#metabrainz

/

      • pite has quit
      • pite joined the channel
      • xSkkarf[m] has quit
      • minimal has quit
      • BrainzGit
        [musicbrainz-server] 14mwiencek opened pull request #3500 (03production…unnecessary-genre-loading): Remove some unnecessary code from `Controller::Role::Tag` https://github.com/metabrainz/musicbrainz-serve...
      • d4rkie has quit
      • d4rkie joined the channel
      • vardhan_ joined the channel
      • d4rk-ph0enix has quit
      • d4rk-ph0enix joined the channel
      • lucifer[m]
        bitmap: i see. also, recording core completed fine for me.
      • BrainzGit
        [listenbrainz-server] 14amCap1712 opened pull request #3231 (03master…popularity): Fix popularity data generation - II https://github.com/metabrainz/listenbrainz-serv...
      • d4rkie has quit
      • d4rkie joined the channel
      • pite has quit
      • d4rk-ph0enix has quit
      • d4rk-ph0enix joined the channel
      • Kladky joined the channel
      • d4rkie has quit
      • TOPIC: MetaBrainz Community and Development channel | MusicBrainz non-development: #musicbrainz | BookBrainz: #bookbrainz | Channel is logged and not empty as it is bridged to IRC; see https://musicbrainz.org/doc/ChatBrainz for details | Agenda: Reviews, defending our sites (zas)
      • d4rkie joined the channel
      • d4rk-ph0enix has quit
      • d4rk-ph0enix joined the channel
      • lucifer[m]
        bitmap: also, i fixed the issue you were facing on macOS in the latest changes today. it happens only on macos and windows because multiprocessing using spawn there instead of fork to create new processes. but python intends to change the default for linux to spawn as well starting in 3.14. i have added point fixes but i would prefer to restructure some of sir code later to do this properly. once the initial python migration is done.
      • also, mayhem ^ all of the multiprocessing code that we have in LB or otherwise might break when we migrate to 3.14 and that part is usually not tested so we'll have to do it carefully.
      • reosarevok[m]
        Sounds like you have time in advance to schedule some test writing! 😜
      • lucifer[m]
        or fix it when it breaks in prod ;)
      • its not exactly easy to write tests for, because you have to start multiple processes in tests and then make sure they actually test what you want to test. because some issues don't show up the small test db.
      • but we can probably find a middle ground and add some tests where possible.
      • aerozol[m]
        LB team, if someone has time to look at this bug report via Mastodon: mastodon.social/@leodurruti@puntarella.party/11...
      • lucifer[m]
        aerozol: hmm i see what the issue is but it will take some time to resolve.
      • aerozol[m]
        Oh, nice. Can I help with making a ticket or anything?
      • lucifer[m]
        sure you can just make a ticket with details and assign it to me
      • reosarevok[m]
        bitmap: oh, meh, but if we use child order for `artist_release_group` then we should probably rename the `primary_type` and `secondary_types` columns? Or should we store the type ids as well? Are those used somewhere?
      • lucifer[m]
        reosarevok: do you have access to jenkins admin?
      • i think i forgot my password again
      • reosarevok[m]
        Seems so (I had no idea)
      • lucifer[m]
        alternatively if you have regular access and could fix a line in sir's test config that would work too for now.
      • reosarevok[m]
        Solved :)
      • lucifer[m]
        yup, thanks again!
      • aerozol[m]
        lucifer: done, I will let the user know: https://tickets.metabrainz.org/browse/LB-1767
      • BrainzBot
        LB-1767: Incorrect listen > stat matches
      • mayhem[m]
        <lucifer[m]> "also, mayhem ^ all of the..." <- What multiprocessing do we have besides fast_fuzzy?
      • lucifer[m]
        mbid mapper and some of the metadata caches iirc.
      • although those might be just thread pools and not process pools now that i think of it.
      • mayhem[m]
        Exactly
      • And fast_fuzzy is a dead end currently. Python is too limited to be fast enough
      • lucifer[m]
        oh :/