Commit Graph

4366 Commits

Author SHA1 Message Date
Chris Pearce
9d0254644c Bug 1207000 - Create GMPVideoDecoderTrialCreator on all platforms. r=edwin 2015-09-22 15:12:09 +12:00
JW Wang
acddedcb5b Bug 1206578 - Group public and private functions respectively for MDSM. r=gsquelart. 2015-09-22 11:03:08 +08:00
JW Wang
31e9b2cc17 Bug 1206574 - Remove AbstractMediaDecoder::IsShutdown(). r=cpearce. 2015-09-22 10:56:17 +08:00
Karl Tomlinson
da268d2dbf bug 1205540 account for active inputs and skip processing when streams are inactive r=padenot 2015-09-18 00:03:00 +12:00
Karl Tomlinson
772fb175f8 bug 1205540 initialize mBufferFormat when constructing silent block r=padenot
This makes an AudioBlock valid for code testing mBufferFormat without IsNull(),
without the need for explicit SetNull().

This is useful so that setting AudioNodeStream::mLastChunks each iteration is
not required for inactive nodes.
2015-09-21 11:37:55 +12:00
Karl Tomlinson
26598e610a bug 1205540 mark BufferSource finished only when producing silent output block r=padenot
This allows simpler processing of the finished state to mark the node as an
inactive input of any downstream nodes.  Otherwise the input could not be
considered inactive until after downstream nodes have finished processing,
but ProcessInput() may not be called again on finished streams.

AudioBufferSourceNode now behaves the same as OscillatorNode and similarly
to nodes that release a playing ref.
2015-09-18 00:00:17 +12:00
Karl Tomlinson
8d2f0b8281 bug 1205540 make source stream available during RemoveInput r=padenot 2015-09-09 14:19:05 +12:00
Karl Tomlinson
24f3f6d27c bug 1205540 provide querying whether engines need to continue processing even without input r=padenot 2015-09-09 08:54:03 +12:00
Karl Tomlinson
b5c756132c bug 1205540 don't send more null chunks than necessary to AnalyserNode r=padenot 2015-09-09 08:52:39 +12:00
Jean-Yves Avenard
3a47518da0 Bug 1205470: [MSE] Remove assertion. r=cpearce
Can also remove mInitDone member as that information is redundant with the init promise being present
2015-09-22 08:21:04 +10:00
Christoph Kerschbaumer
1feac968da Bug 1048048 - add preload content policy types for scripts (r=baku) 2015-09-20 14:56:21 -07:00
Wes Kocher
a88e0ba0f7 Backed out changeset 49a711e67d8b (bug 1190379) for various android crashes CLOSED TREE 2015-09-21 09:09:29 -07:00
Wes Kocher
9c40f87100 Backed out 7 changesets (bug 1048048) for android crashes in various chunks CLOSED TREE
Backed out changeset b5abe23a4ea5 (bug 1048048)
Backed out changeset 4f91b10e8be0 (bug 1048048)
Backed out changeset 450d4a13c90e (bug 1048048)
Backed out changeset 6a727c40eb68 (bug 1048048)
Backed out changeset 88c2333ff745 (bug 1048048)
Backed out changeset 740ab1ecd079 (bug 1048048)
Backed out changeset 02c6d6aef163 (bug 1048048)
2015-09-21 09:08:34 -07:00
Christoph Kerschbaumer
a69ba45c22 Bug 1048048 - add preload content policy types for scripts (r=baku) 2015-09-20 14:56:21 -07:00
Eugen Sawin
7928115ce5 Bug 1190379: Use AndroidDecoderModule for VP8/9. r=jya 2015-09-21 16:39:12 +02:00
Paul Adenot
8d3cb40a78 Bug 1203836 - Test that MediaElementAudioSourceNode pipes correct data to an AudioContext. r=karlt 2015-09-21 11:00:12 +02:00
Paul Adenot
38ca16f8aa Bug 1203836 - Properly handle silent chunks in AudioNodeExternalInputStream. r=karlt 2015-09-21 10:14:59 +02:00
JW Wang
3b3895ed85 Bug 1206576 - Dispatch some MDSM functions to hide its internal thread model. r=jya. 2015-09-21 14:01:45 +08:00
JW Wang
a0cb044960 Bug 1204430. Part 2 - mirror MediaDecoder::mMediaSeekable. r=kinetik. 2015-09-21 13:49:01 +08:00
JW Wang
7d3c2064f3 Bug 1204430. Part 1 - dispatch AbstractMediaDecoder::SetMediaSeekable() to the main thread. r=kinetik. 2015-09-21 13:48:39 +08:00
Andreas Pehrson
4916737c63 Bug 1177276 - Pref on canvas.captureStream by default. r=smaug,mt 2015-09-17 22:09:24 +08:00
Jan-Ivar Bruaroey
6c9de039ed Bug 912342 - get Promise out. r=jesup 2015-09-20 10:04:51 -04:00
Jan-Ivar Bruaroey
32e9d893c2 Bug 912342 - Move code SelectSettings to MediaConstraintsHelper. r=jesup 2015-09-19 02:00:48 -04:00
Jan-Ivar Bruaroey
c8450f5db9 Bug 912342 - Pass in Audio/VideoDevice in place of Audio/VideoSource. r=jesup 2015-09-19 00:49:07 -04:00
Jan-Ivar Bruaroey
b2c0974877 Bug 912342 - Change capture resolution. r=jesup 2015-09-20 18:45:57 -04:00
Jan-Ivar Bruaroey
7575061788 Bug 912342 - Move code MediaOperationTask from .h to .cpp. r=jesup 2015-09-16 22:44:14 -04:00
Jan-Ivar Bruaroey
a0aeb846e8 Bug 1195951 - fix heap type on stack error in MediaUtils' Pledge class. r=jesup 2015-09-20 02:26:41 -04:00
Jan-Ivar Bruaroey
6dccc3ae57 Bug 1181896 - make gUM fail w/OverconstrainedError and candidate argument r=jesup 2015-09-18 14:04:41 -04:00
Jan-Ivar Bruaroey
1fedf527e6 Bug 1181896 - rename .constraintName to .constraint in MediaStreamError r=jesup,bz 2015-09-18 14:03:30 -04:00
Maire Reavy
c42278687e bug 1161276 - part 3 - tests for bitrate property in MediaRecorder r=roc
Android 2.3 and B2G disabled
2015-09-18 14:30:02 -04:00
Maire Reavy
a7d86cae5e Bug 1161276: part 2 - pass bitrates to track encoders for MediaRecorder r=roc 2015-06-08 11:47:28 +08:00
Steven Lee
0b5c28b0ff bug 1161276 - part 1 add bits-per-second interface to MediaRecorder r=roc,smaug 2015-05-25 23:49:03 +08:00
Nathan Froyd
63df15ccef Bug 1205825 - part 2 - call Get*CachedData*Internal variants from MediaCacheStream::GetCachedRanges; r=roc
Each of GetNextCachedData and GetCachedDataEndInternal grab the monitor
on entrance, but GetCachedRanges has already grabbed the monitor for us.
We can call the *Internal variants instead, which assert that the
monitor is held.
2015-09-18 00:15:56 -04:00
Nathan Froyd
ee09350fca Bug 1205825 - part 1 - don't reacquire the media cache's monitor in MediaCacheStream::FlushPartialBlockInternal; r=roc
FlushPartialBlockInternal is only ever called by methods that are
already holding the MediaCache's monitor.  So we don't need to reacquire
it every time we call this method.

fixup p1
2015-09-17 23:57:51 -04:00
Carsten "Tomcat" Book
d9728035a5 Backed out 5 changesets (bug 1205540) for Assertion Failurs in m2-e10s tests on a CLOSED TREE
Backed out changeset e89d8182d588 (bug 1205540)
Backed out changeset abace4cdec06 (bug 1205540)
Backed out changeset b3f6e1db7233 (bug 1205540)
Backed out changeset 0d122cb34921 (bug 1205540)
Backed out changeset 4de5f87180fb (bug 1205540)
2015-09-18 16:36:33 +02:00
Gian-Carlo Pascutto
ded8b8934e Bug 1202424 - Free shared memory when IPC system shuts down, not after. r=mrbkap 2015-09-11 09:33:36 +02:00
Karl Tomlinson
dfa37c5b8b bug 1205540 account for active inputs and skip processing when streams are inactive r=padenot 2015-09-18 00:03:00 +12:00
Karl Tomlinson
fa1b519c60 bug 1205540 mark BufferSource finished only when producing silent output block r=padenot
This allows simpler processing of the finished state to mark the node as an
inactive input of any downstream nodes.  Otherwise the input could not be
considered inactive until after downstream nodes have finished processing,
but ProcessInput() may not be called again on finished streams.

AudioBufferSourceNode now behaves the same as OscillatorNode and similarly
to nodes that release a playing ref.
2015-09-18 00:00:17 +12:00
Karl Tomlinson
9d45dcc429 bug 1205540 make source stream available during RemoveInput r=padenot 2015-09-09 14:19:05 +12:00
Karl Tomlinson
e16647df50 bug 1205540 provide querying whether engines need to continue processing even without input r=padenot 2015-09-09 08:54:03 +12:00
Karl Tomlinson
9fc74a7d4a bug 1205540 don't send more null chunks than necessary to AnalyserNode r=padenot 2015-09-09 08:52:39 +12:00
Andreas Pehrson
085f228bac Bug 1152298 - Implement AutoDriver for CanvasCaptureMediaStream. r=mt 2015-09-17 11:46:05 +08:00
Andreas Pehrson
57b146b3a4 Bug 1161913 - Part 3 - Relax requestFrame ordering guarantee in tests. r=mt
We used to fully guarantee the order of requestFrame() and draw calls.
For instance:
```
ctx.draw(red);
stream.requestFrame();
ctx.draw(green);
```
would guarantee that a red frame ended up in the stream, and not the
green unless another frame was requested.

Now with frames being requested and pushed out on next refresh, we can
only guarantee that everything up to the requestFrame() call is included
in the next frame. Everything after the requestFrame() and before the
next refresh (stable state in most cases) will now also be inevitably
included.
2015-09-17 11:37:05 +08:00
Andreas Pehrson
89c2075b5a Bug 1161913 - Part 2. Request canvas to push out its next drawn frame instead of pulling it. r=mt 2015-09-17 12:36:57 +08:00
Chris Pearce
81b148c6a7 Bug 1205582 - Add pref to disable GMP trial create. r=edwin 2015-09-18 18:35:14 +12:00
Gian-Carlo Pascutto
2790328ed3 Bug 1200477 - Allow building with WebRTC disabled. r=glandium,mrbkap 2015-09-18 08:29:14 +02:00
Jean-Yves Avenard
07c2523a11 Bug 1205911: P2. Ensure demuxer is reset before performing a seek. r=edwin
Should there be any demuxing request pending resolution, it would have caused our internal seek to ultimately fail.
2015-09-18 16:03:58 +10:00
Jean-Yves Avenard
bc0f744683 Bug 1205911: P1. Cancel pending demux request when searching for next keyframe. r=edwin 2015-09-18 16:03:58 +10:00
Wes Kocher
ce3d477bc7 Backed out 3 changesets (bug 1161276) for test_mediarecorder_bitrate.html failures CLOSED TREE
Backed out changeset 9187f4d2518f (bug 1161276)
Backed out changeset 4e8f28b17300 (bug 1161276)
Backed out changeset cf05881e89ed (bug 1161276)
2015-09-17 13:06:41 -07:00
Richard Barnes
7f2bacd58c Bug 1205156 - Add telemetry to measure how often getUserMedia is used over non-secure origins r=jib 2015-09-17 08:44:50 -04:00