From Bugzilla Helper: User-Agent: Mozilla/5.0 (compatible; Konqueror/2.99 (3.0 rc3); Linux) Description of problem: current gnumeric is handling with 1char == 1byte. so that inputting the characters on cell is broken because multibytes character is also counted as 1byte. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1.Choose Japanese font like 'Kochi Gothic' 2.input Japanese characters via XIM on input box. 3. Actual Results: the characters only on cell is broken. Expected Results: cell and input box should be shown similar. Additional info:
Created attachment 49624 [details] screenshot
Created attachment 49625 [details] fix patch
This problem is already reported to upstream. See also http://bugzilla.gnome.org/show_bug.cgi?id=67073 This patch is included in that.
The patch leaks the return value of gtk_editable_get_chars()
Created attachment 49966 [details] Oops... please use this instead of.
Jody says this patch (the second version) and Nakai's input method patch on bugzilla.gnome.org are probably safe, if you guys tested them in both C/japanese/Latin-1 locales.
Jody also says that Gnumeric is NOT going to be usable in multibyte locales, even with these patches - bunches of other things will still be broken, though.
Gnumeric 1.0.x will be usable in CJK locale with this patch. It will not as usable as it in English, but the patch will improve it so much. And XIM update position patch will not break other single byte locale. You and Jody don't know how XIM support is important for CJK users. Imagine the sadness that you cannot input your mother language with your keyboard. Oops, sad...
Patch applied, 1.0.5-2