BlackBerry does not currently support hardware-accelerated CSS filter effects. This bug covers neither CSS reference (SVG) filters, nor CSS custom (shader) filters.
Created attachment 155548 [details] Patch
Comment on attachment 155548 [details] Patch Attachment 155548 [details] did not pass efl-ews (efl): Output: http://queues.webkit.org/results/13386881
Comment on attachment 155548 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=155548&action=review nice! please fix the style nits and minor issues. r-'ing since it needs another iteration to fix EFL anyways. > Source/WebCore/platform/graphics/blackberry/GraphicsLayerBlackBerry.h:83 > + virtual bool setFilters(const FilterOperations &); the 'bool' return type here is very unclear. I imagine this class implements this method which is inherited from someone else? > Source/WebCore/platform/graphics/blackberry/LayerCompositingThread.h:215 > + bool filterOperationsChanged() { return m_filterOperationsChanged; } const > Source/WebCore/platform/graphics/blackberry/LayerCompositingThread.h:218 > + WTF::Vector<RefPtr<LayerFilterRendererAction> > filterActions() { return m_filterActions; } const > Source/WebCore/platform/graphics/blackberry/LayerCompositingThread.h:266 > + bool m_filterOperationsChanged; you should manually default initialize it in the ctor. > Source/WebCore/platform/graphics/blackberry/LayerFilterRenderer.cpp:70 > + default: > + return -1; > + } does it cover all cases of the enum? if you ASSERT_NOT_REACHED here, please. > Source/WebCore/platform/graphics/blackberry/LayerFilterRenderer.cpp:89 > +Uniform1f::Uniform1f(int c_location, float c_val) : Uniformf(c_location), m_val(c_val) wrong style, new line here. > Source/WebCore/platform/graphics/blackberry/LayerFilterRenderer.cpp:138 > +{ > + > +} unneeded extra line. > Source/WebCore/platform/graphics/blackberry/LayerFilterRenderer.cpp:142 > +LayerFilterRendererAction::~LayerFilterRendererAction() > +{ > +} omit this. let the compiler private one to us.
Created attachment 155565 [details] Patch
Please disregard the previous patch.
Created attachment 155567 [details] Patch
Created attachment 155578 [details] Patch
Got rid of WTF:: prefixes.
Comment on attachment 155578 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=155578&action=review Looks good, one more iteration needed. > Source/WebCore/platform/graphics/blackberry/LayerFilterRenderer.cpp:371 > + for (int i = 0;i < LayerData::NumberOfCSSFilterShaders; ++i) Add a space after i = 0; > Source/WebCore/platform/graphics/blackberry/LayerFilterRenderer.h:98 > + // see the ping-pong note in LayerFilterRenderer::applyActions Not completely WebKit style comments. > Source/WebCore/platform/graphics/blackberry/LayerWebKitThread.h:102 > + void setFilters(const FilterOperations& filters) { m_filters = filters; m_filtersChanged = 1; setNeedsCommit(); } Inconsistent use of m_filtersChanged, setting to 1 and false. > ChangeLog:11 > + * Source/cmakeconfig.h.cmake: Acknowldge CSS filter effects Typo Acknowldge
Comment on attachment 155578 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=155578&action=review > Source/WebCore/platform/graphics/blackberry/LayerFilterRenderer.cpp:76 > + , m_refCount(1) Now that Uniformf is RefCounted, you can remove m_refCount =)
Created attachment 155589 [details] Patch
Comment on attachment 155589 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=155589&action=review > Source/WebCore/platform/graphics/blackberry/LayerFilterRenderer.h:100 > + bool shouldPushSnapshot() { return m_pushSnapshot; } Could be const. > Source/WebCore/platform/graphics/blackberry/LayerFilterRenderer.h:101 > + void setPushSnapshot() { m_pushSnapshot = 1; } Better use true. > Source/WebCore/platform/graphics/blackberry/LayerFilterRenderer.h:103 > + bool shouldPopSnapshot() { return m_popSnapshot; } Ditto. > Source/WebCore/platform/graphics/blackberry/LayerFilterRenderer.h:104 > + void setPopSnapshot() { m_popSnapshot = 1; } Etcetera.
Created attachment 155596 [details] Patch
Comment on attachment 155596 [details] Patch Good work!
Comment on attachment 155596 [details] Patch Clearing flags on attachment: 155596 Committed r124242: <http://trac.webkit.org/changeset/124242>
All reviewed patches have been landed. Closing bug.