#metabrainz

[musicbrainz-server] 14mwiencek opened pull request #2739 (03master…fix-type-constraint-genre-error): Add another missing Types module import for Genre https://github.com/metabrainz/musicbrainz-serve...

[musicbrainz-server] 14reosarevok merged pull request #2739 (03master…fix-type-constraint-genre-error): Add another missing Types module import for Genre https://github.com/metabrainz/musicbrainz-serve...

yvanzo: I'd love to get https://github.com/metabrainz/musicbrainz-serve... finished and merged, since it's a minor change *and* a definite improvement. If you don't have time to do the same for Data::Editable, I'd happily merge this and send a second PR myself (or send a commit to yours, but I'm not sure I can push to your repo :) )

Thanks for working on refactoring and clarifying!

outsidecontext: did DNB ever answer?

O’Moin

yvanzo: Mooooin

reosarevok: no :(

Aww

yvanzo: given that, should we merge https://github.com/metabrainz/musicbrainz-serve... for now and improve it further if they end up answering? :/

reosarevok: I can try it with some of their other contact addresses. I had written to their metadata services, because it seemed fitting

And also, I merged bitmap's hopeful selenium fix into master, so yvanzo, you should be able to rebase on master to see whether other errors appear now :)

outsidecontext: that would be lovely

reosarevok: Thanks for having fixed the warnings!

[musicbrainz-server] 14reosarevok merged pull request #2738 (03master…fix-daologger-warning): Fix warnings in DAOLogger https://github.com/metabrainz/musicbrainz-serve...

No problem! Sorry, I never saw them just because I don't run DEBUG, or I would have fixed them the first time :)

bitmap: when you're back, please rebase https://github.com/metabrainz/musicbrainz-serve... and https://github.com/metabrainz/musicbrainz-serve... on master to test it again without the breakage, and https://github.com/metabrainz/musicbrainz-serve... while fixing the eslint issue - then maybe we can release them to beta? :)

reosarevok: Refactoring Perl code should be below our lowest priorities, I’d love to get your help with everything else, e.g. SEC-123, see the draft PR. Investigating the 5xx from this night right now. Will get back to PRs after that.

SEC-123: [musicbrainz-server] CVE-2019-20149: kind-of < 6.0.3 https://tickets.metabrainz.org/browse/SEC-123

Well, I'm mostly working on that at the moment because I have a ton of React PRs open and I don't want to keep adding more of those until we review and merge some :)

So you are opening a ton of Perl PRs instead.

Yes, those are a lot easier to review and merge whenever we have time :)

And reviewing stuff as often as I can, too

If you are out of inspiration, just ask, there are a tons of things that can be done without conflicting with current changes.

I'm not out of inspiration, I just care about improving old code

What help do you want with the SEC one? Should I be researching the CI errors or?

There is a blocking Ecmascript error.

I copied the log in the last commit.

(and don't worry, my plan with those Perl ones is to merge your stuff, then adapt my stuff accordingly)

(not make you deal with all the conflicts)

FWIW, I tried to use Moose with properly on controllers and it broke, so we're doing shit wrong somewhere, I'll figure that out eventually :)

Ok, I'll take a look at those errors, see what I can figure out

[listenbrainz-server] 14chinmaykunkikar opened pull request #2255 (03master…css-updates): css: Bump Bootstrap to 3.3.7 https://github.com/metabrainz/listenbrainz-serv...

yvanzo: the issue is because of this line in init.js: /^[\p{White_Space}\p{Punctuation}]+$/u.test(self.rootField.editNote()) ||

In theory that should be a solved issue? But I can't find where and how it's solved tbh

Alternatively, https://github.com/thematters/matters-web/pull/... is a possibility

reosarevok: I mentioned that RegExp set notation in the commit message, and took inspiration from https://github.com/babel/babel/issues/15193#iss...

reosarevok: However I don’t understand the first warning (maybe it is usual, I didn't check), I tried several alternative export/import notations unsuccessfully.

This one?

export 'DEVELOPMENT_SERVER' (imported as 'DEVELOPMENT_SERVER') was not found in './common/DBDefs-client.mjs' (possible exports: default)

I've been getting that locally for weeks

I expect https://github.com/mathiasbynens/regexpu-core/p... is meant to fix that issue?

yes this one warning

we don’t use the 'v' flag which this PR is about

Yes, but once it's easy we probably should, since it's a straight upgrade of u AFAICT

Anyway, left a comment on the PR

reosarevok: you can totally push commits to others’ pull requests unless they unticked "Allow edits by maintainers"

I just didn't want to push without you checking the solution, in case you weren't happy with it :)

But thanks

Would you be ok with me doing that wrt Data::Editable? :)

(I know it's not a big priority, but it's a small, clear improvement on crappily-named packages)

reosarevok: Thanks, answered on the PR.

For the rest, I’d prefer to have your input on MBS-12552.

MBS-12552: Define usage of the term “entity type” https://tickets.metabrainz.org/browse/MBS-12552

zas: hi! can you please apply this same config to beta.lb and test.lb? https://github.com/metabrainz/openresty-gateway...

Main seems fine to me

But I can also live with Central

I'll comment a bit later

moooin!

lucifer: how did things go last night? did you restart the -web container? Did similarity data get generated?

yes i updated the containers but didn't request new data because didn't know what params you wanted.

ok, I'll put in more requests in a minute.

👍

https://listenbrainz.org/art/ is still broken, because the URLs do not point to api.lb.org -- some magic *is* needed.

uh right, i just saw that as well. it worked on test.lb because gateways config was missing.

i'll fix it by not using url_for like we do in all other places in LB

when did we stop using url_for? I thought that was expected...

we don't use url_for in frontend because its in react.

and the api is mostly queried by react parts, the jinja2 parts don't usually query apis directly

ah, makes sense.

https://test.listenbrainz.org/art/ works for now and is currently release group cover arts.

weirdly that increases repetitions for akshaaatt's grid

at least the page works now!

yvanzo, bitmap: I'm going to update test with https://github.com/metabrainz/musicbrainz-serve... for kellnerd

Since test is already anyway loaded with an old relationship editors branch, it shouldn't be a problem

Thanks!

[listenbrainz-server] 14amCap1712 opened pull request #2256 (03master…fix-caa): Hardcode urls in listenbrainz.org/art https://github.com/metabrainz/listenbrainz-serv...

lucifer: I truncated the similarity.recording table and am requesting piles of new data, but with higher thresholds, so each set should be considerably smaller.

👍

and as I was trying to brainstorm how to move forward with similarities and artist radio, my brain delivered a Picard AutoTag feature that might making tagging MUCH faster and acoustic fingerprints... useless. thanks brain. not what I needed.

[listenbrainz-server] 14amCap1712 merged pull request #2256 (03master…fix-caa): Hardcode urls in listenbrainz.org/art https://github.com/metabrainz/listenbrainz-serv...

lucifer: any idea why the prefetch_count thing works?

mayhem: it ensures that only a fixed numbers of messages are sent to ts writer at time. without prefetch count, rmq/kombu will try to send as many as possible. when it was failing, this was as large as 10k at a time.

why it silently crashes at 10k messages is not known though because logs were empty. probably a bug in kombu i'd assume.

ok, that does kinda make sense then.

for some reason the metadata viewer opted to show this scan as opposed to the more commonly used front, which is only the right half. huh.

can you share the recording mbid of the listen?

its listening now, so the listen hasn't been recorded yet.

[listenbrainz-server] 14amCap1712 merged pull request #2254 (03master…fix-ts-writer): Timescale Writer Fixes https://github.com/metabrainz/listenbrainz-serv...

ah ok

it'll probably be this api call: https://api.listenbrainz.org/1/metadata/lookup/...

yep

yeah preferred caa in MB is https://musicbrainz.org/release/c06d5b18-af19-4...

where are you listening to this from? something that submits release mbids?

spotify, daily jams.

weird, that doesn't submit release mbids and the one returned by the api is the expected one

the listens page shows the art I would expect:

to confirm you are listening this on spotify and not on BP?

correct. I just closed all LB tabs, music still playing.

Hm, The cover you see does have the "front" attribute: https://musicbrainz.org/release/c06d5b18-af19-4...

it may very well be an MB data issue, for sure.

We should probably filter ones that have front but also other attributes?

Or prefer only front

No, I don't think it's an issue in MB, the attribute is correct as it does show the front.

hmm i see but MB website does show the correct cover.

maybe an issue in the query I copied from CAA.

However my code for fetching the image from CAA json response probably doesn't do more than "get an image with 'front' attribute"

iiuc, this mbid should have been resolved by the backend. frontend would be getting a caa id.

mayhem: This is displayed in the Listening Now standalone deb, yes?

Ah, odd

Well, maybe the same mistake I thought I had in the JS code exists in the python code, i.e. "get the first 'front' image" rather than get the first image with only 'front' attribute

uh yeah, the query is missing an order by.

monkey: yes