[Webkit-unassigned] [Bug 174314] RTCPeerConnection's close method should update signalingState

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=174314

--- Comment #1 from Mark Roberts <mroberts at twilio.com> ---
RTCPeerConnection's `close` method should also update `iceConnectionState`. Perhaps this issue could be tracked here, too, instead of a separate ticket.

Run the following code:

    (async () => {
      console.log('1. Constructing an RTCPeerConnection')
      const pc = new RTCPeerConnection()
      const iceConnectionStateChangeEvent = new Promise(resolve => pc.oniceconnectionstatechange = resolve)
      console.log('2. Closing the RTCPeerConnection')
      pc.close()
      if (pc.iceConnectionState !== 'closed') {
        console.error(`Expected the RTCPeerConnection's ICE connection state to be "closed", not "${pc.iceConnectionState}"`)
      }
      console.log('3. Awaiting a "iceconnectionstatechange" event')
      await iceConnectionStateChangeEvent
      console.log('Success!')
    })()

Expected output (Firefox):

    1. Constructing an RTCPeerConnection
    2. Closing the RTCPeerConnection
    Expected the RTCPeerConnection's ICE connection state to be "closed", not "new"
    3. Awaiting a "iceconnectionstatechange" event
    Success!

Actual output (Safari):

    1. Constructing an RTCPeerConnection
    2. Closing the RTCPeerConnection
    Expected the RTCPeerConnection's ICE connection state to be "closed", not "new"
    3. Awaiting a "iceconnectionstatechange" event

In the actual output, "Success!" never prints because `iceConnectionStateChangeEvent` never resolves. This behavior seems specified in the Step 4 of `close` and the description of "closed" in RTCIceConnectionState:

    * http://w3c.github.io/webrtc-pc/#dom-rtcpeerconnection-close
    * http://w3c.github.io/webrtc-pc/#rtciceconnectionstate-enum

(Note: In Chrome, the error prints, too, I guess because `iceConnectionState` isn't updated synchronously?)

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20170710/8b7a0026/attachment-0001.html>