Switch to using accessibilityEnabled and inlineTextBoxAccessibilityEnabled from settings

Source/core/rendering/RenderBlockFlow.cpp

Index: Source/core/rendering/RenderBlockFlow.cpp
diff --git a/Source/core/rendering/RenderBlockFlow.cpp b/Source/core/rendering/RenderBlockFlow.cpp
index df4eb4890dd809149065a29b96bf5c1a9cf5b5d1..bfbc5e05ffed94e77631ab2aa8ca63d3878a6d01 100644
--- a/Source/core/rendering/RenderBlockFlow.cpp
+++ b/Source/core/rendering/RenderBlockFlow.cpp
@@ -33,6 +33,8 @@
 
 #include "core/accessibility/AXObjectCache.h"
 #include "core/frame/FrameView.h"
+#include "core/frame/LocalFrame.h"
+#include "core/frame/Settings.h"
 #include "core/rendering/HitTestLocation.h"
 #include "core/rendering/RenderFlowThread.h"
 #include "core/rendering/RenderLayer.h"
@@ -1783,7 +1785,8 @@ RootInlineBox* RenderBlockFlow::createAndAppendRootInlineBox()
     RootInlineBox* rootBox = createRootInlineBox();
     m_lineBoxes.appendLineBox(rootBox);
 
-    if (UNLIKELY(AXObjectCache::accessibilityEnabled()) && m_lineBoxes.firstLineBox() == rootBox) {
+    Settings* settings = frame()->settings();
+    if (UNLIKELY(settings && settings->accessibilityEnabled()) && m_lineBoxes.firstLineBox() == rootBox) {

[abarth-chromium, 2014/08/23 05:59:11]

This is sad.  We're adding a ton more branches to this codepath.

[aboxhall, 2014/08/27 15:26:33]

I don't follow.

[dmazzoni, 2014/08/27 16:17:15]

I think Adam is saying that this function (createAndAppendRootInlineBox) is
called thousands of times during layout, and testing whether accessibility is
enabled every time is expensive - and this change potentially makes it worse.

I think it might be possible to get rid of this test entirely. I filed
http://crbug.com/408159, let me see what the implications are if we get rid of
this code and the accessibility logic that depended on it.

         if (AXObjectCache* cache = document().existingAXObjectCache())
             cache->recomputeIsIgnored(this);
     }