This bug was filed from the Socorro interface and is report bp-de255c5b-9266-4254-88b1-2e1170171130. ============================================================= Top 10 frames of crashing thread: 0 xul.dll mozilla::MediaPipeline::MediaPipeline media/webrtc/signaling/src/mediapipeline/MediaPipeline.cpp:591 1 xul.dll mozilla::MediaPipelineReceive::MediaPipelineReceive media/webrtc/signaling/src/mediapipeline/MediaPipeline.cpp:1997 2 xul.dll mozilla::MediaPipelineReceiveAudio::MediaPipelineReceiveAudio media/webrtc/signaling/src/mediapipeline/MediaPipeline.cpp:2130 3 xul.dll mozilla::TransceiverImpl::InitAudio media/webrtc/signaling/src/peerconnection/TransceiverImpl.cpp:81 4 xul.dll mozilla::TransceiverImpl::TransceiverImpl media/webrtc/signaling/src/peerconnection/TransceiverImpl.cpp:56 5 xul.dll mozilla::PeerConnectionMedia::AddTransceiver media/webrtc/signaling/src/peerconnection/PeerConnectionMedia.cpp:1148 6 xul.dll mozilla::PeerConnectionImpl::CreateTransceiverImpl media/webrtc/signaling/src/peerconnection/PeerConnectionImpl.cpp:1224 7 xul.dll mozilla::PeerConnectionImpl::CreateTransceiverImpl media/webrtc/signaling/src/peerconnection/PeerConnectionImpl.cpp:1252 8 xul.dll mozilla::dom::PeerConnectionImplBinding::createTransceiverImpl dom/bindings/PeerConnectionImplBinding.cpp:380 9 xul.dll mozilla::dom::GenericBindingMethod dom/bindings/BindingUtils.cpp:3042 ============================================================= There are 21 crashes in nightly 59 starting with buildid 20171129111030. In analyzing the backtrace, the regression may have been introduced by patch [1] to fix bug 1290948. [1] https://hg.mozilla.org/mozilla-central/rev/ab0fc6ccf118

This would be bug 1418367 rearing its ugly head. Bug 1397793 should fix this, and looks very close to ready to land. I could prevent it from crashing when this happens, though.

Looking at the crash stack don't you think we should handle failures to create a audio or video conduit more gracefully then we do right now? Sorry if this is what you were asking for regarding a workaround for bug 1397793 before landing transceivers. I think in the long run it would be helpful not to crash when audio/video conduit init failed.

[Tracking Requested - why for this release]: This is the #4 Windows topcrash in Nightly 20171201100115.

(In reply to Nils Ohlmeier [:drno] from comment #2) > Looking at the crash stack don't you think we should handle failures to > create a audio or video conduit more gracefully then we do right now? > > Sorry if this is what you were asking for regarding a workaround for bug > 1397793 before landing transceivers. I think in the long run it would be > helpful not to crash when audio/video conduit init failed. The only reason we ever noticed this problem is because of the crash, so I'm a little conflicted on this. I wish there was a "make this show up in crash-stats, but don't crash the browser" thing we could do here.

Comment on attachment 8934258 [details] Bug 1421965: Avoid nullptr crash when conduits cannot be created. https://reviewboard.mozilla.org/r/205178/#review210712 LGTM

Pushed by bcampen@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/178376765b9e Avoid nullptr crash when conduits cannot be created. r=drno

No more crashes in nightly 59 since the patch landed.

Tracking in case this reopens.