#metabrainz

ok, so then we would clone the repo and add a submodule to our pyproject?

yes

ok.

but uh actually

this is a c++ project with python bindings

indeed.

hmm i am unsure if this would work then even with submodules

could we upload a locally built wheel with a patched nmslib to pypi?

I have no idea. I wonder if gcrk (he/him) would know.

for instance, https://pypi.org/project/lb-matching-tools/#files

if we could have a branch where nmslib dep points to the github repo and then we could build wheels on that branch and publish to pypi.

and publish the source distribution using the normal branch.

it might work, not very desirable ofc.

lets see what gcrk (he/him) says about this tomorrow and then take it from there.

alternatively, we could remove nmslib from requirements. and ask users to install it manually.

sure

lucifer (IRC): do you happen to know where in the LB docs we document command line options, so I can copy that?

daaaaaamn, those pictures.

https://github.com/metabrainz/listenbrainz-serv... and https://github.com/metabrainz/listenbrainz-serv...

thx

ok, that's done too. PR updated, lucifer. lets look at packaging tomorrow.

mayhem[m]: sounds good

aerozol: would you prefer the sharebrainz idea to be on LB or MB or MeB page?

lucifer: I think it has more to do with MusicBrainz, even though it's aimed at a more casual audience (which makes us think 'LB').

bitmap: Is there a bug or is it correct? “This doesn't seem to resolve any bug, since the old IDs were still deleted correctly by `_delete_and_redirect_gids`.” It seems that the first commit is just a refactoring.

MB project (with a prominent link to LB in the page itself)

ahan cool, makes sense.

yvanzo[m]: IIRC it was just refactoring, but I don't remember the context (it might've been to simplify something in a later commit)

aerozol: Does my comment to your ticket makes sense? Should we elaborate a GSoC idea from that?

bitmap: The citation is from the commit message.

yes

if you're asking if the commit message is correct, then it looks correct

Ok, I just could not connect the two parts of the sentence.

But the code change itself is good.

I think because I initially thought there would be a bug due to this line in sub merge: $self->_delete_and_redirect_gids('artist', $new_id, @$old_ids);

[listenbrainz-server] 14amCap1712 opened pull request #2759 (03master…fixes): Fix db call on artist_entity page https://github.com/metabrainz/listenbrainz-serv...

[listenbrainz-server] 14amCap1712 opened pull request #2760 (03master…fix-dump-imports): Fix dump imports https://github.com/metabrainz/listenbrainz-serv...

er, sorry, that's the working line

I thought I was fixing a bug there, because the old ids wouldn't be passed there correctly (they would be in an array ref)

which is correct, but also (explicitly) delete them from the cache via $self->_delete_and_redirect_gids('artist', $new_id, @$old_ids);

lucifer (IRC): are you on this already with the latest PRs? https://community.metabrainz.org/t/unable-to-lo...

verified, btw.

so the bug is accidentally worked around because the "old IDs were still deleted" by that line

mayhem: oh :/, can you please revert lb prod to v-2024-01-31.0 ?

on it.

I just logged off 5 mins ago

Thanks!

k

Of course was one thing I didn't test and it had to break.

*login was

hahahahahaha, of course.

I didn't see it either.

fixed.

thanks

[musicbrainz-server] 14mwiencek opened pull request #3165 (03master…flushsync-type): Update react-dom libdef with flushSync type https://github.com/metabrainz/musicbrainz-serve...

hello

https://listenbrainz.org/user/TreatnHerRight/ spotted this in the feed

seems like someone's scrobbler is bugging out in a big way

that's all I have to show. bye for now