Re: Update of RTCRtpSender "doohickey" proposal

On 29 April 2014 00:52, Harald Alvestrand <harald@alvestrand.no> wrote:
> If the stream object is null, the receiver can choose to add the track to an
> existing stream or to create a stream for it.

That sort of non-deterministic behaviour needs to be stamped out if we
are going to have an interoperable spec.

I think that a new stream for every track that lacks an explicit
stream is a simple enough thing.  It might even interoperate.

Received on Tuesday, 29 April 2014 16:50:54 UTC