From: Botond B. <bo...@si...> - 2003-06-26 14:45:40
|
On Thu, 26 Jun 2003 18:33:54 +0900 (KST) 酵走昔 <mit...@dr...> wrote: > >But I doesn't mean this signal (or a nicer solution) cannot be > >implemented. > >For example when gaim gets the Enter keyevent, it should pass it > >through gtk_im_context_filter_keypress. If this returns FALSE, it > >shouln't send the message. I haven't looked at gaim code yet though. > gaim use key_press_event to process various keyinput include enter. so > getting text occur before call gtk_im_context_filter_keypress(). > and I have no good idea to make process enterkey after call > gtk_im_context_filter_key_press() Ok, here it is what I was talking about. Patch is attached. I tested it and it works ok. IMHO your patch is not quite correct, because gtk_im_context_reset() was made for a diferent purpose. > however (of course you know,) this is not only problem of gaim but all > similiar messenger or etc.. which use text_view widget and all > input-method which has preedit string. Once they will have to realize that there are users who want to input CJK. And there are a lot of 'em! ;) ps: I'm CC-ing the gaim developers, so that they can integrate this patch. |