Re: delete-event of GtkWidget

[Enrico TrÃger <enrico troeger uvena de>]

On Mon, 22 Jan 2007 09:13:03 +0100, "Dan H." <dunno stoptrick com>
wrote:

> Enrico TrÃger wrote:
>
> > Yes, I misused gtk_widget_hide and already fixed my code.
>
> [...]
>
> > Obviously, I just wondered about the different behaviour of
> > different systems. I tested it also under Windows and there it is
> > like in Debian, the widget is still functional even though only by
> > accident or some other magic ;-).
>
> That's the problem (or actually, the beauty) of Undefined Behaviour:
> there's no need to even speculate on why things seem to work on one
> system but not on others.
Ok convinced. I just had luck(that my app didn't crashed) and bad luck
(because I wasn't aware of the misuse).

> BTW, why didn't your compiler warn you about the incompatible function
> pointer argument in your callback? I always have ALL compiler warnings
> on maximum bickering level to avoid this sort of problem.
Hm, I use -Wall for compiling with gcc. And this code doesn't produce a
warning:
g_signal_connect (dialog, "delete_event", G_CALLBACK(gtk_widget_hide),
NULL);
And how could it? The gcc can't know which return type the definition
of the delete_event expects, can it?


Regards,
Enrico

--
Get my GPG key from http://www.uvena.de/pub.key