Re: Modifier keys

[Owen Taylor <otaylor redhat com>]

Tim Janik <timj gtk org> writes:

> On Fri, 1 Mar 2002, Owen Taylor wrote:
> 
> > 
> > Tim Janik <timj gtk org> writes:
> > 
> > > On Wed, 20 Feb 2002, Owen Taylor wrote:
> > > 
> > > > I'm planning to commit this patch tomorrow morning after I have
> > > > a night or so to think of any problems that it might have. If any
> > > > body wants to look this patch over or try it out, I'd certainly 
> > > > appriciate it.
> > > 
> > > seems you haven't comitted yet, right?
> > 
> > Committed over a week ago.
> 
> yeah, forget that, screwed my tree.
> 
> at this point i only got some minor questions,
> - didn't you claim you had to remove usage of signals in accel groups?

I did claim at one point, but decided later that it was easiest for
GtkWindow to do the conflict resolution and then simply activate the
right keyval/modifier combination in GtkAccelGroup, leaving
GtkAccelGroup more or less untouched

> - why are you storing "gtk-window-key-hash" per object data?
>   considering the key hash is created at latest once a key is pressed
>   within a window, it should be a struct member

Header file cleanliness, mostly. 

 - gtkkeyhash.h isn't public, so can't be included from gtkwindow.h
 - gtkkeyhash.h shouldn't be including gtkwindow.h; GtkKeyHash really
   is basically independent of the rest of Gtk.
 
So, I'd have to had 'gpointer key_hash'. Using object data kept
everything nice an encapsulated in one section of gtkwindow.h. Not
something I feel strongly about though.

> - why isn't _gtk_window_activate_key() simply calling gtk_window_get_key_hash() ?

Looks like I probably never finished some code reorganization. 

Regards,
                                        Owen