wine/gecko
0
0
Fork 0
mirror of https://gitlab.winehq.org/wine/wine-gecko.git synced 2024-09-13 09:24:08 -07:00
Code Issues Packages Projects Releases Wiki Activity
263,714 Commits 1 Branch 166 Tags 1.2 GiB
ca00cd643b
Commit Graph

4366 Commits

Author SHA1 Message Date
Robert O'Callahan
602c99f084 Bug 1201393. Remove usage of FLAG_BLOCK_INPUT from MediaStreamAudioSourceNode. r=jwwang 
There's no reason why WebAudio should block an incoming MediaStream.
 2015-08-31 16:56:33 +12:00
Robert O'Callahan
f7d4a03fe1 Bug 1201393. Remove usage of FLAG_BLOCK_INPUT from MediaRecorder. r=jwwang 
FLAG_BLOCK_INPUT is not needed on mPipeStream because nothing should cause
mPipeStream to block.
 2015-08-31 16:54:39 +12:00
Robert O'Callahan
0fe3d71130 Bug 1201393. Remove usage of FLAG_BLOCK_OUTPUT from MediaRecorder. r=jwwang 
Blocking on the source stream will be treated as silence/no video change.
 2015-08-31 16:50:35 +12:00
Robert O'Callahan
cc6d903f10 Bug 1201393. Remove usage of FLAG_BLOCK_* from OutputStreamData::Connect. r=jwwang 
We don't want to block stream decoding on the output MediaStream, or vice
versa.
 2015-08-19 14:12:16 +12:00
JW Wang
cc480b29f6 Bug 1202533 - Fix naming convention of MediaSink::PlaybackParams. 2015-09-08 10:58:11 +08:00
Matthew Gregan
f722a68a07 Bug 1202332 - XiphExtradataToHeaders miscalculates final header length. r=derf 2015-09-08 11:33:16 +12:00
JW Wang
97825cb60f Bug 1199155. Part 3 - rename mDecodedStream to mStreamSink. r=roc. 2015-09-07 19:32:16 +08:00
JW Wang
aec83ebec8 Bug 1199155. Part 2 - fix includes and forward declarations. r=roc. 2015-09-07 19:32:04 +08:00
JW Wang
31e4b70339 Bug 1199155. Part 1 - move dom/media/DecodedStream.* to dom/media/mediasink/ and implement the interface of MediaSink. r=roc. 2015-09-07 19:31:51 +08:00
Jan-Ivar Bruaroey
1534994a7b Bug 1194640 - add NSPR logging of camera capabilities, r=jesup 2015-09-03 15:07:45 -04:00
Jan-Ivar Bruaroey
8fec18ab44 Bug 1201197 - add dedicated listener to enumerateDevices. r=jesup 2015-09-03 19:09:34 -04:00
JW Wang
d5a2e4ac81 Bug 1199121. Part 3 - remove unused code. r=kinetik. 2015-09-07 11:58:27 +08:00
JW Wang
8a8754ad00 Bug 1199121. Part 2 - handle the case where audio ends before video and switch to system clock for calculating playback position. r=kinetik. 2015-09-07 11:58:11 +08:00
JW Wang
47d2cfc104 Bug 1199121. Part 1 - add the ability to handle video-only streams to AudioSinkWrapper. r=kinetik. 
Note AudioSinkWrapper can now report correct playback position when the media is video-only or audio-only.
We will handle the case where audio ends before video in next patch where it needs to switch to system clock when audio reaches the end.
 2015-09-07 11:56:52 +08:00
JW Wang
84cc47fe2f Bug 1195158. Part 5 - 1. Fix insufficient includes and sort out include order. 2. Only disconnect |mTimedMetadataListener| when the state machine is created. r=cpearce. 2015-09-07 11:39:46 +08:00
JW Wang
d080b4eb18 Bug 1195158. Part 4 - remove unused code. r=cpearce. 2015-09-07 11:39:16 +08:00
JW Wang
70e1c2be8a Bug 1195158. Part 3 - connect listeners. r=cpearce. 
a. MediaMetadataManager is connected to MediaDecoderReader::mTimedMetadataEvent to receive TimedMetadata events.
b. OggReader publish TimedMetadata events through MediaDecoderReader::mTimedMetadataEvent.
c. MDSM calls MediaMetadataManager::DispatchMetadataIfNeeded to publish metadata if playback positoin reaches the publish time.
d. MediaDecoder is connected to MediaMetadataManager::mTimedMetadataEvent to receive TimedMetadata events.
e. MediaDecoder updates its metadata when TimedMetadata events are received.
 2015-09-07 11:38:34 +08:00
JW Wang
cbe9acb08a Bug 1195158. Part 2 - Have OggReader send TimedMetadata events through a event source instead of direct calls to AbstractMediaDecoder::QueueMetadata. r=cpearce. 2015-09-07 11:37:32 +08:00
JW Wang
1b01738902 Bug 1195158. Part 1 - Have MediaMetadataManager listen to an event source to receive TimedMetadata events. OggReader will send TimedMetadata events through an event source. This will break OggReader's dependency on AbstractMediaDecoder::QueueMetadata which then can be removed for it is against our goal to run all MediaDecoder's methods on the main thread. r=cpearce. 2015-09-07 11:37:27 +08:00
Jan-Ivar Bruaroey
ffe45a44b9 Bug 1187315 - Refactor out Constraints.webidl. r=bz 2015-07-29 01:16:19 -04:00
Andrew McCreight
8f1c29e597 Backed out changeset 351d5f864f9e for not compiling. 2015-09-04 09:52:24 -07:00
Andrew McCreight
d400b16911 Bug 1201271 - Warn about unused results for more methods of nsTArray. r=froydnj 
This leaves alone the AppendElement methods.
 2015-09-04 09:45:44 -07:00
Michael Froman
969f1bdddc Bug 1051052 - Made mid an outparam in JsepSession::AddLocalIceCandidate. r=bwc 2015-09-01 16:40:00 -05:00
Matt Woodrow
0fb8c742d7 Bug 1198202 - Increase hardware video decoding fallback threshold. r=ajones 2015-09-03 20:37:31 -04:00
Karl Tomlinson
c53cf6f454 back out 0b4230b29db5,c8d987aeaece from bug 1197028 2015-09-02 23:44:37 +12:00
Karl Tomlinson
1927dc5ad7 bug 1193922 skip SetIsOnlyNodeForContext after shutdown r=padenot 
This works around bug 1200514 but also avoids some unnecessary work.
 2015-09-01 18:20:24 +12:00
Benjamin Chen
73c1cf908b Bug 1197669 - Part3 - Ensure the start event comes before any blobs. r=jwwang 2015-08-27 15:35:30 +08:00
Benjamin Chen
38e06eae22 Bug 1197669 - Part2 - 1. Ensure the EOS is sent to MediaCodec. 2. mEndOfStream should be protected by monitor. r=sotaro 2015-09-01 11:25:35 +08:00
Benjamin Chen
bcc65c5e98 Bug 1197669 - Part1 - 1.Enable the testcase on B2G. 2. fix the mimetype check in testcase. r=jwwang 2015-08-27 11:23:12 +08:00
Gian-Carlo Pascutto
95f412e884 Bug 1200614 - Protect the capture engines array from concurrent access (during shutdown). r=jesup 2015-09-02 18:41:06 +02:00
Robert O'Callahan
ec9072a5d4 Bug 1200099. Stop using a distinct mPlaybackStream to play a media stream through an HTMLMediaElement. r=jwwang 2015-08-31 23:33:53 +12:00
Robert O'Callahan
713f74c164 Bug 1200099. Relax test assumptions to accommodate streams not blocking. r=jwwang 2015-09-01 10:47:48 +12:00
Robert O'Callahan
c427135a0f Bug 1200099. Add a test that captured cross-origin video streams render black. r=jwwang 2015-08-31 14:47:44 +12:00
Robert O'Callahan
4a6ce600ad Bug 1200099. Ensure mSameOriginMedia is propagated to DecodedStream even if we don't get a watch notification. r=jwwang 2015-08-31 14:50:54 +12:00
JW Wang
d94443302a Bug 1199104. Part 3 - use AudioSinkWrapper in MDSM. r=kinetik. 2015-09-02 14:14:40 +08:00
JW Wang
343c52fd45 Bug 1199104. Part 2 - create AudioSinkWrapper. r=kinetik. 2015-09-02 14:11:28 +08:00
JW Wang
caef5118a1 Bug 1199104. Part 1 - create MediaSink. r=kinetik. 2015-09-02 14:11:22 +08:00
Nils Ohlmeier [:drno]
e4ebc26da2 Bug 1200823 - s/test.steeplechase/test.testOptions.steeplechase/. r=bwc 2015-09-01 13:59:03 -07:00
Paul Adenot
5bb3175151 Bug 901633 - Part 16 - Remove another allocation in the sending side. r=jesup 2015-09-01 14:26:14 +02:00
Paul Adenot
69eaf292b1 Bug 901633 - Part 15 - Remove an allocation on the sending side, out of the packetizer. r=jesup 2015-09-01 14:26:13 +02:00
Paul Adenot
ba33a9b932 Bug 901633 - Part 14 - Add testing for our audio processing functions. r=jesup 2015-08-11 13:49:29 +02:00
Paul Adenot
6d221a8d51 Bug 901633 - Part 13 - Teach the resampler at the input of the MSG to dynamically change its channel count if needed. r=jesup 
When the audio comes from a PeerConnection, we don't know how many channels the
audio will have, and it can change anyways.
 2015-09-01 14:25:48 +02:00
Paul Adenot
3aa96e1277 Bug 901633 - Part 12 - Add a function to deinterleave and convert an audio buffer. r=jesup 2015-09-01 14:25:48 +02:00
Paul Adenot
a757399b09 Bug 901633 - Part 7 - Update AudioNodeExternalInputStream to use the new AudioChunk methods. r=karlt 2015-07-29 18:37:18 +02:00
Paul Adenot
3ac9a021b5 Bug 901633 - Part 6 - Update DelayBuffer to use the new AudioChunk methods. r=karlt 2015-07-30 11:27:26 +02:00
Paul Adenot
cc593f939c Bug 901633 - Part 4 - Update AudioNodeStream to use new chunk methods. r=roc 2015-07-29 18:36:47 +02:00
Paul Adenot
eb0c3b0588 Bug 901633 - Part 3 - Fix TrackEncoder to use the new AudioChunk methods. r=jesup 2015-07-29 18:25:34 +02:00
Paul Adenot
95425b9d7f Bug 901633 - Part 2 - Make AudioChannelFormat and AudioSegment more generic. r=roc 2015-07-29 18:24:15 +02:00
Paul Adenot
e0fe4ef9f5 Bug 901633 - Part 1 - Implement a generic audio packetizer. r=jesup 2015-07-30 13:51:57 +02:00
Julian Seward
f81b7f59e6 Bug 1199654 - DecodedStream::DecodedStream does not initialise mSameOrigin. r=jwwang. 2015-09-01 11:52:19 +02:00
Makoto Kato
3bdf3184ac Bug 1197673 - Use float type for SetAudioOutputVolume. r=eitan 2015-09-01 17:51:07 +09:00
Makoto Kato
0407933ac7 Bug 1191667 - Part 3. Call NotifyStartedPlaying to show audio indicator. r=eitan 2015-09-01 17:50:59 +09:00
Makoto Kato
f533a9e387 Bug 1191667 - Part 2. Call onVolumeChanged when changing volume. r=eitan 2015-09-01 17:50:55 +09:00
Makoto Kato
427140ed27 Bug 1191667 - Part 1. Add onVolumeChanged method to nsISpeechTaskCallback. r=eitan 2015-09-01 17:50:51 +09:00
Edwin Flores
38b8fab61e Bug 1172396 - Fix bustage... again. CLOSED TREE. - r=bustage 2015-09-01 18:40:36 +12:00
Edwin Flores
1dc30ef343 Bug 1172396 - Check for GMP resources from chrome process instead of content - r=cpearce 2015-09-01 17:31:48 +12:00
Edwin Flores
498e05de0d Bug 1172396 - Update GMP trial creation pref from chrome process - r=cpearce 2015-09-01 17:31:48 +12:00
Edwin Flores
467d4b55c2 Bug 1172396 - Make GMP initialisation synchronous - r=cpearce 2015-09-01 17:31:48 +12:00
Edwin Flores
0df1cb967d Bug 1172396 - Use a fixed node ID for GMP trial creation - r=cpearce 2015-09-01 17:31:48 +12:00
Edwin Flores
5c14fc51e3 Bug 1172396 - Use GMPVideoDecoderTrialCreator when e10s is enabled - r=cpearce 2015-09-01 17:31:48 +12:00
JW Wang
a6c017958a Bug 1179547 - release decoders when tests finish as well as cubeb streams to reduce the number of concurrent cubeb streams and the chance of OpenCubeb() errors. r=kinetik. 2015-09-01 12:02:55 +08:00
Jean-Yves Avenard
99fb2040fe Bug 1199878: [webm] P2. Hold one frames for which the duration can't be known or estimated. r=kinetik 2015-09-01 10:47:07 +12:00
Jean-Yves Avenard
661217ec10 Bug 1199878: [MSE/webm] Properly calculate media segment duration. r=kinetik 
We can know with certainty the duration of a block if we have a following one. We do not have to always rely on having a previous segment to estimate the duration.
 2015-09-01 10:47:07 +12:00
Ryan VanderMeulen
e8bfc92a9c Bug 1096400 - Skip webspeech recognition tests on B2G. 2015-08-31 09:37:10 -04:00
Gian-Carlo Pascutto
ee8084f185 Bug 1104616 - Proxy video capture access to the main process. r=jesup,glandium,mrbkap 2015-08-31 14:51:29 +02:00
Paul Adenot
6ad7466f95 Bug 1190676 - Part 11 - Fix consumers: Web Speech. r=anatal,eitan 2015-08-25 10:30:49 +02:00
Paul Adenot
cad916280a Bug 1190676 - Part 10 - Fix consumers: AudioDestinationNode. r=karlt 2015-08-25 10:30:43 +02:00
Paul Adenot
5a7579c7cd Bug 1190676 - Part 7 - Fix consumers: InputPort and HWMediaStream. r=jacheng 2015-08-25 10:29:51 +02:00
Paul Adenot
e862413472 Bug 1190676 - Part 6 - Fix consumers: MediaManager. r=jesup 2015-08-25 10:29:51 +02:00
Paul Adenot
61e3feb475 Bug 1190676 - Part 4 - Fix consumers: CanvasCaptureMediaStream. r=pehrsons 2015-08-25 10:29:50 +02:00
Paul Adenot
9f365dce46 Bug 1190676 - Part 3 - Make getting a MediaStreamGraph for a channel more explicit. r=roc 
Hopefully this wil also prevent getting the wrong graph.
 2015-08-25 10:17:31 +02:00
Paul Adenot
0585da350e Bug 1190676 - Part 2 - Force explicitely passing in a MediaStreamGraph when creating a DOMMediaStream. r=roc 
Hopefully this will prevent cross-graph connections.
 2015-08-25 10:16:41 +02:00
Chris Pearce
61f42f840e Bug 1176145 - Expose CDM version number in keySystem string. r=edwin 2015-08-31 16:08:51 +12:00
Chris Pearce
7d7325df97 Bug 1197050 - Move GMPDecoderModule up in PlatformDecoderModule::CreatePDM(). r=jya 2015-08-31 16:08:51 +12:00
Chris Pearce
e69cd0b2df Bug 1197064 - Don't assume positive audio frame offset in GMPAudioDecoder. r=jya 2015-08-31 16:08:51 +12:00
Chris Pearce
7f4a07237b Bug 1200059 - Make Adobe Primetime EME usable on MacOSX if available. r=edwin 2015-08-31 16:08:03 +12:00
Nicholas Nethercote
69d088e45f Bug 1198334 (part 1) - Replace the opt-in FAIL_ON_WARNINGS with the opt-out ALLOW_COMPILER_WARNINGS. r=glandium. 
The patch removes 455 occurrences of FAIL_ON_WARNINGS from moz.build files, and
adds 78 instances of ALLOW_COMPILER_WARNINGS. About half of those 78 are in
code we control and which should be removable with a little effort.
 2015-08-27 20:44:53 -07:00
Ryan VanderMeulen
ab22efa987 Bug 967606 - Skip test_call_start_from_end_handler.html on B2G opt. 2015-08-30 19:17:48 -04:00
Ryan VanderMeulen
7171b37fea Backed out changeset db8aa4fdcba4 (bug 1142455) for making bug 1179547 nearly permafail on OSX 10.6. 2015-08-30 18:38:14 -04:00
Ryan VanderMeulen
73e7010f8e Bug 1191270 - Skip test_audio_capture_error.html on B2G emulator builds for frequent failures. 2015-08-30 17:56:35 -04:00
Ryan VanderMeulen
c7f0c6c21a Bug 1193351 - Disable various failure-prone EME tests on B2G. 2015-08-30 17:52:04 -04:00
Ryan VanderMeulen
9b5547c8be Bug 1186406 - Skip test_eme_playback.html on Windows due to frequent intermittent failures. 2015-08-30 17:47:36 -04:00
Karl Tomlinson
81c893e29b bug 1199561 delay offline buffer allocation until non-null input is received r=padenot 
The avoids the allocation and copying altogether when the output buffer
is silent.
 2015-08-24 19:34:15 +12:00
Karl Tomlinson
cd84679515 bug 1199560 finish offline audio context processing even when allocation fails r=padenot 
ProcessBlock() needs to return aFinished = true at some point.
 2015-08-24 19:29:15 +12:00
Karl Tomlinson
97133980d7 bug 1199559 remove now unused SetRawChannelContents r=padenot 2015-08-25 11:46:30 +12:00
Karl Tomlinson
d313bdd34a bug 1199559 write audioprocess input buffer in a format suitable for direct use by AudioBuffer r=padenot 
saving copying.

This also avoids graph thread allocations when input is null.
 2015-08-28 18:30:17 +12:00
Karl Tomlinson
70dab929a1 bug 1199559 write decodeAudioData buffer in a format suitable for direct use by AudioBuffer r=padenot 
saving a copy.
 2015-08-28 11:15:39 +12:00
Karl Tomlinson
acdaad764e bug 1199559 write offline buffer in a format suitable for direct use by AudioBuffer r=padenot 
saving a copy.
 2015-08-27 21:32:41 +12:00
Karl Tomlinson
0883e6d3cb bug 1199559 add a factory method to accept generated buffer contents in a format suitable for direct use r=padenot 
This API lets AudioBuffer generators avoid a copy.

Future modifications to AudioBuffer will remove the unnecessary initial
allocation and zeroing.
 2015-08-28 18:30:49 +12:00
Karl Tomlinson
612514ebc9 bug 1199559 add a helper to fallibly allocate ThreadSharedFloatArrayBufferList with buffers r=padenot 2015-08-27 20:35:43 +12:00
Karl Tomlinson
d81f47ace2 bug 1199559 permit writing to ThreadSharedFloatArrayBufferList when not shared r=padenot 
Being able to write to the channel data in the buffer list saves the creator
from needing to manage its own pointers to the channel data.
 2015-08-25 08:41:14 +12:00
Jean-Yves Avenard
d0386ee336 Bug 1199904: Only start decoding ahead after explicitly requesting data. r=gerald 
Decode ahead will be stopped when seeking or when switching resolution.
 2015-08-30 20:48:09 +10:00
Jean-Yves Avenard
681637a427 Bug 1199879: [MSE] Use latest demux end time to detect discontinuities. r=gerald 
The ContainerParser doesn't always return an accurate end time.
 2015-08-30 20:48:07 +10:00
Jean-Yves Avenard
a0d7c39db3 Bug 1199573: [MSE] Properly handle partial media header received prior a discontinuity. r=gerald 2015-08-29 12:01:23 +10:00
Anthony Jones
e9a5132afe Bug 1142455 - Tweak MediaCache parameters; r=roc 2015-08-29 13:09:33 +12:00
Geoff Brown
f2b5e5a8ff Bug 1189784 - Disable more dom/media mochitests on Android for intermittent failures; r=me, test-only 2015-08-28 17:07:51 -06:00
Jan Gerber
14f2919cc1 Bug 1199531 - Disable test_BufferingWait.html on Android due to timeouts. r=rillian 
CLOSED TREE
 2015-08-28 08:21:00 -04:00
Nils Ohlmeier [:drno]
51d8639903 Bug 1167443 - Fix verification of end-of-candidates in mochitests. r=mt 2015-08-24 14:49:41 -07:00
Jean-Yves Avenard
986e495fea Bug 1197125: [webm]. Don't load the entire webm in RAM. r=kinetik 
Instead we parse it in chunks of 1MiB max.
 2015-08-28 14:06:35 +10:00
Jean-Yves Avenard
c746d650c6 Bug 1199032: [MSE/webm] P2. Construct WebMDemuxer with special MSE handling. r=kinetik 2015-08-28 14:06:34 +10:00
Jean-Yves Avenard
27fcb73603 Bug 1199032: [webm] P1. Explicitly differentiate WebM usage for mediasource. r=kinetik 
Relying on the data being fully buffered or not turned out to not be such a great idea.
 2015-08-28 14:06:33 +10:00
Karl Tomlinson
a9b63c13ee bug 1197028 re-use DelayNode output buffer r=padenot 2015-08-21 13:36:31 +12:00
Karl Tomlinson
f68937bd63 bug 1197028 release shared buffers from downstream so that upstream can re-use r=padenot 2015-08-21 13:26:58 +12:00
Ryan VanderMeulen
5c48f6d73a Merge m-c to inbound. a=merge 
CLOSED TREE
 2015-08-27 21:34:20 -04:00
Ryan VanderMeulen
df18bc931b Merge inbound to m-c. a=merge 2015-08-27 21:31:35 -04:00
Geoff Brown
d2fcbf75e6 Bug 1189784 - Disable some dom/media mochitests on Android 4.3 Debug; r=gcp 2015-08-27 11:29:02 -06:00
Geoff Brown
b53fd7faeb Bug 1189784 - Enable most dom/media mochitests on Android 4.3; r=gcp 2015-08-27 11:29:01 -06:00
Munro Chiang
991e55a8ee Bug 1178214 - Return INIT_ERROR when video resolution exceeds hw codec capability. r=sotaro 2015-08-24 17:39:34 +08:00
JW Wang
0667c9e3b3 Bug 1068596 - Pass SHARE_DELETE when opening a file so the file can be deleted without waiting for GC. r=roc 2015-08-27 08:14:07 -04:00
Ryan VanderMeulen
5e97b0c17b Backed out changeset d2e4ce298bab (bug 1068596) for Windows crashes. 2015-08-27 07:29:06 -04:00
JW Wang
ffdd7ed73d Bug 1068596 - pass SHARE_DELETE when opening a file so the file can be deleted without waiting for GC. r=roc. 2015-08-27 14:45:41 +08:00
Phil Ringnalda
810b604296 Back out 30a8392db2c2 (bug 1198094) for mQueuedSamples assertion failures 2015-08-26 22:09:40 -07:00
Jean-Yves Avenard
f5e672a07b Bug 1197977: [webm]. Ensure resource length calculation never underflows. r=kinetik 2015-08-27 12:28:20 +10:00
Jean-Yves Avenard
c5d0cf2b04 Bug 1198094: P1. Limit rate at which InputExhausted could be called by mac decoder. r=rillian 2015-08-27 12:28:19 +10:00
Kilik Kuo
e8231d7da0 Bug 1188268 - Correct the logic against |MaybeStartPlayback()| due to rebase mistake. r=jwwang 2015-08-27 14:20:18 +08:00
Gerald Squelart
733d1a4bb5 Bug 1194518 - Part 5: Using std::deque instead of nsTArray to store delayed frames. r=jya 2015-08-26 18:50:00 -04:00
Gerald Squelart
b3a0804e0e Bug 1194518 - Part 4: Using delaying wrapper according to pref 'media.fuzz.video-decode-minimum-frame-interval-ms'. r=jya 2015-08-26 18:49:00 -04:00
Gerald Squelart
21ef5b1d48 Bug 1194518 - Part 3: Delaying decoder wrapper, ensures a decoder appears consistently slow. r=jya 2015-08-26 18:47:00 -04:00
Gerald Squelart
45a65bc063 Bug 1194518 - Part 2: Using passthrough wrapper if pref 'media.fuzz.video-decode-passthrough' is true. r=jya 2015-08-26 18:45:00 -04:00
Gerald Squelart
e93d207d79 Bug 1194518 - Part 1: Passthrough decoder wrapper, useful to spy on MediaFormatReader-decoder calls. r=jya 2015-08-26 18:43:00 -04:00
Jan Gerber
391ef0ee53 Bug 1190776 - Re-enable test_BufferingWait.html. r=rillian 2015-08-25 12:51:17 +02:00
Eugen Sawin
1703cb2f08 Bug 1194014 - Remove redundant includes. r=jya 2015-08-27 15:03:49 +02:00
Eugen Sawin
6ddb881dd2 Bug 1194014 - Fix erroneous resource lookup and potential element overwrite. r=jya 2015-08-27 15:03:49 +02:00
Eugen Sawin
4d8b64229c Bug 1194014 - Fix MediaResource leaking caused by redundant AddRef. r=jya 2015-08-27 15:03:49 +02:00
Jean-Yves Avenard
335d11f041 Bug 1199193: Ensure DrainComplete() is called once all decoded frames have been output. r=rillian 2015-08-28 10:00:35 +10:00
Jean-Yves Avenard
d64335fd70 Bug 1198094: P1. Limit rate at which InputExhausted could be called by mac decoder. r=rillian 2015-08-28 10:00:34 +10:00
Matthew Gregan
d206238fdb Bug 1199069 - Remove ancient deopt from AudioStream required for MSVC2010. r=jwwang 2015-08-27 15:20:32 +12:00
Matthew Gregan
93c6746c0c Bug 1198588 - Remove unused MSG-specific code from AudioStream. r=padenot 2015-08-27 14:49:17 +12:00
Andrea Marchesini
9a2836bb5f Bug 1184292 - Remove some Wakelock warning messages when unlocking. r=smaug 
CLOSED TREE
 2015-08-26 09:21:00 -04:00
Jan-Ivar Bruaroey
86c0e0370b Bug 1064223 - Remove support for constraint-like RTCOfferOptions predecessor. r=mt 2015-08-20 21:49:09 -04:00
Jan Gerber
a1ba162255 Bug 1197777 - Use libavformat-ffmpeg.so.56 if available. r=k17e 2015-08-24 14:08:09 +02:00
Andre Natal
fe3e0ee924 Bug 1198540 - Add translation of digit 0 to word oh on cmudict. r=kdavis 2015-08-25 20:16:00 -04:00
Chris Peterson
cc68dfd296 Bug 1197563 - Polyfill __func__ for MSVC 2013 and earlier. r=froydnj 2015-08-20 23:39:18 -07:00
Anthony Jones
bfc809d025 Bug 1180101 - Test 0 length atom inside moov; r=jya 2015-08-26 18:13:01 +12:00
Jean-Yves Avenard
51689374b6 Bug 1197075: P4. Reject skip promise on cancellation or shutdown. r=edwin 
Also remove redundant code.
 2015-08-26 15:36:54 +10:00
Jean-Yves Avenard
38e604beda Bug 1197075: P3. Decode frames ahead of MDSM requesting them. r=edwin 
This makes the media.*-decode-ahead pref performs more according to its name.
We decode audio and video in advance so a MediaDataPromise can be resolved almost instantly.
Default is 2.
 2015-08-26 15:36:54 +10:00
Jean-Yves Avenard
f8a39b0726 Bug 1197075: P2. Revert "Bug 1171257 - Add force decode ahead to MediaFormatReader r=jya,bholley". r=edwin 2015-08-26 15:36:53 +10:00
Jean-Yves Avenard
66282b3076 Bug 1197075: P1. Revert "Bug 1197051 - Don't try to decode ahead while seeking - r=jya". r=edwin 2015-08-26 15:36:53 +10:00
JW Wang
e90127bc95 Bug 1198565 - prepend the file name to the token so it is easier to know which file times out. r=jya. 2015-08-26 12:48:57 +08:00
JW Wang
1ddbb43629 Bug 1198568 - Fix build error in non-unified build for DecodedAudioDataSink.cpp. r=kinetik. 2015-08-26 10:30:56 +08:00
Jean-Yves Avenard
f800a05d89 Bug 1196696: Always check that track demuxer was successfully created. r=jwwang 2015-08-26 12:50:06 +10:00
Jean-Yves Avenard
30ee771327 Bug 1196558: Don't assert should output task not be dispatched. r=edwin 
We are in a racy situation where the task queue could start flushing after we've tested if it was being flushed. We do not care if the task isn't going to be dispatched as we want the frame to be dropped when that happens anyway.
 2015-08-26 12:50:06 +10:00
Roman Reiss
90ef519755 Bug 875573 - Add video/x-m4v mime type. r=kentuckyfriedtakahe 2015-08-21 20:52:19 +02:00
Karl Tomlinson
03f25624e5 bug 1196109 keep memory allocation for mixed input channel pointer array r=padenot 2015-08-20 18:41:25 +12:00
Karl Tomlinson
af2284d6f1 bug 1197043 use flags to distinguish between external streams and events r=padenot 
MediaStreamAudioDestinationNode does not need any main thread events because
mDOMStream provides GetCurrentTime to consumers.

MediaRecoder also does not use main thread current time.
 2015-08-13 16:13:34 +12:00
Karl Tomlinson
f0bd706a0c bug 1197043 move AudioNodeStream creation to stream class r=padenot 2015-08-12 11:26:24 +12:00
Karl Tomlinson
0744ddb0e1 bug 1197043 introduce MediaStreamGraph::AddStream() r=padenot 2015-08-12 11:46:56 +12:00
Karl Tomlinson
27611775b5 bug 1197043 rename Add/RemoveStream to Add/RemoveStreamGraphThread r=padenot 2015-08-12 11:29:35 +12:00
Karl Tomlinson
393d2224b6 bug 1197043 remove unnecessary aSampleRate parameter for AudioNodeStream creation r=padenot 2015-08-12 11:04:13 +12:00
Karl Tomlinson
de8e0bbe82 bug 1198100 remove mNotificationMainThreadRunnable r=baku 
If AddMainThreadListener() were called multiple times after
mFinishedNotificationSent is set then we'd get some extra NotifyRunnables but
NotifyMainThreadListeners() clears mMainThreadListeners anyway so we still get
only one notification per listener.

mNotificationMainThreadRunnable is an unnecessary optimization, so better not
to add storage to every MediaStream.
 2015-08-24 11:54:24 +12:00
Benjamin Chen
5f85a4d2b4 Bug 762774 - Increase the timeout and reduce the testcase loop times. r=jwwang 2015-08-20 18:05:28 +08:00
JW Wang
eda90e10c3 Bug 1196112 - Part 2: Remove DecodedStream::mMonitor. r=roc 2015-08-24 21:05:32 +08:00
JW Wang
b332b62029 Bug 1196112 - Part 1: Ensure all members except |mShuttingDown| and |mOutputStreamManager| are accessed on the worker thread only. r=roc 2015-08-24 21:05:22 +08:00
Ryan VanderMeulen
dddfd9b6c7 Backed out changesets b2eb913e58c9 and c678e1317fa0 (bug 1196112) for suspicion of causing bug 1197977. 2015-08-24 16:44:14 -04:00
Jean-Yves Avenard
86fdd87d74 Bug 1197664: Report the total number of decoded frames. CLOSED TREE r=kentuckyfriedtakahe 2015-08-24 23:46:35 +10:00
Jean-Yves Avenard
9412fd8a00 Bug 1195073: [MSE] P7. Determine a WebM Media Segment end as early as available information permits. r=kinetik 
Most cluster contains information about their size. When known, we don't need to wait until the next media segment is received to determine its size.
 2015-08-24 23:13:50 +10:00
Jean-Yves Avenard
be2837a7a6 Bug 1195073: [webm] P6. Calculate cluster's end offset if its size is known in advance. r=kinetik. 
This allows to detect the end of a webm media segment without having to wait for the start of a new one.
Also record where an init segment (EBML) starts as this will be required by the WebM ContainerParser.
 2015-08-24 23:13:49 +10:00
Jean-Yves Avenard
98864bb30c Bug 1195073: [MSE] P5. Detect out of order appends and recreate demuxer. r=gerald 
The webm demuxer will only handle data where frames's a monotonically increasing.
 2015-08-24 23:13:49 +10:00
Jean-Yves Avenard
693ecb89cc Bug 1195073: [MSE/webm] P4. Limit nestegg reads to the last block's boundaries. r=kinetik 
This prevent entering into an unrecoverable error state when parsing incomplete data as often seen with MSE.
 2015-08-24 23:13:48 +10:00
Jean-Yves Avenard
d6a6fb71ca Bug 1195073: [webm] P3. Initialize members. r=kinetik 2015-08-24 23:13:48 +10:00
Jean-Yves Avenard
803201660b Bug 1195073: [webm] P2. Add WebMBufferedState::GetLastBlockOffset method. r=kinetik 
MSE may input partial media segment, which could cause the WebMDemuxer and libnestegg to error upon encountering an incomplete block which can't be recovered from.
this will allow to limit read to known complete blocks.
 2015-08-24 23:13:47 +10:00
Jean-Yves Avenard
a6b5df3bb3 Bug 1195073: [MSE/webm] P1. Detect individual webm clusters. r=kinetik 
WebMContainerParser was incorrectly reporting webm blocks rather than clusters, causing the webm demuxer to later fail to parse the remaining data.
 2015-08-24 23:13:47 +10:00
JW Wang
fa2ad1a8b3 Bug 1196112. Part 2 - Remove DecodedStream::mMonitor. r=roc. 2015-08-24 21:05:32 +08:00
JW Wang
3fe6ed90c0 Bug 1196112. Part 1 - ensure all members except |mShuttingDown| and |mOutputStreamManager| are accessed on the worker thread only. r=roc. 2015-08-24 21:05:22 +08:00
JW Wang
6190450338 Bug 1194112. Part 6 - add a test case to test the exclusive mode. r=kinetik. 2015-08-24 10:38:42 +08:00
JW Wang
91209aa8b6 Bug 1194112. Part 5 - remove dead code. r=kinetik. 2015-08-24 10:38:37 +08:00
JW Wang
7486cc31f6 Bug 1194112. Part 4 - reimplement Listener/ListenerImpl to support Move. r=kinetik. 2015-08-24 10:38:33 +08:00
JW Wang
a6b8a5478f Bug 1194112. Part 3 - use perfect forwarding in MediaEventProducer::Notify() so MediaEventSource can decide whether to copy or move according to its ListenerMode. r=kinetik. 2015-08-24 10:38:28 +08:00
JW Wang
2567281e4a Bug 1194112. Part 2 - small code refactoring to reduce typing. r=kinetik. 2015-08-24 10:38:24 +08:00
JW Wang
01fa4f5b2b Bug 1194112. Part 1 - extract event dispatch code from ListenerImpl to its own class. r=kinetik. 2015-08-24 10:38:16 +08:00
JW Wang
139088ba2c Bug 1195632. Part 2 - Have DecodedStream listen to push events of the media queues and call SendData() on its own without the help of MDSM. r=roc. 2015-08-24 10:04:21 +08:00
JW Wang
3c85795637 Bug 1195632. Part 1 - Let DecodedStream have a worker thread and asset some funtions on the worker thread. r=roc. 2015-08-24 09:55:30 +08:00
Alfredo Yang
22a0b98824 Bug 1197151 - Correct TaskQueue asertion. r=jya 2015-08-23 19:36:00 -04:00
Kelly Davis
76c551e0fa Bug 1197455 - Call ps_set_search() only after successful grammar compilation. r=anatal 2015-08-22 10:08:00 -04:00
Timothy B. Terriberry
94ebb79c9f Bug 1196353 - Use standard Xiph extradata format to pass headers from demuxers to decoders. r=jya 2015-08-21 10:17:00 -04:00
Kelly Davis
b30cf3972d Bug 1193183 - Correctly implement SpeechRecognitionAlternative::confidence using ps_get_prob(). r=anatal 2015-08-11 07:03:00 -04:00
Ehsan Akhgari
72b1b417a3 Bug 1197185 - Give NonE10s a virtual destructor; r=jesup 
The Parent template class inherits from both NonE10s and PMediaParent,
and the latter has a virtual destructor.  This confuses clang-tidy when
it wants to rewrite ~Parent() which is currently marked as virtual, and
it makes ~Parent() override which breaks the build since ~NonE10s() is
not virtual.  The easiest way to work around this seems to be making
~NonE10s() virtual.  See the discussion in
https://llvm.org/bugs/show_bug.cgi?id=24496 as well.
 2015-08-21 17:53:35 -04:00
Alastor Wu
e0bc553231 Bug 1187092 - Refactor the suspend process. r=jwwang 2015-08-19 12:43:25 +08:00
Ryan VanderMeulen
804493e8e2 Bug 1187903 - Skip various failure-prone EME tests on Windows debug. 2015-08-21 09:41:52 -04:00
Andrea Marchesini
c11c96942e Bug 1191814 - WebSpeech synthesis API and AudioChannelService, r=eeejay 2015-08-21 11:58:06 +01:00
Sebastian Hengst
9e9295909c Backed out changeset 6beb23f39237 (bug 1195073) because other parts of the bug have been backed out. r=backout 2015-08-21 11:31:59 +02:00
Sebastian Hengst
727f812368 Backed out 5 changesets 507a508aea7, 16669eed518d, 6f6fc1a91d07, 1e983ccb61cc, c5afa29ea85a (bug 1195073) for M2 and W5 on OSX and Linux and R(R2, Ru2) bustage on Linux. r=backout 
Backed out changeset 507a508aea70 (bug 1195073)
Backed out changeset 16669eed518d (bug 1195073)
Backed out changeset 6f6fc1a91d07 (bug 1195073)
Backed out changeset 1e983ccb61cc (bug 1195073)
Backed out changeset c5afa29ea85a (bug 1195073)
 2015-08-21 10:46:05 +02:00
Edwin Flores
c91e20ebc9 Bug 1197051 - Don't try to decode ahead while seeking - r=jya 2015-08-21 17:25:46 +10:00
Jean-Yves Avenard
5ead6d0dc2 Bug 1195073: [MSE] P6. Fix B2G build. a=bustage 2015-08-21 16:40:18 +10:00
Jean-Yves Avenard
5170ad0348 Bug 1195073: [MSE] P5. Detect out of order appends and recreate demuxer. r=gerald 
The webm demuxer will only handle data where frames's a monotonically increasing.
 2015-08-21 16:15:36 +10:00
Jean-Yves Avenard
c7afdd2af9 Bug 1195073: [MSE/webm] P4. Limit nestegg reads to the last block's boundaries. r=kinetik 
This prevent entering into an unrecoverable error state when parsing incomplete data as often seen with MSE.
 2015-08-21 16:15:35 +10:00
Jean-Yves Avenard
9e256f207f Bug 1195073: [webm] P3. Initialize members. r=kinetik 2015-08-21 16:15:35 +10:00
Jean-Yves Avenard
2073826184 Bug 1195073: [webm] P2. Add WebMBufferedState::GetLastBlockOffset method. r=kinetik 
MSE may input partial media segment, which could cause the WebMDemuxer and libnestegg to error upon encountering an incomplete block which can't be recovered from.
this will allow to limit read to known complete blocks.
 2015-08-21 16:15:34 +10:00
Jean-Yves Avenard
c92cbef598 Bug 1195073: [MSE/webm] P1. Detect individual webm clusters. r=kinetik 
WebMContainerParser was incorrectly reporting webm blocks rather than clusters, causing the webm demuxer to later fail to parse the remaining data.
 2015-08-21 16:15:34 +10:00
Jean-Yves Avenard
ddfed3bab2 Bug 1196398: [mp4] Do not allocate memory spanning across ftyp and moov atom. r=kentuckyfriedtakahe 
A typical non-fragmented mp4 would have the ftyp atom located at the beginning of the mp4 and the moov at the end. We would to calculate the location of the metadata by spanning the byte range of the two atoms.
As such, we would typically allocate an amount of memory equivalent to the size of the mp4.

Instead we now reconstruct the metadata to only have the ftyp and moov atoms contiguously.
 2015-08-21 15:35:31 +10:00
Jean-Yves Avenard
c8fcf47e9e Revert "Bug 1196398: [mp4] Do not allocate memory spanning across ftyp and moov atom. r=kentuckyfriedtakahe" 
This reverts commit 73156610be5f
 2015-08-21 14:48:23 +10:00
Jean-Yves Avenard
51f5dbcd81 Bug 1196398: [mp4] Do not allocate memory spanning across ftyp and moov atom. r=kentuckyfriedtakahe 
A typical non-fragmented mp4 would have the ftyp atom located at the beginning of the mp4 and the moov at the end. We would to calculate the location of the metadata by spanning the byte range of the two atoms.
As such, we would typically allocate an amount of memory equivalent to the size of the mp4.

Instead we now reconstruct the metadata to only have the ftyp and moov atoms contiguously.
 2015-08-21 13:10:34 +10:00
Chris Pearce
cb8809ed04 Bug 1197022 - [EME] Disable dormant mode for EME videos on all channels. r=kentuckyfriedtakahe 2015-08-21 13:34:21 +12:00
Karl Tomlinson
79c8a43d05 bug 1196632 pass ownership of runnables to NS_DispatchToMainThread r=jesup 
The already_AddRefed<nsIRunnable>&& overload saves some ref-counting and
provides an assertion if called too late in shutdown.
 2015-08-20 09:52:56 +12:00
Karl Tomlinson
ed0a197b68 bug 1196111 don't keep AudioContext alive from AudioBuffer r=baku 
AudioBuffers are often generated from OfflineAudioContext to be used in other
contexts, when the OfflineAudioContext is no longer required.
 2015-08-18 11:48:38 +12:00
Wes Kocher
f29ff3a92d Backed out changeset f51c32dbea00 (bug 875573) for test_can_play_type_mpeg.html orange 2015-08-20 16:19:39 -07:00
Roman Reiss
eb781d5629 Bug 875573 - Add video/x-m4v mime type to MP4Decoder r=kentuckyfriedtakahe 2015-08-20 14:14:11 -07:00
Wes Kocher
2cac68a0b2 Backed out changeset 91cc337bef3b (bug 1191814) for m-e10s(3) failures in test_global_queue_pause.html 2015-08-20 12:05:37 -07:00
Andrea Marchesini
254bfbb384 Bug 1184807 - better management of the dispatching of events when shutting down, r=rjesup 2015-08-20 19:54:50 +01:00
Andrea Marchesini
e883acb20b Bug 1191814 - WebSpeech synthesis API and AudioChannelService, r=eeejay 2015-08-20 12:48:57 +01:00
Andrea Marchesini
57e1b137f0 Bug 1184807 - TextTrackList should check if we are shutting down before dispatching events, r=mccr8 2015-08-20 17:45:34 +01:00
Matt Woodrow
8125c31adc Bug 1196417 - Make video software fallback only affect the current video instead of the entire browser. r=cpearce 2015-08-20 11:43:36 -04:00
Matt Woodrow
be988f21f8 Bug 1196408 - Make sure we only report a corrupt/slow video frame once. r=cpearce 2015-08-20 11:39:49 -04:00
Matt Woodrow
ec6dd28401 Bug 1196411 - Disable DXVA on 60fps 1080p videos for AMD cards that can't decode quick enough. r=jya 2015-08-20 11:37:26 -04:00
Matt Woodrow
5a374180aa Bug 1196409 - Disable D3D11-DXVA for resolutions not supported in hardware. r=jya 2015-08-20 11:36:04 -04:00
Jean-Yves Avenard
68a1fb84c6 Bug 1196591: Initialize member. r=me 
Small correction introduced in bug 1190238
 2015-08-20 21:04:28 +10:00
Andrea Marchesini
8271863acf Bug 1196514 - remove dom.messagechannel.enabled pref, r=smaug 2015-08-20 08:34:38 +01:00
JW Wang
eb572bfae7 Bug 1195185. Part 4 - remove dead code. r=roc. 2015-08-20 13:02:54 +08:00
JW Wang
8a296370d8 Bug 1195185. Part 3 - align the life cycle of mData with {Start,Stop}Playback. r=roc. 2015-08-20 13:02:43 +08:00
JW Wang
4b83ebaf7d Bug 1195185. Part 2 - Check if all output streams are from the same graph because we don't support connecting streams from different graphs. r=roc. 2015-08-20 13:02:31 +08:00
JW Wang
d02abb9d1a Bug 1195185. Part 1 - rename Connect/Remove to AddOutput/RemoveOutput. r=roc. 2015-08-20 13:02:20 +08:00
Ehsan Akhgari
559da3d613 Bug 1196361 - Remove the media.useAudioChannelService pref; r=baku 
We need to rely on the audio channel service for all products now,
so supporting running without it seems pointless at this point.
 2015-08-19 21:00:18 -04:00
Benjamin Chen
1045493205 Bug 1071375: Re-enable some MediaRecorder tests. r=jwwang 2015-08-17 17:31:14 +08:00
Kilik Kuo
029f56edd6 Bug 1188268 - Make AudioSink a base class, create DecodedAudioDataSink to act as original AudioSink, and move sink-related files to dom/media/mediasink. r=jwwang, r=cpearce 2015-08-18 11:55:01 +08:00
Jean-Yves Avenard
1ea110cc89 Bug 1180214: P3. Do not override known duration with approximated one. r=edwin 
Some gstreamer plugin return nonsensical values
 2015-08-19 15:51:01 +10:00
William Chen
b802d90437 Bug 1131470 - Part 1: Rename existing use of ScreenOrientation to ScreenOrientationInternal. r=baku 2015-08-18 14:55:09 -07:00
Matt Woodrow
ad26ccc5d5 Bug 1195527 - Part 2: Add D3D11 video TextureClient recycler. r=jrmuizel 2015-08-18 21:04:46 -04:00
Matt Woodrow
0337a4f687 Bug 1195527 - Part 1: Move D3D11 texture allocation into the TextureClient. r=jrmuizel 2015-08-18 21:04:14 -04:00
Bobby Holley
9a4849b006 Bug 1195867 - Hoist StateWatching and StateMirroring into XPCOM. r=froydnj 2015-08-18 15:37:06 -07:00
Eugen Sawin
b4cfbe01e9 Bug 1194085 - Abort frame parsing if no valid audio stream is found. r=jya 2015-08-18 20:49:58 +02:00
Matt Woodrow
3678c767ad Bug 1194753 - Wait longer when decoding the first frame of D3D9 DXVA video before deciding it is invalid. r=cpearce 2015-08-18 14:19:45 -04:00
Ryan VanderMeulen
ef346df0b3 Backed out changeset ee800c0e234f (bug 1187092) for media crashes/asserts. 2015-08-18 13:11:05 -04:00
Alastor Wu
b5ce2f0d01 Bug 1187092 - Refactor the suspend process. r=jwwang 2015-08-17 17:25:07 +08:00
Alfredo Yang
d353da94e5 Bug 1195625 - Use correct TaskQueue in SharedDecoderManager and H264Converter promise. r=jya 2015-08-17 23:47:00 -04:00
Alfredo Yang
cc4a6ca8df Bug 1179667 - Use MozPromise to initialize Gonk PlatformDecodeModule. r=jya 2015-08-17 23:05:00 -04:00
Sebastian Hengst
87833bf6d3 Backed out changesets d0eee19c83cd, 0c54ee53678f, 5b202008a431, 81291b4e6dc3, acbc0d7e21cd, a7ceb6162a63 (bug 1194112) for Linux x64 Hazard failure. r=backout 
Backed out changeset d0eee19c83cd (bug 1194112)
Backed out changeset 0c54ee53678f (bug 1194112)
Backed out changeset 5b202008a431 (bug 1194112)
Backed out changeset 81291b4e6dc3 (bug 1194112)
Backed out changeset acbc0d7e21cd (bug 1194112)
Backed out changeset a7ceb6162a63 (bug 1194112)
 2015-08-18 11:11:38 +02:00
Sebastian Hengst
6b20336f46 Backed out changesets 7610baf4a3ae, 6226b99f19bd, 157e41e32906, 60a3b1862f71 (bug 1195158). r=backout 
Backed out changeset 7610baf4a3ae (bug 1195158)
Backed out changeset 6226b99f19bd (bug 1195158)
Backed out changeset 157e41e32906 (bug 1195158)
Backed out changeset 60a3b1862f71 (bug 1195158)
 2015-08-18 10:52:09 +02:00
JW Wang
026c90ac3d Bug 1195158. Part 4 - remove unused code. r=cpearce. 2015-08-18 11:12:39 +08:00
JW Wang
6522811795 Bug 1195158. Part 3 - connect listeners. r=cpearce. 
a. MediaMetadataManager is connected to MediaDecoderReader::mTimedMetadataEvent to receive TimedMetadata events.
b. OggReader publish TimedMetadata events through MediaDecoderReader::mTimedMetadataEvent.
c. MDSM calls MediaMetadataManager::DispatchMetadataIfNeeded to publish metadata if playback positoin reaches the publish time.
d. MediaDecoder is connected to MediaMetadataManager::mTimedMetadataEvent to receive TimedMetadata events.
e. MediaDecoder updates its metadata when TimedMetadata events are received.
 2015-08-18 11:12:39 +08:00
JW Wang
623b3a7a28 Bug 1195158. Part 2 - Have OggReader send TimedMetadata events through a event source instead of direct calls to AbstractMediaDecoder::QueueMetadata. r=cpearce. 2015-08-18 11:12:39 +08:00
JW Wang
7de6530d5a Bug 1195158. Part 1 - Have MediaMetadataManager listen to an event source to receive TimedMetadata events. OggReader will send TimedMetadata events through an event source. This will break OggReader's dependency on AbstractMediaDecoder::QueueMetadata which then can be removed for it is against our goal to run all MediaDecoder's methods on the main thread. r=cpearce. 2015-08-18 11:12:38 +08:00
JW Wang
7c975ccea1 Bug 1194112. Part 6 - add a test case to test the exclusive mode. r=kinetik. 2015-08-18 11:12:18 +08:00
JW Wang
a5033dff75 Bug 1194112. Part 5 - remove dead code. r=kinetik. 2015-08-18 11:12:06 +08:00
JW Wang
1b8f5246eb Bug 1194112. Part 4 - reimplement Listener/ListenerImpl to support Move. r=kinetik. 2015-08-18 11:11:53 +08:00
JW Wang
cf0ef654b9 Bug 1194112. Part 3 - use perfect forwarding in MediaEventProducer::Notify() so MediaEventSource can decide whether to copy or move according to its ListenerMode. r=kinetik. 2015-08-18 11:11:42 +08:00
JW Wang
d1553650cb Bug 1194112. Part 2 - small code refactoring to reduce typing. r=kinetik. 2015-08-18 11:11:29 +08:00
JW Wang
fb130c6016 Bug 1194112. Part 1 - extract event dispatch code from ListenerImpl to its own class. r=kinetik. 2015-08-18 11:11:16 +08:00
JW Wang
1a51eb5ad9 Bug 1195187. Part 2 - add a new class OutputStreamManager for managing output streams. r=roc. 2015-08-18 10:58:13 +08:00
JW Wang
68bebbd414 Bug 1195187. Part 1 - Move output stream connection/disconnection code to OutputStreamData. r=roc. 2015-08-18 10:57:56 +08:00
Jean-Yves Avenard
cfa2a7e843 Bug 1180214: P2. Returns empty interval until we have a start time. r=cpearce 2015-08-18 09:21:56 +10:00
Jean-Yves Avenard
c1fbfa26fe Bug 1180214: P1. Don't attempt to estimate duration until read metadata completed. r=cpearce 
This revert to the behaviour prior bug 1175768
 2015-08-18 09:21:56 +10:00
John Lin
8c237a861d Bug 1186845 - Let MP4Decoder handle 3GPP files if Gonk PDM is available. r=cpearce 2015-08-17 15:12:43 -07:00
Benjamin Chen
75800e9b3e Bug 1192733: add new testcase for dormant. r=jwwang 2015-08-17 10:22:48 +08:00
Benjamin Chen
47d214758e Bug 1192733: fix the MediaFormatReader can not back from dormant state. r=jya 2015-08-14 12:02:15 +08:00
Bobby Holley
6c899edb65 Bug 1188976 - Hoist MozPromise into xpcom. r=froydnj 2015-08-17 14:54:45 -07:00
Bobby Holley
f703fcf697 Bug 1188976 - Use mozilla::Tuple for InvokeAsync. r=froydnj 2015-08-17 14:54:44 -07:00
Bobby Holley
bf29d53b6d Bug 1188976 - Rename ProxyMediaCall to InvokeAsync. r=froydnj 2015-08-17 14:54:42 -07:00
Bobby Holley
39adf73c68 Bug 1188976 - Fix up include. r=me 
Not sure how this didn't get fixed before.
 2015-08-17 14:54:41 -07:00
Ehsan Akhgari
0588ca1483 Bug 1195051 - Part 4: Fix a null pointer crash happening after the destination node gets CCed 
Landed on a CLOSED TREE
 2015-08-17 11:45:25 -04:00
Ehsan Akhgari
edbca9d897 Bug 1195051 - Part 2: Mute the destination node when the AudioContext is suspended, and unmute when resumed; r=padenot 2015-08-17 08:55:00 -04:00