In-Reply-To: <16306.21520.114875.995598 at akasha.ijm.jussieu.fr>
References: <16306.4493.522709.269560 at akasha.ijm.jussieu.fr>
	<20031112110126.GA24987@cs.unibo.it>
	<16306.21520.114875.995598@akasha.ijm.jussieu.fr>
Subject: Re: GTree row selection event
To: andrieu at ijm.jussieu.fr
Cc: lablgtk at kaba.or.jp
Mime-Version: 1.0
Content-Type: Text/Plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-Id: <20031113090754N.garrigue at kurims.kyoto-u.ac.jp>
Date: Thu, 13 Nov 2003 09:07:54 +0900
From: Jacques Garrigue <garrigue at kurims.kyoto-u.ac.jp>
Lines: 27

From: Olivier Andrieu <andrieu at ijm.jussieu.fr>
>  Claudio Sacerdoti Coen [Wednesday 12 November 2003] :
>  >  Anyway, since a couple of students of mine already asked me how to
>  >  obtain an event_ops from a generic widget, I think that this function
>  >  should be inserted in lablgtk2 (with type <as_widget : ... ; ...> ->
>  >  event_ops).
> 
> Or we could simply put the event method in the base widget class, with
> or without the NO_WINDOW flag check ; it's never an error to connect
> these event-related signals, they are just never called if the widget
> doesn't have a GdkWindow.

The goal of the #event sub-object is to be able to tell from its type
whether a widget handles events or not.
If a widget is able to handle events and has no #event method, this is
a bug of lablgtk and should be corrected.

The dynamic approach is also possible, but means that you will only
know whether a widget handles events or not at runtime. Religious problem.

No check at all is a bad idea: your program compiles, but you callback
is never called, and you don't know why.

(Whether a widget handles events or not is not very explicit in the
Gtk2 documentation)

Jacques