Initial support for the desktop media pipeline on Android.

media/ffmpeg/ffmpeg_common.cc

Index: media/ffmpeg/ffmpeg_common.cc
diff --git a/media/ffmpeg/ffmpeg_common.cc b/media/ffmpeg/ffmpeg_common.cc
index cb806c74812aba84ee66885a36f746f161ad0129..13c1f5f26f2f84d3262480846be7fa25c9b9f2d6 100644
--- a/media/ffmpeg/ffmpeg_common.cc
+++ b/media/ffmpeg/ffmpeg_common.cc
@@ -407,11 +407,12 @@ void AVStreamToVideoDecoderConfig(
   }
 
   VideoFrame::Format format = PixelFormatToVideoFormat(stream->codec->pix_fmt);
-  if (codec == kCodecVP9) {
-    // TODO(tomfinegan): libavcodec doesn't know about VP9.

[liberato (no reviews please), 2015/07/10 17:15:09]

maybe keep the TODO.

+  // The format and coded size may be unknown by FFmpeg for VP9, and on Android
+  // where FFmpeg is compiled without parsers.
+  if (format == VideoFrame::UNKNOWN)

[liberato (no reviews please), 2015/07/10 17:15:09]

setting any unknown format to YV12 seems like it could go pretty wrong.  could
mess up UMA stats if nothing else.

at least, it needs to be protected for the android case.  i'm a little surprised
that it's okay for VP9.  doesn't VP9 support alpha?

how much does this setting actually affect?  it looks like there are a fair
number of consumers.

[watk, 2015/07/10 18:51:26]

AFAIK this is ok, but I'm not sure enough, so I'll make it Android only. I think
the right longer term fix is to allow VideoDecoderConfigs to contain unknowns,
but that's a larger change which requires auditing the consumers.

     format = VideoFrame::YV12;
+  if (coded_size == gfx::Size(0, 0))

[liberato (no reviews please), 2015/07/10 17:15:09]

also seems scary to do for any unknown size.

[watk, 2015/07/10 18:51:25]

Done.

     coded_size = visible_rect.size();
-  }
 
   // Pad out |coded_size| for subsampled YUV formats.
   if (format != VideoFrame::YV24) {