This will help debugging.
Created attachment 311200 [details] Patch
Created attachment 311201 [details] Patch
Comment on attachment 311201 [details] Patch Clearing flags on attachment: 311201 Committed r217436: <http://trac.webkit.org/changeset/217436>
All reviewed patches have been landed. Closing bug.
Comment on attachment 311201 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=311201&action=review > Source/WebCore/platform/mediastream/mac/RealtimeIncomingVideoSource.cpp:134 > + if (!(++m_numberOfFrames % 30)) > + RELEASE_LOG(MediaStream, "RealtimeIncomingVideoSource::OnFrame %zu frame", m_numberOfFrames); I guess this logs about once per second. Can we reduce it a bit? > Source/WebCore/platform/mediastream/mac/RealtimeOutgoingVideoSource.cpp:151 > + RELEASE_LOG(MediaStream, "RealtimeOutgoingVideoSource::sendOneBlackFrame"); How verbose will this be? > Source/WebCore/platform/mediastream/mac/RealtimeOutgoingVideoSource.cpp:178 > + if (!(++m_numberOfFrames % 30)) > + RELEASE_LOG(MediaStream, "RealtimeOutgoingVideoSource::sendFrame %zu frame", m_numberOfFrames); I guess this logs about once per second. Can we reduce it a bit?
(In reply to Simon Fraser (smfr) from comment #5) > Comment on attachment 311201 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=311201&action=review > > > Source/WebCore/platform/mediastream/mac/RealtimeIncomingVideoSource.cpp:134 > > + if (!(++m_numberOfFrames % 30)) > > + RELEASE_LOG(MediaStream, "RealtimeIncomingVideoSource::OnFrame %zu frame", m_numberOfFrames); > > I guess this logs about once per second. Can we reduce it a bit? What would be a good tradeoff? > > Source/WebCore/platform/mediastream/mac/RealtimeOutgoingVideoSource.cpp:151 > > + RELEASE_LOG(MediaStream, "RealtimeOutgoingVideoSource::sendOneBlackFrame"); > > How verbose will this be? 1 per second
For incoming frames, we could only log frame size changes, which would reduce to usually ensure that we received one frame out of libwebrtc. Then we could punt on getStats logging that we could trigger from time to time.
<rdar://problem/32479732>