[webkit-reviews] review requested: [Bug 91023] [chromium] Remove the RenderPass pointer from RenderPassDrawQuad : [Attachment 152010] Patch

[bugzilla-daemon at webkit.org]

Dana Jansens <danakj at chromium.org> has asked  for review:
Bug 91023: [chromium] Remove the RenderPass pointer from RenderPassDrawQuad
https://bugs.webkit.org/show_bug.cgi?id=91023

Attachment 152010: Patch
https://bugs.webkit.org/attachment.cgi?id=152010&action=review

------- Additional Comments from Dana Jansens <danakj at chromium.org>
- removed fixme for global renderpass id since we'll map it in the host
compositor
- changed ownership around for RenderPasses so I can remove skippedPasses from
FrameData. now the HashMap owns all passes. The renderPasses list is just
pointers that will actually be drawn. This will be more friendly in ubercomp
land, where we want to hold onto passes we didn't generate across frames and
possibly use them/not use them depending on cached textures.

The ownership change exposed a bug that a unit test causes - a root layer with
opacity < 1 tries to make another RenderSurface, causing it to appear in the
RSLL twice. Fixing this separately.