Re: [webrtc-pc] Promises on the operations chain may never settle (#2973)

Can you clarify what edge case needs to be handled, and provide support for the behavior being "unexpected"?

It's possible we made the wrong [decision](https://github.com/w3c/webrtc-pc/issues/150#issuecomment-152333247) years ago, but this has shipped in all browsers, so changing behavior now would come at a web compat cost of triggering unexpected errors on close where none are triggered today. So I think we'd need a pragmatic reason to change it at this point.

If this is causing a lot of difficult to debug memory leaks, that would be concerning and might be a reason to reconsider this design, but I'm not sure the example given is sufficient to conclude this is happening.

In practice, applications are encouraged to use `peerConnection.onnegotiationneeded` to isolate its negotiation code, and the application knows when it calls `close()` which the browser never calls except on document unload.

-- 
GitHub Notification of comment by jan-ivar
Please view or discuss this issue at https://github.com/w3c/webrtc-pc/issues/2973#issuecomment-2127913561 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Thursday, 23 May 2024 19:52:54 UTC