#metabrainz

Good morning. Got a mail reagarding org application deadline(19 feb) so putting it here :)

mo’’in’

bitmap: CAA-indexer 'master' branch has an unreleased/undeployed patch (about S3), so I’m not sure it is okay to merge a new PR onto it, override the current image and deploy. Your review would be welcome too, see comments.

bitmap: created https://tickets.metabrainz.org/browse/MBH-550 :)

MBH-550: Upgrade pg_amqp to set the correct maximum number of channels

good morning!

mooooin!

nearly!

I have the editor open as we speak

1 week and 2 days behind, not too bad for me

for once that wasn't a prod for action, alastairp. :)

done! https://wiki.musicbrainz.org/Development/Summer...

thanks for the stats review ruaok, lots of good small things to improve

yay, very believable. thanks!

ruaok: I was looking at the PR that you opened for metadata on playlists

what was the motivation for introducing the query parameter to skip lookup? Is this something that you absolutely want to make available on the public endpoint?

main motivation was for making it easier to write tests, but as I thought about it, I can see use cases where someone might skip the metadata lookup step.

if we think that this is useful, then that's fine. however I'm not sure it makes the tests any easier to understand, we're just just swapping "remember to mock this method on a test that uses get()" with "remember to pass in metadata=false on a test that uses get"

and theoretically we should have a test that includes the metadata test anyway (mocked) to ensure that this code flow works. I think this is what you mentioned about delaying until we stop using this webservice lookup in this codeflow?

it is certainly easier than writing a freaking mock. tests are now 2 lines shorter, which feels like a win to me. (and shorter simpler tests is something that you keep emphasising)

mas o menos. given that we need to go back and clean up the service layer, if I were to write a test right now, we'd toss that test out and have to replace it with something better later.

not enthused to write tests that didn't exist before and that won't live long.

sure

for me, I really can't see a difference between the mental effort required to remember to copy a @mock line from an existing test and remembering to add fetch_metadata="false" to the get

but the tests are still working, and we have new useful functionality, so that's fine for now

honestly, if we're going to quibble over sufficiently testing the playlist feature, there are large aspects that have zero tests right now. perhaps we should focus on those before getting into the weeds on the existing tests.

Have you considered using a testing library that stores HTTP requests and replays them, or is that not feasible?

Much easier than mocking IME

atj: we don't have one of those yet. it's an option, but would require integration of a new feature anyway

right now, I want to be doing less infrastructure improvements and get back to working on some actual user facing features. which I haven't done this year, so it seems

you mean you don't enjoy writing bash scripts ruaok?!

hate hate hate hate bash scripts.

I got that impression :)

Understandable

Project bookbrainz-site build #3647: passed in 4 min 15 sec: https://travis-ci.org/bookbrainz/bookbrainz-sit...

Moin!

moin, pristine___ !

I don't have admin on https://github.com/metabrainz/brainzutils-python/. Can someone with admin add the issue-link magic so that we can link to jira with BU-nn, or give me admin on the github repo?

ruaok: yvanzo: ^

you should be admin now

I am, thanks

_lucifer: careful, I might have to ask you to slow down on new PRs...

ruaok: do you still need 'ruaok' user for RabbitMQ?

yes, it is useful for cleaning out queues and whatnot.

we could create an 'admin' user in docker server configs instead if you prefer

sure, lets do that.

ok, will do

iliekcomputers: I'm considering removing messybrainz-server from livegrep as the repo is archived, the code is in the LB repo so search on msb-specific code should keep working. it keeps on throwing up false positives when searching for old code/dependencies

Picard has a proposed hindi project. Nice :)

alastairp: sounds good to me

zas: can we remove messybrainz-server repo from livegrep? It's been archived and the code moved into another repository

this is the file you want to add to

nice. I had a look in docker-server-configs but didn't find anything related

I don't have access to that repo. I can make the change if someone gives me access

invited

thanks!

wait

what

:D

i don't even remember that lol

what server does it live on?

zas will probably have to deploy

ok, will wait for him then, thanks

I'm around

so we merge, and deploy?

i merged it, just need to deploy

I see no docs in syswiki, is this something that I could have done if there were instructions?

alastairp: :DD

shivam-kapila: outsidecontext : the hindi translation of Picard is actually not a project we can accept. any accepted project HAS to include some coding.

we either remove it or couple it with something that requires coding.

alastairp: I updated https://github.com/metabrainz/livegrep-install/...

zas: cool, thanks

zas: should I add a page on syswiki with link to the repo and server login details?

yup

zas: what's the server?

alastairp: i'm not sure if you have access but livegrep.metabrainz.org would work for ssh, no?

I dont even have access. :)

#leastprivilege

:)

yvanzo: shall we pick a time for the release today?

ruaok: whenever you are free, can be now.

now is good.

let me update the tag on my image.

ok

do you have the upgrade steps documented somewhere?

I will compile notes for LB

thx

ruaok: ok, I'll remove it. I haven't edited the page yet, the project was there from last year. I wasn't involved in the discussion, was it acceptable back then?

naw, it should've never been there, but no big deal.

[brainzutils-python] alastair opened pull request #57 (master…docs): BU-36: Sphinx documentation https://github.com/metabrainz/brainzutils-pytho...

ruaok: if we can merge this, I'll use it to add a proper docs page for the stats PR

ok, I'll look at it when I'm idle during this release process.

yvanzo: I'm sanity checking your PRs and I've found that listenbrainz-websockets-{beta|prod|test} is also a consumer. can you please add that to the consumer PR?

I also found another producer. should I update the PRs instead? your call.

listenbrainz-websockets-{beta|prod|test} -> consumer and listenbrainz-api-compat-prod -> producer.

thanks, I will add them both

thank you!

ok, both look good.

ruaok: updated

approved.

so, just to confirm, the other containers are not using RabbitMQ? labs-api-prod, api-compat-nginx, cron, timescale, mbid-mapping, typesense, redis

confirmed.

ok, sounds good.

let me prepare.

ok, I’m ready

should be good to go in a few minutes.

Do you have a way to test messages go through? Like making a specific request to a test server?

(For sir-test, I just made edits.)

yes. once we merge the consumers PR, I can update the beta container, and submit a listen to the beta container for testing.

let me do that.

merged.

beta container deployed. working on submitting listen. encountering bitrot.

submitted.

submission successful: https://usercontent.irccloud-cdn.com/file/4l61H...