Array error in compiled version but NEVER when run from editor

[Jdop]

Getting that cryptic array pointer error with a -1 line id (apparently from a library call ) in COMPILED but I never get the error when I run the same code from the editor. Seems odd, and since its fairly long and complex code, hard to debug.

Would be nice if the libraries actually passed the reason for crashing.

How to deal with this?

[JohnOne]

I cannot help you, but others who may be able to will probably be able to do so easier if you give a full error you are getting, and version of autoit, and scite.

using latest versions of both, I always get an error with a line number from compiled script.

[trancexx]

Update your AutoIt first, then ask.

[PsaltyDS]

The message is not cryptic, you made an invalid array reference. Review the array references in your script. Put in a few _FileWriteLog(), TrayTip(), or ToolTip() debug statements in your script to see how far it got and what it was doing at the time.

If the script is too "long and complex" for you, then it's time to climb a little higher on the learning curve.

Don't give up, you can figure it out.

[kaotkbliss]

If I know approximately where my script went wrong, I like to add msgbox for each time a variable is called or changed in that part of the script. That helps me find where the problem is

[Jdop]

I updated to the latest version of Autoit. Lets see if that cures the problem.

Whats odd is , is never crashes when run in 'edit mode' only the compiled version. In theory that should not happen. The program loads live data from the internet so its constantly changing and therefore hard to debug if there's a random error.

[martin]

I updated to the latest version of Autoit. Lets see if that cures the problem.

Whats odd is , is never crashes when run in 'edit mode' only the compiled version. In theory that should not happen. The program loads live data from the internet so its constantly changing and therefore hard to debug if there's a random error.

You can possibly find the line number if you compile the script as converted by the code below.

NB This was only written for 1D arrays and so would need changing for more dimensions, but the basic idea is there.

#include <file.au3>

$PresentScriptname = "tproblemscript.au3";<------------put your file name here
$DebugScriptName = "debugfile.au3";<--------put the name of the copy here
$ArrayName = "$brray";<-----------put your array variable name here


$Newf = FileOpen($DebugScriptName, 2)
Global $text
_FileReadToArray($PresentScriptname, $text)
For $n = 1 To $text[0]
    If StringLeft(StringStripWS($text[$n], 1), 3) <> "Dim" Then
        
        $sl = StringRegExp($text[$n], "\" & $ArrayName & "\[.*?\]",3)
        if isarray($sl) then 
            for $k = 0 to ubound($sl) - 1
                FileWriteLine($Newf,StringRegExpReplace($sl[$k], ".*(\" & $ArrayName & "\[)(.*)?(\])", "CheckArray(\" & $ArrayName & ",\2," & $n & ")",1) )
            Next
        EndIf
        
    EndIf
    FileWriteLine($Newf,$text[$n])
    
Next
FileWriteLine($Newf, @CRLF & "Func CheckArray($aA,$iS, $iLn)")
FileWriteLine($Newf, @TAB & "if UBound($aA) < $iS + 1 then ")
FileWriteLine($Newf, @TAB & @TAB & 'MsgBox(262144,"ERROR on line" & $iLN, "Array has " & UBound($aA) & " elements," & @CRLF & "but " & $iS & " called for")')
FileWriteLine($Newf, @TAB & @TAB & 'exit')
FileWriteLine($Newf, @TAB & 'Else')
FileWriteLine($Newf, @TAB & @TAB & 'Return $aA[$iS]')
FileWriteLine($Newf, @TAB & 'EndIf')
FileWriteLine($Newf, 'EndFunc')

FileClose($Newf)

Note that the line number reported with the error is the line number in the original script not the debug script.

I have only tried it on a simple example so it probably needs some improvements.

EDIT: Changed to allow for the Array variable appearing more than once on a line.

But really you should write your script so that the error doesn't occur, but since you haven't shown any code we can only guess.

Edited March 7, 2010 by martin

[Jdop]

I'm getting the same error, only now it reports Error at line 13333. Obviously this is not much more helpful than line -1

[Melba23]

Jdop,

Obviously this is not much more helpful than line -1

Why not? It tells you exactly where the error is, which is why trancexx told you to upgrade your Autoit version.

Now your problem is to work out which line of code 13333 actually is - as the compiler strips out all comments and blanks. Did you run the code through Obfuscator when compiling? If you had, you would find a file "filename_Obfuscated.au3" in the same folder as your script which is what the compiler actually used to produce the final script. If you had used the /StripOnly parameter with Obfuscator you would have automatically stripped all the comment and blank lines, so the line given in the error message would refer directly to that line in the Obfuscated file. And then you would know where the problem lies!

If want to try this, you need the full SciTE4AutoIt3 package, which you can download from here. You get lots of other goodies as well, so it is well worth it.

M23

[martin]

Jdop,

Why not? It tells you exactly where the error is, which is why trancexx told you to upgrade your Autoit version.

Now your problem is to work out which line of code 13333 actually is - as the compiler strips out all comments and blanks. Did you run the code through Obfuscator when compiling? If you had, you would find a file "filename_Obfuscated.au3" in the same folder as your script which is what the compiler actually used to produce the final script. If you had used the /StripOnly parameter with Obfuscator you would have automatically stripped all the comment and blank lines, so the line given in the error message would refer directly to that line in the Obfuscated file. And then you would know where the problem lies!

If want to try this, you need the full SciTE4AutoIt3 package, which you can download from here. You get lots of other goodies as well, so it is well worth it.

M23

The code I provided gives the line in the original script. So line 1333 is just that. It also tells you the index and the array UBound. You can eassily extend the code I gave to tell you more, for example the values of other variables.

EDIT: I suppose I have to assume the OP has not actually tried my code.

Edited March 8, 2010 by martin

[Jdop]

I'm getting pointers to line numbers that are not array related, even though the error message says "array error". Is there some 'wiggle room' in the line number thats reported in the obfuscated code.

[trancexx]

Are you using /striponly only?

[Jdop]

Yes, I am. I've gotten some line pointers that make sense, and some that don't. I'm working on it.

[MariusN]

I usually get that error when i forget to declare my variable...

[Jdop]

I usually get that error when i forget to declare my variable...

Are you saying you get a bogus error or line number when you use an undeclared var?

The var on the error I am getting was undeclared, so I added it to the DIM's. Let's see if that cures it.

Wondering if that 'declaration' issue would cause a difference in compiled vs. editor operation, as I never seem to get that error running off Scite.

Edited March 13, 2010 by Jdop

[MvGulik]

whatever Edited February 7, 2011 by MvGulik