[Cairo] virtualizing the surface interface

[Carl Worth]

On Sep 27, James Henstridge wrote:
 > How about cairo_surface_class_t ?  It sounds like the "impl" struct 
 > essentially defines a class of surfaces (in the object oriented
 > sense).

The problem with that is that then I would naturally use:

	cairo_surface_class_t *class;

at which point I start feeling like I'm setting myself up for
complaints from a C++ user about "class" being a reserved word.

-Carl