[webkit-reviews] review requested: [Bug 76015] [chromium] Use region reported painted opaque for draw culling : [Attachment 123192] Patch

[bugzilla-daemon at webkit.org]

Dana Jansens <danakj at chromium.org> has asked  for review:
Bug 76015: [chromium] Use region reported painted opaque for draw culling
https://bugs.webkit.org/show_bug.cgi?id=76015

Attachment 123192: Patch
https://bugs.webkit.org/attachment.cgi?id=123192&action=review

------- Additional Comments from Dana Jansens <danakj at chromium.org>
For enne review again:

- LayoutTest fail: tile's opaque rect needed to be intersected with the
layerVisibleRect
- Unit test fail: Solid color quads can now be considered opaque if their color
is opaque so the CCLayerTreeHostImpl.
- Unit test fail: Solid color quads are not opaque if alpha != 255, not if
alpha != 1.
- Better distinction between tileBounds() and tileRect() from m_tiler when
deciding if the tile is fully opaque or not.
- Turn on opaque tracking always. This is reflected in the
LayerTextureUpdaterTest unit test. - Increasingly concerned about not tracking
paintedOpaque areas for opaque layers, as its all kinds of unclear what happens
if the opaque flag is flipping on/off. This should simplify future code and
potential problems a lot.

Passes all DRT and unit tests for sure on my end now.