Remove Extra Lookup in PeerConnection::forwardMedia()

[SE2Dev]

Optimizes media forwarding somewhat.

[paullouisageneau]

While it's a good idea to optimize the dispatching by SSRC, I'm pretty sure the proposed changes won't work as-is for all use cases: since incomingTrack() is only called in processLocalDescription() when generating a local answer description, SSRCs present in a remote answer after a local offer won't be inserted in mTracksBySsrc. Have you checked the media-receiver example still work with the changes?

[SE2Dev]

While it's a good idea to optimize the dispatching by SSRC, I'm pretty sure the proposed changes won't work as-is for all use cases: since incomingTrack() is only called in processLocalDescription() when generating a local answer description, SSRCs present in a remote answer after a local offer won't be inserted in mTracksBySsrc. Have you checked the media-receiver example still work with the changes?

You're right, I checked against media-receiver and it doesn't work for that use-case. Do you think populating the map in PeerConnection::processLocalDescription and PeerConnection::processRemoteDescription would be a better method? I'm not entirely sure at what point the tracks are actually populated so I'm not really sure what a safe time to set up the mapping would be.

[paullouisageneau]

You're right, I checked against media-receiver and it doesn't work for that use-case. Do you think populating the map in PeerConnection::processLocalDescription and PeerConnection::processRemoteDescription would be a better method? I'm not entirely sure at what point the tracks are actually populated so I'm not really sure what a safe time to set up the mapping would be.

Actually, this is quite independant from tracks creation: for instance, an SSRC will commonly be added in the remote description for an already-existing track. I think updating the map in PeerConnection::processLocalDescription (after local tracks are added to the description) and PeerConnection::processRemoteDescription would be good.

[SE2Dev]

I've re-implemented the changes to take place in PeerConnection::processLocalDescription and PeerConnection::processRemoteDescription, and checked that it does still work with the media-receiver example.

[paullouisageneau]

Thank for the changes, I still have a few comments but now the approach looks good.

[paullouisageneau]

Looks good, thank you!

[paullouisageneau]

My bad, I approved a bit too fast, an additional call to updateTrackSsrcCache() was actually missing after creating incoming tracks. It is necessary because these tracks don't exist when processing the remote description. I fixed it in ef4a05a.