*** Original commit ***
Rollback of 70273a0361
*** Original commit ***
Assert incoming buffers are little endian in the DefaultAudioSink.
***
***
PiperOrigin-RevId: 310889382
When operating the MediaCodec in asynchronous mode, we are seeing
NPEs raised from the DefaultAudioSink because output buffers are pushed
to the DefaultAudioSink with the DefaultAudioSink not configured yet.
This is possible to happen if an output buffer is passed with
MediaCodecAudioRenderer.processOutputBuffer() before an output format
change.
One possible code path to trigger this is when an output format change
is pending and we flush MediaCodec (e.g. for a seek): the callback is
waiting in the looper's queue but we ignore all currently queued
callbacks after a flush().
This commit checks for a pending output format change during a flush():
if one exists, and the next MediaCodec output callback right after the
flush() is an output buffer (and not a new output format), then the
pending output format is propagated first.
The only adapter that needs to change is the
AsynchronousMediaCodecAdapter which previsouly deleted all pending
callbacks immediately on flush(). The AsynchronousMediaCodecAdapter now
needs to also handle every enqueued callback in order to identify is
there's a pending output format change.
Testing: added unit tests for the new code. I have verified that if we
pass an output buffer to the audio renderer before the output format, it
will result on the specific NPE, but I cannot reproduce the base
scenario (an output format change is pending when flushing) on a real
device because we can't know if an output format is indeed pending.
PiperOrigin-RevId: 310885283
We currently propagate only ColorInfo, but propagating the whole Format
is generally useful (e.g., to get the frame-rate on the output side of
the decoder).
PiperOrigin-RevId: 310359650
This allows properties to propagate when switching view types
(e.g. bottomPaddingFraction).
It also allows the style-stripping code to be pushed up to SubtitleView
and therefore shared.
PiperOrigin-RevId: 310353534
Fixes AsynchronousMediaCodecBufferEnqueuerTest broken tests by
enqueueing input buffers that have previously been dequeued from the
MediaCodec.
The test assumes that the shadow MediaCodec implementation can dequeue
at least 10 input buffers before queueing them back. Although fragile,
it seems to work with the current robolectric shadow MediaCodec. This is
at the moment preferred compared to making the test more complicated.
PiperOrigin-RevId: 310325096
Some part of the audio pipeline in the DefaultAudioSink
(PCM processors) are expecting little endian buffers.
As a result, in order to behave like MediaCodec,
make sure the passthrough MediaCodec-bypass pipeline
also processes little endian output buffers.
PiperOrigin-RevId: 310172464
Previously the input format values were used,
but it could be incorrect if two format change were
occurring in quick successions.
PiperOrigin-RevId: 310142675
These methods operate on a thread-safe collection and are thus
inherently thread-safe. This simplifies some player setup configurations
where initial listeners are added on a background thread (e.g. when
using a dependency injection framework).
PiperOrigin-RevId: 310113181
Users of addTextOutput should instead query the current cues if they
need them. This is more consistent with how other listeners are handled.
PiperOrigin-RevId: 310112241
We currently have DecoderVideoRenderer and VideoDecoderRenderer, which
is very confusing! This one is package private, so we can rename it to
remove some of the confusion.
Also fix some nullness issues.
PiperOrigin-RevId: 309964088
In passthrough MediaCodec is not used except if
the format has a DRM.
Nevertheless the code was still requiring that a
passthrough decoder be present even if it was not going to be
used (aka no drm).
PiperOrigin-RevId: 309947271
- Fix DecoderAudioRenderer to re-init codec if the DRM session changes.
- Add canKeepCodec to DecoderVideoRenderer. Previously it was assumed
that the decoder could be re-used, but this will not be true in all
cases for FfmpegVideoRenderer.
Issue: #7079
PiperOrigin-RevId: 309935278
If the condition isn't fulfilled, they currently block until the
test runner times out the test. Our usual approach is to timeout
in the test itself so that the error message is clearly showing the
blocked condition.
Also clean-up some documentation.
PiperOrigin-RevId: 309930198
Now that MediaCodec is not use in passthrough, no
MediaCodec should be created in this mode.
Additionally, do not instantiate a MediaCodec in passthrough
#exo-offload
PiperOrigin-RevId: 309916131
DownloadManager doesn't know enough about the Downloader implementations to
know that interrupting the thread is the right thing to do. In particular,
if a Downloader implementation wants to delegate work to additional worker
threads, then it will probably not want its main thread to be interrupted
on cancelation. Instead, it will want to cancel/interrupt its worker threads,
and keep its main thread blocked until work on those worker threads has
stopped (download() must not return whilst the Downloader is still touching
the cache).
This change moves control over what happens to the individual Downloader
implementations.
Issue: #5978
PiperOrigin-RevId: 309419177
- Stop throwing InterruptedException from CacheUtil. When a CacheUtil operation
throws or returns, the caller should always check its own state to determine
whether they canceled the operation. If a caller is trying to catch
InterruptedException separately to IOException to do something different in
that case, they're probably doing the wrong thing. So it's simpler, and probably
less error prone, just to throw an IOException in the case of interruption.
- Throwing InterruptedIOException is also consistent with what our Extractor and
DataSource implementations do.
Issue: #5978
PiperOrigin-RevId: 309411556
We keep an index hint for the next pending player message. This hint
wasn't updated correctly when messages are removed due to a timeline
update.
This change makes sure to only use the hint locally in one method so
that it doesn't need to be updated anywhere else and also adds the "hint"
suffix to the variable name to make it clearer that it's just a hint and
there are no guarantees this index actually exists anymore.
issue:#7278
PiperOrigin-RevId: 309217614
testSubtitleEventTimes/IndicesHelper make assertions that only happen
to be the same because values in two different constants match up. It
seems much better to explicitly put the assertions in each test.
The other assert methods are just obscuring the underlying call to
Truth.
PiperOrigin-RevId: 309022044
Steps 4-10 of https://www.w3.org/TR/webvtt1/#cue-computed-line
This part is harder to fit into our code structure because it depends on
how many cues are simultaneously visible - so it has to go in
WebvttSubtitle not WebvttCueParser (which only deals with individual
cues in isolation).
This removes the `isNormal()` method that was trying to approximate
the correct behaviour.
PiperOrigin-RevId: 309021686
adPlaybackState is now non-null, and the uninitialized case is covered by a new
boolean hasAdPlaybackState. Position progress updates are now non-null and
initialized with IMA's VIDEO_TIME_NOT_READY constant.
Also fix some misc code issues:
- Remove empty branch for SmoothStreaming (Android Studio warns about this).
- Tidy onTimelineChanged and onPositionDiscontinuity and the methods they call
to improve naming.
- Remove logging for IMA events after release, as these methods are expected to
be called in the current IMA SDK behavior.
PiperOrigin-RevId: 308977116
The player's initial PlaylistTimeline has no ad cue points, so its first
MediaPeriodId has no nextAdGroupIndex. When the AdsMediaSource provides its
first timeline, the cue points become known. For the case where a preroll cue
point appeared, the preroll was detected due to isAd changing on the
MediaPeriodId. For the case where a midroll/postroll cue point appeared, the
MediaPeriodId was actually treated as the same, which leads to keeping the
unclipped original MediaPeriod.
Fix this behavior by checking for nextAdGroupIndex becoming set or decreasing.
PiperOrigin-RevId: 308974490
Its only real purpose is to encapsulate a download ID, but it doesn't actually save
any code, and arguably makes it more complicated by having multiple lists of
Downloader instances, indexed by key, and also because it's another (undocumented)
class to understand.
This CL retains the multiple Downloader lists, but they will be flattened in a
child CL.
PiperOrigin-RevId: 308838026
This is equivalent to the method ExoPlayer provides. It's nice for consistency,
and for retrieving the looper from test code.
PiperOrigin-RevId: 308830288
DownloadRunner.getDownloader was creating Downloader instances before the actual
DownloadManager under test would have created them. Tests were then asserting and
manipulating these Downloader instances, which is quite confusing. In particular
FakeDownloader.assertDoesNotStart() is an assertion on a Downloader instance that
only makes sense when called on an instance that DownloadManager would not have
created by itself.
This change replaces FakeDownloader.assertDoesNotStart() with an assertion on
DownloadRunner that no Downloader instance has been created.
PiperOrigin-RevId: 308822398
- Assertion chaining is generally discouraged. For example, because it's harder
to determine which assertion failed given a line number.
- Also removed chaining of the form obj.actionX().assertY(), because it's easy
for someone editing the test to accidentally delete actionX() when deleting
assertY(), where-as actionX() may often be important for subsequent assertions.
PiperOrigin-RevId: 308820503
- Remove assertReleased and replace it with a proper condition variable
that's opened when Downloader.download or Download.remove finish. As
far as I can tell assertReleased was basically implementing "sleep for
10 seconds after the Downloader starts". Note fixing this properly
also makes the tests run much faster!
- Use ConditionVariable instead of CountDownLatch(1).
- Use AtomicInteger instead of volatile int because it's clearer and
allows removal of explanatory comments.
PiperOrigin-RevId: 308819204
- Improve documentation explaining the benefits of ExoPlayer's ConditionVariable
over the one that the platform provides
- Allow Clock injection
- Create TestUtil method for obtaining a ConditionVariable whose block(long)
method times out correctly when used in a Robolectric test
- Add basic unit tests for ConditionVariable
PiperOrigin-RevId: 308812698
A previous change added these getters to CacheDataSource, but it can
also be useful to retrieve these components directly from the factory.
This is useful for tasks where we're going to need to build multiple
CacheDataSource instances (e.g., to make requests in parallel), and
also need to operate directly on the same components. It's a bit more
natural to retrieve them from the factory than from an arbitrary
CacheDataSource in this case, since it can avoid unnatural code where
you create a CacheDataSource instance earlier than you would otherwise
just to use its getters, and/or create one just to use its getters and
then throw it away.
PiperOrigin-RevId: 308606020
- Executor is a superclass of ExecutorService, so this is arguably a little
more flexible.
- It removes the need to use null for direct execution, because Runnable::run
is a direct executor that can be trivially used instead.
- Removing the error-prone "cannot be a direct executor" restriction in the
parallel version of SegmentDownloader requires not relying on the futures
returned from ExecutorService.submit() anyway.
Issue: #5978
PiperOrigin-RevId: 308586620
Currently, this method is only supposed to be called before removing
the listener from the player or when releasing the player.
If called at other times, it will throw an exception later when
a playback session is ended automatically.
issue:#7193
PiperOrigin-RevId: 308254993
The method is used to finish a brief "seeking" state that can be removed entirely
now state changes for seeking are masked.
PiperOrigin-RevId: 308237009
Unmarshal from json to MediaItem instead of Sample. Further the playlist
of MediaItems is converted to Intent extras which are read by the
PlayerActivity.
PiperOrigin-RevId: 308141231
While most ExoPlayer code parsing ByteBuffers is called with buffers in big
endian, in certain situation, buffers in little endian are used too.
MediaCodec produced ByteBuffers are in little endian, while buffers
receive from the sources are in big endian (ByteBuffer's default).
As a result, some code called from AudioSink in passthrough parsed
bytebuffer in little endian. This is not correct because those
format are specified in BigEndian.
Changing the endianness of the ByteBuffer returned from MediaCodec
would impact a lot more code that can currently be tested in the
current COVID lockdown situation.
As a result, this patch instead make the parsing code independent
of the ByteBuffer.order() set. All the code that is called from
DefaultAudioSink now parses the buffer explicitly in Big Endian.
Additionally, the MPEG big endian header data of size 4 bytes was
retrieved with ByteBuffer.get, which only returns one byte.
PiperOrigin-RevId: 308116173
The first session should only be created once we have the media items
and/or called prepare. Otherwise the first session is created with
an EventTime having an empty timeline making it less useful.
issue:#7193
PiperOrigin-RevId: 308100555
Something that helps a constructor always seemed a bit strange.
It's now possible to use CacheDataSource.Factory directly instead.
PiperOrigin-RevId: 307661930
In the same way it made sense to tie CacheKeyFactory and CacheDataSource
together in
7ea83d7167,
it makes sense to tie the PriorityTaskManager to the CacheDataSource as
well. This prevents error prone scenarios where one can end up being
passed around without (or with the wrong instance of) the other.
This change also internalizes use of PriorityDataSource directly into
CacheDataSource, rather than requiring the caller to chain things
themselves.
PiperOrigin-RevId: 307647290
Audio processors are now flushed twice after reconfiguration.
The second flush call cleared the pending trim start bytes so
transitions between tracks were no longer gapless.
Fix this by removing logic to clear pending trim bytes on flush.
As a result we may trim data incorrectly if there is a flush
before any data has been handled for seeking to a non-zero
position, but this edge case will happen rarely and the effect
shouldn't be noticeable.
PiperOrigin-RevId: 307344357
This is the missing attribute to support all features of the Sample with MediaItem. Hence PlayerActivity can use the setMediaItems() method directly without creating actual media sources in the app code.
PiperOrigin-RevId: 307102036
