Šarūnas Nejus

> > Is there a reason we're using `BLOB_TYPE = memoryview` instead of `bytes`? I tested querying using `bytes` and I see to get the expected results > > https://github.com/beetbox/beets/blob/92fb830559530a4913b6966059c053fdd1e46bcf/beets/library.py#L41-L44...

> Maybe this is intentional? > > Is it expected behaviour that when `chroma` is used it fetches from MusicBrainz anyway? (Because otherwise chroma configured with a disabled musicbrainz would...

Tested a couple of variations of album and singleton import and using `chroma` - now I remember the intentions behind that change. ```yaml musicbrainz: enabled: no ``` disables the data...

I see what you mean - I can definitely imagine a use case where I want to reimport/re-tag some of my music if I start using a new data source....

> Neato!! This is very exciting; I'm really glad we (as a project) are finally getting around to getting serious about making flexattr queries fast. When we built flexattr support...

@sampsyo see the new commits: aggregation ended up significantly simplifying the machinery behind the new queries - this now uses just one query. We can now query multiple flex attributes!...

@sampsyo I apologize - I used completely wrong examples in my previous comment 🤦🏽‍♂️ see the clarification below > The drawback is that we're losing the ability to query related...

Just had a couple of runs comparing `master` to this branch with various queries, see the results below  * Previously, the more filters were provided on the command line,...

@sampsyo would you still want this to be merged? I could try and rebase it off master

Superseded by #5240