To: Claude.Marche at lri.fr
Cc: stalkern2 at tin.it, lablgtk at kaba.or.jp
Subject: Re: animation: pixmap transparency?
In-Reply-To: <16138.54573.314788.280376 at mailhost.lri.fr>
References: <200307080926.20760.stalkern2 at tin.it>
	<200307081555.38754.stalkern2@tin.it>
	<16138.54573.314788.280376@mailhost.lri.fr>
Mime-Version: 1.0
Content-Type: Text/Plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-Id: <20030709092440G.garrigue at kurims.kyoto-u.ac.jp>
Date: Wed, 09 Jul 2003 09:24:40 +0900
From: Jacques Garrigue <garrigue at kurims.kyoto-u.ac.jp>
Lines: 27

From: Claude Marche <Claude.Marche at lri.fr>

> Hi Ernesto,
> 
> I think you should deactivate the clip mask before drawing your
> background. May be such a possibility is missing in lablgtk: directly
> in C, one may call Gdk_GC::set_clip_mask without argument to unset the
> clip mask, but I see no way to call it without argument from lablgtk. 
> 
> lablgtk maintainers, in ideas ?

You're damn right. This is indeed a stupid oversight.
The only workaround I see is to use another GC.
If you are using a GDraw.pixmap, this can be done by
  let pixmap2 = new GDraw.pixmap pixmap#pixmap ?mask:pixmap#mask
This way you can use two GCs: one with no clip mask, and the oter with
a clip mask you can change as needed.

I'm not sure how this should be fixed in the library.
One way it to make the argument to #set_clip_mask an option, but this
is incompatible, and not so nice.
Another idea would to add a #new_gc method, which would return a new
object with a different GC. That is, the above workaround would become
  let pixmap2 = pixmap#new_gc
This looks conceptually cleaner to me.

     Jacques