cc: Video layers without a frame should not occlude

cc/layers/video_frame_provider_client_impl.cc

Index: cc/layers/video_frame_provider_client_impl.cc
diff --git a/cc/layers/video_frame_provider_client_impl.cc b/cc/layers/video_frame_provider_client_impl.cc
index 5fab4e2f04c4d0cc01dc081ccc59a0a86e79880a..8e85fa7d8b8cb9c2780253a8f902b94d31fb6792 100644
--- a/cc/layers/video_frame_provider_client_impl.cc
+++ b/cc/layers/video_frame_provider_client_impl.cc
@@ -98,6 +98,11 @@ void VideoFrameProviderClientImpl::ReleaseLock() {
   provider_lock_.Release();
 }
 
+bool VideoFrameProviderClientImpl::HasCurrentFrame() {
+  base::AutoLock locker(provider_lock_);
+  return provider_ && provider_->GetCurrentFrame();

[danakj, 2015/05/18 17:19:52]

You need to match calls to GetCurrentFrame() with PutCurrentFrame(), but I think
it might be nicer to add a HasCurrentFrame() to the VideoFrameProvider as well.
Or is that horrible?

[hendrikw, 2015/05/18 17:46:58]

That's an weird requirement.  WebMediaPlayerMS::GetCurrentFrame changes state :/
 I don't like. I'll add the HasCurrentFrame() to the provider.

+}
+
 const gfx::Transform& VideoFrameProviderClientImpl::StreamTextureMatrix()
     const {
   DCHECK(thread_checker_.CalledOnValidThread());