There is no point in having this parameter as the tag should always be a single
immutable object instantiated at the time the Timeline is created or earlier.
So there is no preformance benefit and it's error-prone in case people
forget to set setTag=true.
PiperOrigin-RevId: 264117041
This matches the documentation on MetadataDecoder.decode:
"@return The decoded metadata object, or null if the metadata could not be decoded."
PiperOrigin-RevId: 263767144
We no longer need two modules as AndroidX-Test takes care of the system
abstraction and we no longer have Robolectric Handler/Looper workarounds.
PiperOrigin-RevId: 262363201
If we keep streams in chunk sources after selecting new tracks, we also keep
a reference to a stale disabled TrackSelection object. Fix this by updating
the TrackSelection object when keeping the stream. The static part of the
selection (i.e. the subset of selected tracks) stays the same in all cases.
Issue:#6256
PiperOrigin-RevId: 261696082
- Remove manifest argument from callbacks of Player.EventListener and
SourceInfoRefreshListener. Instead make it accessible through
Player.getCurrentManifest() and Timeline.Window.manifest.
- Fix all MediaSource implementation to include the manifest in the
Timeline instead of passing it to the SourceInfoRefreshListener.
- Refactor ExoPlayerTestRunner, FakeTimeline, FakeMediaSource to
reflect these changes and make tests pass.
PiperOrigin-RevId: 257359662
Removes the need for duplicate calls to SampleQueue#read when
implementing DecryptionResources acquisition in the MediaSources.
PiperOrigin-RevId: 250298175
We are currently defaulting to targetSdk=1 as no targetSdk is specified. Only
tests which explicitly ask for another SDK use another test SDK. With the
versioned manifest, all tests run using the targetSDK by default.
PiperOrigin-RevId: 249060796
Also re-enable two of them, although note that the fix here is
not related to the flakiness that caused them to be disabled.
I'm re-enabling them since much has changed in DownloadManager,
and the hope is that these tests are no longer flaky.
PiperOrigin-RevId: 244826225
- We had buildAddRequest and sendNewDownload. Converged to
buildAddDownload and sendAddDownload.
- Also fixed a few more inconsistencies, and brought the
action constants into line as well.
PiperOrigin-RevId: 244274041
- Listener based reporting of progress allows the content length
to be persisted into the download index (and notified via a
download state change) as soon as it's available.
- Moved contentLength back into Download proper. It should only
ever change once, so I'm not sure it belongs in the mutable part
of Download.
- Made a DownloadProgress class, for naming sanity.
PiperOrigin-RevId: 244242487
ShadowLooper -> ShadowLegacyLooper
ShadowRealisticLooper -> ShadowPausedLooper
ShadowBaseLooper -> ShadowLooper
And all public methods from ShadowLegacyLooper get pushed up to ShadowLooper
Pull Request: https://github.com/robolectric/robolectric/pull/4868
Copybara: OK
Also adjust Google3 tests using custom looper shadows where necessary.
Convert exoplayer to paused looper to eliminate reliance on custom shadows
PiperOrigin-RevId: 243839311
- Allows enforcing immutability, which in a future CL will allow
avoiding allocating a new array/list on every call.
- Also some left over doc cleanup from the DownloadState rename.
PiperOrigin-RevId: 243627352
The old domain automatically forwards to the new one. For consistency, change
all doc and code references regardless.
Also adds GitHub CNAME config file which configures our page for the custom
domain.
PiperOrigin-RevId: 243592110
If the returned retry delay is unset, we should use a DONT_RETRY_FATAL action
instead and dispatch the event with the "canceled" flag.
PiperOrigin-RevId: 243251554
Both manifest loads currently use the default retry delay although a
LoadErrorHandlingPolicy is provided to the MediaSource.
PiperOrigin-RevId: 242852672
- Fix ID being dropped when DownloadAction is serialized and
de-serialized as it's passed to DownloadService
- Properly set DownloadAction.id when building an action from
a DownloadState
- Make ID a required constructor argument. Else it's too easy
to not propagate it by accident.
PiperOrigin-RevId: 242457831
The one functional change here is to require subclasses to implement getForegroundNotification. It seems easier and less error prone to
force developers to implement it than it does to document the cases
in which they need to override it (which is most cases).
PiperOrigin-RevId: 240566737
This replaces the deprecated usages of RobolectricTestRunner and
RuntimeEnvironent and fully migrates the tests to androidx.
PiperOrigin-RevId: 238011667
Stop encoding/decoding presentation time as part of the message.
What's actually in emsg boxes is a presentation time delta,
which is why it's only 32 bits, and hence why it doesn't handle
large absolute timestamps. We were using this field to hold
absolute timestamps only for the purpose of passing presentation
times from DashManifestParser.parseEvent back to the calling
method. After this change, we return Pair<Long, EventMessage>
instead.
Issue: #5490
PiperOrigin-RevId: 233561731
This requires to prepare the media source and the periods in a small helper similar
to the metadata retriever. It also gets rid of the need to have abstract protected
methods to load the manifest, to extract the track groups and to convert to stream keys,
as this can now be handled by the media period.
PiperOrigin-RevId: 231385590
That's the same position set in MediaPeriod.prepare (where it may be removed
in the future).
Having the position at an earlier point is necessary to fix an
issue with lazy preparation in ConcatenatingMediaSource where the prepare
position was assumed to be known but MediaPeriod.prepare hasn't been called
yet.
Issue:#5350
PiperOrigin-RevId: 229756637
When the extra adaptation set of a switch group isn't defined in the manifest, we
currently assume it's the first adaptation group. This either leads to wrong grouping
or duplicate track groups.
Such a case may easily happen if the manifest is filtered such that only one of the
switch adaptation sets will be present in the manifest.
PiperOrigin-RevId: 229365379
They are not longer needed anywhere, are error-prone (because of threading
requirements), and complicate testing and using MediaSources without a player.
PiperOrigin-RevId: 227871157
- Enable GZIP for media playlist + encryption key chunk requests in
HLS, as we do during playback
- Pass around DataSpecs rather than Uris. This will be needed for if
we add manifest cacheKey support (which seems like a good idea for
completeness, if nothing else)
PiperOrigin-RevId: 224057139
Moving most of the logic to the base DownloaderHelper helps to implement track
selection for downloading in a single place instead of multiple places.
PiperOrigin-RevId: 223964869
A remove action will eventually just be the unique ID of the download
that should be removed. This is a step toward that goal.
PiperOrigin-RevId: 222832456
1. Pull up all subclasses of DownloadAction into DownloadAction
2. Add DownloaderFactory for Downloader instantiation, and DefaultDownloaderFactory
to replace the instantiation logic being removed from the DownloadAction
subclasses.
This change will upgrade existing action files gracefully (i.e. it does not
lose compatibility with the existing offline implementation, other than some
minor breaking changes to the API).
TODOs:
1. Move test methods from the XDownloadActionTest classes into DownloadActionTest.
This will be done in a subsequent CL. There's a lot of consolidation that can
be done here, including de-duplicating some of the test code added in this CL.
2. Look at merging DownloaderConstructorHelper into DefaultDownloaderFactory.
3. Use customCacheKey in DASH/HLS/SS Downloaders, for completeness.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=222258983
The remaining work is to split Window.isDynamic so that it's
possible to represent a window that wont be appended to, but
may still be removed from.
Issue: #4780
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=221439220
Robustness fix to make sure we ignore tags with known names, but
which are nested inside of unknown tags. For example we don't want
to parse the third period in:
<Period>...</Period>
<Period>...</Period>
<CustomTag>
<Period>...</Period>
</CustomTag>
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=217101630
- Always clip to period duration for the last chunk. We previously
did this only when the last chunk explicitly exceeded the period
end time. We now also do it when the chunk claims to end at the
period boundary, but still contains samples that exceed it.
- If pendingResetPositionUs == chunk.startTimeUs == 0 but the
chunk still contains samples with negative timestamps, we now
clip them by setting the decode only flag. Previously we only
clipped such samples if the first chunk explicitly preceeded the
start of the period.
Issue: #4899
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=215763467
Currently there is no way to disable (or reduce) the logcat output generated
by ExoPlayer.
Issue:#4665
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=213421072
The default character set is always UTF-8 anyway on Android, but
we don't want our code to behave any differently where it's not
(e.g. robolectric test runs could potentially run in an environment
where UTF-8 isn't the default?).
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=211953885
This provides the list of currently buffered media chunks and iterators over
the potential next chunks to the track selection. Having these two parameters
enables more advanced decision logic based on this data.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=210551812
The response headers of the last load are available from the loading source
when creating media source events and can be easily forwarded.
Issue:#4361
Issue:#4615
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=209900693
The MediaPeriodId with index is only properly defined together with a
timeline containing the index. Changing it to the period uid allows to use
the MediaPeriodId independent of the corresponding timeline.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=209430257
Also fix an NPE in FfmpegAudioRenderer if sampleMimeType wasn't set.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=208230947
For now, just allow any prefix (and don't check the namespace).
Checking the namespace requires turning on the XmlPullParser
namespace awareness feature, but this can only be done before
parsing, and I think requires us to specify the namespace when
querying other attributes. This would also risk breaking services
who currently don't set namespaces correctly.
It's possible we'll run into similar issues with the few other
tags (e.g. mspr:pro). We should revisit enabling namespace
awareness if/when we see this.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=207526883
EventSampleStream.readData required eventStreamUpdatable to be
false for END_OF_STREAM to be returned, but it was only being
set to false when the manifest became static. This change also
sets it to false if the period is not the last one in a dynamic
manifest. It also renames it the parameter to appendable, since
that's what it really represents. Events may still be removed
from the start of the EventStream as the period starts moving
out of the live window.
Issue: #4492
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=207517265
All known instances use DataSource as generic type and thus code can be simplified
by removing the generic type altogether.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=205798542
As the manifest and media data sources can be the same now, we can provide
a simplified constructor with just one data source factory.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=205658046
This removes the need to populate the entire period for the common usage of
getting the uid from the period.
Also add default implementation to get period by uid.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=205638265
Until now, the streams were released and re-enabled for each type of stream
(primary, event, embedded) in that order. That leads to problems when replacing
streams from one type to another (for example embedded to primary).
This change restructures the track selection to:
1. Release and reset all streams that need to be released or replaced.
1(a). Including embedded orphan streams.
2. Select new streams.
Issue:#4477
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=203751233
This wires up recent changes of passing a transfer listener to the MediaSource and
allowing DataSources to accept new listeners.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=203280818
This allows to use this queue in the track selection in the future.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=202954871
This will allow to provide available information about future chunks to the
track selection.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=202940994
In the future, this allows to register the BandwidthMeter (managed by the player)
as a listener to all media transfers related to this media source.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=202643946
This url is readily available when creating media source events (from the
data source) but so far not published to external listeners. This change
adds a new field to LoadEventInfo which corresponds to DataSource.getUri().
Issue:#2054
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=202459049
Also moved shared code to SegmentDownloadAction between its subclasses.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=202294880
Both values are helpful for event reporting, but are only available while
the data source is open. Similar to bytesLoaded, they need to be reported
through the Chunk.
Issue:#2054
Issue:#4361
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=201664907
The helper method only reported DATA_TYPE_UNKNOWN even if the actual type
is known.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=200067296
Using new Handler(this) in a constructor potentially leaks an uninitialized
object. This is mostly not a problem because we don't use the Handler within
the constructor. Added a Util method to keep the warning suppression local.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=199605377
If a MediaPeriod uses a Loadable, then there are typically
reference chains of the form:
LoadingThread[GCroot]->Loadable->MediaPeriod->Player
Where the player is the MediaPeriod callback. When the
player is released, this reference chain prevents the
player from being GC'd until Loadable cancellation
completes, which may not always be fast. This in turn
will typically prevent the application's activity from
being GC'd, since it'll normally be registered as a
listener on the player (directly or indirectly via
something like a view).
This change mitigates the issue by removing references
that the MediaPeriod holds back to the player. The
MediaPeriod will still not be eligible for GC, but the
player and application activity will be, which in most
cases will be most of the leak (in terms of size).
Issue: #4249
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=199143646
The lint error suppression only works locally and not for external developers
who still see the lint error and need to suppress it themselves.
This changes 'implementation' to 'compileOnly' in gradle to prevent the
dependency from being exported. Also removes the local lint suppression.
Issue:#4234
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=196251407
This is useful to get hold of the manifest to then obtain DRM init data in
the download flow for protected content (without having to download the
manifest again).
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=196168938
This type of flicker was visible (at least on my S8) whenever starting
the demo app process (from being properly killed) when other notifications
(e.g. completed download notification) was present.
This change passes whether the service was started as foreground through
the intent, and only applies the show-notification-on-stop hack if it
was. It's only necessary to start as foreground if your app is not already
in the foreground, so it's not necessary to do this from activity/ui
components.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=195594930
Also convert them to exposing periods and track groups, like
regular MediaSources do. This gets us much closer to being
able to use standard track selection components during offline
initialization. The helper is responsible for reverse mapping
selected tracks onto physical streams when generating the
download action. This is trivial except for the HLS case, which
is a TODO for now.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=195500826
This may be preferable for using custom binary data,
and it's still easy to store Strings if needed.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=195486817
- Made parts of DownloadActions public that were previously hidden.
This is necessary so that when an app is listening to the
DownloadManager, it can tell something about the tasks.
- Simplified DownloadActivity some more. The helpers may migrate
into their corresponding modules at some point, and also be
converted to return structured TrackGroupArrays rather than a
flat list of Formats.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=195269824
Also add a build rule to simplify the test setup with a custom blacklist.
In addition, the test path pattern are made less specific to capture all the
static nullness tests in the /main/ directories.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=195230334
This includes only the (hopefully) non-debatable changes for the DASH module
and all needed changes for call into the core library.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=195097791
A single master version prevents app developers from providing
custom download functionality (because they can't increment the
version). It's also error prone to expect someone modifying the
DASH action to remember to update the ActionFile master version.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=194981886
This was adding a lot of code, and the multiple use cases
for Downloader was pretty confusing (in particular the
ordering of method calls was unclear). It's also not
performant (e.g. it requires loading/parsing manifest(s)
and initialization segments from disk).
In practice I think apps will need to keep a record of
what's offlined in their app's database (or equivalent),
which they can update by registering as a listener on
DownloadManager. This will be done for the demo app in
a subsequent change.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=194932876
- It's much cleaner to split completed/error notification methods.
- Make error notification show the content title in the demo app.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=194920507
This change is intended to resolve overloading of "Download",
where a DownloadTask could be an actual download task, or a
remove task.
Also cleaned up some documentation.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=194815058
- Redefine Scheduler interface to better describe what implementations
do. The previous version was too general, in that it allowed concrete
DownloadService implementations to pass different Requirements to
the base class and to the Scheduler. It's also difficult to see how
that version could ever support dynamic updates to Requirements, which
is probably a feature we'll need to add quite soon.
- Fix a (probably theoretical) problem where static fields in
DownloadService assumed only a single concrete implementation.
- Stop using PlatformScheduler pre-API-21 in demo app, because it will
fail.
- Define default Requirements that require network.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=194785751
- When calculating the downloaded percentage in DASH, there was no
way to disambiguate between 0 of 0 segments being downloaded because
there are no cached indexes (i.e. 0% downloaded) and 0 of 0 segments
being downloaded because the index defines 0 segments (i.e. 100%
downloaded).
- Also replace use of NaN with a named constant.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=194453202
When we play downloaded content, we rely on the manifest filters
to produce a manifest that contains only the content that was
downloaded. It makes sense just to use the same filters during
download too, so we don't have to worry about any implementation
differences.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=194380704
ProgressiveDownloader never implemented this properly, and we
don't use it in DownloadManager/DownloadService, both of which
use a polling model. A polling model is also what's used elsewhere,
for example to query the current playback position.
This is effectively doing the TODO in ProgressiveDownloader.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=194214579
This removes a fair chunk of code. The line delta would
be more negative if it not for adding all the /* x= */
comments in DashDownloadActionTest!
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=194136689
This makes DownloadService easier to use in general and when only single
notification is used for all downloads.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=193165982
These don't seem to be needed anymore. All tests run without them in gradle
and Blaze.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=191867518
This field (formerly "id") is almost impossible to use so far. Having setters
in the factories allows to specify custom tags for all media sources.
Also added a ExoPlayer.getCurrentTag() method to retrieve the tag of the
currently playing media source in a convenient way.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=191738754
We currently refresh repeatedly in this case. According to the
DASH spec omitting minUpdatePeriod indicates that the manifest
does not change, and therefore we should not refresh. I think
it might be valid to omit minUpdatePeriod in a dynamic manifest
if relying exclusively on EMSGs to trigger manifest refresh.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=191420247
This change enables feeding decoders from the closest sync frame
before a specified seek position, where-as previously we'd
always feed decoders from the start of the chunk. This avoids
decoding and discarding many audio samples during each seek. The
same benefit also applies to video chunks containing more than
one key-frame.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=190539547
This adds callbacks for creating, releasing, and starting to read from media
periods.
Such events allow listeners to keep a list of active media periods. This is
useful to determine when no further events for a certain media period are
expected. It also allows listeners to associate renderer events unambigiously
with a reading media period.
-------------
Created by MOE: https://github.com/google/moe
MOE_MIGRATED_REVID=190462717
