Motown: Ffmpeg video decoder

Motown: Ffmpeg video decoder
This CL gets the ffmpeg video decoder working. The player and associated
components have been modified to move decoded video as far as the sink,
which drops the frames on the floor. The decoder has been tested with an
ogg/vorbis file. Play/pause/seek etc works, and decoded frames are
produced. The correctness of the decoded frames has not been verified.
Currently, the presentation timestamps on the frames are in the timebase
established by the decoder, which is undiscoverable. That needs to get
fixed before we can synchronize A/V.

R=kulakowski@chromium.org

Committed: https://chromium.googlesource.com/external/mojo/+/619a74f464c055c0be14d3766d97275150d0d0bf

[dalesat, 2016-04-27 00:01:26]

Description was changed from

==========
Motown: Ffmpeg video decoder
This CL gets the ffmpeg video decoder working. The player and associated
components have been modified to move decoded video as far as the sink,
which drops the frames on the floor. The decoder has been tested with an
ogg/vorbis file. Play/pause/seek etc works, and decoded frames are
produced. The correctness of the decoded frames has not been verified.
Currently, the presentation timestamps on the frames are in the timebase
established by the decoder, which is undiscoverable. That needs to get
fixed before we can synchronize A/V.
==========

to

==========
Motown: Ffmpeg video decoder
This CL gets the ffmpeg video decoder working. The player and associated
components have been modified to move decoded video as far as the sink,
which drops the frames on the floor. The decoder has been tested with an
ogg/vorbis file. Play/pause/seek etc works, and decoded frames are
produced. The correctness of the decoded frames has not been verified.
Currently, the presentation timestamps on the frames are in the timebase
established by the decoder, which is undiscoverable. That needs to get
fixed before we can synchronize A/V.
==========

[dalesat, 2016-04-27 00:01:26]

dalesat@chromium.org changed reviewers:
+ jeffbrown@google.com, kulakowski@chromium.org

[dalesat, 2016-04-27 00:01:41]

Please take a look. Thanks!

[kulakowski, 2016-04-27 02:16:30]

Rad!

lgtm with a couple nits inline.

https://codereview.chromium.org/1923763002/diff/1/services/media/framework/st...
File services/media/framework/stages/active_multistream_source_stage.h (right):

https://codereview.chromium.org/1923763002/diff/1/services/media/framework/st...
services/media/framework/stages/active_multistream_source_stage.h:56: bool
packet_request_outstanding_ = 0;
s/0/false/

https://codereview.chromium.org/1923763002/diff/1/services/media/framework_ff...
File services/media/framework_ffmpeg/ffmpeg_video_decoder.h (right):

https://codereview.chromium.org/1923763002/diff/1/services/media/framework_ff...
services/media/framework_ffmpeg/ffmpeg_video_decoder.h:49: PayloadAllocator*
allocator_;
Are PayloadAllocators some global-ish things? Or otherwise, who owns them?

https://codereview.chromium.org/1923763002/diff/1/services/media/framework_ff...
services/media/framework_ffmpeg/ffmpeg_video_decoder.h:54: // TODO(dalesat): For
investigation purposes only...remove one day.
Does this mean all three of the following members?

[dalesat, 2016-04-27 15:40:14]

Thanks, George!

https://codereview.chromium.org/1923763002/diff/1/services/media/framework/st...
File services/media/framework/stages/active_multistream_source_stage.h (right):

https://codereview.chromium.org/1923763002/diff/1/services/media/framework/st...
services/media/framework/stages/active_multistream_source_stage.h:56: bool
packet_request_outstanding_ = 0;
On 2016/04/27 02:16:30, kulakowski wrote:
> s/0/false/

Done.

https://codereview.chromium.org/1923763002/diff/1/services/media/framework_ff...
File services/media/framework_ffmpeg/ffmpeg_video_decoder.h (right):

https://codereview.chromium.org/1923763002/diff/1/services/media/framework_ff...
services/media/framework_ffmpeg/ffmpeg_video_decoder.h:49: PayloadAllocator*
allocator_;
On 2016/04/27 02:16:30, kulakowski wrote:
> Are PayloadAllocators some global-ish things? Or otherwise, who owns them?

PayloadAllocator implementations come in two forms: a component like
MojoProducer exposes PayloadAllocator so it can allocate out of a shared buffer,
and the vanilla allocator, which is singleton. I get that it's weird to use a
raw pointer these days. I'll probably change that, but as part of a general
overhaul of how allocators and packets are done. We're going to need things like
packets that refer to texture memory, etc.

https://codereview.chromium.org/1923763002/diff/1/services/media/framework_ff...
services/media/framework_ffmpeg/ffmpeg_video_decoder.h:54: // TODO(dalesat): For
investigation purposes only...remove one day.
On 2016/04/27 02:16:30, kulakowski wrote:
> Does this mean all three of the following members?

Yes. Changed the comment.

[dalesat, 2016-04-27 15:40:48]

Description was changed from

==========
Motown: Ffmpeg video decoder
This CL gets the ffmpeg video decoder working. The player and associated
components have been modified to move decoded video as far as the sink,
which drops the frames on the floor. The decoder has been tested with an
ogg/vorbis file. Play/pause/seek etc works, and decoded frames are
produced. The correctness of the decoded frames has not been verified.
Currently, the presentation timestamps on the frames are in the timebase
established by the decoder, which is undiscoverable. That needs to get
fixed before we can synchronize A/V.
==========

to

==========
Motown: Ffmpeg video decoder
This CL gets the ffmpeg video decoder working. The player and associated
components have been modified to move decoded video as far as the sink,
which drops the frames on the floor. The decoder has been tested with an
ogg/vorbis file. Play/pause/seek etc works, and decoded frames are
produced. The correctness of the decoded frames has not been verified.
Currently, the presentation timestamps on the frames are in the timebase
established by the decoder, which is undiscoverable. That needs to get
fixed before we can synchronize A/V.

R=kulakowski@chromium.org

Committed:
https://chromium.googlesource.com/external/mojo/+/619a74f464c055c0be14d3766d9...
==========

[dalesat, 2016-04-27 15:40:49]

Committed patchset #2 (id:20001) manually as
619a74f464c055c0be14d3766d97275150d0d0bf (presubmit successful).