Re: Proposal: gtk_main() quit callbacks

[Tim Janik <timj gimp org>]

On 28 Feb 1998, Owen Taylor wrote:

> 
> Tim Janik <timj@gimp.org> writes:
> 
> > hello everyone,
> > 
> > i'd like to propose the addition of a gtk_main() quit callback
> > mechanism like so:
> > 
> > gint       gtk_quit_add            (guint	       main_level,
> > 				    GtkFunction        function,
> >                                     gpointer           data);
> > gint       gtk_quit_add_full       (guint	       main_level,
> > 				    GtkFunction        function,
> >                                     GtkCallbackMarshal marshal,
> >                                     gpointer           data,
> >                                     GtkDestroyNotify   destroy);
> > 
> > function will be called when gtk_main () is left if main_level ==
> > gtk_main_level() or main_level==0.
> 
> Hmmm... the main loop I use for ILU never calls gtk_main()
> at all. (It has its own gtk_main() where gtk_main_quit() indicates
> the level). gtk-doom probably never touches gtk_main() at all
> either. 
> 
> So you should at least provide an alternate method to exit out
> of GLE, if people want to use it for such programs. (And
> a way of disabling the quit_add(), if somebody wants to 
> pop up a dialog and call gtk_main() in their doom-equivalent)

sure, but that are demands that the GLE side needs to take care about
(once it becomes usable ;)

> As long as the check is pretty fast though, there shouldn't
> be any harm in having the callbacks there.

if(quit_functions){} at the end of gtk_main(), that should be sufficient ;)

btw, i noticed that the gtk timeouts don't support remove_by_data, would
anyone actually need such a functionality?
(you could argue it should be there for completeness, but then i'd say the
same applies to gtk_init_remove, gtk_init_remove_by_data also and the returned
value of a gtk_init function should be honoured as well).

> 
> Regards,
>                                         Owen
> 

---
ciaoTJ