Use the solid color detection to create solid layers

cc/layers/picture_layer.cc

Index: cc/layers/picture_layer.cc
diff --git a/cc/layers/picture_layer.cc b/cc/layers/picture_layer.cc
index 4207aa248142ba0174101773a7b38bd89ff61af4..68c416011b336f2db8ed6a50626af9bcfa1296af 100644
--- a/cc/layers/picture_layer.cc
+++ b/cc/layers/picture_layer.cc
@@ -36,6 +36,8 @@ void PictureLayer::PushPropertiesTo(LayerImpl* base_layer) {
   Layer::PushPropertiesTo(base_layer);
   PictureLayerImpl* layer_impl = static_cast<PictureLayerImpl*>(base_layer);
 
+  layer_impl->SetSolidColorState(pile_->IsSolidColor(), pile_->GetSolidColor());

[danakj, 2014/09/10 18:40:50]

can the layerimpl not pull this off its own pile instead of adding these class
members which just double the ones on the PLI's pile?

[hendrikw, 2014/09/11 00:27:40]

PLI doesn't know if it's solid or not (only PL knows this).  Besides, I need
this function to override the solid state for the test code.

[danakj, 2014/09/11 23:36:37]

Can't the test set the bool in the pile? Take a look at the is_mask_ in pile for
an example. PLI uses that to change behaviour. And tests set it on the pile to
change behaviour.

+
   if (layer_impl->bounds().IsEmpty()) {
     // Update may not get called for an empty layer, so resize here instead.
     // Using layer_impl because either bounds() or paint_properties().bounds