- 27 May, 2021 10 commits
-
-
PiperOrigin-RevId: 375992816
aquilescanta committed -
PiperOrigin-RevId: 375992627
aquilescanta committed -
PiperOrigin-RevId: 375979170
bachinger committed -
PiperOrigin-RevId: 375919564
olly committed -
Issue: #8985 #minor-release PiperOrigin-RevId: 375913914
olly committed -
The method openSocket in RtspMessageChannel does not actually open a socket. The 'open' term refers more to opening the message channel. #minor-release PiperOrigin-RevId: 375908999
claincly committed -
The callbacks received RTSP messages and RTSP sending errors are now invoked directly from RtspMessageChannel's internal threads. It's up to the handler implementation to decide which thread to handle the messages. #minor-release PiperOrigin-RevId: 375908282
claincly committed -
Previously, RTSP interleaved binary data is posted onto the playback thread for handling, the playback thread then adds the received data to a queue. A loader thread will later dequeue the data and process it. In this CL, the binary data is sent through a separate listener, on RtspMessageChannel's RTSP receiving thread. #minor-release PiperOrigin-RevId: 375907609
claincly committed -
Some RTSP servers will offer multiple WWW-Authenticate options. We wanted to be able to pick them up. #minor-release PiperOrigin-RevId: 375907276
claincly committed -
Currently acquireSession() fails with an NPE from checkNotNull(exoMediaDrm). A follow-up change will result in exoMediaDrm sometimes being non-null while prepareCount==0 (and in this case we still want acquireSession() to fail). preacquireSession() doesn't currently fail in a way the caller can observe - the same NPE is thrown, but asynchronously and it doesn't propagate out of the background thread. Throwing directly seems preferable since it's a clear bug to be trying to preacquire sessions from an unprepared/released manager. PiperOrigin-RevId: 375906450
ibaker committed
-
- 26 May, 2021 14 commits
-
-
PiperOrigin-RevId: 375883879
gyumin committed -
PiperOrigin-RevId: 375830765
klhyun committed -
PiperOrigin-RevId: 375719035
aquilescanta committed -
PiperOrigin-RevId: 375705247
aquilescanta committed -
PiperOrigin-RevId: 375686596
aquilescanta committed -
#minor-release PiperOrigin-RevId: 375674759
samrobinson committed -
PiperOrigin-RevId: 375515685
aquilescanta committed -
PiperOrigin-RevId: 375514509
aquilescanta committed -
PiperOrigin-RevId: 375484765
Oliver Woodman committed -
Without this the Cue isn't deeply immutable, which can be a bit surprising. PiperOrigin-RevId: 375477571
ibaker committed -
This reduces the null-checking required in a follow-up change. PiperOrigin-RevId: 375468504
ibaker committed -
#minor-release PiperOrigin-RevId: 375436117
ibaker committed -
#minor-release PiperOrigin-RevId: 375435339
samrobinson committed -
PiperOrigin-RevId: 375133184
aquilescanta committed
-
- 21 May, 2021 9 commits
-
-
#minor-release PiperOrigin-RevId: 375097412
samrobinson committed -
Unlike Assertions, the introduced method cannot be disabled and throws a ParserException instead. This method is meant to replace regular assertions (which throw RuntimeExceptions) which check input in the parsing code. PiperOrigin-RevId: 375085160
aquilescanta committed -
Its absence seems to have caused https://github.com/google/ExoPlayer/issues/8969. #minor-release PiperOrigin-RevId: 375058222
krocard committed -
PiperOrigin-RevId: 375054051
olly committed -
We need to avoid reading and skipping into preload chunks as they may need to be discarded. The current code iterates over all chunks, but this can be simplified by just checking the last chunk knowing that the preload chunk must always be the last one. As a result, we avoid calling getFirstSampleIndex on all chunks. This is a bug since the method is not allowed to be called for chunks that have been spliced in. This still leaves the smaller issue of potentially calling this method for spliced-in preload chunks, which will be solved separately. Issue: #8937 #minor-release PiperOrigin-RevId: 375053638
tonihei committed -
Tweak the way directories are passed PiperOrigin-RevId: 374912886
olly committed -
PiperOrigin-RevId: 374885599
aquilescanta committed -
PiperOrigin-RevId: 374880929
aquilescanta committed -
PiperOrigin-RevId: 374874272
aquilescanta committed
-
- 20 May, 2021 7 commits
-
-
Preload chunks may still need to be discarded. However, we don't currently support discarding spliced-in chunks. Thus, we need to avoid loadng a preload chunk that needs to be spliced-in. Issue: #8937 #minor-release PiperOrigin-RevId: 374851661
tonihei committed -
PiperOrigin-RevId: 374847609
olly committed -
Colocate lines that refer to the same module, to reduce probability of accidentally removing one line and not the other (which has happened more than once :)). PiperOrigin-RevId: 374847082
olly committed -
This media source wraps another source and publishes a Timeline with ads. The created MediaPeriods for ad and content are mapped back to the original stream to allow seamless playback. PiperOrigin-RevId: 374836091
tonihei committed -
PiperOrigin-RevId: 374836046
tonihei committed -
There is a newly added condition to help advancing between SSAI ads and content in case the ad group position or ad duration changed. The condition currently doesn't check directly whether it's a SSAI transition but relies on indrect signals. Making this more direct helps to understand the purpose and avoid unintentional bugs where this condition would apply in other cases too. In addition, we need to exclude TextRenderer from the check because its read position doesn't correspond to the actual decode position since the decoding happens in the renderer itself (b/181312195). PiperOrigin-RevId: 374835985
tonihei committed -
We can instead just save this information in MediaPeriodInfo, similar to how we store whether the MediaPeriod is last in the timeline etc. PiperOrigin-RevId: 374835918
tonihei committed
-