Re: [gtk-list] Re: problem with gtk_signal_connect_object

[Tim Janik <timj gtk org>]

On Wed, 20 May 1998, Stefan Wille wrote:

> > >  gtk_signal_connect_object (object, "destroy",
> > >                             GTK_SIGNAL_FUNC (g_free),
> > >                             GTK_OBJECT (some_data_associated_with_object));
> >                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> >                               this should have been
> >                               (GtkObject*) some_data_associated_with_object
> 
> If the function will take a void* as 4th argument, not just a
> GtkObject*, 
> shouldn't the prototype be:
> 
>     guint  gtk_signal_connect_object        (GtkObject          
> *object,
>                                              const gchar         *name,
>                                              GtkSignalFunc        func,
>                                              gpointer           
> *func_data);
> 
> just like for gtk_signal_connect? The required cast to GtkObject* seems
> to be
> unnecessary and misleading (it was for me).

well it should better look like
guint  gtk_signal_connect_object (GtkObject      *object,
                                  const gchar    *name,
                                  GtkSignalFunc   func,
                                  gpointer        slot_object);
to keep the argument names decriptive/consistent...

> It should be possible to make the change without breaking existing code, 
> because GtkObject* should be compatible with gpointer.

but, you are probably right and we can savely make that change...

> 
> Stefan
> 

---
ciaoTJ