What version of Hls.js are you using?
development (v1.7 alpha)
What browser (including version) are you using?
Chrome and Safari
What OS (including version) are you using?
Mac OS
Test stream
https://playertest.longtailvideo.com/adaptive/captions/playlist.m3u8
Configuration
Additional player setup steps
No response
Checklist
Steps to reproduce
- Play content with 608 captions or VTT subtitles
Expected behaviour
All captions and subtitles are added to TextTracks.
What actually happened?
Initial cues added to TextTrack are deleted without "cuechange" being evoked on the TextTrack. This bug is not present in v1.6. This only happens in development, where DOM TextTracks are being used. This is a regression introduced with #7515, but it appears to be specific to DOM TextTracks in Chrome and Safari. It seems like the browser removes cues after isetting the mode to "hidden" and appending media. Setting the mode to "hidden" when creating the track, and then waiting until content is appended before adding cues, prevents them from being ejected.
The fix should be to defer cue insertion until after the media append is started or complete.
Console output
Chrome media internals output
What version of Hls.js are you using?
development (v1.7 alpha)
What browser (including version) are you using?
Chrome and Safari
What OS (including version) are you using?
Mac OS
Test stream
https://playertest.longtailvideo.com/adaptive/captions/playlist.m3u8
Configuration
Additional player setup steps
No response
Checklist
Steps to reproduce
Expected behaviour
All captions and subtitles are added to TextTracks.
What actually happened?
Initial cues added to TextTrack are deleted without "cuechange" being evoked on the TextTrack. This bug is not present in v1.6. This only happens in development, where DOM TextTracks are being used. This is a regression introduced with #7515, but it appears to be specific to DOM TextTracks in Chrome and Safari. It seems like the browser removes cues after isetting the mode to "hidden" and appending media. Setting the mode to "hidden" when creating the track, and then waiting until content is appended before adding cues, prevents them from being ejected.
The fix should be to defer cue insertion until after the media append is started or complete.
Console output
Chrome media internals output