Simplify subtree relayout scheduling a bit.
Created attachment 210462 [details] Patch idea.
Comment on attachment 210462 [details] Patch idea. View in context: https://bugs.webkit.org/attachment.cgi?id=210462&action=review > Source/WebCore/page/FrameView.cpp:2445 > + if (m_layoutRoot == &relayoutRoot) > + return; The argument could use a better name. newLayoutRoot? > Source/WebCore/page/FrameView.cpp:2466 > + InspectorInstrumentation::didInvalidateLayout(&frame()); There is some repetition with this. I wonder if splitting this up to two functions would look nicer. The inner one would return bool if it actually did invalidate layout for example.
Created attachment 210467 [details] Patch for landing...
Comment on attachment 210467 [details] Patch for landing... Clearing flags on attachment: 210467 Committed r155046: <http://trac.webkit.org/changeset/155046>
All reviewed patches have been landed. Closing bug.
This change made many (maybe all?) SVG tests crash with an assertion in debug builds. See <rdar://problem/14908967> for an Apple internal bug.
(In reply to comment #6) > This change made many (maybe all?) SVG tests crash with an assertion in debug builds. See <rdar://problem/14908967> for an Apple internal bug. Fix here: bug 120689
More fallback from this change: bug 121524.