[Scilab-Dev] Test report of Chinese (or multibytes languages in general) locale mesages with Scilab 5-beta-1

[Yung-Jang Lee]

Also,

Under Windows, it's not as easy as Linux to switch locale just by setting
LANG.  And most  Chinese users use Scilab under Windows,   so I must face to
the problem not just workaround.

YungLee

2008/5/29 Sylvestre Ledru <sylvestre.ledru at inria.fr>:

> Hello Yung-Jang,
>
> I haven't forget you. Just busy on other stuff.
>
> Note that I am working under Linux Debian with only-required-packages
> (understand: I am not sure to have all the font for the traditionnal
> chinese language).
>
> After playing with internal & external bugs, I have been able to
> reproduce your bug and to fix it (you can call it a workaround)
>
> When I launch this way:
> # LANG=zh_TW ./bin/scilab
> I get "you-failed.jpg"
>
> but when I launched it with:
> # LANG=zh_TW.UTF-8 ./bin/scilab
> I get "Working.jpg"
>
> Does it mean anything to you ?
> (for those who don't know, Java is full Unicode)
>
> Is there zh_TW charset which would not be UTF-8 ?
>
> Sylvestre
>
>
>
> > Hi, everybody :
> >
> > I am just beginning to  translate message PO files (in Scilab
> > 5-beta-1) into Chinese language (zh_TW),  and hope that will meet the
> > next release schedule of Scilab 5.
> >
> > To test my translation (only few PO file right now), I generated
> > scilab.mo and scilab.po files in loclate\zh_TW\LC_MESSAGES with
> > script  generatePoFile in directory  toos\localization after setting
> > language to 'zh_TW'.
> >
> > The attached files are screehshots  after startup  of Scilab  showing
> > banner lines of my Chinese messages by   scilex.exe and Wscilex.exe
> > respectively in Windows-XP.
> >
> > The first screehshot (scilexChinese.png generated by scilex.exe )
> > correctly displays Chinese characters , while the second  one
> > (wscilexChinese.png by Wscilex.exe) fails.
> >
> > My Windows-XP locale setting is zh_TW.BIG5 , while  messages  in
> > core.po file  are in UTF-8 encoding.
> >
> > It appears to me that  Windows  will transform messages back to  its
> > default encoding (in my case, BIG5) and DOS console (scilex) handle
> > this correctly.  But Java console expect an UTF-8  messages and the
> > received  BIG5 encoding messages cause its failure in showing Chinese
> > banner lines.
> >
> > I think this is not a specific problem associate with Chinese
> > language, any multibytes language with  locale setting not in UTF-8
> > will probable  fail in the same way.
> >
> > I try to trace the source codes of Java console module to find a way
> > to attack this problem.  Transform messages back to UTF-8 in
> > ConsolePrint.cpp ( module console) may help , for example :
> >
> > int ConsolePrintf(char *line)
> > {
> >   lineUtf8=localeToUtf8(line); // <------ transform back to utf8
> >   CallScilabBridge::display(getScilabJavaVM(), line);
> >   return 0;
> > }
> >
> > In this way, we need a function 'localeToUtf8'  which transform output
> > strings from locale to UTF-8 encoding. This function is provided in
> > Glib and may be implemented form iconv API within current Scialb
> > developement environment.
> >
> > Hope my test can help and looking forward for suggestions or
> > instructions to attack this problems.
> >
> > Best regards,
> >
> > YungLee
> >
> >
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.scilab.org/pipermail/dev/attachments/20080530/a09b0101/attachment.htm>