#metabrainz

yes, indeed. ok, thanks! I'll have to play with that.

Zastai: have you taken a look at laptops with Ryzen 4000 processors? They go up to 8 cores/16 threads

not yet. wanting a high end gfx card tends to automatically mean intel from what I've seen, but I'm not specifically looking for intel over amd

ishaanshah: we can do the meeting early today, if you have time

Yep, sure I am around

Great

Thanks

How are you?

Doing well, how about you

Good good

Done with manual testing right now

For the release thing?

Yeah

Updating the tests now

Nice!

Hopefully it works this time

Will you get time to deploy and test today>

?*

If it's possible in the next hour or so, sure.

Cool, I will try to get it done

Great.

If it works then I can proceed for recording and the small bug fixes tomorrow

In terms of priorities, I think we should fix the small bug fixes first and then start the recording work.

But yeah, that sounds like a plan!

Yep, yep those are important

Great, I don't have much else, do you wanna talk about something?

Nope, nothing else

Cool. I'll try to test the release thing in prod, once you open the PR.

[musicbrainz-android] amCap1712 opened pull request #43 (master…view-binding): Replace findViewById with view binding. https://github.com/metabrainz/musicbrainz-andro...

ruaok: I am currently free now. What exact behaviour do we want to test for more than 3 week spaced listens?

to make sure we don't miss any.

alastairp: thanks for the feedback!

Making a new ticket for it would be nice, we'll try to fix those things this week.

sure, I wasn't sure where to put it

that is, the specific issue you raised is fixed, yeah, but this seems to be all involved somehow

I just had a big discussion with Mr_Monkey about it, he said he'd give some feedback when he got home

Yeah, this kind of feedback is really valuable. And small things like this are harder to notice for me and ishaanshah because we're too deep into the implementation. Thanks!

it seems like this kind of page interaction is going to be really important in LB, so it makes sense to work out how to do it well, so that we can just copy/paste onto a new page

Mr_Monkey showed me how the flask app sends data to the react on the listens page, this seems like a neat way of doing the interaction

Totally agree about getting it right early. I think react-router should reasonably handle all of these issues.

I do like using the API in javascript better than rendering the data into the html template though.

Is this about the back button not working on graphs page

?

LB-610: The URL in the browser should update with parameters for the history page

Oh

Although yeah the back button doesnt work

I tried to get it to work, but its difficult to sync state and parameters

as I said, I had a bit of a chat with Mr_Monkey about it, he said that we could spend some time working out a standard for how do to this, synchronising the state between the server and the frontend

this will be useful for some parts of AB too

as iliekcomputers said React Router provides a solution for that

*For back button

ishaanshah: a suggestion for next time - feel free to add a comment to any pull request or ticket saying if you found a problem but were unable to find a solution

for example if you knew that the back button wasn't working in this case, just a comment "I know the back button isn't working, but can't work out how to handle this nicely" would have been great

Yep, will do that next time

> I do like using the API in javascript better than rendering the data into the html template though.

iliekcomputers: The ideal world is doing both, IMO. Load the page with initial data (like we do for listens page), and then the next/previous buttons make call to the API (like the stats page does).

Currently those two page are missing what the other one has…

ishaanshah: I'm interested in all the solutions, quirks and issues you/we find for url parameters, as it's fairly high on my list for BookBrainz (now that we have better pagination)

Is react router really needed? is it not overkill for just changing url params and browser history?

Mr_Monkey: what's the benefit of loading the page with initial data? It'll need more code which we could avoid.

Better user experience, less API calls from the client

yeah, it prevents the browser from having to do _another_ round trip for the API data

If I have a really slow connection and I navigate to page 2, instead of the page1 flashing I't'll slowly show page 1, then slowly do another API call to show the correct page. Not ideal.

Also not showing a blank page for people with JS disabled.

Nothing to stop us using the API server endpoint in flask to get the data for the right page

Or, alternativaly, if any of the URL parameters are not the default ones, maybe skip the initial API call?

Yeah, that's what I was thinking, we should just skip the initial API call

Freso: That's another story… we're talking about sending data with the page rather than rendered html

Freso: our suggestion isn't a complete fix to this, the page data will still be rendered by js

[musicbrainz-server] yvanzo merged pull request #1540 (master…MBS-10469-delay-showing-CAA-on-front-page): MBS-10469: Add 10 minute delay to shown CAA on front page https://github.com/metabrainz/musicbrainz-serve...

MBS-10469: Only show release with processed cover art on front page https://tickets.metabrainz.org/browse/MBS-10469

Ah.

iliekcomputers: I do like how snappy the listens page is at first load. Having all the data you need before render starts is a plus.

the timescale branch is a bit slow for the first load of a user -- that is to calculate the total listen count.

[musicbrainz-android] amCap1712 closed pull request #43 (master…view-binding): Replace findViewById with view binding. https://github.com/metabrainz/musicbrainz-andro...

That being said, skipping the API call will already solve that flash of incorrect content issue

I'm inclined to fire off the calculation of that on page load, but to load it from the browser via a JS call so the page load doesn't need to wait.

While I would love for it to be as snappy as possible, I don't really feel good about having to maintain two paths from which data could come into the component unless absolutely necessary

The initial flash can be solved by fetching data in componentDidMount seperately

I feel that sentiment.

Right now I am using other functions to do this to avoid duplication

The flash is definitely an issue we need to fix

> I don't really feel good about having to maintain two paths from which data could come into the component unless absolutely necessary

I'm not convinced that this is as much of a problem as it seems

I'd be happy to have a discussion/hack day soon to see if we can work out how to do it without repeating code

We'll have to reuse the code that fetches data in the API endpoint in the profile view. The frontend should be ok.

I'm not sure what you mean by "reuse code"

do you mean copy some code and have it duplicated in 2 places, or "call a function"?

Call a function, ideally.

gah.

bitmap: I hope you're staying home. CTA is closed. when did that happen last? (besides yesterday, I guess)

What's CTA?

ruaok: never that I can remember. pretty crazy

chicago transit authority.

all public transit is shut.

I was visiting my parents' so I'll be stuck here for a while I guess

so... bitmpa, when are you moving to the EU?

Crazy times, stay safe, bitmap. :(

good question lol

likely not today

LOL

not anytime soon, really. the EU is largely still closed. esp to the US.

[listenbrainz-server] vansika opened pull request #896 (master…send-params-train-model): send params with spark request to train models https://github.com/metabrainz/listenbrainz-serv...

Didn't know iliekcomputers was into train models. ^

Or model trains

lol

Sometimes it just takes a spark to set the kindling interest ablaze

[listenbrainz-server] ishaanshah opened pull request #897 (master…release_stat_iterate): LB-603: Fix spark crash on calculation of all_time release stats https://github.com/metabrainz/listenbrainz-serv...

LB-603: All time release stats overload the spark cluster https://tickets.metabrainz.org/browse/LB-603

iliekcomputers: ^

Mr_Monkey: you have the S trap which we cut the top off, right?

this one that I have doesn't fit in the cooler :)

uh-oh

It is Global Monday of Parents!

People up for reviews: pristine__, reosarevok, Mr_Monkey, ruaok, zas, Freso, bitmap, yvanzo, alastairp, diru1100, CatQuest, ishaanshah[m], Cyna[m], shivam-kapila, _lucifer – anyone else, please let me know ASAP!

pristine__: Go!

pristine says…

"""

Worked on sending params with spark requests, on storing recommendation metadata in HDFS and worked on documentation for LB cf recording

"""

reosarevok: Go!

Hi!

I mostly just updated my PRs last week and merged a bunch

Also worked on MBS-10499

MBS-10499: Entity cleanup banner not shown when add edit is still open https://tickets.metabrainz.org/browse/MBS-10499

So hopefully that'll help confuse less newbies

This week I'm updating PRs again, since bitmap made a huge-ass change that means almost anything that adds flow-typed things needs updating

And then I'll see what I do :)

sorry!

It's ok!

Fin, go CatQuest

CatQuest says...

"""

Meow! `😺`

Tested merging on BookBrainz!

Wrote Even moar tickets!

Wrote some on the rubab-senirebab-etc instrument descriptions, they'll be published shortly!

Here society is lowly returning to normal so I've been outside walking/hpwu-ing etcetra

♉