[Mojo Video Capture] Replace method OnIncomingCapturedVideoFrame with OnIncomingCapturedBufferExt

media/capture/content/thread_safe_capture_oracle.cc

Index: media/capture/content/thread_safe_capture_oracle.cc
diff --git a/media/capture/content/thread_safe_capture_oracle.cc b/media/capture/content/thread_safe_capture_oracle.cc
index 9fc8f93b0f7e283eeee40ef89b188cd05d846b82..0ec69e047dfd15a0de82c9771a163da78f84d474 100644
--- a/media/capture/content/thread_safe_capture_oracle.cc
+++ b/media/capture/content/thread_safe_capture_oracle.cc
@@ -229,7 +229,12 @@ void ThreadSafeCaptureOracle::DidCaptureFrame(
     frame->metadata()->SetTimeTicks(VideoFrameMetadata::REFERENCE_TIME,
                                     reference_time);
 
-    client_->OnIncomingCapturedVideoFrame(std::move(buffer), std::move(frame));
+    media::VideoCaptureFormat format(frame->coded_size(),
+                                     params_.requested_format.frame_rate,
+                                     frame->format(), media::PIXEL_STORAGE_CPU);

[mcasas, 2016/12/14 21:04:27]

Are we sure that media::PIXEL_STORAGE_CPU is always
the case? IIRC, the capture machinery could produce
textures, but probably that path is gone now. Perhaps
we could add 
  DCHECK(frame->IsMappable());
to document this?

[chfremer, 2016/12/14 21:42:29]

Good point. Right now we are sure, because it is the only possible value for
this enum [1]. 

Added the DCHECK.

[1]
https://cs.chromium.org/chromium/src/media/capture/video_capture_types.h?q=PI...

+    client_->OnIncomingCapturedBufferExt(
+        std::move(buffer), format, frame->visible_rect(), reference_time,
+        frame->timestamp(), *frame->metadata());
   }
 }