Mark private method layoutBlockFlow as inline

Source/core/rendering/RenderBlockFlow.cpp

 /*
  * Copyright (C) 2013 Google Inc. All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are
  * met:
  *
  *     * Redistributions of source code must retain the above copyright
  * notice, this list of conditions and the following disclaimer.
  *     * Redistributions in binary form must reproduce the above
@@ skipping 233 matching lines @@
 {
     m_hasOnlySelfCollapsingChildren = RenderBlock::isSelfCollapsingBlock();
     return m_hasOnlySelfCollapsingChildren;
 }
 
 void RenderBlockFlow::layoutBlock(bool relayoutChildren)
 {
     layoutBlockFlow(relayoutChildren);
 }
 
-void RenderBlockFlow::layoutBlockFlow(bool relayoutChildren, LayoutUnit pageLogicalHeight)
+inline void RenderBlockFlow::layoutBlockFlow(bool relayoutChildren, LayoutUnit pageLogicalHeight)

[Inactive, 2014/01/20 18:57:50]

Yes, looking at the original patch, it seems like the extra function call in the
layoutBlock() virtual function is the only difference that might impact
performance.
Hopefully, inlining it will fix the regression.

 {
     ASSERT(needsLayout());
     ASSERT(isInlineBlockOrInlineTable() || !isInline());
 
     // If we are self-collapsing with self-collapsing descendants this will get set to save us burrowing through our
     // descendants every time in |isSelfCollapsingBlock|. We reset it here so that |isSelfCollapsingBlock| attempts to burrow
     // at least once and so that it always gives a reliable result reflecting the latest layout.
     m_hasOnlySelfCollapsingChildren = false;
 
     if (!relayoutChildren && simplifiedLayout())
@@ skipping 2557 matching lines @@
 RenderBlockFlow::RenderBlockFlowRareData& RenderBlockFlow::ensureRareData()
 {
     if (m_rareData)
         return *m_rareData;
 
     m_rareData = adoptPtr(new RenderBlockFlowRareData(this));
     return *m_rareData;
 }
 
 } // namespace WebCore