#metabrainz

cool.

FWIW, this particular case is also of a bad mapping. but that's a different issue.

mapped to https://musicbrainz.org/release/4269c0ab-06cd-4... instead of https://musicbrainz.org/release/78abe772-e46f-4...

So we only do this when 1. user submitted a release MBID and 2. it is not the same as the mapped release MBID

because release date

monkey: yes sounds good.

mayhem: the redesign PR, why not release today?

friday.

err, thursday.

thursday. yeah

big shit like this is best to go out monday.

ok, I really want to get some datasets work done. does anyone else need me before I put my head down for an hour or so? speak up now.

I'd like our resident designer to have another pass, give another few days for users to chime, and have a blog post ready to be posted when we do deploy.

[musicbrainz-server] 14santiagofn opened pull request #2926 (03akshat/footer-revamp…akshat/footer-revamp): Updating footer revamp https://github.com/metabrainz/musicbrainz-serve...

sorry, i couldn't find a way to add my commits to the already existing PR 😢

you likely dont have the rights to do that. :(

but this should work fine, methinks.

reosarevok: for when you return to work: https://test.metabrainz.org/datasets -- the datasets page flow is done now and can be checked for all manners of mishappen things. the only thing left on my list is redirecting users who took the signup plunge to the download page, but that is a project for the morning.

[listenbrainz-android] 14akshaaatt merged pull request #136 (03main…fixes-listenservice-yim-jasje): Blacklist UI added and minor tweaks https://github.com/metabrainz/listenbrainz-andr...

Hi jasje

Two things from #137. 1) Where are we picking up the now playing feature in this PR? 2) I am not sure if maintaining our own timer is the right solution.

Have we checked what impacts it has on performance?

akshaaatt: no playing now will be done later.

You have it linked in the description of the PR

2) Timer is just one solution, since it executes a little amount of code after 1 sec

actually the reporter mixed up the title and description

just read the description for details

the exact thing timer solves is that it keeps track of how much a track is played. lets say i were to skip to some part of the song, the timer will keep track of total listen time

but ofcourse this a very initial approach and I would like some suggestions on it

this approach is also used due to the unstability of android's APIs

[listenbrainz-server] 14faraz16iqbal opened pull request #2463 (03master…refresh-playlists-page): fix: refresh playlists page on while creating, editing and deleting a… https://github.com/metabrainz/listenbrainz-serv...

I think we can just use the delayed message thing to submit the listen after 1min or so has been played jasje

What’s the harm in doing that?

akshaaatt: pausing and playing again doesn't extend delay.

we might have to cancel listen, then note down the remaining appropriate duration left

and post the message again when the user plays with calculated duration.

in the end

we still keep track of the duration

what we can do is create a custom timer and set its period to 2 or 3 seconds so that instructions are executed once per 2-3 secs

but that would mean i have to create a timer class of my own

which is a bit too much of work

and not that fruitful.

plus I don't see any visible performance issues as of now.

I agree that we will have to implement a timer

Just went through the possibilities

In that case, your solution looks reasonable to me.

We could optimise it since we don't require precision

but that would mean copying the libs code and a lawsuit xD

we can always fork the repo and work out our own implementation and improvements to the library jasje

I did and actually posted a Pr to that guys repo but he looks dead to me

Depends on the license of the repo. And this one is Apache License 2.0 so should be fine

oh

but wait how is that supposed to work out

It's not necessary to only submit PRs. You can carry on with your own stuff as well

A lot of commercial orgs use open source code in their proprietary stuff

If only he would have made his class open

things would be alot easier

lol. you do it then

I haven't ever created a github lib but looks like Ill do it now

It's the easiest thing ever tbh. Use jitpack according to me

alright ill give it a try

[musicbrainz-android] 14akshaaatt opened pull request #221 (03master…refactor): Refactor entire codebase https://github.com/metabrainz/musicbrainz-andro...

oh we still had this one

I think MusicBrainz app should just be converted to a sdk type of thing

akshaaatt^^

SDK for what jasje?

for search apis?

An app is just an interface jasje

The app itself uses apis from the backend

all apis musicbrainz has to offer

MB already offers them

but i would be simpler if we just made the calls. none of the retrofit stuff

it would*

Lol

probably create a module inside that others can access

some suggested this in tickets :P

Ouu

I don’t think that’s what we are aiming for currently

tru but still putting the idea out

The mb app is still great for searching, barcodes thing, picard thing and collections

I am not focusing on the mb app currently but I do think we should continue that in future

I want to experiment with some ML stuff soon on both MB and LB based on user interactions in the app

its not that helpful if adding artist takes you to web anyways

Sure we can look into that soon. We just need apis made for those

all the apis are so clunked together with the website

Plus some hard UI/UX decisions which are hard to make currently. But we can do so in future

alot of work indeed

Tbh we already have the base ready for things now. It’s just a matter of pumping features out

I’m assuming MeB will soon come out of that hell hole where we were just refactoring stuff everywhere. The important decisions have already been made and are in progress

less workforce

The only major thing left is the MB revamp… let’s see if santiagofn and I can work through some of it

It’s not just about the workforce jasje. Tbh there are only a few people who understand the tech we are working on and have experience with it

mb sure does need the revamp

akshaaatt: no no not mb/lb, the apps

android is pretty abundant

The fact is that our user base MB wide is not a fan of editing on the app yet because of obvious reasons

but still very less people here

Opening browser for stuff and all..

agree on that

LB android has a huge chance of winning people according to me because it is relevant to the use case

But LB lacks a brand yet overall

Which we need to create via the revamp, app, and more exciting features like yim

Recommendation engine is the key here

true

Once we have the apps and things there for us, we can focus closely on user interactions and work on stuff which people want more

I guess recommendation engine could use some help from users

Right. Exactly that

I want to use ML on the app side to recommend songs better to the users

like dislike a recommendation

Yeah. Tiktok does that

use listen time

Yeah

Even on server side more of what your following listen to and stuff

Plus better engagement with users through stuff like yim or push notifications..

push notifications is a must

FCM is the way

Let’s see…

can we replicate FCM tho?

Lol no

how on earth do they do it

Google uses the play services for it internally

They don’t want to expose push notifications to users now because a lot of people used to misuse it earlier

Like reading other ap’s notifications

it should be just repeated API calls?

Yeah but that drains the battery jasje

what extra do they do then?

Google and these smartphone guys don’t want it to seem like their phones are shitty and that they drain in 4hrs just because some shady slim wanted to do repeated api calls every 2secs throughout the phone’s lifetime

xD

The thing with push notifications are that they run even when the app is closed

That’s the whole point of them

is there competitor of FCM?

Talking about MB revamp, it's the first time I see a backend done using Perl. It's also the first time I encountered Flow instead of typescript

Ikr santiagofn ! Same feelings