Replace all uses of webrtc::VideoRendererInterface by rtc::VideoSinkInterface.

remoting/protocol/webrtc_video_renderer_adapter.cc

Index: remoting/protocol/webrtc_video_renderer_adapter.cc
diff --git a/remoting/protocol/webrtc_video_renderer_adapter.cc b/remoting/protocol/webrtc_video_renderer_adapter.cc
index ec38ed5a020d84b683e3d15862a7b6e1f2b24f02..1dcd4ac0086d359be1ab8645f4654a408c715374 100644
--- a/remoting/protocol/webrtc_video_renderer_adapter.cc
+++ b/remoting/protocol/webrtc_video_renderer_adapter.cc
@@ -41,23 +41,23 @@ WebrtcVideoRendererAdapter::WebrtcVideoRendererAdapter(
     LOG(WARNING) << "Received media stream with multiple video tracks.";
   }
 
-  video_tracks[0]->AddRenderer(this);
+  video_tracks[0]->AddOrUpdateSink(this, rtc::VideoSinkWants());
 }
 
 WebrtcVideoRendererAdapter::~WebrtcVideoRendererAdapter() {
   DCHECK(task_runner_->BelongsToCurrentThread());
 }
 
-void WebrtcVideoRendererAdapter::RenderFrame(const cricket::VideoFrame* frame) {
-  // TODO(sergeyu): WebRTC calls RenderFrame on a separate thread it creates.
+void WebrtcVideoRendererAdapter::OnFrame(const cricket::VideoFrame& frame) {
+  // TODO(sergeyu): WebRTC calls OnFrame on a separate thread it creates.
   // FrameConsumer normally expects to be called on the network thread, so we
   // cannot call FrameConsumer::AllocateFrame() here and instead
   // BasicDesktopFrame is created directly. This will not work correctly with
   // all FrameConsumer implementations. Fix this somehow.
   scoped_ptr<webrtc::DesktopFrame> rgb_frame(new webrtc::BasicDesktopFrame(
-      webrtc::DesktopSize(frame->GetWidth(), frame->GetHeight())));
+      webrtc::DesktopSize(frame.GetWidth(), frame.GetHeight())));
 
-  frame->ConvertToRgbBuffer(
+  frame.ConvertToRgbBuffer(
       output_format_fourcc_, rgb_frame->data(),
       std::abs(rgb_frame->stride()) * rgb_frame->size().height(),
       rgb_frame->stride());