[Webkit-unassigned] [Bug 70533] [Chromium] Cull occluded tiles in tiled layers

Thu Oct 20 14:45:21 PDT 2011

https://bugs.webkit.org/show_bug.cgi?id=70533

--- Comment #8 from Adrienne Walker <enne at google.com>  2011-10-20 14:45:21 PST ---
(In reply to comment #6)

> In the opaque case the parent will already occlude all its children.

I don't believe this is true.  The layer tree is not a bounding box hierarchy.  An opaque parent may or may not occlude its children.

> I'm hoping we can just assume we're considering all the correct layers for occlusion and discuss other aspects of the code for this patch.

There aren't false positives here, but given the number of empty layers that get created in WebKit, I'd really worry about all the potential occluders (mostly child -> grand*parent relationships) that we're missing.  Maybe I'm overstating this worry?

I'm interested in Antoine's previous comment about where this might need to go in the future to support those TODOs.  If you constructed a flattened list of layers that get drawn into a render surface, the set of potential occluders for a layer in that list is all the (opaque) layers that come after it in the list.  If you didn't have such an ordered list, it seems like it'd be O(n^2) to set all the potential occluders directly on each layer.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.