FATAL_ERROR after second seek near end of the stream
Prerequisites: Google Chrome 87.0.4280.66 BigScreen player: 4.4.3 Stream: https://dash.akamaized.net/akamai/bbb_30fps/bbb_30fps.mpd (default)
Steps to reproduce: In devtools execute below commands on video tag:
- $0.play()
- $0.currentTime = 634.1
- Wait for video to ends
- $0.currentTime = 0
- $0.play()
- $0.currentTime = 634.1
- The second seek should fail with FATAL_ERROR, if not repeat points 4-6
The bug is a some kind of a race in the JS. Bug reproduce more frequently after enabling Verbose log level. Error occures also in Vewd Tv browser.
Note: Near-end seek is extensively used in BBC Iplayer. When user perform the seek to the end of a stream in the UI, the BBC Iplayer set the current time right before end of a stream to play last seconds of the video. The issue does not reproduce on Dash Reference Client 3.1.3
Log from first succeeded seek: first_seek.log
Log from second failed seek: second_seek.log
Thanks @jarek-vewd we will take a look.
Hi @jlks Have you had time to see the error?
@jarek-vewd Hi. We've now released an updated version of bigscreen-player along with Dash.js v3.2.0.
Let me know if this resolved your issues 😄
Hi @domjacks I check it and it looks that the "second seek problem" is solved in some way, no fatal error. However, another problem occurs, IMO related to this solution. After seeking to the end, it shows the recommendation screen only for series and only if there is next episode available. For movies and last episodes of series, nothing happens for 60sec and then it comes back to the movie/series main screen. As movies are handled as 1 episode series, I think there is a problem with handling the recommendation screen on the last episode of the series.
Hey @wfilipek, thanks for the contact.
This sounds like a BBC iPlayer problem and not specific to Bigscreen-player per se.
If you wouldn't mind, it would be great if you could send your feedback using the BBC iPlayer help webpage: https://www.bbc.co.uk/iplayer/help/contact-iplayer-help
Hi @domjacks It looks that I assumed too quickly that the second-seek problem is solved. As the repro ratio is not 100%, today I perform a re-check and the problem still occurs, the same as described originally.