Summary: | [chromium] Transform animation state should be inherited from parents | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Dana Jansens <danakj> | ||||||||
Component: | New Bugs | Assignee: | Dana Jansens <danakj> | ||||||||
Status: | RESOLVED FIXED | ||||||||||
Severity: | Normal | CC: | backer, cc-bugs, enne, jamesr, nduca, piman, reveman, shawnsingh, vollick, webkit.review.bot | ||||||||
Priority: | P2 | ||||||||||
Version: | 528+ (Nightly build) | ||||||||||
Hardware: | Unspecified | ||||||||||
OS: | Unspecified | ||||||||||
Bug Depends on: | |||||||||||
Bug Blocks: | 81354 | ||||||||||
Attachments: |
|
Description
Dana Jansens
2012-03-16 10:00:50 PDT
Created attachment 132306 [details]
Patch
I'm not sure if this is something we'll need to worry about being problematic, but which layers have RenderSurfaces may be different between the two threads in the face of animating opacity. Created attachment 132371 [details]
Patch
- Renamed the transformIsAnimating() functions based on feedback from Shawn. I think they're more clear and consistent now.
- Moved the bugfix for creation of render surfaces out, as it's a separate issue.
Comment on attachment 132371 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=132371&action=review Looks good in general. R=me, with nits. > Source/WebCore/platform/graphics/chromium/cc/CCRenderSurface.h:95 > + bool targetSurfaceTransformsAreAnimating() const { return m_targetSurfaceTransformsAreAnimating; } > + void setTargetSurfaceTransformsAreAnimating(bool animating) { m_targetSurfaceTransformsAreAnimating = animating; } > + bool screenSpaceTransformsAreAnimating() const { return m_screenSpaceTransformsAreAnimating; } > + void setScreenSpaceTransformsAreAnimating(bool animating) { m_screenSpaceTransformsAreAnimating = animating; } Why is screen space transform plural here? Isn't there just one? > Source/WebKit/chromium/tests/CCLayerTreeHostCommonTest.cpp:724 > + EXPECT_TRUE(renderSurface2->renderSurface()->targetSurfaceTransformsAreAnimating()); Can you add a test where renderSurface()->targetSurfaceTransformsAreAnimating() is false? > Source/WebKit/chromium/tests/CCLayerTreeHostCommonTest.cpp:729 > + // Verify drawTransformsAnimatingInScreen values > + // Can you add a test where screenSpaceTransformIsAnimating ends up being false? Comment on attachment 132371 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=132371&action=review ty :) >> Source/WebCore/platform/graphics/chromium/cc/CCRenderSurface.h:95 >> + void setScreenSpaceTransformsAreAnimating(bool animating) { m_screenSpaceTransformsAreAnimating = animating; } > > Why is screen space transform plural here? Isn't there just one? there is going to be one for the surface and one for the replica. >> Source/WebKit/chromium/tests/CCLayerTreeHostCommonTest.cpp:724 >> + EXPECT_TRUE(renderSurface2->renderSurface()->targetSurfaceTransformsAreAnimating()); > > Can you add a test where renderSurface()->targetSurfaceTransformsAreAnimating() is false? k! >> Source/WebKit/chromium/tests/CCLayerTreeHostCommonTest.cpp:729 >> + // > > Can you add a test where screenSpaceTransformIsAnimating ends up being false? k! Created attachment 132426 [details]
Patch
Thanks for the input, I made that test a lot better. It tests both cases with the 2 render surfaces, and tests propogation of screen space transforms by setting transforms on 2 of the 3 children of root.
Comment on attachment 132426 [details] Patch Clearing flags on attachment: 132426 Committed r111137: <http://trac.webkit.org/changeset/111137> All reviewed patches have been landed. Closing bug. |