The cast in InitWithNode is wrong. AsElement() asserts instead of
checking the flag, so we always pass an element (and if we didn't we'd
have type confusion problems). I audited the callers and we're fine.
Anyhow, always require an element, and add two convenience constructors
for C++ code.
Differential Revision: https://phabricator.services.mozilla.com/D73636
The cast in InitWithNode is wrong. AsElement() asserts instead of
checking the flag, so we always pass an element (and if we didn't we'd
have type confusion problems). I audited the callers and we're fine.
Anyhow, always require an element, and add two convenience constructors
for C++ code.
Differential Revision: https://phabricator.services.mozilla.com/D73636
This is in preparation for running the tail dispatcher off
nsIThreadObserver callbacks, which only work during regular
event processing.
Differential Revision: https://phabricator.services.mozilla.com/D72264
This patch will do :
- rename `ControlledMediaState` to `MediaPlaybackState`
- rename the related functions
The advantage of doing so :
- more consistent with `MediaAudibleState`
Differential Revision: https://phabricator.services.mozilla.com/D72060
This patch will do :
- replace `boolean` with enum class `MediaAudibleState`
The advantage of doing so :
- It's easier to understand what actually meaning of the parameter we set
Differential Revision: https://phabricator.services.mozilla.com/D72058
This patch will do :
- use current broswing context as a parameter when propagate state change to the chrome process.
The advantage of doing so :
- the chrome process can know which browsing context the state change actually comes from.
---
More details about this change :
Currently, when we propagate any controlled media related states, we would find the top level browsing context first, then pass it through IPC in order to get the correct media controller in the chrome process. However, we have implemented [1] which can find the correct media controller even if we are not passing the top level browsing context.
In addition, in bug1627999, we would like to know which browsing context those states come from. Therefore, we should replace the top browsing context with the current browsing context where controlled media exists.
[1] https://searchfox.org/mozilla-central/rev/41c3ea3ee8eab9ce7b82932257cb80b703cbba67/docshell/base/CanonicalBrowsingContext.cpp#511-515
Differential Revision: https://phabricator.services.mozilla.com/D72056
This patch will do :
- remove the inheritance relationship for `ContentControlKeyEventReceiver` and `ContentMediaAgent` and manually implement the methods we need
- `MediaControlEventListener` will inherit from `ContentControlKeyEventReceiver` directly
The advantage of doing so :
- increase the flexibilty of modification of `ContentMediaAgent` and those modification won't affect other classes inherited from `MediaControlKeysEventListener`
---
More details about this change :
As we would like to extend the class `ContentMediaAgent` and allow the `ContentMediaController` to call its extended method, but if we want to do so in current implementation, the extended method would also affect other classes inherited from `MediaControlKeysEventListener` and that is something we don't want to see.
Considering that, I decided to decouple the inheritance relationship and manually create the function I need (which will be implemented in the next patch)
Differential Revision: https://phabricator.services.mozilla.com/D72054
After applying this patch,
- notify `inaudible` state only if we already notified `audible` state before
- update `inaudible` state when audible media is paused
Therefore, in the media controller side, it would always see a pair of `audible/inaudible` notifications like what we do for the pair of `ControlledMediaState::ePlayed/ePaused` notifications. That can help us implement a way of maintaining a detailed counting about how many media is audible in bug1627999.
Differential Revision: https://phabricator.services.mozilla.com/D71429
There is actually possible to start the listener already while running `SuspendOrResumeElement()`, so we should remove the assertion and use a check instead.
Eg. JS can call `play()`, which would start the listener, before we run this method. This situation can be found when browsing Youtube on the GeckoView.
Differential Revision: https://phabricator.services.mozilla.com/D69673
When the docShell's `SuspendMediaWhenInactive` flag is true, no media should be allowed to start playing. Therefore, we add a check in `Play()`, `CanActivateAutoplay()` to prevent media from playing. In addition, we should also prevent the audio channel agant from starting.
Differential Revision: https://phabricator.services.mozilla.com/D69672
If docShell's `SuspendMediaWhenInactive` is true, then we should suspend or resume the media element according to the docshell active state when the docshell changes it active state.
Differential Revision: https://phabricator.services.mozilla.com/D69671
AutoEnter was an attempt around a race between AbstractThread and MessageLoopAbstractThreadWrap that would cause AbstractThread::GetCurrent() to return an incorrect value. MessageLoopAbstractThreadWrapper is no more and as such AutoEnter is no longer required.
Differential Revision: https://phabricator.services.mozilla.com/D71279
There is actually possible to start the listener already while running `SuspendOrResumeElement()`, so we should remove the assertion and use a check instead.
Eg. JS can call `play()`, which would start the listener, before we run this method. This situation can be found when browsing Youtube on the GeckoView.
Differential Revision: https://phabricator.services.mozilla.com/D69673
When the docShell's `SuspendMediaWhenInactive` flag is true, no media should be allowed to start playing. Therefore, we add a check in `Play()`, `CanActivateAutoplay()` to prevent media from playing. In addition, we should also prevent the audio channel agant from starting.
Differential Revision: https://phabricator.services.mozilla.com/D69672
If docShell's `SuspendMediaWhenInactive` is true, then we should suspend or resume the media element according to the docshell active state when the docshell changes it active state.
Differential Revision: https://phabricator.services.mozilla.com/D69671
Video element can start the Picture-in-Picture mode **BEFORE** or **AFTER** we start the listener for the media cotrol, so we have to ensure we always propagate this information to the chrome process via `ContentMediaAgent`.
Differential Revision: https://phabricator.services.mozilla.com/D67712
The media element being used in the PIP mode would always display on the screen even if the the tab it belongs to is in the background, so users would have high chance to interact with it again. Therefore, we don't want to start a stop-timer to cancel media control for that element.
Differential Revision: https://phabricator.services.mozilla.com/D67381
In order to be able to change the audio output device, the device id is exposed to the constructor of MediaStreamGraph. Any component that gets/creates an MTG will be able to provide the device id in order to determine the desired output device. If the provided device id is null the default device will be used.
Differential Revision: https://phabricator.services.mozilla.com/D64133
Intercepting media control keys would also come up with the virtual control interface in most of platforms, and for the notification sound we don't want either to show the interface or control them.
Currently we use 3s as a threshold to filter those short duration media which are possible to be a notification sound.
Differential Revision: https://phabricator.services.mozilla.com/D64478
This patch is used to prevent intercepting media keys from other applications which should really receive media keys. For example, if user finishes watching a video but doesn't close that tab, and switch to other tasks. Then if they have other applications which are playing music, when they press media keys, the application which should receive media keys should be those background music app, not Firefox.
Therefore, if the media has been paused for a while, then that might indicate that user won't need to control this media anymore. Therefore, we create a timer, which default expire time is 60s, to stop listening media control key events for those kinds of media. But if the media starts again during this period, we would destroy timer and continue to control that media.
Differential Revision: https://phabricator.services.mozilla.com/D63918
It's useless to update the audible state before starting listening to the media control keys event, so we make it consistent with other functions which could only be used after starting the listener.
As the audible change could happen before we start the listener, we have to manually update audible state to the controller after we start the listener.
Differential Revision: https://phabricator.services.mozilla.com/D63943
When media is not audible yet, or it's inaudible at all (maybe it's used as a background video or GIF-like image), or the media is in the tab which has been muted totally, we don't want to use media control to control those media, so we should postpone starting listening to the media control keys for them event until they become audible.
Once they become audible, we would keep control them all the time even if they become inaudible again.
Differential Revision: https://phabricator.services.mozilla.com/D63915
Those suspend and resume related logic were used for implemeting media control and audio focus on Fennec. After removing those things, the only purpose of `AudioChannelAgentCallback` would be muting and unmuting media element from audio channel (eg. clicking sound indicator to mute/unmute tab)
Differential Revision: https://phabricator.services.mozilla.com/D65264
This attribute is used to expose the internal suspend state of media element to see if we set `SUSPENDED_PAUSE` and `SUSPENDED_PAUSE_DISPOSABLE` correctly. As we don't use those suspend state anymore and have removed the related tests, no need to keep this attribute anymore.
Differential Revision: https://phabricator.services.mozilla.com/D65263
`media-playback-resumed` was a notification we used on Fennec to indicate media playback state changes and we should update the UI button on the virtual control interface. As it's not being used by anyone since we deprecate Fennec, we can remove all related code.
Differential Revision: https://phabricator.services.mozilla.com/D65258
At the time CC unlink happens, we would no longer need to listen to the media control key events. Therefore, we should remember to stop and clear the listener to avoid its any method ran after CC.
Differential Revision: https://phabricator.services.mozilla.com/D64787
This covers most cycle collected objects which support weak references, but
not the ones which inherit from a cycle collected class and don't do any cycle
collection on their own.
Differential Revision: https://phabricator.services.mozilla.com/D63962
The check here was used to prevent calling play or pause during playing or pausing.
However, according to the spec [1], even if media session is playing, we would still want to trigger its play action handler when play action happens. So we should forward all media control key events to the content, not to block any of them.
[1] https://w3c.github.io/mediasession/#handle-media-session-action
Differential Revision: https://phabricator.services.mozilla.com/D60918
When media element is inactive, asking MDSM to seek is actually useless, because all the data we decode won't be showed to user. In addition, we have to store extra pending events for `seeking` and `seeked`, which might result in memory overflow if the inactive page is calling `seek()` all the time.
Therfore, we should delay all seeking tasks while media is inactive, and perform the latest seeking task when media becomes active.
Differential Revision: https://phabricator.services.mozilla.com/D58918
Two parameters in `SuspendOrResumeElement()` are acutally the same, they are both related with `IsActive()`, so using one parameter is enough.
In addition, we only suspend or resume element by checking element's active state, so rename `mPausedForInactiveDocumentOrChannel` to `mSuspendedForInactiveDocument` which fits the actual situation more.
Differential Revision: https://phabricator.services.mozilla.com/D58749
