Fix race condition with full reconnect sequence during server restart (#663)
* Fix race condition with full reconnect sequence during server failure
When all participants are reconnecting at the same time, it creates an
opportunity for participant updates included in JoinResponse to be applied
after ParticipantUpdate events. This is due to await returning control
to us when we are performing `await RTCEngine.join`
Checking ParticipantInfo version prior to updating would guarantee we do
not wipe out any existing tracks that were received
* changeset