Gary Katsevman

This is likely because the values in the spec changed but vtt.js hasn't been updated with for that. Hadn't noticed that positionAlign here was still using the old values.

We definitely should fix this but my focus is currently on things like #50 because the webvtt spec currently has position and line alignment options as at-risk https://htmlpreview.github.io/?https://github.com/gkatsev/webvtt/blob/at-risk-june/archives/2019-06-19/Overview.html#status.

Actually, looking at the [implementation report](https://www.w3.org/wiki/TimedText/WebVTT_Implementation_Report) of webvtt again, it looks like Firefox may have an implementation and if we fix it here, we can claim it as another implementation...

Yes, I'd like to get this done some time this year, though, it likely needs to be re-done based on the refactor in master. Another thing is to decide what...

Unfortunately, we don't expose those errors to Video.js users. It's available in the console or if you are doing the parsing yourself.

What is the media sequence number before and after this origin change? If it gets reset to zero, then the stream isn't valid and that's likely why it stops working.

can you please post a sample of the m3u8s from before the origin switch, during the origin switch, and after the origin switch? Even better if you can post the...

In the HLS spec, it says that the media sequence "MUST NOT decrease or wrap" https://datatracker.ietf.org/doc/html/draft-pantos-hls-rfc8216bis#section-6.2.2 Behavior like this makes the stream invalid and likely will have issues in other...

v7 should have both: https://github.com/videojs/video.js/blob/3a128d1d238a3110f1d90c59eb81996e991ba446/src/js/video.js#L435

Ignore me, @mister-ben is correct. Needs a fallback for v7 as it doesn't have `.time`.