This came up in a discussion....
in the RTCWEB WG, we spent a lot of time making sure it was possible to operate without telling the other guy what your SSRCs were ahead of time, because this was important for interworking with various stuff that allocated SSRCs "late in the game".
Now .... is anyone using that functionality now, in an operational WebRTC-based service where they set up connections without signalling SSRCs before sending data in at least one direction?
I'm not convinced anything in web-platform-tests will test this, but arguably it should. Perhaps someone has a nice little test to offer?
Harald