* Do not hold lock while invoking resolver.
Resolver is in room and it will grab its lock.
It is called from partcipant when checking permissions.
Permissions processing uses the participant lock.
So, not a good idea to call a room function with
participant lock held. Avoid that.
Also, use a full lock in the add/remove subscription
path. This is to ensure that permissions path and
subscription path (like subscribing to new participant's
tracks) do not race. As subscriptions are queued now
on the subscriber side, this should be fine.
* Revoke path resolution in ops queue goroutine
* fix test
* Keep track of pending subscriber operations.
This is required to determine if a receiver does not have
any subscription.
* correct spelling of queuing
* lock around hasPermission
* Move subscribe/unsubscribe queue to participant.
As subscribe/unsubscribe operation can come from both
local media track or remote media track, participant
needs to have it.
* Remove comment
* Stop reneg timer on close
* address comments
* Send permissions update on subscribe.
The permission `allowed` update was happening only when
processing pending subscriptions (which happens only on
subscription permissions update).
It is possible that subscription happens through other
paths (like subscribing new participant to tracks).
In that path, we were checking if the track has permissions
and adding to pending. But, we were not checking if
the track is in pending and if it is in there, removing
in on successful subscription and sending an update.
Fix that.
* log more fields in error
As there is a queue to send dynacast update, forcing
an update on unmute should be fine. That will send
the current state. If the subscribers change it,
an update will be sent as necessary.
This addresses the case of subscription changes happening
when the published track is muted. Dynacast updates are
not sent when publisher tarck is muted. If on unmute,
if subscribers do not have any changes, an update is missed
(i. e. the changes that happen when publisher track is muted
is not sent).
* Use a queue for add/remove subscribe operations.
If subscribe/unsubscribe happens very quickly, the subscription
state gets mixed up as things are keyed off of subscriberID.
Use a queue of subscribe operations and process it serially.
* set up callback for down track added
* move the queue on unexpected type
* move the queue if removeSubscirber does not have a subscribed track
* WIP commit
* Clean up
* spelling mistake
* Run subscribed track onBind in a go routine
* Address comments and more safety net
* Cache and restore forwarder state on resume
* conflicts
* mage generate
* Log reason for participant close
Please suggest better naming for different scenarios if something comes
to mind.
* group service requests
* incorporate feedback
* Change names in tests
* Reinstating a function used by cloud
* Use go map for `subscribedTo`.
Range over a sync.Map could produce unexpected results if
there are concurrent changes (change in value or delete).
While we do not change value (the value is a place holder in this map),
just moving to go map to avoid hard to find concurrency issues.
The lock scoping is well defined for the map. So, making the change
to be more defensive.
* Set DtxDisabled from TrackInfo in score calculation.
Also, fix sending connection quality upate on a new subscription.
* comments tweaks
* Move TrackInfo into StreamTrackerManager as this is used by cloud as well
* WIP commit
* WIP commit
* Remove debug
* Revert to reduce diff
* Fix tests
* Determine spatial layer from track info quality if non-simulcast
* Adjust for invalid layer on no rid, previously that function was returning 0 for no rid case
* Fall back to top level width/height if there are no layers
* Use duration from RTPDeltaInfo
Multiple parts of the system relies on a Track's ID (egress, telemetry, etc).
Track ID changes when a track was unpublished, then republished with the
exact same attributes. This PR would allow us to re-use a previously
unpublished Track ID
* Fixed unclean DownTrack close when removed before bound.
When a DownTrack is closed before it had a chance to be bound to a
transceiver, we'd skip close and leave it hanging. This is unlikely in
normal operations. However, it can be seen with permissions and
subscription APIs.
* remove remaining peerID references
