Re: Gtk+ Core TODO Item

[Tim Janik <timj gtk org>]

On Mon, 25 Sep 2000, Eric Lemings wrote:

> Tim Janik wrote:
> > 
> > 2) will mostly amount to changes similar to:
> > 
> >  void
> >  gtk_foo_set_barstring (GtkFoo      *foo,
> >                         const gchar *string)
> >  {
> >    g_return_if_fail (GTK_IS_FOO (foo));
> > 
> >    g_free (foo->barstring);
> >    foo->barstring = g_strdup (string);
> > +  g_object_queue_param_changed (G_OBJECT (foo));
> >  }
> 
> g_object_queue_param_changed currently only notifies the object being
> changed.  Other objects may also be interested/dependent on this
> change.  This would most likely require emitting a signal.  Will this
> capability go into GObject?  If so, how will it be implemented?

queue_param_changed doesn't _notify_ the object, but adds the
pending notification to a queue. the queue will then later be processed
by an idle handler to deliver actuall notification to all entities that
announced interest in it (listeners).
as for the actuall API, yes, we will use signals for this.

> capability go into GObject?  If so, how will it be implemented?

the actuall implementation will tell you soon as enough ;)
(i plan on getting that into CVS HEAD as soon as i
return to germany and get some time to do actuall
coding which should be around end of this/next week)
> 
> Eric.
> 
> 

---
ciaoTJ