W3C home > Mailing lists > Public > public-webrtc-logs@w3.org > January 2019

Re: [webrtc-pc] Need something such as rtpSender.reset() (#2087)

From: Iñaki Baz Castillo via GitHub <sysbot+gh@w3.org>
Date: Thu, 31 Jan 2019 15:11:23 +0000
To: public-webrtc-logs@w3.org
Message-ID: <issue_comment.created-459379663-1548947482-sysbot+gh@w3.org>
> stop() + addTransceiver(track, params) will create a new transceiver. It uses code already present.
> A reset() that doesn't create a new transceiver is a new code path.
> 
> If you keep the MID, then you're telling the other guy to keep his transceiver, even though it's being renegotiated - which means that you can't change # of layers - which was the original point.
> 
> again: Why do you think you need this?

In my simplified use case, I've a PeerConnection I just use for sending media. I want to send a video track, "close" it, then send a new video track without having to reuse previous RtpSender parameters (`encodings`, `mid`, etc). Well, I don't care much about reusing previous `a=mid` or not. In addition I'd like that my local SDP does not grow every time I close a transceiver and add a new one.

So if I understand your text above, calling `transceiver.stop()` (not implemented in Chrome BTW) and later call `pc.addTransceiver()` I would get a new `RtcTransceiver` instance (of course) and the previously stoped m= section would be reused, now with a fresh RtpSender and even with a new `a=mid`. Is that right?

-- 
GitHub Notification of comment by ibc
Please view or discuss this issue at https://github.com/w3c/webrtc-pc/issues/2087#issuecomment-459379663 using your GitHub account
Received on Thursday, 31 January 2019 15:11:25 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 9 October 2019 15:15:01 UTC