Bring it back in line with Sonarr. I was originally worried that we
were caching all Artists and this would be a resource hog. But the
caching will happen on every RSS scan so I think we're better off just
holding the cache and not worrying.
* Cache result of GetAllArtists
* Fixed: Manual import not respecting album import notifications
* Fixed: partial album imports stay in queue, prompting manual import
* Fixed: Allow release if tracks are missing
* Fixed: Be tolerant of missing/extra "The" at start of artist name
* Improve manual import UI
* Omit video tracks from DB entirely
* Revert "faster test packaging in build.sh"
This reverts commit 2723e2a7b8.
-u and -T are not supported on macOS
* Fix tests on linux and macOS
* Actually lint on linux
On linux yarn runs scripts with sh not bash so ** doesn't recursively glob
* Match whole albums
* Option to disable fingerprinting
* Rip out MediaInfo
* Don't split up things that have the same album selected in manual import
* Try to speed up IndentificationService
* More speedups
* Some fixes and increase power of recording id
* Fix NRE when no tags
* Fix NRE when some (but not all) files in a directory have missing tags
* Bump taglib, tidy up tag parsing
* Add a health check
* Remove media info setting
* Tags -> audioTags
* Add some tests where tags are null
* Rename history events
* Add missing method to interface
* Reinstate MediaInfo tags and update info with artist scan
Also adds migration to remove old format media info
* This file no longer exists
* Don't penalise year if missing from tags
* Formatting improvements
* Use correct system newline
* Switch to the netstandard2.0 library to support net 461
* TagLib.File is IDisposable so should be in a using
* Improve filename matching and add tests
* Neater logging of parsed tags
* Fix disk scan tests for new media info update
* Fix quality detection source
* Fix Inexact Artist/Album match
* Add button to clear track mapping
* Fix warning
* Pacify eslint
* Use \ not /
* Fix UI updates
* Fix media covers
Prevent localizing URL propaging back to the metadata object
* Reduce database overhead broadcasting UI updates
* Relax timings a bit to make test pass
* Remove irrelevant tests
* Test framework for identification service
* Fix PreferMissingToBadMatch test case
* Make fingerprinting more robust
* More logging
* Penalize unknown media format and country
* Prefer USA to UK
* Allow Data CD
* Fix exception if fingerprinting fails for all files
* Fix tests
* Fix NRE
* Allow apostrophes and remove accents in filename aggregation
* Address codacy issues
* Cope with old versions of fpcalc and suggest upgrade
* fpcalc health check passes if fingerprinting disabled
* Get the Artist meta with the artist
* Fix the mapper so that lazy loaded lists will be populated on Join
And therefore we can join TrackFiles on Tracks by default and avoid an
extra query
* Rename subtitle -> lyric
* Tidy up MediaInfoFormatter
New: New releases that fail will be retried a second time after waiting 1hr (configurable)
Fixed: Blacklisting releases with the same date and vastly different ages