Plumb activation time to main

cc/trees/single_thread_proxy.cc

Index: cc/trees/single_thread_proxy.cc
diff --git a/cc/trees/single_thread_proxy.cc b/cc/trees/single_thread_proxy.cc
index 0386145c7c55fea08e4f34362dd5f3536b1b1a02..050c153b756501d54db39b903034035f39e64d5a 100644
--- a/cc/trees/single_thread_proxy.cc
+++ b/cc/trees/single_thread_proxy.cc
@@ -624,6 +624,13 @@ void SingleThreadProxy::SendBeginMainFrameNotExpectedSoon() {
   layer_tree_host_->BeginMainFrameNotExpectedSoon();
 }
 
+int SingleThreadProxy::SyncTreeSourceFrameNumber() {

[vmpstr, 2017/04/05 19:11:33]

Do we have pending trees in single thread proxy? (I might not know of all the
uses of this :) ).

[brianderson, 2017/04/06 23:43:15]

SingleThreadProxy is only used in the Browser, where we don't need this feature
- so we could just return 0. But it does seem to break the Proxy abstraction to
expose the concept of a SyncTree.

Alternatively, we could pass in the source_frame_number as an argument to
NotifyReadyToActivate. ProxyImpl would pass in the pending tree's number.
SingleThreadProxy would pass in the active_tree's number (since we commit to
active tree).

[panicker, 2017/04/08 00:29:07]

Updated to passing argument to NotifyReadyToActivate and passing pending-tree in
ProxyImpl.
Passing active-tree in SingleThreadProxy triggers a ton of NotifyReadyToActivate
without corresponding BeginMainFrame. 
Are you sure this should be active-tree here? I'll take a closer look.

[panicker, 2017/04/10 22:20:05]

Discussed offline and updated to passing 0 here.

+  if (layer_tree_host_impl_ && layer_tree_host_impl_->pending_tree())
+    return layer_tree_host_impl_->pending_tree()->source_frame_number();
+  else
+    return 0;
+}
+
 void SingleThreadProxy::BeginMainFrame(const BeginFrameArgs& begin_frame_args) {
   if (scheduler_on_impl_thread_) {
     scheduler_on_impl_thread_->NotifyBeginMainFrameStarted(