Remove preserves3D() from CCLayerDelegate
Created attachment 107755 [details] Patch
Created attachment 107757 [details] Patch
Created attachment 108436 [details] Patch
Rebased patch (and friendly ping)
Comment on attachment 108436 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=108436&action=review > Source/WebCore/platform/graphics/chromium/GraphicsLayerChromium.cpp:367 > + platformLayer->setPreserves3D(m_preserves3D); I don't think we need to do this - the canvas contents layer can't have preserves3D set since it's a CSS property and would only apply to the container layer, and even if it could we should be able to rely on the normal property propagation code > Source/WebCore/platform/graphics/chromium/GraphicsLayerChromium.cpp:400 > + layer->setPreserves3D(m_preserves3D); same here - i don't think media layers can have this bit > Source/WebCore/platform/graphics/chromium/GraphicsLayerChromium.cpp:600 > + if (m_contentsLayer) > + m_contentsLayer->setPreserves3D(m_preserves3D); it looks like we used to propagate the preserves3D bit to the contents layer. this change preserves that behavior but I think it's actually wrong. Can you see if any tests break if you only propagate the preserves3D bit to the m_layer? > Source/WebCore/platform/graphics/chromium/GraphicsLayerChromium.cpp:652 > + m_contentsLayer->setPreserves3D(m_preserves3D); do we need this?
Comment on attachment 108436 [details] Patch R=me since this is preserving the existing behavior, so it's fine to land as-is. I think we can simplify, however.
Created attachment 108551 [details] Patch
View in context: https://bugs.webkit.org/attachment.cgi?id=108436&action=review >> Source/WebCore/platform/graphics/chromium/GraphicsLayerChromium.cpp:367 >> + platformLayer->setPreserves3D(m_preserves3D); > > I don't think we need to do this - the canvas contents layer can't have preserves3D set since it's a CSS property and would only apply to the container layer, and even if it could we should be able to rely on the normal property propagation code Done. >> Source/WebCore/platform/graphics/chromium/GraphicsLayerChromium.cpp:400 >> + layer->setPreserves3D(m_preserves3D); > > same here - i don't think media layers can have this bit Done. >> Source/WebCore/platform/graphics/chromium/GraphicsLayerChromium.cpp:600 >> + m_contentsLayer->setPreserves3D(m_preserves3D); > > it looks like we used to propagate the preserves3D bit to the contents layer. this change preserves that behavior but I think it's actually wrong. Can you see if any tests break if you only propagate the preserves3D bit to the m_layer? Done. Layout tests still pass. >> Source/WebCore/platform/graphics/chromium/GraphicsLayerChromium.cpp:652 >> + m_contentsLayer->setPreserves3D(m_preserves3D); > > do we need this? Removed.
Comment on attachment 108551 [details] Patch R=me
Comment on attachment 108551 [details] Patch Clearing flags on attachment: 108551 Committed r95886: <http://trac.webkit.org/changeset/95886>
All reviewed patches have been landed. Closing bug.