General Tips for using AutoIt

[GEOSoft]

Based on Valuaters excellent AutoIt Wrappers thread (Still not a sticky??) I'm going to attempt something similar here with examples that cover things that you can do to avoid problems when you are using AutoIt. It will include both GUI and non-GUI tips. Please feel free to add your own but please make sure they are accurate. If not you will be getting a PM to either Edit or Remove the reply.

If you see any errors in my replies, please PM me so it can be corrected.

If our supreme commander or any of The mods fell this thread should be removed or moved then please do so and PM me to let me know.

Edited May 9, 2009 by GEOSoft

[Qousio]

I'm not exactly super experienced with Autoit; however I am quite experienced in structured programming languages like C++ and Pascal.

Since Autoit is a structured language, it shares allot of common things with other languages, so learning the basics of structured programming will help you be good(better) at Autoit.

I would suggest all new Autoit users to start learning loops, variables and arrays then moving on to things like pixelsearch and controlsend. I just had to say this, because there’s tons of threads like these every day, when a person doesn't understand how a loop works and tries to make a valid script.

Another great tip to learning Autoit is using the helpfile, it has detailed explanations on the syntax and every function in Autoit. If you want to send a key, try typing "Send" in the help search. It will show Function Reference to any commands related to send.

I hope these simple tips will help

Edited May 9, 2009 by Qousio

[GEOSoft]

How to avoid GUI flicker issues in loops.

When setting the control state always check the current state first.

If BitAnd(GUICtrlGetState($hMyControl), $GUI_DISABLE) Then GUICtrlSetState($hMyControl, $GUI_ENABLE)
;

When setting control data

In this case I want to make sure the control is always lower case during the message loop.

This will cause flicker every time the string is checked to see if it's lower case when the control contains punctuation or whitespace.

;
$sString = GUICtrlRead($hMyControl)
If NOT StringisLower($sString ) Then GUICtrlSetData($hMyControl, StringLower($sString ))
;

This won't

;
$sString = GUICtrlRead($hMyControl)
If StringRegExp($sString ,"[[:upper:]]") Then GUICtrlSetData($hMyControl, StringLower($sString ))

For the same thing using upper case it would be

$sString = GUICtrlRead($hMyControl)
If StringRegExp($sString ,"[[:lower:]]") Then GUICtrlSetData($hMyControl, StringUpper($sString ))
;

Here is a script which will demonstrate the lower case scenario

;
$frm = GUICreate("Flicker Demo", 170, 70)
$in1 = GUICtrlCreateInput("TYPE SOME TEXT HERE", 10, 10, 150, 20)
$in2 = GUICtrlCreateInput("Now TyPe Some Text Here", 10, 40, 150, 20)
GUISetState()

While 1
   If GUIGetMsg() = -3 Then Exit
   $str1 = GUICtrlRead($in1)
   $str2 = GUICtrlRead($in2)
   If NOT StringIsLower($str1) Then GUICtrlSetData($in1, StringLower($str1))
   If StringRegExp($str2,"[[:upper:]]") Then GUICtrlSetData($in2, StringLower($str2))
Wend
;

Thanks to SmOke_N for suggesting a slight change to the RegExp

Edited July 31, 2009 by GEOSoft

[GEOSoft]

Using a Regular expression to match a unicode character.

The only way I have found so far is using hex values (\xnn) unfortunatly that is limited to 2 hex characters and unicode requires up to 4, so after some experimenting I discovered that you can still do it if you use curly brackets

\x{nnnn}

To use the full range of Unicode characters for ChrW(256) on upwards use the range as below

[\x{0100}-\x{FFFF}]

Enjoy

[jchd]

Using a Regular expression to match a unicode character.

The only way I have found so far is using hex values (\xnn) unfortunatly that is limited to 2 hex characters and unicode requires up to 4, so after some experimenting I discovered that you can still do it if you use curly brackets

\x{nnnn}

To use the full range of Unicode characters for ChrW(256) on upwards use the range as below

[\x{0100}-\x{FFFF}]

Thank you for sharing this useful trick.

Just a point: the 0x0100..0xFFFF is just a part of Unicode, albeit by far the most used part. Unicode has been specifying blocks well above 0xFFFF many years ago, in the post-UCS-2 age (Unicode version > 1.1).

The full list of allocated blocks is here. Unicode will never exceed the 0x10FFFF upper bound.

I don't believe that common AutoIt users would need to routinely handle highly rare blocks like 'Byzantine Musical Symbols' or 'Old Persian', whose code blocks are beyond 0xFFFF.

Issues could arise with the use of some "high" private areas (0xF0000..0xFFFFD and 0x100000..0x10FFFD) for ad hoc purpose by third-party applications. There are more than 137,000 codepoints for such private use.

You gave me the incentive to try > 0xFFFF codepoints. The good news is that the \x{nnnnn} trick works in the pattern matching part of AutoIt regexes even when nnnnn > FFFF. Still, built-in functions like StringLen see 2 separate characters instead of 1. Indeed the internal (UTF-16) representation uses two 16-bit words, but they code for just one single Unicode character.

This is something that concerned [advanced] users should be aware of.

[GEOSoft]

Thank you for sharing this useful trick.

Just a point: the 0x0100..0xFFFF is just a part of Unicode, albeit by far the most used part. Unicode has been specifying blocks well above 0xFFFF many years ago, in the post-UCS-2 age (Unicode version > 1.1).

The full list of allocated blocks is here. Unicode will never exceed the 0x10FFFF upper bound.

I don't believe that common AutoIt users would need to routinely handle highly rare blocks like 'Byzantine Musical Symbols' or 'Old Persian', whose code blocks are beyond 0xFFFF.

Issues could arise with the use of some "high" private areas (0xF0000..0xFFFFD and 0x100000..0x10FFFD) for ad hoc purpose by third-party applications. There are more than 137,000 codepoints for such private use.

You gave me the incentive to try > 0xFFFF codepoints. The good news is that the \x{nnnnn} trick works in the pattern matching part of AutoIt regexes even when nnnnn > FFFF. Still, built-in functions like StringLen see 2 separate characters instead of 1. Indeed the internal (UTF-16) representation uses two 16-bit words, but they code for just one single Unicode character.

This is something that concerned [advanced] users should be aware of.

Thanks for that Info. I'll look into it and possibly change the regex a bit. For now I think it's covered for normal usage.

[herewasplato]

[autoit]; Avoid hogging the CPU - use the Sleep function

; Avoid hitting the recursion limit

; If you must call a function from within itself

; then either add a check that gracefully handles

; reaching the recursion limit or your script will

; exit as illustrated in the code below.

Global $i = 0

_test_recursion()

Func _test_recursion()

SplashTextOn("", $i, 60, 40, Default, Default, 17)

;~ Sleep(9) ; <<< Avoid hogging the CPU - use the Sleep function

$i += 1

If $i > 4707 Then Sleep(2000)

_test_recursion()

EndFunc ;==>_test_recursion

; Recursion is useful and needed at times

; See _PathFull

Edited July 21, 2009 by Jon

[CodyBarrett]

here is a little thing about array's for new users

UBOUND() returns the MAX amount of slots available in an array... for Ex

Dim $array1[3]
$array1[0]=''
$array1[1]=''
$array1[2]=''
Ubound ($array) ;Will = 3 because this array is ZERO based

and alittle bit about GUI's

KODA helps alot for new users and advanced users alike, personaly i don't use it i do it from scratch, its all on preference

but obviously for experienced scripters... and not so obvious for nonscripters

NEVER have Guicreate() in an uncontroled loop because that will open endless windows haha its happened to me

[MattyD]

Just like to add Koda (for designing forms) is in the standard install for autoIT and can be found in C:\Program Files\AutoIt3\SciTE\Koda by default - can be tricky to find

Also to talk to a cmd prompt check out the Named Pipe examples that also come bundled with autoIt.

Thirdly a very fast way learn is to look at other peoples scripts - even the standard include UDFs. Most people should find better ways of organising code in some part of their own projects.

[Richard Robertson]

; Avoid hogging the CPU - use the Sleep function

; Avoid hitting the recursion limit
; If you must call a function from within itself
; then either add a check that gracefully handles
; reaching the recursion limit or your script will
; exit as illustrated in the code below.

Global $i = 0
_test_recursion()

Func _test_recursion()
    SplashTextOn("", $i, 60, 40, Default, Default, 17)
;~  Sleep(9) ; <<< Avoid hogging the CPU - use the Sleep function
    $i += 1
    If $i > 4707 Then Sleep(2000)
    _test_recursion()
EndFunc   ;==>_test_recursion

; Recursion is useful and needed at times
; See _PathFull

That code will result in an infinite recursion.

Edit: Just realized that was the point...

Edited May 16, 2009 by Richard Robertson

[GEOSoft]

How do I get the filename only from a path? Easy

$sFile = "C:\some\path\somefile.au3"
$sFile = StringRegExpReplace($sFile, "^.*\\(.*)$", "$1")
MsgBox (0,'',$sFile)

[MDCT]

How do I get the filename only from a path? Easy

$sFile = "C:\some\path\somefile.au3"
$sFile = StringRegExpReplace($sFile, "^.*\\(.*)$", "$1")
MsgBox (0,'',$sFile)

Additional information regarding path manipulation using StringRegExpReplace:

expandcollapse popup

;
Local $sFile = "C:\Program Files\Another Dir\AutoIt3\AutoIt3.chm"

; Drive letter -                                     Example returns     "C"
Local $sDrive = StringRegExpReplace($sFile, ":.*$", "")

; Full Path with backslash -                         Example returns     "C:\Program Files\Another Dir\AutoIt3\"
Local $sPath = StringRegExpReplace($sFile, "(^.*\\)(.*)", "\1")

; Full Path without backslash -                     Example returns     "C:\Program Files\Another Dir\AutoIt3"
Local $sPathExDr = StringRegExpReplace($sFile, "(^.:)(\\.*\\)(.*$)", "\2")

; Full Path w/0 backslashes, nor drive letter -     Example returns     "\Program Files\Another Dir\AutoIt3\"
Local $sPathExDrBSs = StringRegExpReplace($sFile, "(^.:\\)(.*)(\\.*$)", "\2")

; Path w/o backslash, not drive letter: -             Example returns     "Program Files\Another Dir\AutoIt3"
Local $sPathExBS = StringRegExpReplace($sFile, "(^.*)\\(.*)", "\1")

; File name w/0 ext -                                 Example returns     "AutoIt3"
Local $sFilenameExExt = StringRegExpReplace($sFile, "^.*\\|\..*$", "")

; Dot Extension -                                     Example returns     ".chm"
Local $sDotExt = StringRegExpReplace($sFile, "^.*\.", ".$1")

; Extension -                                         Example returns     "chm"
Local $sExt = StringRegExpReplace($sFile, "^.*\.", "")

MsgBox(0, "Path File Name Parts", _
        "Drive             " & @TAB & $sDrive & @CRLF & _
        "Path              " & @TAB & $sPath & @CRLF & _
        "Path w/o backslash" & @TAB & $sPathExBS & @CRLF & _
        "Path w/o Drv:     " & @TAB & $sPathExDr & @CRLF & _
        "Path w/o Drv or \'s" & @TAB & $sPathExDrBSs & @CRLF & __
        "File Name w/o Ext " & @TAB & $sFilenameExExt & @CRLF & _
        "Dot Extension     " & @TAB & $sDotExt & @CRLF & _
        "Extension         " & @TAB & $sExt & @CRLF)
;

Cannot remember the author name, but it can come in handy sometimes.

[GEOSoft]

Q: How do I make sure a folder path ends in a backslash ("\")?

A: Standard method

$sPath = "C:\Some\folder"
If StringRight($sPath, 1) <> "\" Then $sPath &= "\"

Better yet is this one

$sPath = "C:\Some\Folder"
$sPath = StringRegExpReplace($sPath, "(?i)^([a-z]:.*?)(?:\\)*$", "$1\\")

[Mat]

Examples and descriptions of common autoit errors.

http://code.google.com/p/m-a-t/wiki/commanerror

This was posted by me earlier in the Chat section, I've just updated it into my wiki.

Mat

problem was with case sensitivity in the link. It is now fixed and the below post was deleted to save on crap accumulation.

Edited November 14, 2009 by Mat

[AdmiralAlkex]

Examples and descriptions of common autoit errors.

http://code.google.com/p/m-a-t/wiki/Commonerror

This was posted by me earlier in the Chat section, I've just updated it into my wiki.

Mat

That page gives me an error!

Page "Commonerror" Not Found

[GEOSoft]

Problem: You attemt to write a string of unicode characters to an ini file using

$sStr = "Αυτό είναι μια δοκιμή"
IniWrite($sIni, "Test", String", $sStr)

And when you open the ini file to view it the string is simply a bunch of characters including a lot of "?"s. Ini files don't play nice with strings containing characters > Chr(255).

Solution: Write it to the ini as binary. I used this method with great success in a language translator.

$sStr = StringToBinary("Αυτό είναι μια δοκιμή")
IniWrite($sIni, "Test", String", $sStr)

Read it with

$sRtn = IniRead($sIni, "Test", "String", "")
$sRtn = BinaryToString($sRtn)

Edited February 15, 2010 by GEOSoft