[cairo] pycairo subclass

[Steven Chaplin]

On Mon, 2007-12-10 at 05:01 -0500, Behdad Esfahbod wrote:
> On Sun, 2007-12-09 at 20:25 -0500, Steven Chaplin wrote:
> > 
> > The cairo bindings doc 'Memory Management' page describes why its not a
> > good idea to subclass cairo surfaces. So should cairo.Context be kept
> > consistent with cairo surfaces, and not allow subclassing, or is
> > cairo.Context a special case which should allow subclassing?
> > How useful would subclassing cairo.Context be?
> 
> As you said both pango and gdk subclass cairo.Context, for good.  So
> that one should stay like that.
> 
> One reason that subclassing cairo.Surface doesn't make sense, other than
> the issues pointed out in the Memory Management page, is that you can't
> override its methods.  That's a real limitation.  For example,
> subclassing the show_page() method can be quite useful.

I'm not sure I understand. Do you mean that because cairo itself is not
object-oriented there is little point subclassing a surface because you
won't be able to override cairo's show_page() function?

> For cairo.Context this is less of an issue because no other cairo object
> takes a cairo_t and work on it, but it's not true about the other cairo
> types.  It still is a problem with for example pango.  That is, I can't
> subclass a pango.CairoContext and override its stroke() method and have
> pango underlines use my brand new stroke implementation...

I've thought about it a bit more. Since pango and gdk subclass
cairo.Context I think we should allow subclassing from Python too to be
consistent. It took ages to find out how to get it working, but all the
required code was already there, I just had to change one 'NULL' to
'type'! I've updated pycairo in CVS, so the next version of pycairo
should allow subclassing cairo.Context from Python.

regards,
Steve