You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In the case when hls_fragmentloadskipaftermaxretry is true and the playlist is live or very long it's desirable to get an error after a specified amount of load errors on subsequent fragments in a row. This option will allow us to reload the playlist from another source or just log an error to the user.
Current behavior is like this:
// config
hls_fragmentloadmaxretry : 1,
hls_manifestloadmaxretry : 1,
hls_fragmentloadskipaftermaxretry: true
// logs
loading playlist -> OK
loading fragment 1 -> failed
loading fragment 1 -> failed
skip fragment 1
loading fragment 2 -> failed
loading fragment 2 -> failed
skip fragment 2
loading fragment 3 -> failed
loading fragment 3 -> failed
skip fragment 3
and so on to the infinity.
It will never raise an error even when we reach the last fragment
on the playlist (when it's VOD)
This is illustration of behaviour I would like to get:
// config
hls_fragmentloadmaxretry : 1,
hls_manifestloadmaxretry : 1,
hls_fragmentloadskipaftermaxretry: true,
hls_maxskippedfragments: 3
// logs
loading playlist -> OK
loading fragment 1 -> failed
loading fragment 1 -> failed
skip fragment 1
loading fragment 2 -> failed
loading fragment 2 -> failed
skip fragment 2
loading fragment 3 -> failed
loading fragment 3 -> failed
skip fragment 3 -> skipped fragment limit reached
loading fragment 4 -> failed
loading fragment 4 -> failed
I/O ERROR raised
In case when the next fragment after skip is successfully loaded the skip counter is reset, so it's only applicable to the cases when the errors occur in a row.
The text was updated successfully, but these errors were encountered:
In the case when
hls_fragmentloadskipaftermaxretry
istrue
and the playlist is live or very long it's desirable to get an error after a specified amount of load errors on subsequent fragments in a row. This option will allow us to reload the playlist from another source or just log an error to the user.Current behavior is like this:
This is also related with #368.
This is illustration of behaviour I would like to get:
In case when the next fragment after skip is successfully loaded the skip counter is reset, so it's only applicable to the cases when the errors occur in a row.
The text was updated successfully, but these errors were encountered: