The video animation is clearly visible around 33s in the attached trace: just look for 'VideoPlayback' below the 'Mousewheel Animation' RAIL stage. |
||
Issue descriptionThe video animation is clearly visible around 33s in the this trace (https://console.developers.google.com/m/cloudstorage/b/chrome-telemetry-output/o/trace-file-id_49-2017-03-22_01-54-17-98522.html): just look for 'VideoPlayback' below the 'Mousewheel Animation' RAIL stage. That should clearly be showing up as a Video expectation, but it's not.
,
Mar 28 2017
It looks like the ProtoExpectations are indeed getting added for the VideoPlayback, but the video proto expectations are somehow getting filtered out in mergeIntersectingAnimations. The bug has to be somewhere in that code path, because multiple video playback expectations should never get filtered down to zero of them (which is what's happening).
,
Mar 28 2017
Interesting: the way that we identify a video protoexpectation is that we call pe.containsTypeNames([INITIATOR_TYPE.VIDEO]), which seems to be returning false even for ProtoExpectations. I wonder why?
,
Mar 28 2017
It looks like the way that we were detecting whether something was a video expectation when determining whether to merge intersecting expectations: var isVideo = pe.containsTypeNames([INITIATOR_TYPE.VIDEO]); always returned false, because INITIATOR_TYPE.VIDEO isn't a type name, its the initiator type. More broadly, we should always be able to just look at the initiator type of the proto expectation to determine what type of proto expectation it is. I'm going to send out a quick CL fixing this.
,
Mar 31 2017
This was fixed in https://codereview.chromium.org/2778993004/
,
Mar 31 2017
|
||
►
Sign in to add a comment |
||
Comment 1 Deleted