -------
Binary 0bN 0..1 0b01010101
Octal 0oN 0..7 0o13652104
Hexadecimal 0xN 0..9,a..f 0xdeadbeef
Hexadecimal 0XN 0..9,A..F 0xDEADBEEF
Quoted values (escape character):
Type Prefix Range Example
---- ------ ----- -------
Octal \0N 0..7 0o13652104
Hexadecimal \xN 0..9,a..f 0xdeadbeef
Hexadecimal \XN 0..9,A..F 0xDEADBEEF
The Macro PreProcessor (MPP) will attempt to transcode non-decimal numbers
into decimal equivalents where possible during the preprocessing phase of
script loading or via the CLI. The CLI may not translate all numbers and
if in doubt, [expr] may be used to ensure they are translated
Non-decimal numbers are not translated where embedded within double-quotes
In such cases, explicit conversion using [expr] must be used
The commonly used convention of numbers which are prefixed with zero (0)
being assumed as octal is not adhered to. Octals require a "0o" prefix
Binary Numbers
--------------
Often useful for bit manipulation/bit-shifting algorithms
puts 0b110000000 # Not translated within the CLI
Also available via the ticol_bignum.dll plugin
Hexadecimal Numbers
-------------------
The most common alternative to decimal format numbers and understood by
most Ticol commands. Prefixed with "0x"
# Prints ASCII code for the letter 'A' (65)
puts 0x41 # Not translated within the CLI
puts 0X41 # Not translated within the CLI
Octal Numbers
-------------
Less commonly used than hexadecimal numbers (0..7) and prefixed with "0o"
Roman Numbers
-------------
These may be used via the ticol_r2d.dll plugin and due to late-binding
of the lib load are not processed by the MPP but via direct commands
Roman numbers are integer-only and have a small, positive range
puts [roman_to_decimal mdcclxxvi]
puts [decimal_to_roman 1776]
Results:
1776
MDCCLXXVI
See also: cli, data type ranges, roman numbers
double [now ?-ctime?]
Returns the current date/time as a VB-compatible 8-byte double value
or, if the -ctime argument is used, as a C/Unix long integer
a
Example:
puts [now] # VB data format
puts [now -ctime] # C/C++ data format
Result:
42933.993946759263000
1500331879
[double_to_ctime] can also convert...
Example:
clock format [now -ctime]
clock format [double_to_ctime [now]]
Result:
Tue Mar 03 22:57:35 2015
Tue Mar 03 22:57:36 2015
See also: double_to_ctime, ctime_to_double, date, time, clock
Number formats understood by various commands including [eval] and [expr]
are as follows:
Type Format Signed? Bytes Interpretation
-------------------------------------------------------------------
decimal integer ?+|-0..9? signed 8 64-bit
decimal float ?+|-0..9.NN? signed 8 64-bit double
hexadecimal 0x?0..9|A..F? unsigned 8 64-bit
octal 0o?0..7? unsigned 8 64-bit
binary 0b?0|1? unsigned 8 64-bit
All numbers are evaluated as 64-bit wherever possible. Conversion to
floating point is automatic where necessary for real numbers. Floating point
conversion is handled as an 8 byte double which has a precision of 15
decimal places.
Tcl v9 octal number identifiers are implemented, which require 0oNN format
See: http://wiki.tcl.tk/498
Real numbers are converted on the following format basis:
[whitespace] [sign] [digits] [.digits] [ {d | D | e | E }[sign]digits]
All Tcl numbers are stored internally as strings and converted in real time
Casts are possible to coerce and convert between number types and conversion
routines are available
See also: cast, type, math, expr, eval, functions, ticol
inttooct, intothex, inttobin
result [on integer {script} ?{script...}? ?-default {script}?]
A faster equivalent to switch which may be more efficient in tight loops
which tests showed to use only 80% of the time of an [if] ladder
[on] performs slightly faster than [switch]
Similar to BASIC's 'ON..GOTO' or 'ON GOSUB' statement. The integer value
indexes directly into a series of clauses. At least one clause must be
present
[on] is designed to work only with integer values. Other values, e.g.
characters could be resolved into an integer using say [lindex] or
[strchr] / [strstr]. If double values are presented, they will be
truncated to integer
The index value is a base 1 value which indexes into the series of
arguments with 1 as the first clause, 2 as the 2nd and so-on.
These clauses should be presented individually as with [if] and not
wrapped in a single set of braces
If the index is < 1 then that index value is ignored
If the index is > the highest script case then an error is generated
unless a '-default' case is included
Example:
on [rnd 0 10] {
puts "Case 1"
} {
puts "Case 2" # Highest valid index is 2
} -default {
puts "Case Default"
}
Result:
<Randomly prints "Case 1", "Case 2" or "Case Default">
Example:
on 10 { # Trigger an exception
puts "Case 1"
} {
puts "Case 2" # Highest valid index is 2
} {
Result:
<A catchable exception is raised>
Example Using Non-Integer Values
--------------------------------
set ch "c"
on [strchr "abcde" $ch] {
puts "Letter a was found"
} {
puts "Letter b was found"
} {
puts "Letter c was found"
} -default {
puts "Some other character was found"
}
Letter c was found
Incorrect Syntax
----------------
Note that the script cases MUST be presented individually.
Don't present as follows; i.e. with all clauses wrapped within a
pair of braces...
on [rnd 0 10] {
{ # This brace is incorrect ...
puts "Case 1"
} {
puts "Case 2"
} -default {
puts "Case Default"
} # So is this
}
See also: loop, switch, optimising
handle [open filename mode]
handle [file open filename mode]
Returns an integer handle to the open file if successful, otherwise 0
The mode parameter is a string made up of character flags r, w, a and +
and works in the same way as for C/C++. [try...catch] can be used to detect
exceptions opening invalid files
r read mode
w write mode
a append mode
+ update mode
Example:
set fp [open "filename" r]
close fp
Example:
set fp [open "input.txt" w+]
puts $fp "test"
close $fp
set fp [open "input.txt" r]
set data [read $fp]
puts $data
close $fp
Example:
try {
set fp [open "nosuch.txt" w+]
} catch {
puts "File not found"
}
Result:
File not found
Example:
set filename "licence.txt"
try {
set fp [open $filename "rb"]
} catch {
die "Can't open $filename"
}
set line 0
while {! [file eof $fp]} {
file gets $fp s
puts $s
++ line
}
close $fp
For the Windows "open file" / "save file" common dialogs, See: get_fileopen
get_filesave
See also: file gets, file eof, close, gets
handle [dll_open dllname]
dll_close handle
Loads a DLL by name and returns a handle to the loaded DLL file. This handle
can then be used to call the calldll* series functions
Note that handle variables should be set to zero after a call to dll_close
Example:
set handle [dll_open mydll.dll]
set s [calldllstr handle some_string_function]
dll_close $handle
See also: dll_open, calldll
Optimising [if]...[elseif]...[else] Ladders |
Index |
Very large [if..elseif] ladders can be time consuming when high speed
is required. When processing very large [if] ladders within loops
the entire [if] block is copied, split and parsed for each iteration.
This can be avoided. It is possible to take advantage of Tcl's [eval]
command and array speed to make substantial speed gains with this
type of problem
Define each separate [if] clause in an array element and index this
by subscript matching the if conditions, either by number or string
set some_global_var 0
set if_block("foo") {
puts "in block foo: $some_global_var"
++ some_global_var
}
set if_block("bar") {
puts "in block bar: $some_global_var"
-- some_global_var
}
set if_block("exit") {
puts "* Exiting *"
set done 1 # Will cause the main loop to exit
}
array set cmd { # Create integer to array mappings, N -> foo
0 foo # We could use string to array element mappings
1 bar # or string to integer mappings
3 exit # Exit
}
set done 0
while {! $done} {
##################################################################
# We will iterate, mapping a random integer, but we could just as
# easily map to a sequence of strings parsed from an input file
# If strings can be mapped directly then $cmd() is not required
##################################################################
set op [rnd 0 3] # Simulate varying IF ladder conditions
set code [array item cmd $op ""]
if {ne $code ""} {
eval $if_block($code)
}
}
Results:
Will vary depending on the random sequence in this example
in block bar: 0
in block bar: -1
in block foo: -2
in block bar: -1
* Exiting *
To test against a range of strings, put the matching strings into an array
and return an integer mapping as follows. Then use [array item] with a
default return value of 0. This can rapidly match a very large range of
cases and efficiently default to 'false' if there is no match
option expression off
array set valid { # A potentially huge list of strings to check
foo 1
bar 1
}
proc check {s} {
if {[array item valid $s 0]} {
puts "We have a match on $s"
} else {
puts "No match on $s"
}
}
check "foo" # These checks can be rapidly performed in
check "moo" # a loop
Results:
We have a match on foo
No match on moo
See also: if, on, switch, proc, sub, gosub, goto
option (Runtime Configuration Options) |
Index |
Programmatically set various Ticol configuration options at runtime
Command options given after the 1st argument may appear in any order
option autoexec ?on|off|push|pop|reset?
option breakpoint ?on|off|push|pop|reset?
option echo ?on|off|push|pop|reset?
option escape ?on|off|push|pop|reset?
option expression ?on|off|push|pop|reset?
option preprocessor ?on|off|push|pop|reset?
option stacktrace ?on|off|push|pop|reset?
Note that using Ticol as a manual Windows CLI shell requires that "C"-like
escape characters be correctly processed. Thus if autoexec is ON then
C-escapes will not be interpreted but passed-through literally. Thus paths
such as "\first\second\third" will not encounter an embedded \t tab
character
A script cannot change the current state of the CLI settings. These will
be restored after a [run] command
To use C-like escape sequences at the CLI use the following:
option autoexec off
option escapes on
Autoexec should be OFF when running Tcl scripts. Use [exec] or [spawn] to
launch programs safely
In case of strange behaviour always check the [option expression] setting
first
option
------
option with no arguments displays the current full status:
Subcommand Arg Comment
-----------------------------------------------------------------------
autoexec on: Autoexec unknown (disables CLI C-like escapes if ON)
echo off: Echo command and error results
Should always be set to ON during debugging
escape off: C-like escapes. (Disabled in CLI by autoexec)
expression on: Flow control uses [expr]
preprocessor on: Macro preprocessor
option <item>
-------------
[option <item>] returns the current status in a string as one of either
"OFF" or "ON"
push and pop
------------
The push and pop options save or restore the current setting
push may be combined with on or off. e.g. [option echo push off], or
may be used separately, [option echo push], [option echo off]
This is useful for procedures which may need to alter the global setting
temporarily then restore it to its prior condition on exit
Only one save level is currently supported. Any previous push operations
will be discarded if called a second time. Future versions of Ticol will
allow for multiple depth push/pop.
Example:
option echo push on
# Some code...
option echo pop
Combined commands
-----------------
Combined commands are possible which save and set a setting
Example:
proc foo {} {
option expression push on # Save current state and set to ON
# ... some Tcl code here
option pop # Restore the previous state
}
Command Arg Comment
------------------------------------------
autoexec on: Autoexec unknown
echo off: Echo command and error messages at the console
escape off: C-like escapes
expression off: Flow control uses eval
PreProcessor on: Macro PreProcessor
stacktrace on: Debug error stacktrace (when enabled)
breakpoint on Debug breakpoints (when enabled) via [halt]
option <command>
----------------
Returns the current status as either 'on' or 'off'
e.g. 'option echo' -> 'on'
option autoexec
---------------
Turns the automatic execution of Windows programs on or off
If turned off then this enables the Tcl [unknown] feature
If turned on then this will disable "C" escape characters in the CLI
See: help unknown
option echo
-----------
Turns the Command Line Interpreter (CLI) console echo results on or off
This command has no effect within scripts, where there is no local echo
[option echo off] also disables display of fatal error messages
and you must therefore trap and handle all error cases in order to
generate sensible user feedback
The [catch] command requires option echo off for effective use
If you prefer to see error messages within program output as well as
command output, enable option echo on within the INI configuration file
Without [option echo off] all console output could be echoed twice
option escape
-------------
Turns processing "C"-like escape characters such as \t, \n etc. on or off
If turned on will disable autoexec since backslashes can no longer be
safely passed to Windows
[run] will force [option escape] to ON, [source] leaves this setting
unchanged. Entering the CLI will force this option OFF so using [source]
from the CLI will require correctly setting [option escape] in a script
Thus, any scripts which depend on processing Tcl backslash escape
sequences may not work properly if called using [source] from the CLI
See: help escape
option expression (USE CAREFULLY)
---------------------------------
Turn on or off the implicit use of the [expr] command when evaluating
while, for and if flow-control statements
Caution: When set to "on", [expr] with infix operators will be used instead
of the default, [eval], which uses prefix operators. Incorrectly setting
this option scheme can cause unpredictable behaviour with flow control
commands.
In an interpreter, expression handling is more complex and far-slower
than using native Tcl commands but expressions may be more convenient to
use. Thus a method is offered of switching between the two formats
Strings should be wrapped in double-quotes wherever possible. In particular
strings containing math symbols will be interpreted as a math expression
with option expression on.
Example:
# Examples of bad and good expression presentation
option expression on # Enable 'expression on'
if {file exists test} { # BAD. This is not an expression
if {[file exists test]} { # CORRECT. Command return is evaluated
With [option expression on] expression {== $s 123} would be interpreted as
'if {expr "== $a 123"}' and will therefore fail. This behaviour can be
corrected by using a bracketed argument: if {[== $a 123]} or by toggling
option expression off then on again
[option expression] is less efficient than [eval] (default) due to the
text-processing overheads of the expression-handler. Time-critical code
should use prefix notation and [option expression off]
[option] with the 2nd argument is omitted will display the current status.
With [option expression on] both infix and prefix expression styles may
be used. The overhead of selecting between the two styles is trivial
This option is also configured in ticol.ini [config] as UseEval=<Boolean>
Infix expression example:
[option expression on] - Uses infix [expr] for initialiser and test
option expression on
for {[set i 0]} { $i < 10 } {incr i} { puts $i }
for {set i 0} { $i < 10 } {incr i} { puts $i }
while { $i < 10 } { puts $i; incr i }
if { $i < 10 } { puts $i } elseif { $i > 5 } { break }
Prefix expression example:
[option expression off] - (Default): Used prefix [eval] for init and tests
option expression off
# Note that the test expression differs from [option expression off]
for {set i 0} {< $i 10} {incr i} {puts $i}
while {< $i 10} { puts $i; incr i }
if { < $i 10 } { puts $i } elseif { > $i 5 } { break }
Option expression problems:
option expression off
for {[set i 0]} { $i < 10 } {incr i} { puts $i } # Loop never exits
option expression on
if {file exists some.tcl} {... # FAILS ('file' is tested as a var)
option expression off
if {file exists some.tcl} {... # WORKS OK ('file' is a command)
option expression on
if {[file exists some.tcl]} {... # WORKS OK (context is enforced)
Note that with option expression off, if you use the form below then
[eval] will be called twice. This will slow down performance slightly
option expression off
if {[proc $x]} { ... # [eval] is called twice
option expression off
if {proc $x} { ... # [eval] is called once
See also: flow-control, for, while, if, configuration
option preprocessor
-------------------
Turns on or off the source-code Macro PreProcessor
The default is ON and is also applied to CLI commands
option stacktrace
-----------------
Turns on or off the debugging stacktrace for fatal errors
The default is OFF and this option is also configurable in TICOL.INI
This command is ignored in CGI mode unless forced via ticol.ini
This shows the execution stack-trace when a fatal error is displayed.
The first 256 characters of each command name or parameter will be shown
If option echo is OFF then no stacktrace (or error message) is shown
This can help track down the sequence of errors should a bug occur
Tracing the execution stack will slow down Ticol slightly so should be
enabled only during script debugging
Example:
Execution is proc level1->proc level2->puts (invalid [puts] argument count)
Error stack trace:
puts: Invalid stream handle
In: puts x y z
In: level2 x
In: level1 10
option breakpoint
-----------------
See: option breakpoint
See also: option breakpoint, autoexec
Debugging With option breakpoint, halt, resume |
Index |
option breakpoint
-----------------
option breakpoint ?on|off?
Enables conditional and non-conditional breakpoint handling. Execution will
halt prior to each command call. When a [halt] instruction is encountered
you have the option to single-step to further commands, view the execution
stack, resume execution (turn off breakpoints) or to quit the script
[option breakpoint on] will also enable the stack trace on errors
Debug mode is enabled using: ticol.exe <script> /debug
This command is ignored in CGI mode unless forced via ticol.ini
Example:
option stacktrace on # Enable stacktracing
option breakpoint on # Enable breakpoints
watch add handle # Watch variable $handle
halt # Halt and enter the debugger
set handle [dll_open "sqlite3.dll"] # Code continues...
resume # Resume non-debug execution
halt
----
The [halt] and [resume] commands may be called conditionally within flow-
control statements. [halt] and [resume] only have effect with the /BP
command line argument or with [option breakpoint on], otherwise they
are ignored. [halt] will activate the breakpoint debugger
Example:
if {== $i 10} { halt } # Halt and debug if i is 10
Example (at the CLI):
option breakpoint on
set s "1+2*3"
halt
puts [expr $s]
# Debugger is triggered to single-step execution
Enabling breakpoints will also enable stacktrace which provides breakpoint
trace information
Once breakpoints are activated, execution will stop at the first [halt]
command. You need at least one [halt] command in your script. If breakpoints
are not enabled any [halt] command within code will be ignored. You may have
any number of [halt] instructions in a Ticol script, however you should
remove them in any final release due to a small performance cost
When a [resume] command is encountered any previous [halt] instruction will
be cancelled and normal script execution will resume until another [halt] is
encountered
Breakpoints may be enabled immediately in a script by using
[option breakpoint on] followed by [halt]
The debugger menu will halt with the next pending command shown along with
the current command-count and other options. The [textcolor] command will
be ineffective during debugging
You may also enable breakpoints from the Windows command line using the
argument: ticol.exe /BP
The following menu options are displayed for breakpoints:
Info Shows the current execution stack
Next Steps to the next command without showing the stack
Trace Steps to the next command and shows the execution afterwards
Resume Halts debugging and resumes normal execution until another [halt]
command is encountered, in which case the debugger is re-entered
Vars Show variables in the current execution level
Watch Toggle configured watches on/off
Quit Exits the current code. Returns to the O/S unless launched from
the CLI, in which case control is returned to the CLI
Example:
Breakpoint stack item Near local line 1:
At: / 10 0
At: if == $y $b
error "Error generated by error" "Info for error" 401 [- $line 1]
else
return [/ $y $b]
At: Div x
At: level2 10
At: level1 10
(142): Breakpoint at [/]:Info Trace Next Watch(OFF) Resume Quit
Using Stack Trace to Analyse GPFs
---------------------------------
If they occur, GPFs can be analysed by re-running the script with the
command-line argument /ST to enable stack-tracing. When the GPF is triggered
the stack place will show the command stack and which command triggered the
GPF. This may be useful when using [calldll*] commands. /ST will not trace
into plugin libs
The debugger stack will trace the last 100 events
See also: watch, halt, array, arrays, upvar, global, ticol.ini, debugging
See: help PreProcessor
See also: watch, at_exit, cli, autoexec, unknown, autoexec.tcl
Passing Variables to Procedures by Name (by Reference) |
Index |
Variables may be passed to a procedure by name (often called passing by
reference) within procedures using [upvar]. The variable name is passed
in the proc argument and [upvar] is used to create a reference to the
parent variable. Whilst this passes a variable name it emulates 'passing
by reference' since the name is a reference to a variable in another stack
frame level
Pass by value foo $var # Dereference and pass the value
Pass by reference/name foo var # Pass the variable name
The proc argument should be a different name tag value to the referenced
variable (see example below)
Passing by name offers more flexibility than passing by value
Example:
array set colours {
1 red
2 green
3 blue
4 white
}
proc list_by_key {var} {
puts "list_by_key: Listing $var"
# Upvar is required as all we have here is a name
upvar $var # Upvar is required for pass by name
foreach key [lsort [array names $var]] {
puts "Key: '$key $colours($key)'"
}
}
puts "Listing colours() sorted by key"
list_by_key colours # Pass the variable name 'colours' (reference)
Result:
Listing array colours() sorted by key
list_by_key: Listing colours
Key: '1 red'
Key: '2 green'
Key: '3 blue'
Key: '4 white'
See also: pass by value, proc, variables, passing by value
As Ticol arrays operate via efficient hash tables there wasn't much point
in implementing a separate [hash] command
See: arrays
Passing Variables to Procedures by Value |
Index |
Variables may be passed to a procedure by value within procedures
The variable is de-referenced in the procedure call by prefixing with a
dollar sign. i.e. the variable is dereferenced before the [proc] call
Discussion:
Passing by value has limitations. Complex objects such as stacks, structs
or arrays cannot be passed by value to a proc; these must be passed by
name. Simple objects such as strings, integers or real numbers may be
freely passed by value
The value will be passed via the proc argument variable. This argument
variable must still be de-referenced using a dollar sign
Example:
set s "Hello world"
set i 123456789
set f $pi
proc print_arg {value} {
puts "Argument is: $value"
}
puts "Print argument passed by value"
print_arg $s # Pass the variable by value
Result:
Argument is: Hello world
Argument is: 123456789
Argument is: 3.14159265358979323
See also: calling by name, pass by reference, proc, variables
Console Redirection |
Index |
Redirect to a file in Windows by appending the redirect ">" character
at the end of the command. Any arguments such as /NA must appear before
this character
Example
# List all Ticol commands and redirect into a text file...
# Run this from the Windows console command-line
ticol.exe ; "foreach x [commands] {puts $x}" /na > cmdlist.txt
See also: spawn, exec, ticol.exe
Returning Struct Variables from Procedures |
Index |
Ticol structs were written to offer a binary interface with external DLLs
via the calldll_* library. However they may be used internally and within
Ticol procs. Returning values from Ticol procs via structs is possible but
using them in this way isn't as easy as using them in C/C++ etc.
Discussion:
Within procedures, abstract/complex data types such as struct and stack
require a field de-reference using [struct item] to get a reference to
the field variable from the struct variable. Simple types don't require
this
Outside of procedures and where the struct variable is 'in scope' access
to the variable field member requires [ofaddress struct.member] to
retrieve the struct member address and dereference it back to a variable
reference. This is because Ticol structs emulate C/C++ structs by mapping
Tcl variables to address-offsets in a block of memory space. Each field
member variable has a value which stores the memory address of the field.
Where struct members hold a reference (name) of a variable in another
scope such as root/level 1, it will be necessary to use [set] and the
scoping prefix or [upvar] to manipulate that variable. Where a struct
member field contains a literal variable and is therefore passed by
value, no such complex indirect reference is necessary
Example:
Here, a struct with member field 's.i' is passed in var 's'. s.i contains
the name of a variable in global scope. It is echoed using [set] as the
'$' dereference will be evaluated too late...
puts "struct s.i: '[set ::[struct item $value ${s}.i]]'"
Here, struct 's' member field 's.f' contains a literal double value
puts "struct s.f: '[struct item $value ${s}.f]'"
Variables which reference a struct member in struct.member format will
require the use of curved braces to isolate the struct name so that it
may be dereferenced separately from the trailing field name, i.e.
${struct}.member v's $struct.member
Example Code:
proc ret_val {value} {
upvar $value
textcolor green
if {[eq [type $value] stack]} {
puts "ret_val: Stack: '[stack item $value 0]'" # Show stack top
stack push $value "s1 Stack item return" # Return stack val
} elseif {[eq [type $value] array]} {
textcolor red
puts "ret_val: Array: [array item $value 1]"
array set $value 1 "Ret-array" # Return array item
} elseif {[eq [type $value] struct]} {
textcolor cyan
puts "ret_val: Struct $value item(s):"
array set ::[struct item ${value}.a] 1 "a1"
struct set ::${value}.f "f2"
struct set ::${value}.i "i3"
# ADTs require a field dereference using [struct item]
stack push ::[struct item ${value}.k] "k4"
# Simple variables do not require field dereference
struct set ::${value}.r "r5"
struct set ::${value}.s "s6"
} else {
# Set literal var
puts "ret_val: String or number $value is '[set $value]'"
set $value "sn4"
}
textcolor
}
struct ts { # Test struct: holding various data types
a 20 # Array
f 40 # Double/Float Literal
i 20 # Integer Literal
k 20 # Stack Reference
r 20 # String Reference
s 256 # String Literal * 256 bytes
}
# // ASSIGN //
set i 123456789
struct set ts.i 123456789 # By value
set f $pi # Use const pi
struct set ts.f f # By reference
set s "Hello I am s" # Simple var
set r "Hello string"
struct set ts.r r
struct set ts.s "Hello literal string"
set a(1) "This is array item 1"
struct set ts.a a
stack create stk 200
stack push stk "Hello stack inbound"
struct set ts.k stk
# // CALL //
ret_val s # Pass a variable by name
textcolor yellow
puts "Returned string literal is:\t'$s'"
ret_val r # Pass a variable by name
textcolor yellow
puts "Returned string reference is:\t'$r'"
ret_val i # Pass a variable by name
textcolor yellow
puts "Returned integer literal is:\t'$i'"
ret_val f # Pass a variable by name
textcolor yellow
puts "Returned float literal is:\t'$i'"
ret_val a # Pass an array by name
textcolor yellow
puts "Returned array a(1) is:\t\t'$a(1)'"
ret_val stk # Pass a stack by name
textcolor yellow
puts "Returned stack is:\t\t'[stack pop stk]'"
# // Struct Translation and Returns //
newline
ret_val ts
textcolor yellow
# Structs with name references require [ofaddress]
# Global values which were set can be addressed directly
puts "Returned stack is:\t\t'[ofaddress ts.i]'"
puts "a(1) is '$a(1)'"
puts "a(1) is '[set a(1)]'"
puts "ts.a '[array item [ofaddress ts.a] 1]'"
puts "ts.f '[ofaddress ts.f]' ($ts.f)"
puts "ts.i '[ofaddress ts.i]' ($ts.i)"
puts "ts.k '[stack pop [ofaddress ts.k]]'"
puts "ts.r '[ofaddress ts.r]' ($ts.r)"
puts "ts.s '[ofaddress ts.s]' ($ts.s)"
# END #
See also: pass by value, pass by name
pause ?string? ?-quiet?
Temporarily halt execution with a prompt. The prompt message may be
customised by an argument. The -q argument will suppress prompt output
See also: exit
CGI Application Back-End Configuration |
Index |
Using Ticol Directly
--------------------
Ticol can be used directly as a CGI application. It will not be as fast
or efficient as PHP or PERL. It will automatically detect if running in
CGI mode by checking for the environment variable HTTP_HOST.
If no script argument is given then it will try to serve an HTML man
file called 'ticol.htm' if it is found in the same folder
Configuring Apache Webserver
----------------------------
Ticol TCL can be configure either as a direct CGI application where the
embedded URLs are directed to /cgi-bin/ticol.exe?scriptname.tcl where
ticol.exe and ticol.ini plus any scripts are copied to that folder, or
it is possible to have all files of the type ".tcl" automatically
interpreted by Ticol Tcl if configured as a registered type.
To do this either
a) Use a 'shebang' as the first line of each file as follows:
#!C:/path/to/for/example/cgi-bin/ticol.exe
and the following Apache configuration
# httpd.conf
LoadModule cgi_module modules/mod_cgi.so
ScriptAlias /cgi-bin/ "cgi-bin/"
LoadModule actions_module modules/mod_actions.so
AddType application/x-httpd-tcl .tcl
AddHandler cgi-script .tcl
ScriptInterpreterSource Script
Action application/x-httpd-tcl "cgi-bin/ticol.exe"
# Append index.tcl to DirectoryIndex
DirectoryIndex index.html index.html.var index.tcl
this must come before any other content or comment in the file
or,
b) use the following configuration to avoid the need for a shebang
# httpd.conf (requires a registry modification)
LoadModule cgi_module modules/mod_cgi.so
ScriptAlias /cgi-bin/ "cgi-bin/"
LoadModule actions_module modules/mod_actions.so
AddType application/x-httpd-tcl .tcl
AddHandler cgi-script .tcl
ScriptInterpreterSource Registry
Action application/x-httpd-tcl "cgi-bin/ticol.exe"
# Append index.tcl to DirectoryIndex
DirectoryIndex index.html index.html.var index.tcl
Windows Registry Changes:
; --------------------------------------------------
Windows Registry Editor Version 5.00
[HKEY_CLASSES_ROOT\.tcl\Shell\ExecCGI\Command]
@="C:\\path\\to\\example\\cgi-bin\\ticol.exe"
; --------------------------------------------------
The web server must be restarted after making the above changes
Example HTML file: Save as: "index.tcl"
puts "Content-type: text/html\r\n\r\n"
puts "Hello world"
die
Troubleshooting can be aided by entering the following setting into
ticol.ini which will enable command-logging
; ticol.ini configuration for CGI
[Config]
StartupLog=True ; Command logging for CGI to ticol.log
If you install to other than the standard /cgi-bin/ folder you may need
to configure the directory: Consult the Apache manual for full details
<Directory "C:/path/to/ticol">
AllowOverride None
Options +ExecCGI
Order allow,deny
Allow from all
</Directory>
Finally, ensure you have an antivirus scanning exception for ticol.exe or
you may notice substantial delays (up to 30 seconds) before page load
Using TCX Files
---------------
To use TCX files as web page scripts, duplicate the registry entry and
httpd.conf entries for .tcl files and name as .tcx
# httpd.conf (requires a registry modification)
LoadModule cgi_module modules/mod_cgi.so
ScriptAlias /cgi-bin/ "cgi-bin/"
LoadModule actions_module modules/mod_actions.so
AddType application/x-httpd-tcx .tcx
AddHandler cgi-script .tcx
ScriptInterpreterSource Registry
Action application/x-httpd-tcx "cgi-bin/ticol.exe"
# Append index.tcl to DirectoryIndex
DirectoryIndex index.html index.html.var index.tcl index.tcx
Windows Registry Changes:
; --------------------------------------------------
Windows Registry Editor Version 5.00
[HKEY_CLASSES_ROOT\.tcx\Shell\ExecCGI\Command]
@="C:\\path\\to\\example\\cgi-bin\\ticol.exe"
; --------------------------------------------------
Ticol.ini File Changes for Webserver CGI
----------------------------------------
; File: ticol.ini
; This file should reside in the same folder as ticol.exe
[Config]
StartupLog=False ; Enables command logging for CGI mode
CGIDebug=False ; Enables CGI mode debugging
AutoExecScript=FALSE ; Should be set to FALSE for CGI mode
PreprocessorVerbose=FALSE ; Should be set to FALSE for CGI mode
StackTrace=FALSE ; Should be set to FALSE for CGI mode
Using PHP with Ticol
--------------------
PHP can be used within an HTML web page to launch a Ticol Tcl CGI back-end
script with the aid of some PHP as follows:
HTML code:
-- Snip --------------------------------------------------------
<!DOCTYPE html>
<html>
<body>
<?php
function print_procedure ($arg) {
echo exec("ticol.exe test.tcl");
}
$script_name='test.tcl';
print_procedure($script_name);
?>
</body>
</html>
-- Snip --------------------------------------------------------
Tcl script:
-- Snip --------------------------------------------------------
set a 10
set b 20
set c [expr $a + $b]
return c
-- Snip --------------------------------------------------------
If the Tcl output spans multiple lines it can be captured into a PHP
array as follows:
PHP Code:
-- Snip --------------------------------------------------------
$array = array();
function print_procedure ($arg) {
echo exec("ticol.exe test.tcl",$array);
}
-- Snip --------------------------------------------------------
See also: cgi debugging, html, http, Ticol, ticol.ini
CGI Script Debugging |
Index |
Debugging CGI scripts can be problematic. The main problem with CGI scripts
is that errors may result in a fatal server "500" error and generate no
useful output to aid diagnosis.
However, there are a number of things you can try to locate the source of
any elusive bug
First of all you should look at the server's error.log and access.log files
Add this code as the first line of any CGI script during development to
guarantee errors are visible and don't trigger a "500" error
puts "Content-type: text/html\r\n"
or:
puts "Content-type: text/html\r\n\r\n" -nonewline
or:
printf "Content-type: text/html\r\n\r\n"
Run the CGI script in console mode using the /CD (CGI Debug) argument. Then
place a halt command near the suspected cause and run the script from the
command-line using the /BP (Breakpoint) argument and single-step execution
Enable the following parameters in TICOL.INI during CGI code development
[config]
Debug=True
CGIDebug=True
these values will create additional output and also send early HTTP
headers back to the browser
See also: cgi setup, cgi
pid, tid (Process and Thread ID) |
Index |
integer [pid]
integer [tid]
Return the interpreter's process ID (PID) or thread ID (TID)
Example:
puts "PID is [pid]"
puts "TID is [tid]"
Result:
PID is 616
TID is 33456
See also: info, username
PowerShell has a feature which allows commands to be chained together or
"piped". Tcl does this naturally with nested commands but the resultant
nesting of brackets may be cumbersome if nesting is deep.
Given four commands, foo bar moo and par, Tcl would chain these left to
right as follows, where command or proc [foo] is executed first, followed
by [bar] using the results of [foo], next followed by [moo] using the
result of the previous] commands and so on...
set result [par [moo [bar [foo]]]]
See also: K combinator
Comparing Ticol/Tcl With PHP |
Index |
Tcl is similar in appearance to PHP, Java, Javascript and other C-like
languages in its use of curly braces and with similar language constructs.
However, Tcl is a command-based language which has minimal syntax. This
makes it easily extensible
Ticol has extended the Tcl language to add various useful commands.
Here is a brief summary of the differences between standard Tcl and PHP
Item Tcl PHP Ticol
-----------------------------------------------------------------------
Release date 1988 1994 2015
Easy to learn Yes Yes Yes
Reference variables prefixed with $ Yes Yes Yes
Variables are always prefixed with a $ No Yes No
Braced flow-control constructs Yes Yes Yes
Intended mainly for web-development No Yes No
Complex variable dereferencing No No Yes
Can be byte-precompiled Yes Yes No
Literal strings must be quoted always No Yes No
C-like for statement Yes Yes Yes
Suitable for web CGI application? Yes Yes Yes
Natural expression math format? Yes Yes Yes
(Tcl can use prefix notation: "+ 1 1")
Associative arrays Yes Yes Yes
Standard (downwards) variable inheritance No (upvar) Yes No
Static variables No Yes Yes
New flow control constructs possible? Yes No Yes
Suitable for desktop application? Yes (Tk) No No
Usable as a general IT script language Yes No Yes
Usable as an interactive O/S shell? Yes No Yes
Webserver CGI use on Windows Yes Yes Yes
See also: Ticol, Tcl
play_wav filename?.wav? ?-loop? ?-var v?
stop_wav
Plays a Windows WAV file, optionally looping playback. Playback can be
halted after a fixed period by calling sleep, followed by stop_wav
If play_wav is immediately followed by stop_wav or if the script exits
before the WAV has been played then the file will not be audible
[play_wav] should be followed by a [sleep] to allow the WAV file to play
then a [stop_wav] command
Stored base64-encoded data read from a WAV file by a program such as
base64.exe can be played by allocating to a variable and passing via the
'-var v' argument
Example:
puts "Played WAV? :"[bool [play_wav ".\\sample.wav"]]
sleep 4000 # Allow sufficient time to play
stop_wav # Halt playback
See also: beep
Integer to English Word Conversion Plugin |
Index |
To load, use: lib ticol_numword ?-verbose? ?-nocomplain?
string [integer_to_words integer integer ?-noand? ?-hyphen? ?-zero?]
integer [words_to_integer string]
Converts a signed 64-bit integer to an English word representation and
back from words to a 64-bit integer
Arguments:
-noand Don't include the word 'and'
-hyphen Replace hyphens with spaces
-zero Include leading zeroes
Acceptable values range from:
INT_MIN+1 (-9223372036854775807)
to
INT_MAX (9223372036854775807)
Signed negative numbers will be prefixed with the word, "minus". Zero
values will be represented by the word "zero". If overflow occurs the word
"overflow" is returned without error.
By default, numbers are output in English (UK) format with 'and' separators
between the hundred and tens values. This may be disabled using the '-noand'
argument.
[words_to_integer] will accept either English numeric and magnitude words,
or numeric values or. The use of 'and' is optional and all instances are
ignored. Where numerals are used, they must include the correct magnitude,
for example: "9 2 thousand" does not represent "92 thousand". Each separate
number group will be substituted by the matching English key word, so
"200 20 two" will be evaluated as 222.
Examples:
lib ticol_numwords
puts [integer_to_words 12345]
puts [integer_to_words -1234567]
puts [integer_to_words 9223372036854775807]
newline
puts [words_to_integer {minus nine hundred and eighty six}]
puts [words_to_integer ninety two thousand one hundred and twenty three]
puts [words_to_integer 92 thousand 1 hundred and 23]
puts [words_to_integer [integer_to_words 987654321]]
puts [words_to_integer "1 million 200 20 two thousand 8 hundred and 2"]
lib ticol_numwords -unload
Results:
twelve thousand three hundred and forty five
minus one million two hundred and thirty four thousand five\
hundred and sixty seven
nine quintillion two hundred and twenty three quadrillion three\
hundred and seventy two trillion thirty six billion eight hundred\
and fifty four million seven hundred and seventy five thousand eight\
hundred and seven
-986
92123
92123
987654321
1222802
Where multiple words of the same magnitude are encountered, these will be
added (summed)
Example:
puts [words_to_integer twenty three twelve] # Add 23+12
puts [words_to_integer one two three four five] # Add 1+2+3+4+5
puts [words_to_integer one hundred twenty three forty two] # Add 123+42
Result:
35
15
165
QEMU Notes
----------
Some problems were experienced during testing with FP math on early
versions of QEMU emulator and NT4.0. Since it is unlikely that anyone
will be using this combination the problem has not been addressed.
The cause may be in the FPU emulation on some versions of QEMU
See also; plugins
To load, use: lib ticol_ping ?-verbose? ?-nocomplain?
string [ping hostname ?timeout-ms? ?-list?]
string [ping ipv4-address ?timeout-ms? ?-list?]
string [ip_from_host hostname]
Can ping either a hostname or an IPV4 address. The default timeout period is
4000 ms but this may be specified as an optional 2nd argument.
By default the IPV4 address of the pinged node is returned but a list may
be returned by appending the -list argument. In which case a list containing
the IPV4 host address and round-trip response time in ms will be returned.
All versions of Windows are supported from Windows NT4/2K to Server 2016
ReactOS is not currently supported
Example:
set hostname "localhost"
set i 1;
while {$i < 11} {
set r [ping $hostname -list]
printf "%3i\t%s\t%i ms\r\n" $i [lindex $r 0] [lindex $r 1]
incr i
sleep 1000
}
The command [ip_from_host] will perform a hostname DNS lookup to return
the IPV4 address
Example:
puts [ip_from_host www.google.com]
Result:
123.234.123.234
Example:
puts [ip_from_host www.google.com -list]
Result:
123.234.123.234 301
See also: plugins, lib
string [pipe <expr> ?, <expr>?...]
[pipe] strings together a series of statements or commands without
requiring complex or deeply nested brackets. Most Tcl commands will
'chain' or 'pipe' to to each other but may require nested brackets
A comma is used as the expression (or command) separator. This works
in a similar fashion to PowerShell's | pipe separator.
Expressions are evaluated or 'piped' left-to-right with no precedence.
Example:
# Left -> to -> right evaluation
# 2*3 = 6
# 6/1.5 = 4
# 4*atan(1) = PI
puts [pipe 2 , * 3 , / 1.5 , * [funct atan 1]]
Result:
3.141592653589792
Where a [proc] receives [pipe] arguments or participates in a pipe chain
it should use the following definition
# Arg 1: arg is the chained accumulator (always the first argument)
# Arg 2: op is an optional operator
# Arg 3: val is an optional trailing value (args may also be used)
proc foo {arg} {
return [* $arg 23.0] # Return in Tcl eval order (op arg val)
}
proc bar {arg op} {
return [$op $arg 5] # Return in Tcl eval order (op arg val)
}
proc moo {arg op val} {
return [$op $arg $val] # Return in Tcl eval order (op arg val)
}
# Usage is equivalent to: [moo [bar [foo 2] +] / $pi]
# foo arg
# bar arg op
# moo arg op val
#
puts [moo [bar [foo 2] +] / $pi]
puts [pipe 2, foo, bar +, moo / $pi]
Result:
16.233804195373324
16.233804195373324
See also: dot, expr, eval, math
Plugins Extensions for Ticol Tcl |
Index |
Plugins are external modules which can be loaded to add additional
functionality to Ticol. The capabilities of Ticol Tcl can be expanded
by the addition of external DLL routines as linkable plugins
A plugin is loaded using:
integer [lib <plugin-name>?.dll? ?-verbose? ?-nocomplain?]
which returns a count of the number of exported commands successfully
loaded into the interpreter or zero on load failure. The return value
may be ignored
Plugin info is returned using
string [lib <plugin-name> -info]
See: help lib for more information
A small number of plugins are supplied. Most of these were developed
for personal use and for specific projects so may contain only the
initially required functionality. e.g. ticol_json was designed to
parse only HASP licensing data.
Some of the plugins are new, experimental or feature-incomplete:
Title Lib DLL Description Help Topic
-------------------------------------------------------------------------
ANSI ticol_ansi.dll ANSI text output See:ansi
ARP ticol_arp.dll ARP and RARP commands See:arp
ASM ticol_asm.dll Pseudo assembler See:asm
Big Numbers ticol_bignum.dll Arbitrary precision math See:big math
Big Text ticol_bigtext.dll Large Console Text See:big text
Brainfuck ticol_bf.dll Brainfuck interpreter See:bf
Bitmap ticol_bmp.dll Bitmap encoding See:bmp
Call DLL ticol_calldll.dll Call an external DLL See:calldll
CArray ticol_carray.dll C-Like fixed arrays See:carray
CGI ticol_cgi.dll Basic CGI interface See:cgi
CPU ticol_cpu.dll Basic CPU information See:cpu
CRC ticol_crc.dll CRC calculation See:crc
CScript ticol_cscript.dll VBScript/CScript See:cscript
Database ticol_db.dll Random access database See:db
DHCP ticol_dhcp.dll DHCP commands See:dhcp
DNS ticol_dns.dll DNS lookup See:dns
Event Logging ticol_event.dll Windows Event Logging See:event
FTP ticol_ftp.dll A simple FTP client See:ftp
Hasp ticol_hasp.dll Hasp Key querying See:hasp
Highlight ticol_highlight.dll HTML code highlighting See:highlight
HTML ticol_html.dll HTML code processing See:html
INI ticol_ini.dll INI file handling See:ini
JSON ticol_json.dll Convert JSON to Tcl list See:json_to_list
Line Get ticol_line.dll Retrieve lines from text See:ticol_line
Linked List ticol_list.dll Binary Linked Lists See:linkedlist
Map Drive ticol_mapdrive.dll Drive Letter Mapping See:mapdrive
Miscellaneous ticol_misc.dll Misc routines See:misc
NumWords ticol_numword.dll Convert integer to words See:numwords
Ping ticol_ping.dll Simple ICMP ping See:ping
R2D ticol_r2d.dll Roman to decimal convert See:r2d
REG ticol_reg.dll Simple registry services See:reg
RLM ticol_rlm.dll RLM licence querying See:rlm
RTF ticol_rtf.dll Simple RTF text output See:rtf
Service ticol_svc.dll Windows service query See:service
Sci Numbers ticol_sci.dll Scientific format number See:sci
SMTP Email ticol_mail.dll SMTP emailer See:mail
SockPing ticol_sockping.dll Socket ping a host See:sockping
SQLite ticol_sqlite.dll SQLite 3 plugin See:sql
TCX ticol_tcx.dll Embedded TCX handling See:tcx
TicASM ticol_asm.dll Assembler interpreter See:ticol_asm
Variant ticol_variant.dll Variants for calldll See:variant
VB ticol_vb.dll Some VB-like commands See:vb
Windows GDI ticol_win.dll Windows GDI graphics See:win
Wordcount ticol_wordcount.dll Word or char count See:wordcount
XBase ticol_xbase.dll Dbase 3/4 Database See:xbase
ZIP ticol_zip.dll Simple ZIP package See:zip
Tests for these plugins are included with the unit test scripts
See: lib interface - for information about how to write plugins in C/C++
See also: lib, Ticol, lib interface
Bitmap Encoding Plugin |
Index |
A plugin to encode text in Windows Bitmap (BMP) format
To load, use: lib ticol_bmp ?-verbose? ?-nocomplain?
Command and Syntax
bool [bmp write filename string|var ?-var? ?-fore rgb? ?-back rgb?\
?-encrypt?]
bool [bmp read filename newvar ?-nocomplain?]
list [bmp info filename]
void [bmp graph filename ?-w N? ?-r? ?-crop? ?-asterisk? ?-bytes N?\
?-fore colour_code? ?-back colour_code?]
void [bmp graph -var name ?-w N? ?-r? ?-crop? ?-asterisk? ?-bytes N?\
?-fore colour_code? ?-back colour_code?]
A crude form of steganography. Converts text-based data in a string or
variable into Windows BMP format file, which allows the data to be sent
via email or other messaging systems without tampering. The data is not
encrypted but [encrypt] can be used to encode data passed to [bmp]
[bmp read] will read back the original file text to a variable
bmpconv.exe can also be used to convert back to the original file
[bmp write] is binary-compatible. If a file contains binary data, use -var
and pass a [bmp read] binary variable name. The binary data can be verified
using [md5 varname -var]. The resulting file must not be edited using a
graphics editor as BMP 'row-padding' is not followed
[bmp write -encrypt] will perform basic level symmetric encryption on
the data payload. Decryption using [bmp read] is automatic as long as
the file is not edited
[bmp] encryption is low-strength and serves only to obfuscate (obscure)
the data contents. You can combine [bmp -encrypt] with the more secure
[encrypt] command or call an external encryption API using [calldll]
[bmp graph] will show a graphic dump of pixel data, optionally in reverse
(-r) order. Note that BMP images typically store the data 'upside-down'
Red symbolises 1, grey 0. Reverse (-r) shows the data as stored, otherwise
how it is visible in a paint package is shown. -w width bytes can be shown
per line. This command does not produce decoded visual representations of
general graphic bitmaps, just the bit sequence of an encoded file
If -bytes N is specified then display is capped to N bytes, otherwise
all bytes are displayed
[bmp graph] can read data either from file (default) or from a variable
by using the ?-var varname? argument
The foreground and background image colours should be specified as RGB
values in the format RRGGBB. Hex is usually the best way to do this:
0x000000 # (black)
0xFFFFFF # (white)
0xFF0000 # (red)
0x00FF00 # (green)
0x0000FF # (blue)
[bmp info] returns a Tcl list containing information about a file as
follows:
filesize-bytes data-block-size width height colours bits-per-plane
bytes-per-pixel padding-space
Example:
bmp write qbf.bmp $qbf -fore 0xFF0000 -back 0x0
@ start ticol_qbf.bmp
bmp read qbf.bmp s
puts $s
Result:
Displays bitmap in associated paint package
The quick brown fox jumps over the lazy dog
Example:
randomstr 4096 5 -var s -nocomplain # Set some binary data
puts "Starting data MD5: [md5 s -var]"
bmp write "randomstr.bmp" s -var # Write binary
unset s # Clear s
bmp read randomstr.bmp s # Read binary back to s
puts "\[binary length s\]=[binary length s]"
puts "Read-in data MD5: [md5 s -var]"
Result:
Starting data MD5: b46bf481af4fee5b08a9d3d76c135301
[binary length s]=4096
Read-in data MD5: b46bf481af4fee5b08a9d3d76c135301
Example: (external obfuscation)
bmp write test.bmp [encrypt $qbf secret] -fore 0xff0000 -back 0x0
bmp read test.bmp s
puts [decrypt $s secret]
Result:
The quick brown fox jumps over the lazy dog
Example: (internal obfuscation)
bmp write test.bmp $qbf -encrypt -fore 0xff0000 -back 0x0
bmp read test.bmp in
puts $in
assert $in {eq $_ $::qbf} -v -line #__LINE__
Result:
The quick brown fox jumps over the lazy dog
assert: Line 4: PASS 'eq $_ $::qbf'
Displaying Bitmap Images
------------------------
Although intended for debugging display of encoded data, small,2-colour
graphic bitmap images may be displayed at the console using [bmp graph],
The '-crop' argument will force line breaks on and -asterisk can be used
to produce printable output which may also be redirected to a file.
Otherwise, console graphics characters will be used
-asterisk may be used to draw using * characters
The background character will be shown as 75% shaded unless -solid or
-asterisk is used
Console colours can be set using standard Ticol (Windows console) screen
colour integers or colour names using -fore and -back
The console width must be set large enough to avoid wrapping at the
configured font size
This feature could be used to add a graphic banner for a Tcl or batch
script.
Example:
# hello.bmp is 60 x 20 px / 2-colours
lib ticol_bmp
bmp graph hello.bmp -crop -asterisk
Result:
************************************************************
**** ******** *************** **** ***************
***** *********** **************** **** ****************
***** *********** **************** **** ****************
***** *********** **************** **** ****************
***** *********** **************** **** ****************
***** *********** **************** **** ****************
***** *********** **************** **** ****************
***** *********** ******* ***** **** ****** ******
***** ****** ** **** **** **** *** *****
***** *********** ****** *** **** **** **** **** ****
***** *********** ***** *** **** **** *** ***** ****
***** *********** ***** ** ****** **** *** ***** ****
***** *********** ***** ******* **** *** ***** ****
***** *********** ***** ********* **** *** ***** ****
***** *********** ***** ***** ** **** *** **** *****
**** *********** ****** *** **** **** *** *****
**** ********** ****** **** ** *** *******
************************************ ***** *****************
************************************************************
Displaying Internally Stored BMP Images
---------------------------------------
Example:
# Generate using base64.exe
set bmp "Qk3eAAAAAAAAAD4AAAAoAAAAPAAAABQAAAABAAEAAAAAAKAAA
AASCwAAEgsAAAIAAAACAAAAAAAAAP///wAAAAAAAAAAAAAAAAAAAEAAAEA
DB8HgeAAAPAMP4xhgAAAwAx4nCGAAADADHBYMYAAAMAMYBgxgAAAwAxgGD
GAAADADGAYMYAAAMAMIAhxgAAAwDwxjGGAAADADA+DwYAAAMAAAAABgAAA
wAAAAAGAAADAAAAAAYAAAMAMAAABgAAAwAwAAAGAAEDAAAAAAYAAf/+AAA
AAwAAAAAAAAAAAA"
# Display from var 'bmp', crop and rotate from stored orientation
bmp graph -var bmp -crop -r
Result:
Displays the Ticol logo
See also: plugins
Network Drive Map Plugin |
Index |
A plugin to map and unmap network drive letters
To load, use: lib ticol_mapdrive ?-verbose? ?-nocomplain?
Command and Syntax
bool [mapdrive letter?:? unc-share ?-user u? ?-pass? p ?-persistent? \
?-prompt? ?-delete?]
string [mapdrive * unc-share ?-user u? ?-pass? p ?-persistent? \
?-prompt? ?-delete?]
bool [mapdrive *|letter?:? -delete ?-force?]
Maps a drive letter to a UNC share via the Windows mpr.dll API interface
[mapdrive letter] will return boolean 1 or 0
[mapdrive *] will return either a drive letter as "x:" on success or ""
on failure
The share must exist and be in the format:
\\server\share-name?\path...? # String must be escaped
or
//server/share-name?/path...? # Preferred format
A colon character is optional when specifying a drive-letter
Backslash and Tcl reserved characters must be escaped unless calling from
the Ticol CLI command prompt
Unix-style forward slash path separators are supported
Examples:
# Where the highest drive allocated before running is O:
puts [mapdrive * \\\\snoopy\\c\$] # Note the escaped $
puts [mapdrive * //snoopy/c$]
puts [mapdrive r: \\\\snoopy\\c\$] # Note the escaped $
puts [mapdrive s: //snoopy/c$]
puts [mapdrive * //snoopy/c[chr 36]] # Avoiding literal $
Results:
p:
q:
1
1
r:
Errors
------
Use [win_error] to determine the cause of any error
See also: plugins
Miscellaneous Routines Plugin |
Index |
A plugin containing miscellaneous useful routines
To load, use: lib ticol_misc ?-verbose? ?-nocomplain?
Command and Syntax # Comments
ip_match mask ip-address # Match an IP address to a mask (wildcards ok)
# Returns a boolean 1/0
ip_filter list ip-address # Match an IP address to a list of addresses
cidr_match mask ip-address # Match an IP address to a CIDR mask ?.?.?.?/N
# address / bitmask
# Returns a boolean 1/0
# e.g. [cidr_match 192.168.3.0/24 192.168
is_mac mac-address ?-f? # Checks if string is a formatted MAC address
# Returns a boolean 1/0
is_mouse_present # Returns a boolean 1/0
is_networked # Returns a boolean 1/0
is_safe_mode # Returns a boolean 1/0
is_vhdboot # 1 if booted from VHD (Win8/Server 2012+ only)
# Use [info winver -vernum] > 6.1 to control
monitorcount # Returns a count of connected monitors
rot13 string # Perform rot-13 transformation on string
# and return the result (source is unchanged)
sysmetrics N # Call GetSystemMetrics API with argument N
# Returns an integer value
# See an API guide for calling values+returns
uptime # Uptime in seconds since last boot (Win XP+)
volume serial X: # Return the serial of a disk as a hex value
volume filesystem X: # Return the file system type of a disk volume
wallpaper get # Get the path+filename of the wallpaper
wallpaper set file ?align? # Set the current wallpaper with alignment...
# -centre, -tile, -stretch (returns a boolean)
Examples:
IP match can use wildcards * or ? to represent octets or individual digits
octets may be omitted from the right-hand end of the IP address mask
puts [ip_match 192.168.3.* 192.168.3.1]
puts [ip_match 192.168.3.* 192.168.101.1]
puts [ip_match 10.* 10.11.12.13]
Results:
1
0
1
Example:
CIDR (Classless Inter-Domain Routing)
[cidr_match] requires a full IP address with a bitmask value /N where
N specifies significant bits from the left hand end of the IP address
puts [cidr_match 192.0.0.0/8 192.168.101.1]
puts [cidr_match 192.168.3.0/24 192.168.101.1]
Result:
1
0
Example:
puts [is_mac abcdefgh] # Too short + invalid characters
puts [is_mac 001122aabbcc] # Valid unformatted MAC
puts [is_mac 00-11-22-aa-bb-cc -f] # Check valid formatted MAC
Result:
0
1
1
Example:
puts [rot13 "Hello world"]
Result:
Uryyb jbeyq
Example:
puts [volume serial c:]
Result:
A1B2C3D4
Example:
puts [volume filesystem c:]
Result:
NTFS
Example:
set bt [- [now] [/ [uptime] [expr 24*3660.0]]]
set ut [uptime]
puts "Uptime is: $ut seconds"
puts "Boot time: $bt"
puts "\[now\] is [now]"
Results:
Uptime is: 183206 seconds
Boot time: 43089.035953438070000
[now] is 43091.121631944443000
assert: Line 82: PASS '> $_ 0'
Boot date+time: 2017/12/21 15:08:46
See also: long_to_ip, ip_to_long, Plugins, Windows Version Numbers
Registry Plugin (ANSI) |
Index |
To load, use: lib ticol_reg ?-verbose? ?-nocomplain?
Path/Key arguments must be supplied in Tcl "backslash-escaped" format.
The result (if string or list) will be returned in the same "backslash-
escaped" format. Only ANSI strings are supported by [reg]. Unicode/MBCS
strings are not supported
Supported in Windows 2k/XP or higher
Syntax:
reg <command> path ?value? ?-type?
bool [reg set]
string [reg get]
bool [reg del]
regtype [reg type]
For [reg type] the return will be a string equivalent of one of the
following keywords
enum {
REG_NONE
REG_SZ
REG_EXPAND_SZ
REG_BINARY
REG_DWORD
REG_DWORD_BIG_ENDIAN
REG_LINK
REG_MULTI_SZ
REG_RESOURCE_LIST
REG_QWORD
REG_UNKNOWN;
}
Use either 'reg' or 'registry'
reg get "HKEY_VALUE\\path" keyname ?-binary var? ?-type?
for [reg get] the default type is -auto
reg set "HKEY_VALUE\\path" keyname value ?-binary var? ?-type?
or [reg set] the default type is -string
reg del "HKEY_VALUE\\path" keyname ?value?
if ?value? is omitted then the entire subkey and values are deleted
reg type "HKEY_VALUE\\path" keyname
Where the optional '-type' argument is one of:
-auto (default)
-dword Store a 32-bit HEX value
-qword Store a 64-bit HEX value
-list Decompose/reassemble a Tcl list (REG_MULTI_SZ)
-multiline Stores/retrieves CRLF delimited text as REG_MULTI_SZ
-string Stores/retrieves a Tcl list as REG_SZ
The type argument must be the last item in the command
Binary strings cannot be passed via the Tcl command line. Instead, for
[reg set] and [reg get] an argument reference to the variable-name is
passed using "-binary varname" and any preceding variable argument is
ignored
If DWORD or QWORD values are supplied with HEX values, they must include
a leading 0x - e.g. 0xffffffff
Example:
puts [bool [reg set HKEY_CURRENT_USER\\Console ColorTable06 \
0x00508dfc -dword]
puts [inttohex [reg get HKEY_CURRENT_USER\\Console \
ColorTable06 -dword]]
puts [reg get HKEY_CURRENT_USER\\Console ColorTable06 -dword]]
Result:
True
0x508dfc
5279228
Where types are mixed between binary and non-binary, strings will be
truncated at the first null character
Example: (Sets partial string)
binary set a \xfe\xffhello\x21\x00\x22world\x02\xff\xfe -length 16
reg set HKEY_CURRENT_USER\\Software\\Ticol BINDATA -binary a
reg get HKEY_CURRENT_USER\\Software\\Ticol BINDATA -binary b
puts [binary tohex b]
puts [reg get HKEY_CURRENT_USER\\Software\\Ticol BINDATA]
Result:
feff68656c6c6f210022776f726c6402
¦ hello!
Note that -multiline and -string affect the interpretation of sent/returned
Tcl data. Both types are stored as REG_MULTI_SZ but will be encoded or
decoded as appropriate. You can store in Tcl list format and retrieve as
a CRLF-delimited multi-line string (and vice-versa)
Example: (Using line-wrap, where a space is added after the \)
puts "Country: '[reg get {HKEY_CURRENT_USER\\Control\
Panel\\international} sCountry]'"
Result:
'United Kingdom'
Example: (from https://www.tcl.tk/man/tcl8.4/TclCmd/registry.htm#M9)
set ext .tcl
# Read the type name
set type [registry get HKEY_CLASSES_ROOT\\$ext {}]
# Work out where to look for the command
set path HKEY_CLASSES_ROOT\\$type\\Shell\\Open\\command
# Read the command
set command [registry get $path {}]
puts "Filetype \"$ext\" opens with $command"
Example:
reg set HKEY_CURRENT_USER\\Software\\Ticol TestValue "Hello world" -string
puts [reg get HKEY_CURRENT_USER\\Software\\Ticol TestValue]
puts [reg type HKEY_CURRENT_USER\\Software\\Ticol TestValue]
puts [bool [reg del HKEY_CURRENT_USER\\Software\\Ticol TestValue]]
Results:
Hello world
REG_SZ
true
See also: plugins, calldll
A plugin which offers rudimentary ANSI SQLite 3 database capability
To load, use: lib ticol_sql ?-verbose? ?-nocomplain?
handle [sql open database_file]
variant [sql query handle querystring]
sql close handle
sql free variant|variant-address
To run a query you must first use [sql open] to obtain a database handle
This handle is passed to [sql query]. You should close the handle using
[sql close] after you have finished querying the database
SQL queries return a VARIANT object. It is vital that this is freed by a
call to [sql free] using either the dereferenced variable ($var) or name
(var) and this must be called before the variable goes out of scope.
Do NOT use [varray unset var] to release structures returned by this
plugin (see example below)
The returned VARIANT table may be manipulated using the [varray] commands
If a database file does not exist then [sql open] will create it
Example:
set handle [sql open c:\\path\\to\\my.db]
Example;
set v [sql query $handle "SELECT * from Customers"]
Example:
sql free v
Example:
sql close $handle
Full Example:
lib ticol_sqlite.dll
lib ticol_varray.dll
newline
set handle [sql open ".\\albums.db"]
if { $handle } { # Null is not returned for file not found
set v [sql query $handle "SELECT * from Artists"]
set rows [varray size $v]
set cols [varray size $v 2]
for {set j 0} {< $j $rows} {incr j} {
for {set i 0} {< $i $cols} {incr i} {
puts "[varray get v $j $i]\t" -nonewline
}
newline
}
newline
puts "Retrieved table has $rows row(s) and $cols col(s)"
sql free v # Ensure [sql free] is used on the query
sql close $handle
} else {
die "Can't open SQL database"
}
Result:
ArtistID ArtistName
1 Puddle of Mudd
2 Frank Zappa
3 Love/Hate
4 Cheap Trick
5 Megadeth
6 Foo Fighters
Retrieved table has 183 row(s) including header, and 2 column(s)
See also: varray, plugins
To load, use: lib ticol_crc ?-verbose? ?-nocomplain?
value [crc value ?-hex? ?-noprefix?]
value [file_crc filename ?-hex? ?-noprefix?]
Extends [file] by adding: [file crc filename] into the namespace
Calculates the 32-bit CRC checksum of an ANSI string or file and is tested
on files up to 10Gb in size
By default a decimal value is returned, prefixed with "0x" unless -noprefix
is used. A hex value can be returned by appending the argument -decimal
[file_crc] will calculate the CRC for a file. This is quite efficient even
on fairly large files of up to 1Gb. The maximum file size is 4Gb as this
is the largest which may be accommodated by the 32-bit integer range
Example:
puts [crc $qbf]
Result:
0x414fa339
Example:
puts [file_crc VS97PRO3.ISO]
Result:
0x22ffdfcc
See also: plugins
CScript - VBScript Console Link Plugin |
Index |
To load, use: lib ticol_cscript ?-verbose? ?-nocomplain?
void [cscript script-string ?-echo? ?-nocomplain?]
Offers basic level calling of the BASIC script language from the
Windows scripting VBScript host which can allow a number of scriptable
objects to be called including, for example, WMI or registered
applications such as MS Word. This plugin may be useful where the use
of an intermediate script file is not desired.
At present, it is not possible to return a value directly to Ticol as
VBScript, other than for 'Expr()' is not designed to pass results back
to a caller under normal circumstances. Output could be redirected
to a file or database.
To run a script file, load the file using [readfile]
The script is converted to Unicode before running but will be handled
by Ticol as ANSI. '-echo' may be used to echo the Unicode script
Tcl strings containing escaped characters and in particular, escaped
double quotes or paths with backslash characters, MUST be escaped using
[unescape] before passing to [cscript]. This will translate escaped
double-quotes embedded within strings back to literals
A non-Tcl script read directly from file may not require to be escaped.
VBScript code called by [cscript] is currently run in the main thread
The '-nocomplain' allows errors to be ignored. The script call will
return with no error if any exception occurs, otherwise the error will
be fatal and halt the script, displaying the error number and message.
In some circumstances it may be better to use [spawn] or [exec] to
call the WSH directly via cscript.exe or wscript.exe as this will allow
capture and return of text output back to a Ticol variable.
Example:
lib ticol_cscript
set s "msgbox \"Hello World\"" # String contains escaped double-quotes
cscript [unescape $s] # so we must ensure these are unescaped
lib ticol_cscript -unload
Result:
<A MsgBox displays 'Hello world'>
Console Output
--------------
By default, VBScript passes printed output to a messagebox to display.
The following scriptlet may be used in order to provide console output
using stdout.WriteLn:
Set fso = CreateObject ("Scripting.FileSystemObject")
Set stdout = fso.GetStandardStream (1)
Set stderr = fso.GetStandardStream (2)
Example:
set s "\
Set fso = CreateObject (\"Scripting.FileSystemObject\")
Set stdout = fso.GetStandardStream (1)
stdout.WriteLine \"Hello world\""
lib ticol_cscript
cscript [unescape $s]
lib ticol_cscript -unload
Result:
Hello world
Example:
# Note that Tcl requires embedded dquotes to be escaped in source
# We must use [unescape] to pass to VBScript
set s "
Set fso = CreateObject (\"Scripting.FileSystemObject\")
Set stdout = fso.GetStandardStream (1)
strComputer = \".\"
Set objWMIService = GetObject(\"winmgmts:\\\\.\\root\\cimv2\")
Set cols = objWMIService.ExecQuery(\"SELECT * FROM Win32_ComputerSystem\")
For Each objItem In cols
stdout.WriteLine \"Name: \" & objItem.Name
stdout.WriteLine \"Manufacturer: \" & objItem.Manufacturer
stdout.WriteLine \"Model: \" & objItem.Model
Next"
lib ticol_cscript
cscript [unescape $s]
lib ticol_cscript -unload
Enabling WScript (WSH)
----------------------
The Windows Scripting Host (WSH) is installed by default in most
versions of Windows. It may be disabled for security reasons.
If you wish to re-enable and are having problems accessing the
WScript object, try this registry setting
Windows Registry Editor Version 5.00
[HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Windows Script Host\Settings]
"Enabled"=dword:00000001
Available Objects
-----------------
The following objects (ProgIDs) are examples of those known to be
available to use with VBS -> CreateObject()
(depending on 3rd-party application installation in some cases);
InternetExplorer.Application
Scripting.Dictionary
Scripting.FileSystemObject
Shell.Application
System.Collections.ArrayList
System.Collections.Queue
System.Collections.SortedList
WbemScripting.SWbemLocator
Word.Application
Word.Window
Unavailable Objects
-------------------
The following objects (ProgIDs) are not available:
Scripting.* Other than listed above
WScript (Root) e.g. Wscript.Version
Wscript methods are not inherited when launched as a
VBScript engine. The common reason to include is to
gain access to console printing using WScript.Echo
https://blogs.msdn.microsoft.com/ericlippert/2003/10/08/why-
cant-i-create-the-wscript-object/
Platforms
---------
Should work on all Windows platforms which have WSH installed
from NT4/SP6a upwards. Does not work on ReactOS
CGI Use
-------
[flush] should be called before calling [cscript], particularly in
order to flush any HTTP headers sent using [puts]. [cscript]
cannot flush pending console output.
# CGI Server Script
# Where variable 's' contains valid VBScript code
puts "Content-type: text/html\r\n\r\n" -nonewline
flush
lib ticol_cscript
cscript [unescape $s]
lib ticol_cscript -unload
Recommended IDE
---------------
VBSEdit
https://www.vbsedit.com
Alternative Solutions
---------------------
A more flexible alternative is to use [spawn] with cscript.exe as this
will offer the full range of commands including WScript.*
Windows cscript.exe accepts only script files. Where direct commands are
preferred rather than an intermediate script files, the author's VBS.EXE
may be used together with [spawn]
Example:
unset x -nocomplain
spawn x cscript.exe //NoLogo myscript.vbs
puts $x
spawn x vbs.exe /w Console.WriteLine \"Hello world\"
spawn x vbs.exe /w Console.WriteLine 4*atn(1)
More Information
----------------
ActiveX objects may be viewed using the NirSoft ActiveXHelper utility
http://www.nirsoft.net/utils/axhelper.html
The VBScript is documented here:
https://docs.microsoft.com/en-us/previous-versions/windows/
internet-explorer/ie-developer/scripting-articles/d1wf56tt(v=vs.84)
https://www.w3schools.com/asp/asp_ref_vbscript_functions.asp
See also: spawn, plugins
Random File Database Plugin |
Index |
To load, use: lib ticol_db ?-verbose? ?-nocomplain?
bool [db create filename {list of fields} ?comment? ?-y?]
handle [db open filename ?mode?] # Open a database
string [db time handle] # Last write update time
long [db count handle] # Count of records written
string [db get handle N] # Get record. Record is base 1
bool [db put handle ?record?] # Write record. Record is base 1
string [db field handle N] # Field data. Field is base 0
short [db fieldcount handle] # Count of fields in database
string [db fieldname handle N] # Field name. Field is base 0
long [db fieldsize handle N] # Field size. Field is base 0
short [db comment handle string] # Add a database comment
bool [db close handle] # Close the database
string [db info handle -list] # Get database info as Tcl list
bool [db is_db handle] # Validates a file as Ticol DB
void [db info handle] # Get database info
Offers XBase/Dbase III like random-file capabilities with a similar
command-set. Unlike the XBase plugin this is not limited to restrictive
field lengths. DBase is limited to a field length maximum of
<unsigned char> or 254 bytes per field
A field in the 'db' plugin may be any integer value up to <signed long>
bytes although, in practice it is strongly recommended that field lengths
be kept to under say 1,000 or so bytes per field. The enhanced storage
capacity makes it feasible to store whole web pages or scripts as database
records. XBase and DB files may be used together
Record offsets are a 'base 1' value. Field offsets are a 'base 0' value
That is, the first record is record '1'. The first field is field '0'
Fields are treated as an array and array indices (as we all know) start
at '0'. All field types are 'character' and the DB plugin performs no
data-validation. Binary data must be ASCII encoded (See: [base64] and
[tohex])
[db create] creates an opens a new DB file. It requires a Tcl list
argument in the form of a series of tuples as {name length}.
Example: db create $filename {{Name 20} {Address 20}}
Example: db create $filename {{Name 20}
{Address 20}
}
Example:
set handle [db create $filename {{Name 20} {Address 20} {Phone 20}} \
"This is a comment" -y]
db set $handle 0 "Biffa Bacon"
db set $handle 1 "49 Letsby Avenue"
db set $handle 2 "(0123) 4567890"
db put $handle 1 # Write record 1
db set $handle 0 "Fatha Bacon"
db set $handle 1 "49 Letsby Avenue"
db set $handle 2 "(0976) 543210"
db put $handle 2 # Write record 2
db get $handle 1 # Read record 1
puts "Field 0: '[db field $handle 0]'"
puts "Field 1: '[db field $handle 1]'"
puts "Field 2: '[db field $handle 2]'"
db close $handle
Results:
Field 0: 'Biffa Bacon'
Field 1: '49 Letsby Avenue'
Field 2: '(0123) 4567890'
Database Structure (in C/C++)
-----------------------------
+----------------------------+
| Ticol DB Record Structure |
+============================+ <- 0
| Header prolog |
| 276 bytes |
+----------------------------+ <- 272
| Header suffix * N records |
| field_count * DB_RECORD |
+----------------------------+ <- 272 + (field count * sizeof(DB_RECORD))
| Field data... |
| record_size * record_count |
+----------------------------+
_DB_HEADER_TAG="DBX1" // 4 bytes header ID tag
const _DB_COMMENT_MAX=200; // A maximum of 200 character comment
typedef struct DB_HEADER_TYPE { // +- Header Prolog --------------------+
char tag[4]; // | 4 File type ID tag (magic number)|
long version; // | 4 > 0 (_DB_HEADER_TAG) |
long prolog_length; // | 4 Prolog header length |
struct tm ts; // | 36 Time struct sizeof(struct tm) |
long record_count; // | 4 How many records in the DB |
long highest_write; // | 4 Highest write so far |
long record_length; // | 4 Record length (All fields) |
long header_length; // | 4 Total length of header |
long field_count; // | 4 Field count |
char comment[_DB_COMMENT_MAX+1];// | 201 200 byte descriptive comment |
char reserved[3] // | 3 Reserved for byte-alignment |
} DB_HEADER, *PDB_HEADER; // +------------------------------------+
// | 272 Total (34 x 8 bytes) |
// +------------------------------------+
// field_count * DB_RECORD (N x 85)
const _MAX_DB_FIELD_NAME=80;
typedef struct DB_RECORD_TYPE { // +- Record Header Suffix -------------+
long length; // | 4 Field length |
char name[_MAX_DB_FIELD_NAME+1];// | 81 The field name + 1 |
char reserved[3]; // | 3 Reserved (byte alignment) |
} DB_RECORD, *PDB_RECORD; // +------------------------------------+
// | 88 Total (11 x 8 bytes) |
// +------------------------------------+
// Header Suffix Size = sizeof(DB_RECORD) * field_count
// Record count = ((Entire Header size - 272) / 88)
// Rest of file = header_length byte(s) + (record_length * record_count)
Notes
-----
CGI response limits are unknown and untested
See also: plugins, xbase, sqlite
DHCP Plugin (Simple DHCP commands) |
Index |
To load, use: lib ticol_dhcp ?-verbose? ?-nocomplain?
bool [dhcp enabled adapter_index]
bool [dhcp release adapter_index ?-verbose?]
bool [dhcp renew adapter_index ?-verbose?]
double [dhcp start]
double [dhcp end]
The [dhcp] plugin offers extra commands which may be used to query
and/or renew a workstation's IP address
Windows has a concept of an adapter index within an array of adapters.
We iterate from 0 to adapter_max-1 but we must use the adapter index
value to manipulate it via [dhcp] as adapters may be unplugged or
inserted, thus rendering the actual, live array 'sparse'.
The required 'adapter_index' value is obtained from [info adapter_index N]
where N iterated from 0 to [info adapter_count]-1
[dhcp start] and [dhcp end] return the DHCP lease start and expiry dates
as a double value. These double return lease date values can be converted
to a readable date format using the [date] command
Once the adapter is unbound from TCP/IP it will no longer be discoverable
using its index value. [info adapter_index N] should be used to retrieve
the current index value for an adapter which we handle via an index of
0..N from [info adapter_count]
The [info] command provides a number of arguments which are useful for
handling DHCP requests and querying adapter properties
Notes:
An undefined date is returned in the "C" epoch as 0 or 1970/01/01 00:00:00,
which is interpreted in the Ticol (VB) date scheme as 25569.0
Example:
option expression off
# CAUTION: This script will release and renew your IP address
for {set i 0} {< $i [info adapter_count]} {++ i} {
if {dhcp enabled $i} {
puts "ID:[info adapter_id $i]"
puts "Index: [info adapter_index $i]"
# Show the current DHCP server and lease dates
puts "DHCP Server =[info dhcp_server $i]"
puts "DHCP Lease Start=[date [dhcp start $i]]"
puts "DHCP Lease End =[date [dhcp end $i]]"
if {> [dhcp end $i] 25569.0} { # If more recent than 1970/01/01
set index [info adapter_index $i]
textcolor magenta
puts "Renewing $i (index $index)"
puts "Released?\t[bool [dhcp release [info adapter_index $i]]]"
puts "Renewed?\t[bool [dhcp renew [info adapter_index $i]]]"
# Show the new DHCP server and lease dates
puts "DHCP Server =[info dhcp_server $i]"
puts "DHCP Lease Start=[date [dhcp start $i]]"
puts "DHCP Lease End =[date [dhcp end $i]]"
textcolor
newline
break
}
newline
}
}
See also: plugins, dns, info
DNS Plugin (Simple DNS / Reverse DNS lookups) |
Index |
To load, use: lib ticol_dns ?-verbose? ?-nocomplain?
list [dns hostname|ipv4-address ?-mx?]
string [rdns ipv4-address]
Performs a DNS or reverse DNS lookup on the hostname or IP address argument.
[dns] returns a list containing the IPV4 address(es) of the host. More than
one item may be returned. [rdns] returns the hostname matching a given IP
address if that address has an associated record
Examples:
lib ticol_dns
puts [dns yahoo.com]
puts [dns yahoo.com -mx]
foreach x [dns yahoo.com] { puts \t$x }
puts [rdns 8.8.8.8]
puts [rdns 127.0.0.1]
Results:
98.138.253.109 206.190.36.45 98.139.183.24
mta6.am0.yahoodns.net mta7.am0.yahoodns.net mta5.am0.yahoodns.net
98.138.253.109
206.190.36.45
98.139.183.24
google-public-dns-a.google.com
snoopy
See also: plugins
FTP Plugin (Simple FTP Client) |
Index |
To load, use: lib ticol_ftp ?-verbose? ?-nocomplain?
ftp hostname get|put|ls filepathspec ?-u username? ?-p password?
?-t targetpath? ?-pasv? ?-overwrite? ?-port port? ?-v?
ftp hostname|exists size filespec ?filespec...? ?-u username?
?-p password? ?-port port? ?-pasv?
ftp ls:
Accepts one filename. Returns a list containing the contents relating to
the filespec
ftp size:
Accepts multiple filenames. Will return a list containing the discovered
filenames and their sizes.
Example:
ftp localhost size /tools/zip-7100a80516.zip /tools/netdiag.zip
Result:
{/tools/zip-7100a80516.zip 1292683} {/tools/netdiag.zip 661806}
ftp exists:
Accepts multiple filenames. Will return a list containing the supplied
filenames and a boolean 1 or 0 depending if the file exists or not.
Example:
ftp localhost exists /tools/zip-7100a80516.zip /nosuchfile.zip
Result:
{/tools/zip-7100a80516.zip 1} {/nosuchfile.zip 0}
See also: plugins
XBase - A Simple Dbase III/IV Plugin |
Index |
To load, use: lib ticol_xbase ?-verbose? ?-nocomplain?
handle [xbase open filename]
long [xbase create filename {tcl-list-of-sets}]
bool [xbase close handle]
string [xbase info handle]
long [xbase count handle]
string [xbase get handle N] # record is base 1
short [xbase fieldcount handle]
string [xbase fieldname handle N] # field is base 0
long [xbase fieldsize handle N] # field is base 0
string [xbase field handle N] # field is base 0
bool [xbase set handle fieldnum value] # field is base 0
bool [xbase put handle ?record?] # record is base 1
bool [xbase delete handle record] # record is base 1
bool [xbase undelete handle record] # record is base 1
bool [xbase is_deleted handle record] # record is base 1
bool [xbase pack handle]
bool [xbase setver handle integer]
bool [xbase binary set handle integer value]
bool [xbase binary field handle N]
bool [xbase index handle index_name ?field_number?]
A basic XBase/Dbase plugin to handle Dbase III/IV files (and equivalents)
This plugin offers simple manipulation of generic XBase database files.
[xbase] can handle DBF files produced by OpenOffice, Excel or XBase
programs
No advanced features are offered such as memo fields or indexing
Standard DBase files suffer from what is, in the 21st century, a fairly
severe field size limitation of 254 characters. A record can have up to
1024 fields and the field name is limited to 31 characters.
See: ticol_db for a database which can handle larger records.
Note that record indices are "base 1" values, whilst field indices are the
same as arrays (base 0)
[xbase open] will return a handle to an open database. This must be saved
and re-used for other xbase commands
[xbase create] requires a Tcl list one or more braced arguments of field
specifiers as follows:
{field_name length ?type? ?decimals?} where field_name can be no longer
than 11 characters (for DBase III emulation) and optional 'type' is one
of:
'C' - Char/string (default)
'D' - Date (data is stored as text)
'L' - Logical
'M' - Memo
'N' - Numerical
Each field set must be separately braced
Example:
xbase create "test.dbf" {{name 30} {address 25} {age 3}}
The Ticol [xbase] lib doesn't enforce data-typing on field-write. However,
it will create files which are compatible with the XBase format. Memo
fields are unsupported and will be treated as text. It is up to the
programmer to write code which enforces data types as necessary
XBase error codes:
enum {
DBERROR_SUCCESS # 0 No error
DBERROR_ALREADY_OPEN # 1 The file is already open
DBERROR_FILE_NOT_FOUND # 2 Can't find existing file
DBERROR_OPEN_FAIL # 3 Failed to open existing file
DBERROR_INVALID_PARAM # 4 Invalid argument
DBERROR_UNKNOWN_OPEN_MODE # 5 Bad file open mode
DBERROR_VERSION # 6 Wrong DB version on existing file
DBERROR_HEADER_LEN # 7 Bad record header length
DBERROR_RECORD_LEN # 8 Bad record length found
DBERROR_FIELD_COUNT # 9 Bad field count found
DBERROR_FIELD_LEN # 10 Bad field length found
DBERROR_FIELD_TOO_BIG # 11 Max field length is 256 chars
DBERROR_MEMORY # 12 Out of memory
DBERROR_SEEK_FAIL # 13 Disk access error failed
DBERROR_DISK_SPACE # 14 Low disk space. Cannot write
DBERROR_FILE_EXISTS # 15 Can't create new file
DBERROR_EXCEPTION # 16 Critical exception occurred
DBERROR_UNHANDLED # 17 Unhandled exception
DBERROR_ERROR # 18 Default error code
}
[struct] may be used to write binary data
[xbase create] returns 0 for success or an error code > 0.
On success, [xbase create] returns a handle to an open file
Example:
set handle [xbase create {{Name 30 C} {Address 40 C} {Phone 20 C}}]
[xbase close] Closes a database and previously opened handle. It destroys
all resources associated with that handle. It returns a boolean 1 if the
handle is valid and the database was closed successfully
[xbase info] returns a Tcl list containing the following:
record-count field-count {{fieldname length type} ... }
[xbase count] returns a count of the number of records in an open xbase
database file
[xbase get] retrieves a record from an open database. The record must exist
If the record is retrieved then the whole record buffer is returned. Use
[xbase field] to retrieve individual field values. The record index value
is a "base 1" value with the first record being 1
[xbase fieldcount] returns a count of the number of fields in an open xbase
database
[xbase fieldname] returns the field name for a given xbase field and takes
a base 0 index value
[xbase fieldsize] returns the size of an allocated field in an existing
database file. It takes a base 0 index value
[xbase field] returns individual field data for an open xbase database at
the position of the current record (see [xbase get]). [xbase field] takes
a base 0 index value
[xbase set] set the value of a given xbase database field. The index value
is array base 0 and the first field is 0. [xbase set] does not actually
flush (write) the record to disk file (See [xbase put]).
[xbase set] returns a boolean success return value
[xbase put] writes the pending values in the record buffer to disk file.
If a record number is omitted then data is written to the current record.
The record index value is a "base 1" value, [xbase put] returns a boolean
success return value
[xbase delete] does not physically delete a record. The XBase convention is
to tag a hidden field with a delete flag. Likewise [xbase undelete]. The
command [xbase is_deleted] can be used to test whether a record is deleted
or not
[xbase pack] will 'pack' a database, removing all deleted records and
truncating the file as required. This may be a slow process
[xbase setver] will set the version flag field according to the integer
value passed. Use of this command is not recommended other than for
experimental use. Common version values are as follows:
In this value, bits 0-2 indicate version number: 3 for dBASE Level 5,
4 for dBASE Level 7. Bit 3 and bit 7 indicate presence of a dBASE IV or
dBASE for Windows memo file and bits 4-6 indicate the presence of a
dBASE IV SQL table; bit 7 indicates the presence of any .DBT memo file
(either a dBASE III PLUS type or a dBASE IV or dBASE for Windows memo file)
The default value is 0x03
[xbase binary] is able to write data with embedded NULL (0x00) characters
to and from database fields. Note that strings with embedded NULLs will
be truncated at the first NULL with most Ticol commands. You can use
commands such as [binary] to manipulate binary fields (See: binary)
Indexes
-------
Standard XBase database indexes are not provided
[xbase index] provides built-in indexing using Ticol associative arrays
It is trivial to write one's own routines to use the efficient Ticol
associative arrays to index an XBase database file by using the subscript
as the record key value and the array value as the stored record number
value. A proc could be provided to match on partial index key matches.
Duplicate index key entry cases would need to be handled to avoid
collision. This is the method used by [xbase index]
set rc [xbase count $handle]
for {set i 1} {<= $i $rc} {++ i} {
xbase get $handle $i
set text [string left [xbase field $handle 0] 20] # Field 0
set index(text) $i # Usual subscript/value assignment inverted
}
set search "Hello world"
puts [xbase get $handle [string left $index($search) 20]]
See: xbase_index.tcl
Combined Read Database Example
------------------------------
option expression off
lib ticol_xbase -verbose
set filename "employee.dbf"
set handle [xbase open $filename]
set rc [xbase count $handle]
set field_count [xbase fieldcount $handle]
puts "------------------------------------------------------"
puts [xbase info $handle]
puts "------------------------------------------------------"
for {set i 1} {< $i 10} {++ i} {
xbase get $handle $i
printf "-- Record %4i -------------------------------------\r\n" $i
textcolor yellow
for {set j 0} {< $j $field_count} {++ j} {
printf "%4i '%10s\t%-44.44s'\r\n" $j\
[xbase fieldname $j $handle]\
[xbase field $j $handle]
}
textcolor
puts "----------------------------------------------------"
}
xbase close $handle
Results
# 29 records
# 11 fields
------------------------------------------------------------------
29 11 {{SUPPLR_ID C 10} {COMPANY C 36} {CONTACT C 25} {CON_TITLE C
25} {ADDRESS C 42} {CITY C 13} {REGION C 11} {ZIP_CODE C 10} {COU
NTRY C 12} {PHONE C 15} {FAX C 15}}
-- Record 5 ---------------------------------------------------
0 ' EMPLOY_ID 666 '
1 ' LAST_NAME Ismay '
2 'FIRST_NAME Bruce '
3 ' EMP_TITLE Sales Manager '
4 ' BIRTHDATE 19711107 '
5 ' HIRE_DATE 20120123 '
6 ' ADDRESS 23 Widget Hill '
7 ' CITY London '
8 ' REGION '
9 ' ZIP_CODE '
10 ' COUNTRY UK '
11 'HOME_PHONE (044) 0123-4848458 '
12 ' EXTENSION 3453 '
13 ' NOTES Graduated from St. Elsewhere '
14 'REPORTS_TO 333 '
------------------------------------------------------------------
See: xbase_inspect.tcl
Create Database Example
-----------------------
option expression off
lib ticol_xbase.dll
set handle [xbase create new "{Name 30} {Address 40} {Phone 10}" -y]
set info [xbase info $handle] # Get database info
puts "Info Info: $info"
puts "Info Records:\t[lindex $info 0]"
puts "Info Fields: \t[lindex $info 1]"
puts "Info Structure: [lindex $info 2]"
foreach {x} [lindex $info 2] { # Iterate fields
puts "'$x'"
}
newline
set rc [xbase count $handle] # Get record count
puts "Database record count for new.dbf is: $rc"
# set handle field value
xbase set $handle 0 "Biffa Bacon"
xbase set $handle 1 "49 Letsby Avenue"
xbase set $handle 2 "(0123) 4567890"
xbase put $handle 10 # Write to record 10
newline
set i 10
xbase get $handle $i # Get the record
set field_count [xbase fieldcount $handle]
for {set j 0} {< $j $field_count} {++ j} { # Iterate fields
printf "%4i %-10s\t\[%s\]\r\n" $j\
[xbase fieldname $j $handle]\
[xbase field $j $handle]
}
xbase close $handle
Results:
Info Info: 0 3 {{NAME C 30} {ADDRESS C 40} {PHONE C 10}}
Info Records: 0
Info Fields: 3
Info Structure: {NAME C 30} {ADDRESS C 40} {PHONE C 10}
'NAME C 30'
'ADDRESS C 40'
'PHONE C 10'
0 NAME [Biffa Bacon ]
1 ADDRESS [49 Letsby Avenue ]
2 PHONE [(0123) 456]
Notes
-----
CGI response limits are unknown and untested
See: xbase_man_create.tcl
See also: plugins
string [xbase get handle N] # record is base 1
[xbase get] retrieves a record from an open database. The record must exist
If the record is retrieved then the whole record buffer is returned. Use
[xbase field] to retrieve individual field values. The record index value
is a "base 1" value with the first record being 1.
set rc [xbase count $handle]
for {set i 1} {<= $i %rc} {++ i} {
xbase get $handle $i
set text [string left [xbase field $handle 0] 20] # Field 0
set index(text) $i # Usual subscript/value assignment inverted
}
set search "Hello world"
puts [xbase get $handle [string left $index($search) 20]]
See also: xbase, plugins
To load, use: lib ticol_zip ?-verbose? ?-nocomplain?
integer [zip {filespec list} ?-name filename??.zip?? ?-r?
?-p password? ?-verbose? ?-y?]
If '-name' is omitted then a ZIP file with a random filename will be created.
The filespec-list may contain multiple wildcard filespecs. These will be
expanded automatically. The filename is returned in the results list
if -r is specified then any given folders will be searched recursively and
this will apply to all filespecs given. The resulting ZIP file may be
encrypted using relatively weak but standard ZIP encryption
To overwrite existing files use '-y' (yes)
[zip] returns a list containing the ZIP filename, size of the ZIP file in
bytes and a count of the number of files successfully added to the ZIP
and any error code (in that order)
The ZIP module, ticol_zip.dll, is based on the InfoZIP code 2.3 and thus
AES encryption is not supported. The library supports password protection
using ZIPCrypt deflate only. You should not use this plugin to create ZIP
files which require high levels of encryption security
Example:
set result [zip {*.tcl} -r -password secret] # ZIP recursively
puts "ZIP filename is [lindex $r 0]"
Copyright
---------
The internal ZIP library code is Copyright
Copyright (c) 1990-1999 Info-ZIP. All rights reserved
The plugin library is not produced or supplied by Info-ZIP and is
not supported by the Info-ZIP team
The plugin library is not supplied by Info-ZIP
See also: plugins, lib, Ticol, big numbers, mail
Scientific Format Number Plugin |
Index |
A plugin to add basic scientific format number support
To load, use: lib ticol_sci ?-verbose? ?-nocomplain?
sci-number [sci decimal-number] # Convert decimal to scientific
double [dec sci-number] # Convert scientific to decimal
Where the scientific format number is in the format:
?+|-?<double-coefficient>?e|E?+|-?<integer-exponent>
and where 'double-coefficient' may be presented in integer format
[dec] is useful to ensure that a data source containing numbers which
randomly appear in scientific format are normalised to decimal. [dec]
safely passes-through real/float and integer valueas and will automatically
convert those which appear in scientific format
Ticol does not natively support scientific format numbers without
conversion. Support for these may be added in future revisions.
However, these two commands offer the ability to support scientific format
numbers using commands
Scientific representation is an 'inexact' representation. That is, a [big]
integer representation may be more precise than that converted by [sci]
The result of [dec] must be within the bounds of the 64-bit integer range
The input to [sci] should be within the 64-bit integer range
The output format for [sci[ will format with no trailing zeroes in the
base with 2 digit exponent and with capital letter E output
Example:
lib ticol_sci
puts [sci -0.002]
puts [dec -2.000000e-03]
puts [dec 2E24]
Results:
-2.0E-03
-0.002
2000000000000000000000000.000000
Note that [atof] and [ftoa] may also be used internally to convert
Note that [printf] also supports scientific number output using the %e
formatting option
Example:
printf "%e\r\n" -0.0002
Result:
-2.000000e-004
https://en.wikipedia.org/wiki/Scientific_notation
Values passed to [dec] will be normalised to x.xx format
Examples:
puts "[dec 12.1e+1]\t[sci [dec 12.1e+1]]"
puts "[dec 12.1e-1]\t[sci [dec 12.1e-1]]"
puts "[dec 3887.1e+1]\t[sci [dec 3887.1e+1]]"
puts "[dec 3887.1e-1]\t[sci [dec 3887.1e-1]]"
puts "[dec 8283887.1e+5]\t[sci [dec 8283887.1e-5]]"
Results:
# Decimal Value Scientific Value
121.000000000000000 1.21E+2
1.210000000000000 1.21E+0
38871.000000000000000 3.8871E
388.710000000000040 3.8871E+2
828388710000.000000000000000 8.2838871E+1
See also: atof, ftoa, plugins, bignum
Windows Service Query Plugin |
Index |
A plugin to query Windows services and retrieve their status
To load, use: lib ticol_svc ?-verbose? ?-nocomplain?
service <registered-service-name> ?-verbose?
The command must be run with sufficient access rights
See: is_elevated, elevate
The following codes or verbose status messages are returned:
Code Verbose Comments
---------------------------------------------------------------------------
0 Running The service is installed and running normally
1 Stopped The service is installed + in a stopped state
2 Stopping The service is installed and running
3 Not found The requested service was not found
4 No error Reserved
5 Can't load advapi32.dll Cannot load the required system DLL
6 Can't locate export Wrong version of advapi32.dll
QueryServiceStatusEx Can't locate function in the DLL
7 OpenSCManager failed Could not open the SCM. Check access rights
Ensure Ticol.exe is run 'elevated'
8 Unknown service status An unknown service status was returned
9 Unhandled error Unhandled error
This offers a simple method of querying Windows services. Only services
with ANSI names are supported. Services typically have two names, an
actual registered service name and a display name. This command requires
the registered service name
This plugin is query only. Service actions are not yet supported
Example:
"Branch Cache" service name is "PeerDistSvc"
"DNS Client" service name is "Dnscache"
"DHCP Client" service name is "Dhcp"
"Distributed Link Tracking Client" is "trkwks"
NET START will show the display name, not the registered service name
The true name may be obtained from the Windows Service Control Manager
(SCM) via Start->Run->services.msc
Services which have spaces in their registered name must be enclosed in
quotes
Example:
lib ticol_svc
puts [service "FileZilla Server"]
puts [service "FileZilla Server" -verbose]
puts [service "Nosuch App"]
puts [service "Nosuch App" -verbose]
Result:
0
Running
3
Not found
See also: plugins, lib, Ticol, big numbers, zip
This is a simple SMTP emailer plugin which exposes the following command:
To load, use: lib ticol_mail ?-verbose? ?-nocomplain?
sendmail smtp_server from_address {to_list}
?-subject {subject}?
?-body {message body}?
?-u username? ?-p password?
?-port port_value?
?-a {attachment list}?
?-cc {cc list}? ?-bcc {bcc list}?
?-priority <0..9>?
?-dsn? ?-nc? ?-debug?
Arguments should be braced, especially as indicated
sendmail returns the SMTP error code. 250 usually indicates success.
See SMTP documentation for more information
The first 4 arguments to sendmail are mandatory and positional.
The remainder are optional and may be supplied in any order
If a username/password is given then SMTP authentication will be attempted
This is a simple SMTP client and SSL/TLS is not supported. Note that SMTP
sends passwords in plain text where there is no complex authentication
A copyright and source of origin signature is added to all emails unless the
-nc (no copyright) argument is passed. Useful debugging information can be
displayed using -debug
Multiple unlimited attachments are supported, but these are not expanded
from wildcards. Wildcards, if found, are ignored. ls filespec var -list
may be used to expand wildcard filespecs. Concatenate using [append] not
[lappend]
Example:
set smtp_server smtp.example.com
set from admin@example.com
set to [list recipient1@example.com admin@example.com]
set subject "This is a test email sent at [date]"
set body_text "Hi there, this is a nice email - $qbf"
sendmail $smtp_server $from $to -subject $subject -body $body_text
The mail module, ticolmail.dll, is not based on 3rd party code and was
written by the author of Ticol
See also: plugins, lib, Ticol, big numbers, zip
TCP Socket Ping Plugin |
Index |
This is a simple TCP 'ping' tester which exposes the following command
integer [sockping hostname port ?-var varName?]
To load, use: lib ticol_sockping ?-verbose? ?-nocomplain?
This plugin was written to test for the availability of services on an
internal LAN. Scripts can be written to advise an administrator in the
event that a critical service is no longer 'up'.
This will test for the presence of live socket connectivity by making a
TCP connection to a remote host on the specified port. If the remote host
responds then 0 will be returned. If the host is up but does not respond on
the given port then 1 is returned. Otherwise an error code is returned as
follows. No data is sent to the remote host and the connection is dropped
as soon as established
Not all services will return a response string (e.g. web servers)
0 Successfully connected to the remote host on the given port
1 Failed to connect to the remote host (timed out/no response)
2 Syntax error
3 Too many command line arguments
4 Socket error
5 WSAStartup error
6 Getaddrinfo error
7 Disconnect error (error while disconnecting from remote host)
Example:
lib ticol_sockping
set r [sockping localhost 25 -var v] # Check if an SMTP server is up
puts "Returned code $r"
puts "Response string: '$v'"
Results:
Returned code 0
Response: '220 mail.kerys.co.uk - MailQ v1.16, HP (2 CPU) on HP Wedne
sday, 20th December 2017 +0000 23:57:22'
See also; plugins
Big Integer Math Plugin |
Index |
Large precision integer math with some float operations
To load, use: lib ticol_bignum ?-verbose? ?-nocomplain?
For information: puts [lib ticol_bignum -info]
This module adds large-precision signed-integer number commands. Standard
Ticol math uses 64-bit integer values. Numbers up to 6500 digits are
handled (subject to recursive stack memory limits. Since these digits are
strings they can be easily interfaced into databases and files
This is primarily an integer library which takes only integer values.
[fdiv] however, can return float format results. Binary operations such as
[big &] will generally operate only on positive integers as integers of
arbitrary width have no fixed position for a sign flag. Some operations
will allow emulation of fixed-bit-width integers but without sign value.
Whilst it is possible to combine math operations on several big numbers
note that calculation speeds of big math are relatively slow and
inefficient, particularly for division. Calculations may need to be
restructured to compensate for this. Division involving numbers with
hundreds or thousands of digits may take many minutes to compute
For binary values, there is no fixed bit-width so, again, calculating
negative values is a problem due to the unfixed bit-width for 2's
compliment. Therefore negative binary numbers are not handled rather than
round up to the next highest power of 8 and lose performance.
For more information ...
See:
http://www.codeproject.com/Tips/1079637/Twos-Complement-for-Unusual-Integer-Sizes
Float values can be handled by multiplying by a precision factor of say 15,
performing math such as division and then dividing back down afterwards or
shifting as a string and inserting a decimal point
Values are truncated, not rounded up or down. Ticol natively uses prefix
notation for commands, not infix, so the syntax is:
[big <operator> value ?value?]
Example:
puts [big / 22 7.0]
Where 'operator' is one of the following integer-only operators:
+ a b Integer addition
- a b Integer subtraction
* a b Integer multiplication
/ a b Integer division (see fdiv for float)
Note is truncated, not rounded up or down
% a b Modulo division (remainder after integer division)
++ a Integer increment
-- a Integer decrement
== a b Equivalence
!= a b Difference
< a b Less-than
> a b Greater-than
<= a b Less-than or equal to
>= a b Greater-than or equal to
& a b ?N? Bitwise AND (result optionally restricted to N bits)
| a b ?N? Bitwise OR (result optionally restricted to N bits)
|| a b Logical OR
^ a b Bitwise XOR
~ a Unsigned bitwise NOT
(May not return valid bit-width results)
! a Logical NOT
+= a b Add (Requires a variable name Lvalue without $)
-= a b Sub "
*= a b Mul "
/= a b Div "
&= a b BitAnd " (result optionally restricted to N bits)
|= a b ?N? BitOr " (result optionally restricted to N bits)
^= a b BitXor "
<< a b ?N? Shift a left by b bits ?bitwidth N? (mul by power of 2)
(result optionally restricted to N bits)
>> a b ?N? Shift a right by b bits ?bitwidth N? (div by power of 2)
(result optionally restricted to N bits)
** a b a to the power of b
abs a Absolute value (remove sign)
bintoint a ?-noprefix?
Convert binary string to big integer (0b prefix optional)
debug on|off Turn on the trace for "/" and "fdiv" (see graph)
div2 a Rapid division by 2 (faster than [big / x 2]
div10 a b Rapid division by 10 (faster than [big / x 10]
fact a Factorial (1..1000)
fdiv a b Floating point division of two big numbers
Precision is 18 digits after the decimal point
fib a Fibonacci number (1..30000)
fmod a b Remainder after float division
fmul a b Float multiplication <NOT YET IMPLEMENTED>
graph Display a coloured graph of a big number using the
standard IEEE electrical colour code. Note that orange
may appear as khaki on some systems is resolved using a
registry tweak. See: console colours
inttobin a ?-noprefix?
Convert from unsigned big positive integer to binary
string. Use [big abs] to ensure +ve input value
is_even a Returns boolean true (1) if number is even
max a b Maximum of two big integers
mean ... Mean value of a sequence of integers (returns a float)
Precision is 18 digits after the decimal point
min a b Minimum of two big integers
mul2 a Rapid multiplication by 2 (faster than [big * x 2]
pow a b Raise integer a to the power of b (alias for **)
pow2 a Raise 2 to the power of x (faster than [big pow 2 n])
rnd lo hi Random integer between range lo and hi
sqrt a Integer square root
Two command offer big floating point operations, [big mean] and [big fdiv]
Usage with conditional expression statements and [option expression on]
RIGHT if {[big == $n 0]} { ... # Expression enclosed in square brackets
WRONG if { big == $n 0 } { ... # Expression has not been enclosed in []
RIGHT while {[big < $n 100]} { # Expression enclosed in square brackets
WRONG while { big < $n 100 } { # Expression has not been enclosed in []
Usage with conditional expression statements and [option expression off]
RIGHT if { big == $n 0 } { ... # Expression enclosed in square brackets
RIGHT if {[big == $n 0]} { ... # Expression has not been enclosed in []
RIGHT while { big < $n 100 } { # Expression has not been enclosed in []
RIGHT while {[big < $n 100]} { # Expression enclosed in square brackets
Examples: Result
lib ticol_bignum # (Loads the library)
big % -793455062 -139725649 # -94826817
big / 123456789012345678567890 987654321 # 124999998873437
big * 101020280607085690 181818 # 18367305379419105984420
big inttobin [big << 7 24] # 111000000000000000000000000
big fdiv 423423423437 23 # 18409714062.478260869565217391
big mean 1234567 987654321 23 7 # 247222229.50000000000000
set a 12345; puts [big += a 1000] # 13345
lib ticol_bignum -unload # (Unloads the library)
Example:
option expression on
proc fac n {
if {$n <= 1} {
return 1
} else {
return [big * $n [fac [big - $n 1]]]
}
}
set s {}
for {set i 1} {$i < 40} {incr i} {
set f [fac $i]
append s [format "|%4s| %50s|%5s|\n" $i $f [string length $f]]
}
puts "+----+---------------------------------------------------+-----+"
puts "| n| Factorial n | Len |"
puts "+----+---------------------------------------------------+-----+"
puts $s -nonewline
puts "+----+---------------------------------------------------+-----+"
puts "| 39| 20397882081197443358640281739902897356800000000| |"
puts "| | See: http://www.rpbridge.net/7z78.htm | |"
puts "+----+---------------------------------------------------+-----+"
Result:
+----+---------------------------------------------------+-----+
| n| Factorial n | Len |
+----+---------------------------------------------------+-----+
| 1| 1| 1|
| 2| 2| 1|
| 3| 6| 1|
| 4| 24| 2|
| 5| 120| 3|
| 6| 720| 3|
| 7| 5040| 4|
| 8| 40320| 5|
| 9| 362880| 6|
| 10| 3628800| 7|
| 11| 39916800| 8|
| 12| 479001600| 9|
| 13| 6227020800| 10|
| 14| 87178291200| 11|
| 15| 1307674368000| 13|
| 16| 20922789888000| 14|
| 17| 355687428096000| 15|
| 18| 6402373705728000| 16|
| 19| 121645100408832000| 18|
| 20| 2432902008176640000| 19|
| 21| 51090942171709440000| 20|
| 22| 1124000727777607680000| 22|
| 23| 25852016738884976640000| 23|
| 24| 620448401733239439360000| 24|
| 25| 15511210043330985984000000| 26|
| 26| 403291461126605635584000000| 27|
| 27| 10888869450418352160768000000| 29|
| 28| 304888344611713860501504000000| 30|
| 29| 8841761993739701954543616000000| 31|
| 30| 265252859812191058636308480000000| 33|
| 31| 8222838654177922817725562880000000| 34|
| 32| 263130836933693530167218012160000000| 36|
| 33| 8683317618811886495518194401280000000| 37|
| 34| 295232799039604140847618609643520000000| 39|
| 35| 10333147966386144929666651337523200000000| 41|
| 36| 371993326789901217467999448150835200000000| 42|
| 37| 13763753091226345046315979581580902400000000| 44|
| 38| 523022617466601111760007224100074291200000000| 45|
| 39| 20397882081197443358640281739902897356800000000| 47|
+----+---------------------------------------------------+-----+
| 39| 20397882081197443358640281739902897356800000000| |
| | See: http://www.rpbridge.net/7z78.htm| |
+----+---------------------------------------------------+-----+
Notes:
See: modulo sign for a discussion of how Tcl interprets signs with % as
symbol usage in Tcl differs from that of other languages
See also: option, plugins, lib, Ticol, zip, mail
Large text display
To load, use: lib ticol_bigtext ?-verbose? ?-nocomplain?
For information: puts [lib ticol_bigtext -info]
Call using:
integer [bigtext string ?-x xpos? ?-y ypos? ?-fore colour?\
?-back colour? ?-nocomplain? ?-slow? ?-q?]
list [bigtext - -list]
This module adds large console text or "calculator" style text handling and
which may be usefully combined with the [box] command, Large text is useful
for banners, status, clock or date display. The caller is responsible for
clearing the console of any previously displayed text
The text is better represented with some Windows console fonts than others.
There are limitations on representing text accurately using ASCII characters
Certain fonts space characters irregularly and with excess spacing.
Additional spaces between characters may be inserted using the '-space'
argument
The following characters are handled:
Numerals: 0..9
Letters: A..Z (lower case is converted to upper case)
Punctuation: . , : ; ! ' " ? <space> <tab>
Symbols: + - * / = % \ @ [ ]
If a character is encountered which is not handled then a trappable error
will be returned unless the '-nocomplain' argument is used, in which case
the request to print an unhandled character will be ignored
The full list of handled characters may be returned as a Tcl list (except
TAB and SPACE) using [bigtext - -list]
Slow text drawing, which can emulate old-fashioned terminal output, is
enabled using the '-slow' argument
Example:
# Display the first 20 characters of $qbf
lib ticol_bigtext
cls
bigtext [string left $qbf 20]
Result:
___ _ _ _ _ _ __ _ _ _ _
| |_||_ | || || | |/ |_||_|| |\ \/ |\ | |_ | |\/
| | ||_ |_\|_|| |_ |\_ |_|| \|_| \/\/| \| | |_|/\
Example:
# Draw a real-time clock using large text
lib ticol_bigtext
cls
while {1} {
is_mod $i 7 {
box 20 20 40 6 white darkmagenta -solid
bigtext [time] -x 22 -y 21 -fore white -back darkmagenta
}
}
Result:
_ _ _ _ _ _
| || |. _||_|.|_ _|
|_||_|.|_ |_|. _| _|
Display Width
-------------
Character widths vary and spaces have been included with some characters to
improve readability. The caller is responsible for limiting the supplied
text to ensure it will fit the console window
[bigtext] returns the total actual character width used either by a string
or a single letter. This can be used to discover both the total string
or individual character length before drawing if the -q (quiet) argument is
used and thus check that the string will fit on the console and adjusting if
necessary before displaying
Example
# Discover character and string display widths
lib ticol_bigtext
puts [bigtext W -q]
puts [bigtext 0 -q]
puts [bigtext $qbf -q]
puts [bigtext [time] -q]
puts [bigtext [date] -q]
Results:
5
3
121
20
28
See also: box
A plugin to perform a wordcount or character count
To load, use: lib ticol_wordcount ?-verbose? ?-nocomplain?
Call using:
integer [wordcount arg ?-var? ?-chars?]
This module offers rapid count of characters in either a literal string
or a variable. Using a variable with the -var argument is more
efficient with very large strings as no internal copy is performed
in dereferencing a variable before analysis
The '-chars' argument will return a count of the number of characters
in the supplied string literal or variable
Example:
lib ticol_wordcount
puts [wordcount $qbf]
puts [wordcount qbf -var]
puts [wordcount qbf -var -chars]
Results:
9
9
43
See also: charcount
Modulo Sign (Discussion) |
Index |
integer [% value]
integer [big % value]
Modulo [%] and [big %] follow C/C++ expression rules in that the sign
is set to the same as the dividend. This may change in future versions if
there is a convincing reason to change
ActiveTcl states that: "The remainder will always have the same sign as
the divisor and an absolute value smaller than the absolute value of the
divisor. This differs from other languages as follows:
a b => r Visual BASIC 5 (sign same as dividend)
-------------------------------------------------
+ + => +
+ - => +
- + => -
- - => -
C++ (sign same as dividend)
--------------------------------------------------
+ + => +
+ - => +
- + => -
- - => -
a b => r ActiveState Tcl (sign same as divisor)
--------------------------------------------------
+ + => +
+ - => -
- + => +
- - => -
a b => r Ticol Tcl [%] (sign same as dividend)
--------------------------------------------------
+ + => +
+ - => +
- + => -
- - => -
a b => r Ticol Tcl [big %] (sign same as dividend)
--------------------------------------------------
+ + => +
+ - => +
- + => -
- - => -
Test Cases:
Divided Divisor Result
----------------------------------
C/C++ mod % sign usage
----------------------------------
123423 % 2323 = 304 +
123423 % -2323 = 304 +
-123423 % 2323 = -304 -
-123423 % -2323 = -304 -
Ticol Tcl mod % sign usage
------------------------------
123423 % 2323 = 304 +
123423 % -2323 = 304 +
-123423 % 2323 = -304 -
-123423 % -2323 = -304 -
Ticol Tcl mod big % sign usage
------------------------------
123423 % 2323 = 304 +
123423 % -2323 = 304 +
-123423 % 2323 = -304 -
-123423 % -2323 = -304 -
Common pitfalls
---------------
"When either a or n is negative, the naive definition breaks down and
programming languages differ in how these values are defined" (Wikipedia)
"When the result of a modulo operation has the sign of the dividend, it
can lead to surprising mistakes. For example, to test if an integer is
odd, one might be inclined to test if the remainder by 2 is equal to 1:
bool is_odd(int n) {
return n % 2 == 1;
}
But in a language where modulo has the sign of the dividend, that is
incorrect, because when n (the dividend) is negative and odd, n mod 2
returns -1, and the function returns false. One correct alternative is to
test that it is not 0 (because remainder 0 is the same regardless of the
signs):
bool is_odd(int n) {
return n % 2 != 0;
}
Or, by understanding in the first place that for any odd number, the
modulo remainder may be either 1 or -1:
bool is_odd(int n) {
return n % 2 == 1 || n % 2 == -1;
}
See:
https://en.wikipedia.org/wiki/Modulo_operation
See also: mod, big mod, math
Brainfuck Interpreter Plugin |
Index |
A Brainfuck interpreter
To load, use: lib ticol_bf ?-verbose? ?-nocomplain?
For information: puts [lib ticol_bf -info]
This module offers a small Brainfuck interpreter which is reasonably fast
but note that Brainfuck is an inherently slow and inefficient language
string [bf script ?-highlight? ?-nocompress? ?-colour? ?-kc?
?-ks? ?-c? ?-blob ?char?? ?-debug?]
Brainfuck characters [ and ] require special attention as these are Tcl
command prefixes and must therefore be escaped unless reading directly
from a file (recommended)
The code will be run in compressed format for speed gains unless the
-nocompress argument is used
The default is for direct character output (recommended), but the output
can be cached up to about 32kb using the -c argument. Use cached mode to
return a value to Tcl. If cached mode is not used then an empty string
will be returned.
The output result can be colourised using -colour (-color)
For pseudo graphic (pixel) output, characters other than CRLF and space
may be replaced with a solid block using the -blob argument. -blob
may specify a value which is the blob character (e.g. -blob [chr 254])
Useful characters include 0xb0 (25%), 0xb1 (50%), 0xb2 (75%), 0xdb (solid)
CRLFs may be retained in the compressed code using -kc
Space and TAB characters may be retained using -ks
If any Tcl special characters are returned to Tcl by a script then they
must be escaped
Example:
# Note that [ and ] are escaped Tcl special characters
lib ticol_bf
set s "
++++++++++\[>+++++++>++++++++++>+++>+<<<<-\]>++.>+.+++++++
+++.>++.<<+++++++++++++++.>.+++.--------------.>+.>."
puts [bf $s -c] # Use -c cached mode to return to [puts]
lib ticol_bf -unload
Result:
Hello world
Example:
lib ticol_bf
bf [readfile mandelbrot.bf] -colour -blob [chr 0xfe]
lib ticol_bf -unload
Result:
<Displays a mandelbrot fractal, coloured with a small, solid blob>
http://en.wikipedia.org/wiki/Brainfuck
See: plugins
Macro Pre-Processor (MPP) |
Index |
The Ticol Macro Preprocesor or MPP is used to enhance the flexibility of
Tcl comments, to add C-like conditional blocks and to preprocess static
aspects of a source script in order to speed up execution
Macro commands apply on a per-file basis
The MPP runs when a file is executed or loaded from the command-line and
the Ticol Command Line Interface (CLI). Ticol is optimised for small
script files of under about 2Mb and some runtime processing is offloaded
to the MPP Ticol will be slower to load larger files
The PreProcessor completely strips all Tcl and "C"-style comments and
superfluous whitespace from a file as well as making basic static syntax
checks for mismatched square and curved brackets, braces and double quotes
giving an indication as to which line may have caused the problem
It not only performs useful substitution and conditional execution but
when combined with [calc] it will pre-solve any constant expressions and
perform as much computation on the source file before runtime (See: calc)
Enhanced C-like macro commands are processed by the PreProcessor.
These include conditional #define, #ifdef, #else, #endif blocks as well
as macro variables. For more information see: help macro
C-like const escape sequences within strings, such as \t or \n, are
translated by the MPP. Tcl non-const escape sequences such as \$ will be
evaluated 'on-the-fly' (see: escape, unescape)
Const hexadecimal, octal and binary numbers such as 0x123, 0o123 or 0b111
are translated to decimal by the MPP. Values which are not to be
translated must be supplied as quoted strings. Other than escape
sequences, string contents are not processed. If you don't want this to
happen, wrap the const values in double-quotes
There is no macro expansion with arguments as in C++ such as:
#define _toupper(_c) ( (_c)-'a'+'A' )
In summary. the Ticol Macro PreProcessor does the following:
* Completely removes all commented code from the source
In either standard Tcl "short" or "C" style 'long' form
* Processes all macro statements such as #if, #else, #endif
* Processes all #define statements and performs substitution
* Processes all macro #exec, #echo, #if etc. statements
* Replaces all standard escape sequences wherever possible
* Performs substitution on standard macro variables e.g. __LINE__
* Translates original source line numbers to the __LINE__ variable
* Replaces all unquoted hex, octal and binary numbers with decmimals
* Replaces all instances of [calc] expressions with prefix Polish form
* Performs resolution and simplification on [calc] expressions
* Checks and reports basic syntax errors such as brace/bracket parity
outside of quoted strings (note that parity within quoted strings
is not checked
The PreProcessor may be disabled on the command-line with the /NP argument
or from the CLI using:
option preprocessor off
Compatible scripts can be run withy the MPP disabled using /NP, but they
must not include macro blocks "C" style comments or other non-standard
aspects of Tcl syntax
The PreProcessor is not usually applied to direct CLI console input, only
to scripts
Include files are not supported by the MPP (there is no #include command)
A special/reserved variable 'NDEBUG' is used to globally disable the
[assert] command (case insignificant). For more information, see: assert
For more details of macro commands etc. See macro
For an example of a good macro system for standard Tcl, see 'Sugar'
http://wiki.tcl.tk/11156
See also: cli, macro, debugging, escape, unescape, assert
Preset Global Variables and Constants |
Index |
The following constants or variables are preset on loading Ticol:
Variable Type Description or Example
----------------------------------------------------------------------
::env() Array Const array of Windows environment variables
Subscripts are converted to upper-case indices
are case-sensitive in Tcl
::argv() Array Const array of command line arguments
::exe_name Const C:\VC5\MyProjects\picol\Release\ticol.exe
::autoexec_ran Var 1 if autoexec.tcl ran successfully, otherwise 0
::false Const 0
::true Const 1
::eof Const -1
::qbf Const The quick brown fox jumps over the lazy dog
::pi Const 3.14159265358979323
::NAN Const -1.#IND00000000000
::win64 Const Boolean. Whether Windows is 64-bit or not
::tcl_config Const String. Path+filename of Ticol config (INI)
::tcl_cgi Const 1 (true) if running in CGI mode, otherwise 0
::tcl_cli Const Boolean. Set to 1 if running from the CLI
::tcl_date Const Ticol compilation (build) date/time
::tcl_threaded Const Boolean. Set to 1 if threaded compile of Ticol
::tcl_version Const Ticol version. Same as [info tclversion]
::tcl_precision Const 2. Set in TICOL.INI
::errorCode Var Tcl error code sent to a the console
::errorLine Var May be preset using the #__LINE__ macro
::errorMsg Var Error message generated by a Ticol error
Fatal exceptions generate no errorMsg variable
::errorInfo Var Additional information
Environment variables may be set as individual constants. This depends on
the Windows version in use
Use the command [vars -const] to display all known constants on loading
Note that constants or variables with embedded # characters must be
enclosed in double-quotes
See also: argv, global variable, global, upvar
See the following topics:
puts Simple print output
format Format strings for output with puts
printf Formatted print similar to C's printf()
newline Output a new line
textcolor Set the colour of the text output
See also: Ticol
int [precision double]
Returns the effective precision length for a 'double' or 'real' number and
which ignores trailing whitespace and trailing zeroes
Comparing real numbers can be difficult if they exhibit different precision
Rounding may not always be useful unless a proper rounding values is used.
This would depend on the precision of both numbers
Improperly formatted values are not valid and will return a precision of
zero. Hexadecimal integers are not supported
Integer values will return a precision of zero
Non-numeric strings will also return a precision of zero
Example:
puts [precision $pi] # Returns 17
puts [precision 100] # Returns 0 (integer)
puts [precision "100."] # Returns 0 (integer)
puts [precision 123.456] # Returns 3
puts [precision 123.4560000] # Returns 3
puts [precision "123.4560000 "] # Returns 3
puts [precision 123.4a6] # Returns 0 (not valid)
puts [precision Foobar] # Returns 0 (not valid)
See also: double, fraction, integral, vfract, data type ranges
integer [printf format-string|literal ?arg? ...]
Print command and wrapper for [format]. Calls [format] and outputs the
result after unescaping backslash escape sequences
[printf] emulates the C/C++ printf function. It can output only to stdout
use [format] to format strings for other output streams
[printf] returns the count of (escaped) characters output to the console
[printf string args] is functionally equivalent to
[puts [format string args]]
[printf] prints only to the console. To print to a file stream combine
[puts] and [format] as follows...
set fp [open filename wb]
puts $fp [format "%s %i %f\r\n" One 2 3.0]
close $fp
Line Endings:
[printf] doesn't issue Windows CRLF pairs (0x0d,0x0a) or '\r\n'
If you wish to print newlines you must append \r\n escape character
sequences within a supplied string argument
Example:
printf "%s %li 0x%x %3.2f\n" "Hello" 12345 987654 $pi
Result:
Hello 12345 0xf1206 3.14
Example: (Source: https://www.tcl.tk/man/tcl8.4/TclCmd/format.htm)
option expression on
puts "+-------------------------+"
puts "| 64-bit printf demo |"
set sep +-[string repeat - $w1]-+-[string repeat - $w2]-+
printf "$sep\r\n"
printf "| %-*s | %-*s |\r\n" $w1 "Index" $w2 "Power"
printf "$sep\r\n"
# Print the contents of the table
option expression on
set p 1
for {set i 0} {$i<=30} {incr i} {
printf "| %*d | %*I64i |\r\n" $w1 $i $w2 $p
set p [expr "wide($p) * 3"]
}
printf "$sep\r\n"
Result:
+-------------------------+
| 64-bit printf demo |
+-------+-----------------+
| Index | Power |
+-------+-----------------+
| 0 | 1 |
| 1 | 3 |
| 2 | 9 |
| 3 | 27 |
| 4 | 81 |
| 5 | 243 |
| 6 | 729 |
| 7 | 2187 |
| 8 | 6561 |
| 9 | 19683 |
| 10 | 59049 |
| 11 | 177147 |
| 12 | 531441 |
| 13 | 1594323 |
| 14 | 4782969 |
| 15 | 14348907 |
| 16 | 43046721 |
| 17 | 129140163 |
| 18 | 387420489 |
| 19 | 1162261467 |
| 20 | 3486784401 |
| 21 | 10460353203 |
| 22 | 31381059609 |
| 23 | 94143178827 |
| 24 | 282429536481 |
| 25 | 847288609443 |
| 26 | 2541865828329 |
| 27 | 7625597484987 |
| 28 | 22876792454961 |
| 29 | 68630377364883 |
| 30 | 205891132094649 |
+-------+-----------------+
Note: The big math plugin module can be used to extend the number range
(See: big math)
See: format and format string for more details
See also: format, echo, puts, format string, file
list [procs ?glob-pattern?]
list [info procs ?glob-pattern?]
Returns a Tcl list of user-defined procedures. These will also be included
in the [commands] list.
Example:
option echo on
procs
procs c??e
procs t*n
Result:
cube rainbow square testmean
cube
testmean
Use: option echo on to display directly to the console
See also: commands, functions, info
Passing Structs to a Procedure By Name |
Index |
Although Ticol structs are designed primarily to facilitate communication
with external DLL routines they can also be used internally within script
procedures (procs)
Due to the non-binary nature of Tcl, dealing with structs is not as easy
or as transparent as when using a language such as C or C++. A struct may
contain direct instances of simple Tcl variables such as string, list,
integer or double as well as references to complex abstract data types
such as Array, Stack or Struct
To pass a struct to a proc you will need to pass it by name value
(See: pass by name)
With Tcl procedures any variable passed by name will need to be referenced
either via the global '::' scope prefix or a local alias created using
[upvar]. Each referenced struct field requires either a variable aliased
using [upvar] or '::' or a command which can make use of the struct
variable itself such as [struct item] or [array item]
Once you have created a local reference to the struct you can also, if
necessary create member field references using [upvar]
Struct member fields can be read as well as written back to and
the original struct members are updated when the proc exits
Structs may contain references to other structs but as this can get
rather complex it is best avoided
Example:
Example of a proc setting and returning various data types
undef print_arg -nocomplain
proc print_arg {value} {
# Incoming arg is passed by name
upvar $value
textcolor green
#puts "print_arg \{$value\}"
if {[eq [type $value] stack]} {
puts "Stack: '[stack item $value 0]'"
} elseif {[eq [type $value] array]} {
puts "Array: [array item $value 1]"
} elseif {[eq [type $value] struct]} {
textcolor magenta
puts "Struct '$value' item(s):"
puts " i: '[set ::[struct item ${value}.i]]'"
puts " f: '[set ::[struct item ${value}.f]]'"
puts " r: '[set ::[struct item ${value}.r]]'"
puts " s: '[struct item ${value}.s]'"
puts " a: '[array item ::[struct item ${value}.a] 1]'"
puts " k: '[stack item ::[struct item ${value}.k] 0]'"
} else {
puts "String or number $value is [set $value]"
}
textcolor
}
struct ts { # Test struct: holding various data types
i 20 # Integer Literal
f 40 # Double/Float Literal
r 20 # String Reference
s 256 # String Literal
a 20 # Array
k 20 # Stack
}
set s "Hello world"
set i 123456789
struct set ts.i i
set f $pi
struct set ts.f f
set r "Hello from string reference"
struct set ts.r r # String reference by name
struct set ts.s "Hello from literal string"
set a(1) "This is array item 1"
struct set ts.a a # Array by name reference
stack create stk 200
stack push stk "Hello from stack"
struct set ts.k stk
print_arg s # Pass a variable by name
print_arg i # Pass an integer variable by name
print_arg f # Pass a float variable by name
print_arg a # Pass an array variable by name
print_arg stk # Pass a stack by name
print_arg ts # Pass a struct by name
See also: struct, proc, calling by value
Access to pointers is not a native feature of Tcl. A modest level of
address manipulation can be carried out in Ticol in order to facilitate
the marshalling of data using structs or binary values via [calldll]
The key commands are: struct, setb and ofaddressb
struct test {f 4} # Declare struct with member 'f' of 4 bytes width
# The struct will have 4 bytes of memory allocated
struct setb f 4321 # Set the struct member field to binary value 4321
puts $f # Show the memory address pointed to by member 'f'
puts [ofaddressb f] # Dereference the struct pointer to get the value
struct unset test # Delete
See also: calldll, addressofb, addressof, ofaddress, malloc, memcpy
puts ?stream|stdout? string ?-nonewline?
puts ?-nonewline? ?stream|stdout? string
Print a string to the console with a trailing newline by default
To enhance performance [puts] returns a NULL value ("") rather than
allocate and return the supplied string merely to be discarded in most
cases
Buffered output is configured for stdout by default. This may be disabled
using the /NB (no buffer) argument when you launch Ticol. However this may
result in no output under certain conditions until a program exits. If
your code produces delayed output then you may also use the [flush]
command at critical locations in order to force output
Line endings:
[puts] issues a standard Windows CRLF pair (0x0d,0x0a) or \r\n unless
the -nonewline argument is used
Use: option echo on to display command returns directly to the console
without using [puts] or [printf]
Alternatives:
Use: set varname with [option echo on] to display unescaped the
variable contents
Use with [gotoxy]:
If using [puts] with absolute screen coordinates via [gotoxy] you should
use the -nonewline argument if possible to avoid screen scrolling issues
Example:
option expression off
for {set i 0} {< $i 10} {++ i} {
gotoxy 5 5
puts "I am at 5,5" -nonewline
}
[puts] examples:
Example:
puts "Hello from Tcl" -nonewline
set a "Hello World"
puts stdout $a
set fp [open "test.txt" wb]
puts $fp # Display the file handle (4582320)
puts fp "Hello file"
close fp
Example: (Source: https://www.tcl.tk/man/tcl8.4/TclCmd/format.htm)
puts "+-------------------------+"
puts "| 64-bit integer printf |"
set sep +-[string repeat - $w1]-+-[string repeat - $w2]-+
puts $sep
puts [format "| %-*s | %-*s |" $w1 "Index" $w2 "Power"]
puts $sep
# Print the contents of the table
option expression on
set p 1
for {set i 0} {$i<=30} {incr i} {
puts [format "| %*d | %*I64i |" $w1 $i $w2 $p]
set p [expr "wide($p) * 3"]
}
puts $sep
Result:
+-------------------------+
| 64-bit integer printf |
+-------+-----------------+
| Index | Power |
+-------+-----------------+
| 0 | 1 |
| 1 | 3 |
| 2 | 9 |
| 3 | 27 |
| 4 | 81 |
| 5 | 243 |
| 6 | 729 |
| 7 | 2187 |
| 8 | 6561 |
| 9 | 19683 |
| 10 | 59049 |
| 11 | 177147 |
| 12 | 531441 |
| 13 | 1594323 |
| 14 | 4782969 |
| 15 | 14348907 |
| 16 | 43046721 |
| 17 | 129140163 |
| 18 | 387420489 |
| 19 | 1162261467 |
| 20 | 3486784401 |
| 21 | 10460353203 |
| 22 | 31381059609 |
| 23 | 94143178827 |
| 24 | 282429536481 |
| 25 | 847288609443 |
| 26 | 2541865828329 |
| 27 | 7625597484987 |
| 28 | 22876792454961 |
| 29 | 68630377364883 |
| 30 | 205891132094649 |
+-------+-----------------+
See also: echo, format, printf, set, file open, pause
A test constant defined as: "The quick brown fox jumps over the lazy dog"
https://en.wikipedia.org/wiki/The_quick_brown_fox_jumps_over_the_lazy_dog
See also: preset globals, const, vars
Decimal to Roman Conversion Plugin |
Index |
Roman to decimal conversion is offered for amusement only. The maximum
value which can be converted from decimal is 10000. Values above this
will raise a Tcl exception which may be trapped using [catch]
To load, use: lib ticol_r2d?.dll? ?-verbose? ?-nocomplain?
Commands:
string [decimal_to_roman integer]
integer [roman_to_decimal roman-value]
Order presentation of Roman numerals need not always be significant
and may need careful presentation to achieve expected results:
Example:
puts [roman_to_decimal IVXCLDM] # Same as MDXLIV (1544)
puts [decimal_to_roman 1544] # Same as IVXCLDM (1544)
Results:
1544
MDXLIV
Examples:
lib ticol_r2d
puts [decimal_to_roman 1776]
puts [roman_to_decimal MDCCLXXVI]
puts [roman_to_decimal [decimal_to_roman 1776]]
Results:
MDCCLXXVI
1776
1776
See also; plugins
long-integer [rnd low high]
long-integer [rndseq low high]
integer64 [rnd64 integer64-low integer64-high]
Return a random number between 'low' and 'high' where either value is a
positive, signed long value in the range 0 to LONG_MAX rather than a
native 64 bit unsigned integer
[rndseq] guarantees sequential random numbers which do not repeat whereas
[rnd] does not. There may be cases where it is undesirable to have a
repeating random number, e.g. list selection or random text-parsing
[int64] returns a random number in a full 64-bit integer range
Not recommended for safety critical implementations
Example:
puts [rnd 10 200]
puts [rnd 10 200]
puts [rndseq 10 200]
puts [rndseq 10 200]
puts [rnd64 $LONG_MAX $INT_MAX]
puts -[rnd 10 200]
Result:
42
42 # [rnd] can repeat the same number
87
23 # [rndseq] will not repeat the same number
8538924581789421291 # Big random number (use: source defs.tcl)
-42 # Generate negative numbers by prefixing with -
The big math library can return larger random numbers (See: big)
See also: rnd function, rand, srand, lrand, randomstr, math
Simple conversion of plain text, outputting to an RTF file
To load, use: lib ticol_rtf?.dll? ?-verbose? ?-nocomplain?
Commands:
rtf write filename text ?-y? ?-font name? ?-fs N? ?-rtf?
rtf write filename var -var ?-y? ?-font name? ?-fs N? ?-rtf?
Where:
-var Data is passed in a variable by name (faster for big data)
-y Overwrite any existing file
-font Name of the font (as it appears in Windows)
e.g. "Courier New" or "Tahoma"
May be repeated to define multiple fonts
Each font will be registered as: \f0, \f1 etc.
The first font will be the default (font 0) \f0
-fs Default font size in points e.g. -fs 10 for 10 point
May be repeated to define multiple fonts (see -font)
-rtf Specifies that the text will be supplied in RTF format
(excluding headers). This allows custom RTF codes to be
inserted into the body and resolve clashes with single-
backslash Ticol escape codes.
If -rtf is omitted and RTF escape-codes given then the
codes will be mis-transformed on output
If -rtf is omitted then \\ codes will not be escaped
If -rtf used then \\ codes will require [unescape \\x]
Thus you can use "\par [unescape \\b\\fs20]" -rtf
where this would otherwise embed "\f" (formfeed) and
"\b" (backspace)
-colourtable <rtf-code> (excluding CRLF)
Example:
-colourtable "\{\colortbl ;\\red255\\green0\\blue0;\}"
The internal default is:
{\\colortbl ;\
\\red255\\green0\\blue0;\
\\red0\\green255\\blue0;\
\\red0\\green0\\blue255;}
Example:
rtf write test.rtf $qbf -y -font "Courier New" -fs 12
Notes:
RTF commands may be entered into the text string
Since backslash sequences can be complex in Tcl and particularly when
mixing with RTF and because inserting backslash sequences correctly can
be problematic, you may use [chr 92] to avoid complications. e.g. for
"\f0 use" "[chr 92]f0"
Where there are conflicts with C++/Tcl escape sequences, literal RTF
command characters present in the source-text may need to be escaped.
e.g. (where $q contains our text)
Remember to use the -rtf argument when calling [rtf write]:
[string replace $q "{" "\{"] # Curly braces
[string replace $q [unescape "\\"] [unescape "\\\\"]] # Backslashes
or (within source file):
append s [unescape "\\page\\f1\\fs32\\b "]
append s $q
append s [unescape "\\b0\\fs18\\f0 \r\n\r\n\\par "]
Due to backslash translation issues, RTF backslash commands should
ideally be executed from source files not the CLI. CLI use would
require: option escape on, option autoexec off
Example:
lib ticol_rtf # Load the plugin library
rtf write testrtf.rtf "[unescape \\fs32\\b] $qbf [\\b0]" -rtf -y
@ start testrtf.rtf # Display the final RTF text
A More Complex Example:
# Write a complex RTF document
# This demonstrates a multi-line string with line-continuation chars
# and trailing comments after multi-line string continuations
set ct "{\\colortbl ;\ # Start custom colour table
\\red255\\green0\\blue0;\ # 1: Red
\\red0\\green255\\blue0;\ # 2: Green
\\red0\\green0\\blue255;\ # 3: Blue
\\red255\\green255\\blue0;\ # 4: Yellow (for highlighting)
\\red128\\green0\\blue255;}" # 5: Violet
lib ticol_rtf # Load the plugin library
# Write the body text
rtf write $filename "The \ul quick\ulnone [unescape \\super\\cf1] \
brown [unescape \\cf0\\nosupersub\\f1] fox \
[unescape \\f0\\fs32\\b\\highlight4\\f2] \
jumps\par over [unescape \\f0\\highlight0] the \
[unescape \\fs64] \
la\cf5 z\cf0 y [unescape \\fs20\\b0] dog!"\
-rtf -y -colourtable [unescape $ct]\
-font Tahoma -font "Courier New" -font "Segoe Script"
@ type testrtf.rtf # Display the raw RTF code
@ start testrtf.rtf # Display the final RTF text
Demo Script:
A demo script is included with Ticol which will convert this man file to
RTF which is called man_to_rtf.tcl. The bundled RTF version of the Ticol
manual is generated using this Tcl script
Useful commands include: readfile, append
See also: plugins, chr, unescape, web search re: RTF control codes
Quickstart Guide for Ticol Tcl |
Index |
Loading Ticol
-------------
From the Windows console prompt type the following command to load the
Ticol Command Line Interpreter (CLI):
ticol.exe
Running Scripts from Windows
----------------------------
Scripts can be run directly from the Windows command line using the
following command:
ticol.exe <scriptname>
Where '<scriptname>' is the name of a TCL or TCX script
Command-Line Help
-----------------
Use the following command:
ticol.exe /?
Quitting Ticol
--------------
The interpreter can be exited either by typing 'exit' at the CLI prompt
or by pressing CTRL+BREAK
The following commands are useful in getting started with Ticol Tcl:
Command Description
----------------------------------------------------------------------
# comment # Standard Tcl comments are a hash character
/* comment */ # Ticol supports enhanced C++ style long comments
ticol.exe # Load the interpreter CLI from Windows
load <scriptname> # Load a TCL or TCX script into the CLI
run # Run a previously loaded script
run <scriptname> # Run a script directly from the CLI
dump # Display a loaded script with highlighting
puts <string> # Print a string
set varname value # Set a variable
$varname # Dereference a variable
expr <expression> # Evaluate a natural expression
eval <command> # Evaluate a Tcl command
if {expr} then {script} # Conditionally evaluate a scriptlet
while {expr} {script} # Evaluate a script while 'expr' is true
source <scriptname> # Interpret a script (used for script inclusion)
option # Display configuration option settings
option echo on # Turn on local command result echo (off to disable)
option expression on # Confgure flow control commands to use expressions
vars # Display existing variables
help <topic> # Display help on a specific topic
find <topic> # Find help based on a string search
See the tutorial and faq sections for more information
See also: tutorial, ticol, faq
string [randomstr length ?style? ?-var newvar? ?-nocomplain?]
Create a random string of characters in varying formats and of a given
length
Style values (where style is a positive integer):
0 Mixed upper/lower case
1 Upper case letters
2 Lower case letters
3 Numbers (0..9)
4 Upper/lower case letters and numbers (0..9)
5 Binary data in the range 0x00..0xFF
6 Hexadecimal numbers 0..9, A..F (capitals)
7 Binary string (even bias)
8 Binary string with bias to 0
9 Binary string with bias to 1
10 Hexadecimal numbers 0..9, A..F (lower-case)
11 Printable ASCII characters 33 to 127 except 96
12 All hexadecimal characters (upper/lower case)
13 Octal characters (0..7)
14 Password style 0..9, a..z, A..Z . - % #
15 Safe-readable (confusing characters avoided)
16 Alphanumeric upper-case 0..9, A.,Z
You may define an [enum] to hold the const values
enum {
RANDOM_STYLE_MIXED
RANDOM_STYLE_LOWER
RANDOM_STYLE_UPPER
RANDOM_STYLE_NUMBERS
RANDOM_STYLE_ALL
RANDOM_STYLE_BINSTR
RANDOM_STYLE_HEX
RANDOM_STYLE_BINARY
RANDOM_STYLE_BINARYBIAS0
RANDOM_STYLE_BINARYBIAS1
RANDOM_STYLE_HEX_LOWER
RANDOM_STYLE_PRINTABLE
RANDOM_STYLE_HEXALL
RANDOM_STYLE_OCTAL
RANDOM_STYLE_PASSWORD
RANDOM_STYLE_SAFE
RANDOM_STYLE_ALNUM_UPPER
}
Example:
# Display a row of 20 chars of each type of random string
option expression off
for {set i 0} {< $i 17} {++ i} {printf "%2i:[randomstr 20 $i]\r\n" $i}
Result:
0:YAscKInuJAjvYtxkPpNr
1:qawbgcfcehedfewghkpc
2:SRLQMENLFPKMLDADKSPL
3:31183734177678368924
4:Ukwj71Pk9LvrBwjbO3p9
5:/ÇÆ,- Ô'Ý?ƒû¨õË?¦t-¶
6:5991668E260A498F6BA5
7:11001000110111001000
8:00100100011011010000
9:00011010111111111110
10:8173dc3d3a0df1533a19
11:>eMTb_L'ADs9_4U;E,$p
12:6E646D5D826C3d5063C6
13:34575216245752571651
14::lGf#M5f#V:UbXM3xcHr
15:fyU2A1JkmS7dpEx6MpJJ
16:SF8NIL3J35Y7OH1729CM
See also: rnd, randomise, rndseq, lrand
string [range charset ?charset? ...]
string [range asc-range ?asc-range? ...]
range returns a string of characters generated from 'charset' where
charset is a hyphenated pair of characters: e.g. A-Z or 0-9
Any character pair can be selected in the printable ANSI character set
Multiple character sets are accepted and descending ranges (Z-A) are
accepted
A numeric ASCII range above 9 may be used. Values below 10 will be assumed
to represent integers. The following are valid.
# Print [asc 32] to [asc 42] and [asc 64] to [asc 80]
# Will return: ' !"#$%&'()*@ABCDEFGHIJKLMNOP'
puts "'[range 32-42 64-80]'"
If the RValue of a character pair is absent then 255 is assumed: e.g. "X-"
A range character may be specified using the [asc] command...
e.g. puts [range [chr 26]-c]
Examples:
puts [range 0-9 z-w F-J]
puts [range 9-0 Z-A]
Results:
0123456789zyxwFGHIJ
9876543210ZYXWVUTSRQPONMLKJIHGFEDCBA
See also: string, randomstr
string [read filenum ?-bytes N? ?-line? ?-tab?]
string [file read filenum ?-bytes N? ?-line? ?-tab?]
Read data from an open file stream, as opened with the [open] command
Either reads the entire file or a specified number of bytes in either block
read or line input modes.
To read lines of a text file into an array use [readfile]
Options
-------
-bytes N Read N bytes from file. If '-line then line buffer is N bytes
-line Line input mode. Uses a line buffer of 1024 or -bytes length
Line mode will remove any trailing CRLF characters from the EOL
-tab Replace tab characters with a space
Examples:
set fp [open "input.txt" rt] # Read a file in text mode
set data [read $fp] # Read the whole file
puts $data # Print the whole file
close $fp
set fp [open "input.txt" rb] # Read a file in binary mode
set data [read $fp -bytes 80] # Read a raw block of 80 bytes
puts $data # Print the data (may include CRLFs)
close $fp
set fp [open "input.txt" rb] # Read a file in binary mode
set data [read $fp -line] # Read a line, stripping off CRLFs
puts $data # Print the input line
close $fp
Notes
-----
[read] is not binary-safe. Ticol is not designed for handling binary files
Line ending/CRLF handling varies depending on which mode a file is opened
with. A [read] after opening in "rt" mode will translate CRLF (\r\n) to
\n only and thus may cause a read skew errors of not handled correctly
Be consistent and use either "rt"/"wt" or "rb"/"wb" modes. Don't mix
It is usually preferable to use "rb"/"wb"
When opening a file in binary mode e.g. set fp [open $filename wb]
be sure force CRLFs on write using: puts $fp "Text\r\n" -nonewline
When opening a file in text mode such as: set fp [open $filename wt]
you may use: puts $fp "Text"
When reading using [file eof $handle] the final valid line of the file will
not return EOF. This will be returned on a subsequent read; so do something
like...
set fp [open $filename rb]
if {$fp} {
do {
set s [read $fp -line -tab] # Use internal default line buffer of 1024
if {! [file eof $fp] } { # Test EOF here
puts "$i: '$s'" # -line input will strip CRLFs
++ i
}
} while {! [file eof $fp]}
close $fp
See also: open, file, readfile, writefile
string [readfile filename ?-max byteCount?]
integer [readfile filename varname|- ?-array? ?-max byteCount?]
Reads data from a file in binary mode, optionally passing into a variable
Where a variable is specified, readfile returns the count of bytes,
otherwise the read-in data block is returned
The line read within [readfile] handles lines up to 8Kb (8192 bytes) long
Longer lines will be split and wrapped
The filename should be given unescaped if it contains backslash characters
Certain variables such as $argv() are stored in escaped format. Direct user
input routines may store an unescaped string which may be passed directly
Example:
readfile [unescape $argv(1)] a -array]
Alternately, Unix-style forward slashes '/' may be used in the filename
If -max is specified combined with -array, then this will limit the maximum
number of lines, otherwise it will limit the total number of bytes read-in
If the -array argument is given then the file will be written to a Tcl array
with an integer subscript of index base 0. The return value in this case
will be the number of lines in the array. You can use [array walk arrayName]
to inspect the array or [for] to iterate it efficiently
Varname may be omitted by specifying a - sign. This will allow a byte-count
to be specified without specifying a variable.
Example:
puts [readfile autoexec.bat q]
Result:
1031
Example:
puts [readfile autoexec.bat q -max 200]
Result
200
Example:
# Read autoexec.tcl to an array variable
option expression on
readfile autoexec.tcl a -array
for {set i 0} {$i < [array size a]} {incr i} {puts $i:$a($i)}
Result:
<prints out file the contents>
See also: read, file
Recursion and Issues Arising |
Index |
Recursion and recursive proc calls should be controlled to avoid excessive
nesting, otherwise the interpreter will run out of memory. Well-designed
routines should not need to recurse deeply
The maximum theoretical command/proc recursion depth is set to 5000 but in
practice this may be lower where proc variable consume memory.
If secondary command calls are nested then this recursion depth limit may
be significantly lower and may result in a non-fatal exception error.
On a side note, the default ActiveState Tcl recursion limit is 1000, this
is perhaps a sensible limit.
A trappable Ticol exception will be raised at a recursion depth of 5000
(use [catch] or [try...catch]) but this may depend on available memory.
Additionally, the breakpoint debugger stack will track only the first 100
command calls, after which no further traces will be stored
The maximum recursion depth is configurable via ticol.ini as
MaxRecursion=<value>
where <value> > 0 && <value> <= 5000
If exceptions occur then try lowering this value to around 1000 by editing
this value in ticol.ini
Example:
# This routine will trigger a recursion-depth exception
proc foo {} {
set ::highest_level [max $::highest_level [info level]]
puts $::highest_level
foo # Uncontrolled recursive call
}
foo # Start a Stack Overflow
See also: proc
regex (regular expressions) are not supported
[string match] ([match]) supports glob style string matching and [scan]
can also be useful in string matching
Example:
set string "#08D03F"
scan $string "#%2x%2x%2x" r g b
puts "$r $g $b"
Results:
8 141 63
https://www.tcl.tk/man/tcl8.5/TclCmd/scan.htm
See: string match, scan
Ticol reserves a small number of variables for internal use
These values are protected against deletion by [clear] and [run]
argv const string
argv0 const string
argc const integer
eof const integer
env array
exe_name const string
false const integer
NAN const string
pi const float
qbf const string
tcl_config const string
tcl_cli const string
tcl_date const string
tcl_threaded const bool
tcl_precision const integer
tcl_version const string
theta const float
this Defined dynamically during evaluation
true const integer
win64 const integer
See also: $this
bool [rename oldname newname]
Rename a procedure (proc). This may be used to temporarily 'hide' or
replace an existing procedure.
Renaming a proc to an empty string "" or {} will delete the proc.
The author is not a fan of this mechanism and has added the [undef]
command to undefine a proc.
[defined] may also be used to test for the existence of a proc before
it is defined or deleted. Rename has less 'functional transparency'
and obviousness of purpose than [undef]
You cannot rename a currently active proc or its parent proc. Attempting
to do so will generate an error
Example:
proc square {x} {
* $x $x
}
puts [square 10]
rename square skware
puts [skware 10]
rename skware ""
Result:
100
100
See also: is_active, undef, defined, file rename, ren
integer [result -code]
string [result]
Return the result of the last command either as an integer code value
See also: eval, break, continue, return, exit, error
return value ?-code number|nmemonic?
return [command result] ?-code number|mnemonic?
or:
return -code number|nmemonic value
[return] exits a procedure or block of code. [return] may optionally
return an error code to the calling routine. There are limits to the use
of [return] and it does not behave in the same way as with other languages
such as C/C++ [return] will exit all nested procedures and unwind the
call-stack down to proc level 0 (the root invocation level)
Example:
proc level_p {} {
puts "in level_p ([info level])"
proc level_q {} {
puts "in level_q ([info level])"
proc level_r {} {
puts "In level_r ([info level])"
return $::true
}
level_r
}
level_q
return $::false
}
puts [bool [level_p]]
Results:
in level_p (1)
in level_q (2)
In level_r (3)
True
[return] will return a code value of 2 (return) for all levels > level 0
At level 0 [return] will return code 0 (ok)
Return Code Values:
------------------------
Number Mnemonic
------------------------
0 ok
1 error
2 return
3 break
4 continue
------------------------
The standard return codes can be overridden using the '-code' argument.
This allows Tcl procedures to be written which extend the language and
interact with system codes
Whilst [return] can be used to exit structured control commands such as
[if], [for], [while], [do] etc., results may be unexpected, although you
may use the following to exit control structures:
return -code break
[return] will exit a control structure and if the code is passed back
to nesting level 0 then the script will exit
Example:
# Will exit the script
puts "I will now exit"
return 23
puts "I cannot be seen"
Example:
# This code will exit the for loop and script when $i == 5
option expression off
for {set i 0} {[< $i 10]} {++ i} {
if {[== $i 5]} {
return $i
puts "I will not be printed if i==5"
}
}
Example: ([return] cannot be used with control block)
# Does not exit using return at level 0
# Will iterate for 10 loops
set i 0
while {[< $i 10]} {
++ i
return
}
Executing a [return] command from a root-level flow-control block structure
such as [if] will not terminate the script, if you wish to do this then use
a '-code exit' argument which will return to the CLI or Windows.
Example:
set a [rnd 0 100]
if {== a 10} {return 23 -code exit}
Result:
Script is exited with 22 if random value is 10
Return will exit from other flow-control constructs
Example:
set i 0
goto_block {
start {
while {< $i 10} {
if {== $i 5} {return $i}
incr i
}
}
}
puts "i is $i"
Result
i is 5
Using [return] with a code "ok" or code 0 is the only way to return a
string-literal from [eval] without it being interpreted as a command:
Example:
set x "Hello"
puts [eval "return $x -code ok"]
Result:
Hello
[return] can return the result of a Tcl command to another command
Example:
# Will set array a to a 4 element array
array set a [return {
1 {foo}
2 {bar}
3 {baz}
4 {quux}
}]
array walk a
-------------------------------------------
hash_table object:0x3c9e10 size:250
-------------------------------------------
Bucket(73) Node: 0x20bd300 chained nodes:1
1: 0x20bd300 key:'2' => 'bar'
Bucket(113) Node: 0x20bd3f0 chained nodes:1
1: 0x20bd3f0 key:'4' => 'quux'
Bucket(165) Node: 0x20bd330 chained nodes:1
1: 0x20bd330 key:'3' => 'baz'
Bucket(220) Node: 0x20bd2d0 chained nodes:1
1: 0x20bd2d0 key:'1' => 'foo'
See also: break, continue, stop, exit, if, while, foreach
integer [replacechar variable search_char replacechar]
integer [replacechar variable search_mask replacechar -mask]
Performs an in-situ replace of characters specified by 'search' with
'replace'. Because [replacechar] operates on a variable name it avoids
the need for internal and external copy-reassignment which makes it more
efficient with very large strings (i.e. you don't need to [set] the
return value back to the variable)
Use [string replace] to replace characters within strings
Only the first character of the 'search' and 'replace' arguments are
significant unless the '-mask' argument is used. Each of these arguments
must either be a string or a character representation of an integer using
[chr]
[replacechar] should be used in preference to [string replace] with
reassignment when replacing single characters
[replacechar] takes a variable name as its first argument, not a string
this also avoids dereference and copy
A count of the number of characters replaced is returned
Example:
set s [readfile ticol.man]
replacechar s "\t" " "
Example:
set s [readfile ticol.man]
replacechar t1 [chr 0x09] [chr 0x20]
See also: string replace, string, strchr, strstr
bool [rewind handle]
Rewind the file stream pointer for a given file handle. The handle must
have been opened using [open] or [file open]
See also: tell, seek, open, file
right, string right |
Index |
string [right string n]
string [right varName n -var]
string [string right string n]
string [string right varName n -var]
Return the rightmost 'n' characters of a string
If n < 1 then an empty string will be returned
If n >= length of 'variable' then the whole string will be returned
If the string contains special Tcl characters []{}$ then call [unescape]
on the string before calling [string right]
Example:
puts [right $qbf 3]
Result:
dog
See also: left, mid, setc, setl, setr, strto, strtor, filter
double [round number ?places?] # Command version
double [funct round number ?places?] # Command version of function
double [expr round(N,places)] # Function version
Rounding to a given number of decimal places
The command [round] and function round() follow PHP's PHP_ROUND_HALF_UP
which rounds values up to 'places' decimal places away from zero.
Thus 1.1 rounds to 1.0, 1.9 rounds to 2.0
The 'tie-breaker' at 1.5 is for 1.5 round to 2.0 and -1.5 into -2.0
Example:
puts [round 1.1]
puts [round 1.5]
puts [round 1.9]
puts [round -1.1]
puts [round -1.5]
puts [round -1.9]
puts [funct round $pi 4]
puts [expr round($pi,4)]
Result:
1
2
2
-1
-2
-2
3.1416
3.1416
Example:
puts [round 1.5 0]
Result:
2
Example:
puts [round -1.5 0]
Result:
-2
See also: math, format, fpart, vfract, fraction
string [setr target-variable source-variable | source-string]
Insert source string or variable into the target, right-aligned
[setr] is based on the common BASIC RSET command and can be useful
used in combination with [makestr]
Example:
set s [makestr 20 *]
setr s "hello"
puts $s
Result:
***************hello
Example:
Blank out part of a string
set s "0849 666 616"
setr s [string repeat * 7]
puts "'$s'"
Result:
0849 *******
See also: makestr, setc, setl, mid, strstr, string
RLM Network Licence Query Plugin |
Index |
The [rlm] command may be used to check, analyse and monitor Reprise RLM
licence files using direct or UNC (SMB) file access.
To load, use: lib ticol_rlm ?-verbose? ?-nocomplain?
Prerequisite:
File read access is required to the licence file being checked
Command and Syntax:
Currently only one subcommand is supported, [rlm check] which can
analyse licence or module expiry
integer [rlm check file module ?-months|-days? ?-min|-avg? ?-var r?]
Example:
option expression off
lib ticol_rlm -verbose
set filespec "C:\\rlm\\product\\*.lic"
ls $filespec -r -array a # Get a list of RLM LIC files
if {array is_set a} {
set count [array size a]
puts "Checking $count file(s)\r\n"
for {set i 0} {< $i $count} {++ i} {
try {
set r [rlm check [escape [lindex $a($i) 3]] module01]
puts "Checking '[escape [lindex $a($i) 3]]' (result $r)"
if {> $r 0} {
set unexpired($found) "[lindex $a($i) 3] ($r days)"
++ found
} else {
textcolor red
puts "module01 has expired in [escape [lindex $a($i) 3]]"
textcolor
}
} catch {
puts "Error $r analysing file"
}
}
puts "Line: "#__LINE__
if {> $found 0} {
set msg "Found $found unexpired RLM files in this tree\r\n\r\n"
for {set i 0} {< $i $found} {++ i} {
append msg "$unexpired($i)\r\n"
} else {
set msg "All files in $filespec have expired module: module01"
}
msgbox $msg $scriptname
} else {
msgbox "No RLM files found in the current folder tree" $scriptname
}
Other Tools
-----------
A dummy version of rlmsign-x86.exe is available which can "sign"
skeleton RLM licence files to produce a visually-convincing but
wholly-invalid licence file for test use with scripts. Such fake
licences cannot be used with the Reprise rlm.exe licence server
to serve valid product licences
This is available from the author's website as: rlmsign-x86.exe
rlmtool.exe can unsign licences or create signed, valid licence
files from a template file, providing it has access to the valid
rlmsign-x86.exe on your system. Alternately it can use a template
to create a valid skeleton RLM licence which can be tweaked and
signed later on
Useful Links
------------
Technical information on the RLM licensing system is freely available
for administrators from Reprise Software
Reprise Licence Admin Bundle
http://www.reprisesoftware.com/admin/software-licensing-download.php
Reprise Licence Admin Guide (PDF)
http://www.reprisesoftware.com/RLM_License_Administration.pdf
See also: ticol_hasp, plugins
run - (CLI Command - Run a Ticol Tcl script) |
Index |
run filename?.tcl? ?argument? ?argument...?
run filename.tcx ?-password password? ?argument? ?argument...?
run # Pre-loaded script (arguments are not allowed)
See also: [clearprocs], [source]
Runs a Tcl script file from the Command Line Interpreter (CLI) shell.
[source] is the standard command to execute external Tcl code
Both [run] and [load] destroy the current program space whereas [source]
or [eval] does not. Use [source] to chain external code where the
current script and variables need to be retained. [run] functions in
much the same was as in legacy BASIC
[run filename] will load and run the specified filename and this form
will accept arguments. It will also decrypt and run obfuscated TCX files
as necessary. It may be necessary to give a password argument for some
files.
[run] will internally call [load] and [clear]. If you wish to keep
existing variables use [eval] or [source] instead.
Use [clearprocs] to clear any currently defined procedures
If a path is not given to the script then Ticol will look for the script
in the current directory
Use [run] without an argument if the file has already been loaded into
the program space using [load]. Arguments cannot be specified in this
case
[run] and [source] must NOT be used to chain from one script to another.
Use [eval] instead. Calling [run] from a running script will generate an
error.
[run] will force [option escape] to ON, [source] leaves this setting
unchanged. Entering the CLI will force this option OFF so using [source]
from the CLI will require correctly setting [option escape] in a script
Thus, any scripts which depend on processing Tcl backslash escape
sequences may not work properly if called using [source] from the CLI
Note that TCX protected files which are encoded with the /C:PASSWORD
argument will require the password to be specified using a
'-password password' argument. Files encrypted using other methods will
be automatically self-decrypting and validated. A password may be a
phrase, user-name, MAC address, workstation or Ticol version.
The resulting TCX file will be locked accordingly when created
Use [dump] to display non-encrypted (obfuscated) loaded source code at the
console.
The [run] command will clear the user-level variable table including non-
system level consts and will reset the interpreter command count
To call a child script and return to the parent use [source scriptname]
To run a script in elevated mode, use: [elevate scriptname]
Example:
ticol> run test.tcl
ticol> dump
Result:
test.tcl is executed
The script is displayed
See also: source, clearprocs, eval, load, dump, exit, call, elevate
Scripts may be run in a number of ways, either with or without arguments
Example:
ticol.exe # Load the CLI
load myscript # Load the script
dump # View it
run # Run it (no args allowed)
Example:
ticol.exe # Load the CLI
run myscript arg1 arg2 # Run script from CLI with args
Example:
ticol.exe # Load the CLI
source myscript.tcl arg1 arg2 # Run using source
Example:
ticol.exe myscript arg1 arg2 # Run from Windows CLI
See also: load, run, eval
integer [sc string1 string2 ?-nocase? ?-verbose? ?-expand?]
Compare two strings and return base 1 index-offset at which there is
a mismatch or return 0 if the strings are identical
[cmp] is a similar command but which returns a boolean value
Comparing complex strings while debugging can be difficult. [sc] simplifies
this process and gives an indication as to the location of a mismatch
'-verbose' shows the characters at the mismatch point (if any).
The '-expand' argument shows more detail regarding the mismatch point.
Example:
set s "The quick brown dog jumps over the lazy fox"
puts [sc $s $qbf -verbose]
Result:
sc: Mismatch at char 17, 'd' != 'f'
17
See also: cmp, string, debugging
string [scan input-string format-mask ?output-variable?...]
[scan] parses an input string using conversion specifiers in the style of
C sscanf and outputs into one or more variables
This command parses substrings from an input string in a fashion similar to
the ANSI C sscanf function and returns a count of the number of conversions
successfully performed or -1 if the end of the input string is reached
before any conversions have been performed
The string parameter is the input to be parsed and the format parameter
indicates how to parse it, using % conversion specifiers. Each variable is
a name of a new variable to be created or set
When a substring is scanned from string that matches a conversion
specifier, the substring is assigned to the corresponding variable. If no
variables are specified, then scan works in an inline manner, returning the
data that would otherwise be stored in the variables as a list. In the
inline case, an empty string is returned when the end of the input string
is reached before any conversions have been performed.
Any prefix or embedded text in the mask/format-specifier string must match
the parsed string (case sensitive)
printf style alignment and precision values are not supported by [scan]
e.g. %-.2s is not allowed. Width values are, for example. %2s
An asterisk (*) following the percent sign indicates that the value is to
be discarded and not stored in a variable
The Tcl [chars] and [^chars] mask forms are not supported
Scan Format Specifier
--------------------
%[*][width][length]specifier
%% Literal % character
* Placeholder. Skip this value
%b Binary value (up to 64 bits)
%c Single character
%d Decimal integer (64 bit)
%f Floating point value
%g Floating point value (general)
%i Integer (64 bit)
%n Number of characters stored in the preceding variable
%o Octal
%s String
%x Lower-case hexadecimal
%X Upper-case hexadecimal
The use of %$ is unsupported
Example:
puts [scan "hello there 1 $pi 12" {%s %*s %i %4f %2x} s t p q]
puts $s,$t,$p,$q
Result:
hello,1,3.140000,12
Example:
set ip 12.34.56.78
scan $ip %d.%d.%d.%d o1 o2 o3 o4
vars o*
Result:
o1 12
o2 34
o3 56
o4 78
Example:
# Scan an ISO date into [date] format
scan 20161122 %4i%2i%2i y m d
puts "$y/$m/$d"
Result:
2016/11/22
Example:
set string "#08D03F"
scan $string "#%2x%2x%2x" r g b
assert "$r $g $b" {== $_ "8 208 63"} -v -line #__LINE__
Example:
set result [scan $qbf "The %c%n%c%n%c%n" a p b q c r]
puts "result: $result $a $p $b $q $c $r"
puts "[chr $a][chr $b][chr $c]" # "qui"
assert "$a $b $c $p $q $r" {== $_ "113 117 105 5 6 7"} -v
assert "[chr $a][chr $b][chr $c]" {eq $_ "qui"} -v
See also: format, printf, clock scan
integer [screen width | height | attribute]
Query screen metrics
Note that Windows 10 can dynamically resize the console window and may
auto-resize a console if you drag it from one monitor to another on a
multiple-monitor system. This may cause screen formatting to become
misaligned
[screen width] and [screen height] are returned as pixels
You can query [screen width] to check for a resized console window
See [console] for specific console functions
Note that when redirecting text at the Windows console using ">" the
true console metrics are not available. [screen] will return generic
default values
Example:
puts [screen width],[screen height]
Result
1280,800
See also: console, gotoxy, box, msgbox
string [set variable-name ?value?]
string [= variable-name ?value?]
Returns, creates, assigns a value to or from a variable or displays a
variable. [=] is a shorthand alias for [set]
If 'value' argument is absent then the contents of variable are returned
(i.e. the variable will be dereferenced without the need for a $ prefix)
The variable will also be returned without backslash escapes translated
If a variable with a given value argument doesn't exist it will be created
Empty arguments may be specified using {}
Early version of Tcl lacked the '$' operator and instead used [set var] to
echo a variable's contents
To concatenate variables efficiently use append instead of chaining set
Code Description
-------------- --------------------------------------
set name # display/return variable contents
# without backslash translation
set array(subscript) # display/return array variable contents
set name {} # if name then set to ""
set name {} # if not name then create and set to ""
set name value # if name then set to var
set name value # if not name then create and set to val
Example:
option echo on
set env(windir) # Echo $env(windir)
Result:
C:\Windows
Example:
= a 23
= a
Result:
23
Struct Members
--------------
Use [struct set] to set a struct field member. See: struct set
Array Promotion
---------------
A standard variable may be promoted to an array by assignment using
[set]. No confirmation of promotion is issued. This helps facilitate
the working of [static] variables
Example:
set a Hello # Define a standard/simple variable
puts $a # Returns "Hello"
set a(10) 23 # Variable a is now type <array>
puts $a # Returns "<array>"
puts $a(10) # Returns 23
Results:
Hello
<array>
23
Variables with Embedded Whitespace
----------------------------------
It is not recommended to create and reference variables with embedded
whitespace but it is perfectly possible to do so.
Tab characters within variable names should be avoided at all cost as
these require escaping, which will be blocked by any necessary braces.
Example:
# Space
set "my var" 23
puts ${my var}
unset {my var}
Cautions
--------
Do not call with dollar prefix '$a' as follows:
set a 10
puts [set $a] # Will raise error 'set: Variable '10' not found'
Empty variables may be set using: set var {}
[set variable] will not echo contents if option echo is set to OFF. Use
[option echo on] to see console echo results
See also: $, dereference, struct set. struct, setb, static, dollar,
append, variable, variables, vars, unset, puts, echo
string [setat var string startpos len]
Sets a string or array variable directly by copying from a string with
parameters pointing to the location and length of the segment to copy from
in the source string. [setat] is similar to [mid] except that it sets a
specified variable. Out of range arguments will return a null string
Invalid startpos and range values will return an empty string
Example:
setat q $qbf 17 3 # Set q by copying from $qbf at 17 for 3 chars
puts "q is $q"
Result:
fox
Example:
setat a(0) $qbf 1 3 # Set a(0) by copying from $qbf at 1 for 3 chars
puts "a(0) is $a(0)"
Result:
The
See also: append, mid, mids, cat, cmp, performance
struct setb member-field value ?width? ?-value?
setb existingVar|newVar variable|string|expression ?width? ?-value?
Set a struct member variable to some binary/integer value which is passed
via the address of a Ticol variable. This will be in Intel 'little-endian'
order. Float values should be converted using [mkn]
An optional width value may be specified:
[addressof] [addressofb],[set],[$]
| |
v v
+---------------------+ +---------------------+
| Tcl Variable Object |->| Variable Data Block |
| | | (Dynamic) |
+---------------------+ +---------------------+
| |
v v
[ofaddress] [ofaddressb]
[addressof] [addressofb],[set].[setb],[struct setb]
| |
v v
+---------------------+ +---------------------+
| Struct member 1 |->| Struct Data Block |
+---------------------+ | (Fixed) |
| Struct member 2 |->| |
+---------------------+ | |
| Struct member N |->| |
+---------------------+ +---------------------+
| |
v v
[ofaddress] [ofaddressb]
[setb] will only operate with an rvalue which is a variable
You cannot use a literal as an Rvalue for [setb]
By default a pointer to the original value is stored in the target field
If '-value' is passed then the literal value will be stored as binary
rather than an address value (store value rather than store reference)
Retrieve [struct setb] values using [ofaddress] to dereference back
Literal integer values greater than the range of DWORD must be retrieved
back using [ofaddressb <var> 8]
You can set a buffer to NULL prefix using e.g.
struct setb buffer [chr 0] 4
Example:
struct s { a 10 b 20 c 30}
struct setb s.a 12345 # Hex is 0x003039
puts "Address:$s.a" # Will return the address-offset
puts "Value:[ofaddressb s.a]" # Will return the binary value
newline
dump s
Results:
Address:36931840
Value:12345
struct 's' level 0 (0x2338940) data @36931840 (0x2338900)
¦ 60 byte(s), 3 member(s)
+ 1:'->s.a' => 36931840 (0x2338900) 10 byte(s)
+ 2:'->s.b' => 36931850 (0x233890a) 20 byte(s)
+ 3:'->s.c' => 36931870 (0x233891e) 30 byte(s)
Address 0 1 2 3 4 5 6 7 8 9 A B C D E F ASCII
-------- ----------------------- ----------------------- ----------------
02338900 39 30 00 00 00 00 00 00 00 00 00 00 00 00 00 00 90..............
__ __ __ __
00 00 30 39 -> 0x3039 -> 12345
Comparing [struct set] with [struct setb]
-----------------------------------------
struct s {a 10}
struct set s.a 23 # Will set to binary 0x32,0x33 (text 23)
dump s
struct 's' level 0 (0x259b9b0) data @39434656 (0x259b9a0)
¦ 10 byte(s), 1 member(s)
+ 1:'->s.a' => 39434656 (0x259b9a0) 10 byte(s)
Address 0 1 2 3 4 5 6 7 8 9 A B C D E F ASCII
-------- ----------------------- ----------------------- ----------------
0259b9a0 32 33 00 00 00 00 00 00 00 00 00 00 00 00 00 00 23..............
struct s {a 10}
struct setb s.a 23 # Will set to binary 0x17 (23)
dump s
struct 's' level 0 (0x259b9b0) data @39434656 (0x259b9a0)
¦ 10 byte(s), 1 member(s)
+ 1:'->s.a' => 39434656 (0x259b9a0) 10 byte(s)
Address 0 1 2 3 4 5 6 7 8 9 A B C D E F ASCII
-------- ----------------------- ----------------------- ----------------
0259b9a0 17 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
See also: setb,
See also: struct, set, addressof, ofaddress, addressofb, ofaddressb, vars,
pointers, is_pointer, struct elements, calldll, calldll_cdecl,
struct dump, type, vars, array, stack, variables,
bool [setenv variable value]
Set an environment variable. This does not alter the main operating system
environment table. Changes are lost when Ticol exits
A variable can be cleared by calling with an empty string {} value
Example:
setenv ticol hello
puts "'[getenv ticol]'"
setenv ticol {}
puts "'[getenv ticol]'"
Results:
'hello'
''
See also: getenv
result [setresult value]
Experimental. For external DLLs only
Single Line Commands |
Index |
Commands may be sent from Windows as a single line provided they are
properly punctuated. The entire command must be wrapped in double-quotes
with individual statements separated by semi-colons. A single semi-colon
must follow this command statement
Since Tcl will evaluate and replace all variables within a single
expression it may be necessary to delay evaluation of variables in, say a
terminating while test of a do..while loop by wrapping the test expression
in extra braces, i.e.
option expression on
set i 0
do {
puts "i is $i"
if {$i == 2}
return
incr i
} while { $i < 10 }
option expression on
set i 0; do {puts "i is $i"; if {$i == 2} return; incr i} while {$i < 10 }
option expression on
set i 0; do {puts "i is $i"; if {$i == 10} return; incr i} while {{$i < 10}}
If the terminating punctuation isn't present then the command may be
executed as an external Windows command or Tcl script. A Tcl script may
follow the single-line command
Example: (Wrapped to fit the help window)
The following command calls the function PingAPI in an external DLL called
ping32.dll with various parameters set and a buffer allocated:
ticol ; "set timeout 0; set datalen 64;set buf [makestr 16]; calldll ping32
PingAPI localhost [addressofb buf]:4 datalen:2 timeout:4 1000; puts $buf"
(Note the requirement for the ";" prefix in the above command)
Result:
127.0.0.1
Example: (Wrapped to fit the help window)
set i 0; array foreach env val ss {printf "$i:\t%-20.20s=%-45.45s\r\n"
$ss $val; [incr i];}
Result:
<lists the environment table>
See also: Ticol, CLI, spawn, exec, shell
integer sgn(N)
sgn() is defined as an expression function. It may also be defined as
a Tcl proc as follows:
proc sgn x {expr {($x>0)-($x<0)}}
The [expr] function sgn() may also be called using [funct sgn N]
See also: expression functions, funct, proc, math
Single quote characters do not group statements into an argument or
characters into a string. Double quotes should be used in most cases to
group characters into a string and curly braces to group words into single
arguments. In most cases you can also use curly braces to group characters
into a string
Example: (wrong)
puts 'Hello world' # [puts] will receive 2 arguments...
# ... "'Hello" and "world'"
Examples: (correct)
puts "Hello world"
puts {Hello world}
See also: double quotes, braces
See:
array
array get
array list
array list
array sort
array foreach
arrays
lsearch
lsort
See also: faq
spawn variable ?-echo? program arg1 arg2...
spawn - ?-echo? program arg1 arg2...
Run a Windows console program with redirected output and interactive input
Output may be directed to a variable or echoed back to the console without
passing into a variable if the dummy variable '-' is given.
Return Value
------------
If a variable is given then spawn returns the number of characters read or
0 for a null return. If -echo is not used then the returnd variable may be
printed using [puts]
CTRL+C
------
CTRL+C is available for external applications, if quit using CTRL+C then
execution will be returned to the Ticol CLI shell
Arguments
---------
[spawn] expects a backslash-escaped input string
[escape] may be used to correctly backslash-format a string
[unescape] may be used to remove backslash-escapes where required
NOTE: Avoid wrapping the entire argument list in double-quotes. This is not
necessary and may result in failure to launch the command. If you quote
then ensure you quote only the command to be executed and its tail
Examples:
From the CLI with backslash escape processing ON
CLI: option escape on; option autoexec off
RIGHT: spawn - $env(comspec) /c chcp
RIGHT: spawn - -echo dir c:\\tools\\*.exe -w -od
RIGHT: spawn - -echo cmd.exe /c "dir c:\\tools\\*.exe -w"
RIGHT: spawn - notepad.exe $filename
WRONG: spawn - "notepad.exe $filename"
From the CLI with backslash escape processing OFF
CLI: option escape off
spawn - -echo cmd.exe /c "dir c:\tools\*.exe -w"
RIGHT: option escape off
spawn - -echo c:\tools\mailcmd.exe
RIGHT: spawn - -echo c:\\apps\\mailcmd\\mailcmd.exe
RIGHT: spawn - -echo $env(comspec) /c c:\\apps\\mailcmd\\mailcmd.exe
WRONG spawn - -echo "cmd.exe /c dir c:\\tools\\*.exe -w"
NOTE that when -echo is used the output text can have no colour attributes
NOTE that some Windows files cannot be spawned. e.g. qwinsta.exe
NOTE that some console applications may not choose to flush output
continuously when and this will affect '-echo'. All programs should have
their output flushed by Windows on exiting.
cmd.exe is not launched by [spawn], this is by design. Some commands such
as chcp or dir will require a command-shell prefix as these are command-
shell commands implemented within CMD.EXE and not external programs.
e.g. for:
cmd.exe /c chcp # chcp is command built into cmd.exe
You may need to input the full, translated COMSPEC value, e.g.:
C:\WINDOWS\system32\cmd.exe chcp
This may be done using the $env(COMSPEC) const instead of the full path
(case is insignificant)...
$env(COMSPEC) chcp
$env(comspec) chcp
For local buffered output control see the [flush] command
Keyboard input is unaffected and is not redirected
Example:
spawn var xping 127.0.0.1 -n:2 ; puts $var
Result:
1 00:31:40 Reply from: 127.0.0.1 00-00-00-00-00-00 (0 ms)
2 00:31:41 Reply from: 127.0.0.1 00-00-00-00-00-00 (0 ms)
Example:
puts [spawn - cmd.exe /c chcp]
Result:
Code page is: 850
Example:
option autoexec off
option escape on
puts [spawn - $env(comspec) /c set /a "_num=(2+3)*5"]
Result:
25
Example:
puts [spawn - havewe.exe c:\\ /comma /gb]
Result:
23.88978
Example:
spawn - -echo xping 127.0.0.1 -n:1 # Echo to the console
Result:
XPing: Pinging 127.0.0.1 [Snoopy] with 32 bytes
1 13:37:22 Reply from: 127.0.0.1 00-00-00-00-00-00 (0 ms)
Example:
spawn -echo xping 127.0.0.1 -n:1 # Return via command body
Result:
<no result> (the returned text is discarded)
Example:
spawn d $env(comspec) /c dir *.dll /b
Result:
Returns <list of dlls> in variable d
puts "Spawning hh.exe to load help topic '$ dereference' ..."
spawn - hh.exe "mk:@MSITStore:ticol.chm::/ticol.man.htm#$+dereference"
Calling External Script Interpreters With [spawn]
-------------------------------------------------
External script interpreters such as VBScrip or PHP may be called using
[spawn] with the console output being redirected back to a variable
Example:
unset x -nocomplain
spawn x cscript.exe //NoLogo myscript.vbs
puts $x
See also: exec, shell, flush, ticol_cscript
list [split <string> ?maskchars?]
list [split <string> maskchars ?-array var?]
list [split <string> string -literal]
list [split <string> string -literal ?-array var?]
Split a string according to mask string and return a Tcl list
If no mask character-set is given then a single space is used
The mask character strings are removed from the returned data
If the first character is matched to mask characters then it is removed
and replaced with an empty brace pair
If the last character is matched to mask character then it is removed
and replaced with an empty brace pair
If two successive mask characters are matched then the characters are
both removed and replaced with an empty brace pair
Otherwise, characters are returned according to Tcl list rules after any
mask chars are removed and the intervening text split by braces
[split] is extended in Ticol Tcl with the '-literal' argument. This adds
the ability to match literal strings rather than character sets and which
modify interpretation of the 'mask' argument. This is useful when using
say CRLF pairs to split text into an array of lines without additional
processing, or if parsing say HTML text by HTML tag strings. Using [split]
with \r\n without the '-literal' argument will mis-parse the resulting
data and introduce empty lines
Note that if you try to split using a single numeric char e.g. 0x0d, it
will not work, as the MPP will substitute the numeric string "13" and
tokenise on that. Instead you must pass [chr 13] or [chr 0x0d]
If you split a string in an array by filtering \r\n (0x0d,0d0a) then note
that because all CRLF pairs are removed, this will also remove all
empty lines comprised of only CRLFs
Example:
puts [join [split $qbf " "] "\n"] # Split and rejoin with \n newlines
Result
The
quick
brown
fox
jumps
Example:
set s "the quick\nbrown fox\njumps over\nthe lazy\ndog!"
puts [join [split $qbf " "] [chr 10]]
Result:
{the quick} {brown fox} {jumps over} {the lazy} dog!
Examples:
split "comp.lang.tcl"
split "comp.lang.tcl" "."
Results:
c o m p . l a n g . t c l
comp lang tcl
Example:
# Split CRLF-delimited text into array using literal string "\r\n"
set s "The quick\r\nbrown fox\r\njumps over\r\nthe lazy\r\ndog!"
split $s [chr 0x0d][chr 0x0a] -literal -array a
for {set i 0} {< $i [array size a]} {++ i} {
puts "$i->'$a($i)'"
}
Results:
0->'The quick'
1->'brown fox'
2->'jumps over'
3->'the lazy'
4->'dog!'
See also: join, chartok, list, array, arrays
Square brackets are used for command substitution. The contents of a pair
of square brackets is evaluated by the interpreter, and the return value
is substituted into the script
See: commands
Native stack data type (built-in / not external)
integer [stack count varName]
bool [stack create varName size]
integer [stack push varName value ?value? ... ]
string [stack pop varName]
string [stack peek varName]
void [stack foreach varName indexVar {script} ?-r?]
string [stack item varName index ?default?]
list [stack list varName]
list [stack listr varName]
integer [stack size varName]
bool [stack is_empty varName]
bool [stack clear varName]
bool [stack unset varName -nocomplain]
The stack command implements a native stack data type. The stack must first
be allocated using a call to [stack create]. The stack may be destroyed
using [stack unset] or emptied of contents and retained using
[stack clear]
Items are added using [stack push] and removed using [stack pop]. the head
of the stack may be inspected using [stack peek] and the current size of
the stack queried using [stack size]..[stack peek] is the equivalent to
[stack head]
It is possible to index the stack in normal (LIFO) or reverse (FIFO) order
by using [stack item]. Note that index values should be array base 0.
Additionally the contents of a stack can be returned as a Tcl list in
either LIFO or FIFO order by using [stack list] and [stack listr]
respectively
An existing stack may be cleared (re-initialised) to an empty stack using
[stack clear]
Stack objects will be destroyed automatically when the stack variable goes
out of scope within a procedure, but they may also be destroyed
programmatically using [stack unset s] if necessary
[stack foreach] will operate in LIFO order unless the -r (reverse) argument
is used, in which case the output will be in FIFO order
Operations on an initialised but empty stack will return an empty list {}
Stack operations on a variable not initialised as stack type will result in
a runtime error which is trappable using [catch] or [try ... catch]
A stack overflow will result in a runtime error which is trappable using
[catch] or [try ... catch]
All [stack] commands are addressed via the variable name
Example:
stack create s 1
catch {stack push s man} err; puts $err
catch {stack push s man} err; puts $err
Result:
0
Stack overflow
Stack variables must be passed to procedures by reference (name) and
accessed using [upvar] or should be accessed as global variables.
Example:
proc display_stack {x} {
puts [stack list $x]
}
stack create a 10
stack push a Hello
display_stack
See also: stack examples, array, catch, try
bool [stack create varName size]
Creates a new stack with a fixed size. It is an error to push into a stack
over the stack size. Use [stack count] and [stack is_empty] to manage
the stack size to prevent error cases. Alternatively, use [try..catch] or
[catch]
Example:
stack create stk 2
puts [stack count stk]
stack push stk Hello
puts [stack count stk]
stack push stk world
puts [stack count stk]
stack push stk Bad!
Results
0
1
2
<error>
See also: stack, stack push, stack count, stack examples
void [stack foreach varName indexVar {script} ?-r?]
Iterates stack contents, applying to a script body in the same manner as
[foreach]
Example:
stack create stk 100 # Create a stack of 100 items
unset x -nocomplain # Ensure iterator x is not set
foreach x $qbf { # Populate the stack
stack push stk $x # Push a word into the stack
}
unset y -nocomplain # Ensure iterator y is not set
stack foreach stk y { # Iterate the stack
puts "Stack item: $y"
}
Results:
Stack item: The
Stack item: quick
Stack item: brown
Stack item: fox
Stack item: jumps
Stack item: over
Stack item: the
Stack item: lazy
Stack item: dog
See also: stack, stack create, stack push, stack examples
Here are a few examples using Ticol stacks:
Example of stack creation
stack create s 1000
Example of a foreach loop pushing words from a string
foreach {x} $qbf {
stack push s $x
}
Example of simple iteration of stack with stack destruction
set i 0
while {! [stack is_empty s]} {
puts "[incr i]\t'[stack pop s]'"
}
puts "Stack empty? [bool [stack is_empty s]]"
Result:
1 'dog'
2 'lazy'
3 'the'
4 'over'
5 'jumps'
6 'fox'
7 'brown'
8 'quick'
9 'The'
Stack empty? true
Example of proc with for loop to iterate stack declared at higher level
option expression on # Use infix flow-control expressions
proc stack_iter {s} {
upvar s local
set limit [stack count local]
for {set i 0} { $i < $limit} {incr i} {
puts [stack item local $i]
}
}
Example of clearing a stack object:
stack clear s
puts "Stack size [stack size s]"
puts "Stack count [stack count s]"
puts "Stack empty? [bool [stack is_empty s]]"
Result:
Stack size 1000
Stack count 0
Stack empty? true
Example of deleting (unsetting) a stack and triggering access error
stack unset s
puts "Stack empty? [bool [stack is_empty s]]"
Result:
stack is_empty: Stack not initialised
Example: stack head:
proc head1 s { # Method 1
upvar $s t
return [stack item t -end]
}
proc head2 s {
upvar $s t
puts "[stack count t]"
return [stack item t [- [stack count t] 1]]
}
stack create s 100
foreach {x} $qbf { # Method 2
stack push s $x
}
puts [head1 s]
puts [head2 s]
Result:
dog
dog
Example: stack head:
proc tail s {
upvar $s t
if {eq [type t] "stack"} {
if {! [stack is_empty t]} {
return [stack item t 0]
}
}
return {}
}
See also: stack, array
static (static variables) |
Index |
static varname ?initialValue? # Declaring with optional initialiser
static varname -unset # Unsetting a protected static variable
A static variable is one which holds its value within a [proc] between
evocations. These are useful for counters or other accumulators within
a procedure
[static] declares a static local variable of either simple or array type
and optionally, initialises it to an initial value. Once the local static
variable has been created any initialiser value will be ignored and [set]
should, thereafter be used to set it. The new variable should be set only
in local scope. Any references to the root scope master copy of a static
variable will be discarded when a procedure exits or when [set] is used
locally. No other type is supported by the [static] command.
Static variables are global in scope with a pseudo-namespace. They are not
cleaned up when a proc exits or is undefined and are 'protected' from
deletion for the lifetime of the script unless forcibly deleted using
[static varname -unset]
The detected 'type' of the variable when first set will be 'empty'.
The variable will be set automatically to standard or array type on first
assignment [stack] and other variable types are not supported by [static]
Using combinations of [uplevel] and [upvar] other data types can be made
static within procs (see examples below)
Static variables, including parent array variables may be unset only by
using [static varname -unset]. This command 'understands' simple and
array variable types. [static <varname> -unset] understands standard
vars, arrays, stack, struct and dict types
Example:
proc foo {} {
static a # [type a] will be 'empty' on declaration
set a Hello # Sets 'a' and global '::foo::a' as string
}
static foo::a -unset
proc foo {} {
static a # [type a] will be 'empty' on declaration
set a(1) Hello # Sets 'a' and global '::foo::a' as array
}
static foo::a -unset
Tcl does not directly support the declaration of static variables
but Ticol offers a [static] command as well as an alternate method of
implementing Tcl code below. Alternatively the [global] command can be
used or [upvar/uplevel] at root level (0)
When run at [proc] level, [static] will create a root level variable with
a name in the form of 'procname::varname' and will set up a local [upvar]
alias as 'varname'. If a default value has been given then the local and
root alias will be set to that value
When run at root callframe level (0), [static] will create a root level
instance of 'varname' and will initialise it to any default value on
creation
[static] should not be used as an alternative to [set] as the argument
is for initialisation, not assignment. Any subsequent initialiser call
will be ignored
The static variable can be addressed as <procname>::<varname> when
declared in a [proc] or just <varname> if declared in root-level scope
Naming example:
# Calling from proc [foo] to declare [static q] will create
foo::q
# Which has a global name reference of
::foo::q
# And a local upvar proc scope alias of
q
# The global root copy can be read (not recommended) using
$::foo::q
# You can set both values from within proc scope using
set q value
# You should not assign directly to the root variable
# as changes will be lost when the proc stack unwinds
Example Using [static]:
option expression on # Required for if {$x}
# [proc] level 1
proc foo {{x {}}} { # Proc [foo] will set var q only if non-zero
static q 0 # Declare static q with initial value '0'
if {$x} {
set q $x
puts "foo->q was set to $q"
}
puts "foo->q::$q"
return $q
}
# [proc] level 0 (root/global scope)
assert [foo 1] {== $_ 1} -v -line #__LINE__
# Pass no arg (test for value persistence)
assert [foo] {== $_ 1} -v -line #__LINE__
static foo::q -unset
Mistakes:
Using [static] to assign where [set] is required
# In root-level scope
static foo 23 # [static] is a declaration with initialiser
assert $foo {== $_ 23} -v # PASS
static foo 11 # [static] is not assignment (init is ignored)
assert $foo {== $_ 11} -v # FAIL
set foo 11 # Use [set] to assign a static variable
assert $foo {== $_ 11} -v # PASS
Example Use:
This shows a static variable being used as an incremental counter between
proc evocations...
proc foo {} {
static x 0
++ x
puts "x is $x"
}
foo
foo
foo
puts $foo::x
static foo::x -unset
Results:
x is 1
x is 2
x is 3
3
Promoting Arrays to Static
--------------------------
Array variables may be 'promoted' to a protected static type by
'overloading; as follows, but note that this command cannot accept
an initialiser value...
array create a # Declare an array variable
static a # Promote the array to static/protected type
# An initialiser throws an error without -nocomplain
static a 23 # Wrong!! Not valid for array types
static a 23 -nocomplain # OK. Will not throw an error
Note that individual array elements may not be promoted to 'static', only
whole arrays
Using [static] within recursive procs
-------------------------------------
As each call to a recursive [proc] unwinds, it updates and erases the
global variable rather than cascading values back to each [proc] parent
and then finally back to the root [proc] as the first proc call exits
Instead use the [global] command as per the following example:
(See: unwrap)
set l "{catalog {cd {{title {Empire Burlesque}} {artist {Bob Dylan}} \
{Company Columbia} {Price 10.90} {year 1985} }}}"
proc unwrap_as_text {l} {
global foo::s # Avoid clashes with root a namespace var
foreach x $l {
if {> [ldepth $l] 2} {
unwrap_as_text $x
} else {
append ::foo::s "$x\t"
}
}
if {<= [ldepth $l] 2} {
set ::foo::s [string trim $::foo::s "\t "]
append ::foo::s "\r\n"
}
}
unwrap_as_text $l
textcolor cyan
puts $foo::s
textcolor
Results:
catalog
cd
title Empire Burlesque
artist Bob Dylan
Company Columbia
Price 10.90
year 1985
[static] implemented as a proc
------------------------------
This sets up an emulated 'namespace' var in the format
<procname>::<varname> within the parent proc. The implementation in Tcl
is relatively trivial...
option expression off
undef static # undef built-in command
# [proc] level 2
if {! [defined static]} { # Avoid errors if re-run
proc static {name {value {}}} {
if {> [info level] 1} { # If not root, will be 1
set scope [uplevel 1 {info procname}] # Save proc in new scope
set varname ${scope}::$name # Create root scope name
if {! [is_set ::$varname]} { # If root var name empty
uplevel 1 "upvar @0 $varname $name" # Run upvar in root
uplevel 1 "set $name $value" # Set local alias
} else {
uplevel 1 "upvar @0 $varname $name" # Run upvar in root ...
} # ... callframe at level 1
} elseif {! [is_set ::$name]} { # We are in root namespace
set ::$name $value # Set on init
}
}
}
# [proc] level 1
proc foo {{x {}}} { # Proc [foo] will set var q only if non-zero
static q 0
if {$x} {
set q $x
puts "foo->q was set to $q"
}
puts "foo->q::$q"
return $q
}
# [proc] level 0 (root/global scope)
assert [foo 1] {== $_ 1} -v -line #__LINE__
# Pass no arg (test for persistence)
assert [foo] {== $_ 1} -v -line #__LINE__
Static Structs
--------------
Structs can be made static within a proc using the following proc
Note that structs are time-costly to set up and require a per-proc-call
fixup
proc static_struct {name args} {
set scope [uplevel 1 {info procname}]
set varname ${scope}::$name
if {! [is_set ::$varname]} {
uplevel @0 "struct $varname {$args}"
}
uplevel 1 "upvar @0 $varname $name"
foreach {x y} $args {
uplevel 1 "upvar @0 ${varname}.$x ${name}.$x"
}
}
proc foo {x y} {
static_struct s {a 20 b 30} # Create a static struct in the proc
struct set s.a $x
struct set s.b $y
}
foo Hello world # Pass 2 arguments
puts "foo::s.a '$foo::s.a'"
puts "foo::s.b '$foo::s.b'"
Result:
foo::s.a 'Hello'
foo::s.b 'world'
Static Stacks
-------------
Stacks can be made static using the proc below
This will handle both proc and root level use
proc static_stack {name args} {
set scope [uplevel 1 {info procname}]
set varname ${scope}::$name
if {> [info level] 1} {
if {! [is_set ::$varname]} {
uplevel @0 "stack create $varname $args"
}
uplevel 1 "upvar @0 $varname $name"
} else {
if {! [is_set $name]} {
uplevel @0 "stack create $name $args"
}
}
}
proc foo {x y} {
static_stack $x 100
puts "foo: Pushing $y onto stack $x"
stack push $x $y
}
foo s Hello
foo s World
set i 1
stack foreach foo::s x {
puts "stack item $i: -$x-"
++ i
}
See:
http://wiki.tcl.tk/1532
See also: is_static, global, set, scope
stop
Halts the execution of a program launched from Windows and returns to the
Ticol CLI. When launched from the Windows console the exit command exits
to Windows. Returns code 6.
Example:
option expression off
set i 0
while {1} {
if {> $i 5} {
stop
}
puts "i is $i"
++ i
}
Results:
i is 0
i is 1
i is 2
i is 3
i is 4
i is 5
See also: exit, return, at_exit, die
Highly efficient buffered storage string builder for large strings
Available as either anonymous or named objects
Anonymous Format:
void [store add string ?-size N?] # Concatenate a string
string [store get] # Return and clear the buffer
integer [store length] # Get the current store length
string [store peek] # Peek the current contents
bool [store exists] # Test for default storage
bool [store exists {}] # Test for default storage
Named Format:
void [store add name string ?-size N?] # Concatenate a string
string [store get name] # Return and clear the buffer
integer [store length name] # Get the current store length
string [store peek name] # Peek the current var contents
bool [store exists name] # Test for specific storage
Storage Management
list [store list] # List vars and storage size
integer [store clear] # Clear *all* storage
[store] works in 2 forms. One with multiple named global storage variable
name references and the 2nd a single default, global reference with no
associated variables (i.e. 'anonymous' storage)
High-efficiency, dynamic-buffered memory concatenation command for large
strings. [store] is much faster than [append] and will pre-allocate memory
in 1 megabyte blocks, auto-resizing as necessary. [store] is about 5 times
faster than [append] due to buffered memory pre-allocation avoiding the
need for memory management calls during storage operations.
Store is useful for rapidly building complex CGI based web pages and can
achive fairly quick page generation and load times
Either a default (unnamed) buffer can be used or a named buffer. Note
that the buffer name is not a Tcl variable but a reference to any number
of buffer stores.
There is no return value from [store add], all errors are fatal and
may be trapped using [catch] or [try..catch]
The data is retrieved and the memory store and cleared using [store get]
which also clears and resets the internal memory buffer and frees all
allocated memory. It is good practise to call [store get] before first
use to ensure the buffer has been cleared. [store get] also removes
the storage variable from the table of known variables. To avoid errors
you can test for a variable's existence using [store exists]
Example:
store add foo # Add 'foo' to the default storage
store add bar $qbf # Add $qbf to storage object 'bar'
puts [store exists] # TRUE
puts [store foo] # TRUE
puts [store exists bar] # TRUE
puts [store exists quux] # FALSE
store clear # Clear all
The default internal buffer cache is 8Kb. For processing large files
such as web-pages you may wish to increase this to about 1Mb. The
cache size is used at the first call to [store add] or at the first
call after a call to [store get]
The current buffered string length can be queried using [store length]
[store length] or [store length name] returns the current length of
a specific storage variable
[store list] will return a Tcl list containing the active storage
variables and their sizes (string lengths)
Example:
store add $qbf
store add foo $qbf
store add "foo bar" Hi
puts "'[store list]'"
Results:
# These can be iterated using [foreach {a b} [store list]]
'{} 43 foo 43 {foo bar} 2'
[store clear] will clear ALL storage without further confirmation.
This includes the default (unnamed) storage. This can be used
on exit from a file called by [source] but beware that the parent
variables will also be cleared
Example:
# Default (unnamed) buffer cache
option expression off
store get # Force clear default storage
readfile ticol.man -array a
for {set i 0} {< $i [array size a]} {++ i} {
store add "$a($i)\r\n"
}
puts "Store length is: [store length]"
set q [store get]
puts "Final var length is: [string length $q] byte(s)"
Example:
# Using a named buffer cache 'b'
option expression off
store clear # Clear all storage
readfile ticol.man -array a
for {set i 0} {< $i [array size a]} {++ i} {
store add b "$a($i)\r\n"
}
puts "Store length is: [store length b]"
set q [store get b]
puts "Final var length is: [string length $q] byte(s)"
See also: append
bool [string equal string1 string2]
Compare 2 strings for equality.
Note that strings which are stored in 'escaped' format are retained in
that format internally and their string length is longer than the
unescaped version. To compare escaped strings with non-escaped strings
call [unescape] first
Example:
set s1 "Hello\tworld"
set s2 "Hello[chr 9]world"
puts "Equal? "[bool [string equal [unescape $s1] $s2]]
puts "Equal? "[bool [string equal [unescape $s1] "Hello[chr 9]world"]]
Results:
True
True
See also: eq, eqn, ne, nen
A variety of string functions are available. Some implement the standard
Tcl 'string' command with parameters. Others replicate BASIC-like
functionality. See each command separately for more details
append s v # Rapidly append variable to value...
asc s # Return the ASCII value of 1st char of a string
char value # String cast
chartok # Split a string based on characters from a mask
chr n # Return character of number N
cmp v s p # Compare strings
comma s # Format a numeric string with commas
concat v s # Concatenate strings
filter s m # Filter string s using mask m
is_alpha s # Test if a string is composed of alpha characters
is_alnum s # Test if a string is composed of alphanumeric characters
index s n # Rapid indexing of strings to return a single character
instr s c # Locate position of c in string s
join l c # Join strings to form lists
left a n # Return the left part of string a for n characters
makestr l c v # Create a new string using malloc
memset v c l # Rapidly set memory
mid a b p n # Place string b within string a at p and for n chars
mids a s n # Return section of string a starting at s for length n
range c # Return a string of characters in the given range
replacechar v s # Performs an in-situ replace of characters
right a n # Return the right part of string a for n characters
setc v s n # Insert string into var c, centrally aligned for length n
setl a b n # Place a string b left-aligned into a for n characters
setr a b n # Place a string b right-aligned into a for n characters
scan s c # Parses a string using conversion specifiers
setat s v # Sets a variable by copying from a variable or string
slice s n # Break up freeform text into fields
split s m # Slice up text
store s # Buffered concatenation
string # See the [string] command
strstr a b # Locate string b within a and return a base 1 offset or 0
# You may also specify '-nocase' as a trailing argument
stripto # Remove characters up to given position from left
striptor # Remove characters up to given position from right
strsplit # Split a string
strstr # Find location of one string in another
strtok # Split a string
trim s c # Trim both ends of a string
trimleft s c # Trim the left hand end of a string
trimright s c # Trim the right hand end of a string
wrap # Line-wrap freeform text string
See each command for more detailed information, e.g. help string trim
See also: string, string to_wchar, wchar_to_string, wchar_length, functions,
array
string is_alpha, string is_alnum |
Index |
bool [string is_alpha s] # Test if composed of alpha chars
bool [string is_alnum s] # Test if composed of alphanumeric chars
Examples:
puts [is_alpha "abcdef"]
puts [is_alpha "abc.ef"]
puts [is_alnum "abc123"]
puts [is_alnum "abc.23"]
Results:
1
0
1
0
See also: string, is_*
wide-string [string to_wchar string|input-variable output-variable]
Convert a standard ANSI string to a wide (unsigned char) string to pass to
an external DLL. It is not recommended that wide strings be used
internally with Ticol as there is no useful means of manipulating them.
string_to_wchar can only pass its output directly into another variable.
If the output variable does not exist it will be created.
Example:
string to_wchar "hello" fred
puts $fred
h
dump fred
var 'fred' at 31752880 (0x1e482b0), 1 byte(s)
Address 0 1 2 3 4 5 6 7 8 9 A B C D E F ASCII
-------- ----------------------- ----------------------- ----------------
01e482b0 68 00 65 00 6C 00 6C 00 - 6F 00 00 20 61 75 74 6F h.e.l.l.o.. auto
Example: (Uses the AutoIT v3 DLL)
# Note that when passing wchar* to a Unicode DLL requiring LPWSTR we use
# addressofb (binary address) not strptr
puts "Getting window text ..."
set buflen 4000
set buf [makestr [* $buflen 2]]
string to_wchar "Untitled - Notepad" title
puts "Check: wchar_length of title is : [wchar_length title]"
call [calldll AutoItX3 AU3_WinGetText "" [addressofb title]:4 [strptr\
buf]:4 $buflen:4]
set result [wchar_to_string buf]
puts "AU3_WinGetText wchar_to_string is: '$result'"
See also: wchar_to_string, wchar_length, calldll
string [stripto mask-string variable|string ?-nocase? ?-var?]
Strip all characters up to the given mask string. If the mask string is not
encountered then a null return will result.
If the -var argument passed then the payload argument will be assumed
to be a variable. The changed value is returned by the variable contents
will be unchanged
If a string literal is passed then the filtered string will be returned
Examples:
set x "51cb\r\n<!DOCTYPE html>"
stripto "<" x
puts "x is now: $x"
puts [stripto :: "class::name"]
puts [stripto :: "class::name" -remove]
Results:
x is now: <!DOCTYPE html>
::name
name
See also: striptor, strstr, left, right, mid, mids, strto, strtor
string [striptor mask-string variable|string ?-nocase? ?-var?]
Reverse strip all characters after the first encountered instance of the
given mask string when searched in reverse direction from the end of the
string. If the mask string is not found then the entire string will be
returned
If the -var argument passed then the payload argument will be assumed
to be a variable. The changed value is returned by the variable contents
will be unchanged
Examples:
puts [striptor bye "Goodbye cruel world"]
puts [striptor :: "class::name"]
puts [striptor :: "class::name" -remove]
Results:
Goodbye
class::
class
See also: stripto, strstr, left, right, mid, mids, strto, strtor
string [strto string maskchars ?-var? ?-nocase? ?-right?]
string [strtor string maskchars ?-var? ?-nocase? ?-left?]
[strto] will return the contents of a string up to and excluding the first
instance of the first character in the "maskchars" set (if any). If a
character is unmatched then an empty string "" is returned
[strtor] will search in the reverse direction
for [strto] the "-left" argument will return the left hand part
for [strtor] the "-right" argument will return the right hand part
A variable name may be passed using the "-var" argument for speed-intensive
code. Case may be ignored using the "-nocase" argument
These command can simplify code and speed up the splitting of strings into
component parts
Longhand code:
[strto] and [strtor] avoid moderately slow and complex longhand code
involving indexes and string-slicing such as that below...
set s "http://foo.com:80"
proc strto {s q} {
foreach x [split $q] {
set p [strstr $s $x]
puts "$s x:$x p=$p"
if {> $p 0} {
-- p
return [left $s $p]
}
}
return ""
}
puts [strto $s ":"]
Result:
http
Examples:
set s "http://foo.com:80"
puts [strtor $s ":" -left] # Get URL from port value
puts [strtor $s ":" ] # Get port value from URL
puts [strtor s ":" -left -var] # Get URL from port value (-var)
puts [strtor s ":" -var] # Get port value from URL (-var)
Results:
http://foo.com
8080
http://foo.com
8080
See also: strtok, left, right, mid, mids, stripto, striptor
integer [strtok string mask array ?-nocase?]
[strtok] 'tokenises' (breaks apart) a whole string as delimited by 'mask'
and returns the individual components in a base 0 integer-indexed array.
A count of the number of tokenised items in the array is returned
Characters from the mask string are not included in the output.
Typically a string will be split by spaces, tabs or CRLF pairs.
If the 'array' argument exists it will be overwritten without warning
Example:
option expression off
set count [strtok $qbf " " a]
for {set i 0} {< $i $count} {++ i} {puts $a($i)}
Result:
The
quick
brown
fox
jumps
over
the
lazy
dog
Example:
strtok $qbf " " a
array foreach a x {puts $x} # Will be unsorted from array a
Result:
The
dog
brown
jumps
over
the
fox
lazy
quick
Example:
strtok $qbf " " a
foreach x [array sort a] {puts $x} # Will be sorted from list
Result:
brown
dog
fox
jumps
lazy
over
quick
The
the
See also: array, arrays, chartok, strto, strtor
See topics:
struct create
struct clear
struct copy
struct dump
struct elements
struct implementation
struct item
struct set
struct setb
struct size
struct unset
->
=>
=>>
arrays of struct
[struct] offers a simple emulation of C/C++ binary structs for the purpose
of interfacing with external DLL routines. C-style Unions are not supported
A struct maps one or more Tcl variables to a fixed memory block of the
defined struct size. This block cannot be changed without destroying and
then recreating the struct. Each struct has a parent Tcl variable and one
or more Tcl 'field' member variables which are referenced by the parent
Each field has not inherent data type and is simply a reference to a
block of memory
The command [class] may be defined as an [alias] for [struct]. Ticol
does not support proper functional class objects, in fact trying to make
structs work like classes is pretty awful (See: class example)
Structs are useful for passing a large collection of variables or
properties to a [proc] when passing by reference (by name) and using
[upvar]
struct variable: [create], [unset], [clear], [copy], [size] etc.
|
v
[addressof] [addressofb],[set].[setb],[struct setb]
| |
v v
+---------------------+ +---------------------+
| Struct member var 1 | -> | "C"-style struct |
+---------------------+ | data block |
| Struct member var 2 | -> | (Fixed) |
+---------------------+ | |
| Struct member var N | -> | |
+---------------------+ +---------------------+
| |
v v
[ofaddress] [ofaddressb]
Use [setb] to write binary values to the struct field offsets or
[struct set] to write plain ANSI text values. You can only write to struct
members, you cannot can write to the host struct variable directly
Data cannot overflow the allocated field area. Any excess is truncated
A Ticol struct is designed to hold binary-compatible data. It cannot hold
Tcl abstract data types such as Stacks or associative Arrays although it
can contain a name reference to any other object type. Such data objects
would not be useful to pass to external DLLs
New structs can be instantiated from an existing struct using [new].
[new] will create a new instance and clear all member fields to NULL
Note that struct field names are limited to 256 characters
Examples of Ticol binary structs:
struct create s {{f1 10} {f2 20} {f3 30}}
struct s {
{a 4}
}
struct s {
a 23
b 45
}
struct create s 256 {
{a 40}
{b 100}
}
The contents of a struct member field is returned by dereferencing
the struct member e.g. $s.a
The field address can be accessed by using [addressofb] and [ofaddressb]
can be used to invert the binary address of the field member
Example:
struct s {a 20}
struct set s.a "Hello world"
puts "Address of member is: '[addressofb s.a] '"
puts "Contents via address is: '$s.a'"
Results:
Address of member is: '32569824'
Contents via address is: 'Hello world'
[unset] cannot be used to delete individual struct members. If you wish to
alter a struct object then you must instead delete the entire struct
object and recreate it
A useful type of struct is one which holds a single member variable.
In such cases the struct itself can be passed as an alias for the member
variable since the struct binary address offset is the same as the first
member offset. This is useful (and safer since the data-block is fixed)
for passing simple, fixed-byte-size variables to external functions via
[calldll*]
Example:
struct x {dword 4} # A single-field struct, size 4 bytes
struct setb x.dword 1234 # Set a binary value in the struct
# struct contains:
# Address 0 1 2 3 4 5 6 7
# -------- -----------------------
# 02339470 D2 04 00 00 00 00 00 00 # 0x000004D2 as little-endian
puts [ofaddressb x.dword] # Display the value
Example:
# Retrieves the true name of a Windows service (ANSI) using API calls
lib ticol_calldll # Load the calldll plugin
set service_name "PeerDistSvc" # The service we will query
const SC_MANAGER_ENUMERATE_SERVICE 0x0004
set scm_handle [calldll Advapi32 OpenSCManagerA [info hostname] 0\
$SC_MANAGER_ENUMERATE_SERVICE]
set sc_handle [calldll Advapi32 OpenServiceA $scm_handle $service_name\
$SC_MANAGER_ENUMERATE_SERVICE]
struct dummy {buflen 4} # Pass buflen by reference as long
struct setb dummy.buflen 100
struct dummy2 {buf 1024} # Pass buffer by reference via struct
struct setb dummy2.buf [chr 0] 4
set display_name ""
set bresult [calldll Advapi32 GetServiceDisplayNameA $scm_handle\
$service_name dummy2.buf:4 dummy.buflen]
calldll Advapi32 CloseServiceHandle $sc_handle
calldll Advapi32 CloseServiceHandle $scm_handle
puts "Result=[bool $bresult] $service_name=='[ofaddressb $dummy2.buf]'"
See also: struct create, struct dereference, struct copy, struct set,
struct size, struct elements, struct dump, struct item
new, pointers, is_pointer, struct setb, struct elements, calldll,
calldll_cdecl, struct dump, type, vars, array, stack, variables,
addressof, ofaddress, addressofb, ofaddressb, class example, =>
-> => =>> (struct or class dereference) |
Index |
-> struct # Dereference a struct member
=> structname.field ?length? val # Set (assign) a field member
=>> struct # Dereference and [eval] the result
[->] is an alias for [struct item]
[=>>] is an alias for [eval {[struct item]}] or [eval [-> {$x}]]
eval "[-> d.set] {[eval [-> c.get]]}"
Is the same as:
=>> d.set [=>> c.get]
Example:
# Won't handle nested procs and different levels
# Uses global scope variables v's upvar
alias class struct # Define [class]
class test {
data 256
get 20 # Sufficient size to hold method
set 20
}
proc test::constructor {x} {
upvar $x
set class_name [info procname]
striptor :: class_name -remove # Remove "::"
struct set ${x}.get "${class_name}::get ::${x}.data"
struct set ${x}.set "${class_name}::set ::${x}.data"
}
proc test::get {x} { # Proc is mapped from struct field
return [-> $x]
}
proc test::set {x} { # Proc is mapped from struct field
# upvar $x # Note that upvar is not required
return [eval "struct set $x"] # Requires {*} expand
# return [eval "=> $x"] # Alternate form
}
new test c # Create a new instance of 'test'
test::constructor c # Call the constructor
=>> c.set {$::qbf} # Call the method
puts "Object contains: [=>> c.get]"
Results:
Object contains: The quick brown fox jumps over the lazy dog
Comments:
These commands are experimental
[return] is, strictly speaking, not required from a [proc]
[struct set] can be simplified as [=>]. See =>
{*} would reduce the need to use [eval] with some operations
See also: =>, struct, class, struct item
address [struct ?create? structname ?size? {{membername size} ...}
?-init <string>?]
Creates a Ticol Tcl struct with one or more member fields
The 'create' and 'size' arguments are optional. size defines the
struct's binary block size in bytes. This is detected from the
field definition list but can be overriden via this argument.
Individual member names can be no longer than 256 characters
Struct fields exist as independent variables and are referenced via a
dot-separated, parent-referenced naming structure as in C++ (e.g. s.f1)
Struct member fields must be assigned-to using [struct set], [setb], or
[struct setb] commands. [set] cannot be used to assign to struct members
Field members may optionally be initalised to a string value which may
be specified by the -init <string> argument. Any excess over the defined
field width will be truncated on a per-field basis.
The default field initialisation is to fill with null characters
Examples:
struct create s {{f1 10} {f2 20} {f3 30}}
struct s {
{a 4}
}
struct s {
a 23
b 45
}
struct create s 256 {
{a 40}
{b 100}
}
Returns address of the new struct's data block if successful. Binary
operations on the return address must be used with extreme caution.
Ticol structs contain binary blocks of memory space but this cannot be
addressed directly other than by [struct copy]
Struct fields are addressed by struct.field naming. So for struct 's'
with field a, one would reference as: s.a
The struct memory layout is as follows:
[addressof] [addressofb],[set].[setb],[struct setb]
| |
v |
+---------------------+ |
| Tcl Variable(s) | v
+---------------------+ +----------------------+
| Struct member 1 | -> | Struct Data Block |
+---------------------+ | (Fixed memory block) |
| Struct member 2 | -> | |
+---------------------+ | |
| Struct member N | -> | |
+---------------------+ +----------------------+
| |
v v
[ofaddress] [ofaddressb]
See also: pointers, is_pointer, struct setb, struct elements, calldll,
calldll_cdecl, struct dump, type, vars, array, stack, variables,
addressof, ofaddress, addressofb, ofaddressb
Tcl doesn't naturally or easily lend itself to "C" style arrays of struct
objects, since arrays are associative and index their subscript based on
textual values; but it is possible to emulate them with some limitations.
The first of these isn't recommended for use for arrays with more than few
hundred elements. This is because this method will create individual struct
variables for each element in a pseudo-array. Ticol isn't optimised for
rapidly handling large numbers of plain variables.
Standard variable lookup uses a non-optimised LIFO/FILO linear search
array with the most recently declared variables at the "top" of the array
and newer variables are accessed first in any search. This is sufficiently
optimal for standard Tcl use but inadequate for a linear lookup of
thousands of variables, Also, the variable table is replicated in
recursive calls which would consume high levels of memory.
The method, which allowed dotted array dereferencing is as follows
This method would work just as well without the curved brackets
struct t {a 10 b 20} # Declare a struct in the standard manner
set i 0
# Create struct elements using a pseudo array which we can dereference
loop i 0 10 {
new t a($i) # This creates standard variables not elements
}
# Enter some data - we will fill with QBF
# a($i) looks like an array but it's actually a struct
# so we can dereference the fields
foreach x $qbf {
struct set a($i).a $x
struct set a($i).b [* $pi [rnd 1 23]]
++ i
}
# Print out the contents using the [->] struct dereference command
loop i 0 10 {
puts "$i .a = [-> a($i).a]"
puts "$i .b = [-> a($i).b]"
}
# Clean-up
struct unset a(*) -glob -nocomplain
# Results. Each struct is comprised of 2 more more variables
# struct a(0) <struct>
# integer * a(0).a 34179792
# integer * a(0).b 34179802
0 .a = The
0 .b = 43.9822971
1 .a = quick
1 .b = 43.9822971
# ...
Update (v1.27/2022)
-------------------
A more effective method now exists
As from November 2022 build of v1.27 array elements with struct member
references will be properly dereferenced by $ (as well as by [set] and
[struct item]) instead of the [->] struct dereference command
Any struct insteance can be mapped into an array by name and then
have struct member dereferenced directly
This lends use to [new] to create a series of unique struct varables
which may be allocated by name to array elements and also to do so
via an allocation loop.
At present this allows only complex dereferences (but not assignment)
in the format:
$arrayname(subscript).structmember
e.g.:
$a(foo).bar
This form of resolution may be blocked by enclosing the array index
in curly braces
e.g.
${a(foo)}.bar
Example:
struct template {a 10 b 20} # Underlying struct template
new template foo_1 # Create a new struct based on 'template'
struct set foo_1.a "Hello" # Assign using [struct set]
struct set foo_1.b "World"
set a(0) foo_1 # Assign this struct to array element 0
puts "$a(0).a $a(0).b" # Dereference
puts "${a(0)}.a ${a(0)}.b" # Don't dereference (see below)
Result:
Hello World
foo_1.a foo_1.b
If you wish to prevent this form of struct dereference then enclose the
array reference in curly braces thus:
${a(0)}.a
You may also set struct members via the array, as follows. Although it is
non-intuitive to dereference the array element in order to set a value
in the struct member
struct set ${a(0)}.a 23
It is not yet possible to assign to or pass such complex references to
commands. Compatibility extends solely to dollar dereferencing
See also: calling array of struct methods, new, struct, array, ->, foreach, loop
Calling methods within arrays of struct |
Index |
Here's an example script which calls methods within an array of struct
using various means.
option expression off
struct template { # Template struct used for instantiation
a 100
b 30
}
# Instantiation
new template foo_1 # Duplicate template struct into a new one
new template foo_2 # Duplicate template struct into a new one
# Script/method definition
set script {static i 1; puts "Hello $i"; ++ i}
sub foo $script # Assign our script to a [sub] called 'foo'
# Binding the array element to a struct instance
# This is a simple demo and we use only one array element
# array set a 0 foo_1 # Alternate method of setting (1)
set a(0) foo_1 # Alternate method of setting (2)
puts $a(0) # Visual confirmation of object link
# Assigning member functionality
struct set ${a(0)}.a $script # Assign a scriptlet as a "method"
struct set ${a(0)}.b foo # Point to a sub/proc aka a "method"
puts ${$a(0).a} # Will resolve to the method source
# Various examples of calling our 'methods'
# Calling member 'a' ...
expand ${$a(0).a} # Method 1 (Call with expand command)
{*}${$a(0).a} # Method 2 (Call via {*} operator)
eval ${$a(0).a} # Method 3 (Call with eval)
=>> ${a(0)}.a # Method 4 (Call using =>>)
@> ${$a(0).a} # Method 5 (Call with 'eval subst')
# ${$a(0).a} # Won't work as '.a' won't be expanded
# Calling member 'b' ...
${$a(0).b} # Will work as it calls a single Tcl word
Results:
Hello 1
Hello 2
Hello 3
Hello 1
See also: arrays of struct, struct, array, class, class example
Ticol does not support objects or classes but these can be emulated via
the [struct] data type
A full, worked example of a pseudo-class using Ticol which a progress-bar
objects and updates it. Multiple such objects may be created and updated
independently. This form of emulation offers no internal methods, only
links to external ones
option expression off
cls
proc progress_bar {s i} {
#################################################
# [in] s Struct (as pseudo-class object)
# [in] i Current progress value to use
#################################################
upvar $s
if {ne [type $s] struct} {
die "[info procname]: arg s is not a struct"
}
# [->] is complex struct member dereference
# $${struct}.member or ${${struct}.member} is proposed
set x [-> ${s}.x] # Alternate to ${${s}.y}
set y ${${s}.y}
set style ${${s}.style}
set init ${${s}.init}
set scale_max ${${s}.scale_max}
set width ${${s}.width}
set colour ${${s}.colour}
if {! $init} {
if {== $style 0} {
box $x [- $y 1] [+ $scale_max 5] 3 black -nofill
gotoxy [+ $x 2] $y
for {set j 0} {< $j $width} {++ j} {
textcolor $colour
if {is_mod $j 2} {
puts \xdb -nonewline
} else {
puts " " -nonewline
}
}
clreol
} else {
box $x [- $y 1] [+ $scale_max 5] 3 grey -nofill
}
=> ${s}.init $::true
}
set n [* [+ $width 1] [/ [double $i] $scale_max]]
set percent [* [/ [double $i] $scale_max] 100.0]
gotoxy [+ $x [+ 2 $n]] $y
if {> $style 0} {
if {> $percent 90} {
textcolor green
} elseif {> $percent 60} {
textcolor yellow
} else {
textcolor red
}
puts [chr 219] -nonewline
} else {
textcolor $colour
puts [chr 219] -nonewline
}
# Print n.nn %
textcolor
gotoxy [/ $width 2.0] [- $y 1]
textcolor white
printf " %.2f %% " $percent
textcolor
if {>= $percent 100} {
=> ${s}.init $::false
}
}
alias class struct
# Vars are declared wide enough to hold the max value
class template {
init 10
style 10
x 10
y 10
colour 10
width 10
scale_max 10
method 20
}
# Create a new struct (class) instance and clear to NULL
new template st
# Use struct member assign [=>] or the verbose [struct set] to assign
=> st.style 1
=> st.x 1
=> st.y 6
=> st.width 74
=> st.scale_max 400
=> st.init 0
=> st.method progress_bar # Declare a method by name only
gotoxy 1 1
textcolor magenta
puts "Pseudo-class demo using a progress-bar"
textcolor
for {set i 0} {<= $i $st.scale_max} {++ i} {
gotoxy 1 3
puts "Scale value: $i" -nonewline
[$st.method st $i]
sleep 5
}
sleep 500
# Reset some style parameters ...
=> st.x 4
=> st.y 12
=> st.style 0
=> st.width 54
=> st.colour cyan
=> st.method {progress_bar st $i} # Set a combined method + args
# And draw again ...
for {set i 0} {<= $i $st.scale_max} {++ i} {
gotoxy 1 3
puts "Scale value: $i" -nonewline
=>> st.method # Member evaluation command
# @> $st.method # Experimental method call command
# eval [subst $st.method] # Cumbersome equivalent to [@>]
sleep 20
}
See also: class, struct, new, =>, =>>, ->
Experimental @> Struct Command |
Index |
@> structname.member
The experimental [@>] command is a shorthand for combined [eval [subst]]
This is required as Ticol v1.26 does not currently offer $ dereference
calling of struct methods, only dereferencing of variable contents.
This mechanism may be deprecated in future relases of Ticol
Example:
class template {
method 20
}
new template st
struct set st.method "puts hello"
=>> st.method # Member evaluation command
# @> $st.method # Experimental method call command
# eval [subst $st.method] # Cumbersome equivalent to [@>]
Result:
Hello
See also: struct, =>>, ->
string [struct set structname.fieldname]
address [struct set structname.fieldname value ?length?]
$structname.fieldname # Dereference field member
Sets a string value to a struct field. This may be less than the field
length and as specified by the optional length argument
[struct set] with no arguments is an alias for [struct item]
[struct set] with arguments returns an integer containing the address
of the struct for use with lib ticol_calldll
[set] also performs the same function
The operator '=>' is an alias for [struct set]
Example:
struct s {{f1 20} {f2 10}} # Declare a struct with 2 fields
struct set s.f1 "Hello world" # set adjusts to assign a pointer var
puts [struct set s.f1] # Return the contents of s.1
puts $s.f1 # Return the contents of s.1
puts [set s.f1] # Dereference the pointer variable
puts [ofaddress s.f1] # Dereference the pointer variable
Results:
Hello world # [struct set] dereference
Hello world # Dollar struct member dereference
Hello world # Plain [set] dereference
Hello world # [ofaddress struct.field] will dereference back
Struct shortcut dereference operators ->, => and =>> are available
See also: struct item. struct, struct dereference, ->, =>, =>>
integer [struct size structname]
Returns the total size of a struct or of a struct's data field
Example:
struct s {{f1 10} {f2 20} {f3 30}}
puts [struct size s]
puts [struct size s.f1]
puts [struct size s.f2]
puts [struct size s.f3]
Returns:
60
10
20
30
Field Sizes and Unknown Data Size
---------------------------------
Fixed size fields, combined with an unknown variable size can be a
problem when assigning to struct fields. In such cases the struct field
can assign the address of a variable instead. This should be no more
than 10 characters long. Pass a variable name to [addressof] to store
the address. Use [ofaddress] to revert from address back to content
The variable must not be reassigned during the lifespan of the struct
field
Example:
# Can't take address of a const
set qb $qbf # So use a variable
struct set foo.a [addressof qb] # e.g. address is 36337072
puts [ofaddress $foo.a] # Dereference the address back
Result:
The quick brown fox jumps over the lazy dog
See also: pointers, is_pointer, struct setb, struct elements, calldll,
calldll_cdecl, struct dump, type, vars, array, stack, variables,
addressof, ofaddress, addressofb, ofaddressb
list [struct elements structname ?-size | -value?]
Returns a Tcl list containing the names of a struct's elements
(fields) which can be used to emulate [struct foreach]
Optionally a list of struct sizes or values (as addresses) can be returned
Examples:
struct s {{f1 10} {f2 20} {f3 30}}
puts [struct elements s]
foreach x [struct elements s] {
puts $x
}
Returns:
s.f1 s.f2 s.f3 # A Tcl list of member elements of s
s.f1
s.f2
s.f3
See also: pointers, is_pointer, struct setb, struct elements, calldll,
calldll_cdecl, struct dump, type, vars, array, stack, variables,
addressof, ofaddress, addressofb, ofaddressb
struct clear structname
Erase a Ticol struct memory block. Cannot be used on binary structs
returned from calldll_* calls
[struct clear] has no return value
Ticol structs are cleared to NULL when created. There is no need to
use struct clear after defining a struct. Typically, a struct is
cleared after a calldll_* API call
lib ticol_calldll
struct s {f1 10}
set ts_ptr [calldll_cdecl msvcrt40.dll localtime t]
struct copy s $ts_ptr
struct clear s
See also: pointers, is_pointer, struct setb, struct elements, calldll,
calldll_cdecl, struct dump, type, vars, array, stack, variables,
addressof, ofaddress, addressofb, ofaddressb
struct copy dest-structname source-address
[struct copy] is provided for use with lib ticol_calldll and in
linking to external DLLs
Use with care: Copies 'struct bytes' count from a struct* memory
address as returned by calldll_*. A copy may be required to use the
return values from some external DLLs or Windows API functions
[struct copy] returns the number of bytes copied to the target struct
Careless use of [struct copy] may result in a fatal exception although
struct field boundaries are protected from being overwritten
Ensure any struct being copied from a DLL address is from a DLL which
is opened persistently by using [dll_open]
Example:
Calling functions time, localtime and asctime in msvcrt40.dll
lib ticol_calldll # Load our calldll plugin
struct tm { # Define a Tcl struct which can be mapped
tm_sec 4 # int seconds after the minute 0-59
tm_min 4 # int minutes after the hour 0-59
tm_hour 4 # int hours since midnight 0-23
tm_mday 4 # int day of the month 1-31
tm_mon 4 # int months since January 0-11
tm_year 4 # int years since 1900
tm_wday 4 # int days since Sunday 0-6
tm_yday 4 # int days since January 1 0-365
tm_isdst 4 # int Daylight Saving Time flag
} # Where 'int' is a 32-bit long
set t 0
# Clear the struct data-block to zero
struct clear tm
# Load the DLL
set handle [dll_open msvcrt40.dll]
# Call the MSVC API to get the integer C time
calldll_cdecl $handle time t
# ((time_t)-1) is an error
if {== $t -1} {
puts "CRT mktime API call failed at line "#__LINE__
exit
}
# Return a struct* tm* address from the time integer
set ts_ptr [calldll_cdecl $handle localtime t]
# Copy the returned binary struct* to our Tcl struct
# Will not copy more bytes than struct can accommodate
struct copy tm $ts_ptr
# Get ASCIIZ string using struct tm: (current time)
set ret [calldllstr_cdecl $handle asctime tm]
# Release the DLL
dll_close $handle
# An ASCIIZ string is now in $ret
textcolor yellow
puts "C++ asctime(now): '[string trim ret [chr 10]]'"
textcolor
Result:
C++ asctime(NOW): 'Sun Sep 04 19:21:40 2016'
One struct can be copied to another using [struct copy] providing that
the struct sizes match. Data is truncated if there is a size mismatch
struct s {a 43}
struct t {a 20}
struct set s.a $qbf # Quick brown fox...
puts $s.a
puts "Copied [struct copy t $s] byte(s)" # $s returns struct address
puts $t.a
Result:
The quick brown fox jumps over the lazy dog
Copied 20 byte(s)
The quick brown fox
See also: pointers, is_pointer, struct setb, struct elements, calldll,
calldll_cdecl, struct dump, type, vars, array, stack, variables,
addressof, ofaddress, addressofb, ofaddressb
struct dump structname
Produce an analysis of the struct structure, member variables,
contents and hierarchy
Example:
struct s {f1 15}
struct set s.f1 "Hello world"
struct dump s
Result:
struct 's' level 0 (0x21e2480) data @35529840 (0x21e2470)
| 15 byte(s), 1 member(s)
+ 1:'s->s.f1' => 35529840 (0x21e2470) 15 byte(s)
Address 0 1 2 3 4 5 6 7 8 9 A B C D E F ASCII
-------- ----------------------- ----------------------- -----------
021e2470 48 65 6C 6C 6F 20 77 6F-72 6C 64 00 00 00 00 45 Hello world.
* Var 0::'s' has no parents
See also: pointers, is_pointer, struct setb, struct elements, calldll,
calldll_cdecl, struct dump, type, vars, array, stack, variables,
addressof, ofaddress, addressofb, ofaddressb
string [struct item structName.fieldname ?default?]
string [struct set structName fieldname ?default?]
[struct item] is an alias for [struct set]
Retrieves an item (member field) by directly accessing the struct object
instead of using $ to dereference. This may be useful when handling complex
or multi-layer variable references
[struct set] without a value is analogous to [set var] and will return the
member field's contents
[upvar] is not required for struct members if the parent struct object
is within scope. A default return may be specified where the field
is not found
Example:
struct foo {
a 10
b 20
}
struct set foo.a "Hello"
struct set foo.b "World"
puts "[struct set foo.a] [struct item foo.b]"
Result:
Hello World
See also: pointers, is_pointer, struct setb, struct elements, calldll,
calldll_cdecl, struct dump, type, vars, array, stack, variables,
addressof, ofaddress, addressofb, ofaddressb, passing structs
Struct Implementation Details |
Index |
Ticol structs are provided primarily to interface with external DLLs via
the calldll_* plugin library. The design maps a series of struct 'member
field' variables, each of which hold a value representing a starting
address-offset in a block of binary memory space. This allocated space
being shared with DLLs etc. and where the actual data is stored.
As such, struct member variables use a form of 'indirection'
Example:
This can be seen via a [struct dump] of a simple struct
struct s {{f1 10} {f2 20} {f3 30}}
struct dump s
Results:
Here we can see struct 's' which has a data-block allocated at starting
address 33082528 (0x1f8cca0). The first struct member (s.f1) points to this
starting address. The second (s.f2) points to address 33082538 (0x1f8ccaa),
which is +10 bytes higher than the first. The last member (s.f3) points to
address 33082558 (0x1f8ccbe), at 20 bytes higher, and has a reserved storage
space of 30 bytes. When calling a DLL the struct address 33082528 is
passed to a DLL along, if necessary, with the size [struct size s], which
in this case is 60 bytes total
--
var type is 3 v->val is 33082528
struct 's' level 0 (0x1f8ccf0) data @33082528 (0x1f8cca0)
¦ 60 byte(s), 3 member(s)
+ 1:'s->s.f1' => 33082528 (0x1f8cca0) 10 byte(s)
+ 2:'s->s.f2' => 33082538 (0x1f8ccaa) 20 byte(s)
+ 3:'s->s.f3' => 33082558 (0x1f8ccbe) 30 byte(s)
Address 0 1 2 3 4 5 6 7 8 9 A B C D E F ASCII
-------- ----------------------- ----------------------- ----------------
01f8cca0 00 00 00 00 00 00 00 00 - 00 00 00 00 00 00 00 00 ................
01f8ccb0 00 00 00 00 00 00 00 00 - 00 00 00 00 00 00 00 00 ................
01f8ccc0 00 00 00 00 00 00 00 00 - 00 00 00 00 00 00 00 00 ................
01f8ccd0 00 00 00 00 00 00 00 00 - 00 00 00 00 74 65 5F 74 ............te_t
--
Write strategy
--------------
When a write is performed to a struct member field the allocated block size
for that field is erased and the incoming data copied to that part of the
struct data-block area. The command [struct setb] (set byte) allows a
specific byte width to be specified on write. This must be < than the struct
member field size.
Write-Back on Proc Exit
-----------------------
Struct updates are written back by updating the whole struct data-block when
the stack unwinds during proc exit
See also: struct, structs, struct dump, dump
struct unset structname ?structname...? ?-nocomplain? ?-const? ?-glob?
Deletes all members of a struct and the enclosing struct, releasing
all memory. An error will be raised if the struct does not exist or
if the variable is not a struct type unless '-nocomplain' is used
A const struct may be deleted using the '-const' argument
Multiple struct names may be passed
You cannot unset an individual struct member field
Using structs with calldll_*
Structs may be passed by name or using [addressof]
See: calldll
You can also delete a struct using [unset]
See also: pointers, is_pointer, struct setb, struct elements, calldll,
calldll_cdecl, struct dump, type, vars, array, stack, variables,
addressof, ofaddress, addressofb, ofaddressb
struct setb variable | value ?byte-width?
Write a variable in binary format to a struct member field
You must correctly supply the byte-width interpretation of the data for
the data to be stored correctly. 4 (long) is the default
'ofaddressb' is the inverse
See also: pointers, is_pointer, struct setb, struct elements, calldll,
calldll_cdecl, struct dump, type, vars, array, stack, variables,
addressof, ofaddress, addressofb, ofaddressb
integer [sizeof variable-type]
Returns the interpreted (estimated) byte-size of a variable. This may be
useful when using the calldll* functions to pass data of a specific size
Example:
set a 10
puts [sizeof [type a]]
Returns:
4
Example:
puts [sizeof integer]
Returns:
4
Note that this value is an estimate and does not represent the actual
storage
See also: type, vars, char, int, int64, double, addressof, addressofb,
strptr
sleep milliseconds
Sleep for a given number of milliseconds
A value of 0 for milliseconds causes Ticol to yield to the operating system
for as long as necessary
A value > 1 will sleep for approximately that number of milliseconds
Negative values are not allowed and will be rounded up to 0
The precise behaviour will be operating system version dependent. Unless
there is a reason to sleep for a precise time then [sleep 0] is recommended
Example:
time {stack create s 100 ; stack unset s; sleep 0} 10000 # Sleeps for 7 ms
Note that [time code] will indicate performance in microseconds
(1/1000th ms)
For more information, see:
https://msdn.microsoft.com/en-us/library/windows/desktop/
ms686298(v=vs.85).aspx
https://randomascii.wordpress.com/2013/04/02/sleep-variation-
investigated/
See also: doevents, ticks
Standard Tcl string processing commands
See individual sections for more information:
string toupper string ?first? ?last? Convert chars in range to upper-case
string trim string ?charmask? Trim string Default mask is space
string trimleft string ?charmask Trim string left only. Mask as above
string trimright string ?charmask? Trim string right only. Mask as above
string tolower | toupper string Convert chars to lower/upper case
string index Return the character at a given index
string initcaps Word 1st chars forced to uppercase
string is_empty string | var -var Returns a boolean
string length string Return length of a string
string left string|var n ?-var? Return left n chars of string or var
string map token-list var|string Remap tokens using a map list
string match ?-nocase? pattern string Glob match a pattern to a string
string mids string|var start n -var Return middle section of string/var
string range string first last Get a sequence of chars from string
string repeat char ?n? Get a string of repeated characters
string replace string this that Replace 'this' with 'that' in string
string reverse string|var -var Reverse a string or variable
string right string|var n -var Get rightmost n chars of string/var
string slice string index... Slice a string into a Tcl list
string toarray string arrayname Convert a CRLF delim string to array
string totitle string Convert a string to initcaps
See also: string functions
See also: strings, strsplit, map, left, right, mid, mids, chr, asc, vars
filter
string [string index string indexValue]
Returns the character of a string at the given index position
Example:
puts [string index $qbf 4]
Result:
q
See also: string, filter
bool [string is_empty string]
bool [string is_empty var -var]
Test whether a variable or string is empty and return a boolean 1 or 0
Example:
puts [bool [string isempty $qbf]]
puts [bool [string isempty {}]]
Results:
False
True
See also; string
bool [string match pattern string ?-nocase?]
Match a string by "glob" matching where the following rules are used
? Match any single character
* Match more than 1 character
<else> else match a literal character
Substrings may be matched by wrapping the search pattern in asterisks,
otherwise an exact match will be performed
Character groups, e.g. [a-z], are not yet implemented
Returns a boolean indicating a successful match
Example:
puts [string match qu??k quick] # Exact match
puts [string match qu??k QUICK -nocase] # Exact match, no case
puts [string match d*g dog] # Exact match on 'dog'
puts [string match b???n brown] # Exact match on 'brown'
puts [string match "*fox*" $qbf] # Substring match - no spaces
puts [string match "*cat*" $qbf] # Substring match - no spaces
puts [string match "* f?x *" $qbf] # Substring match with spaces
Result:
1
1
1
1
1
0
1
See also: string, string trim, string replace, string slice, filter
string [string repeat char ?count? ?separator?]
Returns a repeated string of zero or more characters. If the count value
is omitted then 1 character of the specified type is returned.
If 'char' is s string then the first character of the string is used
If 'count' is zero or less than no characters are returned
An optional separator string may be specified of one or more characters
Example:
puts [string repeat x]
puts [string repeat x 0]
puts [string repeat x 10]
puts [string repeat x 5 -]
puts [string repeat $qbf 3]
Results
x
xxxxxxxxxx
x-x-x-x-x
TTT
See also: string, string replace
string [string replace string this that ?-noescape?]
Replaces instances of 'this' with 'that' in 'string' and returns the
modified string. String replace is inefficient. See [replacechar] to
replace within simple variables
Argument 'that' may be null (empty), in which case, 'this' will be removed
Note that [string replace] takes a string and does not alter the string
argument. You must reassign using any returned result
The default behaviour is to escape Tcl backslash escapes within the find
and replace arguments. This can be overridden using the -noescape argument
Example:
set t1 [readfile ticol.man]
set t1 [string replace $t1 "\t" " "]
puts [expr [string replace 2,100 , ""]*2] # 4200
See also: replacechar, string, filter
string tolower, string toupper, tolower, toupper |
Index |
string [string tolower <string>]
string [string toupper <string>]
string [string tolower <string>]
string [string toupper <string>]
String case conversion, short-form versions can be used as a slightly
more efficient alternative
Examples:
puts [string toupper $qbf]
puts [tolower "HELLO WORLD"]
Results:
THE QUICK BROWN FOX JUMPS OVER THE LAZY DOG
hello world
See also: string, filter
Commands:
A Tcl script is a string containing one or more commands. Semi-colons and
newlines are command separators unless quoted as described below.
Close brackets are command terminators during command substitution
(see below) unless quoted.
Evaluation:
A command is evaluated in two steps. First, the Tcl interpreter breaks the
command into 'words' (arguments or parameters) and performs substitutions
as described below. These substitutions are performed in the same way for
all commands
Secondly, the first word is used to locate a command procedure to carry out
the command, then all of the words of the command are passed to the command
procedure. The command procedure is free to interpret each of its words in
any way it likes, such as an integer, variable name, list, or Tcl script.
Different commands interpret their words (arguments) differently
Words (parameters):
Words of a command are separated by white space (except for newlines, which
are command separators).
Double quotes:
If the first character of a word is double-quote (") then the word is
terminated by the next double-quote character. If semi-colons, close
brackets, or white space characters (including newlines) appear between
the quotes then they are treated as ordinary characters and included in
the word. Command substitution, variable substitution, and backslash
substitution are performed on the characters between the quotes as
described below. The double-quotes are not retained as part of the word
Argument expansion:
If a word starts with the string {*} followed by a non-whitespace
character, then the leading {*} is removed and the rest of the word is
parsed and substituted as any other word. After substitution, the word is
parsed as a list (without command or variable substitutions; backslash
substitutions are performed as is normal for a list and individual internal
words may be surrounded by either braces or double-quote characters), and
its words are added to the command being substituted.
For instance,
cmd a {*}{b [c]} d {*}{$e f {g h}} # is equivalent to
cmd a b {[c]} d {$e} f {g h}
Braces:
If the first character of a word is an open brace ({) and argument
expansion rule above does not apply, then the word is terminated by the
matching close brace ('}'). Braces nest within the word: for each
additional open brace there must be an additional close brace (however,
if an open brace or close brace within the word is quoted with a backslash
then it is not counted in locating the matching close brace).
No substitutions are performed on the characters between the braces except
for backslash-newline substitutions described below nor do semi-colons,
newlines, close brackets, or white space receive any special interpretation
The word will consist of exactly the characters between the outer braces,
not including the braces themselves.
Command substitution:
If a word contains an open bracket ([) then Tcl performs command
substitution. To do this it invokes the Tcl interpreter recursively to
process the characters following the open bracket as a Tcl script. The
script may contain any number of commands and must be terminated by a
close bracket (]). The result of the script (i.e. the result of its last
command) is substituted into the word in place of the brackets and all of
the characters between them
There may be any number of command substitutions in a single word.
Command substitution is not performed on words enclosed in braces.
Variable substitution:
If a word contains a dollar-sign ($) followed by one of the forms described
below, then Tcl performs variable substitution: the dollar-sign and the
following characters are replaced in the word by the value of a variable.
Variable substitution may take any of the following forms
$name
${name}
${<complex-nested-dereference>}
$name(index)
${${<complex-nested-dereference>}(index)}
See:
https://en.wikipedia.org/wiki/Tcl#Syntax_and_fundamental_semantics
See also: dereference, nested dereferences, Ticol, Tcl
Ticol Tcl should not be used in situations which require high levels of
security. It should be treated as alpha or beta quality and used
cautiously in such situations and only after intensive testing
(See licence terms)
The following issues should be considered:
Guard User Inputs
-----------------
Ensure that user inputs from say [input] are guarded from malicious
execution. Either parse returns carefully or wrap in {} braces
to prevent automatic execution.
Command Blocking
----------------
Lock down undesired commands using ticol.ini -> CommandUndefList=
Undefine commands such as: [@], [spawn], [exec], [shell], [run]
e.g. CommandUndefList=spawn exec shell run
Lock down shell execution of unknown command via
ticol.ini -> AutoExecCommand=TRUE
The following security issues are addressed:
TCX Code Obfuscation
--------------------
This uses moderate strength symmetric encryption for script obfuscation
and is not designed to be "uncrackable". It should, however, be secure
enough for most uses
An MD5 feature is available to allow MD5 signatures of TCX code snippets
and obfuscated TCX files to be checked and validated
See: md5, md5 verification
Due to its interpreted nature and the introspective nature of Tcl it is
not easy to restrict access to obfuscated code. Some TCX content could
leak out via individual code-snippets but the full source code should
not be casually available. The /ECHO argument and [dump] command are
disabled for TCX code. Also [load script.tcx varname] is disallowed
Encryption
----------
The [encrypt] command uses moderate strength symmetric encryption and is
not designed to be "uncrackable". It should, however, be secure enough for
most uses and to prevent all but the most determined attacker
CGI Security
------------
Ticol should be used in public-facing CGI mode with extreme caution and
it may be far-more preferable to use PHP or PERL.
The command-line /EXPR is disabled if Ticol is detected as running in CGI
webserver mode, e.g.
http://localhost:8800/cgi-bin/ticol.exe?/expr%2022/7.0
ticol.ini should be configured with an entry to block undesired commands
such as [cli], [exec] or [spawn]. A sample CGI ticol.ini is shown below.
-------------------------------------------
[Config]
AutoExecScript=FALSE ; (autoexec.tcl)
CGIDebug=False
StartupLog=True
PreprocessorVerbose=FALSE
StackTrace=FALSE
CommandUndefList="cli exec spawn"
-------------------------------------------
See: command line
Individual commands: e.g. ticol.exe ; "puts Hello"
are disabled if Ticol is detected as running in CGI webserver mode, e.g.
http://localhost:8800/cgi-bin/ticol.exe?%3b&%20puts%20{hello%20world}
See: command line
In Tcl everything is a string. Interpretation of numbers is automatic
Tcl strings may be unquoted or quoted using double quote characters
puts Hello # Assumed to be a literal unquoted string
is the same as:
puts "Hello" # Explicitly declared and quoted string
However, strings which are wrapped in double quotes are protected from
alteration by the Macro PreProcessor (MPP) and unquoted strings cannot
contain whitespace characters such as " " or \t
puts Hello World # Invalid, will trigger an error. Whitespace not allowed
puts "Hello World" # Valid quoted string with embedded whitespace
Because "everything is a string", (including numbers), these are valid:
puts [+ 1 "1234"] # -> 1235
puts [+ 1 "0x1234"] # -> 4461
puts [+ 1 "0o1234"] # -> 669
puts [+ 1 0x1234] # -> 4461
puts [+ 1 0o1234] # -> 669
puts [+ 1 0b10011010010] # -> 1235
Strings Within Strings
----------------------
Under most circumstances, adjacent strings will merge. Embedding double-
quotes within a quoted string effectively splits the string in two.
Whether or not these halves merge depends on what intervenes between them.
Embedded double-quotes should always be escaped using a backslash.
Embedding double quoted strings within double quoted strings should be
avoided at all cost. There may be cases where it can work but it is
recommended to avoid double-quoting and use braces to delimit strings
that is, if no variable references need be substituted or embedded
command called.
Some Examples:
puts Hello-world # An undelimited string
puts "Hello world" # A quoted string
puts {Hello world} # A braced string
puts "Hello"" world" # Adjacent strings merge
puts "An accidentally "embedded" string" # Not recommended
puts "Goodbye \"cruel\" string" # Properly-escaped \"
puts "Brave [chr 34]new[chr 34] world" # Embedded " chars
puts {Another "embedded" string} # No variables present
puts "Value of s is $s" # $s is substituted
puts {Value of s is $s} # $s isn't substituted
puts [subst {Value of s is $s}] # $s is substituted
puts "Random is [rnd 1 10]" # [rnd] is called
puts {Random is [rnd 1 10]} # [rnd] isn't called
puts [subst {Random is [rnd 1 10]}] # [rnd] is called
puts "Substring is {braced}" # Braces appear in string
See also: quoting hell, string, Macro PreProcessor. subst, braces, escapes
string [string range string first last|end]
Returns a range of consecutive characters from string, starting with the
character whose index is first and ending with the character whose index is
last
The index-reference values are "base 0' with an index of 0 referring to the
first character of the string first and last may be specified as for the
index method. If first is less than zero then it is treated as if it were
zero, and if last is greater than or equal to the length of the string then
it is treated as if it were end. If first is greater than last then an
empty string is returned.
Invalid arguments for first and last, e.g. overlapping, will return SUCCESS
(code 0) but an empty string
Example:
string range $qbf 10 18 # Return chars from char 10 to 18 inclusive
string range $qbf 12 6 # Illegal (overlapping) range == empty list
string range $qbf 0 1024 # Returns the whole string (over length arg)
string range $qbf 0 0 # Returns the first character (0 to 0)
string range $qbf 0 -1 # Invalid range
string range $qbf -1 0 # Invalid range
string range $qbf -1 -1 # Invalid range
string range $qbf 0 # Invalid arguments (arg missing)
Result:
brown fox
{}
The quick brown fox jumps over the lazy dog
T
See also: string, filter
list [string slice <string> startpos ?next-slice? ...]
Breaks up a string into a well-formed Tcl list by 'slicing' it at fixed
offsets. This command may be used to break formatted, tabular text, possibly
delineated by CRLFs into list fields. During the process, whitespace is
removed from the source string and a single space added as a list-separator
The slice is applied to each line where data has CRLFs
The data MUST be in a regular (non-varying) columnar format and without
empty "slices to avoid "spiralling". Tab-separated data may not slice well
If the data is irregular, non-columnar, or has a non-regular header the
data will 'spiral' out of sequence
If no index is given or if the index value is specified as 0 then the whole
string is returned
[string slice] copies characters up to and including each numeric argument
with the first characters assumed, it returns a list element and then
continues to the next numeric argument until the string is exhausted.
To visualise this, imagine slicing a loaf of bread with the knife at the
slice points, leaving a slice to the left
Values must be in ascending order or an error will be raised
Values to the left of the given offset are returned for each offset value
Argument index values larger than the list length are ignored
The index values are a base 1 index-offset into the string not base 0
Example:
# qbf == "The quick brown fox jumps over the lazy dog"
puts [string slice $qbf 3 9 34 39 50]
Result:
3 9 34 39 50 <-- Slice (end) points
| | | | ||
The quick brown fox jumps over the lazy dog
^ ^ ^ ^ ^ <-- Start start points
1 4 10 35 40
|
(start character is assumed)
Character ranges 1-3, 4-9, 10-34, 35-39, 40-50
Output:
The quick {brown fox jumps over the} lazy dog
Example: (Slicing up redirected output from systeminfo.exe)
set filename "sysinf.txt" # Microsoft Windows systeminfo output
readfile $filename data # Read the output file
set q [string slice $data 27 1024] # Slice columns at 27 and 1024 chars
foreach {x y} $q { # Index using Tcl variables
textcolor green
printf "%-26.26s " $x
textcolor
printf "%-51.51s\n" $y
}
Result:
Host Name: Gandalf
OS Name: Microsoft Windows XP Professional
OS Version: 5.1.2600 Service Pack 3 Build 2600
OS Manufacturer: Microsoft Corporation
OS Configuration: Standalone Workstation
OS Build Type: Multiprocessor Free
...
Bad example. Uses tab characters (0x09) instead of space-padding
set x "\
684\t.\Backups
28\t.\Data
7508\t.\Documents"
# Displays with regular spacing
# 684 .\backups
# 28 .\Data
# 7508 .\Documents
set q [string slice $x 7] # Won't slice evenly
See also: string
integer [string toarray string arrayname]
Convert a CRLF-delimited string to an array. The output array
variable must not exist
Example:
set s [readfile ticol.man]
set lines [string toarray $s a]
Result:
<displays the file contents>
See also: string
trim, trimleft, trimright, string trimleft - string trimright |
Index |
string [string trim string ?charset?]
string [string trimleft string ?charset?]
string [string trimright string ?charset?]
string [trim string ?charset?]
string [trim var ?charset? -var]
Returns a string with any trailing or leading characters from the set
given by chars removed.
If 'charset' is not specified then white space is removed (spaces, tabs,
newlines, and carriage returns)
[trim] is a slightly more efficient version of the standard [string trim]
Example:
string trimleft [funct fraction [expr 4 * atan(1)]] 0.
string trimleft [funct fraction [/ 22 7.0]] 0.
Result:
141592653589792
142857142857143
Example:
puts [string trim $qbf " T[range a-w]yz"]
Result:
x
See also: trimprefix, trimsuffix, string, strsplit
integer [trimprefix varName string ?-nocase?]
Strips the string argument from a variable if found. A count of the number
of characters removed (the length of 'string') is returned
[trimsuffix] does not operate on a string literal.
Example:
set s $qbf # "The quick brown fox jumps over the lazy dog"
set r [trimprefix s "The "]
puts $r
puts '$s'
Result:
4
quick brown fox jumps over the lazy dog
See also: trim, trimsuffix
integer [trimsuffix varName string ?-nocase?]
Strips the string argument suffix from varName (if set) and if there is
a corresponding match. [trimsuffix] returns the new length of the variable
if changed, otherwise zero if no changes were made
[trimsuffix] is a command, not a function, and does not operate on a
string literal
If varName does not exist then 0 is returned
See [string trim] for a functional style trim command
Example:
set s $qbf # "The quick brown fox jumps over the lazy dog"
puts [trimsuffix s "dog"]
puts '$qbf'
Results:
40
'The quick brown fox jumps over the lazy'
See also: trim, trimprefix, string trim
list [strsplit string separators ?variable?]
[strsplit] splits a string into a list using 'separators' as a template
Optionally the result can be placed in a variable passed by argument
The separator character(s) is/are not included in the output
Example:
strsplit "hello world" "r" a # Split by char
puts $a
Result:
{hello wo} ld # List of two items
See also: string, list
string [subst ?-nobackslashes? ?-nocommands? ?-novariables? string]
Perform backslash, command, and variable substitutions
subst performs the first stage of Tcl script evaluation, performing
variable substitution and/or command substitution, but stops short of
evaluating the commands contained at the top level of the script.
Subst ignores: option escape on; option autoexec off settings and always
processes Tcl backslash escapes such as \\, \" or \$
Any optional arguments must appear before the string value
Example:
set html {<html><head>$title</head></html>}
set title "Hello, World!"
set output [subst -nocommands $html]
set output
Result:
<html><head>Hello, World!</head></html>
Example:
subst "The quick \"brown fox\" jumps over the lazy dog"
Result:
The quick "brown fox" jumps over the lazy dog
See also: {*}, expansion
swap var1 var2
Swap a pair of standard Tcl variables. Since all Tcl variables are accessed
by name rather than memory address (even when addressing via [addressof]
etc. this is effectively a rename operation which renames each variable with
the other's name with no intermediate placeholder variable
Const types cannot be swapped
Example:
set a world
set b Hello
swap a b
puts "$a $b"
Results:
Hello world
[swap] is useful for quickly exchanging sorted and unsorted arrays
Example:
option expression off
# Using a temporary and [swap] to swap between sorted and unsorted
# arrays, [swap] moves no data, it merely renames the two variables
array sort foo -array out
swap out foo # Swap $foo and $out
set s [array size foo]
for {set i 0} {[< $i $s]} {incr i} {
printf "%4i:\t%-40.40s\r\n" $i $foo($i)
}
Swapping Upvars in Procedures
-----------------------------
The use of [swap] on local upvars within a procedure will have only local
scope effect. The original parent variables remain unchanged unless
assigned directly
set a 2
set b 3
proc foo {} {
upvar a
upvar b
swap a b # Swap local upvars
puts "foo swap 1: locals a=$a, b=$b"
puts "foo swap 1: globals ::a=$::a, ::b=$::b"
swap ::a ::b # Swap parent globals
puts "foo swap 2: locals: a:$a b:$b"
puts "foo swap 2: globals ::a=$::a, ::b=$::b"
}
foo
puts "root: a:$a, b:$b"
Results:
foo swap 1: locals a=3, b=2
foo swap 1: globals ::a=2, ::b=3
foo swap 2: locals: a:3 b:2
foo swap 2: globals ::a=3, ::b=2
root: a:3, b:2
See also: vars
switch (Flow Control) |
Index |
switch -option|-- value { arg {script} ... } ...
switch -option|-- value {
arg {script}
arg {script}
...
}
switch -option|-- value arg {script} ...
Options
-------
The first argument is always an option value
Multiple options may be given
-- No option (required if no option given)
-exact Use exact matching (default)
-glob Use wildcard style matching *,?, <chars>
-nocase Case insensitive match with -glob or string matches
-eval Treat each block lvalue as an expression
[switch] has a high setup cost for each call so it is suggested that
within time-intensive loops you use [if...elseif..else] instead
Switch consists of an introductory flag setting, a value to use as
a reference, which may be the result of an expression or other command,
then a (possibly braced) series of argument + code-block pairs as a Tcl
list.
[switch] does not evaluate it's main argument value. This must be either
a literal, dereferenced variable or a return from some other command
Example:
puts [switch -- [expr "rnd(1,4)"] 1 "return 1" default "return 99"]
Cascade (fall-through) defaults are supported but at present all code
blocks must be wrapped in braces.
An enhancement to [switch] is the $this variable which will either
contain the value of any return statement or the evaluated expression
argument if return is absent.
It is very strongly suggested that you always include a 'default' case
[return] is used to return a value from a [switch] case
Example:
option expression on
for {set j 0} {$j < 10} {incr j} {
puts [switch -- [lindex [list a b c d] [rnd 0 3]] {
a {return "a"}
b {return "b"}
default {return "c or d"}
}]
}
A local variable 'this' is available which represents the variable to be
evaluated. This may be the result of an expression.
If $this is used in a single-line switch then wrap the code in braces ...
Example:
puts [switch [expr "rnd(1,4)"] 1 "return 1" default {return $this}]
This example shows cascade or "fall-through" default arguments for cases
2 and 3 ...
switch -- [expr "rnd(1,4)"] {
1 {puts "is one"}
2 {-}
3 {-}
{puts "is 2 3 or 4 (actually: is $this)"}
}
switch -- [lindex [list a b c d] [rnd 0 3]] {
a {puts "is a"}
b {puts "is b"}
default {puts "c or d"}
}
Example:
set count [lcount $qbf] # Uses global test const $qbf
-- count
switch -glob -nocase [lindex $qbf [rnd 0 $count]] {
the* {puts "is the*"}
qu* {puts "is qu*"}
br* {puts "is br*"}
fo* {puts "is fo*"}
j???s {puts "is j???s"}
?v?r {puts "is ?v?r"}
*y {puts "is *y"}
d?g {puts "is d?g"}
default {puts "is default"}
}
Example: (Performance testing)
time {switch -glob [lindex $qbf [rnd 0 10]] {default {} }} 100000 -verbose
Result:
1 microsecond(s) per iteration
See also: if, on, flow control, proc
sysfree varName
Release a string which is typically allocated by an external DLL call using
the Windows SysAlloc*() API calls (via an external DLL)
Note that [sysfree] takes a variable name not a variable contents. That
variable should hold a valid memory address allocated by SysAlloc*
This is functionally-equivalent to calling Win32 API SysFree((BSTR)address)
[sysfree] returns a boolean success. An exception is raised if the passed
address is invalid.
Where 'calldll' calls return a SysAlloc allocated string and these strings
are not freed then a memory leak will occur. This could have catastrophic
consequences. Tcl variables in such cases will be addresses pointing to
memory space
Calling [sysfree] with an address not returned by SysAlloc* is a fatal
error and will terminate the environment
Example:
# Returns the address of a string
lib ticol_calldll
set datenow [calldllstr "mslib145.dll" "VBDateStr" [now]]
42342.862013888887
puts "PHPDate: "[calldllstr "mslib145.dll" "PHPDate" $datenow "r"]
PHPDate: Fri, 04 Dec 2015 20:41:18 +0100
sysfree datenow # Pass the variable name
See also: calldll
Tcl stands for Tool Command Language. This is a popular and portable script
language which is widely used in Linux environments. The most popular version
of Tcl is also suitable for embedding within C/C++ environments.
It is a simple, but powerful language which can be used to rapidly develop
useful scripts or prototype ideas. Tcl bases everything, even procedures
as 'commands', optionally with one or more qualifying arguments.
It uses braces within the syntax and although Tcl source looks very much
like C/C++, the braces perform a somewhat different function.
Tcl requires adopting a slightly different mindset and reminding one's
self that this is not C.
Tcl is often used in embedded applications and one of the more notable
uses is within Cisco(TM) routers.
Abridged Summary
----------------
All operations are commands, including language structures and code is
written in prefix notation.
Commands accept a variable number of arguments
In standard Tcl there are no keywords so control structures may be
redefined
Data types are essentially string-based and conversion is automatic
where requred
Tcl is usually a cross-platform language
Standard Tcl supports Unicode and Object Orientation (OO)
Can be bytecode compiled
Full socket support
Ticol Tcl
---------
Ticol is a small subset of the much larger Tcl language and with a small
range of language extension which make it useful for Windows scripting and
as a replacement for Windows batch scripts.
Ticol is intended to be used only for personal, hobby or administrative
scripting and not as an equivalent or replacement for languages such as
C/C++/C# etc. Ticol should not be used in high-risk environments without
appropriate acceptance testing.
Abridged Summary
----------------
Self-contained binary with no dependencies (no .net required)
Small footprint
Was designed primarily for older systems
Unicode is not supported. The language is European ANSI only
Object orientation is not natively supported
Plugins are supported and these differ from standard Tcl's TEA
There is no bytecode compiler. Ticol is purely interpreted
May be used as a CGI back-end script handler
Minimal socket/networking support
For more information search the internet for "Tcl"
See websites:
http://antirez.com/articoli/tclmisunderstood.html
http://en.wikipedia.org/wiki/Tcl
http://wiki.tcl.tk
https://www.tcl.tk
See also: Tcl syntax, Ticol
TCL is an acronym which stands for "Tool Command Language"
The language was designed by John Ousterhout who was said to be
frustrated by programmers inventing their own language and who decided to
create a script language which would more appropriately fit the task
TCL is now well-established and with a pedigree which goes back several
decades back to 1988. The official release of TCL is extremely stable and
is quite fast since it includes a byte-code compiler. It is also quite large
and not designed to be a minimalist/portable distribution. There are
TCL-like languages such as "Jim" (and Ticol of course) which aim to fulfil
these demands
Ticol TCL is an interpreter-only implementation of a subset of TCL commands
Ticol is strictly interpreted and does not offer byte-code compilation.
For this reason, Ticol will always be much slower than the official releases
of TCL
The mixed-case acronym "Tcl" is commonly used in place of "TCL"
See also: Ticol, TCL syntax, tcl and php
Tcl has a 'C-like' visual appearance but the design has certain non-C-like
limitations. In particular, routines are either commands or procedures
rather than functions and both of these may be called by dereferencing by
wrapping in square brackets rather than suffixing curved brackets.
Tcl is based on commands with none or more arguments. All language
elements including math operators are available as commands
command arg1 arg2 ...
Example:
/ 22 7.0
puts Hello -noconfirm
Ticol (and standard Tcl) has a separate C-like expression language
evaluator which is usually evaluated using the [expr] command. This
uses infix notation but may also include commands.
Example:
puts [expr "[/ 22 7.0]*3"]
* Tcl as a language is said to have minimal or no syntax
One could perhaps consider that it had a C-like appearance
* Variables are 'typeless' and PHP-like and prefixed by a dollar-sign
* Variables are interpreted on-the-fly
* Square brackets trigger evaluation and are the equivalent of function
calls: C/C++: foo(arg) -> Tcl: [foo arg]
* Tcl knows nothing about C-like math expressions. The [expr] or [calc]
commands are used to this. In Ticol these may also be controlled by
[option expression]
* Square brackets are evaluated first. This is important within strings as
commands (function calls) are evaluated even within quoted strings
unless protected by braces
* Each level of braces prevents/delays evaluation by one level. Each
call to interpret removes one level of braces
* Unlike C/C++, Tcl keywords are commands and may have optional switch args
These perform a similar function to C# and PowerShell attributes in
passing qualifying and configuring information to the command which may
modify its behaviour or outputs
* Variables don't inherit "downwards" as in C/C++. Only globals are visible
(providing they are referenced by the '::' global prefix)
Instead you must use [upvar] or [uplevel] to access variables lower down
the call stack or pass via the proc argument list. You can access global
variables using the global-scope prefix "::" e.g. $::env
* Keywords must generally be separated by whitespace from adjacent symbols
if { [eq $a 10] } { statement...
* C/C++ like multi-part statements separated by semi-colons are supported
* Flow-control keyword braces must open on the same line as the command
if { condition } {
statement ...
* Braces group statement together and prevent evaluation
* Braces may be nested to allow evaluation only at the Nth level
* Heavy use is often made of lists. (spaced/quoted/braced strings)
* Grouping within Tcl lists is enforced by brace-wrapping
* Commands are defined by CRLF line-end structuring rather than C/C++
freeform text. Thus CRLFs may be significant in command braces
For more information, search the internet for: tcl syntax
See also: Tcl
TCP/IP and Networking Functionality |
Index |
Ticol omits the standard Tcl socket support but incorporates a small
number of useful networking routines as follows:
http Remote HTTP data-retrieval
html HTML formatting commands
info <various> Network interface information
ip_to_long IPV4 conversion routine
long_to_ip IPV4 conversion routine
Plugin libraries are available as follows:
arp Simple address resolution protocol
dns Domain Name Service lookup
ftp A simple FTP client library
ping A simple PING library
smtp A simple SMTP client library
etc.
Additionally, external _stdcall or _cdecl DLLs may be called to add TCP/IP
connectivity
See also: plugins
Syntax Checking (Static) |
Index |
Ticol performs minimal syntax checking via the Macro PreProcessor (MPP)
Checks are made only on brace and double-quote parity with some indication
being given as to where the error is located if there is a parity mismatch
Command syntax is not checked and errors may not become evident,
particularly within uncalled code, until that code is eventually called.
Thus, the language requires careful coding and error checking
Perhaps someone may wish to write a static code analyser
See also: syntax, faq
bool [seek handle offset ?whence?]
bool [file seek handle offset ?whence?]
Set the file pointer within an open file opened by [open]
The position is set relative to one of 3 options with 0 being the default
0 seek to an absolute position
1 seek relative to current position
2 seek relative to EOF
Useful defines:
const SEEK_SET 0
const SEEK_CUR 1
const SEEK_END 2
Example:
set fp [try {
open ticol.txt rb
} catch {
die
}]
if {file seek $fp 10 0} {
puts "File position for $fp is at '[file tell $fp]'"
}
close $fp
Result:
File position for 4700672 is at '10'
See also: tell, file
string [tab ?n?]
[tab] returns a string of n spaces which can be used with [puts], [printf]
or [format] to advance the cursor. [tab] mimics the TAB() command in BASIC
although it offers this functionality by returning a string of spaces
rather than advancing the cursor directly
[tab] is shorthand for [string repeat " " $n]
Example:
puts "Hello[tab 10]world"
Result:
Hello world
See also: string
integer [tell handle]
Returns the file pointer position for an open file handle
Example:
See tell
See also: file, open, close, rewind, seek
tcl2bat (Conversion Batch Script) |
Index |
tcl2bat.bat is a batch script which can convert any Ticol Tcl script into
a functioning batch file. See tcl2bat.bat in the distribution pack.
This is achieved by wrapping the original script with a batch file header
which then calls the Ticol interpreter. Converted BAT files can be loaded
and run by Ticol Tcl.
Example:
tcl2bat foo.tcl
Result:
Processing: Infile: 'C:\Tcl\foo.tcl' outfile: 'C:\Tcl\foo.bat'
* Done *
See also: batch files, md5 batch example, exec, spawn, shell
textcolor ?foreground-value|-? ?background-value|-?
Sets the text color
The color arguments may be a symbolic name such as 'blue' or a numeric
value which corresponds with standard Windows console values (e.g. blue,
cyan and these may optionally be prefixed with 'light', e.g. 'lightred')
Either argument may also be omitted by specifying a '-' character
e.g. [textcolor - 13], [textcolor 13 -]
[textcolor] has no return value
The following colour constants are defined. You may also use numeric
values:
0 0x00 black
1 0x01 darkblue
2 0x02 darkgreen
3 0x03 darkcyan
4 0x04 darkred
5 0x05 darkmagenta
6 0x06 darkyellow
7 0x07 grey (gray)
8 0x08 darkgrey (darkgray)
9 0x09 blue
10 0x0a green
11 0x0b cyan
12 0x0c red
13 0x0d magenta
14 0x0e yellow
15 0x0f white
Both 'grey' (UK) and 'gray' (US) spellings are accepted. Colour values are
case-significant. You may use [enum] to define alternate constants if you
wish
A hyphen (-) may be specified to omit a foreground value
The text colour is reset after each command-line command. Use multiple-
commands separated by semi-colons to apply
Bright background colours are supported in the Windows console, since
the top bit is used for this purpose, the blink feature is not supported
See also: cls, box, enum
$this is a local flow control variable and is available as follows ...
Control Structure Type Supported?
---------------------------------
Yes No
---------------------------------
[switch] [if]
[goto_block] [while]
[goto] [for]
[on] [do]
[foreach]
[array foreach]
[exit]
[return]
[try], [catch]
It is provided only in the 'slow' flow control mechanisms and where its
provision is useful
In [switch] control structures this represents the evaluated switch
selection expression. $this is valid in the following constructs.
[switch]
Example:
option expression on
for {set j 0} {$j < 10000} {incr j} {
puts [switch -- [expr "rnd(0,10)"] {
0 -
1 -
3 -
5 -
7 -
9 {return "$this is odd"}
2 -
4 -
6 -
8 -
10 {return "$this is very strange"}
}]
}
Example:
goto_block {
one {puts "this is $this"}
two {puts "this is $this"}
goto four
three {puts "this is $this"}
four {puts "this is $this"}
}
Results:
this is one
this is two
this is four
See also: do, switch
Windows may allocate background threads for various reasons, e.g. mutex
flag control by Windows. For this reason the number of threads in use by
Ticol may rise during execution. After a period of idle time, the thread
count may also decline
This may also increase the amount of memory used by Ticol as shown by
Windows Task Manager
integer [ticks]
Shows the tick count in milliseconds since Windows was loaded
After 9,223,372,036,854,775,807 ms or approximately every 106,751,991,167
days (roughly every 292,471,208 years) this value will wrap around to zero
Example:
while { 1 } {
textcolor [% [/ [ticks] 1000] 0xf]
puts "[info hostname]: Uptime Ticks: [/ [ticks] 1000] (sec)" -nonewline
puts "[expr {int(([ticks] / 1000.0) / (60*60*24))}]" -nonewline
puts " day(s)"
textcolor
sleep 1000
}
See also: clock, date, time, info
time [time ?-ctime? ?-noseconds? ?-separator char?] # Get system time
time [time double ?-ctime? ?-separator char?] # Time from a double
double [time hh:mm?:ss?] # Double from time
integer [time {expression} ?iterations? ?-info | -quiet?] # Script timer
The Ticol [time] command offers 3 modes of operation. It can return the
current system time, convert a double time value to an ASCIIZ time string
or it can time the execution of a Tcl script to microsecond resolution
[time double] requires a full decimal fraction with decimal point
[time] with no arguments returns the system time in HH:MM:SS format
If the -ctime argument is added then a "C" integer time value is returned
-ctime also requires that the resulting date/time be within the C epoch
If seconds digits are not required they may be omitted using -noseconds
[time hh:mm:ss] is unsupported. Use [date hh:mm:ss ?-ctime]
Example:
puts [time]
puts [time -separator [chr 32]]
puts [time -separator .]
puts [time -noseconds]
puts [time -ctime]
puts [time .583333]
puts [time [date now] -ctime] # Within C epoch
puts [date 1970/01/01 00:00:00]
puts [time [date 1970/01/01 00:00:00] -ctime] # C epoch start
puts [time [date 1998/01/30 23:30:00] -ctime] # After C epoch
puts [time [date 1968/04/08 23:50:01] -ctime] # Before C epoch
Result:
11:41:57
11 41 57
11.41.57
11:41
14:00:00
1478346117
1478346117
22569
0
886203000
Invalid C epoch time: 24936.993067
[time] with expression and iteration arguments runs the given expression
for the number of times specified by 'iterations'. The average execution
time per iteration in milliseconds is displayed. This variation uses the
Windows timer. The '-info' switch displays extra timing information.
[time] used as a profiling tool will require [option echo on] in order
to see the output text
The default number of iterations is 1. The fewer iterations, the less
accurate the calculation of elapsed time will be. A suggested minimum
number of iterations to test is 10
The resulting ms time is displayed to 4 decimal places. This value is also
the command return value. The -quiet arg can be used to chain only this
numeric result back into other code for analysis
Note that [time code] will indicate performance in microseconds
(1/1000th ms)
Example:
puts [time {;} 1000000]
Result:
1 microseconds per iteration
Example:
proc square x {expr $x*$x}
puts [time {square 42} 10000]
Result:
94 microsecond(s) per iteration
This may vary when repeated several times ...
108
111
133
108
117
Example:
time {expr 2* asin (1.0)} 100000
Result:
9 microsecond(s) per iteration
[time] as a flow-control command may be nested...
time {
time {
set r [randomstr 100]
store add $r
} 100
set s [store get]
} 100000 -info
Converting Double Date/Time Values
[time double] can be used along with [date double] to return only a time
value from a double which contains a date and time component
Example:
set d [date now]
puts $d
puts [time $d]
Result:
42663.7881250000
18:54:54
See also: date, ticks, clock, timer
integer [timer]
High-resolution millisecond timer. Returns the current tick count in
milliseconds as the time elapsed since Windows was started.
After 9,223,372,036,854,775,807 (µs) or approx. every 106,751,991,167 days
(roughly every 292,471,208 years) this value will wrap around to zero
timer uses the Win32 timeGetTime() API and resolves to that level of
accuracy
Example:
set start_time [timer]
# Do some processing
# for ...
set end_time set [timer]
puts "Run took [- $end_time $start_time] milliseconds"
See also: time, date, ticks, clock seconds
There is no Tk implementation or GUI support in Ticol other than the
[msgbox], [inputbox], [get_fileopen] and [get_filesave] commands
[box], [gotoxy] and [input] may be used to create text-mode dialogs
See also: msgbox, inputbox, get_fileopen, get_filesave
trace add command | * ?-glob?
trace remove command | *
trace list
trace log enabled|disabled
A simplified version of the Tcl [trace] command which traces only commands
The command adds or removes a runtime trace event for the given command.
'*' may be used to trace all commands. When the traced command is called
details are logged to the console as execute "Exec cmdcount (level)>" and
return values as: "Retn cmdcount (level)>"
The traced command or proc need not exist at the time the trace is
registered
Only the first command argument is logged to the console
Only commands can be traced. Variables are not traced
If a [subst] argument contains commands, these are not traced unless a
separate [trace add] registration is made
Tracing Procedures
------------------
Procedures can be traced as follows. Logging can also be added as desired
Note that the [proces] argument in [foreach] must not be wrapped in curly braces
foreach x [procs] {
puts "Adding proc $x to trace"
trace add $x
}
trace log enabled
Examples:
--------
trace add *
trace add puts
trace add if
trace remove puts
trace remove *
trace list
Example:
trace add incr
Result:
Exec 0> incr iters
Retn 0> {incr} -> {1061}
Example:
trace add factorial
assert [factorial 5] {== $_ 120} -v -line #__LINE__
Result:
Exec 1039 (0)> factorial 5
Exec 1043 (0)> factorial 4
Exec 1047 (0)> factorial 3
Exec 1051 (0)> factorial 2
Exec 1055 (0)> factorial 1
Retn 1055 (0)> {factorial} -> {1}
Retn 1051 (0)> {factorial} -> {2}
Retn 1047 (0)> {factorial} -> {6}
Retn 1043 (0)> {factorial} -> {24}
Retn 1039 (0)> {factorial} -> {120}
assert: Line 103: PASS '== $_ 120'
See also: debugging, time, timer
string [type varName | string ?-var?]
Return the type of a variable as interpreted by Ticol. If the -var
argument is used [type] will detect only the variable type and will return
"undefined" if a matching variable is not found, otherwise a string
type will be assumed and it will be analysed for the likely data type
Pointer-variables which hold addresses will be interpreted as integer type
Arrays must be passed by reference (arrayname) not $arrayname
Lists are detected only if braced and must present their enclosing braces
either by being quoted or stored in variables
e.g. [type "{a b c}"] or set a "{a b c}"; puts [type $a]
Known data types
----------------
enum const
const const
array array
stack stack
struct struct
string string
list string
Dereferenced types
------------------
integer
float
Example:
set a 10
type $a
type a
Result:
integer
integer
Example:
set a(1) hi
type $a
type a
Result:
string
array
Example:
puts [type $tcl_config] # Will check the variable's contents
puts [type tcl_config] # Will check the variable type itself
string
const
See also: sizeof, vars
number [trunc number precision]
[trunc] will truncate a real (double) number to a specified precision
without rounding. If an integer is presented then [trunc] will return the
integer will be returned without any change.
Supplying [trunc] with a non-decimal value, e.g. string or hexadecimal, will
result in a trappable error
Example:
puts [trunc 12345 4] # No effect on integers. The integer is returned
puts [trunc $pi 0] # Returns no decimal places (integer)
puts [trunc $pi 2] # Returns 2 decimal places
puts [trunc hi 1] # Raises an exception
Results
12345
3
3.14
<exception>
See also: round, math, double
try ... catch (Flow Control) |
Index |
result [try {script1} catch {script2} ?-quiet?]
Evaluate script1. If an error is returned by script1 then evaluate script2
and in either case, return the result for further processing
[try..catch] handles runtime errors not syntax errors (incorrect arg
counts)
Unlike [catch] fatal error messages are suppressed for a [try...catch]
command because [try] will have its own error-handler in [catch]
The catch-handler must handle any error message output and take any
necessary corrective action such as closing file-handles or deleting
resources
The '-quiet' argument will suppress local error messages and the caller
must rely on using other methods such as $errorMsg to interpret failures
Be careful in that, due to the way the Tcl language operates, a return
value from a command which generates an exception may not be set since
the returning [set] command is not executed. Thus, where [some_command]
raises an exception ...
[try] does not guarantee all variables [catch] may need to address will be
set. Any variables referenced by the relevant [catch] clause must be
set before [try] is called even if set to null
Example:
try {
set r [some_command] # r will not be set on exception
} catch {
puts "r was not set!"
}
You should set any variable first
set r -9 # Set r to a suitable default value
try {
set r [some_command] # r will not be set on exception
} catch {
puts "r was set to $r"
}
The recommended method of setting a return value from a command which
may fail a [try...catch] exception is to set the return value of [try]
Example:
# Success case
set p [try {
/ 22 7.0
} catch {
0
}]
assert $p {== $_ 3.142857142857143} -v -line #__LINE__ # PASS
# Failure case
set q [try {
/ 22 0
} catch {
return -1 # Some arbitrary value which we want set
}]
assert $q {== $_ -1} -v -line #__LINE__ # PASS
[try] differs from standard Tcl in having a [catch] handler instead
of [finally]
The variable ::errorMsg may be set by the trapped command
Differences between [try] and 'catch' clauses in [try]:
try Does not display error messages
catch Displays error messages dependent on option echo off
Example:
try {open fred.txt r} catch {puts "Can't open file"; exit}
Result:
Can't open file
Example:
try {
open nosuchfile r
} catch {
puts "Caught: '$::errorMsg'"
}
Result:
Caught: 'open: Failed to open file 'nosuchfile' for mode r access'
See also: catch, errorMsg
Select from multiple choices:
See either: undef macro Used by the Macro PreProcessor
undef command Used to undefine a proc
See also: proc, macro
undef command (Undefining Commands) |
Index |
undef proc|command ?proc|command...? ?-nocomplain?
undef -glob filter ?-proc? ?-nocomplain?
Undefine a Tcl commands or procs. This is a non-standard command which
would normally be implemented using the rather clumsy [rename command ""]
You cannot [undef] the active proc from within that proc or any parent
proc. Any attempt to undef any active proc in a call chain will raise an
exception. The -nocomplain argument can be used to ignore errors and
continue however, the current proc cannot delete itself.
The first example of [undef] undefined specific commands by name and has
no return.
The second example of [undef] used -glob and returns a Tcl list of
commands or procedures undefined. This can be restricted to undefining
procedures only by adding the -proc argument.
You may undefine any command listed by the [commands] command. This allows
build-in Tcl commands to be re-defined by custom ones of your own choice
You can redefine any aspect of the Tcl language and replace with your own
constructs since all constructs even [for], [while], [return] etc. are
defined as commands and subject to undefinition
All procedures defined with the [proc] command are also present in the
command list and may similarly be undefined and tested using [defined]
If a proc or command does not exist and -nocomplain is omitted then the
script will halt with an error
Aliases
-------
If an attempt is made to undefine a [proc] using [undef] and an alias
still exists then the proc will remain active until the last alias is
undefined. When the last alias to a [proc] is removed the procedure
will be destroyed
[undef] is not to be confused with the #undef macro. See: undef macro
Example:
proc foo {} {puts Hello}
if {[defined bar]} {
undef bar
alias bar foo
}
bar
Result:
Hello
Example:
puts [undef -glob * -proc]
Result:
bar foo moo {my proc}
See also: is_active, rename, commands, proc, procs, undef macro
proc (Tcl Procedures) |
Index |
proc procname { arg ?{arg default-value}? ?args?}... {code}
Define a Tcl procedure. The first code bracket MUST be on the same
line as the proc declaration. A [proc] need not have an explicit [return]
as it will return the value of the last operation or expression executed
Procedures are created in global (static) scope and all procs (even if
definitions are nested within other procs) remain visible in global scope
until they are removed using [undef]
There is a small performance cost to [proc] definition due to the fixup
time. You should avoid using [proc] definitions within loops which need
high performance.
A defined [proc] is called when its name is encountered in a script,
either outside of a string or enclosed in square brackets.
e.g. foo, [foo] or using call procname.
Nested [proc] definitions are supported
proc foo {x} {expr $x-1}
^ ^ ^
| | |
proc name | body
|
list of argument names. wrapped in braces
It is recommended that the argument values are wrapped in curly braces
Multiple arguments are all enclosed in a single set of outer braces
A proc may be defined entirely on a single line, but is more usually defined
over multiple lines:
Example:
proc name
| argument
| | body
| | |
v v v
proc cube {x} {
set r [* $x $x]
* $r $x
}
Optional+default arguments may be specified by braced pairs as follows:
Argument 1
| Argument 2
| | Default value 2
| | | Argument 3
| | | | Default value 3
| | | | |
v v v v v
proc foo {x {y 23} {z skidoo}} {
puts "proc foo: x=$x y=$y z=$z"
}
Here, proc [foo] is described with mandatory argument 'x' and optional
arguments 'y' and 'z'. Each of these two optional arguments has a default
value set. Those default values being, '23' and 'skidoo' respectively.
NULL default arguments are specified by empty double quotes or braces.
e.g.:
Argument 1
| Argument 2
| | Default value 2
| | | Argument 3
| | | | Default value 3
| | | | |
v v v v v
proc http_login {node {user ""} {pass ""} } { ...
proc http_login {node {user {}} {pass {}} } { ...
Procedure arguments must be syntactically valid if provided as unquoted
strings. If unquoted strings must be passed to a proc, then ensure they
are wrapped in double quotes to prevent arguments being interpreted as
commands:
Example:
# 1st arg--+ +---------- 2nd arg
# | |
proc foo {arg *} {... # Wrong (assumed to be command operator [*])
proc foo {arg "*"} {... # Right, but useless (string const, not var)
proc foo {{arg *} {... # Possibly what was intended (default for arg)
# | |
# +--+--------- Argument 'arg' with default value of '*'
Nested procs are permitted and the [upvar] command will reference both
string and array variables in the parent or global context and set the
parent variable accordingly. Note that all procs are declared and available
in global scope (level 0)
Note that the syntax of a proc is not checked until the proc is first
called, thus procs should be tested independently for errors and debugged
using a separate test file which ensures they are called
Procedure Names
---------------
Procedure names should be alpha or alphanumeric and not be numeric values,
this is to avoid a conflict with the Macro Preprocessor expression optimiser
Example:
proc 123 {args} {... # Wrong. Can't be called other than by [call]
proc 3.0 {args} {... # Wrong. Can't be called other than by [call]
proc a12 {args} {... # Correct. Alpha name prefix
proc abc {args} {... # Correct. Alpha name prefix
Procedure Definition
--------------------
Ticol procs are stored as source code snippets which are not translated in
terms of final '\' escapes until evaluated. A proc may be inspected
using [info args procname] and [info body procname] which shows the proc
code unescaped. If you call another command with the result of [info body]
then backslash escapes will be translated before use. An active proc's own
name can be retrieved using [info procname]
It is much more efficient to define all procs in advance rather than
defining during the remainder of script execution, within loops etc.
A recursion depth limit of 5000 is imposed on [proc] recursion which is
configurable via ticol.ini.
Example Test:
# Should be: first 23 skidoo
foo first
# Should be: second 2 skidoo
foo second 2
# Should be: third 3 three
foo third 3 three
Result:
proc foo: x=first y=23 z=skidoo
proc foo: x=second y=2 z=skidoo
proc foo: x=third y=3 z=three
A valid procedure name may also be one or more spaces...
Example:
proc { } {} {puts "Foo!"} # Define procname " "
[{ }] # Call using braces
[" "] # Call using dquotes
Results:
Foo!
Foo!
Example Nested Proc: (where each new [proc] calls a child proc)
proc level_p {} {
puts "in level_p ([info level])"
proc level_q {} {
puts "in level_q ([info level])"
proc level_r {} {
puts "In level_r ([info level])"
}
level_r
}
level_q
}
Standard variables can be 'promoted' or 'cast' to array type within a
procedure using a global reference:
Example:
set q {} # Declare a non-array global
proc array_promote {} {
puts "Enter array promote. Setting global var ::q as array"
set ::q(1) "Hello from array_promote" # Set global q as array element
set ::q(2) "Two"
puts "::q(1) is '$::q(1)'"
}
array_promote
puts "Test var 'q' in global scope: \$q(1) is '$q(1)'"
Result:
Test in global scope: $q(1) is 'Hello from array_promote'
Variable Length Arguments using 'args'
--------------------------------------
The standard Tcl 'args' variable argument is optionally defined as the last
argument of any proc.
This is a special, trailing argument variable which will receive either
none or all of any excess arguments passed to a proc when it is called.
If args is not specified as the final argument of a proc then any excess
arguments supplied when the proc is called will be ignored
Example:
proc demo {first {second "none"} args} {
textcolor darkyellow
puts "Enter \[demo\]"
puts "first = $first"
puts "second = $second"
puts "args = $args"
textcolor
newline
}
demo one
demo one two
demo one two three "four five"
Results
first = one
second = "none"
args = ""
first = one
second = two
args = ""
first = one
second = two
args = three {four five}
See: pass by name, pass by value
Calling Procs from [eval]
-------------------------
Any proc may be called from [eval] regardless of argument count
without specifying arguments. In such case, null or empty arguments
will be supplied by default to the proc.
Example:
proc foo {x y} {
puts "x is '$x' x is '$y'"
}
eval foo
Result:
x is '' x is ''
See also, proc args, args, info args, info body, info procname, unknown,
varname, static, undef, defined, foreach, for, while, if
proc unknown {arguments} {body}
If the [autoexec] option is set to OFF using [option autoexec off] and a
proc called 'unknown' is defined then this proc will be called when a
given command is unrecognised.
Enabling [autoexec off] will prevent the CLI from acting as a command shell
The autoexec option is Ticol specific and not standard Tcl
A useful example is one which uses unknown to redefine and extend the Tcl
language...
Example: (multiple arguments must be quoted or braced)
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
user@host> proc unknown args {return [expr $args]}
user@host> autoexec off # Required
user@host> option echo on # Required
user@host> "1 + 2"
3
user@host> pow(2,20)
1048576.000000000000000
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Example: (Save the following as unknown.tcl and run it)
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
proc unknown {x y} {
puts "In process unknown: $x $y"
}
autoexec off
puts "Running unknown.tcl"
foople woople
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
ticol unknown.tcl
Result:
Running unknown.tcl
In process unknown: woople
Example: (One line calculator)
option autoexec off; option echo on; undef unknown;proc unknown {args} {\
textcolor magenta;\
puts "Evaluating: '$args'\r\n";\
textcolor;expr "$args"}
See:
http://wiki.tcl.tk/795
See also: autoexec, proc
integer [unlink filespec ?-verbose? ?-nocomplain?]
Delete a single or multiple files without prompting *.* is banned
anywhere within a filespec
If -verbose is used then messages are shown
If -nocomplain is used then deletion of multiple files will continue
on any errors. If -nocomplain is omitted, an error will be
returned on failing to delete a file. Unless you want your script to halt
if a file is not found then -nocomplain is recommended.
A count of the number of successfully-deleted files is returned
Files which are open or locked by another process cannot be deleted
Example:
set x [unlink *.tmp]
puts "Deleted $x file(s)"
Result:
Deleted 23 file(s)
See also: file, delete, rename
bool [unset var ?var...? ?-nocomplain? ?-const?]
bool [unset mask ?mask ...? -glob ?-nocomplain? ?-nocase? ?-const?]
Undefine (delete) one or more variables. The -glob argument may be used
with one or more glob search 'masks'. -const is required to delete a
constant. You may also specify -nocase when using -glob.
Use [array unset] to delete entire arrays
Beware of calling [unset -glob] where it may encounter a const variable
and you are not supplying the -const argument. In such cases you must
also add -nocomplain so that an error is not raised when a const is
encountered
Const variables are not intended to be assigned-to but deletion of a const
and subsequent reuse is possible using -const and also with a glob wildcard
using both -const and -glob arguments together
Example:
unset argc -const -nocomplain
unset argv -const -nocomplain
The variable argument(s) may be a string, list, struct, or array element
Example:
set s Hello
unset s
array set a {
1 one
2 two
}
unset a(1)
Where a struct is deleted this will also delete all member variables.
Unsetting an array will delete all elements of the array
Where an array is unset this will destroy the entire array
Where an array element is unset this will remove only that element
Removing all of the elements of an array does not remove the array itself,
you should call [unset arrayname] or [array unset arrayname] to remove
Global variables (including arrays) should be referenced with a "::" prefix
e.g. unset ::globalvar
e.g. unset ::arrayname(1)
The -nocomplain argument will suppress warnings where a variable does not
exist. This may be useful where the script is called (using [eval]) from
another script which may or may not have a variable set
Example:
# Start of script (chained from another script which may have q set)
unset q -nocomplain
set q Hello
puts $q
A range of single-letter variables can be quickly cleared as follows:
unset {*}[split [range a-z]] -nocomplain
Which expands to
unset a b c d e f g h i j k l m n o p q r s t u v w x y z -nocomplain
Example:
unset tcl_version -const # Delete a const
Glob searching
--------------
Glob deletion of variables may be slow and inefficient if you have a lot of
variables declared. For glob searching the -glob argument is required.
An error will be raised if no matching variables are found unless the
-nocomplain argument is specified
The following wildcards are valid
* Any group of characters or all matches
? Any single characters
Example:
unset * -glob # Unset all non-const variables
unset * -glob -nocomplain # Unset all non-const vars without complain
unset a* -glob # Unset all vars starting with 'a'
unset ?h?? -glob # Unset that, this, them etc. (4 letters)
unset a* b* c* -glob # Unset multiple vars starting with a, b, c
unset * -glob -const # Unset all constants
See also: clear
string [unwrap list]
Remove leading+trailing whitespace and then remove x1 level of braces
from a list.
[unwrap] can be replaced with a proc which offers more comprehensive
list-unwrapping features (see below)
[unwrap] as a procedure
-----------------------
Below is a simple proc which unwraps a nested list which is defined with
XML-style nesting. It is implemented to display the contents of each root
level but can be adapted to fill database entries, arrays etc.
Script
------
proc [ldepth] must be defined: See: ldepth
undef unwrap # Undefine the built-in [unwrap]
option expression off # If statement uses [eval]
proc unwrap {l} {
set s {}
foreach x $l {
if {> [ldepth $l] 2} {
unwrap $x
} else {
append s "$x\t"
}
}
if {<= [ldepth $l] 2} {
set s [string trim $s "\t "]
puts $s
}
}
undef unwrap # Undefine the built-in [unwrap]
option expression on # If statement uses [expr]
proc unwrap {l} {
set s {}
foreach x $l {
if {[ldepth $l] > 2} {
unwrap $x
} else {
append s "$x\t"
}
}
if {[ldepth $l] <= 2} {
set s [string trim $s "\t "]
puts $s
}
}
Example:
set l "{catalog {cd {{title {Empire Burlesque}} {artist {Bob Dylan}}\
{Company Columbia} {Price 10.90} {year 1985} }}}"
unwrap $l
Results:
catalog
cd
title Empire Burlesque
artist Bob Dylan
Company Columbia
Price 10.90
year 1985
See also: list processing, list, ldepth, append
uplevel command
uplevel ?level? command ?command...?
uplevel ?@level? command ?command...?
Run a command in a higher call-frame level. By default 'level' is an
integer representing the number of relative levels above the current one
Each nested [proc] call increases the level by 1. [uplevel] has no return
[uplevel] joins each of its arguments together with spaces after trimming
leading and trailing white-space from each
If more than one command is given then a level value must be given (even if
0 for the current level)
NOTE Ticol uses @ instead of # as a level prefix to reserve the use of #
for embedded comment preprocessing
If level value is omitted then 1 will be assumed and the level immediate to
the current one will be selected
If 'level' is prefixed by an @ sign then the reference will be an absolute
level rather than a relative one. Standard Tcl uses the # symbol. The most
common use for '@' will be to refer to the root (global) level @0
Nested Quotes and Uplevel
-------------------------
Occasionally, a need can arise to wrap an uplevel argument in double-quotes
which is within a quoted string already. Nested quotes are not allowed.
However, you may use [chr] to insert quote characters as follows:
In the example below, it is vital to wrap the call to [is_array] in
quotes as the variable 'q' may contain spaces which could lead to an
erroneous return. Here we use [chr 34]. Note that [chr] here may not
be wrapped in braces or this will block the command call.
if {[uplevel @0 "is_array [chr 34]$q[chr 34]"]} {
...
Alternately, a quote character may be defined as a var or const
const QUOTE [chr 34]
if {[uplevel @0 "is_array $::QUOTE$q($i)$::QUOTE"]} {
...
Local Variables and Uplevel
---------------------------
Passing a combined proc-local variable to execute in an uplevel should be
done by wrapping the whole argument in double-quotes rather than braces
If braces are used then the variable cannot be resolved locally before
evaluating at the upper level. Thus
proc foo string {
uplevel @0 {eval $string} # WRONG Can't resolve braced local var
...
proc foo string {
uplevel @0 "eval $string" # CORRECT Can resolve local before call
...
Uplevel Example:
proc clearlocals {level} {
set varlist [uplevel $level {info locals}]
foreach {v} $varlist {
set command "is_const $v"
if {[! [uplevel $level $command]]} {
set command "unset $v -nocomplain"
if {[uplevel $level $command]} {
textcolor green
puts "* unset $v successfully"
textcolor
}
}
}
}
set a Hello
set b world
set c $qbf
set d "Dee"
clearlocals @0 # Absolute level of 0 (The global root level 0)
#clearlocals @1 # Absolute level of 1 (The current proc level 1)
#clearlocals 0 # Relative level of 0 (The current proc level 1)
#clearlocals 1 # Relative level of 1 (Parent of the current proc)
newline
puts "Locals: '[info locals]'"
See also: upvar, set, at_exit, global
upvar ??@?level? uplevelvar ?localvar? ?-nocomplain?
Create a link to a variable in a parent scope and optionally refer to a
named instance in local scope. Ticol uses the '@' rather than '#'
character to specify an optional target level.
The name of the referred-to variable and local alias variable may be
identical.[upvar] has no return. [upvar] is also required when passing a
proc variable by name rather than by value.
The variable named by 'uplevelvar' need not exist at the time of the call,
it will be created as a plain variable in the specified parent scope.
+------> [parent var 'x']
| |
^ v
| proc(x) {
|
+------<--- [local upvar 'x']
}
To refer to a variable called 'a' in global scope from within a [proc] as
'a', use either:
upvar a a
upvar a
[upvar] will search parent contexts until the first match to the given
variable name is found. If a match is found a new variable will be created
with the given new name but with the contents of the located variable
If a local variable name is not given then an alias with the same name as
the parent variable will be created automatically
The level value is optional and if omitted then 1 will be assumed. If level
is prefixed by an @ sign then the reference will be an absolute level
rather than a relative one. Tcl uses the # symbol. Ticol uses @ to release
# for embedded comments via preprocessing
Since the [at_exit] command can take no parameter arguments, upvar may be
required for [at_exit] to access variables in other scopes
Example:
set a(1) Hello
set q a(1)
foo q # Call the proc by name with q as an alias for a(1)
It is an error (and unnecessary) to attempt to call [upvar] for a variable
with a locally-declared '::' global scope prefix. e.g. [upvar ::true]
Example:
proc myproc {} {
upvar a b; # Reference a as local variable b
puts "upvar of a is b->$b"
}
set a 123
myproc
Result:
upvar of a is b->123
Example:
proc myproc {} {
upvar a # Ticol supports lazy shorthand for [upvar a a]
puts "upvar of a is $a"
}
set a 123
myproc
Result:
upvar of a is 123
Example:
set a {}
proc myproc {} {
upvar a # Ticol supports lazy shorthand for [upvar a a]
set a(1) "One"
puts "local a is $a(1)"
}
myproc
puts "global a is $a(1)"
Result:
global a is One
Passing Variables by Name
-------------------------
# A bad example!
# Error: upvar: 'x' already exists at level 1
proc foo {x} { # This uses the equivalent of [upvar x x]
upvar $x # Won't work because new local var 'x' conflicts
set $x 23 # Assign
} # already exists as proc variable via proc
set x 0 # body declaration
foo x # Call proc [foo]
puts $x # Wont' print out 23
# Avoiding name clashes...
proc foo {_x} { # Note that we must dereference the arg var
upvar $_x x # _x avoids conflict with variable name x
set x 23 # Assign to local (will update parent)
} # as new variable 'x' does not already exist
set x 0 # via proc body declaration. We could also
foo x # name the var 'foo' or 'q' etc.
puts $x # Will print out 23
# Variable update cascade back to parent from nested proc calls
proc a {x} {
upvar $x
puts "Enter proc \[a\] x is $x -> $$x"
++ $x
proc b {x} {
upvar $x
puts "Enter proc \[b\] with arg x as $x -> $$x"
set $x 23
}
b $x
}
set q 0
textcolor cyan
puts "Starting level 0 'q' is '$q'"
textcolor
a q # Call the nested proc sequence with variable 'q' "by name"
textcolor cyan
puts "Resultant level 0 'q' is now $q"
textcolor
To pass array elements or arrays see: passing arrays
To pass structs (or a pseudo-class) to a proc or sub, see the help topic
'upvar with structs'
See: static, uplevel, global, vars, set, at_exit, upvar and structs,
upvar and arrays, passing variables by name
Passing Arrays to Procedures Using upvar |
Index |
[upvar] can create a reference for an array or array element and pass
by name.into a [proc] or [sub] while subsequently processing either the
whole array or the element within the [proc] or [sub]
Example:
Passing an array by name using complex dereferencing
proc foo {x} {
upvar $q
puts "Type is: [type $q]"
puts "${${q}(0)}" # Complex dereferencing
}
set a(0) Hello
foo a
Result:
Type is array
Hello
Example:
Passing an array by name simpler dereferencing and a local alias
proc foo {q} {
upvar $q x
puts "Type is: [type x]"
puts "$x(0)" # Simple dereferencing using an alias
}
set a(0) Hello
foo a
Result:
Type is array
Hello
Example:
Passing an array element by name simpler dereferencing and a local alias
This would also allow loop iteration
proc foo {q} {
upvar [varname $q] x # Create a local alias to the array
puts [array size x] # Show the array size
puts "$x(0)" # Simple dereferencing using an alias
}
set a(0) Hello # Set an array element in root
foo a(0) # Pass an array element to [foo]
Result:
1
Type is array
Hello
See also: arrays, upvar, upvar and structs
Using [upvar] with Structs |
Index |
Ticol structs emulate "C" style structs but are not identical in usage
A Ticol struct is a single binary memory block, onto which map one or more
standard Tcl variables. Due to this, the parent struct variable is
disjoint and separate from the field (member) variables although they will
all reside within the same scope.
When [upvar] is used to gain access to a struct it maps onto the struct
parent variable, not the separate field member variables. Most commands
will resolve the mapping between the referred struct and its members
The parent struct variable has a relationship with named members. These
names are fixed and can't be aliased via [upvar]
It may be necessary to separately [upvar] struct member variables
Additionally, the reference [upvar] variable argument passed via [proc]
has to be resolved using a complex/braced dereference. Because of this
additional indirection, upvar-ed structs are more complex to address
Example:
-------
struct s {a 10}
proc foo {x} {
upvar $x # Dereferencing, x gives the struct name
puts "-> ${x}.a" # This requires x to be braced
puts "[set ${x}.a]" # This requires x to be braced
struct set ${x}.a Bye # This requires x to be braced
}
struct set s.a Hello
foo s # Pass the struct via foo
puts "s.a:'$s.a'" # Will display the changed value
Results:
-------
Hello
Hello
Bye
Example:
-------
This demonstrates the need to separately [upvar] struct member field
variables as well as the need to use [struct set] to assign to struct
member fields. It also reminds that the [++] like operators work only
on simple/scalar variables
Double-dereferencing is required to address struct member variables
as per the example below
proc process_rgb {st} {
upvar $st # This won't enable access to struct members
upvar ${st}.r # x.r (red)
upvar ${st}.g # x.g (green)
upvar ${st}.b # x.b (blue)
# Must use [struct set] to assign. Can't use [++] with complex vars
struct set ${st}.r [+ ${${st}.r} 1]
struct set ${st}.g [+ ${${st}.g} 1]
struct set ${st}.b [+ ${${st}.b} 1]
}
struct create template { # Define a template instance
r 5
g 5
b 5
}
new template x # Instantiate x (members are initialised to 0)
new template y # Instantiate y
process_rgb x # Call with x
process_rgb y # Call with y
Example
-------
From v1.27 release the following is recommended but the more complex version
(above) will still work
proc process_rgb {st} {
upvar $st q # This won't enable access to struct members
upvar q.r # x.r (red)
upvar q.g # x.g (green)
upvar q.b # x.b (blue)
# Must use [struct set] to assign. Can't use [++] with complex vars
struct set q.r [+ $q.r 1]
struct set q.g [+ $q.g 1]
struct set q.b [+ $q.b 1]
}
struct create template { # Define a template instance
r 5
g 5
b 5
}
new template x # Instantiate x (members are initialised to 0)
new template y # Instantiate y
process_rgb x # Call with x
process_rgb y # Call with y
See also: upvar, static, struct, struct set, proc
The following commands have been added to make Ticol development easier
array walk # Walk an array's hashtable
at_exit # Call useful debug, trace or cleanup routines on script exit
commands # Returns a filterable list of defined commands
debug # Enable console debugging for debug versions of Ticol
dump # Dump a variable, giving useful debug information
find # Find help keywords
functions # Returns a filterable list of expr functions
help # This help file
mem_used # Returns the current memory used by Ticol
option # Shows the state of optional configuration settings
spawn # Spawn an external program and capture output to a variable
time # Time the execution of a block of code
trace # Enable debug tracing
walk # Walks the variable table showing addresses, type and content
vars # Returns a filterable list of variables by level
#__LINE__ # Macro to aid tracing line numbers in the original source
See also: debugging
string [variable var ?value?]
In standard Tcl this creates and initialises a namespace variable.
Ticol does not implement namespaces, it emulates them
Declared as an alias for [set]. Declared for Tcl compatibility
Variables declared in [proc] scope using set are local to that proc and
in the proc namespace
If value is absent then the contents of 'var' are returned instead
Example:
variable s Hello
puts $s
Result:
Hello
See also: variables, set, vars, upvar, static, dim, array
Tcl automatically detects variable types. There is no need for type-
declaration when declaring variables. Variable declaration is similar to
PHP with variables starting with a dollar sign and having any alphanumeric
name
Standard Tcl allows any character to be used in a variable name. Ticol
Tcl uses a more restrictive character set. Valid variable name characters
include: A-Z a-z 0-9 _ . : () & <tab> and <space>
"abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_.:& \t"
Curved brackets define array variables thus varname(subscript)
Spaces are significant within subscripts so subscripts which contain
whitespace should be quoted or braced
Ticol also allows an underscore "_" and the dot character "." in names
and will automatically create parent->field reference names for struct
member variables which include a dot character (See: structs)
A variable name may also contain spaces. Even comprised entirely of
spaces. Unlike standard Tcl, a Ticol variable cannot be an empty string
A variable name may be any length
Examples:
$a_ticol_variable
$also.a.ticol.variable
set [string repeat x 10240] Hello
set $qbf Hello
puts $$qbf
Variable boundaries can be protected from adjacent recognised characters by
curly-braces around the name.
Examples:
set a "ll"; puts "He${a}o => Hello"
puts ${this_is_a_ticol_variable_reference}
Scoping and Namespaces
Namespaces are emulated by extended variable names only. The global
namespace is implemented and is addressed by prefixing a variable name
with double colon characters "::"
Examples:
::argv(1)
::env(path)
Although essentially typeless, the following variable types are
interpreted:
integer A 64-bit integer value stored as a string
double A floating-point value stored as a string
hex A hexadecimal literal preceded by 0x stored as a string
octal An octal value preceded by 0o stored as a string
binary A binary value preceded by 0b stored as a string
array Standard Tcl hashtable arrays (one true dimension only)
struct Emulation of C-like binary structs for use with calldll_*
Additionally, bit/byte field-widths may be set when using setb or ofaddressb
in order to exchange data using structs via the Windows API or other DLLs
See also: namespace, type, arrays, struct, vars. calldll, setb, ofaddressb
Variant Handling Plugin |
Index |
To load, use: lib ticol_variant ?-verbose? ?-nocomplain?
It is not possible to directly declare a variant. However, to create a
variant in order to interface with an external DLL via [calldll] a
[struct] could be created or a block of memory allocated of the correct
size could be passed.
The following commands are exported (and are documented separately)
[varray <subcommands>] # Main command
[vsplit] # Utility command
[dumpvariant] # Introspection command
See: varray, vsplit, dumpvariant, struct, malloc
Variant Array Functions (Plugin) |
Index |
Contained in lib plugin: ticol_variant,dll
To load, use: lib ticol_variant ?-verbose? ?-nocomplain?
A rudimentary mechanism to manipulate Windows VARIANT objects
[varray] is provided to help interface with external DLLs only
For internal use, use standard arrays or the ticol_carray plugin
handle [varray create size ?lbound?] Create a new 1D variant array
Lower bound can be set
bool [varray set handle element value] Assign an item to array element
integer [varray size handle ?dimension?] Element size of a variant array
string [varray get handle index ?index?] Get element from 1D or 2D array
integer [varray lbound handle] Return lower bounds of an array
integer [varray ubound handle] Return upper bounds of an array
list [varray list handle] Return variant array as a list
varray foreach handle var {script} Iterate elements (in order)
using variable 'var' and {script}
integer [varray unset handle] Free the associated variable
The caller is responsible for
ensuring the variable is
associated with a variant array
Variant arrays are returned from external DLL functions called by the
calldll* interface. Such arrays are not garbage-collected automatically
and should be deleted using [varray unset] (if safe) and when no longer
required, in order avoid memory loss
"Arrays start at 1"
Variant arrays can have any arbitrary lower bound although the default in
this library is 0. [varray set] sets at the actual bounded index rather than
starting at 0 so if the lower bound is 2 then [varray set] element 2 will
set the first element and so on
One and two dimensional arrays are supported to various degrees
Variant array cleanup is normally performed by [varray unset] but arrays
which have not been unset will be cleaned up and deleted when the lib
is unloaded
Example:
option expression off
set v [vsplit $qbf] # Split to Variant (SAFEARRAY) array
set elems [varray size v]
puts "Array has $elems elements:\n"
for {set i 0} { < $i $elems } {incr i} {
puts "$i: [varray get v $i]"
}
varray unset v # Must be deleted!
Result
Array has 9 elements:
0: The
1: quick
2: brown
3: fox
4: jumps
5: over
6: the
7: lazy
8: dog
Example:
option expression off
set v [vsplit $qbf] # Split to Variant (SAFEARRAY) array
set i 0
varray foreach $v x {
puts "$i: $x"
++ i
}
varray unset v
Result
0: The
1: quick
2: brown
3: fox
4: jumps
5: over
6: the
7: lazy
8: dog
Create Example:
set handle [varray create 5 2] # Lower bound is set to 2
dumpvariant $handle
set elems [varray size $handle]
varray set $q $lbound Hello
varray set $q [+ $lbound 1] World
puts "Array has $elems elements:\n"
for {set i $lbound} { < $i [+ $lbound $elems] } {incr i} {
puts "$i: [varray get $q $i]"
}
varray unset $handle
DUMP: Variant 'q' at 36015216 (0x2258c70)
Type is 8204 (0x200c) BYVAL
Variant is an array of type 0xc
Type is 12 byte Variant VT_VARIANT
Array has 5 elements:
2: Hello
3: World
4:
5:
6:
See each varray command for more details
See also: vsplit, calldll, dumpvariant
string [varname variable]
[varname] extracts the variable name prefix from an array variable where
that name is not known in advance but required, e.g. when passed via a
proc argument and we need to handle the root array name
Examples:
puts [varname env(path)]
proc foo x {
puts [varname $x] # x resolves to 'a(0)'
puts [array size [varname $x]]
}
set a(0) Hello
foo a(0)
Results:
env
a
1
See also: proc, upvar, array
varray size (Plugin) |
Index |
Contained in lib plugin: ticol_variant.dll
To load, use: lib ticol_variant ?-verbose? ?-nocomplain?
integer [varray size variable|address ?dimension?]
Returns the array count for a 1D variant string array
The dimension argument is a base 1 value (1st is 1, 2nd is 2)
The default dimension is 1, which represents the first array bound
A 2d array would have a count for dimensions 1 and 2 where 1 would be the
rows and 2 the columns. vvarray lbound and vvarray ubound can be used to
calculate the varray size value
This command currently supports a maximum of 2 dimensions
Example:
# Split to Variant (SAFEARRAY) array
set v [vsplit $qbf]
puts "Varray has [varray size v] element(s)"
Result:
Varray has 9 elements:
See also: arrays, varray get, varray lbound, varray ubound, calldll
varray get (Plugin) |
Index |
To load, use: lib ticol_variant ?-verbose? ?-nocomplain?
string [varray get address row ?column?]
Return a string element from a 1D or 2D Variant string array. Such arrays
may be returned from DLLs such as SQLite queries.
Example:
# .. code to set variant 'vr'
option expression off
set lbound [varray lbound $vr]
set ubound [varray ubound vr]
for {set i $lbound} { < $i $ubound} {incr i} {
puts "\t$i: [varray get vr $i]"
}
See also: arrays, varray size, varray lbound, varray ubound, calldll,
dumpvariant
varray lbound (Plugin) |
Index |
Contained in lib plugin: ticol_variant.dll
integer [varray lbound address|variableName ?dimension?]
Return the lower bounds for a 1 or 2D variant array.
The default dimension is 1
For 2d arrays specify 2
Example:
# Split to Variant (SAFEARRAY) array
set v [vsplit $qbf]
puts "Lower bound is [varray lbound v]"
puts "Upper bound is [varray ubound v]"
puts "Varray has [varray size v] element(s)"
Result:
Lower bound is 0
Upper bound is 8
Varray has 9 elements:
See also: arrays, varray ubound, varray get, varray size, calldll
varray ubound (Plugin) |
Index |
Contained in lib plugin: ticol_variant.dll
To load, use: lib ticol_variant ?-verbose? ?-nocomplain?
integer [varray ubound address|variableName ?dimension?]
Return the upper bounds for a 1 or 2D variant array.
The default dimension is 1
For 2d arrays, specify 2 as the dimension value
Example:
# Split to Variant (SAFEARRAY) array
set v [vsplit $qbf]
puts "Lower bound is [varray lbound v]"
puts "Upper bound is [varray ubound v]"
puts "Varray has [varray size v] element(s)"
Result:
Lower bound is 0
Upper bound is 8
Varray has 9 elements:
See also: arrays, varray lbound, varray get, varray size, calldll
vars ?filter? ?-const | -auto | -all? ?-level level? ?-list? ?-quiet?
Non-interactive tabular display of registered variables
These are colour-coded according to auto-detected type
See: 'info vars' to return a Tcl list of variables by query
Options:
filter A glob filter pattern
-const Show defined constants
-auto Show normal (automatic) variables
-all Show all variables
-level N Show vars at given level N
-list Return a Tcl list of all varnames
(-auto, -const , all and -level may be used)
-q Quiet mode (no text output)
Tcl variables are traditionally strings but Ticol extends to binary structs
Filter is a partial prefix-string-match on the variable name
The '-level level' argument is valid and useful only from within a procedure
Example:
vars a -const
Result:
Defined variables
-----------------
string argv0 ticol
integer argc 1
Example:
puts [vars -const -q]
Result:
18
See also: walk, info vars, info exists, dump, set, proc, upvar, uplevel,
global, debugging
string [vfract double ?rounding? ?-float?]
Converts a decimal remainder value into a vulgar fraction string
Any integer part is stripped and ignored. Sign is ignored
The maximum internal resolution is X/1000 ths
If the fraction cannot be converted then the original value is returned
On error or overflow, "" is returned
In order to ensure or force a return a rounding value can be specified
The default internal rounding is 4 decimal places
If -float is specified then both denominator and numerator are returned
as x.0 floating point values thus making re-evaluation easier
Examples:
puts [vfract .75]
puts [vfract 0.343750000]
puts [vfract [fraction $pi] 4]
puts [vfract .318]
puts [vfract 1.5]
puts [vfract 0.75 -float]
Results:
3/4
11/32
177/1250
159/500
1.2
3.0/4/0
See also: fpart, fraction, fract, round
To load, use: lib ticol_variant ?-verbose? ?-nocomplain?
variant [vsplit string ?mask?]
Split a string according to mask characters and return Variant array
If not specified the mask character is a space
Example:
set v [vsplit $qbf]
puts $v
dumparray v
Result:
32539136
DUMP: Variant 'v' at 32539136 (0x1f08200)
Type is 8204 (0x200c) BYVAL
Type is 4 byte float VT_R4
See also: split, join, varray, variant arrays, list
vwait (Requires threaded compile of Ticol) |
Index |
vwait variablename ?timeout-ms?
vwait undefined-variable
[vwait] will block execution and wait until the variable named as its
argument changes. [vwait] is of use only with the threaded compile of
Ticol and the event command [after]
If scripts are executed concurrently then conflicts may occur. [vwait]
allows for script synchronisation
Where a main script calls [after] events (threaded version of Ticol),
[vwait] should be called to suspend execution of the primary script
until all [after] events which reference 'variablename' are completed
[vwait] should not follow [after 0 {code}] as {code} will be executed
immediately and before [vwait] can be evaluated. Thus [vwait] will
never exit
The [vwait] variable argument need not exist. In such case [vwait] will
wait permanently, whilst processing events. [vwait forever] is a common
Tcl idiom where 'forever' is used as a pseudo-variable
Possible Issues
[vwait] will suspend the main event loop so if there is a bug in the
called script then execution may never return to the CLI or to Windows
Examples
--------
Example:
set wait 0
textcolor magenta
puts "I know..."
textcolor
after 1000 set wait 1
vwait wait # Wait until 'wait' changes
textcolor magenta
puts "Tcl, but not Tk"
textcolor
Results:
I know...
Tcl, but not Tk
Example:
proc foo {} {
puts "Hello"
#for {set i 0} {< $i 10000} {++ i} {doevents}
}
puts "Launching \[foo\] in 4 seconds"
after 4000 {set vret [foo]}
vwait vret # Wait until [foo] has launched and completed
puts "Done"
Note:
Be careful to brace after arguments and not use:
after 4000 set vret [foo]
as [foo] will be called before [after]
Example:
vwait forever # 'forever' is a variable which does not exist
Result:
<loops forever>
See also: after
integer [waitkey ?timeout-ms? ?-flush?]
Wait for keyboard input or return on timeout if no keypress occurs
If no keyboard input occurs, 0 is returned. Otherwise the ASCII code of
the pressed key is returned
Example:
set k [waitkey 4000]
puts $k
Result:
0 # If no key was pressed within 4000 milliseconds
13 # If ENTER was pressed within 4000 milliseconds
See also: inkey, input, gets
walk ?count? ?-level value?
Walk is a debugging tool which will let you traverse or 'walk' the raw
Ticol variable table in the current call frame level, showing variable
addresses and other information
Walk shows all variables, distinguishing storage types but doesn't show
information in as much detail as [vars] or allow filtering by name
Example: (From within proc at, say, level 2)
walk 10 -level 0
Results:
walk: Level 0
Type Address Name Value
------------------------------------------------------------
auto 0x1de8f00 b 'start'
auto 0x1de8c10 a 'x'
auto 0x1de8ab0 sw '80'
auto 0x1de8460 fn_name 'testmean'
auto 0x1de84d0 y '14'
auto 0x1de81c0 errorCode '0'
auto 0x1de8160 errorLine '0'
auto 0x1de8100 errorMsg ''
auto 0x1de7fa0 errorInfo ''
const 0x1e27f20 exe_name 'c:\tcl\ticol.exe'
Example:
if {! [defined bar]} {
proc bar {} {
set r "In proc bar"
set set_level_[info level] [info level]
walk 2
mar
}
}
if {! [defined foo]} {
proc foo {} {
set q "In proc foo"
set set_level_[info level] [info level]
walk 2
bar
}
}
walk 2
foo
Results:
walk: Level 0
Type Address Name Value
-------------------------------------------------------
auto 0x2337940 errorCode '0'
auto 0x23378e0 errorLine '0'
walk: Level 1
Type Address Name Value
-------------------------------------------------------
auto 0x233c670 set_level_1 '1'
auto 0x233c270 q 'In proc foo'
walk: Level 2
Type Address Name Value
-------------------------------------------------------
auto 0x233cd40 set_level_2 '2'
auto 0x233c8b0 r 'In proc bar'
See also: introspection, dump, array walk, vars, info, debugging
watch add varname|filter
watch remove varname|filter
watch on|off
watch list
watch trace glob-pattern ?-nocase? ?-value? ?-halt? ?-line N?
[watch] can add variable watches to the single-step breakpoint mode.
Unlimited watches may be enabled but since these are displayed to the
console more than two or three will become unmanageable
Variables need not be in scope at the time of calling [watch]
[watch trace] will trace a variable to the console along with any value
only if that variable exists in the current level context. Otherwise
no output is generated. Optionally, a line number can be specified
using '-line', case can be ignored using '-nocase' and a value can
be glob-matched instead of matching by variable name using '-value'
The '-halt' argument can be used to halt the script when a match is
found. The matched value will be returned for use with [try/catch]
[watch on|off] has no effect on [watch trace]
Example:
# Halt the script when variable value in current scope matches 6*
watch trace 6* -value -halt -line #__LINE__
[watch] (without trace), is active only when running in single-step/
breakpoint mode after a [halt] instruction has been met. The watched value
is shown only after a [watch] command is encountered. The current status
of a variable is shown. If set then the value is shown
Requires the command: ticol.exe filename /bp
or: option breakpoint on
[watch off|on] can set the watch state prior to entering the single-step
debugger to avoid screen-clutter. [watch off] will disable any active
watches and [watch on] will (re-)enable them
Example:
option expression on # Enable flow-control [expr] syntax
option breakpoint on # Enable the debugger
watch add i
watch add result
watch off
halt # Enter the debugger
proc watch_test {} {
for {set i 0} {$i < 20} {incr i} {
puts "Loop $i"
if {== $i 10} {
watch on # Enable the watches
halt # Enter the debugger
set result $::qbf
puts "Return result is $result"
return $result
}
}
}
watch_test
See also: halt, trace, option breakpoint, dump, walk, debugging
integer [wchar_length variable]
[wchar_length] returns the length of a wide (unsigned char) string.
It accepts only a variable which has been set by [string to_wchar] or via
a DLL return as input. A string literal cannot be used.
For example code see: wchar_to_string
See also: string to_wchar, wchar_to_string
string [wchar_to_string variable]
Convert a wide-character (possibly Unicode) variable returned by an
external DLL to ANSI for use with Ticol. Due to the absence of checking,
particular care should be taken when allocating buffers for string
conversion, e.g. not to confuse variable names with values (buflen with
$buflen).
Note that there may be loss of content text during translation
Example: (This uses the AutoIT AutoItX3 DLL)
# Fake passing a Unicode string
set buflen 4096 # Set buffer length var to 4kb
set buf [makestr [* $buflen 2]] # Allocate an 8kb Unicode buffer
puts "Getting clipboard as Unicode..." # We will masquerade as Unicode
call [calldll AutoItX3.dll AU3_ClipGet [strptr buf]:4 $buflen:4]
set result [wchar_to_string buf] # Convert Unicode to ANSI
puts "wchar_to_string result is: '$result'" # Display ANSI result in Ticol
See also: string to_wchar, wchar_length, cast
while (Flow Control) |
Index |
Syntax:
while {condition} {code-block}
while {condition} {
code-block
}
The following additional commands are available for use with [for]
return # Will exit the script when used at root-level
break # Will be passed back to any enclosing structure
stop # Halt operation, return to the CLI if loaded
exit # Exit the script and return to Windows
goto # Where used with an enclosing [goto_block]
continue # Resume execution at the head of the loop
Standard Tcl [while] loops are supported as well as [do..while] loops
A while loop is about twice as fast as an equivalent [for] loop due to
additional syntax checks necessary for [for]
Examples:
while 1 {puts [rnd 1 10] }
while { 1 } {puts [rnd 1 10] }
while { 1 } {
puts [rnd 1 10]
}
Note that you may not use this form. It is not syntactically-valid as
the CRLF is significant and must be "masked" by enclosing braces
while {condition}
{
code block
}
See also: do while, for, if, foreach, switch, flow control
string [win_error errornumber]
Returns a formatted system error string for a given error number
Use [get_win_error] to retrieve the error number
Example:
puts [win_error [get_win_error]]
puts [win_error 2]
Result:
The operation completed successfully
The system cannot find the file specified
See also: get_win_error, info
A number of issues may require workarounds
Problems with escaped quotes
----------------------------
This may occur with complex, nested strings or with scripts which generate
other scripts. Use [chr 34] instead
Example:
puts "One embedded double quote ->[chr 34]<-"
Problems with escaped backslashes
---------------------------------
This may occur with complex, nested strings or with scripts which generate
other scripts. Use [chr 92] instead
Example:
puts "[chr 92]t[chr 92]ricky"
You may also consider escaping or double-escaping backslashes as appropriate,
either manually or using [escape]; and then using [unescape] to revert
Example:
puts "->[unescape [escape "\t"]]<-"
Result:
-> <-
[upvar] and [uplevel] don't Support the # Character
---------------------------------------------------
This is by design since # is reserved by the Macro Preprocessor (MPP)
The # character is a hard comment character in Ticol Tcl
Use @ instead of # for [upvar] and [uplevel]
See also: cli, FAQ
string [wrap string ?width? ?wrap-string? ?-wordwrap? ?-indent N?]
Format a string by wrapping using line feeds at a fixed length
The default width is 80.
A 'wrap string' may be specified which is appended to each line
The default wrap string is "\n" (a newline)
Word-wrapping will be attempted if -wordwrap is used. If the word is too
large, it will be truncated and wrapped to the next line without
hyphenation
An optional -indent value can be specified which will be applied after the
first line. This is commonly used in formatted files such as email headers,
or RLM licence files etc. The indent is subtracted from the line length
Example:
puts [wrap $qbf 10]
puts [wrap $qbf 15]
Result:
The quick
brown fox
jumps over
the lazy
dog
The quick brown
fox jumps over
the lazy dog
Example:
lib ticol_bignum
puts [wrap [big fact 100] 40]
Result:
9332621544394415268169923885626670049071
5968264381621468592963895217599993229915
6089414639761565182862536979208272237582
51185210916864000000000000000000000000
Example:
lib ticol_bignum
puts [wrap [big fact 100] 40 -indent 3]
Result:
9332621544394415268169923885626670049071
5968264381621468592963895217599993229
9156089414639761565182862536979208272
2375825118521091686400000000000000000
0000000
Example:
puts [wrap $qbf 11 -indent 3]
Result:
The quick b
rown fox ju
mps over th
e lazy dog
Example:
puts [wrap $qbf 11 "\r\n->" -indent 1]
Result:
The quick b
-> rown fox ju
-> mps over th
-> e lazy dog
Example:
puts "'[wrap $qbf 20 '[chr 13][chr 10]' -wordwrap]'"
Result:
'The quick brown fox'
' jumps over the'
' lazy dog'
Example:
puts "'[wrap $qbf[chr 32]$qbf 20 '[chr 13][chr 10]' -wordwrap -indent 3]'"
Result:
'The quick brown fox'
' jumps over the'
' lazy dog The'
' quick brown fox'
' jumps over the'
' lazy dog'
Example:
Truncated words
puts [wrap $qbf 4 -wordwrap]
Result:
The
quic
k
brow
n
fox
jump
s
over
the
laz
y
dog
Example:
Truncate with indent
puts [wrap $qbf 4 -wordwrap -indent 2]
Result:
The
qu
ic
k
br
ow
n
fo
x
...
See also: wrap example, puts, printf, format
wrap example script |
Index |
This is an example script for the [wrap] command which reads prose text
from a text file and produces multi-columnar output with balanced columns
and word-wrap
option expression off
set filename test.txt
set columns 3
set height 300 # 300 lines desired
readfile $filename t # Read to temp string t
set column_width [/ [- [- [screen_width] 1] $columns] $columns]
set t [string replace $t "\t" " "] # Replace tabs
set t [wrap $t [- $column_width 1] -wordwrap] # Word-wrap the text
split $t "\r\n" -array a # Split lines to an array
set height [+ [/ [array size a] $columns] 1] # Balance the columns
set row_max [min $height [array size a]] # Guard row_max
dim row $row_max # Speed-up allocation
for {set i 0} {< $i $row_max} {++ i} { # Create blank array
set row($i) "" # We append to this
}
set quit $false # Set break flag
for {set j 0} {< $j $columns} {++ j} { # Iterate and create rows
for {set i 0} {< $i $row_max} {++ i} {
set s [string repeat " " $column_width ]
set k [+ [* $j $height] $i]
if {< $k [array size a]} {
setl s $a($k) #
append row($i) " $s" # Append columns to rows
} else {
set quit $true # Until we reach the max
break
}
}
if {$quit} { # Inner break cannot
break # quit the outer for
}
}
for {set i 0} {< $i [array size row]} {++ i} {
puts $row($i)
}
Results:
Billy had to go strategy? didn't learn nature.
around by way of You know, we've that A piece that can
the living room been collecting the elevator defend you
to get to where stories as wasn't against
his mother was. people for how frightening. anything,
'I didn't know long, That isn't what including
dragons grew so we don't know, they learned. other dragons.
fast', said really, we don't They learned the So what if this
See: wrap, readfile, screen width, string replace, array size, dim,
for, string repeat, array size, break
bool [writefile filename string|varname ?-var? ?length? ?-append?\
?-unescape?]
Write information directly to a file
Returns a boolean 1 on success, 0 on failure
If the -var argument is used then the 3rd argument will be interpreted as
a variable rather than a literal. This can significantly speed up writes,
where -unescape is used, particularly with large files. For small data
sets and where -unescape is not required, -var is not required
Examples:
writefile "fox.txt" $qbf # Write the contents of var qbf
writefile "fox.txt" qbf -var # Write the contents of var qbf
set q [readfile binary.b64] # Read base 64 data
binary set b [binary base64 x -decode]
writefile "bread.bin" b -var [binary length b] # Write binary var b
The length parameter is necessary only for binary data. If used carelessly
it could trigger an out-of-bounds memory read. Particular care should be
taken when escaping or unescaping strings
Escape Codes
------------
Note that [writefile] does not escape character translation and by default,
performs a faithful binary-write of string contents. If strings need to be
unescaped then the [unescape] command should be used to process the data
before writing. Alternatively, for large strings, the "-unescape? option can
be used to avoid a time-consuming call to the [unescape] command
See also: file, unescape
Creating Script-Embeddable Content from Files |
Index |
Occasionally it might be required to embed an external file's contents
directly into a Tcl script. This can be achieved using [base64] encoding
via a very short script as follows...
# Encoding an input file as embeddable base64 content and saving
# to a file...
writefile output.fil [base64 [readfile input.fil]]
The contents of the resulting file may be copied and pasted into a Tcl
script
See also: readfile, writefile, base64
String Comparison Operators |
Index |
The following string comparison operators are supported either
independently or in an 'expr' statement:
bool [eq a b ?-nocase?] # Equality
bool [ne a b ?-nocase?] # Inequality
bool [lt a b ?-nocase?] # Less-than
bool [gt a b ?-nocase?] # Greater-than
bool [le a b ?-nocase?] # Greater-than or equal to
bool [ge a b ?-nocase?] # Less-than or equal to
bool [is_empty a] # If set but empty
bool [is_set a] # If a variable is set
bool [eqi a b] # Equality (no case)
bool [nei a b] # Inequality (no case)
bool [lti a b] # Less-than (no case)
bool [gti a b] # Greater-than (no case)
bool [gei a b] # Greater-than or equal to (no case)
bool [lei a b] # Less-than or equal to (no case)
bool [eqn a b length ?-nocase?] # Test [eq] over max length
bool [nen a b length ?-nocase?] # Test [ne] over max length
The [eq] and [ne] commands are enhanced non-math commands in that they
will compare empty string values "". If a value is empty, no argument is
supplied and an empty string will be presumed. The following results are
generated where 'var' is any valid (non empty) variable.
Relational operations (>, >= etc.) on empty strings are undefined and will
generate an error. This resolves a traditional Tcl problem in testing for
null command returns
These comparison commands are much faster than chained commands of
[string equal] and [string left]
The result of command expressions containing spaces should be wrapped in
double quotes, e.g.: if { eq "Hello world" "[myfunct]" } { ...
The [eqn] and [nen] operators are efficiency friendly ones which provide
fast and simple length equality tests without cumbersome code. Note that
a length value of 0 will compare 2 null strings {} and {} and will thus
return true
eq "" "" => true
ne "" "" => false
eq var "" => false
ne var "" => true
eqn foo bar 0 => true (compares {} and {})
eqn foo bar 1 => false
Example:
set a "Hello"
puts [bool [eq "" ""]] -> true
puts [bool [ne "" ""]] -> false
puts [bool [eq $a ""]] -> false
puts [bool [ne $a ""]] -> true
puts [bool [eqn $a "Hell" 4]] -> true
puts [bool [nen $a "Fred" 4]] -> true
puts [bool [eqn $a "he" 2 -nocase]] -> true
puts [bool [nen $a "hell" 4]] -> true
puts [bool [eqn $a "hello" 5 -nocase]] -> true
Example:
if { eq [set x [gets stdin]] "" } {
puts "Nothing was entered"
} else {
puts "You entered '$x'"
}
When using relational commands in an expression via [expr] or via a flow-
control command with [option expression on] be aware that qualifying
arguments are unsupported in that scenario. You must embed the command
style within the expression
Example:
puts [expr "Fred eq fred" -nocase] # Wrong
puts [expr [eq Fred fred -nocase]] # Correct
puts [expr "Freddy eqn Freda" 4] # Wrong
puts [expr [eqn Freddy Freda 4]] # Correct
See also: math operators, string
string [escape string]
The interpreter must continually reinterpret source and can't store
permanently unescaped 'special' backslash characters. This can create a
need to escape or unescape strings which contain backslash characters
Console output such as [puts] will unescape strings automatically but
other commands such as [writefile] where there may be a need to retain
literal escape characters do not and will require the use of [unescape]
[escape] will escape a string by inserting backslash characters before
detected special Tcl characters such as $ or [. This may be required for
strings with literal backslash (\) sequences as returned from external
DLLs to be used by Ticol. [unescape] is the inverse
Escape sequences are handled differently in the Command Line Interface to
within running scripts. See: option escape, escape sequences
Commands such as [ls] will return strings containing backslash characters
which will require escaping before printing using [puts] or [printf]
A useful workaround for path problems is to use the Unix forward slash
character - '/' which avoids the need to escape when handling paths
Example:
# The location of ping32.dll is C:\tools
puts "DLL filename: [calldllstr ping32.dll GetDLLFileName]"
puts "DLL filename: [escape [calldllstr ping32.dll GetDLLFileName]]"
Result:
C: oolsping32.dll # BAD - Inadvertent \t TAB sequence in \tools
C:\tools\ping32.dll # CORRECT
Example (From the CLI):
set s "c:\windows\media"
puts $s
set s [escape "c:\windows\media"]
puts $s
Result:
c:\windows\media
c:\\windows\\media
Example (from the Windows CMD line):
ticol.exe ; "puts [cd {\\\\server\\share\\path}]"
if the entire expression is not wrapped then the following errors may
occur:
Error: 1 too many [ brackets found
<argument> is not recognised as an internal or external command
See also: unescape, escape sequences, puts, set, option escape
Ticol Macro PreProcessor Overview (MPP) |
Index |
The macro-PreProcessor emulates some of the block-control behaviour of a
C++ PreProcessor. See also: #if, #echo, #exec, #exit
#ifdef variable
#ifndef variable
#define variable
#undef variable
#if <tcl-expression>
#else
#endif
#exec
#deflist
#echo
#exit
#define variable ?value? Define a macro control variable
Full definition:
#define<whitespace>LValue?<whitespace>RValue?<whitespace|crlf>
Macro variables are not replaced within quoted
strings in a source script
The rvalue may be a Tcl expression but the
expression will be evaluated at run-time.
The expression may contain values which have been
previously defined
Defined values may optionally be wrapped in double
-quotes and may not contain embedded tabs or
spaces unless so wrapped. Only a single wrapper of
quotes are stripped. Any intermediate quotes
remain but the Preprocessor will require such
quote characters to be escaped and be properly
paired
You may use [chr 34] to specify single quotes
Macro variables may be full TCL expressions
providing that they are defined on a single line.
Defined expressions containing spaces must be
wrapped in double-quotes
Variables are literal substring replacements and
not Tcl variables. It is therefore recommended to
use capitals for macro variables to avoid confusion
and unpredictable results.The use of [const] or
[enum] is preferable to #define
The defined rvalue (if any) must be followed by
at least one whitespace or end of line character
#undef variable
Undefine a previously #define(ed) variable/value
#if <tcl-expression>
Conditionally process according to a Tcl expression
Must be paired with an #endif
May have an optional #else
Nested #if... statements are not supported
#ifdef variable
Apply control block if variable is defined
Nested #ifdef... statements are not supported
#ifndef variable
Apply control block if variable isn't defined
Nested #ifndef... statements are not supported
#else
Apply control block if previous block was false
#endif
Mark the end of an if..else control block
#exec <tcl-expression>
Execute a Tcl statement within the MPP context
#echo string ?# comment...?
#echo "string" ?# comment...?
Echo a string from the Macro PreProcessor
This may only be a simple string
Only basic escape.characters are processed,
including \r\n\t
The string may be quoted or unquoted
May not contain another macro command other than
a comment, which will truncate the line
Macro substitution will not occur within strings
#deflist List all current '#define's
#exit
Immediately exit macro processing
No further Tcl source code is output and the
source code is truncated at this point
The macro-PreProcessor is run before program execution and it may be
disabled using the /NP argument
Defined macro variables are not replaced within string constants
Macro variables are case-sensitive
A macro command must be contained on a single line. This line may be
extended by the use of line-continuation characters (\). Due to this
restriction a macro variable cannot contain complete blocks of Tcl code
which depend on newline characters for syntax. Nor can comments appear
after a line-continuation character.
Macro symbol replacement is recursive up to a depth of 100 definitions.
i.e. a #define Lvalue may define an Rvalue consisting of another Rvalue
already defined. This includes Rvalues embedded in subexpressions
It may help code-clarity for defined PreProcessor variables to be in
upper-case
Remember that macro commands are executed during the preprocessing phase
so all macro commands are executed *before* the script is run
Messages emitted by the MPP may also be redirected and used for logging
purposes
Examples:
#define comment
#define ESCAPE_CHAR 27 # Define an integer value
#define ESC ESCAPE_CHAR # Secondary definition
#define HELLO "puts \"Hello \ # Multi-line definition
world\""
#define PI_APPROX "[expr \"22/7.0\"]" # Complex expression
#ifdef PI # Test var presence
#undef PI # Undefine a macro variable
#define PI "[expr 4*atan(1)]" # Define an expression
#endif
#if file exists ticol.exe # Conditional tests
#echo File ticol.exe exists
#endif
Tcl # comment statement are completely stripped out of the source code,
as are invalid #if... blocks. This significantly speeds up execution
Nested #if/#endif expressions are not yet supported
The macro variable "true" is pre-defined but may be #undefine'ed
Note that the Macro PreProcessor operates only on loaded/run source code
and not within the Ticol CLI. Macro variables are not persistent outside of
an individual run session. Macro variables are discarded once execution has
returned to the CLI
A macro #include command is not provided as this functionality is provided
via [eval <script-name>]
Scripts With No Tcl Code
------------------------
It is possible to use the Macro processing language to write a script which
contains no Tcl code at all. For example, the following working CGI script
contains only macro language...
# Configure the CGI instance of Ticol's ticol.ini
# [config] -> PreprocesorVerbose=FALSE to suppress MPP warnings
#echo "Content-type: text/html\r\n\r\n"
#echo "<html><head><title>Tcl MPP Echo test</title></head"
#echo <body>
#echo "<br><center><h1>This file contains no Tcl code</h1></center>"
#echo </body>
#echo </html>
See also: escape sequence, debugging, macro variables,
#__LINE__, #__DATE__, #__TIME__
Macro #if #exec #echo |
Index |
#if <tcl-expression>
#echo unquoted-literal-string
#echo "quoted-literal-string"
#exec <tcl-expression>
#else
#endif
#exit ?errorcode?
Instruct the Macro PreProcessor (MPP) to conditionally process a block of
code depending on the result of <tcl-expression> where this is any valid
Ticol Tcl expression. The expression-handler is local to the MPP so knows
nothing about global variables. Nor can it be used to set variables which
will be used at runtime by the script itself
#if
---
#if <tcl-expression>
#else
#endif
#if must be matched with #endif and may also include an optional #else
Unlike C/C++ the start of any # macro statement may be indented using
whitespace characters
# comments or /*...*/ comments may follow any of these commands and the
expression argument must be contained wholly on the same line
The script is persistent to the pre-processor instance so local variables
set by one #if statement may be use by a later one
Global or other variables are not available
Keyword command filtering is not applied
#if allows a script-block to be executed conditionally according to a
wide-variety of conditions such as the presence of files, the host system,
date/time, local environment settings etc.
The MPP interpreter instance will use default values. You are free to call
[option] (or any other command
Macro variable substitution is performed, including __LINE__
#echo <literal-string>
----------------------
#echo may be used to print a literal text string from the MPP
There is no need to enclose the string argument in double quotes
No escaping is performed and all characters are printed literally
The printed string need not be enclosed in double-quotes and macro
variable substitution is performed, including __LINE__
Strings need not be wrapped in double-quotes, however if the first
and last characters (excluding whitespace) are double quotes then
these will both be stripped before printing. Otherwise, all double
quotes will be ignored
Example:
#exec set pie 3.142857142857143 # Approximately pi
#if {== [/ 22 7.0] $pie}
#echo "Conditional macro SUCCESS"
#exec newline
#else
#echo "Conditional macro FAILED"
#exec newline
#endif
Result:
Conditional macro SUCCESS
Example:
#if eq $env(computername) TORNADO
#echo Running on TORNADO
#else
#echo Not running on TORNADO
#endif
Result:
Running on TORNADO
#exec <tcl-expression>
----------------------
#exec will cause its argument to be evaluated in an MPP interpreter
context as a Tcl expression. This may be any Ticol command or expression
The result will be effective in the MMP context only. The MPP interpreter
instance will be destroyed when the MPP exits
#exec could be a security risk when using scripts which you have not
developed yourself, particularly if scripts are obfuscated as TCX files.
When using scripts written by others this option should be disabled
#exec may be disabled using /NMX at the command-line or by setting the
INI configuration file section [Config] -> PreprocessorExec=FALSE
Macro variable substitution is performed, including __LINE__
#exit ?integer-code?
--------------------
#exit will immediately exit macro preprocessing without further action
source output will be truncated at the point of exit and no further macro
processing performed. All script prior to #exit will be evaluated
If a numeric argument is given then this will be returned to Windows as an
ERRORLEVEL value when Ticol exits after script execution completes
Example of conditional script interpretation:
#if % [clock seconds] 2
#exec textcolor red
#echo Clock is odd
#exec textcolor
#exit 255
#else
#exec textcolor green
#echo Clock is even
#exec textcolor
#exit 254
#endif
See also. MPP, #ifdef, #else. #endif
The following Macro PreProcessor (MPP) constants are available. These
should be placed at the end of any line and will commonly be used in
conjunction with [set] or [format]. Multiple #-prefixed variables
are not permitted.
Note that __FILE__ is not defined when calling a script using [eval]
#__DATE__ Returns an unformatted ISO date YYYYMMDD
#__FILE__ The current full script name/path when called using [run]
#__HOST__ The computer host name of the system being run on
#__LINE__ The current script line number as an integer
#__TIME__ The current time in 24 hour clock format as HH:MM:SS
#__USER__ The currently logged-in username (if any)
These values, which are interpreted as comment fields, may be passed
into Tcl variable by placing them as the last clause of a statement
__LINE___ is available as a direct macro variable including within
#echo, #exec and #if
Example:
set cdate #__DATE__
set cfile #__FILE__
set cline #__LINE__
set ctime #__TIME__
printf "%s" #__DATE__
puts "In script $cfile, line $cline, date is $cdate at $ctime"
Result:
In script C:\tests\test.tcl, line 33, date is 20171007 at 01:00:43
See also: macro, macro preprocessor
Macro Escape Sequences |
Index |
These require [option escape on] or configuration of TICOL.INI
Hex and octal escape sequences inside strings are checked and translated
by the Macro PreProcessor on running a Ticol script
Literal backslash characters within a string should be represented by a
double-backslash character. e.g. "c:\\windows\\system32\\notepad.exe"
Strings which are not within a source file i.e. which are read from file or
input by a user at the Ticol CLI are not passed through the PreProcessor
and are not interpreted. This enables the Ticol CLI to be used as a
general-purpose interpretive Windows shell.
e.g. "c:\windows\system32\notepad.exe"
These should be in the form...
\xN, \xNN for hexadecimal sequences
\N, \NN, \NNN for octal sequences
Example: puts "This is a blob character \xfe."
Result: (square blob character)
Example: puts "\x68\x65\x6C\x6C\x6F"
Example: puts "\150\145\154\154\157"
Result: Hello
Example: puts "Hello\tworld"
Result: Hello world
Processing will halt with a line-number indication if an invalid sequence
is encountered, e.g. "\x ". You can view the translated code using the
echo (/E) argument. The advantage of preprocessing is that such codes
can be checked for validity and code execution speed is increased.
Square brackets (and other characters) may be escaped by using the [chr]
command as [chr 91]->'[' [chr 93]->']' or \[ and \]
The comment prefix character (#) may be escaped using \#. The Macro
PreProcessor will treat \# as a literal # character
Other escape sequences:
Sequence Description Processed by
----------------------------------------------------------
\\ Literal backslash Interpreter
\$ Dollar sign Interpreter
\( Open bracket Interpreter
\) Close bracket Interpreter
\[ Open square bracket Interpreter
\] Close square bracket Interpreter
\{ Open brace Interpreter
\} Close brace Interpreter
\" Double quote Interpreter
\a Bell Macro PreProcessor
\b Backspace Macro PreProcessor
\f Formfeed Macro PreProcessor
\n Newline Macro PreProcessor
\r Carriage return Macro PreProcessor
\t Tab character Macro PreProcessor
\v Vertical tab Macro PreProcessor
Macro PreProcessors escapes may be disabled with the /NP argument
Not supported:
\' Single quote. Preprocesor restriction - use [chr 39]
\" Double quote. Preprocesor restriction - use [chr 34]
\0 Not supported. Directly manipulating null terminators
within Ticol strings should be avoided as it may lead to
memory corruption
Examples:
puts "hello\b\b\b\b\bWorld"
puts "[chr 91]Hello[chr 93]"
Results:
World
[Hello]
See also: escape, macro, option, unc paths
A few features are provided to assist debugging
Pseudo-macro items may be used as the first item within a comment area
to add line number, date and time as follows. (Case is ignored)
puts "Line: "#__LINE__ This is a comment which should be ignored
puts "Date: "#__Date__ So would this be
puts "Time: "#__time__ And this too
The Macro PreProcessor will check a Tcl file on loading for mismatched
brackets/braces and double-quotes. An indication is given as to the
likely location of any braced section with a mismatch
Line numbers may be traced by setting a variable using: set line #__LINE__
Single-step execution with breakpoints is provided by option breakpoint and
the [halt] instruction
array walk Walks though the array hashtable showing nodes and content
assert Assert a value against a Tcl expression
breakpoint option breakpoint. Enable single-step breakpoints via [halt]
cmdcount Count of the # of executed commands
commands List defined commands
debug print Debug output which can be globally disabled in release code
debug Enables or disables debugging of a debug release of Ticol
dummy Dummy command. Just echoes its arguments
dump Performs a binary dump of variables or structs
empty Empty command. Does nothing
explain Expand [calc] and resolve into Tcl commands
functions List known functions
get_win_error Show last Windows error code
halt Enable a breakpoint halt point
inspect Inspect raw variable information
lib... -modules shows loaded modules and -info shows version info
level Return the current [proc] level
#macros Macros allow conditional interpretation for test purposes
mem_used Returns the current memory used by Ticol
procs List defined procs
resume Disable breakpoint single-stepping until next [halt] is met
sc Compare two strings showing the offset of any mismatch
stacktrace option stacktrace. Stack trace dump on fatal error
stack list List a stack's contents and show state
struct dump Show detailed information about a struct's memory
time Times the execution of a block of code, repeating N times
trace Traces the execution of selected commands
vars Lists variables by search prefix and type in any level
walk Walks the variable table showing addresses, type and content
watch Set a variable watch with breakpoints or trace in real time
win_error Describe a Windows error code
Output or string return commands require the use of [puts] or [option
echo on] otherwise no output will be seen
[call] is an empty command which does nothing.
Example:
call [puts Hello]
Result:
Hello
See also: assert, trace, time, breakpoints, halt, dump, dummy, debug, walk,
vars, mem_used, performance, faq, PreProcessor, macro, #if, #else,
#endif, option stacktrace option breakpoint, halt, resume, watch
integer [dummy]
string [dummy -return]
list [dummy -list]
[dummy] is a dummy Tcl command which does nothing other than print out its
own argument list and return the argument count.
Note that if this argument list contains commands then these may be
evaluated before the call to [dummy]
[dummy] may be useful for checking that what is passed into a command
matches expectations and verifying the count of any expanded arguments
If the -return argument is used then the arguments will be concatenated
and returned as a string
If -list is specified then a formatted Tcl list will be returned
Example:
assert [dummy a {*}{b [c]} d {*}{$e f {g h}}] {== $_ 8} -v
Result:
Enter: [dummy] argc == 8
0='dummy'
1='a'
2='b'
3='{[c]}'
4='d'
5='{$e}'
6='f'
7='g h'
assert: PASS '== $_ 8'
Example:
dummy [printf hello]
Result:
Enter: command [dummy] argc==2
dummy: 0='dummy'
dummy: 1='5'
Example:
puts [dummy a {*}{b [c]} d {*}{$e f {g h}} -return]
a b {[c]} d {$e} f g h
See also: empty, args, debugging
void [empty ?...?]
[empty] is an empty Tcl command which does nothing. It accepts any number
of arguments and ignores them. It returns a null ("") result.
See also: dummy
Windows DLL Test Interface (wintest.exe) |
Index |
A simple Windows GUI interface is provided which links to ticol.dll and
is able to either load and run Tcl scripts or to launch the CLI within
ticol.dll. Note that the underlying Ticol interpreter is ANSI only.
This interface is quite rudimentary and intended for testing only. it was
written in a legacy copy of Visual BASIC 5.0 to enable DLL linkage to be
tested.
Wintest has options to enter the CLI, load/save Tcl files, clear the
code viewer and set breakpoints or stacktrace.Rudimentary editing is
possible but not recommended. With the Command Line Interface loaded
([cli]) the interface will behave the same as ticol.exe
Wintest is typically supplied with no installer disk set although one
can be generated if requested. Usually one needs only to have msvbvm50.dll
available in the PATH or current folder
To install: Copy the following files to the same folder:
wintest.exe # The Windows binary (UPX compressed)
ticol.dll # The DLL version Ticol Runtime
msvbvm50.dll # Visual BASIC 5.0 runtime DLL
ticol.ini # Optional configuration file
autoexec.tcl # Optional startup Tcl file
You can pass the name of a Tcl file to load to wintest.exe at the
Windows command line, thus:
wintest.exe foo.tcl
You may also optionally run this script by appending the /R argument to
the command line, thus:
wintest.exe foo.tcl /r
The [exit] command is used to exit any spawned CLI session
Wintest should not be used for in a production environment
See also: ticol, ticol.dll, plugin
Performance and Efficiency |
Index |
Ticol is a pure interpreted script language. It does not perform any byte-
code compilation or dynamic optimisation, so performance will be slower
than other Tcl variants such as ActiveState. However, since Ticol was
designed for day-to-day administrative script work this should not present
any serious problems
If performance is an issue then it is recommended to use another such as
ActiveState Tcl
The Ticol Macro PreProcessor (MPP) is able to statically optimise some
aspects of the script source code but the result will still be interpreted
The MPP strips out comments, so they are not interpreted as commands by
Ticol
Loop-processing constructs will not be as efficient as a compiled language
such as C#. however many of the more complex list-processing or string-
splitting operations are quite efficient. Ticol offers a wide range of
commands and some plugin libraries which are "closer to C++" and therefore
fast and efficient
You should avoid single-character loop-processing within Tcl, particularly
for loops which necessitate complex branched if statements.
Preferably call a dedicated, external DLL function and use Ticol as the
'glue' to hold together more efficient components. A simple plugin
interface library is offered for C++ programmers.
The following commands are offered to enhance performance:
any Avoids slow/nested logical or [||] statements
all Avoids slow/nested logical and [&&] statements
array item Ambiguous indexing of arrays with default value
calc Static, self-optimising equivalent to [expr]
Translates expressions into Tcl commands
chr Return a character by number with optional offset
cmp Efficient compare function
comma Efficient comma-formatting of numbers
concat Efficient string concatenation
eq, eqi Efficient comparison commands
eqn, nen Shorthand equivalence/disequivalence
funct Call any expr function directly
ge, gei *i is the case insignificant version ...
gt, gti Greater than (gti:ignore case)
index Rapid single char index into a string
is_mod Modular arithmetic test
le, lei Less than or equal (ignore case)
lt, lti Less than (lei:ignore case)
makestr Create a new string
ne, nei Not equal (nei:ignore case)
setr Assign to a string and align
setat String manipulation
store Buffered memory storage
strsplit Split a string
strstr Locate a substring
tohex Convert an integer to hexadecimal
The Following Commands Reduce CPU Stress
sleep Pause execution for X milliseconds
doevents Yield briefly to Windows
Move [proc] definition, list evaluation and creation outside of loops if
possible. Pre-evaluate lists into variables.. Make full use of the
macro-PreProcessor to optimise constants during the preprocessing phase
Use [concat] when concatenating large strings. Avoid chaining the [set]
command as the accumulated string passes through the interpreter and
is repeatedly recopied.
POOR:
set out $out$ch
GOOD:
concat out $ch
Use [cmp], this compares string segments in memory, instead of extracting
and then comparing the resulting extract in an expression. It can
potentially avoid copying large blocks of data. 'time' tests show it to be
around 30% faster especially within loops
POOR:
if { eq [mid $data $i 2] "Hi" } {
GOOD:
if { cmp data "Hi" $i 2 } {
When extracting small sections of a string use [setat] to directly slice
the section of a string to a variable, thus avoiding the use of the set call
This sets a variable directly with index values located at a specified
position and length. Timing tests show [setat] is around 50% faster than
using [mids] together with [set]
POOR:
set ch [mids $data $i 1]
GOOD:
setat ch $data $i 1
Character replacement can be performed rapidly using [replacechar] rather
than re-assigning the result of [string replace]
POOR:
set s [readfile ticol.man]
set s [string replace $s "\t" " "]
GOOD:
set s [readfile ticol.man]
replacechar s "\t" " "
To concatenate strings, particularly large ones or large numbers of small
ones use [append] instead of repeated calls to chained [set] commands.
The use of [set] chains creates multiple copies of strings whereas append
appends directly to the original variable. Append is faster than [lappend]
and [store] is about 5 times faster than append but less flexible to use
POOR:
set a "The quick"
set a "$a brown fox"
set a "$a jumped over"
set a "$a the lazy dog"
GOOD:
set a "The quick"
append a " brown fox"
append a " jumped over"
append a " the lazy dog"
BEST:
store "The quick"
store " brown fox"
store " jumped over"
store " the lazy dog"
set a [store get]
A key to interpreter performance is pre-calculating as much as possible
outside of intensive loops and making good use of intrinsic functions
instead of synthesising them from other commands
String comparison for string prefixes should use the [eqn] and [nen]
commands which will safely test equality or inequality for a given length
This would be effective in tight loops when comparing large string prefixes
than using a compound of [eq] and [string]
POOR:
set a "The quick brown fox"
if { eq [string left $a 3] "The" } { do something ... }
GOOD:
set a "The quick brown fox"
if { eqn $a The 3 } { do something ... }
Use [for]. For loops are often far more efficient than while {} loops
For accurate performance timing, use the 'time' command or run Ticol in
"graph" mode with the /G and /NA command line arguments
The performance of individual commands can be estimated more accurately
when running directly and outside of complex scripts, e.g.
ticol.exe ; "time {loop i 0 100 {}} 100" /g /na
ticol.exe ; "time {if {} {}} 100" /g /na
ticol.exe ; "loop i 0 100 {time {}}" /g /na
ticol.exe ; "time {set a 10} 100" /g /na
[let] may be used for multiple, simultaneous assignments
Use [array foreach] to iterate arrays if you need simple behaviour
Because it is a generalised expression-handler with significant checking-
overheads, [expr] may be far slower than decomposing an expression into
dedicated component operations with Tcl 'prefix' syntax. The Macro
PreProcessor will optimise the [calc] command and expand into raw Tcl
statements. If the MPP is disabled then [calc] will be interpreted
equivalent to [expr]
Example:
set src [expr ($x & $y) % 3] # Expression
set src [% [& $x $y] 3] # Decomposed expression (~50% faster)
set src [calc ($x & $y) % 3] # Expression is optimised by the MPP
See:
http://wiki.tcl.tk/348
Move command calls out of loop arguments and bodies where possible to
avoid repeated calls to slower evaluation commands
Example:
option expression off
for {set i 1} {< $i [linkedlist count $handle]} {++ i} { # Slower
set c [linkedlist count $handle] # Evaluate once
for {set i 1} {< $i $c} {++ i} { # Faster
See also: optimising, mpp, calc, performance graph, performance example,
time, timer, clock, cat, cmp, setat, let, Ticol, Tcl syntax, faq
Performance Example |
Index |
The script hanoi.tcl can be used to check performance. An equivalent EXE
version (hanoi.exe) which uses the same libraries is included for
comparison
ticol hanoi.tcl 17
- should take about 4 seconds on an Intel Core i5 with an idle CPU and
about 3 seconds on a core i7 with idle CPU
The precompiled example hanoi.exe will take about 25ms to perform the
same task
See also: performance, performance graph, efficiency, Ticol
Performance Graph (Profiling) |
Index |
The command-line /G option will generate a graphic profile analysis of
command calls so that code can be optimised by substituting more efficient
operators for those command which are called most frequently
The graph will show the total number of calls, a graph of that count and
the average (mean) time taken per command call in milliseconds. Note that
this time includes all polymorphisms of commands such as [time] or [date]
which may have significantly varying execution time.
Functions which typically call script snippets such as [while], [for],
[foreach] etc. will show far higher execution times as these command calls
include many others which are listed separately
The timer resolution is approximately 1 microsecond (us) per command call
and the timing in microseconds (us) per call is should be treated as an
estimate. Due to the accuracy of microsecond timers on Windows some timings
may be too small to estimate and are rounded up to 1 us
Graphing has no affect on performance when disabled and a small effect on
performance whilst enabled. Using [time] with /G may give a misleading
view on actual code-efficiency due to the /G analysis impact on performance
Call time per microsecond may vary significantly from evocation to evocation
and the total times include only function calls excluding interpreter kernel
processing time
Note that the final total time is the sum of all commands and will usually
be greater than the actual runtime. This is useful only as a guide to
performance. Total execution times may differ substantially from actual
runtimes due to the problem of compound commands and procs repeating the
total execution time of endpoint commands they call. To gain a more
accurate estimate, run each command separately
Procedure names are highlighted in magenta, compound flow-control commands
which call [eval] or [expr] are highlighted in yellow. Particularly slow
commands are highlighted in red and have the display value scaled down to ms.
Example:
ticol.exe adapters.tcl /g
Results: (Actual graph is displayed in colour)
Performance Analysis Graph for adapters.tcl
Command us/Call Calls Call Graph
puts 691 114 **********************************************
info 21616 103 ********************************************
incr 1 30 *************
textcolor 67 27 ************
if 108945 23 **********
< 4 12 ******
set 1 11 *****
gotoxy 73 9 ****
eq 0 6 ***
== 0 6 ***
+ 0 6 ***
Totals: 386
Time: 6,054,842 µs (6.0548 seconds)
See also: performance, debugging
Ticol Design and Picol Design Changes |
Index |
Based on the original Picol source this code has been significantly
re-engineered.
Significantly Redesigned Code
------------------------------
Command lookup now uses a sorted list with binary search
The parser system is modified to add new features including "{*}"
[proc] is redesigned and more complex
[eval] code now applies [subst]
[eval] code block applies [trace]
[eval] code block applies /G performance graphing/profiling
[eval] code block applies stack tracing and debugging
Fixed size buffers are eliminated
Code is optimised by preprocessing using the MPP
Variable parent inheritance chains are enabled
Square braced numbers are not interpreted as procs ([123])
Additions
---------
Multi-level variable dereferencing with $$
Macro preprocessor (MPP) added
Arrays are implemented using hashtables
Structs are new and implemented using C-compatible binary structs
The ability to call/exchange data with external DLLs is added
[expr] (and [calc]) expression handling have been added
[subst] is added
[@], [exec] and [spawn] are added to call external apps
[unknown] can shell out to Windows apps
Plugin support added via [lib] with several plugins made available
Command line help added
Ability to call external DLLs using [calldll]
Sub-language plugins for bf, VBScript and pseudo-assembler
Retained
--------
Simple FIFO variable tables are retained as these are efficient
enough and keep the most recently added variables at the top
See [walk]
See also: design goals, MPP, faq
The following design goals were adopted when designing Ticol Tcl.
* Single, tiny EXE file, fully portable with no installation needed
* Final compressed EXE size of about 200 Kb
* Speed not vitally important as it is intended for low-priority tasks
* Plugin extension DLLs, including ZIP, big numbers, email etc.
* Big math library with a range of useful functions
* No inherent system registry dependencies
* To offer a useful IT Administration tool as an adjunct to the CMD shell
* To provide a script language which can ease many routine admin tasks
* Not to emulate the full Tcl language but to focus on a useful subset
* Tk would not be implemented as it was not required
* Fast interpreter performance with minimal code complexity
* Avoid unnecessary complexity and unnecessary features e.g. namespaces
* Minimal socket/internet connectivity to keep the code simple
* No external dependencies other than the MSVC runtime
* Ability to link to 3rd party DLLs to 'lever' Ticol's usefulness
* Provide a basic-level conditional Macro-PreProcessor (MPP) facility
* Offer MPP optimisation of expressions where possible
* A built-in help/man facility with example code and find feature
* Enhance comments and provide a long-comment mechanism
* To improve, where possible and at minimum cost, selected Tcl features
* Add common script idioms as built-in commands for maximum performance
* Ability to be called by BAT/CMD or other script files, e.g. FOR /F
* Ability to spawn console apps and grab text output to a Tcl variable
* Export to a linkable DLL for other programs to utilise Ticol scripting
* A more reliable replacement for SET /A in evaluating script expressions
* Include useful new commands which simplify and speed up Ticol code
* Support hashed Tcl arrays and C++ struct emulation for DLL linking
* The widest range of compatibility across all Windows versions and ReactOS
* Ability to act as a useful Windows console shell
* Syntax-highlighted code-viewer
* To support Windows from NT4/SP6 through Windows Server 2016
* ANSI-only as this is a personal, not a commercial project
Future goals:
* Regular expression handling
* Extended API file copy/move
* Possible limited socket functionality
* Reduced EXE size, more
* More plugins
* Nested #if...#endif macro blocks
See also: Ticol, design, disadvantages, faq
Ticol Disadvantages |
Index |
Lack of Unicode / Non Western Character Sets
--------------------------------------------
Ticol was designed to meet a particular design spec, the majority of which
was for personal use. One of the goals was to reuse all existing C
library code, all of which was ANSI (ASCII) based in order to get the
project up and running as quickly as possible. The other was that it
was mainly for home and small office use in the UK only. Consequently
there was no need to invest time rewriting the base library code to
handle non-UK characters. This was a design constraint, not an oversight
The language fulfilled particular requirements and was in daily use
on a number of Windows servers for several years and a LAN monitoring
and diagnostic web page was running under Ticol CGI for at least 2 years
Ticol is shared with anyone who might find it useful but it should be
acknowledged that if they have non-UK (e.g. "unicode" / "wide" character
requirements then this program was not designed for them.
If you have a pressing need for Unicode / non-UK character support then
I can strongly recommend ActiveState Tcl or "Jim" Tcl.
Interpreters are Slow
---------------------
Ticol is fully-interpreted which means that opportunities for optimisation
are limited and it will never run as fast as a compiled language. The
Ticol Pseudo Assembler language does run quite fast though. No part of
Ticol code is precompiled into say "byte code". Despite this Ticol
isn't exactly a slouch and can perform many scripting tasks quite
quickly. Plugin code is compiled in C++ which means it runs at full
speed.
Lack of Threading Support
-------------------------
I didn't need this and other than some experimental code and an
experimental threaded version very little time has been invested in
making Ticol threaded. ActiveState is recommended if you need threading
Lack of Full Socket Support
---------------------------
Other than a few specific network based commands I didn't need full
socket support so if you want to write a webserver in Tcl then again
you will need to use ActiveState
See also: Ticol, design goals, design, faq
Ticol file functions will handle UNC paths such as Windows SMB share paths
The handling of share paths may require careful use of [escape] and
[unescape] if Windows backslash '\' separator characters are used
It is recommended to use Unix forward-slash '/' characters instead
Examples:
# [unescape] is required for all escape sequences
set filename "\\\\server\\share\\c\$\\path\\to\\file.txt"
readfile [unescape $filename]
# [unescape] is required for the dollar $ character escape
set filename "//server/share/c\$/path/to/file.txt"
readfile [unescape $filename]
See also: file, readfile
string [unescape string]
Unescape a string by de-duplicating backslash character pairs. This may be
required for strings with literal backslash (\) sequences as returned from
external DLLs to be used by Ticol, or when generating, say RTF output.
[escape] is the inverse
Ticol requires scripts to have strings in "escaped" format if they contain
literal backslashes. This is achieved by using a double-backslash as with C++
Standard escape characters such as \r \n \b are processed by the MPP
Only a small number of commands intrinsically unescape backslashes
(e.g. [puts]). Other commands such as file/string-handling commands may need
[unescape]
Example:
append s [unescape "\\b0\\fs18\\f0 \r\n\r\n\\par "]
^ ^ ^ ^ # Escaped backslashes
^ ^ ^ ^ # Escaped by the MPP
Result:
# $s contains the sequence "\b0\fs18\f0 CRLFCRLF\par "
Where CRLF is the byte-sequence 0x0d0x0a
See also: escape, puts, set, option escape
urlencode, urldecode |
Index |
string [urlencode string]
string [urldecode string]
Encode a string containing an HTTP URL
Example:
puts [urlencode "http://www.mysite.com/page with space"]
Result:
http%3A%2F%2Fwww.mysite.com%2Fpage+with+space
See also:
White space is relevant in Tcl outside quoted strings
See also: Ticol, tcl, syntax
INI Handling Plugin |
Index |
Load using: lib ticol_ini?.dll?
string [ini read inifile section key]
bool [ini write inifile section key value]
string [ini read bool inifile section key]
bool [ini write bool inifile section key value]
integer [ini array read inifile section key array-variable ?-max N?]
bool [ini array write inifile section key array-variable]
Various INI file handling routines. The 'section' value must not be
enclosed in square brackets. The file may have comments, delimited by a
semi-colon (;)
[ini read] will read a standard INI file value from an INI file
[ini write] will write an ANSI variable to an INI file with the given
section and key
[ini read bool] performs the same as [ini read] but returns a boolean
value of 1 or 0 depending on whether the key has a value of 1/0,
True/False, or Yes/No
[ini write bool] writes to file the same as [ini write] but evaluates the
value argument and writes either "True" or "False"
[ini array read] will read up to 1000 items by default with an integer
sequence matching the 'key' prefix (e.g. Item1...ItemN) with an array base
of 1 (not 0). This maximum can be altered by using the -max N argument,
e.g. puts [ini array read $inifile config Item q -max 750].
[ini array read] returns an integer count of items read into the array.
Sparse items in the source INI file are not counted
[ini array read] can be used to count the items in the INI array.
Alternatively you can set a value in the file which explicitly tracks the
count
[ini array write] will write a Tcl array which is indexed with an integer
subscript starting at array base 0 or array base 1. The INI file will have
items written with a prefix with the key value having a suffix at array
base 1, e.g. Item1...ItemN. Sparse items in the source array are not
counted.
Example:
textcolor cyan
lib ticol_ini -verbose
textcolor
newline
set inifile test_ini.ini
ini write $inifile config testKey Hello
ini write bool $inifile config testKeyBool 1
ini read $inifile config testKey
ini read bool $inifile config testKeyBool
Results:
1
1
Hello
1
Example:
lib ticol_ini -verbose
set inifile test_ini.ini
array set a { 1 One
2 Two
3 Three
} -const
ini write array $inifile config item a
ini read array $inifile config item a
Results:
# Writes the values to the INI file
Example:
lib ticol_ini -verbose
set inifile test_ini.ini
puts [ini array write $inifile config Item a]
puts [ini array read $inifile config Item q -max 10]
Results:
1
3
See also: plugins
ticol.ini - Configuration file for Ticol.exe |
Index |
A TICOL.INI configuration file is optional and not required
This file permits custom configuration options and is particularly
relevant when using Ticol in CGI mode
The ticol.exe folder is searched first, followed by %APPDATA%\Ticol\
(Default values are shown)
Setting Description/Comment
----------------------------------------------------------------------------
[Config]
AutoExecCommand=FALSE Ticol autoexec [unknown] command (ON|OFF)
AutoExecScript=TRUE Whether to run autoexec.tcl if found
Equivalent to argument /NA if FALSE
Set this to FALSE for Server CGI mode
BackslashEscapes=TRUE C/C++ style escape characters: \n \t etc.
CGIDebug=FALSE Send an HTTP 1.1 header to stdout for CGI
CommandUndefList=<list> Space-separated list of commands to
undefine on startup
ConsoleEcho=FALSE Echo evaluation results to the console
Equivalent to [option echo on|off]
Debug=FALSE Debug command (ON|OFF)
EnablePreProcessor=TRUE /NP (No PreProcessor) command line arg
FlowControlUsesExpression=TRUE /EXPREVAL and Option expression ON|OFF
ForInitAllowsExpression=FALSE; [for] {init} statement is controlled by
FlowControlUsesExpression v's [eval]
HashTableIncrement=0 Increment when resizing array hash tables
If set to 0 then tables will auto-resize
HashTableDefaultSize=250 Starting element size for array hash tables
HashTableResizeThreshold=2.5 Array hash table depth resize threshold
When mean depth exceeds this the array is
resized according to HashTableIncrement
HashTableMinResizeThreshold=200 Min number of array nodes before resize
MaxRecursion=<value> The maximum recursion depth (default 5000)
where <value> > 0 && <value> <= 5000
OptimiseExpressions=TRUE PreProcessor optimise [calc] (/NEO)
Pause=FALSE Pauses after execution (/PAUSE)
PreProcessorVerbose=FALSE PreProcessor verbosity
Prompt=$u@$h$g CLI prompt. 'User@Hostname>'
StandardComments=FALSE Toggles end-of-line comments (;#)
StackTrace=FALSE Toggles the error debug stack trace
Verbose=FALSE Sets /V verbose command line argument
^ It is recommended that the following
hash-table values are not altered
Notes
=======================================================================
CommandUndefList
----------------
If you undefine certain commands you may not receive a bad command
error if autoexec is set to ON since these commands (e.g. EXIT, CALL) will
be passed to Windows and exist as valid commands within Windows
The undef list may be used to make the interpreter instance 'safe'
e.g.
CommandUndefList=foo bar baz
Prompt
------
For more information see: prompt
StandardComments
----------------
For more information see: comments
Hash Table Tuning
-----------------
It is recommended that the hash table settings be left unchanged as
the default values should give best performance
See also: Ticol, Frequently asked questions, cli prompt
Known Bugs and Feature Anomalies |
Index |
Does Ticol Have any Bugs?
-------------------------
Very likely, but every effort has been made to try and chase these out
Using a language for real-world tasks usually flushes out a number of bugs
Many such bugs have been found during practical applications of the language
Project Status
--------------
Builds of v1.27+ from 2022 onwards are now in the beta stage of development
with only a few features left unfinished.
There are some features which I am less than happy with, such as the
overall handling of escape characters, the problem of double-unescaping
and the workaround of using the [escape] and [unescape] commands.
Hopefully, this will be improved in future versions
I Found a Bug, What Do I Do?
----------------------------
First of all, check you are using the most up-to-date version
File a bug report. However, ensure you attach supporting documentation, test
or example code otherwise this will be a waste of time.
Send bug reports to: ticol@kerys.co.uk
No Unicode Support
------------------
Ticol is a Western ANSI character-set only script language. Unicode is not
natively supported. This is not a bug but by design but built-in conversion
routines are available to connect with external Unicode/wide-character DLLs
Major Issues
------------
Escape Sequence:
String escape sequences are a headache since the underlying source
code is continually reinterpreted. Constant sequences such as \n
are handled by the Macro PreProcessor but literal backslash characters
and escaped double-quotes are problematic. Workarounds for double-
escaping are to use the [escape] and [unescape] commands to control
escape handling as preferred
Quoted Strings Within Strings or Within Array Subscripts:
The parser has limited extended handing for embedded double-quoted
strings within strings or array subscripts: Double quotes are tolerated
in Ticol if escaped properly. These are not accepted in ActiveState Tcl
e.g. set q a("with spaces") is valid
e.g. set q a( "hello" ) is not valid
A workaround is to wrap the entire string in curly braces, either at
inner or outer levels
An option would be to trap for escaped strings in every call or every
call to a range of commands (e.g. [eq]). At present only command
endpoints such as [puts] or [printf] escape strings automatically.
Trapping all calls will slow down the interpreter
The following are not valid in ActiveState Tcl
set a({Hello world}) x
set a("Hello world") x
set a( "Hello\ world" ) x
set a(Hello world) x
Valid in ActiveState Tcl:
set a(Hello\ world) x
Additionally. Array subscript string are stored in their unescaped
form so the following two subscripts are different...
set q (has\ space) x
set q ("has space") x
Note that this also works!
array unset "a(Hello world)"
Long Math Keywords:
The math keywords, 'add', 'sub', 'and', 'or' have been removed
by design as detection and processing of these are time-costly to the
interpreter and they serve no useful benefit
Calldll Byte Widths
[calldll] byte widths rely on parsing a colon ":" character. This is
less than ideal and may cause issues with strings such as paths
which contain a colon. An alternate solution will be looked at in
due course
Math Overflow Handling
Overflow exception handling in string to integer conversion requires
improvement. Some conversion overflow cases are not trapped
Multi-Threading
---------------
The stable release of Ticol is not multi-threaded. A test/development
version offers rudimentary threaded execution but it not stable for
anything other than experimentation. Ticol is unlikely to be released
in a stable multi-threaded form as it would require a significant
rewrite of the core
Unfinished Features
-------------------
The following features are unfinished, feature-incomplete, have limitations
or are pending possible future implementation:
after Currently non-threaded
Implemented experimentally only
Would require a threaded rewrite
dictionary Double derefrencing not implemented
[dict] does not preserve ordering
due to underlying hash tables
vwait Experimental, see above
timezone Not implemented
array unset Glob search spec not implemented
Use [unset array -glob]
Variant arrays Partially implemented
array_create name elems ?dims?
array_set name ix ?ix? value
array_add name ix ?ix? data
fconfigure Not implemented
The following [file] commands are not yet implemented:
lreplace
Some file commands are of minimal use and can either be emulated
or called as external commands
file atime name ?time?
file attributes name ?option value option value...?
file channels ?pattern?
file copy ?-force? ?--? source ?source ...? targetDir
file copy ?-force? ?--? source target
file join
file join name ?name ...?
file link ?-linktype? linkName ?target?
file lstat name varName
file makelink
file mtime name ?time?
file nativename name
file owned name
file pathtype name
file rename ?-force? ?--? source ?source ...? targetDir
file rootname name
file separator ?name?
file system name
file tail May be implemented later
file tail name Best done as a plugin
file tempfile ?varName? ?template?
Namespaces Emulation only by name prefix
Desirable Features Which May Be Implemented
--------------------------------------------
Multi-line paste into the CLI
Rudimentary editor plugin module for the CLI
Regular expressions (via a plugin perhaps)
Features Which Are Unlikely To Be Implemented
---------------------------------------------
If the following features are required then ActiveState is recommended
Threading (other than experimentally)
Threaded [after] (other than experimentally)
Threaded [vwait] (other than experimentally)
Network socket commands
Unicode/MBCS/Multilanguage string support
Object orientation / classes (other than via simple struct emulation)
See also: FAQ, Ticol, Tcl syntax, workarounds, error
* Ticol is an interpreter not a compiler. It will perform much more slowly
than a compiled language such as C++, C#. Often around 100 times slower
You must therefore write code carefully if you need it to perform quickly.
Ticol was designed to be used as a script for non time-critical jobs such
as IT support scripting
* There is a C++ API provided for the development of external plugins which
can offer greater flexibility and speed
* You can get help on the Ticol command-line using: ticol.exe /?
* ticol.exe can be used as an MD5 generating tool using the /MD5 argument
* Use the command [option debug on] for development level debugging
* Enable [option echo on] when encountering problems and debugging
* C/C++ style backslash escape sequences are not translated from within the
command-line interface. This is to enable the CLI to be used as a Windows
shell.
You can enable backslash escape translation using:
option autoexec off
option escape on
* You can use ticol.exe as an expression evaluator for batch scripts
Use the /EXPR command-line argument (See: help /expr)
* Using logical and bitwise operators with [expr]
Bracket complex sub-expressions when using logical or bitwise operators
(&&/||, &/|)
Example:
proc is_power {x} {
expr "!(($x-1) & $x)"
}
* All data required for a proc may be declared locally within that proc
however, handling initialisation will slow down each [proc] call
* Although the expression-handler routine implements BEDMAS you are
recommended to bracket expressions fully, especially with and/or
* Avoid [expr] if maximum speed is required. Use [calc] or expand manually
Commands or bit operations may be faster than expression math
* Use [option expression off] and eval statements in flow-control for speed
Eval/commands are faster than calling the expression-processor
* Some non-standard commands may be faster than longhand Tcl commands
e.g. setat, append, cmp, eqn, concat, store
* Precalculate values wherever possible, especially outside of loops
* Use precalculated lookup tables and arrays (e.g. for Fibonacci values)
* Use conditional statements in the Macro PreProcessor rather than live
code to handle static code changes
* Optimise performance and expand Ticol by writing a __stdcall DLL
* You can call the Microsoft Visual C++ runtime library functions via
calldll*_cdecl
* Division desiring a float result must supply at least one float operand
* You can use casting operators, e.g. [double] to cast from integer in a way
which aids readability
e.g.
puts [/ [double 22] 7] => 3.142857142857143
puts [/ 22 [double 7]] => 3.142857142857143
puts [/ 22 7.0] => 3.142857142857143
* Arrays may be iterated efficiently using [array foreach]
This is quicker than translating to a list using [array get] with
[foreach], but note that this iteration is un-ordered
* When using with arrays take care to wrap multi-variable arguments in
braces
WRONG:
foreach x y [array get env] {puts $x=$y}
RIGHT:
foreach {x y} [array get env] {puts $x=$y}
* Examine the workings of an array's hashtable using [array walk name]
* Variables (including structs) can be examined using [dump varname]
* All standard variables are strings in Tcl
* Tcl lists are well-ordered strings
* An empty string in Tcl is represented as the empty list {} not ""
e.g. set q {}
* Ticol expands comparison operators to use "" as well as {}
e.g. set q {}; eq $q ""
* Reduce CPU usage in loops using [sleep] or [doevents]
* Increment numbers using either ++ var or incr var (or -- var or decr var)
* For tips on marshalling data via the [calldll] interface see: pointers
* Windows system errors can be decoded using [win_error]
* If a program appears not to enter loops or exits unexpectedly check for
option expression ON. You may need to use: option expression OFF
Eval-based flow-control logic will not work with the former
* Execution stack-tracing and breakpoint single-stepping can be useful
* Try using [calc] instead of [expr] as it may be more efficient since it
can pre-optimise at the preprocessing stage
* Long comments may be toggled as follows:
/*
Commented code which is inactive ...
# */
# /*
Commented code which is active ...
# */
* Take care commenting out 'if' statements because the short comment
character # is also a macro prefix, and '#if;' is a macro command
See also: performance, faq
No, you need to type an actual help topic!
For example:
help cli
help command line
help faq
help load
help print
help run
help tips
help troubleshooting
Topic and Picnic are English chocolate bars which became popular in
the 1960s ...
.................................
/ -.- .--. .-. . .-. /|
/ / / / /-' / / / |
/ ' `--' ' ' `-' / |
/.................................../ /
| | /
| Milk Chocolate Bar with Hazelnuts | /
| Price 3d |/
'-----------------------------------'
https://en.wikipedia.org/wiki/Topic_%28chocolate_bar%29
Try again!
See also: help help
Principle of Least Surprise (POLS) |
Index |
Principle of Least Astonishment (POLA)
Hopefully, most of the Ticol language follows the Principle of Least
Astonishment (POLA). Some experimental features have been added but
these have been designed to work as sensibly as possible and to be
well documented.
I found some of the features of standard Tcl to slightly violate
'POLA'. When considering how to improve or add new features thought
has been given to leaning on existing concepts and paradigms from either
C/C++ or BASIC in order to avoid violating POLA.
Non-decimal numbers follow common or standard conventions
0NN Decimal value with leading zero (not octal values)
0xNN Hexadecimal values
0oNN Octal values
0bNN Binary values
https://en.wikipedia.org/wiki/Principle_of_least_astonishment
See also: why write ticol
Why I Wrote Ticol Tcl? |
Index |
Why write a Tcl interpreter at all?,
Why not learn to use PowerShell, PHP, PERL etc?
I was interested in Tcl, and I wanted to experiment writing a script
language which would be able do a number of tasks attempted with batch
scripts. When I started the project I had a lot of free time outside work
and thought, (rather foolishly as it turns out) that it might only take a
couple of months to complete something usable and as a fun hobby project.
That was a few years ago. I've since been adapting the language constantly
to add new ideas and to solve new tasks. At the time, I was also looking
for a script language back end for a mail server I was writing and
thought this would fit the bill.
It was an interesting challenge expanding and transforming the original
500-line Picol source code and transforming it into something with well
over 150,000 lines of C++ code (excluding plugins). To add extra interest
I decided to write many of the low-level functionality such as hashtables
and expression-handler from the ground up rather than use public available
standard libraries. Obviously this wouldn't be recommended for a
commercial project as it would consume far too much time.
I'm not a professional C++ programmer, so in many ways Ticol will fall
quite short of some people's expectations. It doesn't have regular-
expressions, a bytecode compiler, static syntax validation, full socket
support, multi-threading, and doesn't link into standard Tcl. Some
features have been completed only to the degree that they satisfied
personal requirements.
I have no pretensions about Ticol. As an end-product it originally written
to do some fairly mundane scripting tasks I needed where I found batch
script too limiting, PowerShell often too bulky or cumbersome, and
standard Tcl too problematic for my specific requirements. Some features
may not be implemented perfectly or fully. I wanted something tiny and
truly portable with easy use of optional plugin expansion to extend the
language and quickly add new features as required
Drawbacks
---------
Ticol has drawbacks, the big one is that, unlike recent versions of
standard Tcl, it doesn't handle Unicode, but I really didn't need it and
also didn't want to have to rewrite dozens of underlying C++ libraries
I had which were available for use. Had I had to rewrite these for
Unicode/MBCS support then the project would never have been realised.
Also, Ticol does not perform in-depth code syntax analysis before
execution so scripts need to be written and tested very carefully.
Perhaps someone will write a static code-analyser in Tcl?
For any of my own special requirements I can quickly write a DLL in C++
to perform a more specific task and then use Ticol as a wrapper or "glue"
make the DLL useful. I think this is what John Ousterhout intended
Writing a custom plugin DLL for a specific job isn't particularly
difficult but it does require following a set of rules via an API
There are some features in Ticol Tcl which may be transient and which,
being less than satisfactory, may be removed or modified.
ActiveState Tcl falls back to big-integers when numbers exceed 64-bit
integer range. Ticol does not. ActiveState Tcl has full socket and
multi-threading support as well as regex and Unicode support.
Real World Uses
---------------
What have I used it for?
A number of small but moderately complex automation tasks, including
a script which ran a detailed 24/7 LAN diagnosis console with
recommendations of course of action both as a console and web CGI
application. This also queried a Safenet HASP licence server and emailed
notice of impending licence expiries and performed a number of log
filtering and housekeeping tasks, combined with batch scripts. It also
provided real time discovery of network HASP licence keys and was used for
a few years as part of an RLM licence expiry check and automated licence
renewal system. It filled in the gap between Windows batch scripts and
PowerShell without needing dotnet dependencies to be dealt with.
In more recent years Ticol has been used extensively to generate
historical replica railway signalling simulator and as a back-end
script language to animate a physical replica illuminated diagram.
The HTML and RTF versions of the Ticol online man (help) file are also
exported using Ticol, complete with syntax highlighting provided by the
ticol_highlight.dll plugin I also have Ticol working nicely as a
substitute for PHP for simple Web server CGI back-end scripts. In fact
most of the active work performed by the language has been CGI based
running from a portable Apache webserver.
Alternatives
------------
There are certainly far better small Tcl and similar languages out there
such as "Jim", "Little" or "Ch C++" (and I really like Ch)
This discussion about why Tcl is misunderstood may be useful
http://antirez.com/articoli/tclmisunderstood.html
If you don't like Ticol and don't find it very useful then perhaps you
would like to try one of these excellent alternatives instead?:
http://www.little-lang.org
http://jim.tcl.tk/index.html/doc/www/www/index.html
https://www.softintegration.com
Why Not Use PowerShell?
-----------------------
Why not use Tcl if it does the particular job you require?
I disliked having to use PowerShell. It seemed unwieldy, over-complex,
messy, fragile and far too system-dependent or system-interwoven. For
many tasks it was overkill and ill-suited to tasks which were beyond
the scope of batch scripts but still not terribly complex, I also needed
to be able to apply some standard admin scripts to a wide variety of
Windows systems with varying versions of the underling dotnet runtime
and often without PS installed. A prime motivator was continually
hitting the frustrating limits of Windows batch scripting, which was often
almost (but not quite) good enough to do the job.
There were occasions where PowerShell was the perfect tool for the job
such as 365 or mail server admin and where I had full ownership of
the system in order to ensure PS was fully updated but I always had an
eye on keeping things as simple as possible, at least system configuration
wise.
Some of these systems would have been software or engineering team test
machines and with a fixed/stable environment which I couldn't tamper with
and where a truly portable script interpreter was needed which would have
zero systemic modification demands on the host.
On most of the systems the underlying dotnet configuration could not
be altered by system admins as these test hosts required specific
dotnet configurations. This placed restrictions on PowerShell usage.
Batch scripts would have been ideal as they had the largest flexibility
and commonality across all O/S from XP to Server 2012 and with fewest
additional dependencies but batch scripts very quickly become unwieldy
for even moderately complex tasks. Ultimate real world solutions were
often a mixture of batch script, PowerShell, VBScript and Tcl.
Another issue was that while I had time at home to work on a hobby project,
free time in the workplace was a luxury and pragmatism was usually the
order of the day. Re-use code, keep things simple and easy to maintain
I found it less time-wasting to fix a bug or add a new feature to Ticol
than spend ages trying to get a PS script working properly.
Moreover, it was fun writing something to do a specific task even if the
tasks, relatively speaking, were fairly trivial.
These articles may be interesting:
https://tcl.apache.org/html/why.html
https://arstechnica.com/civis/viewtopic.php?f=17&t=154990
https://en.wikipedia.org/wiki/Principle_of_least_astonishment
https://outsourcedguru.wordpress.com/2016/02/22/why-powershell-sucks-so-badly/comment-page-1/
See also: faq, licence
FAQ - Frequently Asked Questions |
Index |
In this section the most anticipated questions about Ticol Tcl are covered
About Ticol
Arrays
Bugs
Classes
Configuration
Debugging
DLL Files and Linking to External DLLs
Documentation
Editor, CLI and IDE
Efficiency and Performance
Errors
File and I/O Issues
Flow Control
Graphics and Windows GUI
International
Language and Syntax
Licence and Copyright
Linked lists
Macro Language
Math
Plugins
Procedures
Security, Script Tampering and Intellectual Property
Strings
Structs
Testing
Threading
Tricks
Useful References
Variants
Windows Console
See also: Some Common Mistakes
About Ticol
-----------
Q) What is Ticol? What is its intended purpose?
A) Ticol is a personal project to write a compact and portable Tcl
interpreter for general workplace IT administrative script and hobby use
It has not been specifically designed for public use, but it is offered
for use by anyone who finds it useful. Some may find shortcomings, such
as the lack of Unicode support a problem, but it should be understood
that the author does not require this feature, so this and other
unnecessary features will NOT be added. Ticol makes extensive use of the
author's own C++ libraries developed over many years of hobby C/C++
programming
Ticol is not specifically recommended for use in a high-risk production
environment without extensive suitability testing
Q) How do you pronounce the name?
A) Ticol == "tickle". This is a word-play on the original source project
and the fact that the Tcl icon is a feather. The name derives from the
original source project - 'Picol'
Q) Why write a Tcl interpreter, why not learn to use PowerShell, PERL etc?
A) See: why
Q) In what major ways does Ticol differ from standard Tcl?
A) Ticol is designed to be Windows-specific and is not intended to ever be
a cross-platform language. Additionally, as this is a personal project it
has been written to be ANSI-specific and primarily for use on legacy
platforms such as Windows NT, 2000 and XP. Compatibility is, however,
maintained for the most recent versions of Windows on 32 and 64-bit
platforms
Q) Does Ticol support Unicode
A) No. It is ANSI-specific. If you'd like a Unicode version then I invite
you to get coding and write your own Tcl version
Q) I really don't like Ticol. What shall I do?
A) Either use another flavour of Tcl or have a go at writing your own using
the original and excellent 500-line Picol source as a starting point
I recommend ActiveState Tcl
Q) Can I contribute to Ticol?
A) Yes, you could make a small donation to help with equipment costs
A) Good (well-detailed) bug reports are welcomed. Example scripts should
be included if possible
Q) Are there any alternatives to Ticol?
A) Yes, the "Jim" programming language is pretty cool
http://jim.tcl.tk
...and ActiveState Tcl is excellent!
Q) Does Ticol include a full socket library?
A) No. Although I wrote the SMTP libraries I'm not a particularly expert
socket programmer and didn't really need much in the way of TCP/IP
sockets so haven't added general socket functionality other than HTTP
retrieval. It would be feasible to write a socket command plugin
though
Q) What is Tcl anyway and who uses it?
A) Tcl is a simple and usually interpreted language with some visual
similarities to "C"
It is used by Cisco(TM) within routers, by ANSYS, Boeing, NIST and
many others:
http://wiki.tcl.tk/37389
Q) Isn't Tcl just "toy" language?
A) No, see this thread:
http://wiki.tcl.tk/3970
Q) Does Ticol Tcl perform bytecode compilation
A) No. It is strictly an runtime interpreter although the Macro Pre-
Processor will perform some optimisations to increase speed
Q) What order does Tcl perform evaluation in?
A) Evaluation for square brackets is performed in the reverse order to that
which might be expected. For example, the statement:
puts [call [return]]
Will be evaluated in the order of innermost-to-outermost brackets:
1. return (evaluated first)
2. call (evaluated 2nd)
3. puts (evaluated last)
Evaluation arguments which are not enclosed in square brackets is
evaluated in normal (left to right) order. For example:
eval {22/7.0}
Will be evaluated in the following order:
eval (last)
22/7.0 (first)
Arrays
------
Q) Does Ticol support arrays?
Q) Does Ticol support multi-dimensional arrays?
A) Basic, single-dimensional arrays are supported and true arrays of arrays
Multi-dimensional arrays e.g. $a(1,2,3) aren't directly supported in the
Tcl language, and they are typically emulated by embedding commas (or
some other separator) in concatenated array index parameters. As this is
a common Tcl "kludge" and you have to take care to avoid carelessly
introducing unquoted whitespace into the array subscript.
a(1,2)
True, arrays of arrays are also supported via nested dereferencing.
So, ${$a(1)(0)} will reference array element 1 of a(), subscript
item 0
See: help arrays
See: help arrays of arrays
Simple Variant arrays are supported for use with CallDLL* and structs
See: help varray
Q) Can fixed, multi-dimensional arrays be emulated?
A) Yes, using the ticol_carray.dll plugin. See the 2d_carray.tcl example
# Very basic example
lib ticol_carray
set dim_x 5 # A 5 x 3 matrix (addressed 0..14)
set dim_y 3
set h [carray create [* $dim_x $dim_y]] # Allocate NULL array
proc array_set {handle x y value {rvalue}} {
# [in] x and y should be array-base 0
carray set $handle [+ [* $y $::dim_x] $x] $value $rvalue
}
proc array_get {handle x y} {
# Return Lvalue only
# [in] x and y should be array-base 0
return [carray get $handle [+ [* $y $::dim_x] $x]]
}
array_set $h 1 0 "1,0"
puts "1 0 -> '[array_get $h 1 0]'"
Q) I see blank lines where comments were included in the source code when
I use the /ECHO command. Why is this?
A) If comments are indented using whitespace then that whitespace will be
retained. This is by design since Ticol cannot know if the whitespace is
significant or not (/ECHO also tries to preserve code readability)
Bugs
----
See: help bugs
Classes
-------
Q) Does Ticol support classes and object orientation
A) No, this is not part of the design
However, structs offer some operators which may be used to emulate some
aspects of class-like behaviour but is very cumbersome. Since proper
object-oriented principles aren't supported the use of class-like
behaviour in Ticol is not encouraged and true object-orientation
concepts such as data privacy, inheritance, constructors and destructors
are unsupported. Structs are permitted and these may be passed to
procedures ([proc])
See: class, ->, =>, ->>, static, args
See: struct
Configuration
-------------
Q) How can I find what version of Ticol Tcl I am running?
A) Use: puts $tcl_version
Q) Does Ticol have a configuration (INI) file?
A) Yes. ticol.ini
It also has an autoexec file (autoexec.tcl) which can also be used to
configure
See: help ticol.ini (from the Ticol CLI (Command Line Interpreter))
See: autoexec.tcl
Q) Is there a script which can be run at startup to configure standard procs
for the CLI?
A) Yes. Create a file called 'autoexec.tcl' in the same folder as ticol.exe
This is enabled by default. You can disable using the command-line using
the /NA argument or permanently within ticol.ini.
See: help autoexec.tcl
Q) Can I undefine commands automatically at startup?
A) Yes. Either by including undef commands within an autoexec.tcl file or
by using a CommandUndefList entry in TICOL.INI
Debugging
---------
Q) Does Ticol support runtime expression assertion for debugging?
A) Yes, use the [assert] command
See: assert
Q) Can [assert] be disabled globally as in C/C++?
A) Yes, use: #define ndebug
at the start of your script to disable, the same as in C++
See: assert
Q) I get no output on either a console or CGI script, why?
A) Code will be generating an error but you have [option echo off]
Use [option echo on] and if running as a CGI script then enable the
following in TICOL.INI. Always enable 'option echo on' if a script
halts unexpectedly.
[config]
Debug=True
CGIDebug=True
See also: cgi, debugging
Q) I have a recursive procedure which is failing or triggering an
exception. How can I see what is happening?
A) Use: [trace]. e.g. for a proc called 'foo', use; trace add foo
which will show the proc call trace along with any arguments
You can also use single-step debugging with [halt] and [watch]
See: trace, breakpoint, halt, watch
Q) Can I single-step or use breakpoints within Ticol?
A) Yes. enable breakpoint/debugging mode using either ticol.exe /BP,
ticol.exe /DEBUG or using the command:
option breakpoint on
You must then issue one or more [halt] instructions within your script
or at the Ticol command-line interface (CLI)
When the [halt] instruction is met you may single-step and trace the
execution stack
See: option breakpoint, breakpoints, halt
Q) Do errors show the execution stack?
Not by default. Stack-tracing may be enabled separately using the command
option stacktrace on
To provide detailed execution stack trace information should an error be
raised
See: option stacktrace, breakpoints, halt
Q) How can I trace used memory within a script?
A) Use the following example code
Note that it is normal to see a small amount of memory allocated at the
start of a script and periodically, for object allocation. However,
memory used should not show a constantly increasing trend:
proc show_mem_used {m} {
textcolor magenta
global line
puts "[++ line]: Used=[expr [mem_used]-$m] byte(s)"
textcolor
}
set start_mem [mem_used]
puts "Start memory: $start_mem"
# Do some tasks here...
# e.g. time {struct create s {f1 10}; struct unset s} 1000
show_mem_used $start_mem
Start memory: 5013504
1: Used=0 byte(s)
17 microsecond(s) per iteration
2: Used=4096 byte(s)
Q) How can I trace function calls?
A) Use:
trace add <function-name>
to add a trace item, and:
trace remove <function-name>
to remove a trace item
See: trace
Q) How can I debug a problematic Ticol CGI web script?
A) The problem with CGI scripts is that errors may result in a server "500"
error and generate no useful output.
You should look at the server's error.log and access.log files
Add this code as the first line of any CGI script:
printf "Content-type: text/html\r\n\r\n"
or:
puts "Content-type: text/html\r\n"
Run the script in console mode using the /CD (CGI Debug) argument
Place a halt command near the suspected cause and run the script from the
command-line using the /BP (Breakpoint) argument and single-step
execution. Enable the following in TICOL.INI during code development
[config]
Debug=True
CGIDebug=True
See also: cgi, debugging
Q) The /G (graph) performance trace shows 1 ms for all commands in Windows
NT, 2000 or ReactOS, why is this?
A) The necessary precise nanosecond (ns) timing functions are not available
in these versions of Windows and are thus implemented as microseconds
(ms). When timed in nanoseconds, the time for most command calls will
show as '1' for all but the slowest functions and on the slowest systems
DLL Files and Linking to External DLLs
--------------------------------------
Q) Can a DLL be loaded permanently for the duration of a script?
A) Yes, use dll_open to return a handle to the DLL and dll_close to close it
See: calldll
Q) What return types can be called?
A) Long (DWORD), string (LPSTR), double and COM Variant struct pointer
Q) Are there any examples showing the use of calldll?
A) Yes, search: help calldll examples
Q) What types of DLL linkage can calldll link to?
A) __stdcall and __cdecl only. Other linkage styles are unsupported
Q) Is linking to COM interface DLLs supported?
A) No, COM DLLs are not supported and there are no plans to add this
If you want this you will have to write a COM wrapper
Q) How does [calldll] load DLLs?
A) DLLs can be loaded for the duration of the script using
dll_open/dll_close or they can be loaded/unloaded on a per-call basis
Note that the Windows IPHLPAPI.DLL will leak memory if loaded 'per call'
this is a known Microsoft DLL problem
See: help calldll
Q) Are plugin DLLs supported which can add new commands?
A) Yes. Using the [lib] command and if the plugin programming guidelines
are adhered to. Lib will load exported commands in the interpreter's
namespace. By default all are loaded by this can also be filtered
Documentation
-------------
Q) Is there a manual for Ticol?
A) Yes, this is it. Also available in Windows Help (CHM) and HTML format
Although Ticol has additional commands, the underlying language is the
same as for standard Tcl. It is recommended that you search online and
visit the site at
http://wiki.tcl.tk
You should also read LICENCE.TXT and any accompanying README.TXT
Editor, CLI and IDE
-------------------
Q) How can I load a file into the interpreter?
A) From the ticol.exe CLI, use the command: load filename
A) From the command-line, you can run using: ticol.exe filename
Note that in both cases the '.tcl' filetype is optional
If you are loading/running a TCX file then the filetype must be
specified
Q) How can I tell what file is loaded?
A) It is displayed on the console Window title bar if loaded
Q) Can I view a loaded script from the CLI?
A) Yes, use the [dump] command to show the script with syntax
highlighting
See: dump
Q) What is the maximum command-line length for the CLI?
A) 4096 characters (ANSI/ASCII text only)
Q) What is the maximum file size which can be loaded/evaluated?
A) This is limited only by memory but very large files (> 1mb) may be
unwieldy and slow to use. Consider breaking up into smaller files
and calling these using [eval]
Q) What editor is recommended for Ticol?
A) Notepad++ is highly-recommended and is used by the author
Website:
https://notepad-plus-plus.org/
See: help editor
This topic also explains how to input the Ticol keywords
Q) Some commands give no output even under error conditions. Why?
A) Unlike ActiveState Tcl the CLI echo is OFF by default
You may use [try/catch] to intercept and handle error conditions
Standard Tcl echoes command results of all commands by default
Thus a command issued from the CLI would echo a result twice
Within scripts you may wish to enable: 'option echo on' if you don't
intend to use [puts] or [printf] to display command results
Use the command: option echo on :to enable this feature
See: help option
See: puts, printf
See: try, catch
Q) Some commands issue output twice. Why?
A) When running in the CLI you will see an additional results echo if you
have 'option echo on'.
You may disable this using 'option echo off'
See: help option
Efficiency and Performance
--------------------------
Q) Do you have any tips for writing big programs?
A) Yes,
See: big programs, efficiency
Q) Is there any way to analyse script performance?
A) Yes. Run ticol.exe with a script and the /G (Graph) argument to generate
a performance analysis graph. You can then discover which commands are
called the most and substitute more efficient commands or expressions
Example: ticol myscript.tcl /g
See: performance, graph
Q) How can I improve performance
A) Write a plugin DLL LIB module which uses C++ code to perform time-
critical calculations. The C++ interface is quite simple and the
C++ source code for the interface can be made available.
See: performance
Q) Are there any commands which can enhance performance?
A) Yes, several command are supplied with Ticol which implement common
routines which might be implemented longhand in Tcl. For example [is_mod]
offers a rapid way to test modular division
See: performance
Errors
------
Q) I get a "missing operand" error in [expr], What does this mean?
A) Ticol behaves the same as ActiveState Tcl when handling null (empty)
"" strings in that empty strings don't present themselves as arguments
to [expr] and some other commands. If a variable is set to "" and then
evaluated then no argument can be passed and the expression handler
will see a missing argument or "operand"
Example error using ActiveState Tcl with 'a' set to an empty string:
% set a ""
% puts [expr $a eq "hello"]
missing operand at _@_
in expression " _@_eq hello"
A workaround is to wrap any strings likely to be "" in braces as {""} or
to avoid using string operations within expressions and to use commands
instead
% set a ""
% puts [expr {$a} eq "{}"]
1
% set a ""
% puts [expr {$a} ne "hello"]
1
Q) A [goto_block] doesn't execute at all
Q) I get an unexpected error relating to command syntax in a [goto_block]
A) Check that any initialisation commands at the head of the [goto_block]
are wrapped within a dummy [goto_block] entry
Example:
goto_block { # WRONG
set a 1 # Can't initialise outside a defined block
set b 2
one { # First goto label ...
goto_block { # RIGHT
0 { # Dummy block correctly used as an initialiser
set a 1
set b 2
}
one { # First goto label ...
Q) I have a loop which hangs and doesn't exit. I can't see anything wrong
with the expression argument. Why is this?
A) First, check for the correct [option expression] statement which
matches your coding style. If this setting does not match you could end
up with infinite loops since Ticol will try to evaluate a constant value
Example: # Will evaluate only the first argument
option expression off
set i 0
while {$i < 10} {... # Never executes ($i evaluated and is FALSE)
option expression off
set i 1
while {$i < 10} {... # Loops indefinitely ($i evaluated and is TRUE)
Correct:
option expression on
set i 0
while {$i < 10} {... # Works OK ([expr] used to evaluate [<])
option expression off
set i 0
while {< $i 10} {... # Works OK ([eval] used to evaluate [<])
See: option expression, flow control, expr, eval
Q) I have a statement (e.g. [switch]) which has the correct argument count
but I still get an error message relating to invalid syntax, why?
A) Check for arguments which are uninitialised (empty) variables or which
have not been declared in advance. If auto-declared as arguments they
will be initialised to {} (the empty list/empty string)
Q) [for], [if] or other statements don't seem to work. Why is this?
A) Check your expression types. If you have used 'prefix expressions' then
you cannot use 'option expression on'. Either wrap the prefix expression
in square brackets within the flow-control statement or use
option expression off
Ensure you use [puts], [printf] or [option echo on] to see output
Q) I get "foreach requires matching pairs of lvalue and rvalue (3 found)"
with the following statement
foreach x y [array get env] { puts $x=$y }
A) You need to wrap the values x and y in braces as follows
foreach {x y} [array get env] { puts $x=$y }
Q) I get a 'procedure already defined' error when running a script
A) Either you have defined a procedure with the same name twice within your
Tcl script or you are trying to restart and re-run a Tcl script which has
halted. You may undefine unwanted commands or procs by using the 'rename'
command. It is recommended that proc definitions are wrapped with a
'defined' statement
See: help defined
Q) I get an error '<expression>' is not recognized as an internal or
external command. Why?
A) You have a flow-control (if, for, switch etc.) expression which is in
expression format rather than eval format. Expression format is 'infix'
while eval format requires a leading command name. Hence the lvalue will
be evaluated as a command with 'option expression off'. To resolve the
problem change the configuration in TICOL.INI or set
'option expression on'
A) You have produced a result which is, somehow being called as a command
See: help option
Q) Strings with C-like escape sequences don't work from the CLI
A) CLI interaction with Windows requires C-escapes to be disabled
You can enable escapes within strings if you set the following options:
option autoexec off
option escapes on
Strings with escape sequences such as \n will work fine other than from
the CLI interface
See: help option
Q) A script halts with no error message, why?
A) Use: [option echo on] to ensure error information is shown
A) Wrap procedures and commands in [catch] statements and trap errors
A) Use a [halt] statement before the offending code and single-step using
ticol.exe scriptname /bp
Q) When I run a script from the CLI sometimes I get extra output displayed
when the script finishes
A) When running from the CLI the results of the script are displayed
This is suppressed when running from the Windows command-line
Q) A macro variable isn't getting replaced, why?
A) Macro variables aren't replaced within quoted strings
Check that the variable is outside a quoted string
File and I/O Issues
-------------------
Q) How can I print to an open file?
A) Use [puts] with the file handle of an open file as the first paramter
set fp [open filename wb]
puts $fp "Hello\r\n" -nonewline
close $fp
Q) How can I print to a file using formatted [printf] style output?
A) Combine [format] with [puts] as follows
set fp [open filename wb]
puts $fp [format "%s %i %f\r\n" One 2 3.0]
close $fp
Flow Control
------------
Q) How can I exit or 'return' from a child script which has been spawned
by a parent e.g. using [source script] ?
A) The spawned child script may be exited prematurely by [return -code exit]
[return] or [return -code ok] at root level will not cause a child script
to exit and return to the parent
Graphics and Windows GUI
------------------------
Q) Is Tk graphics capability supported?
A) No, Ticol is designed for use as an administrative (ANSI) console script
language there is no Tk or GUI support in Ticol other than [msgbox],
[inputbox], [get_fileopen] and [get_filesave]. You may, however, call
external DLLs or the Windows API directly via the calldll plugin
International
-------------
Q) Does Ticol support non-US/UK or international languages?
A) No. It is an ANSI only language. However external libraries can be called
such as AutoIT and these can handle Unicode strings.
You are recommended to use Active-state Tcl or some other Unicode
alternative version of Tcl
Q) When will Unicode/International capability be added to Ticol?
A) Never. This is a program primarily for my own use which I have chosen to
share with anyone who finds it useful
Language and Syntax
-------------------
Q) Can nested [||] or [&&] statements be made easier to read?
A) Yes. Avoid the need for complex, nested and bracketed logic statements
by using either [expr], [all] or [any]
if {|| $a [|| $b [|| $c $d]]} ... # Nested logical or
if {any $a $b $c $d} ... # Inline logical or
option expression on
if {$a || $b || $c || $d} ... # Inline logical or
if {&& $a [&& $b [&& $c $d]]} ... # Nested logical and
if {all $a $b $c $d} ... # Inline logical and
option expression on
if {$a && $b && $c && $d} ... # Inline logical and
See: help any, all, help expr
Q) Does Tcl perform 'lazy evaluation' on [&&] expressions?
A) Ticol cannot perform lazy evaluation. Tcl always evaluates every []
statement within an expression. Thus expressions which contain invalid
statements within some context will always be executed
e.g.
if {&& [valid command ...] [invalid command ...]} { ...
will execute [invalid command ...]
Q) Tcl is alleged to be "typeless" how come you refer to numeric types?
A) Tcl is *internally typeless* (generally speaking). The values represented
at source code level are not typeless. There is a category difference
between internal representation and non-internal presentation of data
Q) Is backslash handling in Tcl a nightmare?
A) Yes. See: backslash
Q) [day_of_week] returns unexpected results for dates before 1583. Why?
A) To keep code simple, the Gregorian system is used for all dates and
in lines with changeover rules for 1583 which are commonly used
The actual dates used around the world and over time are varied and
quite complex. The command does not use the Julian calendar system
Q) Can a variable name contain a space?
A) Yes. Either enclose double-quotes or follow normal backslash
escaping rules. However, variable names with spaces are not recommended
Q) Can a variable name be one or more spaces as per standard Tcl?
A) Yes. To assign you must wrap the space in curly-braces, thus:
set { } Hello
puts ${ }
Q) Can I do lambda or 'lambda-like' anonymous functions/procs?
A) Yes
See: apply, lmap and lambda
Q) Are const enums supported?
A) Yes
See: enum
Q) How do I pass a reference to variable up a hierarchy of [procs] ?
A) Pass by variable name from each level
You must ensure that the inbound var in each proc is dereferenced using
$ and that this dereferenced var is set using [upvar]
# root->[foo]->[bar]->[baz]
proc baz x {
upvar $x
puts "Endpoint in baz $x is $$x (Level [info level])"
}
proc bar x {
upvar $x
baz $x
}
proc foo x {
upvar $x
bar $x
}
set s Hello
foo s
Q) Are there differences between Ticol and Standard Tcl?
A) Yes, many
Ticol was developed to implement a subset of standard Tcl with
a few design differences and a few features I wanted to experiment with
Tcl doesn't have much in the way of syntax, it is command-based
An early decision was made not to support Unicode and to be ANSI-only as
the author had no use for Unicode support and had a large selection of
existing ANSI C++ library code
See: help differences
Q) How does Ticol handle special dereferencing cases (e.g. double)?
A) Ticol Tcl supports multi-level dereferencing either using [set] or
PERL-style multiple $ sign prefixes i.e. $$varname
See: double dereference
Q) How can I set the following variable which embeds whitespace?
set q a("Hello world")
A) You cannot embed a double-quoted string within an unquoted string
You can, however, embed escaped double-quotes within a double-quoted
string as follows:
set q "a(\"Hello world\")"
Note that this will create an "escaped" string which is not unescaped
by the macro-preprocessor (MPP). You can unescape this as follows:
puts [unescape $q]
Q) Are there any common coding mistakes which can be made when moving to Tcl
from C/C++ ?
A) Yes. A few quite common ones
See: help mistakes
Q) What standards have been applied when developing Ticol?
A) Every attempt has been made (given limited time and resources) to
implement a subset of the Tcl language as documented at
http://www.tcl.tk/
As this is a hobby program development will be ongoing but may be
sporadic, intermittent and often experimental
Q) I get a syntax error when executing a script with an if statement?
A) Check that the first brace of the if' body statement is on the same line
as the if statement itself. If not then you will get an error. The Tcl
syntax requires the opening brace to come before any CRLF after the
opening if command
RIGHT:
if { conditional-test } { ... # Correct brace at and of line
WRONG:
if { conditional-test } # Misplaced brace
{ ...
Check if you have flow-control commands set to use expr or eval. Check
the setting for 'option expression'
See: help if
See: help option
Q) Does Ticol support infix flow-control conditional tests?
A) Yes, both infix and prefix operator tests are supported and these apply
to if, while and for flow-control constructs.
To enable infix expression logic. Use the command:
option expression on
To enable Tcl prefix expression logic. Use the command:
option expression off
See: help option expression
Q) How can I comment-out a block of code?
Q) Are long-comments available similar to C/C++ ?
A) Use C/C++ style /* ... */ long comments
A) Use a [goto_block] with a [goto] 'skip' command
A) Use if {0} { ... }
A) Use #ifdef with a macro-variable which has not yet been set
#ifdef comment
... Some Tcl code or non-Tcl content to be ignored ...
... this may include syntactically invalid code
... all content here is removed before execution
#endif
To enable the block later, #define comment before the #ifdef statement
See: help #ifdef, help goto, help comment
Q) Are short C++ style comments e.g. '//' available?
A) No, but you can emulate them as follows:
proc // args {
# C++ style comment - do nothing. Discard all args on this line
}
Q) Is the content of a Ticol comment syntactically-significant?
A) No. Since Ticol uses a Macro PreProcessor comments are stripped out
before syntax is analysed
Q) How can I include a literal hash (#) character in source code as this is
usually treated as a comment prefix?
A) A hash character may be used in quoted strings
A) Use the escape sequence \#
Translation and removal of comments is performed by the Macro
PreProcessor (MPP) but the MPP will recognise and accept escaped hash
symbols
See: escapes, Macro PreProcessor
Q) How can I iterate an array which has a non-contiguous subscript
A) You may use [array foreach] or [foreach] as follows
foreach {x y} [array get env] { puts $x=$ }
Q) Are short (one-line) comments allowed at the end of a line?
Q) What is the short-comment syntax?
A) Yes. Unlike standard Tcl, one-line comments are allowed anywhere on a
line. The remaining line from the first # character onwards will be
ignored. For this reason the '#' character is replaced by '@' in
the [upvar] and [uplevel] commands.
A prefixing/separating semi-colon is not mandatory (i.e. ' ;# comment')
A comment start # symbol is valid only outside of quoted strings
See: help #ifdef, help #define
Q) Do comments or #ifdef..#endif commenting affect or slow down execution?
A) No. The PreProcessor will strip out all macro and comment lines before
they are evaluated. This feature is also used to reduce the size of
obfuscated TCX files produced using: ticol.exe filename /C
Q) Can I have nested #ifdef..#endif blocks?
A) No. These are not supported at present but planned for the future.
Ticol macro support is quite basic and doesn't extend to nested
#if..#endif constructs yet
See: help macro
Q) Does the Ticol MPP support complex macro expressions?
A) Yes, via the #if statement which can take Tcl expressions but these
may be costly to process since each #if loads the script interpreter
See: macro if
Q) Can I verify or see the preprocessed code content?
A) Yes. If the code has not been encrypted/obfuscated you may use the
/ECHO command-line argument to echo the resulting code to the console
This may be redirected to a file using the standard > operator
Note that code which has been obfuscated cannot be viewed
See: debugging
Q) How can I add the source code line-number into the program?
A) This can be done via macro variables
The macro variable #__LINE__ will echo the source code line
You can assign or display as follows:
set line #__LINE__
puts "Line: $line"
or
puts "Line: "#__LINE__
See: help macro
Q) Do Ticol structs support instantiation?
A) Yes. Basic instantiation is supported
See: struct, new
Q) Does Ticol support a [goto] command
A) Yes. But only as part of a localised [goto_block] command
The use of [goto] is not recommended and is provided for entertainment
purposes only
See: goto_block
Q) How can I call [asc] to get the value of a space?
A) Use [asc " "] or, if the call is embedded in a string use [asc { }]
A) Use [asc [chr 32]] or [asc [chr 0x20]]
Q) How do I go about freeing Windows allocated memory returned from calldll?
A) Pass the return value to a Tcl variable and call [sysfree] or pass via
another calldll call to free the block
Windows memory allocation will require the ticol_calldll.dll plugin
See: calldll, sysfree
Q) How to I zeroise a buffer before passing to the Windows API using
calldll?
A) Allocate the buffer using [makestr] then use [memset] to set the buffer
Memset will clear memory only up to the allocated length of the string
[makestr] and [memset] are intended only for DLL interface use only
Windows memory allocation will require the ticol_calldll.dll plugin
See: makestr, memset, calldll
Q) Are nested [proc] blocks permitted?
A) Yes. However procs are not evaluated or tested for errors until they are
first called at runtime which may make producing error-free code
difficult and debugging problematic
See: proc, upvar, uplevel
Q) What Abstract Data Types does Ticol support natively?
A) Standard Tcl strings. Since "everything is a string" in standard Tcl this
includes string, numeric, dictionary and list data. Ticol additionally
supports binary structs for use with [calldll] only as well as standard
Tcl arrays and stacks. Arrays are implemented as hashtables
A plugin is also available which offers standard linked-lists
See:linkedlist
See also: array, list, stack, linkedlist
Q) How can I save/restore [option] states when calling external scripts?
A) Use the 'push' and 'pop' options
A) Use the following:
Save:
set expression_state [option expression]
option expression off
set echostate [option echo]
Restore:
option echo $echostate
option expression $expression_state
Licence and Copyright
---------------------
Q) Is Ticol free software? (Freeware)
A) Yes. It is free for both personal and business use, but it may not be
resold. You agree to use it at your own liability with no warranty
whatsoever and indemnify the author against any claims of liability
(See the attached licence)
Q) Is Ticol recommended for business-critical applications?
A) No, Ticol is under development and should not be relied on for business-
critical applications. Businesses may use it for evaluation and testing
as long as no liabilities are created
Q) Is Ticol open source, can I have the source code?
A) No and no
Ticol is a derivative work of a BSD project which permits commercial
/closed-source derivations to be created
The original Picol source code is freely-available but requires
some work to fix memory leaks and expand functionality. If you want the
original source code it can be easily found using a search engine if
you search for 'picol'
Q) Awwww... this isn't fair, I *want* the source code, I'm going to tell
my mum|dad|teacher|brother|etc..
A) See other (and far better) open-source Tcl/Tcl-like projects instead
where source is freely available (and of far better quality)
Q) What is the purpose of these licence terms?
A) To avoid unreasonable demands, for support, annoyance, or legal
liabilities being placed on the author
Q) What licence is Ticol supplied with?
A) See: help copyright and
http://www.linfo.org/bsdlicense.html
Q) My virus scanner reported an infection. Why is this?
A) Ticol contains no intentionally-malicious code. All primary binary
distributions will have been virus-scanned before release using a
commercial virus scanner such as Kaspersky A/V
However, the binary could have become infected during redistribution by
a 3rd party or your virus-scanner may be reporting a false-positive.
Please cross-check using several other virus-scanners. If in doubt,
discard the binary distribution you have and obtain a bone-fide copy
Ticol should NEVER be bundled with a binary 'wrapper'/installer program
other than as a clean ZIP or similar archive
Q) Is Ticol guaranteed or warranted?
Q) Ticol trashed my data, my wife divorced me, boss hates me etc. wrecked
my programming skills, can I sue?
A) No. No warranty of any kind, either express or implied, is offered, and
by using it you agree to indemnify and hold the author free from any and
all liabilities either direct or indirect arising from its use
Every reasonable effort has been made to ensure Ticol is stable, given
that this is purely a hobby-project and not a commercial one
Q) Can I sell copies of Ticol?
A) No. Resale is prohibited. If you are reselling it you are in breach of
licence terms and must immediately cease and desist
Q) Why is resale prohibited?
A) The author does not wish to accept contractual liabilities for
support or to warranty of application in any way
Q) I paid for a copy of Ticol from someone other than the author
Shouldn't you to provide me free support?
A) No. And you should not have been charged for a copy other than the cost
of media (CDROM or USB stick)
And you should ask for your money back
Q) Can I force you to provide support or a warranty?
A) This software is offered gratis and 'as-is' under the licence terms and
clearly stated to be at used entirely at your own risk
Q) Can I redistribute Ticol from my website?
A) Generally, no. But contact me as admin @ the program's documented
web-domain and I will get back to you and work something out if and
only if this involves no liabilities, risk or annoyance to the author
Q) Is the documentation included with the main software licence
A) No. The extensive manual (over 900 pages if printed) is covered by
normal international copyright for written works, but it may be
supplied free of charge with the binaries
Q) What if I don't like the licence terms?
Q) What if I dislike the Ticol program?
Q) What if I refuse to abide by the licence terms?
A) Stop using this software immediately.
You are no longer covered by the licence terms
Find another product. There are many excellent alternatives
Q) I'd like to negotiate a different licence term, is this possible
A) Contact me and I might discuss
Linked Lists
------------
Q) What type of lists are available?
A) Standard Tcl lists with most key features implemented. Such lists are
no more than formally-organised (well-defined) strings
Traditional linked-lists are offered via the ticol_list.dll plugin
See: linked list
Macro Language
--------------
Q) Is there an #include command for the Macro PreProcessor language?
A) No. It is not required. Use [source scriptname] to include a script
at runtime
Math
----
Q) How can I use numbers with leading zeroes?
A) This is problem for standard Tcl not for Ticol Tcl which doesn't
interpret numbers with leading zeroes as octal. Octal values require
zero followed by the lowercase letter 'o' - i.e. 0oNN
..code
puts [expr 0123*2] # Returns '246'
..end
Q) How can I force a variable to a "float" (double) so it evaluates
correctly in an expression?
A) Cast to double using the [double] command
Example:
set a 22
set b 7
puts [/ $a [double $b]]
Result:
3.142857142857143
A) Wrap the variable name in braces and append ".0"
Example:
set a 22
set b 7
puts [/ $a ${b}.0]
Result:
3.142857142857143
Q) Are "scientific" format numbers such as 4e23 or 2e64 supported?
A) No, other than via the [atof] and [ftoa] commands and via
the ticol_sci.dll plugin
This may be added to ticol_bignum.dll plugin and, natively, later on
See: atof, ftoa
Q) What are the number ranges used by Ticol?
A) Integers are manipulated as 64 bit values with a signed range
64 bit signed int 9,223,372,036,854,775,808 to 9,223,372,036,854,775,807
64 bit unsigned int 10 to 8,446,744,073,709,551,615
The default integer type is signed. Numbers may be coerced to unsigned
using the [unsigned] cast command
Larger integers (up to 6000 digits) may be used via the big math plugin
ticol_bignum.dll
Real numbers are detected automatically by the presence of a decimal
point value. Coercion can be forced by either appending "/0" or using the
[float] command
Floating point values are in the following range:
double 64 bit signed 1.7E +/- 308 (15 digits of precision)
See: data type ranges, atof, ftoa
Q) Do I have to bracket math expression when using the [expr] command?
A) No. Operator prioritisation is assigned automatically using standard
BEDMAS rules with remaining operators (e.g. +, - being evaluated Left To
Right expressions are dynamically rewritten to prioritise * and /
operators. However, expression within source code may be easier to read
and maintain if properly bracketed
See: help expression
Q) Complex expression results produce fractionally different results to
those of cscript or other interpreted languages such as Visual BASIC
Why?
A) Tests show that some expressions in cscript or wscript (and VB) are
rounded ups to the nearest integer. This may be due to these languages
using a 'float' value for calculation. Ticol uses 'double' for all non-
integer calculations and 64-bit integers for integer calculations and
does not automatically promote integers to reals. You can override this
within expressions by using number-casting functions e.g. double()
Comparing float/double values is a well-known programming problem
See: help expression
Q) Are there any mathematical accuracy issues with Ticol?
A) Ticol will manifest any or all of the underlying accuracy issues present
in the MSVC runtime etc. when using large floating point values, but the
language is not intended for scientific precision but as a script
language. In particular, very large 64-bit values may require the use of
an [unsigned] cast
See: help math, big math
Q) When applying a math operator such as * or / to mixed integer and real
numbers will the result be an integer?
A) If one or more operands of an operator expression are real numbers
then this operator expression will be evaluated as a 'double'
This is in the design specifications for standard Tcl
See: help math
Q) I am calling [expr] with a complex math expression and it either returns
zero or a wholly inaccurate result. Why?
A) Tcl, by language definition interprets all numeric values automatically.
There is no concept of fixed typing but various means of coercing are
provided by Ticol including non-standard type-cast commands
This may cause problems with expressions which involve operations which
expect floating point (real) numbers such as division. Real numbers are
interpreted as so by the presence of a decimal point. Otherwise, integer
type will be assumed
Example:
expr "22/7" returns 3 but 22/7.0 will return 3.142857142857143
expr 123/(0.123*10)
Q) Very large numbers come out negative. Is there a fix for this?
A) Ticol uses signed 64-bit integers for all integer calculations. As this
is a script language high-speed performance is not expected and is thus
less of an issue. You can coerce these values to be unsigned using the
[unsigned] cast command or expression function
Preferably use the ticol_bignum.dll big math library which offers the
[big] command and supports up to 4000+ digit numbers
See: help unsigned, big math
Q) puts [expr "not 0"] does not work, why?
A) The 'not' command is not implemented within [expr]
Likewise for 'and', 'or', 'xor' long-form commands
Use the '!' operator
puts [expr "!0"] # Outputs 1
Memory
------
Q) Does Ticol support heap memory allocation?
A) Yes, using standard "C" malloc, free, memcpy and memset
These are provides to help interface with external DLLs via [calldll]
and for use with structs and not for regular script use
Extreme care should be taken with the used of these commands as
careless use will produce unstable code and may also make the
interpreter unstable
These commands were added only for the interest and amusement of the
author. If you object to these features, don't use them
See: malloc, free, memcpy, memset
Plugins
-------
Q) Is it possible to take a screenshot of a image or photograph of a bmp-
encoded file and recreate the payload contents from it?
A) Yes The BMP plugin uses the bitmap image as a simple data-container
and does not follow byte-padding rules but if you follow these steps
you should be able to reproduce
1 Paste the image into a paint editor
2 Resize to the correct bit width. (Rows must be a multiple of 4)
3 Decrease colour depth to 2 colours (the palette is ignored)
4 You may need to produce a negative image
5 Save as a Windows BMP image
Procedures
----------
Q) How can I share procedures between multiple tcl applications?
A) Place the desired procedure definitions in a TCL or TCX include file
Be sure to control definition using if {[defined procname]} or
forcefully undefine using undef procname -nocomplain before definition
Q) Can I find out the name of the procedure I am 'in' at runtime?
A) Yes, use [info procname]
proc foo {} {
puts "Currently executing proc \[[info procname]\]"
...
Q) How do I get the list of arguments or other information inside a proc?
A) Use the following:
[info args procname] to get a list of the proc's arguments
[info body procname] to get the script text for the proc
Q) What are valid procedure names?
A) A procedure may consist of the upper and lower-case letters A-Z as well
as spaces. Where a proc has spaces, that name must be wrapped in double-
quotes. Example:
proc "a b" {} {puts "Hello from proc a b"}
["a b"] # Either
"a b" # Or
See also: procedures, info
Security, Script Tampering and Script Protection
------------------------------------------------
Q) Are security issues discussed in the help file?
A) Yes, see: security issues
Q) Can Ticol scripts be protected from snooping or tampering by users?
A) Yes. Scripts can be encrypted, or, more accurately, 'obfuscated' since
the encryption code is NOT 'military-grade'. This should be sufficient
to prevent all but the most determined snooper. Scripts can be "bound"
(or, "locked") to execute via password, workstation name or username
restrictions
Additionally, obfuscated snippets of code can be embedded within scripts
by using the [tcx] command
See: help protection
Q) Is code-obfuscation strong encryption?
A) No. It is intended to be sufficient to prevent casual code tampering,
not to deter determined attackers or encryption experts
Code obfuscation should not be used to protect secure information
Q) I've forgotten the password to an obfuscated script. Can it be decoded?
A) No. The encryption is not 'military grade' but decrypting password
protected scripts is not a trivial task. If you think you can crack
it then good luck. The author can offer no assistance in decrypting
files. Keep backups of unencrypted scripts and records of encryption
criteria for important files. Assume you won't be able to decode them
and keep regular backups of development code (or use VCS/version
control)
See: help encryption
Q) What filetype is an encrypted Ticol script?
A) TCX
Q) Can encrypted TCX scripts be loaded in the Command Line Interface (CLI)
console?
A) Yes, but they cannot be viewed from the ticol.exe command line
Q) Can encrypted TCX scripts be identified as TCX scripts?
A) No. The data will appear as purely random
They may be converted to a bitmap to visualise this
Q) Can I embed TCX encrypted code within another script?
A) Yes, use the ticol_tcx.dll plugin and the [tcx] command to execute
the block of code
You will need to "compile" using: ticol.exe /C /64
Or use /C then use base64.exe to convert to base64 format which will be
readable by [tcx]
Q) Can encrypted TCX scripts be echoed to the console using /ECHO or
from the CLI using [dump] ?
A) No
Q) Can I get the source code from an encrypted TCX script?
A) No, the intact script is not visible, although some minimal information
relating to procs may be displayed when single-stepping using the
debugger
Q) Can any information leak from a TCX script?
A) Some limited information could be revealed via the debugger and /G
performance trace option, e.g. command names and values
[info proc]
Strings
-------
Q) How can I handle and dereference a string which contains Tcl commands?
As soon as $var is encountered the commands within are evaluated
A) To test for non-null use: if {is_empty varname}
To print out, use echo varname -var
Q) Does Ticol use counted or uncounted ASCIIZ strings?
A) Ticol uses uncounted ASCIIZ strings except for strings created by
[binary] commands, which are counted and which currently have limited
scope
See: binary
Q) I am having problems with very long numeric strings giving incorrect
results. Why is this?
A) If using Notepad++ (and possibly some other editors) then very long
strings may be broken up by the insertion of a carriage-return+line-
feed pair. This will break the string into 2 or more parts
To resolve this you should break the string into shorter parts, wrap
in double-quotes and use line continuation characters.
See: troubleshooting
Q) Are nested quotes allowed?
A No but you can emulate them using [chr 34]
Q) How can I effect nested quotes?
A) Nested quotes are not allowed in Tcl, however, you may use [chr] to
insert quote characters. Alternately, a quote character may be defined
as a var or const
const QUOTE [chr 34]
if {[uplevel @0 "is_array $::QUOTE$q($i)$::QUOTE"]} {
...
See: uplevel
Q) I am having issues escaping within strings. What can help?
A) Use a recommended 'minimal escaping style'. Quotes are not always
required in Tcl since (just about) everything is a string anyway
Don't use quotes where braces could be used
Don't use braces where no backslash substitutions are otherwise needed
See:
https://wiki.tcl-lang.org/page/Tcl+Minimal+Escaping+Style
Structs
-------
Q) Can Ticol structs be used freely as in C/C++?
Q Can I used Ticol structs to develop complex ADTs?
A) No. Structs are mainly provided to aid the marshalling of data to/from
external DLLs. They may have some limited use inside Ticol scripts but
as they cannot be instantiated as unnamed blocks of data it would be
difficult to use them to create complex Abstract Data Types.
The intention is that any routines which require complex ADTs would
be written as a plugin DLL in C++
Q) How do I get the memory address of a struct data block?
A) Use [addressofb]
Example:
struct x {a 10}
puts [addressofb x.a] # Gives the address of struct x and x.a
Q) How do I retrieve the binary contents of a struct field?
A) Use [ofaddressb]
Example:
struct x {a 10}
struct setb x.a 12345 # Set binary value
puts [ofaddressb x.a] # Will return 12345
Testing
-------
Q) What testing has been performed on Ticol?
A) Reference tests has been made against ActiveState Tcl. Visual BASIC 5,
MSVC++ and CScript/WScript. Additionally, around 200 test scripts have
been used to help validate against regression or other errors during
development. These will usually be included with the distro and can
all be run from a single tests.tcl using the command: ticol tests
See: security issues
Threading
---------
Q) Can Ticol Tcl handle multi-threaded operations?
A) In the general release compile version, no, it doesn't
An experimental/development version is available which will thread but
not terribly well. The underlying interpreter is really not designed
for threading and a proper threaded version would most likely require
a rewrite from the ground-up. The experimental version queues complete
execution threads through a single interpreter instance. It does not
run multiple interpreters and synchronise results
I may do more work on the threaded version once the unthreaded release
becomes more stable. Any threaded release is offered only as a test
platform
If you require a stable, multi-threaded Tcl then I would recommend
ActiveState Tcl
Tricks
------
Q) Is it possible to link Ticol Tcl to any speech engine?
A) Yes, via VBScript and an intermediate temporary file:
set vbstring {Set Speak=CreateObject("sapi.spvoice")\r\n\
Speak.Speak $r} # We will [subst] r
set r $pi # Assign some numbers
set filename [escape [mktemp vbs]] # Create a unique file
writefile $filename [subst $vbstring] # Write with subst $r
exec cmd /c $filename # Works OK
# spawn - cmd /c $filename # This works OK too
unlink [escape $filename # Remove temp file
A) It is also possible to use lib ticol_cscript.dll without a
temporary file as follows:
set txt {Hello world} # Assign some text
set s "Dim Speak # VBS Skeleton
Set Speak=CreateObject(\"sapi.spvoice\")
Speak.Speak \"$txt\""
lib ticol_cscript # Load cscript plugin
cscript [unescape $s] # Call [cscript]
lib ticol_cscript -unload
Useful References
-----------------
http://wiki.tcl.tk
http://www.invece.org/tclwise/
Variants
--------
Q) Are VARIANT data types supported
A) Yes. Ticol structs can be used to emulate variants
A) A plugin, ticol_variant.dll offers a small number of variant handling
routines. The intention is that any complex code requiring variant use
would be implemented as a plugin DLL written in C++
See: plugins, ticol_variant, ticol_calldll
Windows Console
---------------
Q) How can I stop Ticol trying to run unknown commands as a Windows command?
A) Use the 'option autoexec off' command to disable this.
You may re-enable using 'option autoexec on".
This allows Ticol to be used as a Windows command shell or to call a
procedure called 'unknown'
See also: help option
See also: help ticol.ini
See also: help unknown
See also: help autoexec
-o-
See also: non standard tcl commands, help, tips, Ticol, ticol.ini, bugs,
design goals
Ticol-Specific, Non Standard Tcl Commands |
Index |
Ticol offers the following useful non-standard commands to make the
language more flexible and efficient, either natively or via a simple
plugin system which can be called from scripts. These were developed by
the author out of necessity for various tasks and written in C++ to
avoid using longhand Tcl
/* ... */ C/C++ style long comments
. String head dereference (Dot) Similar to C++ '*string'
#if <expression> A basic C++ like Macro PreProcessor which allows for
#ifdef conditional execution. #if can take a Tcl command and
#ifndef execute conditionally depending on the result
#endif This can regulate execution to a particular time/PC etc.
#define #define may include complex Tcl substitutions
#undef Standard macro variables may also be defined or undefined
#else and used to control execution
#endif Also, you may execute Tcl commands directly or echo
#echo output to the console using #exec and #echo
#exec Execute a Ticol statement
#exit Immediately halt the Macro PreProcessor
@<command> Alias for exec. Run a Windows program
addslash Append a slash to a path (if required)
addressof Return the address of a var as a number
addressofb Return the binary address of a string
after Run a command after a given number of milliseconds
alias Create an alias of a proc or command
all Test multiple booleans
any Test multiple booleans
arity Return the arity value of a function (See [funct])
arp ARP lookup of IPV4 address (plugin)
array foreach Loop foreach element in an array
asc Return the ASCII value of a character
assert Debugging routine
at_exit A routine to be called when a script exits
autoexec Control the execution of unknown commands
base64 Base64 encode/decode a string or variable
baseconv Convert from one base to another
beep Play a system beep with frequency/duration
big Large number arithmetic (plugin)
bigtext Large text display (plugin)
binary Handle binary strings (experimental)
bintoint64 Convert a binary number to an integer
bool Handle boolean flags and values
box Draw a graphic box
calc Macro PreProcessor (MPP) optimised [expr]
call Call a Tcl function discarding any return value
calldll Call a __stdcall DLL and return a long (plugin)
calldll_dbl Call a __stdcall DLL and return a double (plugin)
calldll_str Call a __stdcall DLL and return a string (plugin)
calldll_variant Call a __stdcall DLL and return a variant (plugin)
calldll*_cdecl Call a __cdecl DLL (e.g. msvcrt40.dll) (plugin)
calldll_cdecl Call an external __cdecl DLL (plugin)
calldll_dbl_cdecl "
calldll_str_cdecl "
calldll_variant_cdecl "
calldll_open Get handle to a DLL (plugin)
calldll_close Close DLL handle (plugin)
carray C-like arrays (plugin)
cat Concatenate values
char Cast a value to an 8-bit (1 byte) integer
charcount Count the number of characters in a string
chartok Transform string to array by characters in a mask
chomp Efficiently slice up large strings
chr Return a character 0..255
clear Erase all variables
cli Open a new command line interface
clipboard Manipulate the Windows clipboard
cls Clear the screen
cmp Efficient string comparison
concat Efficiently concatenate strings
console Control the console
const Declare a constant string
convbase Convert from one base to another
comma Comma-format a numeric value
cos Calculate the cosine
cosh Calculate the hyperbolic cosine
copyfile Copy a file using Windows API
cpu Get CPU information (plugin)
crc CRC checksum (plugin)
ctime_to_double Convert from Unix C time to double
date Date routines
debug Turn debug mode On or Off
decimal_to_roman Plugin command in ticol_r2d
decrypt Decrypt a string encoded by [encrypt]
degrees Convert radians to degrees
die Exit to Windows returning an optional errorlevel
dim Declare an array (allocate an empty hash table)
diskfree Show free disk space (64-bit integer)
diskstat Show disk status as a code or return letter/UNC path
disktype Return the disk type
dll_close Unload a DLL loaded with dll_open
dll_open Load a Windows DLL (__cdecl or __stdcall)
dns Simple DNS lookup (A, MX) (plugin)
do do-while loop flow-control
doevents VB equivalent yield control to Windows
double Cast a value to an 64-bit (8 byte) real number
double_to_ctime Convert from a double to Unix C time
dump Debugging analysis of variables or source code
dumpvariant Debugging analysis of a variant object
echo Print a literal string without backslash translation
elevate Self-elevate if running non Windows XP or 7+
encrypt Encrypt a string or variable using moderate encryption
enum Declare a series of const enums
eq String comparison equals (-nocase)
eqn String comparison equals for given length (-nocase)
escape Escape a string by adding backslash characters
event Windows event log messages (plugin)
exists Test if a file exists
exit Exit the Ticol interpreter
explain Explain (expand) a [calc] expression
extract Retrieve characters from the left side of string by mask
fib Compute fibonacci series value
file_crc Calculate the CRC32 checksum of a file
filesize Return the size of a file
filter Filter characters from a string by mask
fmod Floating point remainder after division
fpart Return the fractional part of a double as an integer
fraction Return the integer part of a real number
free Release memory assigned by malloc
fromhex Convert hex characters to a string
ftp Simple FTP client services (plugin)
funct Call an expr function directly
functions List expr functions
ge Greater-than or equal-to string comparison
get_fileopen Common-dialog file-open menu
get_filesave Common-dialog file-save menu
get_win_error Retrieve the Windows error number
getcwd Return the current working directory
getenv Return an environment variable
getsid Retrieve a SID matching a given username
goto Flow-control goto. Valid within a goto_block
goto_block Block of code which contains line-numbered statements
gotoxy Locate the cursor
gt Greater-than string comparison
halt Enter the single-step debugger
help This help file
hostname Return the configured host name
html HTML manipulation commands
html header_field Get an HTTP response field from the header
html remove_header Remove the HTTP response header block
html chomp Parse and strip HTML source code
http Retrieve a page of HTML from an HTTP server
in Check a list for an item
index Rapid single character index into a string
ini INI file handling (Plugin)
inkey Get a keypress
input_csv Read standard CSV data from file and return a Tcl list
inputbox Text input window
inrange Test if an integer is within a given percent range
instr BASIC-like string matching with case-ignore arg
int Cast to integer (64-bit)
integral Return the integer part of a real number
inttobin Convert from integer to binary
inttohex Convert from integer to hex
inttooct Convert a number to octal format
ip_to_long Convert an IPV4 address to unsigned long
is_admin Return true (1) if user is an admin
is_command Detect if a command or proc is implemented
is_const Test if a variable is a constant
is_dir Check if a filename is a directory
is_dst Test for Daylight Savings Time
is_elevated Detect if Ticol is running in Win7+ elevated mode
is_empty Test if a variable is both set and empty
is_list Test if a string is a valid Tcl list
is_mod Test for modular division safely
is_pointer Test if a variable is a pointer (address value)
is_proc Test if a command is a proc or not
is_set Test if a variable is set or not
iselevated Returns true (1) if Ticol is running elevated
json_to_list Convert JSON format data to a Tcl list
K K-combinator. See functional programming
le Less-than or equal-to string comparison
left Return the leftmost characters of a string
len Alias for [string len]
let Variant of set which accepts a Tcl expression
lib Load commands in a Ticol extension DLL
link_to_path Resolve a Windows LNK shortcut file to a path
linkedlist Standard (non Tcl) linked lists
list Returns a well-formed Tcl list
list_to_array Convert a well-formed list to a Tcl array
load Load a Ticol script to the CLI or a variable
long Cast a value to a 32-bit (4 byte) integer
long_to_ip Convert a long value to IPV4 string
loop Efficient [for] loop
lrand Return a random element from a list
ls Perform a simple directory listing
lsearch Search one list for a match to another list
lt Less-than string comparison
makestr Create a string with optional fill character
malloc Heap memory allocation
map Tcl map function. Map and replace strings
max Math test
md5 Generate an MD5 checksum string
md5_file MD5 checksum of a file
mean Calculate the mean of a series of numbers
mem_used Show the memory used by Ticol
memset Erase/set a buffer for use with calldll*
mid Set part of a string
mids Return part of a string
min Math test
mkdir Create a folder
movefile Move a file using the Windows API
msgbox Display a Windows MessageBox
mul Math operator
ne Compare string inequality
nen Compare string inequality for a given length
newline Print a new line to the console
ni Test if string not in a list
now Return the current date/time as a double in VB6 format
null NULL value
octtoint Convert from octal to decimal
ofaddress Inverse of addressof. Returns the original data
ofaddressb Binary inverse of 'struct setb'
on Emulation of BASIC's ON..GOSUB
option Configure Ticol options at runtime or at the CLI
pause Wait for a keypress
peek Check the top of a global stack
pid Return the interpreter's process ID (PID)
ping Plugin ICMP ping module (plugin)
pipe Pipe together a series of commands without nesting
play_wav Play a WAV file
printf Calls the format command. Mimics C/C++ printf
precision Returns the effective precision of a real number
procs List user-defined procedures
put Writes n bytes value to a struct variable safely
pwd Prints the current working directory
rand Return a random number
randomstr Return a random string
range Return a string in a given range
rarp Reverse ARP lookup of MAC address (plugin)
readfile Read a file to a variable
registry Registry commands (plugin)
replacechar Search and replace single characters
resume Cancel a previous debug-trace [halt] command
right Return the rightmost characters of a string
rnd Return a random number in given range
roman_to_decimal Plugin command in ticol_r2d
rot13 Apply rot13 transformation to a string
round Round a value
rtf Simple RTF text output (plugin)
run Aliased to [source]. Run a [load] script
run Run a Tcl file from the interpreter shell
scan Scan a string to one or more variables
screen Screen metrics
sendmail SMTP send mail command (plugin)
service Simple service view command (plugin)
setat Set a variable from a string segment
setb Set a binary value in a struct field
setenv Set an environment variable
setl Insert source string to target, left-aligned
setr Insert source string to target, right-aligned
setresult Function for plugin DLLs to set the Tcl return value
shell Spawn a Windows Command shell
short Cast a value to a 16-bit (2 byte) integer
signed Convert unsigned integer to signed value
sin Calculate sine
sinh Calculate hyperbolic sine
sizeof Return the estimated byte size of an integer
sleep Suspend execution for N milliseconds
spawn Run a console program with output capture
sql Basic SQLite database commands (plugin)
sqr Return the square of a number
sqrt Return the square root of a number
srand Set the random number seed
stack Native stack objects
stop Halts execution and returns to the Ticol CLI
stop_wav Halt playing a WAV file
store Rapid buffered string concatenation
strchr Index a character into a string
string to_wchar Convert an ANSI string to a wide character string
stripto Strip LH string to a given character
striptor Strip RH string to a given character
strptr Return the binary address of a string
strsplit Split a string into a list
strstr Returns the search index for one string in another
strtok Transform a string to an array by string separator
struct Declare a binary struct (for calldll)
swap Swap two standard Tcl variables
sysfree Release a string returned from a DLL
tab Return a string of n spaces
tan Calculate the tangent
tanh Calculate the hyperbolic tangent
tcx Run obfuscated TCX code within a script
textcolor Set the text color
ticks Shows the tick count in milliseconds
time Shows the current system time
tohex Convert a string to hex characters
trim More efficient [string trim]"
true Boolean value
try Try / catch structured error handling
type Show the interpreted type of a variable
unknown Standard Tcl but interacts with autoexec on|off
unlink Delete (unlink) a file
unsigned Cast a variable to unsigned integer
username Return the current Windows user
username Return the current Windows user
varray Declare a variant array (for calldll)
varray get Return a 1D or 2D variant string element
varray lbound Return the lower bound of a variant array
varray size Returns the count of a 1D string variant
varray ubound Return the upper bound of a variant array
vars Display a list of variables
vsplit Split a variant array
waitkey Wait for a keypress or time out
watch Trace variables when in single-step debug mode
wchar_length Get the length of a wide-character string
wchar_to_string Convert a wide-character string to ANSI
win_error Display Windows error messages
wrap Format a string to a given column width
write Write to file
xbase DBase III/IV database plugin
zip Simple ZIP module (plugin)
-o-
See also: Ticol
Ticol is intended as a casual tool for occasional toolbox scripts and
may have some use as a learning tool for students.
Ticol is not recommended for use as a continually-running script. It is
intended for relatively short lifespan tool scripts which run for
possibly less than 100,000 cycles before exiting.
Memory testing will have been performed to check stability on a routine
basis up to around 100,000 operational cycles.
Short scripts which are called intermittently and on a regular basis
but which exit then restart should be ok. This would most likely
include CGI scripts although Ticol is not recommended as a production
CGI script engine, particularly as there are far better alternatives such
as PHP and PERL.
Although the size of each script is limited only by available memory it
is strongly recommended to keep scripts not much longer than a few
thousand lines. Larger scripts can be broken up and run on a modular
basis using the [source] command
Ticol Tcl Interpreter for Windows |
Index |
"Inside every large program there is a small program trying to get out"
(Sir Tony Hoare)
"Inside every small program there is a large program trying to get out"
(Ticol Author)
A compact Tcl subset-compatible ANSI script language for Windows which is
a fully portable app needing no dependencies on Windows. Ticol covers a
subset of Tcl with many additional features:
Windows versions from NT4.0/SP6a to Server 2012R2 and recent versions of
ReactOS are supported
Code obfuscation using several methods of endpoint execution-locking is
available. Regular-expressions and pattern-matching are not yet supported
Unicode and non-ANSI character languages are not supported
Ticol is also regularly tested on ReactOS with a view to compatibility
The language adds many additional features to basic Tcl command list
See: ticol specific commands
Ticol was written using MSVC++ 5.0. (Still useful!)
See also: Ticol design goals, alphabetic list, Tcl, help, PreProcessor,
Ticol command line, debugging, performance, functions, Tcl syntax,
flow control, comments, workarounds, bugs, commands by function
tips, faq, tutorial
-o-
A Brief Introduction to Programming in Tcl
At first glance, Tcl code looks very much like C/C++ but the syntax is
nevertheless quite different. If you are a C/C++ programmer then you will
need to unlearn a few habits possibly dispel the odd assumptions about
how you might write code when using Tcl
Tcl uses curved braces to delineate blocks of code, but the equivalent to
C/C++ functions are 'commands'. Commands are evoked, either 'naked' by
using the name of the command openly in a script, or by enclosing within
square brackets, thus:
command # Call 'command'
[command] # Call 'command'
Commands may have zero or more arguments. Some may be optional and
command modifiers a typically prefixed with '-' e.g. -nocomplain
# puts is a command to output to the console
puts "This string won't print a newline" -nonewline
Tcl comments begin with a hash # character but Ticol also supports C-
style /* .. */ long comments The two may be combined.
Various flow-control constructs are available including 'if' and 'for'
which most programmers will already be familiar with. These constructs are
also implemented as Tcl commands and obey the same rules as in many other
languages such as [if], [for], [while], [switch]. [do], [try..catch]
Curly braces delay evaluation and are used to enclose both blocks of code
to be evaluated and strings. In most cases you should enclose strings
within curly braces rather than double quotes but neither is obligatory
unless the string contains whitespace. Since evaluation is delayed, various
blocks of code can be passed around as arguments to commands without being
interpreted while being transferred. The use of curly braces looks a lot
like C/C++ but the function is quite different.
Command arguments are terminated by an unbraced line ending so braces
must 'protect' line-ends using K&R style bracing. Failure to adhere to this
style will cause errors in your script
proc foo x { # Correct K&R Tcl style
... # [foo] receives the body argument correctly
}
proc foo x # Incorrect Allman/BSD C/C++ style
{ # [foo] will fail to receive the body argument
... # contained within the braces
}
Curly brace and square bracket counts must match within a script
[if], for example, can take a varying number of arguments including
'elseif', each of which must be properly braced if they are to be split
across multiple lines. Note that you could, if you wished, put the command
on a single line but this is not particularly readable! ...
if {something} {
do-something
} elseif {something-else} {
do-something-else
}
The expression arguments used by various flow-control constructs can be
either of 'expression' type with infix command-operators or in Tcl prefix
style with prefix command operators. This option is set either within
ticol.ini or by using the 'option expression' command
option expression off # Commands will use Tcl prefix notation
option expression on # Commands will use natural/infix notation
Tcl commands should be separated by whitespace. Tab characters should
normally be avoided other than for indenting code for the sake of
readability. You can't 'butt' commands against each other unless they are
both wrapped in square brackets.
![rnd 0 1] # Invalid. Requires a space between ! and [rnd]
! [rnd 0 1] # Correct style with whitespace as separator
puts [rnd 0 1][rnd 0 9] # Correct. Results are concatenated
Operators in C/C++ such as +, -, == etc. are implemented in Tcl as commands
and can be called as such. These command-operators may be used as arguments
for flow-control constructs. So commands such as [if] and [for] may take
relational commands such as [>] or [==]
option expression off # Use Tcl prefix syntax
set i [rnd 0 1] # Set i to either 0 or 1
if {== $i 0} { # [==] is used in prefix format
puts "i is 0" # [puts] is used to print output
} else { # [else] is a subcommand of [if]
puts "i is 1" # [puts] is called as a 'naked' command
} # Indent properly for readability!
There is no [=] assignment in standard Tcl although Ticol Tcl defines this
command as an alias to [set]. Instead, the [set] command (similar to the
BASIC LET statement), is used either to assign values or to return them to
some other command or a variable.
As an alternative to using [set] to return the contents of a variable, the
dollar prefix may be used to dereference (access) a variable's contents.
set s "Hello world" # Assign variable s using the [set] command
puts $s # Dereference variable and print using [puts]
We could also enclose the string in the example above in curly braces, thus
and in many cases this is preferable. See brace coding styles in Tcl.
set s {Hello world}
Console output is generated either using [set] (standard Tcl) or [printf]
(non standard Ticol command). The [printf] command is merely a wrapper
for the [format] command which is a standard Tcl one. Both [printf] and
[format] offer a C/C++ style range of formatting options and share common
flag characters etc. such as %s or %f
By default [puts] prints a newline whereas [printf] does not
set s "world"
puts "Hello $s" # Print using [puts]
printf "Hello %s\n", $s # Print using [printf]
Tcl commands may be nested. In which case the innermost command is executed
first, followed by the next innermost and working outwards. Tcl commands
are evaluated within double-quoted strings since evaluation is is not
delayed by double quotes. If an expression is wrapped in curly braces
such commands are not called immediately
puts [chr [rnd [asc A] [asc Z]]] # Prints a letter from A to Z
puts "[chr [rnd [asc A] [asc Z]]]" # Prints a letter from A to Z
puts {[chr [rnd [asc A] [asc Z]]]} # Prints [chr [rnd [asc A] ...
Complex expressions may be constructed using Tcl commands but the
syntax may take some getting used to. An alternative is to use
a natural expression supplied to the [expr] command:
puts [* [+ 23 57] [/ [+ [* 3 10] 3] 13.0]] # 203.076923076923040
puts [expr (23+57)*(((3*10)+3)/13.0)] # 203.076923076923040
The [expr] command will prioritise expressions in BEDMAS order and it is
good practice to always brace expressions supplied to [expr]
Note that [expr] is less efficient than using raw Tcl commands but is
much easier to read. Commands may also be inserted into expressions
providing they return a numeric result
puts [expr [+ 23 57]*(([* 3 10]+3)/13.0)] # 203.076923076923040
Variables are deleted using the [unset] commands. Since it is an error
to attempt to delete a non-existent variable you may wish to use the
'-nocomplain' argument. Many commands have a '-nocomplain' option which
will cause any fatal errors to be ignored. [unset] can also take many
variable names as it's argument and can also perform 'glob' style matching
if required:
set s "Hello world"
unset s -nocomplain # s exists and is deleted
unset foobar -nocomplain # foobar does not exist
In fact, most commands have optional arguments such as '-nocase' or
'-nocomplain' which modify their behaviour. These are documented
within the built-in help file. To find out more use the command
help help at the Ticol command line.
Various iterator commands are available and the [for] command looks
very similar to that in C/C++ but again, can operate using Tcl prefix
syntax or expression (infix) syntax as defined by [option expression]
option expression off
for {set i 0} {< $i 10} {++ i} { ... }
option expression on
for {set i 0} {$i < 10} {++ i} { ... }
It is a good idea to pick one style and stick to it within a script to
avoid hard-to-debug errors
Ticol also offers a [while] command, [do...while] and a useful [loop]
command as follows
loop i 0 10 { # Sets variable i to zero and iterates to 9
puts $i # Prints numbers from 0 to 9
}
Procedures are Tcl's equivalent to C/C++ functions. Once defined you can
call procedures in the same way as built-in commands. Nested commands
are evoked via square brackets...
proc foo {x} {
return [* $x 2]
}
puts [foo 23] # Call foo and print the result (Returns 46)
Procedures support calling arguments by name or by value. If you call by
name you will need to use [upvar] to reference the variable within the
parent scope. For a discussion, see: upvar
proc foo {x} {
upvar $x y # Create a link to the parent variable as 'y'
return [* $y 2]
}
set i 23
puts [foo i] # Returns 46
Separate script files may be executed using the [source] command
source \path\to\some-other.tcl
and scripts may also be loaded and executed at the Ticol command line
using [load] and [run] (or just [run scriptname]). You can also inspect
the currently loaded script (if any) using [dump]
c:\>ticol.exe # Run Ticol from the Windows console prompt
load foo # Load foo.tcl
dump # Show the contents of foo.tcl colour highlighted
run # Run foo
exit # Exit the interpreter and return to Windows
To halt a running script use the CTRL+BREAK key sequence (not CTRL+C)
To get a better understanding, have a look at the example scripts in this
help file
See also: info, faq, troubleshooting
-o-
Windows GDI Graphics Plugin |
Index |
To load, use: lib ticol_win ?-verbose? ?-nocomplain?
This plugin provides basic access to Windows GDI drawing functionality
including the ability to open a Window and draw various shapes or text
The plugin was developed to provide a solution for a specific task
rather than as a general purpose windowing solution. This task was
to provide an enriched form of console mode drawing.
This is a new and largely experimental feature and is subject to future
revisions and amendments. The range of commands is not yet complete
but has been widely tested.
See the screensaver sample source code: 'help win example'
This can also be extracted from the screensaver SCR binaries
using: 'screensavername.scr /echo'
The following commands are supported
# Return an RCB hex value as RRGGBB (values are optional)
[rgb ?r? ?g? ?b?]
# Return window active status (e.g. mouse exit)
[win active]
# Draw a character mode box with fill option
[win box x y w h ?fore? ?back? ?-single? ?-solid? ?-nofill?]
# Draw a filled circle
[win circle handle x y ?colour? ?width?]
# Close the window and release any resources
[win close]
# Clears the graphic window, optionally to a given colour
[win cls ?colourname|#RRGGBB?]
# Return the number of columns
[win columns]
# Create a graphic window (only one window currently supported)
[win create -w width -h height ?-fullscreen? ?-smallfont? \
?-largefont? ?-movequit? ?-closeonlostfocus?]
# Default is show
[win cursor show|hide]
# Drawn an ellipse
[win ellipse handle x y w h ?color? ?width? ?-dot|-dash|-dotdash|-solid?\
?-fill ?colour??]
# Select a drawing font
[win font -name FONTNAME -size NN -bold -italic -underline -normal -clear]
# Locates the cursor at character X/Y coordinate using character coords
[win gotoxy x y ?-t?]
# Return the window height in pixels
[win height]
# Draw a line
[win line x1 y2 x2 y2 ?colour? ?width? ?-dot|-dash]-dotdash?]
# Plot from the stored [moveto] coords to the given x/y point
[win lineto x y ?colour? ?width? ?-dot|-dash]-dotdash?]
# Move the line plotting start point to x/y coords
[win moveto x y]
# Test if WM_PAINT was called and if we need to repaint our window
[win paint ?n?]
# Draw a parallellogram
[win para origin-xpos origin-ypos width height rot_angle1|90 \
splay_angle2|90 colour|grey pwidth]
# Set up for Windows Screensaver preview (Pass handle from Windows)
[win preview HANDLE]
# Draw a single pixel with optional colour value
[win pset x y ?color?]
# Draw a text string to the graphics console window
[win puts string ?-size HEIGHT? ?-rot ANGLE?]
# Draw a rectangle
[win rect x y w h ?color? ?width? ?-dot|-dash|-dotdash|-solid?\
?-round? ?-fill ?colour??]
# Return a random HTML colour as "#RRGGBB" hex
[win rndcolour ?lo|-? ?hi|-?]
# Return the number of screen rows
[win rows]
# Set absolute cursor position (pixel coordinates)
[win setpos x y]
# Blocking sleep
[win snooze ?milliseconds?]
# Set foreground/background using Ticol constants or "#RRGGBB" values
[win textcolor fore|- ?back?]
# Get the current cursor X location in columns
[win wherex]
# Get the currnt cursor Y location in rows
[win wherey]
# Get the window width in pixels
[win width]
# Get the absolute pixel position for character coordinate X-axis
[win xpos]
# Get the absolute pixel position for character coordinate Y-axis
[win ypos]
The cycle of coding involves creating a new window and providing a local
event loop to service it. Without an event loop the window will hang.
No handles are required other than to test that the window opened
correctly since only one graphics output window is allowed
The returned window handle may be passed to Windows API calls
The basic cycle of operation is as follows...
option expression off
# Load the required plugin
lib ticol_win -nocomplain
# Create the window
set h [win create -fullscreen -smallfont]
if {== $h 0} {
console show
puts "Failed to open console output window"
pause
die
}
# Optionally, hide the text mode console
console hide
# Optionally, hide the main cursor
win cursor hide
# Service the window
while {win active} {
# Allow user cancel (if desired)
if {inkey} {break}
# Service the loop here ..
# main_service_proc # Call any main service proc
win snooze 100 # Call win snooze if desired
# snooze 100 # Call any alternate/custom snooze
}
# Clean up
win cursor show
console show
win close
lib ticol_win -unload
Closing the parent console window will also close the graphics
window. You may wish to hide the parent console using [console hide]
but it will usually be vital to ensure that [console show] is called
on script exit to avoid losing control of the console.
If execution of your script has to pause you should create a
suitable 'snooze' procedure or call [win snooze] at regular intervals.
(See: win snooze)
Linux WINE and ReactOS Compatibility
------------------------------------
The ticol_win plugin is not fully compatible with WINE or ReactOS due to
minor differences in the underlying API.
Windows Compatibility
---------------------
This plugin has been tested with all NT family variants of Windows from
NT4.0/SP6a to Windows Server 2012R2
See also: plugins, win snooze, win active, console, lib
hex-string [rgb ?red? ?green? ?blue?]
[rgb] takes optional integers for red green and blue components and returns
a formatted hexadecimal string as "RRGGBB"
Example
-------
puts [rgb 8 4 68]
Result
------
080444
See also: win, hex
bool [win active]
Returns true while a window opened using [win create] is active
[win active] may also be called passively within long loops to ensure that
the underlying Windows event loop is polled and the window avoids being
'frozen' by Windows assuming it is non-responsive.
Example:
# The terminating conditions here depend on [win create]
option expression off
lib ticol_win
win create -fullscreen
while {[win active]} {
if {inkey} {break}
win snooze 100
win gotoxy 1 1
win textcolor blue
win puts [time]
win textcolor
}
win close
lib ticol_win -unload
See also: win, win create, win snooze, sleep, plugins
void [win box x y w h ?fore? ?back? ?-single? ?-solid? ?-nofill?]
Draws a GDI box with optional border style and fill in a window opened
using [win create]
Coordinates are pixel coordinates, not character coordinates
See also: win, win create, plugins
void [win circle x y ?border-colour? ?diameter? \
?-fill ?fill-colour??|?<style>?]
Where <style> is one of: -dot, -dash or -dotdash
Draws a GDI circle with optinal border style and fill in a window which
was opened using [win create]
Coordinates are pixel coordinates, not character coordinates
The smallest diameter which can be drawn is 2 pixels. This is a limitation
of the underling API.[win pset] can be used to draw single pixel points
See also: win, win create, plugins
bool [win close]
Closes a window which was opened using [win create]
Only a single window is permitted so there is no requirement to pass the
window's handle
See also: win, win create, plugins
void [win cls ?colourname|#RRGGBB?]
Clears window which has been created using [win create], optionally
the command may specify a given colour
See also: win, win create, plugins
integer [win columns]
integer [win rows]
Returns the number of text mode character columns or rows for a window
which has been created using [win create]
The columns and rows value will depend on which font size was selected
when the window was crated
See also: win, win create, plugins
handle [win create -w width -h height ?-fullscreen? ?-smallfont? \
?-largefont? ?-movequit? ?-closeonlostfocus? ?-center?]
Creates and opens a graphics mode window. The width and height may be
specified in pixels or a full-screen window may be specified.
A full screen window will occupy all monitors on a multi-monitor system.
Small character fonts may be specified or larger than standard fonts via
the -smallfont or -largefont arguments. The font size selection will
affect the column and row values
Optionally, the window can be set to close on mouse movement or if
focus is lost.
A window opened by [win create] requires an event loop which must
continuously poll [win active]. If this is not done then the window
will 'hang'.
The new window will be positioned by Windows unless the -center option is
given
See also: win, win active, plugins
void [win cursor show|hide]
Either shows or hides the Windows graphic cursor. This does not affect
the console mode (text) cursor
See also: win, win create, plugins
void [win ellipse handle x y w h ?color? ?width? \
?-dot|-dash|-dotdash|-solid? ?-fill ?colour??]
Draws an ellipse in a window which has been created by [win create]
This can have a varying colour, border width or border shading style
as well as a differing fill colour
See also: win, win create, plugins
void [win font -name FONTNAME -size NN -bold -italic -underline\
-normal -clear]
Selects a font style for use with [win puts] in a window which has been
created by [win create]
Varying font style attributes may be specified by the arguments
See also: win, win create, win puts, plugins
void [win gotoxy x y ?-t?]
[win gotoxy] works analogously to [gotoxy]. It locates the cursor at
character coordinates specified by x and y
Note that these are block character rather than pixel coordinates
unless the -t argument is used.
The -t argument will force translation from character mode addressing
to pixel mode addressing so that the selected X/Y location is set up
ready for pixel addressing in the same manner as [win setpos]
See [win setpos] to set absolute pixel coordinates for use with -rot
Example:
# Assuming [win create] and an event loop
win gotoxy 10 5
win textcolor magenta
win puts "Hello World"
Result:
Will print "Hello World" in magenta at x=10, y=5 in character block mode
Do not confuse [win gotoxy] with [gotoxy]
See also: win, win create, win setpos, plugins
integer [win height]
integer [win width]
Returns the width or height of a window created by [win create] in
pixels
See also: win, win create, plugins
void [win line x1 y2 x2 y2 ?colour? ?width? ?-dot|-dash]-dotdash?]
Draws a line from x1,y1 to x2,y2 with optional colour, width and style
values in a windows which has been created by [win create]
See also: win, win create, win lineto, plugins
void [win lineto x y ?colour? ?width? ?-dot|-dash]-dotdash?]
Draws a line from the position last stored by [win moveto] to the x,y
coordinates specified in a window which was created by [win create]
Varying colour, width and line style attributes may be specified
See also: win, win create, win line, win moveto, plugins
void [win moveto x y]
Store the pixel x/y coordinates for use with [win lineto] in a window
created by [win create]
Sequences of [win moveto] followed by [win lineto] enables 'logo' like
'turtle' graphics.
See also: win, win create, win lineto, plugins
bool [win paint ?n?]
Test if a window created by [win create] has received a WM_PAINT message
and if it should be redrawn (repainted). Optionally the count of
required WM_PAINT messages may be specified. The default is 1.
See: win snooze for example code
See also: win, win create, win active, plugins
void [win para origin-xpos origin-ypos width height \
rot_angle1|90 splay_angle2|90 colour|grey pwidth]
Draws a parallelogram in a window created by [win create].
The left corner x/y position must be specified as well as the height,
rotation angle and splay angles
Example:
The paralllogram is located with one point at x,y with a possible rotation
angle about that point. The width and height can be specified and the
'splay' angle which distorts from a rectangular shape can be specified
(rotation)
x,y width
+--/-------------+
\/ splay \ height
\ \
+----------------+
Optionally the border colour and width may be specified
See also: win, win create, win circle, win rect, plugins
void [win preview HANDLE]
Set up for Windows Screensaver preview in a window object which was
created by [win create]. You should pass the handle returnd by Windows
via the /P command line callback argument.
This will create a preview window within the control panel's screensaver
preview area
The following snippet may be used to handle Windows screensaver
preview mode...
for {set i 0} {< $i $argc} {++ i} {
if {&& [eqi $argv($i) "/p"] [< [+ $i 1] $argc]} {
set preview_handle $argv([+ $i 1])
win preview $preview_handle
set i 0
set start [timer]
draw_preview # Call your preview proc
while {win active} {
if {> $i 20} {
draw_preview # Call your preview proc
set i 0
}
loop 10 {
if {! [win active]} {break}
doevents
sleep 10
}
++ i
# Destroy the console screen after 1 minute
if {> [timer] [+ $start 60000]} {break}
}
console show
exit # You must now exit this instance
}
}
See also: win, win create, plugins
void [win pset x y ?colour?]
Will plot a single pixel in a window opened by [win create] at a given
x and y location. The x/y location must be within the bounds of the current
window
Optionally a colour can be specified using the same format as
[win textcolor]. If a colour value is omitted then the value set by
[win textcolor] foreground will be used
The colour argument may use Ticol colour constant names or "#RRGGBB" values
[win circle] can be used to draw larger points
See also: win gotoxy, win puts, win textcolor, win circle
void [win puts string ?-size HEIGHT? ?-rot ANGLE?]
Draws a text string to in a graphics console window created by [win create]
By default, the string is drawn in text mode unless -rot is used and
will therefore locate at the currently set character block mode coordinate
Rotated text can be specified by passing a rotation angle using the -rot
argument.
Note that -rot requires x/y location to be set using [setpos] rather than
[gotoxy]
Example:
# Assuming [win create] and an event loop
win gotoxy 10 5
win textcolor magenta
win puts "Hello World"
Result:
Will print "Hello World" in magenta at x=10, y=5 in character block mode
See also: win, win create, win setupos, plugins
void [win rect x y w h ?color? ?width? \
?-dot|-dash|-dotdash|-solid? ?-round? ?bordersize?? ?-fill ?colour??]
Draws a rectangle in a window which was created by [win create]
Optionally the colour, line width and line style may be specified, as
well as an optional fill and fill colour
Example:
x,y
+------------------+
| |
| | Height
| |
+------------------+
Width
See also: win, win create, win circle, plugins
integer [win rndcolour ?lo|-? ?hi|-?]
Returns a random HTML formatted colour as "#RRGGBB" hex
Optionally this may be restricted to the 'lo' or 'hi' ranges
Values are not bit-masked per RR/GG/BB range but will follow a continuous
range from 'lo' to 'hi'.
Example:
puts [win rndcolor]
puts [win rndcolor 0xff 0xfff]
puts [win rndcolor 0x1f000 0xff0000]
Result:
#d0401a
#0006a6
#1e1c23
See also: win, win create, plugins
void [win setpos x y]
Sets the absolute cursor position (pixel coordinates) in a window which
was created using [win create].
[gotoxy] sets the drawing location based on character block locations
May be used with [win xpos], [win ypos] to query the pixel location
See also: win, win create, win xpos, win ypos, plugins
void [win snooze ?milliseconds?]
A safe sleep which should be used by code interacting with a window which
was created by [win create] in preference to [sleep]
[win sleep] will permit polling of [win active] and [win paint] whereas
[sleep] will block polling of these events.
Execution delays should call [win sleep] or a custom, internal snooze
routine rather than calling [sleep] since the latter will block execution
Such a routine would need to handle cleanups as required and also trap
and call any repaint events
The code sample below allows for user cancellation to be turned on/off
option expression off
#######################################
# An example [snooze] procedure
# [global] ::cancel_allowed
#######################################
proc snooze {milliseconds} {
set start [timer]
set limit 0
if {&& [== $::cancel_allowed $::true] [> $milliseconds 100]} {
set limit [/ $milliseconds 100]
loop i 0 $limit {
set k [inkey]
if {== $k 27} {
console showcursor
textcolor red
newline
puts " * Cancelled *"
textcolor
win cursor show
win close
console show
lib ticol_win -unload
die
}
if {! [win active]} {
win cursor show
win close
console showcursor
console show
lib ticol_win -unload
die
}
if {[win paint 1]} {
# repaint_proc # Call any repaint procedure here
}
loop 5 {
if {>= [- [timer] $start] $milliseconds} break
if {! [win active]} break
win snooze 20
}
}
} else {
if {! [win active]} {
win cursor show
win close
console showcursor
console show
lib ticol_win -unload
die
}
win snooze $milliseconds
}
}
See also: win, win create, win active, win paint, sleep, plugins
void [win textcolor fore|- ?back?]
Sets the default foreground or background colour for text output in a
window which was created using [win create]
The arguments may use Ticol colour constant names or "#RRGGBB" values
See also: win, win create, plugins
integer [win wherex]
integer [win wherey]
Returns the current character-mode cursor X or Y location in rows in a
window which was created by [win create]
[win wherex] and [win wherey] interact with [win gotoxy]
See also: win, win create, plugins
integer [win xpos]
integer [win ypos]
Returns the absolute pixel Y or Y drawing position
These commands interact with [win setpos] and can be used to override
character block-mode coordinates. These commands should not be usd in
conjunction with [win gotoxy]
See also: win, win create, win setpos, win wherex, win wherey, plugins
Windows GDI Graphics Plugin - Example (LabelMania.scr) |
Index |
This is the full and complete Ticol Tcl source code for the Ticol
LabelMania Windows screeensaver. The Tcl source is embedded within a
Windows SCR (EXE) screensaver binary
This script displays a configured text phrase in embossed-type characters
randomly across the screen in various colours and orientations,
periodically clearing the screen and redrawing.
option expression off
sub at_exit {
console show
}
lib ticol_win
console hide
const MAX_STRING_LENGTH 60
const SCREENSAVER_TIMEOUT_MS 60000
set opt_preview $false
set font_size 18
set text " Hello from Ticol Tcl! "
const APP_NAME "LabelMania"
const CSIDL_LOCAL_APPDATA 28
set appdata [escape [get_csidl $CSIDL_LOCAL_APPDATA]]
if {ne $appdata ""} {
set ini_file "$appdata\\$APP_NAME\\${APP_NAME}.ini"
if {! [file isdirectory "$appdata\\$APP_NAME"]} {
file mkdir "$appdata\\$APP_NAME"
}
if {! [file exists $ini_file]} {
ini write $ini_file "Config" "Text" $text
}
}
if {ne $ini_file ""} {
set s [ini read $ini_file "Config" "Text"]
if {ne $s ""} {
set text $s
}
}
if {> $argc 0} {
for {set i 0} {< $i $argc} {++ i} {
if {eqn $argv($i) "/C" 2 -nocase} {
set r [inputbox "Enter a short phrase of up to \
$MAX_STRING_LENGTH characters:" "Configuration" $text]
if {ne $r ""} {
set r [string left $r $MAX_STRING_LENGTH]
ini write $ini_file "config" "text" $r
}
console show
lib ticol_win -unload
die
}
}
}
win create -fullscreen -closeonlostfocus -movequit
win cursor hide
proc display {s} {
set start [timer]
set text [string left $s $::MAX_STRING_LENGTH]
while {win active} {
win cls "#[rnd 0 0xffffff]"
for {set i 0} {&& [< $i 300] [win active]} {++ i} {
sleep 100
win setpos [rnd 0 [win width]] [rnd 0 [win height]]
win textcolor "#[rnd 0 0xffffff]"
win puts " $text " -rot [rnd 0 3700] -size $::font_size
}
loop 40 {
if {! [win active]} break
sleep 100
}
if {$::opt_preview} {
if {>= [timer] [+ $start $::SCREENSAVER_TIMEOUT_MS]} {
return
}
}
}
}
if {> $argc 0} {
for {set i 0} {< $i $argc} {++ i} {
if {&& [eqi $argv($i) "/p"] [< [+ $i 1] $argc]} {
set opt_preview $true
set preview_handle $argv([+ $i 1])
win preview $preview_handle
set font_size 10
break
}
}
}
display $text
console show
win close
lib ticol_win -unload
See also: win, win create, plugins
Potential Plugins and Extensions |
Index |
Although time and resources are limited, the following plugins might be
nice to do: Source for an example plugin project might be released.
Also a sample project and which will link to a static LIB file
Logo A Logo like language plugin
Compiler A pseudo-compiler which can insert Ticol scripts into an
exectable binary
This would require Microsoft Visual C++ and use of the
Ticol LIB file
Plugin sample Source code for very basic plugin example
See also: plugins, ticol
Appendix - Windows Version Numbers |
Index |
Major version numbers as reported by [info winver -vernum]
3.1 Windows NT 3.1 Server and Workstation (3.10.528)
3.5 Windows NT 3.5 Server and Workstation (3.5.807)
3.51 Windows NT 3.51 Server and Workstation (3.51.1057)
3.51 Windows NT 3.51 Newshell Preview plus IE5 (4.0.1057)
4.0 Windows 95 (4.0.950)
4.0 Windows 95 OEM SR1 (95A) (4.00.950)
4.0 Windows 95 OEM SR2 (95B) (4.00.1111)
4.0 Windows 95 OEM SR2.1 (4.03.1212-1214)
4.0 Windows 95 OEM SR2.5C (4.03.1214)
4.0 Windows NT 4.0 Server and Workstation (4.0.1381)
4.10 Windows 98 (4.10.1998)
4.10 Windows 98 SE (4.10.2222)
4.90 Windows ME (4.90.3000)
5.0 Windows 2000 (5.00.2195)
5.0 Windows 2000 Professional (5.0.2195)
5.0 Windows 2000 Server (5.0.2195)
5.1 Windows XP RC1 (5.1.2505)
5.1 Windows XP (5.1.2600)
5.1 Windows XP SP1 (5.1.2600.1105-1106)
5.1 Windows XP SP2 (5.1.2600.2180)
5.1 Windows XP SP3 (5.1.2600)
5.2 Windows XP Pro x64 Edition (5.2.3790)
5.2 Windows Server 2003 (5.2.3790)) --
5.2 Windows Home Server (5.2.3790)
5.2 Windows Server 2003 SP1 (5.2.3790.1180)
5.2 Windows Server 2003R2
6.0 Windows Vista RTM (6.0.6000.16386)
6.0 Windows Vista (6.0.6000)
6.0 Windows Vista SP2 (6.0.6002)
6.0 Windows Server 2008 (6.0.6001)
6.1 Windows 7 RTM (6.1.7600.16385)
6.1 Windows 7 (6.1.7600)
6.1 Windows Sever 2008R2
6.1 Windows 7 SP1 (6.1.7601)
6.2 Windows 8 RTM (6.2.9200.16384)
6.2 Windows 8 (6.2.9200)
6.2 Windows Server 2012 (6.2.9200)
6.3 Windows 8.1 (6.3.9600.16384.130821-1623)
6.3 Windows Server 2012R2 (6.3.9600) (Misreports as 6.2)
See also: info winver