Currently the local and remote candidates/pairs show up in no particular order in the "ICE stats" section (why is it called like that??? - would "ICE connections" make more sense). I think it would make sense if we would at least group candidates and pairs in ICE components, to make it easier to understand which candidates are from several interfaces but for the stream vs candidates for multiple streams.

Comment on attachment 8947533 [details] Bug 1414171 - pt 1 - Add componentId field to RTCIceCandidatePairStats. https://reviewboard.mozilla.org/r/217234/#review223032 ::: dom/webidl/RTCStatsReport.webidl:141 (Diff revision 1) > unsigned long long bytesSent; > unsigned long long bytesReceived; > DOMHighResTimeStamp lastPacketSentTimestamp; > DOMHighResTimeStamp lastPacketReceivedTimestamp; > boolean selected; > + unsigned long componentId; // moz this needs [ChromeOnly], since this is Gecko internal thing. As far as I see, about:webrtc uses chrome privileges. with that, r+

Comment on attachment 8947534 [details] Bug 1414171 - pt 2 - Sort ICE stats by componentId on about:webrtc. https://reviewboard.mozilla.org/r/217236/#review223126

Comment on attachment 8947616 [details] Bug 1414171 - pt 3 - Add componentId to expected fields in test_peerConnection_stats.html. https://reviewboard.mozilla.org/r/217314/#review223128

Pushed by mfroman@nostrum.com: https://hg.mozilla.org/integration/autoland/rev/281f1e0d4f69 pt 1 - Add componentId field to RTCIceCandidatePairStats. r=smaug https://hg.mozilla.org/integration/autoland/rev/07eb80c26253 pt 2 - Sort ICE stats by componentId on about:webrtc. r=drno https://hg.mozilla.org/integration/autoland/rev/40d90dc4aa87 pt 3 - Add componentId to expected fields in test_peerConnection_stats.html. r=drno

https://hg.mozilla.org/mozilla-central/rev/281f1e0d4f69 https://hg.mozilla.org/mozilla-central/rev/07eb80c26253 https://hg.mozilla.org/mozilla-central/rev/40d90dc4aa87