Autoit3Wrapper character encoding problem

[Jos]

It could very well be that there is still one in "C:\Program Files (x86)\AutoIt3\SciTE" and it all depends how you have installed.

In the past, that was the standard location which was changes to %SCITE_USERHOME% in case this exists. Simple way to figure out what is the correct one is using the Options\Open user Options file  in the SciTE Menu.

Jos

Edited August 22, 2015 by Jos

[cetipabo]

OK thank you very much !

for me the problem is solved.

Edited August 22, 2015 by cetipabo

[kyo]

Update:

When using ConsoleWrite("Effacé") -  I get this in the console: (darnit, can't paste what I see)

Looks like this, with the "xE9" on black background.

EffacxE9

Looks like the encoding is messing things again...

[Jos]

This is a known issue and can be "corrected" by this code from @jchd:

ConsoleWrite(BinaryToString(StringToBinary($s, 4), 1))

Jos

Edited August 22, 2015 by Jos

[kyo]

Ok thanks, I'll use this instead.

Sometimes I wish I could write all my software in english, it would be much simpler!

[jchd]

OK, it seems the dust has settled about the encoding issue. Warm thanks to Jos for being that responsive.

Guys, make the line of code Jos mentionned a function named _ConsoleWrite and use it whenever you expect non-ANSI chars, more precisely when you expect chars outside the system/user codepage you're using. Remember it's working for displaying Unicode chars provided you use the setting output.code.page=65001.

To sum up:

• either you leave output.code.page=0 and you're stuck to using std ConsoleWrite with ANSI-only chars (chars not in your codepage will display as ?)
• or you set output.code.page=65001 and then need to use _ConsoleWrite when displayed strings may contain chars > 0x7F. With this settings chars in [0x80..0xFF] display like blacked out xE9 with std ConsoleWrite. You may as well use _ConsoleWrite at all times without thinking.

Hope this clears things up. If you still experience issues, just speak up.

[Iczer]

another "UTF-8 - related" problem:

"NewFileEncoding=UTF8 + utf8.auto.check=4" work for "main script", but if you have includes with non-english characters, then all of them should be converted to UTF-8 also.

If not - compiled script will not work as expected - any non-english characters in menus/titles will not be recognised and autoit-written tray menus will go blank.

 

[Jos]

You will have to be more specific what you are doing and which utility is corrupting.
So what are you doing for it to break?

Jos

[Iczer]

if i'm not doing anything - it break. If i'm reencoding written by me includes from "code page" to "utf8" - everything OK.

if TrayCreateItem()  is placed in "code page" encoded include and have any non-english characters in "text" part then those characters will not be visible if you run compiled script.

Same with $iIndex = _GUICtrlMenu_FindItem ($hMenu, $sRequired_Text) - it will not find required text

 

 

 

[Jos]

Yes you are doing something and that something you need to tell me.

Are you running, compiling?
From SciTE or external?
Are you using tidy/au3stripper?
Need input ... all the gory details!!!!

Jos

Edited August 24, 2015 by Jos

[Iczer]

i'm running compiled script, externally, of course. Au3Stripper options:

#AutoIt3Wrapper_Run_Au3Stripper=y
#Au3Stripper_Parameters=/mo

 

[Jos]

I give up ....
Let me know when you are able to provide all details for me to be able to replicate else others can perform the miracle you are expecting.

Jos

EDIT: p.s. ensure you test with the available beta's as there are some fixes there  for include files and au3stripper.

Edited August 24, 2015 by Jos

[mLipok]

@Jos

I think this is somehow related to this:

But this was fixed .

btw.
I still have some other problems but actually after my vacation I'm litlle off. So I think I can find my issue in my case in some near feature.

 

[Allow2010]

I also noticed today that au3stripper breachs all my german "umlauts" like äöü when cimpiling scripts. Without au3tripper the exes work just fine..

I tried with rightclick ->compile eith options and the tried with and without au3stripper, no options for au3stripper used

i tried with release and beta autoit and with the latest scite beta files...

Edited August 26, 2015 by Allow2010

[Jos]

Folk,

I appreciate feedback but please remember I am sailing blind unless somebody shares the files causing the problem and tell me the exact versions they are using.

Jos

[Allow2010]

ok, i will prepare a sample...

[Allow2010]

Versions:

Autoit: v3.3.14.1

(as the problems comes up without the beta as well, i did install only final version)

Editor: 29-7-2015 and also latest beta files

[[ ]] Au3Stripper.exe 14-Aug-2015 13:14 137K  
[[ ]] AutoIt3Wrapper.au3 22-Aug-2015 22:10 264K  
[[TXT]] History.html 22-Aug-2015 22:14 104K  
[[ ]] SciLexer.dll 04-Aug-2015 16:34 765K  
[[ ]] SciTE.exe 04-Aug-2015 16:35 919K  
[[ ]] Tidy.exe 04-Aug-2015 16:25 148K  
[[TXT]] au3.properties 02-Aug-2015 15:56 16K  

use the attached au3 and right click. select compile with options, enable au3stripper, compile

the resulting stripped au3 has the errors, but the exe still works (not sure why, i also have problems with the exe with my real code)

i am still trying to prepare an example where the exe also displays the broken characters...

 

 

 

test.zip

Edited August 27, 2015 by Allow2010

[Allow2010]

finally i found the problem...

It is a very unique combination of things:

• The au3 is UTF8-DOS
• the include is DOS (no utf)
• both have umlaute (öäü...)
• the include is used in the au3

when you run it or compile it without au3stripper it works fine, when you use au3stripper it creates problems with the text

When i convert the include from DOS to U8-DOS (Ultraedit can do this) the problem is gone.

One more thing: When i open existing script, scite opens them different than it did in earlier version or some setting got lost:

File-Encoding is set to Code Page Property.

this results in displayerrors for my scripts (see first picture)

When i change to UTF-8 ist look OK, but i have to change it everytime i open the script...any idea on this? Never had this before...

example.zip

Edited August 27, 2015 by Allow2010

[Allow2010]

i tried to get rid of the problem by converting my au3 files to DOS ASCII instead of UTF-8.

This works fine (no au3stripper problems and not strange characters in scite), until i use rightclick "compile with options". This converts the au3 back to utf-8 and my problems start again...

Edited August 27, 2015 by Allow2010

[Jos]

So this is happening when this combination of files is merged by au3stripper?
Au3stripper currently recognises the file encoding when it has a BOM, else it simply merges the files together in a file with a file encoding used by the master script.
Wondering If I need to do any conversions along the way in case of different encoded files.

Jos