Some Questions about the implementations

I am curious about some of the implementations.

  1. It seems that the demo-app.js wrote different code for different browser version. Could it use adapter.js to handle the compatibility issue of different browsers?

  2. The webrtc-internal shows the client generates two PeerConnections. Is there any reason to use two PCs instead of one?

  3. Since the client uses PCs, I expect the server also use PCs.
    I wonder where is the server-side code for creating PCs and negotiation offer/answers.

Thanks.

Why? We have mediasoup-client that adapts to each different browser. Why would we need adapter.js?

Yes, to avoid the pain of managing SDPs with bidirectional m=sections.

No, mediasoup (server side) does not use SDPs but RTP parameters. It’s mediasoup-client who builds “remote SDPs” to conform to the WebRTC API of browsers.

BTW: You should read the mediasoup documentation instead of making random questions. Everything is explained in there.