Re: balsa internationalization

[ymnk jcraft com (Atsuhiko Yamanaka)]

Hi,

  |On Thu, 21 Sep 2000 19:23:38 +0200
  |Pawel Salek <pawsa@TheoChem.kth.se> wrote:
  |
  |I attach the patch that hopefully solves the internationalization issue
  |completely. It uses setting languages and fontsets. 
  |
  |THIS IS IMPORTANT, it needs to be tested for:
  |b. languages other than English, Polish and Swedish (I have tested these
  |and seem to be ok). Multibyte charsets are particularly interesting.

I tried this patch for reading/writing Japanese email.

At first, I got next error message,
message/text:: font not found: -*-fixed-medium-r-normal-*-*-*-*-*-c-*iso-2022-jp
It seems a wrong XLFD font name has been generated.
If I added one line as follows, that warning message disappeared.

  static struct {
    gchar *charset, *font_postfix;
  } charset2font[] = {
    {"iso-8859-1", "iso8859-1"},
    {"iso-8859-2", "iso8859-2"},
    {"iso-8859-7", "iso8859-7"},
+   {"iso-2022-jp", "jisx0208.1983-0"},
    {"euc-jp",     "jisx0208.1983-0"},
    {"euc-kr",     "jisx0208.1983-0"},
    {"koi-8-r",    "koi8-r"},
    {"koi-8-u",    "koi8-u"},
    {"us-ascii",   "iso8859-1"}
  };


However, Japanese characters were not displayed on balsa.
I guess there are two problems.

1. I guess that 'gdk_fontset_load' must be used to display 
   Japanese characters (and other mutlibyte charsets?) on Gtk widget.
   For example,
     gdk_fontset_load("-*-fixed-medium-r-normal--14-*-*-*-*-*-*-*");

2. To display Japanese characters on Gtk widget, 
   character codes must be converted before insering text to widgets.
   For example,
     iconv_t cd=iconv_open("EUC-JP", "ISO-2022-JP");
     iconv(cd, (const char **)&in, &insize, &out, &outsize);
   I guess code-conversions are also need for 
   other multibyte character e-mails;Chinese, Korean, etc.

Anyway, thank you for awesome hack!

Regards,
--
ymnk, JCraft Inc., mailto: ymnk@jcraft.com