Re: [gtk-list] Re: Interpreter requests

[Tim Janik <timj gtk org>]

On 9 Jun 1998, Marius Vollmer wrote:

> Tim Janik <timj@gtk.org> writes:
> 
> > GtkType
> > foo_enum_get_type (void)
> > {
> [...]
> > }
> 
> Wouldn't it be better to just add gtk_type_register_enum /
> gtk_type_register_flags / etc functions to Gtk, so that
> foo_enum_get_type could just look like:
> 
>     GtkType
>     foo_enum_get_type (void)
>     {
>       static GtkType type = GTK_TYPE_INVALID;
>       if (type == GTK_TYPE_INVALID)
>         type = gtk_type_register_enum ("FooEnum", foo_enum_values);
>       return type;
>     }

sure, that is probably a good idea!

> gtk_type_register_enum would be
> 
>     GtkType
>     gtk_type_register_enum (char *name, GtkEnumValues *vals)
>     {
>       GtkTypeInfo enum_info =
>       {
> 	name,
> 	0,
> 	0,
> 	(GtkClassInitFunc) NULL,
> 	(GtkObjectInitFunc) NULL,
> 	(GtkArgSetFunc) NULL,
> 	(GtkArgGetFunc) NULL,
>       };
> 
>       GtkType type = gtk_type_unique_no_hook (GTK_TYPE_ENUM_TYPE, &enum_info);
>       gtk_type_enum_set_values (type, vals);
>       gtk_type_run_creation_hooks (type);
> 
>       return type;
>     }
> 
> gtk_type_enum_set_values could then be removed from the exported
> functions.
> 
> Other useful registration functions (and associated getters) would
> certainly be
> 
>     GtkType gtk_type_register_boxed (char *name, copy, free);
>     GtkType gtk_type_register_foreign (char *name, copy, free, ...);
>     GtkType gtk_type_register_object (...);
> 
> Ask me if you want to know more about what I have in mind here.

yes, please extend on this subject.

> 

---
ciaoTJ