#metabrainz

alastairp: oh looks nice!

for flask testing, i guess we find another testing lib or fork and fix the stuff we need. the repo seems unmaintained so submitting patches there is likely not useful.

lucifer: pip wasn't happy with upgrading just flask - there were quite a number of dependencies that we have that weren't compatible and so I had to update them too

i see, makes sense

when I did redirect tests in LB I did this: https://github.com/metabrainz/listenbrainz-serv...

not sure if this was in response to not being able to get self.assertRedirects working, or I just didn't know about it?

i see, i do remember using assertRedirect somewhere but don't recall distinctly

bitmap: the only other merged thing is https://github.com/metabrainz/musicbrainz-serve... which is just docs

So I guess I might as well merge master to beta again later if so

but I don't know how much of flask-testing we actually use - asserMessageFlashed, assertTemplateUsed, ...

But there's that weird ISE with annotations we would ideally also put out a fix for :)

yeah

I'm guessing that's related to the diffing, I'm looking at it

yeah, probably wouldn't be hard to add a common test case for stuff we use in BU.

so it's possible that we could replace it with 5 lines in the base test class

*CB, but yes

i have often wanted to not create flask app after each method but once for a class

ah no, i meant BU. so that we can reuse in LB as well.

oh, sure. now I understand what you mean

it's possible that all we actually need to do (given the tests that we have) is create the app and push a request context

e.g., this test client already exists in flask, it's just a helper to move it into the class instance

yes makes sense

bitmap: I guessed the same, but why would it work just fine locally :/

i think the renderer socket is just timing out in prod because it goes so long without receiving any data

then it fails to parse because it only received partial json

So the main issue is just how horribly long the diffing takes?

Not sure what to do about it

yeah, not sure yet. I'm measuring which bit is the slowest

I think generic-diff just isn't optimized for large strings. they suggest to use fast-diff in their readme, so I'll try that

monkey: hi, this dropdown (top right) is from a recent PR by Ansh. Any feedback on it? Specifically, if it's currently set to "newest", should the option in the dropdown be selected/disabled in any way?

akshaaatt: aerozol: your feedback on this welcome too!

alastairp, Ansh: yes, I think showing the currently selected option (which is just adding the "active" css class if i'm not mistaken) would be useful

Not much other feedback, it's clear and the placement makes sense.

monkey: thanks. only other change is to remove Popularity and replace with 'Most Popular' and 'Least Popular' (so that we have the matching ascending/descending options for that)

Yes, that sounds great

riksucks: I'm around now

bitmap: oh, neat

Hopefully that will help

What's the loss compared with the usual one?

it is way faster, but we lose the ability to customize how the strings are split before diffing (e.g. in some cases we use \s+ to implement word diffing)

Hmm

Would it make sense to use it only for annotations?

So, either pass a useFastDiff or call a different FastDiff component?

[troi-recommendation-playground] 14mayhem opened pull request #47 (03main…general-cleanup): General cleanup https://github.com/metabrainz/troi-recommendati...

I think so, either special case annotations or any text larger than n characters

lucifer: quickie for ya, so we can get the troi gsoc condidates to be less befuddled. :) https://github.com/metabrainz/troi-recommendati...

bitmap: I just thought it might be annoying to check len, but maybe not

Can you send a quick PR?

sure

Ideally we'd release that today as well (with a quick test in beta to see nothing is broken) because the whole thing seems ridiculous at the moment

mayhem: done

Hi monkey, I wanted to ask, how do you guys plan to code the new frontend of LB

Should I be keeping something in mind while writing code, so that minimal changes need to be done

When migrating the codebase ofc (specially asking this so that I can choose the right drag and drop library for the soc project)

lucifer: thanks!

I don't think there's anything you would need to keep in mind, except maybe one: I think we will eventually end up with a sidebar to show the current user's followers on any page, so don't assume by default that the new functionality will only ever be used on the feed and home pages

No specific considerations with the drag&drop library.

I've been using react-sortable for playlist items but I don't think it does full drag/drop features

might be worth havign a quick look though

[musicbrainz-server] 14mwiencek opened pull request #2484 (03production…mbs-12302): MBS-12302: Speed up large annotation diffing https://github.com/metabrainz/musicbrainz-serve...

Some unsolicited advice on the side: probably best to start with and get a fully working end-to-end feature using the ListenCard context menu before doing anything drag&drop.

alastairp: I think the popularity fix you suggest makes sense. Other than that monkey has already suggested on how to make the selected option visible :)

[troi-recommendation-playground] 14mayhem merged pull request #47 (03main…general-cleanup): General cleanup https://github.com/metabrainz/troi-recommendati...

reosarevok: opened a PR for that though there seem to be some issues / test failures with certain diffs, will check in a bit

bitmap: thanks!

Also, hmm

Opening instrument for editing also creates links to all the instrument attributes, heh

(not from the list, but from its doc page)

I see two options: 1) have a check for id = 14 that shows a string for "Possible values" like "The possible values for this relationship can be seen from the {instrument list}"

2) have the possible values here, but block editing attributes if they have a root ID of 14 and not an id of 14

bitmap: what sounds better to you?

both, I think? :) linking to /instruments makes sense but also making sure you can't edit them even if you type the url

lucifer: I made 2 small changes to https://github.com/metabrainz/critiquebrainz/pu..., if you could have a quick look at them and see if they're sane then that would be welcome

Ok

bitmap: also, maybe you can help jesus figure out https://github.com/jesus2099/konami-command/iss...

alastairp: lgtm, thanks!

awesome

I need to remember how to re-upload CB translations too. lots of work done recently changing strings

[critiquebrainz] 14alastair merged pull request #420 (03master…bad-urls): Fix some incorrect url_for definitions https://github.com/metabrainz/critiquebrainz/pu...

[critiquebrainz] 14alastair merged pull request #394 (03master…sort_reviews): CB-347: Allow users to browse reviews by new, old and other criteria https://github.com/metabrainz/critiquebrainz/pu...

[critiquebrainz] release 03v-2022-04-11.0 has been published by 14github-actions[bot]: https://github.com/metabrainz/critiquebrainz/re...

Ansh: your review ordering PR is up, thanks! https://critiquebrainz.org/review/?sort=popular...

lucifer: CB-418

CB-418: Improve display of releases on recording review page https://tickets.metabrainz.org/browse/CB-418

"Currently, we show the first recording only." - I think you meant to say the first release only?

alastairp: Thanks for your review.

uh yes, alastairp.

fixed

I think we should also update the documentations.

Appreciate the thought Ansh :)

Ansh: which documentation?

that should be automatic i think. if you update the docstrings in flask endpoint, it gets updated in doc.

Hi, I had a few questions that I posted late on Saturday. I'll post them again..

Hello @monkey @lucifer, I had a few doubts about send track project.

1. I am thinking that when a user clicks on "Recommend to user" option from the ListenCard component, they will see a modal with a populated list of users along with check boxes. What should this list populate?

a. Followers + Search box to search MusicBrainz users

b. Followers + Following + Search box

c. Followers only

d. Followers + Following only

2. When a recommendation is sent, should the sender see his sent recommendation in their feed? eg. "You recommended 'Song' to User B".

3. Should a third user (say User C) be able to see who recommended what track to whom? eg. "User A recommended 'Song' to User B".

I have done some manual HTML edits. The ListenCard component's options will look something link this after adding option 'Recommend to user'. https://usercontent.irccloud-cdn.com/file/ujiYg...

A User B will see something like this in their timeline when User A recommends them a track. https://usercontent.irccloud-cdn.com/file/P0Wx5...

chinmay: its going to be regular meeting time now, you'll probably need to wait an ~hour.

No worries

Mind if I join?

lucifer: I don't think it is getting updated automatically.

Ansh: I'll check if the docs are building automatically

chinmay: sure. its here in the channel. if you want to give a review as well, when Freso asks let him know.

Speaking of…

Oh alright :)

It’s International Louie Louie Monday!

I have one review from CatQuest, so:

CatQuest says…

"""

I was mostly working on BookBrainz tickets. This week will be spent preparing for Easter. 🐟

"""

atj: Go!

(Others up: akshaaatt, reosarevok, yvanzo, lucifer, monkey, zas, bitmap, mayhem, alastairp, Freso – anyone else who want to give review, let me know ASAP!)

No atj?

akshaaatt: Go!

Hi everyone!

Last week I continued work on MusicBrainz.

bitmap and I were also discussing how we could automate the release process, hence I started to understand jenkins and how we could potentially migrate to github actions.

This involved some research work.

Other than that I reviewed yellowhatpro's GSoC proposal and continued some work on the android app.

I also fixed some work in the Listenbrainz revamp work.

That's it for me! fin. Go reosarevok!

Hi!

Last week I worked on a bunch of tickets, including MBS-11178 (to document entity types and whatnot)

MBS-11178: User-facing entity type details lists https://tickets.metabrainz.org/browse/MBS-11178

And finally getting Freso's requested MBS-11866 hopefully in a reviewable state