Re: imhangul is not correct..

[tee kuis kyoto-u ac jp (Yusuke Tabata)]

cwryu debian org writes:
>> It will be useable if it handles backspace properly and
>> 3-bulsik Hangul keyboard type (hangul3.el for Emacs).
I am preparing fix, but it can not be well tested before gtk+-2.0.
So I entered a bug at
 http://bugzilla.gnome.org/show_bug.cgi?id=72817
 


>> > You can toggle input method by shift+space key or option menu.
>> Is it an immodule's duty to handle this state?  
To write a base class which handles input state is in my future plan.
However, in current gtk+ you should switch between native input
method and defalut input method(for latin chars)
by mouse click and it is tedious for many people.

>> 
>> Hmm..  Then, IMO XIM-based input module, like the Japanese one, should
>> be the best solution for Korean.
Yes. Generally, native input method system is the best for native people.
It is well tuned, documented, maintained...
; In Japan the situation is different, kinput2(Japanese XIM server) sucks.
But gtk+ is a cross platform and internationailzed toolkit, 
so I think having gtk+'s own input method system is worthy.

>>  Usually all input entries in one
>> toplevel window (sometimes in one program or in the whole desktop)
>> share one Hangul toggle state.  And most users still prefer to see the
>> Hangul toggle state from the panel, rather than besides the input
>> entry.
Hmm, sharing IM contexts in a process is trivial,
mozilla does it in XIM framework.
But whole desktop is difficult... Does it needed?

--
Yusuke Tabata
Graduate School of Informatics, Kyoto University
 tee kuis kyoto-u ac jp