SDK / exoplayer

Issue:#1782

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=136025847

A blocking call is necessary where we want to guarantee that
the player wont access the surface after the method call has
returned. We currently only do this for the case:

Surface->Null

But we should also do it for the case:

SurfaceA->SurfaceB

Since the caller may reasonably do something like destroy
SurfaceA immediately after it's been replaced.

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135921296

This allows the injectable reader factory to be a stateless factory, allows
the seeking to be consistent and will allow multiple CC channel support later
on.

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135909712

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135819142

Issue: #1907

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135798950

- Fix issue in ExoPlayerImpl where the timeline was null'd
  but onTimelineChanged was not fired.
- Add the ability to not reset the timeline. This is useful
  for retries where you know the timeline will be the same
  as it was previously.

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135797577

playingPeriodHolder can be null in the case that the first
period is still being prepared. We need to make sure we
release the period that's being prepared in such cases,
which is loadingPeriodHolder.

Issue: #1914

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135793472

- Move nearly all logic onto the calling thread (i.e. the thread
  calling open/read/close), to make threading correctness more
  obvious.
- Document which variables are read/written from which thread, and
  why the call sequences are safe.
- Fix thread safety issue that I think could probably cause data
  corruption in the case of a read timeout followed by another
  request into the DataSource.

Also:

- Relaxed content length checking to be consistent with the other
  http DataSource implementations, and avoided parsing the headers
  where they're not used.
- Fixed missing generics in CronetDataSourceFactory.
- Added TODO to work with servers that don't support partial range
  requests.

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135702217

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135700280

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135692709

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135682130

- Make sure no events are posted on PlaybackControlView
  if it's not attached to a window. This can cause leaks.
  The target hide time is recorded if necessary and
  processed when the view is re-attached.
- Deduplicated PlaybackControlView.VisibilityListener
  invocations.
- Fixed timeouts to be more intuitive (I think).
- Fixed initial visibility of PlaybackControlView when
  used as part of SimpleExoPlayerView.
- Made some more attributes configurable from layout xml.

Issue: #1908
Issue: #1919
Issue: #1923

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135679988

This is a minor cleanup. The main thing it fixes is that
the "Inconsistent headers" and "Unexpected Content-Range"
log messages were printing List<String> objects, rather
than the actual headers they are supposed to print.

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135378074

The access in fillCurrentRequestPostBody wasn't protected with
synchronization. Furthermore, just synchronizing it wouldn't be
sufficient, since what we really need to check is whether the
Content-Type header has been added to the UrlRequest.Builder.
The contents of requestProperties may have changed between the
headers being added to UrlRequest.Builder and the call to
fillCurrentRequestPostBody.

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135376904

- Allow null Content-Type in response headers.
- Inline validateResponse, just because it makes it clearer
  what thread it's being executed on when inlined.

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135375063

- Fix bug in getCurrentRequestStatus where we weren't
  blocking on the condition variable.
- Make methods static where possible.
- Clean up getUri implementation.

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135373586

This change is a no-op reodering, as a precursor to
further cleanup. The public methods are grouped by
the class/interface they implement. The private methods
are ordered with things that will become static in
a subsequent change at the bottom.

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135372629

Use Call.Factory instead of OkHttpClient

This allows using alternate implementation of an HTTP client. We can use
OkHttpClient as before as it implements Call.Factory.

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135353593

Without this developers which reuse a SurfaceHolder with multiple instances of
SimpleExoPlayer need to call simpleExoPlayer.setVideoSurfaceHolder(null) to get
the SimpleExoPlayer.ComponentListener removed from the surface holder. If they
don't, the component listener is still registered and as a member class leaks
an instance of simpleExoPlayer.

Issue #1855

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135292439

Playback would fail if a renderer is toggled from consuming from
one child to another in a single step.

Issue: #1900

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135270356

https://github.com/google/ExoPlayer/issues/1827

Three different modes available: fit (default), fixed_width, fixed_height
Developers need to use wrap_content for the dimension which is not fixed:

app:resize_mode="fixed_width"
android:layout_width="320dp"
android:layout_height="wrap_content"

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135264861

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135260806

Configuration of retry currently fails if all of the
following are true, which is highly unlikely but does
occur in the ref'd issue.

1. Loading/extraction fails
2. Neither length of stream of a seek map is known
3. At least one track has been output by the extractor

Issue: #1899

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135228687

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135215320

Issue: #1895

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135211179

Ref: https://blog.bintray.com/2015/02/09/android-studio-migration-from-maven-central-to-jcenter/

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135211072

- Enfroce read returns 0 if readLength==0 everywhere.
- Fixes and simplifications for CronetDataSource.

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135138232

- Make read return 0 if readLength==0
- Return RESULT_END_OF_INPUT properly in the case that
  bytesRemaining is unset (this was broken previously,
  but only applies for assets > 2^31 bytes, so it's
  unlikely anyone ever hit this issue)

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135136541

It'll be a TODO to clean up some of the commonality between
publishing this and the core library (e.g. the version number
should really be spec'd in one place only). Just getting this
submitted for now so that we don't lose track of how I created
https://bintray.com/google/exoplayer/extension-okhttp

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135102459

Issue:#1875

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135100986

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135100455

- Correctly null out streams[j] in the case that a renderer
  is being disabled.
- Read discontinuities from all children, not just enabled
  ones. This fixes a failure when reading a discontinuity
  with all renderers disabled.
- Add in some assertions to make incorrect stream selection
  failures obvious and immediate.
- Relocate subtitles so they're above the shutter (needed so
  they continue to be visible when video is disabled but
  text is still enabled).

Issue: #1854

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=135089944

This CL adds support for initialization segments in HLS. This is required mainly for(but not limited to) usage of fMP4. The fMP4 support only consists in creating the required extractor if the extension is .mp4, provided the initialization segment is correctly
loaded and passed to the extractor.

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=134997636

This prevents a large amount of memory from being held
in the case that a player instance is released, but the
application is holding dangling references to the player
that are preventing it from being garbage collected.

Issue: #1855

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=134992945

- Fix NPE issue in SingleSampleMediaPeriod.
- Delay handling of EOS in TextRenderer until the last
  subtitle is fully played out.

Issue: #1882

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=134979286

Added Basic CEA-708 caption support to ExoPlayer V2, roughly equivalent to the current state of CEA-608. It's still missing window (and the associated positioning and formatting) support, emulating a 3-line rollup.

Currently this has only been tested with the RawCC container, though it should work with anything that makes use of SeiReader (i.e. MPEG2-TS) with minimal changes to SeiReader.

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=134781660

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=134776852

When reading the last period, the readingPeriodHolder was set to null in
updatePeriods if it was the last period. (This would occur almost immediately
when playing a single-period source.) seekToPeriodPosition suppresses reusing a
loaded/prepared period if the reading period and playing period did not match,
which meant that the whole timeline was recreated when seeking in the last
period.

Leave readingPeriodHolder non-null. This means that at all times either
playingPeriodHolder == readingPeriodHolder (and they could be null or
non-null), or playingPeriodHolder and readingPeriodHolder differ and are both
non-null.

Also fix an issue where streams were never forced to be recreated during track
reselection when reading ahead.

-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=134774238