When the publisher stops publishing, the individual receivers would close
attached DownTracks first before notifying MediaTrackReceiver callbacks.
This means #1454 does not fix the issue entirely since there is still a window
when we can subscribe to a closing track.
Due to the order of events in MediaTrackReceiver and friends, SubscribedTrack
will be closed before the track is removed from RoomTrackManager.
Because of this, when a track is unpublished, it's possible to be subscribed
to the track as it's closing.
By introducing a closing state, we'd prevent accidental subscription to
closing tracks.
Addressing edge case where a layer stopped before bitrate could be
measured. Purely bit rate based change deduction missed this as
the before and after did not have bit rates.
Use available layers to look for changes, especially currently
forwarding layer going away.
Also, simplifying bits. Only in the optimal allocation path,
these things are required. When congested, bitrate is always needed.
So, for optimal path, just look at available layer changes and adjust.
Don't need to look for bitrate based layer changes. Clean up that code.
The following sequence caused early migration complete declaration
1. Audio track received
2. Audio track published callback in progress
3. Video track received, this clears the pending track
4. Audio track published callback finishes. This checks for pending
tracks. As nothing is pending migration complete declared.
5. Due to the above, the remote video track is closed as not resuming.
That causes an unsubscription.
Fix
- Wait till publish callback to finish to remove a track from pending
fully.
- Introducing a new map as pending tracks is used in OnClose too. So,
did not want to delay removing from it as a close could happen while
publish callback is happening.
Also, moving the publish callback to a go routine (just like the recent
change for running those in a go routine for migrated muted tracks)
Not a good design. There is not an easy way to filter messages
before it hits media node. Without that, there is not a lot
of advantage.
And there are sequences that are not handled correctly in this
deleted implementation.
So, deleting code to prevent use.
Even if an add track has been queued and can be used immediately
when the previous incarnation unpublishes, send the unpublished
callback as the track was technically unpublished and republished.
The re-publish will pick the same track SID when the pending track
is queued as it will get the SID from an existing published track.
* let panics crash
* Revert "let panics crash"
This reverts commit 8027cccadd.
* catch and log panics then os.Exit
* Recover only recovers, caller can exit
* only exit on pacic, still need Recover calls in goroutines
UpdateSubscription had a shortcoming where when it couldn't find the
participant, it ignored the request.
This PR further removes the reliance of current publisher state from
subscribers.
- SubscribeToTrack only takes in a trackID
- Introduced RoomTrackManager to maintain all published tracks to a room
- Added TrackUnpublished event to clearly indicate when a track has been removed
- SubscribeRequested event no longer include information about the publisher
* Return early if already subscribed.
When already subscribed, returned `subTrack` is nil.
Return early, but do not return an error.
* check for nil subTrack
* check for nil as well
* Notes on wht to do
- Should targetLayers be altered while doing opportunistic locking
- Should targetLayers be altered in any other path than stream allocator path?
- Lock to layer as long as it is <= opportunistic layer
- When not congested, opportunistic can be highest
- When congested, opportunistic could be nil or lowest if paused is not allowed
- When muting, can we hold on to current layers (or keep it as previous) and
restore on unmute.
- Store current/target in forwarder state and restore on seeding
- Watch for looking for targetLayers, etc. when looking to insert padding
packets. There may be an assumption about restarting on key frame and hence
okay to insert padding when target layers are invalid. This may not be true
any more when doing opportunistic forwarding.
- Can we distinguish between publisher mute or dynacast (i. e. publisher side
stopping) vs subscriber mute and do something useful? Publisher side mute
could mean continuity in sequence numbers on a restart (might be able to
catch it with opportunistic forwarding). But, there is the challenge of
unmute from publisher via signalling channel vs media. If media is arriving,
should subscribers do opportunistic forwarding before publisher mute state
update happens?
- Maybe introduce a mode where forwarding continues to a frame end (of course
with a time limit just in case the end of frame packet is lost) and then
insert silence/padding packets?
- Ensure that audio blank frame insertion does not suffer from frame boundary
issues.
* pub/sub mute separate + more notes on things to check
* WIP commit, more notes
* WIP commit
* WIP commit
* WIP commit
* WIP commit
* WIP commit
* WIP commit
* clean up
* slightly better comments
* Do not stop on unmute
* do not inject blank frames when pub muted
* do not forward on audio publisher mute
* Add Timer to detect dtls failure quickly
* Fix pc state check in timeout after ice
* More strict conditions to switch candidate type
* log for signal interuppt
* typo
* Ignore inactive media.
Seeing some errors of
"track info not published prior to track".
Happens when trying to configure publisher answer for audio
for DTX and stereo and trying to find a pending track.
There are cases where there are inactive m-lines in SDP
that is hitting this.
* fix sense
When AdaptiveStream is enabled, default the subscriber to LOW quality stream
we would want LOW instead of OFF for a couple of reasons
1. when a subscriber unsubscribes from a track, we would forget their previously defined settings
depending on client implementation, subscription on/off is kept separately from adaptive stream
So when there are no changes to desired resolution, but the user re-subscribes, we may leave stream at OFF
2. when interacting with dynacast *and* adaptive stream. If the publisher was not publishing at the
time of subscription, we might not be able to trigger adaptive stream updates on the client side
(since there aren't any video frames coming through). this will leave the stream "stuck" on off, without
a trigger to re-enable it
