[Cairo] cairo API suggestion

[Bill Spitzak]

Don't such ideas make assumptions that Cairo is running locally?

I would prefer a system by which all "objects" *could* in fact be id numbers 
that refer to unaccessable and unreadable objects in a different memory 
space. Of course this is not true of Cairo right now, but it would be nice to 
allow it.

Cairo is supposed to be a *drawing* library, not another user-interface 
toolkit.

On Friday 12 September 2003 09:11 am, James Henstridge wrote:
> While going over the Cairo API, I found another area that would be good
> to change in order to make it more friendly for language bindings: it
> would be nice if the cairo_surface_create_for_image() took user data and
> destroy notify arguments (similar to APIs in GTK).
>
> This would allow me to make the cairo_surface_t own a reference to a
> Python object.  This way, the Python binding would be able to make sure
> the Python object owning the buffer won't get destroyed before Cairo is
> finished with the surface.

-- 
                   ,~,~,~,~ ~ ~ ~ ~
     /\_       _|_========___         Bill Spitzak
 ~~~/\/\\~~~~~~\____________/~~~~~~~~ spitzak at d2.com