Re: XIM key interpretation fix for 2.2.x?

[Takuro Ashie <ashie homa ne jp>]

Hi all.

I'm also considering about this problem.
I don't have complete solutions yet, but at least I agree with
Tokunaga-san's opinion mostoly.

At Fri, 5 Sep 2003 04:37:40 +0900,
TOKUNAGA Hiroyuki wrote:
>  I agree with you that IM want keystrokes only when inputting. So active im
> context variable should set to NULL when focused text input widget loses
> the focus, I think.

To determine whether the state is inputing or not, I think input
widgets should have a vertual method like "get_im_context" or
"get_active_im_context".

If input widgets has such a method, toplevel doesn't always need to
have im context variable. Instead, we can determine the state by
querying to focused widget through such method.

Here is the sample patch:

  http://www.homa.ne.jp/~ashie/linux/files/gtk+-2.2.4-get_im.diff

But, I'm not sure such a change is appropriate or not.

Regards,

--
Takuro Ashie
Mail: ashie homa ne jp
Web:  http://www.homa.ne.jp/~ashie/