That makes sense. Do you store the cname/msid/etc. on the focus, for when you want to do renegotiations? I don't know how XMPP exactly works, but when using SDP I should send signaling "a=inactive" (for example), to temporarily mute a client. In this case, I would have to repeat the same cname, msid, etc. in the follow-up SDP.
From: users [mailto:firstname.lastname@example.org] On Behalf Of Boris Grozev
Sent: Sunday, July 31, 2016 08:40
To: Jitsi Users <email@example.com>
Subject: Re: [jitsi-users] cname not supplied in REST interface
What we do is have Jicofo (the component that controls the bridge) read the bridge's ssrc and insert the cname/msid/etc in the offer to the client. The bridge does not perform offer-answer.
On Saturday, July 30, 2016, Oliver Hausler <firstname.lastname@example.org<mailto:email@example.com>> wrote:
As explained here https://github.com/jitsi/jitsi-videobridge/blob/master/doc/rest-videobridge.md, when using the REST interface, I need to convert the JSON response from Jitsi to SDP.
Problem: Jitsi only supplies an SSRC in [Sources], but not a cname attribute.
To create an SDP compatible with WebRTC, I need to add the SSRC to the SDP:
Assuming that Jitsi does not use cnames because it never changes an ssrc, is it a valid workaround to add the SSRC from sources redundantly as cname, as shown below?