1210960 - Intermittent test_peerConnection_basicH264Video.html | application crashed [@ mozilla::WebrtcGmpVideoEncoder::Encode_g(webrtc::I420VideoFrame const*, webrtc::CodecSpecificInfo const*, std::vector<webrtc::VideoFrameType, std::allocator<webrtc::VideoFrameTy

Status: RESOLVED INCOMPLETE

Description

[Wes Kocher (:KWierso)]

https://treeherder.mozilla.org/logviewer.html#?job_id=15107896&repo=mozilla-inbound

Comment 1

[Randell Jesup [:jesup]]

Dup?

Comment 2

[Chris Pearce (:cpearce)]

I don't know of a dupe.

You're hitting Unexpected VideoFrameType here:

https://dxr.mozilla.org/mozilla-central/source/media/webrtc/signaling/src/media-conduit/WebrtcGmpVideoCodec.cpp#130

That's not code media playback ever hits.

Comment 3

[Randell Jesup [:jesup]]

(In reply to Chris Pearce (:cpearce) from comment #2)
> I don't know of a dupe.
> 
> You're hitting Unexpected VideoFrameType here:
> 
> https://dxr.mozilla.org/mozilla-central/source/media/webrtc/signaling/src/
> media-conduit/WebrtcGmpVideoCodec.cpp#130
> 
> That's not code media playback ever hits.

Thanks.  And looking over that code it seems pretty impossible; Encode() in generic_encoder in webrtc.org is preceded by ConvertFrameType, which asserts on anything that isn't a key or delta frame.  The initialization of nextFrameTypes in video_sender.cc is wrong ("1, kVideoFrameDelta" - 1 is kAudioFrameSpeech, but that should trigger an assert() when converted -- and if it doesn't, it'll end up Delta anyways.  And in GMP, Encode() dispatches Encode_g() via DISPATCH_SYNC.

So this should be truly impossible.