#metabrainz

[musicbrainz-server] 14reosarevok opened pull request #2812 (03master…MBS-10610): Mbs 10610 https://github.com/metabrainz/musicbrainz-serve...

... good job remembering to set the title, me

[musicbrainz-server] 14reosarevok opened pull request #2813 (03master…MBS-12847): MBS-12847: Allow searching for edits without any notes https://github.com/metabrainz/musicbrainz-serve...

[bookbrainz-site] 14MonkeyDo merged pull request #921 (03master…dependabot/npm_and_yarn/cookiejar-2.1.4): chore(deps): bump cookiejar from 2.1.3 to 2.1.4 https://github.com/metabrainz/bookbrainz-site/p...

discourse is down, upgrade didn't go well, on it

[bookbrainz-site] 14MonkeyDo merged pull request #920 (03master…dependabot/npm_and_yarn/superagent-8.0.8): chore(deps): bump superagent from 7.1.1 to 8.0.8 https://github.com/metabrainz/bookbrainz-site/p...

[bookbrainz-site] 14MonkeyDo merged pull request #916 (03master…dependabot/npm_and_yarn/classnames-2.3.2): chore(deps): bump classnames from 2.3.1 to 2.3.2 https://github.com/metabrainz/bookbrainz-site/p...

[bookbrainz-site] 14MonkeyDo merged pull request #917 (03master…dependabot/npm_and_yarn/babel/preset-react-7.18.6): chore(deps-dev): bump @babel/preset-react from 7.16.7 to 7.18.6 https://github.com/metabrainz/bookbrainz-site/p...

[bookbrainz-site] 14MonkeyDo merged pull request #918 (03master…dependabot/npm_and_yarn/swagger-ui-express-4.6.0): chore(deps): bump swagger-ui-express from 4.3.0 to 4.6.0 https://github.com/metabrainz/bookbrainz-site/p...

still on discourse upgrade, it resumed, there's an upgrade from pg12 to pg13 that went bad

bitmap, yvanzo: I think I figured out why my first instinct is to just blank the PR template: the entirety of the content that shows when a PR loads is part of the long introductory welcome to MBS comment

So it's like "newbie stuff, go away" :D

And it's a lot easier to clear all than to carefully remove only the long comment

[musicbrainz-server] 14reosarevok opened pull request #2814 (03master…MBS-12839): MBS-12839: Allow searching for edits by kind of edit type https://github.com/metabrainz/musicbrainz-serve...

reosarevok: that's an interesting point. I see that you can have mutiple templates, maybe there's a case for a "beginners" one (the default), and a "reo" one

reidel1!Y

Freso: around? the discourse upgrade is everything but smooth, I may need your help

the migration was stuck in between, mostly because of following

I disable this plugin in the web.yml config and it seems it resumed, but I guess we'll have something to fix regarding oauth

it is https://github.com/metabrainz/discourse-musicbr...

I'll finish the app upgrade, then re-enable this plugin to see (but I suspect the file will still missing)

[listenbrainz-server] 14MonkeyDo opened pull request #2344 (03master…combine-playlists-collaborations): Combine playlists and collaborations tabs https://github.com/metabrainz/listenbrainz-serv...

lucifer, alastairp: I just noticed https://github.com/metabrainz/mb2wikidatabot/pu... - what's with that?

reosarevok: good question. I guess we started and never got around to fixes?

the changes look pretty small, probably something I can fix this week

Thanks

mayhem: looking at that Wikidata survey. "Please estimate the number of people using your tool / application"

zas: ^ maybe you're a good option for this too :)

I can list number of editors but I guess we would kinda care more about number of data users, which of course we don't know, but do we have a way to get a general ballpark?

I guess maybe it's on zas' summit presentation

how do you define 'data users' ?

No clue :D

But I can say we get 10 billion queries per year

(from your presentation)

yes

But that we have no way to estimate how many of them interact with any Wikidata-related stuff

I think that'll be good enough tbh

this discourse upgrade is an unexpected mess...

making progress though

Is this the 3.0 one?

yup

reosarevok: I normally say: 1 going on 20k.

I'm not sure what that means :D

One developer and 20k yearly users.

Oh

Let them sort it out for their own use.

Good thing we're not doctors

Anyway, sent the survey :)

It was actually interesting, unsurprisingly they have a very clear idea of what is a mess in their DB, I'd bet they have a less clear idea of how to fix it

#relatable

I had to fix duplicated keys manually in the database, and I disabled oauth plugin for now, after fighting to upgrade from pg12 to pg13. I may need help to fix the oauth thing though.

So it's mostly "this is a list of things we know are a mess, how much does each of them affect your org?"

alastairp: unless something has changed, multiple PR templates seem to be a mess: https://stackoverflow.com/questions/73771068/mu...

Anyone with OAuth + Ruby fu around?

We need to update this plugin, it uses a deprecated feature it seems

zas: I could probably take a look. My Ruby fu is a bit rusty, but at least I have a discourse development setup already running and some experience with its plugin system. How urgent is it? I could take a look later this week

btw, the metabrainz/discourse-musicbrainz-auth repo was archived few months ago, but it is clearly used by our discourse instance (proof: it blocks the upgrade), I unarchived it.

outsidecontext: well, rather urgent urgent, if we can fix it now, I don't have to revert the upgrade, but to be frank, I'm not sure this is possible, I have no idea which changes it needs to work again

zas: I see. Ok, let me power up the discourse thing. I see the new method is to use omniauth. I used this once. Let me check this and I'll get back to you

k thanks!

you know that when you created a local dev user for some system with these "easy to remember credentials", but a few weeks later you don't and none of the obvious choices works?

oh yes... ;)

i'm a rubyist, just had a look and it seems involved

yup, if we can't find a solution very quickly, I'll have to roll back the upgrade. So any help is welcome.

honestly I don't think this is something we test on prod

we need to write a new clas that inherits from Auth::Authenticator

*class

This change doesn't stand out from changelog, because the actual removal happened in intermediary version (2.9.x) and suggested upgrade was from 2.8.14 to 3.0.0.

this indicates a migration would be needed due to a change in SQL structure https://meta.discourse.org/t/adding-a-new-manag...

I guess it is safer to roll back for now

I'll start the process, and we'll postponed upgrade to 3.x until this is solved. We also need a way to actually test those upgrades.

am I doing something wrong here? I am using strawberry v1.0.9 on void linux, it shows that I am logged in to listenbrainz, and I have had 0 tracks scrobbled

I have been logged in for at least a month and have listened to plenty of music

hmm. what is your username on LB DeepT ?

zas: I think I get this to run today, have the plugin now loading here locally, currently updating it

"pheni". I just found my issue, it was a PEBKAC https://github.com/strawberrymusicplayer/strawb... , but now im getting errors

specifically <DeepT> "pheni". I just found my issue, it was a PEBKAC https://github.com/strawberrymusicplayer/strawb... , but

shoot sorry bad paste

Scrobbler Libre.fm error: Error transferring https://libre.fm/2.0/ - server replied: (499)

oh thats for libre. i need to wake up and stop clogging the chat with nonsense

:P

seemed to have done the trick. :)

ya sorry for the bother

grrrr, trying to revert discourse back to v2.8.14 is a pain too ... they upgraded to ruby 3.1 and apparently previous stable versions don't work with it. So I have to downgrade discourse_docker (at our own risk). Setting version to v2.8.14 (as documented) doesn't work, I have to use commit id instead. Everything takes ages. Lesson: we need to test future upgrades, even between stable versions.

CatQuest: hello there, before I run off to sleep, would you like to discuss the IRC documentation updates?

You can check out the complete thing here: https://prathamesh-g.notion.site/IRC-Documentat...

akshaaatt: that ui test bug is solved we can begin now :)

here's the thread that solved it

zas: I'm a bit thrown back here as well :( there is an open issue for newer omniauth oauth2 that affects the MB oauth setup: https://github.com/omniauth/omniauth-oauth2/iss... . there is a workaround, but needs more changes

on my side, reverting is also a pure mess, it looks like we are stuck in the middle of nowhere

^ping alastairp for the docs (If you're free that is :D)

Just tried to visit https://community.metabrainz.org/t/plug-ins-not... and I get a "site down" message.

Sophist-UK: yes, forums are down due to a failing upgrade, we are on it (since hours)

Ok - thx.

These things happen.

Just wanted to make sure someone knew.

!m zas

You're doing good work, zas!

mayhem: Unfortunately if I did that for every time zas did good work, BrainzBot would likely have a nervous breakdown from having to repeat it so often.

Hi, community.metabrainz.org seems to be down right now. Is it a known issue, or is it just a problem on my end?

it's a known issue, the admins are working on it

elomatreb[m]: Thank you. Hopefully it's nothing too serious.

Sophist-UK: true dat!

I'm running out of options with discourse, something broke and starting from a fresh install using backups doesn't look an option before we can fix oauth plugin. I tried to restore previous version, and it fails to start properly even when using pg data backup (and the process is hacky, because previous stable is somehow now unsupported (it was out 3 weeks ago) because docker_discourse moved to ruby 3.1 (and previous stable runs on 2.7).

I managed to use previous docker image (it requires changes to launcher script which assumed stable=latest stable discourse repo)

but even after full rebuild of all containers, something is missing

each rebuild or pg data copy takes a long time, which slows down attempts

I'm trying yet another rebuild right now

but the real solution lies in upgrading the oauth plugin and complete the move to 3.0

why not skip the oauth plugin, bring the site back up and let us have moment to think about it?

not sure people who have accounts with oauth will be able to log in. maybe the accounts can be recovered via email?

well, may be possible