Notes from #MetaBrainz Meeting 2025-12-01

[aerozol]

Meeting start: IRC Logs for #metabrainz | MetaBrainz Chatlogs
Meeting wiki page: MetaBrainz Meeting - MusicBrainz Wiki

Agenda

• Reviews
• LLM code contribution guidelines

Last meeting: 2025-11-24
Next meeting: 2025-12-08

@reosarevok:
Hi, and welcome to another MetaBrainz Monday Meeting!

Reviews

@hemang-mishra

[mailed in] [Hemang Mishra] Nothing much to share this week. Mostly was busy with my end semester exams.
Went through a few queries from new contributors.
Now finally holidays have started. So traveling back home…

@m.amanullah7

[mailed in] [mamanullah7] Hello everyone, nothing much from my side last week, my end sems exams are going on and tomorrow i’ve my last theory paper.
Last week i’ve tested the latest brainzplayer issue [#3390] and reported multiple datasource playing songs at same time to monkey and i can see that issue is fixed [#3414]
i’ll test this and some other previous issue if it solved or still exist again when i’ve some moments.
added funkwhale and navidorome local setup documentation, so new contributors don’t find any difficulties to set it up locally!
[LB-1849, LB-1203, LB-1869: Improve BrainzPlayer activation strategy and failure recovery #3390]
[BrainzPlayer: fix double playback issue #3414]

@_Junaid

[mailed in] [fettuccinae] Last week I was a bit busy with exams, so I couldn’t do much. Worked on PR feedback and did some more local testing for the notifications branch.
So, that’s a lot of exams! Hope they went / are going well, good luck to everyone!

@rob

[mailed in] [mayhem] Still feeling crap, but myself yet more antibiotics, so hopefully should be better soon. I only did the most essential MeB tasks in the last week. Ya está.

@ApeKattQuest_MonkeyPython

[ApeKattQuest] last week I:

• majorly fucked up my extHD AGAIN so had to (and continue to deal with) that
• tagged a lot of releases an did editing!
• still waiting for things to arrive!
• ate food!

This week I’ll:

• continue work on instruments
• pack
• relax a bit, eat good food
• fret because Soon™ I’ll be travelling (my most ambitious endeavour yet!)
• wait for stuff I ordered to arrive so I can tag more releases

Next week is my birthday! and after that I’ll be afk tidying and packing and such, probably won’t be doing any more work this year
however! if anyone wants me to test anything, or has like a really well researched instrument ping me and i’ll deal with it ASAP!

@JadedBlueEyes

[Jade] Finishing up the final course work and stuff
I also did a little bit of work on a plug-in for Picard that allows looking up a release in harmony, although that needs some tweaking to unhardcode some stuff
The mail service also now has Dutch translations!
And my personal website/self-hosting repository has reached 750 commits since I created it almost exactly 2 years ago now

@zas

Last week, I fixed few issues with bot-detector, tuned it a bit, handled few unblock requests
And I also started to tackle with the Picard plugins V3 task: wrote specs, documentation, implemented the CLI, and registry maintenance tool, I’m waiting for phw/outsidecontext to show up so we can decide few things
kudos to rdswift for testing
the current branch has more than 450 commits…
but the plugin system is really taking shape now
I wrote a migration script that will help devs to convert their plugins from V2 to V3
I also spend a bit of time on an issue we have with SSL/HTTP2 on gateways, still investigating

@julian45

spent most of last week doing family related things (drove brother and myself a few hundred miles to visit parents, stayed with them a few days, drove a few hundred miles back last night)
but on the side i tweaked the anubis rules in front of our blog to better handle rss reader tools
and continued thinking about better & more efficient ways to manage our server fleet

@reosarevok

This week I worked on MBS-14178
[MBS-14178: Indicate the edited release on recording and RG edits entered from the release editor]
And worked some more on the label React conversion
Plus a bunch of small PRs, code review, etc

@bitmap

last week I was spending time with family, but did manage to finish the code for syncing metabrainz accounts to the MB editor table
also worked on preserving POST parameters through the oauth login flow, building an in-page login UI to handle cases where the session expires, and documenting some minor issues on the metabrainz side
I helped a tiny bit with the possible-duplicates fetching code for MBS-12758 too
[MBS-12758: Convert label edit form to React]

@anshgoyal31

[ansh] Last week, I picked up some tickets for ListenBrainz
And reviewed couple of PRs
And continued working on fixing some failing tests on the MB Homepage

@lucifer

last week, i worked on adding a rejection queue to timescale writer so that listens that fail to insert for any reason do not block rest of the queue.
also, worked on fixing a bug in incremental stats but that turned out to be a dead end. need to rethink how to address it.
some minor fixes for misc LB tasks.

@mr_monkey

[monkey] Last week I reviewed PRs and looked into some security tickets for BookBrainz, updating inter-connected dependencies
On the LB side I continued fixing issues with brainzPlayer, and reviewing improvements to the history importers
I also hotfixed an error that took the LB ingestion pipeline down
And finally, worked on the new YIM25 design implementation

@kellnerd

I finally got back to writing code myself for Harmony this weekend.
So in addition to the new mora.jp provider by serial_ata, today’s release also includes a few minor bug fixes for Tidal and Bandcamp.
And I’ve finally created official documentation for the accepted Release Lookup parameters.

LLM code contribution guidelines

@reosarevok

Through last year’s GSOC, we discussed our views on using LLMs in the context of GSOC, leading to defining this policy:

We discourage the use of AI tools in communications and contributions.
The goal of GSoC is for you to learn and improve your skills, not just to finish your project at any cost.
However, if you do use any AI or LLMs, you must disclose their use in your PRs and/or conversations with your mentor. Additionally, you must fully understand and be able to explain any AI-generated code you submit.
Any undisclosed use of AI or Large Language Models, or inability to explain AI-generated contributions, will be grounds for instant disqualification.

(From Development/Summer of Code/Getting started - MusicBrainz Wiki )

This discussion topic is in two parts: 1. reviewing/confirming/amending the above policy and 2. a proposal to add a similar policy to the MeB github contribution guidelines.
We currently have no such general policy, nor any mention anywhere to what degree we are comfortable with LLM-enhanced code contributions.

The topic was recently being discussed in the GSOC mentors mailing list, and I too found myself this week thinking about the implications. Other OSS projects have come up with policies that could provide a helpful base, if only for discussion (for example Zulip: zulip/CONTRIBUTING.md at main · zulip/zulip · GitHub )

I find that many of us on the team are on the fence regarding LLM-coding, neither convinced nor entirely rejecting modernity (return to monke), but all of us feel some level of unease and mistrust that is hard to put into words. Along with a more direct policy wording, I would like to propose something along those lines to explain our point of view, which I adapted from a recent conversation I had with a contributor:

If you are going to use an LLM as part of your workflow, all of understanding the issue and problem solving should be done without using LLMs, as they are not thinking models and often get the intention wrong for lack of context and understanding/reasoning. This is where we humans shine, and this is the core of our work as developers.

The job market definitely does not agree with us, and will allow/encourage/require you to use AI for increased productivity.
However this little open-source bubble is rather humanistic in its philosophy, and prioritizes community, critical thinking and robust and long-lasting code over productivity and speed, so our priorities are very different from those of a financially motivated team or project.

I’m not asking for a decision to be taken right now, but if you have strong opinions on the topic, feel free to share them. For the next step, I will be editing a more suitable proposal this week in a GDoc, and requesting feedback from the whole team in a more asynchronous manner. That’s it from me, thank you for coming to my ted talk.

A little bit of chatter (see the logs) but no big disagreements, monkey says he will come back with a concrete proposal next week.

absent: outsidecontext (2), jase (1), aerozol (sick)

[Weblate]

@zas

the current branch has more than 450 commits…

Jesus Chrysler Superstar!
It will be a perfect, final, ideal work.