Hmm, we could modify frames_pending() to not count any surface that is occluded. That would resolve alf's concerns, but I'm not totally sure it's a good idea to make a special case of occlusion. If someone writes a shell that changes its view of surfaces without changing the Scene, then "fixing" this bug would cause problems of visible lag and even indefinite freezes.
I think we're better off declaring this not-a-bug. But I also think it could be easy to fix if necessary.
Hmm, we could modify frames_pending() to not count any surface that is occluded. That would resolve alf's concerns, but I'm not totally sure it's a good idea to make a special case of occlusion. If someone writes a shell that changes its view of surfaces without changing the Scene, then "fixing" this bug would cause problems of visible lag and even indefinite freezes.
I think we're better off declaring this not-a-bug. But I also think it could be easy to fix if necessary.