On Thu, 2010-09-02 at 13:29 -0400, Havoc Pennington wrote: > Each pixbuf could keep a cairo_surface_t which just pointed to the > pixbuf pixels and that cairo_surface_t would be used in paint() or for > other compositing/scaling/etc. kind of operations. Are you thinking of the loaders populating an image surface (and thus being in application memory) or populating an xlib surface (and thus being in X server memory?) People seem to have gone to a lot of effort over the years to find the right mix between GdkPixmap and GdkPixbuf and a fair bit of code has been written (using ie cairo_surface_create_similar()) to do caching, etc. Between your ideas here and Benjamin's work, I get the impression that the status quo is well and truly smashed [great!] but now I'm definitely vague about what's being loaded where, and when someone is going to have to worry [differently] about the performance implications of large copies between app and X server, etc. AfC Sydney
Attachment:
signature.asc
Description: This is a digitally signed message part