- From: Ingar Arntzen via GitHub <sysbot+gh@w3.org>
- Date: Tue, 30 Apr 2019 20:09:32 +0000
- To: public-web-and-tv@w3.org
On second thought. I'm not so sure about this after all. The sequencer implementation (open source JS alternative to track elements) guarantees enter/exit events for cues with length 0, and guarantees that the active cues list is always consistent with the history of event callbacks from the sequencer. However, the sequencer will deliver all due events in a single batch, implying that the enter and exit events for length 0 cues will always be delivered together (in the right order - depending on playback direction). So, the cue will not be a part of the active cues lists (unless playback is paused on exactly this point). Fixing this, would be possible (but complicated). I would still be reluctant to actually do this for the sequencer, as it might have performance implications for huge event batches. We have applications operating on datasets > 100K cues, and framerate reevaluation of large sets of active cues, which means that the batch oriented nature of the sequencer is crucial. So from a theoretical point of view, I guess this may be a bug. From a practical point not so much? -- GitHub Notification of comment by ingararntzen Please view or discuss this issue at https://github.com/w3c/me-media-timed-events/issues/41#issuecomment-488096936 using your GitHub account
Received on Tuesday, 30 April 2019 20:09:33 UTC