#metabrainz

thoughts on how that should be done? 1) redirect -> about if not logged in else user page 2) render about page without changing url but redirect for user page.

We could keep the index page as is, visible to users not logged in, and duplicate or include the contents of the index page at the top of the about page

(I'm talking about the new about page https://test.listenbrainz.org/about/ )

"new"

I'm not sure about duplicating stuff

the include at top sounds better to me

I guess we can do that conditionally too? Include the content of the index page in about page only if user is logged in?

Maybe I'm overthinking this

i am fine either way on that one.

monkey: https://tickets.metabrainz.org/browse/MBS-12205

MBS-12205: Add "play list listenbrainz button" to release, release group and collection pages

[listenbrainz-server] 14akshaaatt merged pull request #1863 (03master…akshat/clean-codebase5): Refactor Codebase Part Five https://github.com/metabrainz/listenbrainz-serv...

if you want to add any more detail to the ticket.

Thank you

monkey: i just saw multiple `invalid request block size: 18387 (max 8192)...skip` in web prod logs. any endpoint comes to mind that could cause this issue?

Can't say for sure

lucifer: http://livegrep.metabrainz.org/search/livegrep?...

ah no, i actually meant monkey because he had found this issue with feedback endpoint some time ago.

sounds like a UWSGI issue.

mayhem, right. that's what i meant any endpoint that could cause so much large url.

the instant playlists could.

Ah, thanks for context. Well it could be that someone is requesting feeback for too many listens at a time?

we recently doubled the buffer size from 4096 to 8192 to accomodate upto 100 mbids in feedback endpoint.

37 characters per recording mbid. Sending about 220 tracks could trigger this.

oh, i was under the impression that we also added chunking in requesting feedback.

*chunking on api calls from frontend.

I doubt it's from the front-end, we don't show that many listens on any page afaik

i see.

it could be the data set hoster.

Although… let me look at code real quick

let me see if I can reproduce

let me look in nginx logs then. since these requests never reach flask, sentry won't be helpful.

monkey, I'm thinking of pausing any UI changes in LB to be carried out until we have introduced the sidebar because that could double our work and isn't worth it at the moment.

I think that's wise akshaaatt, some of the issues we're trying to fix will be gone after redesigning

Yuss

Also, I tried working on bumping bootstrap but it's gonna be a huge PR the way I'm thinking

How do you suggest we minimize the work?

What makes it huge?

(and how huge?)

We may not be able to break that one down

Introducing design system and then potentially fixing the entire front end because of the changes made to the latest versions

Actually we could, making PRs to be merged in a separate branch

Right, we have stuff like `className="col-md-8"` everywhere, don't we.

Yeaahh

monkey for example https://github.com/metabrainz/listenbrainz-serv...

This just replaces stuff but breaks down the entire frontend

I can't think of merging this anywhere. The only thing comes to my mind is fixing everything and then merging

But then the PR would get chunky

I think if we want to make separate PRs, (which will be easier to review) we're going to have to break everything and go page by page (for example) while the rest of the website is broken. Hence a separate branch

That sounds doable but we would have to focus on getting this done soon

So for example merge this initial PR into a new branch, then open PRs for each page

Because the longer it takes, the messier it will be

Yep

Okay then, I can plan on focusing on this

I wonder if it's worth moving to using react-bootstrap if we're in there reworking the layout

I could go with it but in MusicBrainz we have decided to go ahead with normal bootstrap

So it could spin my mind a little

lucifer: did you get the error again just now?

twice?

Looking at our react pages, it shouldn't be that much work so maybe we can do initial PR and then a PR for all the front-end pages

Alright monkey

mayhem: yup.

Ah, let's not complicate it then, if we are moved to BS5 we'll be able to easily move to react-bs if we decide to

but the size is half than previous errors.

ok, I'll limit my calls to that endpoint to 100 recordings.

lucifer: yeah, to be expected. that was the dataset hoster.

👍

mayhem: got that error thrice again? again dataset hoster?

possible.

I just sent it three more times just for good measure. :)

let me push the fix and then it should STFU

makes sense, thanks!

[bookbrainz-site] 14MonkeyDo closed pull request #545 (03master…dev): Initial commit for ravamping footer https://github.com/metabrainz/bookbrainz-site/p...

[listenbrainz-server] 14mayhem opened pull request #1866 (03master…playlisty-endpoints-docfix): Add note about max number of recording_mbids https://github.com/metabrainz/listenbrainz-serv...

[bookbrainz-site] 14MonkeyDo closed pull request #597 (03master…trishna): Improve[UI]: Responsive footer, overlapping of footer and div in contribute page and buttons spacing on collection page for better UI https://github.com/metabrainz/bookbrainz-site/p...

tandy1000: validate-token fix released.

huh. a URL length of 3749 still gives a 502. :(

does the error say anything different now, lucifer ?

mayhem: invalid request block size: 8633 (max 8192)...skip . i still see this in error logs.

which logs are you tailing?

listenbrainz-web-prod

huh, so the request does hit the back end.

it reaches till uwsgi, which errors and does not forward to flask.

make sense. what doesn't make sense is why a URL that is 3749 still trips this.

unrealted, but do we know which users these are ?

dns_server is some part of our monitoring system iirc. but the other two are requesting the same thing very frequently.

not monitoring

mayhem, uwsgi buffer includes url, cookie, headers etc. everything except request body. could any of those be present there?

oh

I guess the browser add a bunch of stuff.

lol @ username dns_server

doubling it is unexpected but not impossible. can you try with a few less mbids and see if that succeeds ?

sorry, was in some meetings. I just saw that I was mentioned. reading backlog now

uh oh. bulk lookup url formats :(

50 seems to work. I'll go for that.

nice to have eagle eyed lucifer on the team. <3

hehe :)

i also found the entry in the nginx logs. its from bono so that confirms again it was dataset hoster :D

general rule: if it could be my fault, it was my fault.

lolol. but i guess we could apply that to all of us.

quickie for you, lucifer : https://github.com/metabrainz/data-set-hoster/p...

should be good to go now.

thx

[data-set-hoster] 14mayhem merged pull request #4 (03master…add-playlist-support): Add playlist support https://github.com/metabrainz/data-set-hoster/p...

monkey, akshaaatt: the prod image build has broken down. https://github.com/metabrainz/listenbrainz-serv...

Funtime

it seems something went wrong during merging #1863

That's all on the player page, let me fix it

[listenbrainz-server] 14mayhem merged pull request #1866 (03master…playlisty-endpoints-docfix): Add note about max number of recording_mbids https://github.com/metabrainz/listenbrainz-serv...

Lol I'll fix that lucifer

Or monkey are you on it already?

On it

Thanks

Issue is we merged a PR with a new react page, and then merged another PR that moves some things. The second PR was not aware of the new react file

oh makes sense.

Yuss

Pushed the fixed paths to master

mayhem, monkey, akshaaatt: in case you find it useful and didn't know about it yet, the github UI has now added a option to update the branch.

Yeah, I was aware of it but still haven't tried it

nice

Yuss saw that

[musicbrainz-server] 14mwiencek merged pull request #2416 (03master…cdtoc-type-fixes): cdtocs type fixes https://github.com/metabrainz/musicbrainz-serve...

[musicbrainz-server] 14mwiencek merged pull request #2409 (03master…flow-0.171.0): Upgrade to Flow 0.171.0 https://github.com/metabrainz/musicbrainz-serve...

https://usercontent.irccloud-cdn.com/file/DQmij... wow that's awsome. this needs to be wider known

[12:43] <mayhem> reosarevok: the idea is that it will fall back to youtube if spotify isn't available.

ftr I tried this userscript for monkey some time earlier this week, and it worked fine in waterfox classic (with youtube and everything)

btw i like the button where it is (above art)

.. or wait maybe it was vivaldi... hmm

no waterfox, works for me https://test.listenbrainz.org/player/?recording...

I do get

Could not save this listen

Bad Request

error tho

...

[listenbrainz-server] 14akshaaatt opened pull request #1867 (03master…akshat/clean-codebase6): Final Refactoring of the Frontend Codebase https://github.com/metabrainz/listenbrainz-serv...

I like how there is like several of these where there is aa clip of a cat making sound https://www.youtube.com/watch?v=6cFo0XDuLiI

like catbased music

people making music ontop of it

ha

"This is my absolute favorite thing about humanity, the fact that ten thousand years of history and innovation all led up to a dozen-odd people who have never met coming to make an absolutely fantastic song together over the sound of a strangely rhythmic cat meowing, all put together by yet another person that none of us have met, purely for the enjoyment of the few thousand people that happen to find it