Changes subtreeShouldRenderToSeparateSurface logic to account for explicit clipping

cc/layer_tree_host_common.cc

Index: cc/layer_tree_host_common.cc
diff --git a/cc/layer_tree_host_common.cc b/cc/layer_tree_host_common.cc
index 0a204ceb6914df96a5bdede2b6c8c9ee7e1c6915..37364ec7cc60b54384b8b8d8df1c5681838f1a9f 100644
--- a/cc/layer_tree_host_common.cc
+++ b/cc/layer_tree_host_common.cc
@@ -239,6 +239,7 @@ static inline void markLayerAsUpdated(Layer* layer)
 {
 }
 
+

[shawnsingh, 2012/12/18 19:11:00]

nit: no need to add extra line here

 template<typename LayerType>
 static bool subtreeShouldRenderToSeparateSurface(LayerType* layer, bool axisAlignedWithRespectToParent)
 {
@@ -274,7 +275,7 @@ static bool subtreeShouldRenderToSeparateSurface(LayerType* layer, bool axisAlig
         return true;
 
     // If the layer clips its descendants but it is not axis-aligned with respect to its parent.
-    if (layerClipsSubtree(layer) && !axisAlignedWithRespectToParent && numDescendantsThatDrawContent > 0)
+    if (layerClipsSubtree(layer) && !axisAlignedWithRespectToParent && !layer->drawProperties().does_not_require_scissoring)
         return true;
 
     // If the layer has some translucency and does not have a preserves-3d transform style.
@@ -435,15 +436,25 @@ template<typename LayerType>
 static void preCalculateMetaInformation(LayerType* layer)
 {
     int numDescendantsThatDrawContent = 0;
+    bool doesNotRequireScissoring = true;
+    bool sublayerXformPreventsClip = !layer->sublayerTransform().IsPositiveScaleOrTranslation();

[shawnsingh, 2012/12/18 19:11:00]

like Dana, I also prefer Transform instead of Xform.  I know we used it in
gfx/transform_unittests but I think everywhere else our convention has been to
spell out this word

 
     for (size_t i = 0; i < layer->children().size(); ++i) {
         LayerType* childLayer = layer->children()[i];
         preCalculateMetaInformation<LayerType>(childLayer);
+
         numDescendantsThatDrawContent += childLayer->drawsContent() ? 1 : 0;
         numDescendantsThatDrawContent += childLayer->drawProperties().num_descendants_that_draw_content;
+
+        if ((childLayer->drawsContent() && !childLayer->canClipSelf()) ||
+            !childLayer->drawProperties().does_not_require_scissoring ||

[shawnsingh, 2012/12/18 19:11:00]

I feel like my proposed variable name in earlier comments got overlooked, but I
really think it will help readability here.  The point we want readers to
realize is that this boolean tracks that information for all a layer's
descendants, but not the layer itself.  That's the key hint that makes this
if-statement more readable, IMHO.  Something like
"descendants_can_clip_themselves".  With that, the if-statement reads like:
"check the layer, check it's descendants, check the transforms related to the
childLayer"

+            sublayerXformPreventsClip ||
+            !childLayer->transform().IsPositiveScaleOrTranslation())
+            doesNotRequireScissoring = false;
     }
 
     layer->drawProperties().num_descendants_that_draw_content = numDescendantsThatDrawContent;
+    layer->drawProperties().does_not_require_scissoring = doesNotRequireScissoring;
 }
 
 // Recursively walks the layer tree starting at the given node and computes all the