#metabrainz

🤔

or "sign in with google" i guess that works?

seems so! whoo

_lucifer: you signed up for this: should i "transger to metabrainz foundation" or "keep separate from metabrainz foundation" ?

CatQuest, i don't remember dropbox asking me that.

hm

I don't want to fuck it up, ruaok?

Dunno.

ah it was because i didnt have a thing

Project bookbrainz-site build #3770: passed in 4 min 16 sec: https://travis-ci.org/bookbrainz/bookbrainz-sit...

[bookbrainz-site] MonkeyDo merged pull request #579 (master…dependabot/npm_and_yarn/ssri-8.0.1): chore(deps): bump ssri from 8.0.0 to 8.0.1 https://github.com/bookbrainz/bookbrainz-site/p...

Project bookbrainz-site build #3771: passed in 4 min 35 sec: https://travis-ci.org/bookbrainz/bookbrainz-sit...

[bookbrainz-site] MonkeyDo closed pull request #580 (master…fix/entity): Refactored footer of entity display component https://github.com/bookbrainz/bookbrainz-site/p...

Spotify and Youtube - Access Token + Refresh Token

Deezer - Non-expiring Access Token

Apple Music - Expiring Access Token, refresh manually (need to confirm this one, Mr_Monkey found something in Apple docs which suggest that the usual refresh token process might be possible)

Mr_Monkey, ruaok, alastairp: I think we can use this schema for storing user tokens to external services.

so you'd put specific implementation details (token, refresh token, etc) in auth_details?

yes because they differ across most services

if the user opts for only playback say using deezer or apple music, we probably do not need to store their access tokens in the database so that column is optional

I don't see that optional column; am i missing something?

use of refresh tokens and permissions also varies across services

ah, you mean the whole auth_details is optional

for a service with no expiring token, would `token_expires` be null?

deezer uses 0 for that

we could probably use the same?

[musicbrainz-server] reosarevok opened pull request #2010 (master…MBS-10726): MBS-10726: Use "medium", not "disc", as a generic term in RE buttons https://github.com/metabrainz/musicbrainz-serve...

now that I look at it again, even `token_expires` can be moved into `auth_details`. I am not sure whether its better to have a column allowing NULL values or move it into `JSONB`. thoughts?

the same logic can be applied to `access_token` and `refresh_token` as to whether they should have their own columns with NULL allowed or kept in JSONB.

I'm ok with `access_token` and `refresh_token` being in JSONB.

[bookbrainz-site] akashgp09 opened pull request #581 (master…overlapping-button): FIX(BB-593): Overlapping buttons https://github.com/bookbrainz/bookbrainz-site/p...

reosarevok, yvanzo: is there a good reason we don't store the email until it's verified? seems like it'd make the system less confusing

Hi bitmap, I don’t see any but then we should check permissions always depend on having verified email, not just on having email.

zas: looks like I can't be around for the meeting today :(

outsidecontext: can you send a mail to freso @ meb to summarize weblate advantages over transifex? At least it will give people matter to think about, and we can wait you're around to discuss this in details

alastairp, around?

hello!

are you available to debug the jenkins issue?

in ~30 minutes?

works for me 👍

zas: yes, will do

outsidecontext: thanks for that. is this related to the forum thread?

yes

I think this might have been started by me making a mistake with deploying jesus' CB translations

I'll have a quick look at the thread later today and see if I can make another release

alastairp: we wanted to try weblate for Picard as a pilot project, basically.

zas: if it's ok let's move this discussion to next week, dealing with some work trouble here right now and won't be able to get something written down. I can prepare something for the next meeting

ok, np

zas: would it be feasible to host our own instance of weblate?

bitmap: I dunno, tbh. I expected it was a conscious decision not to store data that might be made up or whatever

But it might just be a historical accident

I did test weblate with docker last year but did not complete using MB OAuth with it.

_lucifer: sorry, time got away on me

what's up

hi!

honestly the faster we move ovv from transifex the better

I looking on the pending PRs and thought to discuss and move ahead the Jenkins image cleanup one. Last time, we were discussing filtering on name so that different projects don't clean each others images.

yes, right

(we might have to postpone this discussion during meeting time, are you planning on finishing for the day after meeting or can you stay around?)

I dunno, transifex isn't that awful

But I'm not against a change either

i can stay around for another hour after the meeting

great

see this also: https://www.jenkins.io/doc/book/pipeline/jenkin...

we use $BUILD_TAG in our test setup

COMPOSE_PROJECT_NAME_ORIGINAL="listenbrainzunittest_jenkinsbuild_${BUILD_TAG}"

but this turns into 'listenbrainzunittest_jenkinsbuild_jenkis_[job jame]-1238'

which is huge

yeah, the job name seems redundant.

I think we should use BUILD_TAG, then. we might just be able to do COMPOSE_PROJECT_NAME_ORIGINAL=${BUILD_TAG}, because I think it contains all of the information

I bring it up just to clean up the names of the containers/images

makes sense.

so now going back to the filter, maybe we should just filter on label _and_ project name

we also have some none:none in the output when we list images.

do you have a `docker image ls` command that reproduces this?

yes, i also have a build on ci showing that. let me find that

reosarevok: It’s not that awful indeed but it has a few hiccups and doesn’t allow for MB auth.

this is the --format {{.Repository}}:{{.Tag}} bit, so it strikes me that there's an image layer around which matches the label filter, but doesn't have a tag for some reason

yvanzo: sure, not saying we can't improve on it, just wanna make sure it's an improvement :)

filtering based on the name will solve this, but it also means that we might leave it behind and not pick it up with the filter

yeah

docker images --filter "label=org.label-schema.name=ListenBrainz"

see this (ssh into paco and run it manually)

see the last item. I'm not sure why it's there. it doesn't have the same image id as anything else, which is weird

reosarevok: right, it's not sure because it depends on how features actually work, so we have to try.

yeah right

<BANG>

It’s World Water Monday!

People up for reviews: yvanzo, bitmap, zas, alastairp, Mr_Monkey, Freso, reosarevok, _lucifer, ruaok, shivam-kapila, CatQuest – anyone else who want to give review, let me know ASAP!

I haven’t received any mailed in reviews, so…

yvanzo: Go!

I'm going to add to this PR right now a filter to include the repository too

Hi all!

we can continue after meeting

👍

Sorry alastairp and _lucifer for interrupting ;)

Last week I helped reosarevok with scheduling schema changes.

Also fixed a regression I caused in annotation formatting.

Plus reviewed some PRs and that’s all.

Go reosarevok?

Hi!

Last week I spent quite a bit of time putting together schema change tickets and preparing a blog post for them (with help from the team, thanks all)

I also fixed a bunch of ISEs that slipped past testing / review from a huge bitmap PR and hotfixed them

Hi folks!

And started working on some smaller schema change tickets

Hi c1e0!!

Plus the usual triaging

And fixing small old stuff

I plan to do the same this week, plus hopefully some hiking and birdwatching

Since I need to take some days half-off or I'll go over my hour limit, lol

CatQuest: what about you? :)

whatevs.

hi

Meow! `😺`

🎉 INST-977 🎊

INST-977: Improve ektara/gopichan https://tickets.metabrainz.org/browse/INST-977

Last week I celebrated spring equinox!