> > If such change is possible, solving the bug will be easy. But I'm not > sure this is good appropirate, too... > Just to keep the discussion going, this patch is surely one way to solve it, but i think it has some disadvantages. - it has to be made clear to a person using a im context that they need to override this added method. An ideal solution should not require a creator of a widget that uses an IM context to have to implement a separate method so that the IM will get keystrokes, it should happen automatically - a bug in behaviour should not reuire an API change to fix, or at least it should not require apps using the library to change their code to fix. - if we add this function, widgets that use an im context but not nec. included in the gtk-2.4 release will not be fixed. They all have to add their own get_im_context-like overriding method I think the spirit of the original snooper patch is more appropriate. The IMContext should be modified so that when it is or becomes active, it tells someone "i need keystrokes", and the keystroke handling mechanism should be modified so that when it knows a particular context needs keystrokes, they will get there. This ensures that people who make their own widgets and use an IM context, will get proper behaviour, without having to override any methods. ------ +----------------------------------------------+ ( Ken Deeter (Kentarou SHINOHARA) ( ) ) ( "If only God were alive to see this.. " ( ) -Homer Simpson ) +----------------------------------------------+ _______________________________________________ gtk-list@xxxxxxxxx http://mail.gnome.org/mailman/listinfo/gtk-list