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

[Sylvestre Ledru]

Hi,
Thanks but I haven't received your previous email :|

If you checkout the source from the svn, could you do:
# svn add your-new-files
# svn diff > for-sylvester.diff
=> mail it to me.

Thanks
Sylvestre

Le mardi 03 juin 2008 à 20:33 +0800, Yung-Jang Lee a écrit :
> Sorry,
> 
> The files I send have two bugs due to copy and paste mistake, 
> 
> ----------localetoutf.c-----------------------------------
> BOOL  openLocaleToUTFConverter(char *sysLocale,char *lang)
> {
>       // Assume sysLocale in lang_contry.charset or  lang_contry
> format
>       //
>       // Ex. en_US, zh_TW.CP950, zh_CN.UTF-8
>       //
>       char encoding[5];
>       char * pch=strchr ( sysLocale,
> '.');//<------------------------------- this line missing
>       if(pch){    
> -----------------------------------------------------------------
> 
> and 
> --------------------getLocaleInfo_Windows.c-----------------------------
>             // BY YJLee get Windows CODE Page Information\ <----
> should delete last  character 
>             int ret = GetLocaleInfo(LOCALE_USER_DEFAULT,
>                         LOCALE_IDEFAULTANSICODEPAGE,
> -----------------------------------------------------------------------
> 
> Please check it.
> 
> YungLee
> 
> 2008/6/2 Yung-Jang Lee <yjlee123 at gmail.com>:
>         Sylvestre,
>         
>         Enclosed is my small patch for Java console to display chinese
>         characters and the result PNG screenshot. 
>         
>         The test in under Windows XP, with BIG5 encoding (Windows
>         CodePage 950).
>         
>         This test confirm that it need a locale to utf-8 convert
>         before send string to Java console for locale not compatible
>         with  UNICODE.
>         
>         The patch contains following files
>         
>              1.    localization\src\c :   localetoutf.h,
>                 localetoutf.c
>                        provide functions
>                     void localeToUTF(char** buffer) <--- doing locale
>                 to utf convert 
>                     openLocaleToUTFConverter(char *sysLocale,char
>                 *lang) <--- open or close converter
>                         This two functions calls functions from iconv
>                 library which is contained in Scilab5  prerequreiment.
>                 
>              2. localization\src\c :   setlanguage.c, 
>                  add a line to call openLocaleToUTFConverter ,
>                 
>              3. localization\src\c :  getLocaleInfo_Windows.c, 
>                 modified return value of getLocaleUserInfo in
>                 lang_contry.codepage format t,codepage  information is
>                 used inside openLocaleToUTFConverter 
>                 
>              4. console\src\cpp: consolePrint.c : 
>                 add a line to call localeToUTF
>         
>         This is just a small patch to test locale to utf encoding
>         concept, it need further enhancement.
>         
>         I have two questions. First , in my  chinese Java console ,
>         the menu items are disabled ( unlike your Linux case)  is
>         this  a  bug under Windows ? Second, menu items are displayed
>         in chinese under Linux but in english  under Windows, what
>         make them different ? 
>         
>         
>         YungLee
>         
>         2008/5/30 Yung-Jang Lee <yjlee123 at gmail.com>:
>         
>         
>                 Sylvestre:
>                 
>                 Let me try. Give me a week.
>                 
>                 YungLee
>                 
>                 2008/5/30 Sylvestre Ledru <sylvestre.ledru at inria.fr>:
>                 
>                 
>                         
>                         > To solve this problem, we must decide locale
>                         setting in runtime  and
>                         > transform  characters  back  to  UTF8  if
>                          the locale setting  is not
>                         > in UTF8.
>                         
>                         OK, then, a way of doing this could be:
>                         * if the LANG/LC_MESSAGES/whatever variable
>                         says that the language is
>                         UTF-8, don't do anything.
>                         * if not, convert it
>                         However, that sounds like a workaround to me,
>                         not a really bug fix.
>                         There must be a way to deal with this.
>                         
>                         > Hope these can help.
>                         It does!
>                         
>                         S
>                         
>                         
>                 
>         
>