Handle the case when a mediastream track doesn't have extra data.

content/renderer/media/media_stream_renderer_factory_impl.cc

Index: content/renderer/media/media_stream_renderer_factory_impl.cc
diff --git a/content/renderer/media/media_stream_renderer_factory_impl.cc b/content/renderer/media/media_stream_renderer_factory_impl.cc
index 929519efdb5fdf922effe50a5d61fda0299a5c13..3163953062809700771066625430c89c483ddc6d 100644
--- a/content/renderer/media/media_stream_renderer_factory_impl.cc
+++ b/content/renderer/media/media_stream_renderer_factory_impl.cc
@@ -141,7 +141,15 @@ MediaStreamRendererFactoryImpl::GetAudioRenderer(
   // and mixes audio from all the tracks that belong to the media stream.
   // For now, we have separate renderers depending on if the first audio track
   // in the stream is local or remote.
-  if (MediaStreamTrack::GetTrack(audio_tracks[0])->is_local_track()) {
+  MediaStreamTrack* audio_track = MediaStreamTrack::GetTrack(audio_tracks[0]);
+  if (!audio_track) {
+    // This can happen if the track was cloned.
+    // TODO(tommi, perkj): Fix cloning of tracks to handle extra data too.

[perkj_chrome, 2015/12/10 12:44:44]

Is this an audio track problem in general or only webaudio? 

Cloning is triggered from here:
https://code.google.com/p/chromium/codesearch#chromium/src/third_party/WebKit...

And implemented in chrome here.
https://code.google.com/p/chromium/codesearch#chromium/src/content/renderer/m...

[tommi (sloooow) - chröme, 2015/12/10 13:37:37]

In general and I don't think this is an issue with webaudio tracks actually.
Yes I know and the problem with that implementation is that the extra data is
not cloned and there's no way to get to the original source of the original
track.  For remote tracks for example, there will never be a "native" source or
track.

+    LOG(ERROR) << "No native track for WebMediaStreamTrack.";
+    return nullptr;
+  }
+
+  if (audio_track->is_local_track()) {
     // TODO(xians): Add support for the case where the media stream contains
     // multiple audio tracks.
     return CreateLocalAudioRenderer(audio_tracks[0], render_frame_id, device_id,