Bug 1799222 - Prevent accidental dispatches to threadpool and timer threads, r=xpcom-reviewers,necko-reviewers,geckoview-reviewers,media-playback-reviewers,jesup,m_kato,padenot,kmag Differential Revision: https://phabricator.services.mozilla.com/D161349

Bug 1803834 - part3 : enable wpt 'autoplay-hidden.optional.html'. r=media-playback-reviewers,padenot Depends on D163742 Differential Revision: https://phabricator.services.mozilla.com/D163743

Bug 1803834 - part2 : pause/resume non-audible autoplay video when its visibility changes. r=media-playback-reviewers,padenot https://html.spec.whatwg.org/multipage/media.html#ready-states:eligible-for-autoplay-2 Depends on D163741 Differential Revision: https://phabricator.services.mozilla.com/D163742

Bug 1803834 - part1 : rename `mAutoplaying` to 'mCanAutoplayFlag'. r=media-playback-reviewers,padenot Our current attribute `mAutoplaying` seems be equal to the `can autoplay flag` defined in the spec, so rename it to the spec name. https://html.spec.whatwg.org/multipage/media.html#can-autoplay-flag Differential Revision: https://phabricator.services.mozilla.com/D163741

Bug 1702483 - Check for timeCode under-/overflow only when it is time to cast. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D163678

Bug 1702483 - Add gtest. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D163700

Bug 1796069 - Register AppleVTDecoder callback thread with the profiler. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D160774

Bug 1796069 - Ignore registrations in CallbackThreadRegistry if refcnt logging is enabled. r=padenot Callback threads that need to be registered with CallbackThreadRegistry are by definition externally managed; otherwise they wouldn't need to be registered in the first place. We may not have any control of the lifetime of the externally managed thread, which, when registered with the profiler has an nsThread wrapper created around it. If the thread outlives all xpcom shutdown stages, the nsThread wrapper instance will leak (it implements nsIThread and nsISupports!). If refcnt logging is enabled when it leaks, this particular instance will be flagged as a leak, and could mark a run on try as failed. To work around this potential leak, this patch makes CallbackThreadRegistry registration and unregistration noops when refcnt logging is enabled. I will also note, that the lifetime issues found with externally managed threads were most notable on Mac with Grand Central Dispatch's Dispatch Queues. Over time a large number of real threads could be seen, and the risk of one of them outliving xpcom shutdown was close to 100% for a common use case involving camera capture (though timing dependent). Differential Revision: https://phabricator.services.mozilla.com/D163207

Bug 1796069 - Move CallbackThreadRegistry definitions to cpp file. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D161325

Bug 1796069 - Record camera/screen capture with tracking id. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D160772

Bug 1796069 - Remove unused BasicTrackSource. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D160755

Bug 1796069 - Annotate MediaStreamTrackSources with a tracking id and apply it to CopyVideoStages. r=padenot This will allow time consuming tasks (copies, capture, etc.) along the pipe to be grouped per-track-source when visualizing profiler markers. Differential Revision: https://phabricator.services.mozilla.com/D160754

Bug 1796069 - Mark video copies with PerformanceRecorder<CopyVideoStage> throughout the gecko realtime video pipe. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D160753

Bug 1803022 - VideoFrame's codedRect should be null after it's closed r=padenot According to the spec [1], VideoFrame::GetCodedRect() should return nullptr when [[Detached]] is true. [1] https://w3c.github.io/webcodecs/#dom-videoframe-codedrect Differential Revision: https://phabricator.services.mozilla.com/D163288

Bug 1799205 - enable video seamless looping. r=media-playback-reviewers,padenot Differential Revision: https://phabricator.services.mozilla.com/D163251

Bug 1778510: Cut over AudioContext to new fine-grained Timer Precision calls r=padenot Depends on D151304 Differential Revision: https://phabricator.services.mozilla.com/D151305

Bug 1802943 - Include missing header for memset. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D163281

Bug 1802649 - Fix C++20 -Wdeprecated-volatile warning by replacing postfix increment with addition. r=media-playback-reviewers,padenot C++20 deprecated decrement/increment of object of volatile-qualified types, e.g. v++. https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2019/p1152r3.html warning: compound assignment with 'volatile'-qualified left operand is deprecated [-Wvolatile] In this case, mInt doesn't need to be atomic. The test only cares that the Foo::Notify5() member function is volatile. Depends on D163164 Differential Revision: https://phabricator.services.mozilla.com/D163167

Bug 1802649 - Fix C++20 -Wdeprecated-volatile warning by replacing volatile variable with an atomic. r=media-playback-reviewers,padenot C++20 deprecated decrement/increment of object of volatile-qualified types, e.g. v++. https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2019/p1152r3.html warning: dom/media/systemservices/CamerasParent.cpp:1073:20: '++' expression of 'volatile'-qualified type is deprecated [-Wvolatile] If counter variable was incremented on the main thread, we could just remove the volatile. But it's incremented off the main thread, so we need to protect against races from multiple CamerasParent objects. Differential Revision: https://phabricator.services.mozilla.com/D163164

Bug 1802225 - Remove Layers.{h,cpp}. r=tnikkel,geckoview-reviewers,jgilbert,media-playback-reviewers,padenot,m_kato Where appropriate, `#include "Layers.h"` is replaced with more specific inclusions. Differential Revision: https://phabricator.services.mozilla.com/D162934

Bug 1778510: Cut over AudioContext to new fine-grained Timer Precision calls r=padenot Depends on D151304 Differential Revision: https://phabricator.services.mozilla.com/D151305

Bug 1802798 - Fix WebAudio crash on x86_32 CPUs without DAZ mode. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D163180

Bug 1778510: Cut over AudioContext to new fine-grained Timer Precision calls r=padenot Depends on D151304 Differential Revision: https://phabricator.services.mozilla.com/D151305

Bug 1801692 - always check other conditions before requesting data in order to ensure that we don't request data during another request or seeking. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D162790

Bug 1801818 - Make VideoFrame.timestamp non-null r=padenot,emilio Keep up with https://github.com/w3c/webcodecs/pull/437 Differential Revision: https://phabricator.services.mozilla.com/D162808

Bug 1799826 - ensure no further sample request when transitioning from looping decoding state to the completed state. r=padenot I assume one possible situation of causing crash is that the looping state is transitioning to the completed state due to both tracks can't get new data after reaching EOS [1]. In this case, we should also finish queue and cancal any ongoing sample request in order not to trigger EOS again after entering the completed state. [1] https://searchfox.org/mozilla-central/rev/650c19c96529eb28d081062c1ca274bc50ef3635/dom/media/MediaDecoderStateMachine.cpp#3048 Differential Revision: https://phabricator.services.mozilla.com/D162676

Bug 1797557 - Ensure storing VideoFrames in IndexedDB throws r=padenot,smaug Serializing a VideoFrame when `forStorage=true` should throw a `DataCloneError` (See [1] for more detail). [1] https://w3c.github.io/webcodecs/#videoframe-transfer-serialization Differential Revision: https://phabricator.services.mozilla.com/D162679

Bug 1798959 - Ensure all VideoColoSpaceInits work r=padenot Add a wpt ensuring the VideoColorSpace can be constructed from all kinds of VideoColorSpaceInit values. The new wpt is modified from an existing mochitest and it will replace the old mochitest. Differential Revision: https://phabricator.services.mozilla.com/D162525

Bug 1798959 - Set VideoColorPrimaries to PlanarYCbCrData r=jgilbert,bradwerth,padenot When constructing a VideoFrame from the buffer data, the given VideoColorPrimaries passed through VideoFrame's Constructor should be set to PlanarYCbCrData's mColorPrimaries. Depends on D161225 Differential Revision: https://phabricator.services.mozilla.com/D161226

Bug 1798975 - Map new VideoColorSpace values to gfx's ones r=padenot Convert the newly added VideoColorSpace values to gfx's values Depends on D161224 Differential Revision: https://phabricator.services.mozilla.com/D161225

Bug 1798975 - Add new values to VideoColorSpace r=webidl,jgilbert,padenot,smaug Add more color spaces introduced by https://github.com/w3c/webcodecs/pull/566 Differential Revision: https://phabricator.services.mozilla.com/D161224

Bug 1800547 - add some profiler markers for media engine playback. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D162055

Bug 1800233 - part2 : add test cases to ensure the position won't be changed to zero incorrectly after seeking. r=padenot Depends on D162025 Differential Revision: https://phabricator.services.mozilla.com/D162026

Bug 1800233 - part1 : only seek back to the head when all tracks reach to EOS already. r=padenot This patch fixes the incorrect condition for the single track playback, because the EOS flag for the nonexistent track would always be true. Eg. for audio-only playback, `mIsReachingVideoEOS` is always true. Differential Revision: https://phabricator.services.mozilla.com/D162025

Bug 1800302 - Fix a null-pointer dereference r=padenot Differential Revision: https://phabricator.services.mozilla.com/D161972

Bug 1799995 - correct wording for LoopingDecodingState. r=media-playback-reviewers,padenot Differential Revision: https://phabricator.services.mozilla.com/D161742

Bug 1799826 - cancel any data request when leaving looping decoding state if playback has reached to the end. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D161745

Bug 1799816 - Remove symbols that are optionally part of libmozavutil. r=media-playback-reviewers,padenot Differential Revision: https://phabricator.services.mozilla.com/D161667

Bug 1799695 - enable seamless video pref for test_seamless_looping_video.html. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D161617

Bug 1764293 - Add test to validate playback of muxed streams with mismatched start times up to 0.5 seconds. r=padenot Technically, this patch only needs test files that cover up to a 0.5 second stream start time mismatch, but I've included files up to 1.1 seconds in case we want to loosen the tolerance in the future (spec suggests ~1 second is OK). For now, I left the tolerance at 0.5 seconds so as not to change the value too drastically at once. Differential Revision: https://phabricator.services.mozilla.com/D160981

Bug 1764293 - Separate fuzzing values used for stream start/seek to better support muxed streams with mismatched start times as seen on Twitter/DokiDoki, etc. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D149586

Bug 1782366 - part2 : mapping audio extension to 'video/mp4' as well. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D157685

Bug 1782366 - part1 : don't use space in the pattern. r=padenot Although standard ftyp type contains spaces sometimes, in reality, not all of files follow this rule and they sometime skip the space. In order to increase the tolerance of parsing, we should allow ftyp without having spaces but with correct name. Eg. "M4A " v.s. "M4A". Differential Revision: https://phabricator.services.mozilla.com/D157684

Bug 1760920 - pt 10. Count only the media resources belonging to the document r=media-playback-reviewers,padenot Differential Revision: https://phabricator.services.mozilla.com/D146893

Bug 1799335 - enable seamless video pref in the test. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D161491

Bug 1799464 - enable seamless video for autoplay-hidden.optional.html. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D161493

Bug 1262276 - part18 : add a pref to control video seamless looping. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D160869

Bug 1262276 - part17 : disable late time check for seamless looping. r=padenot This doesn't work for seamless looping because the video end time can be larger than the media time. We would need to revisit these mechaniam later to see if we can have a better way to skip to next frame. Differential Revision: https://phabricator.services.mozilla.com/D160864

Bug 1262276 - part15 : disable seamless looping for test_video_low_power_telemetry.html in order to fix the test failure. r=padenot Differential Revision: https://phabricator.services.mozilla.com/D160700

Bug 1262276 - part14 : video-only seek should use original time as a seek target. r=padenot When requesting data from demuxer, it should always use normalized data, which is within [0, original-track-duration]. So we need to adjust seek target time in order not to use clock time which has grown over the duration due to multiple times of looping (that time is larger than duration). Differential Revision: https://phabricator.services.mozilla.com/D160699