Rename Frame::renderer() to contentRenderer() There was a fixme in the code. I fixed it, it was bothering me (and causing some minor confusion).
Created attachment 20638 [details] patch I'd like hyatt or darin to look at this. This is not just a rename. I also simplified a few cases where the code did things like this: m_frame->document()->renderer()->view() instead of m_frame->view() or other such nonsense.
Comment on attachment 20638 [details] patch r=me!
Comment on attachment 20638 [details] patch + // FIXME: is root always a RenderView? if so, static_cast at initialization + ASSERT(root->isRenderView()); This assert is wrong and the FIXME should be removed. The root is always a RenderView when subtree is false, but never is when subtree is true. Everything else looks fine.
(In reply to comment #3) > (From update of attachment 20638 [details] [edit]) > + // FIXME: is root always a RenderView? if so, static_cast at > initialization > + ASSERT(root->isRenderView()); > > This assert is wrong and the FIXME should be removed. The root is always a > RenderView when subtree is false, but never is when subtree is true. > > Everything else looks fine. Interesting. I removed the ASSERT and changed the if as part of: https://bugs.webkit.org/show_bug.cgi?id=18585 (when I found the ASSERT was being hit by the layout tests). I really should squash these two bugs into one, and the patches likewise.
Committing to http://svn.webkit.org/repository/webkit/trunk ... M WebCore/ChangeLog M WebCore/WebCore.base.exp M WebCore/editing/Editor.cpp M WebCore/loader/FrameLoader.cpp M WebCore/page/DragController.cpp M WebCore/page/EventHandler.cpp M WebCore/page/Frame.cpp M WebCore/page/Frame.h M WebCore/page/FrameView.cpp M WebCore/rendering/RenderTreeAsText.cpp M WebCore/svg/graphics/SVGImage.cpp M WebKit/mac/ChangeLog M WebKit/mac/Misc/WebCoreStatistics.mm M WebKit/mac/WebView/WebRenderNode.mm Committed r32231