Classical works detection broken for arrangements of other composers

[krefson]

The current algorithm for detection of classical "Works" insists that the composer field be identical for all movements in a work. This is not always the case, and it fails in quite a few cases, the most prominent being Mozart's first four piano concertos, which are arrangements of movements by other composers:

https://en.wikipedia.org/wiki/Piano_Concertos_Nos.1%E2%80%934(Mozart)

in which case any attempt to tag any of the complete a complete cycles of these always fails. For example the MusicBrainz release Murray Perahia playz Mozart, which has been mis-entered in MusicBrainz with multiple composer relationships including "Wolfgang Amadeus Mozart".

Ideally the audio files of these recordings ought to be tagged with COMPOSER="Hermann Raupach", for example and ARRANGER="Wolfgang Amadeus Mozart".

Another example is the F. A. E. sonata

I suggest that the "same composer" criterion be relaxed in two respects:

1. If there are multiple composers listed, consecutive movements with the same composer at any position in the list should be considered part of the same work. This will allow tagging the work with a COMPOSER list containing the arranger.
2. I suggest that tag ARRANGER be considered as equivalent to COMPOSER for the purposes of works detection only. This would further allow the tagging to better reflect reality.