<div dir="ltr"><br><br><div class="gmail_quote">2008/9/13 Colin Walters <span dir="ltr"><<a href="mailto:walters@verbum.org">walters@verbum.org</a>></span><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Hi,<br>
<br>
We had a short discussion about this on #gtk+ a few days ago.<br>
Basically, it would greatly help the GObject-based stack above Cairo<br>
(GTK+, HippoCanvas, GooCanvas, Clutter, etc.) if we had official<br>
GObject "boxed" types for Cairo objects.<br>
<br>
This is necessary because the GTypes are used in things like signals<br>
to specify argument types, and language bindings need that data. Now,<br>
one possibility Matthias Clasen suggested would be to put these in<br>
GDK, which would be somewhat tricky from a layering perspective<br>
(neither Pango nor Clutter for example currently depend on GDK as far<br>
as I know), but might be workable. Another possibility is to put them<br>
in Cairo itself, optionally enabled.</blockquote><div><br>A reflection of this problem will appear at language bindings level. If we put this in Cairo itself, then pycairo should add support for this and optionally depend on pygobject. It sounds fine to me, but I hope the pycairo maintainer is OK with it too! ;-)<br>
<br>But I think this is a good idea; currently GooCanvas is defining its own private GTypes for Cairo stuff, which clearly are not specific to GooCanvas...<br></div></div><br>-- <br>Gustavo J. A. M. Carneiro<br>INESC Porto, Telecommunications and Multimedia Unit<br>
"The universe is always one step beyond logic." -- Frank Herbert<br>
</div>