Bug 91769

On what platform is this happening? I"m running this test on Mac Safari and Mac Chromium just fine.

Is ENABLE_CSS_FILTERS even turned on the Mac port?

Regardless, this is occurring on the Mountain Lion test bots, which unfortunately are still not public-access. :(

(In reply to comment #2) > Is ENABLE_CSS_FILTERS even turned on the Mac port? Both ENABLE_CSS_FILTERS and ENABLE_CSS_SHADERS are turned on the Mac port. That's at least on Lion. CSS Shaders also need WebGL enabled and also activated. The error you see happens only when ENABLE_CSS_SHADERS is disabled at compile time, as we enable the runtime flags using the overridePreference call. I can only suppose that Mountain Lion has ENABLE_CSS_SHADERS turned of by default at compile time. In that case, can you please add the whole the css3/filters/custom folder to the ignored list on Mountain Lion? And maybe a new bug to track the status of that compile time flag?

Mountain Lion doesn't have any differences from Lion when it comes to the ENABLE_CSS_SHADER flag. And other tests in css3/filters/custom pass fine; it's just css3/filters/custom/custom-filter-animation.html that is failing.

Added the expected failing result to the mac/TestExpectations list in <http://trac.webkit.org/changeset/123136>.

There are two issues: 1. Safari Mac on Mountain Lion has accelerated animations support for filters. It is only enabled on ML at runtime. However, CSS Shaders are not composited in hardware yet, so it fallbacks to software rendering. CSS animations will still try to apply using platform code meaning that WebKit will be stuck forever waiting for the animation started event. 2. The framebuffer is not attached again in repaints. That's because of a GraphicsContext3D bug: https://bugs.webkit.org/show_bug.cgi?id=93625 I will post patches for both bugs soon.

Created attachment 157574 [details] Patch V1

Comment on attachment 157574 [details] Patch V1 View in context: https://bugs.webkit.org/attachment.cgi?id=157574&action=review > Source/WebCore/ChangeLog:13 > + Also the GraphicsContext3D doesn't update it's own internal m_boundFBO when a framebuffer is deleted, typo it's -> its > Source/WebCore/ChangeLog:17 > + The man for glDeleteFramebuffer says that the bound framebuffer is unbound before deleting it. I suggest "The documentation for glDeleteFramebuffer..." > Source/WebCore/ChangeLog:20 > + and GraphicsContext3DEfl.cpp, to make sure that m_boundFBO is not used for other purposes. I think you left GraphicsContext3DEfl out of this patch. I'll r+ but cq- to either fix that or this ChangeLog.

Created attachment 158414 [details] Patch for landing Thanks Dean! The efl change was there, but I've used the bad filename for it in the Changelog -> efl/GraphicsContext3DPrivate.cpp. I've updated the Changelog.

Comment on attachment 158414 [details] Patch for landing Rejecting attachment 158414 [details] from commit-queue. Failed to run "['/mnt/git/webkit-commit-queue/Tools/Scripts/webkit-patch', '--status-host=queues.webkit.org', '-..." exit_code: 2 Last 500 characters of output: ebCore/ChangeLog Hunk #1 succeeded at 1 with fuzz 3. patching file Source/WebCore/platform/graphics/ca/GraphicsLayerCA.cpp patching file Source/WebCore/platform/graphics/efl/GraphicsContext3DPrivate.cpp patching file Source/WebCore/platform/graphics/filters/FECustomFilter.cpp patching file Source/WebCore/platform/graphics/opengl/GraphicsContext3DOpenGLCommon.cpp Failed to run "[u'/mnt/git/webkit-commit-queue/Tools/Scripts/svn-apply', u'--force']" exit_code: 1 cwd: /mnt/git/webkit-commit-queue/ Full output: http://queues.webkit.org/results/13489879

Created attachment 158421 [details] Patch for landing Rebased.

Comment on attachment 158421 [details] Patch for landing Clearing flags on attachment: 158421 Committed r125617: <http://trac.webkit.org/changeset/125617>

All reviewed patches have been landed. Closing bug.