slimserver icon indicating copy to clipboard operation
slimserver copied to clipboard

LMS incorrectly handles tracks added to playlists when players are playing from the buffer

Open 986ster opened this issue 2 years ago • 6 comments

If a track is added to the end of a playlist during the time that the player is playing music from its buffer only (i.e. LMS is no longer sending music to that player), the newly added track will not be played and the music will stop at the end of the buffered track(s). The track skip button has to be pressed in order to resume the playlist.

Similarly, if a track is inserted into a playlist using the Play Next button, it will be correctly played at the end of the current track if LMS is still sending current track music to the player, otherwise it will be inserted after the following track if the current track is playing from the buffer, meaning that a track ends up being inserted in the wrong place.

986ster avatar Dec 19 '23 17:12 986ster

More details in this forum post: https://forums.slimdevices.com/forum/user-forums/logitech-media-server/1658961-tracks-added-to-current-playlist-not-playing-if-player-buffer-has-already-been-filled

986ster avatar Dec 19 '23 17:12 986ster

See https://github.com/Logitech/slimserver/pull/963

mherger avatar Dec 20 '23 07:12 mherger

What player do you use for your tests? If you have a squeezelite and some real SB, I'd appreciate if you check for a difference

philippe44 avatar Dec 20 '23 08:12 philippe44

I do have some old Receivers still that I could dig out and revive, despite having replaced all of my in-use hardware with Squeezelite. I'm away over the Christmas break, but I could endeavour to stand one of these up for some parallel testing after I'm back. Does anyone know the buffer size in a Receiver, so I can make sure that if I do any testing I'll have a fair idea of whether I'm actually using the buffer or not?

986ster avatar Dec 20 '23 16:12 986ster

There is a bug in squeezelite that I discovered at the occasion of this discussion, so it's not a good test vehicle for this use case. I'm trying to see if there is a fix, but I'm not sure. It's something that has been here since the beginning: squeezelite is not able to correctly process flushing of the track being streamed (not played) and it confuses LMS in its response.

philippe44 avatar Dec 20 '23 17:12 philippe44