[Fix]Speedup Songs node by skipping pointless fetch of cuesheet data #11785
Chipping away at what makes songs node on large library so slow.
The way embedded cueshseets are implemented is taking 7% of the total time taken to process the songs node, even when there are no embedded cuesheets in the music collection. Even worse the cuesheet processing in
It is a minor change to stop fetching embedded cuesheet replay gain data by default, and I suggest we backport it for an immediate performance gain.
I am looking at fixing the cuesheet data design and replaygain issues for Leia in a separate PR.
referenced this pull request
Mar 1, 2017
Jenkins fail is libavcode issue, nothing to do with this PR.
Looks like embedded cuesheets have been broken by the unification of music lib and file views #8011 back in Jarvis. Not only is the lib data ignored, but when not in the lib play from file view results in a corrupted current playlist