#metabrainz

[listenbrainz-android] 14akshaaatt opened pull request #163 (03main…additions): Fix token insertion and cleanup code https://github.com/metabrainz/listenbrainz-andr...

rescrobbler takes ts when it submits the listen

on servers which have ~/.shh/docker_server_configs_key, I separated them

zas: id_rsa on all hosts do match these deploy keys

("all" that I checked)

maybe not on newer servers

for example

the listen Small by Portishead

That would explain it.

I'm fine with having to specify GIT_SSH_COMMAND in any case, it was just unexpected

is 6.45 minutes long

Do our docs state which is correct?

6.45/2 ~ 3 mints

Gtg. Bbib.

ah got it, I usually use a specific key and change config to have a shortcut

yeah, if you don't specify a shortcut, it uses id_rsa

Negative, here's what the LB docs say:

>The payload should also include the listened_at element, which must be an integer representing the Unix time when the track was listened to. The minimum accepted value for this field is LISTEN_MINIMUM_TS. playing_now requests should not have a listened_at field

yes, the main issue with this approach is that you have to clone with the "shortcut" instead of github.com

so I guess we can't really avoid that (after it does what the user asks)

when should the ts be taken then?

Ok, I will review hosts listed in docker-server-configs at least

start seems good to me

bitmap: It's not needed on other hosts (which just have their identity file as id_rsa)

but doesn’t accurately represent when the actual listen was submitted

jasje: IMO at the start is correct

yvanzo: ok, I thought I had created a separate deploy key on a few other servers too, but maybe it was for docker-postgres-cluster

Ingested listens will also have an `inserted_at` field added to them

what do you mean by shortcut btw?

bitmap: a config like this https://www.irccloud.com/pastebin/kJYL9NNK/

start it is then ig

just an entry in ~/.ssh/config

as above ^

ah right

docs updation required tho

I will update the sys docs too

perhaps there are better options, but basically it links a deploy key to a specific name (hence the "shortcut"), since the "Host" can be used in place of normal one (usually git@github.com)

atj: ^^ aware of better approach?

[listenbrainz-android] 14akshaaatt merged pull request #160 (03main…add-newsbrainz-activity-android-manifest): Fix: Declare NewsBrainzActivity in androidManifest https://github.com/metabrainz/listenbrainz-andr...

[listenbrainz-android] 14akshaaatt merged pull request #163 (03main…additions): Fix token insertion and cleanup code https://github.com/metabrainz/listenbrainz-andr...

yvanzo: are you moving the key on pink for now or?

bitmap: it's working on pink for now

ok

zas, cc atj: it's the recommended approach at least: https://docs.github.com/en/authentication/conne...

git config --local core.sshCommand 'ssh -i /path/to/key'

mayhem, I’ve triggered a beta release of the app with new ui/ux which should fix the issue of configuring the scrobbler and ease the pain that user goes through

Do join the beta program for the app on the playstore

I’ll inform you and others when it’s available

how do I join the beta program? I've got the app store page open

on the question as to when to submit the listens, we should do the same that last.fm does --that way scrobble programs need to have less customization

reosarevok (or whoever leads today's meeting): I won't be around at meeting time since it's a public holiday here and we're barbecuing.

My review https://www.irccloud.com/pastebin/ZwYeGbQe/

happyrbecue

Thanks yvanzo!

[listenbrainz-ios] 14gauravbhardwaj7 opened pull request #3 (03main…update-listens): Integrated user-listen api https://github.com/metabrainz/listenbrainz-ios/...

mayhem https://youtu.be/RiWixaMp9Zs

https://youtu.be/dQw4w9WgXcQ this might also help

I couldn't find the damn button, because I am already signed up. time to slow down and read. lol.

Welcome to Monday Mayhem with me, and I guess also mayhem

Today everyone is being lazy (good for you!) and we have not one, not two, but three mailed in reviews!

ara?

kellnerd says:

'''

Last week I've fixed two BB bugs before the release, one was caused by Safari's inability to keep up with browser standards, the other was my own ;)

CatQuest: now

Time to fully get back into BBusiness this week by starting with my GSoC project, yay!

'''

?

atj said:

'''

!

No he did not say !

Please, you two, let me post these :)

atj said:

'''

Last week I continued working on preparations for the consul migration with zas. This included writing a new role to install and configure unbound, resolving an issue that caused consul not to be started on boot and implementing new functionality to manage consul script checks. We also did a fair bit of testing and it looks like we're pretty close to being ready.

I also spent some time looking for summit accommodation and discussing options with mayhem.

'''

And our usual mailer aerozol said:

'''

Bits and pieces this last week.

Clarified the auto-editor election process with reosarevok. Drafted some dataset attribution guidelines with mayhem. Tweaked some Weblate colours and icons with yvanzo.

A bit of community stuff, some more MB guideline drafting.

A short review from me this week, who would have thought! Go… bitmap!

'''

so, go bitmap I guess? :)

heh

hello

last week I was mostly working on improving our postgres cluster for high availability so we can perform switchovers with no downtime

I spent time testing and improving the (currently unused) haproxy container in the docker-postgres-cluster repo, plus the associated health checks for determining whether a node is primary or standby

this is still in the testing phase, but perhaps we could deploy it next time we ugprade pg (maybe in lieu of next schema change)

[insert "i know some of those words" meme]

I also submitted a pr for MBS-13015 and did a small amount of code review

MBS-13015: Tracks with (only) URL relationships do not show in the relationship editor https://tickets.metabrainz.org/browse/MBS-13015

fin. go reosarevok !

Hi!

I spent quite a bit of time improving our JS tests

Still a lot to go

Other than that, tiny bit of code review, a fair amount of community management

Few small tickets

I also sat on the board meeting for Q2 which was uneventful

I think that's most of it - CatQuest, you?

Arararara~! hi!

Figured out some monkey stick/stump fiddle mess

Uploaded quite a few images with reo

working on adding another instrument right now

and tested the new ✨ Alias Locales!✨ ♉

CatQuest: now?

thanks yvanzo and reo for such fast deployment

Oh, yeah, I also helped upload some images ^ :)

I'll be a bit more afk in the coming weeks. But i'l be around.

and ya are jsut going to have ot get used to me going "ara"

ut it's JP not Ca

uh next can be ...

fin, go Pratha-Fish 🐟

hmm

zas then

what does ara mean in your context? around here it means 'now'

uhm..

monkey: can explain better thna me i think

and if zas is not around, thne monkey can go

Hello !

Last week I merged some BB PRs, did a bunch of beta testing, some fixes, and deployed a new version of BB

Also reviewed PRs for BB, particularly working on the revision reverting

I'm around btw

Oh, I forgot the list for today: zas, mayhem, yvanzo, lucifer, akshaaatt, Pratha-Fish, riksucks, jasje, jivte (GSoC students, if you're here ping me and I'll add you too)

On the LB front, I worked on improving and fixing playlists

Did some ticket triage, fixes tiny bugs

And talked more about the redesign and new recommendations page

That's the gist of it. zas can go now :)

Hey

last week was mostly about preparing the most to new consul cluster

ara ~= "oh dear" I guess

as atj said, we're getting very close, I expect to do it this week

OR "oh is that so?"

or waait? huh? i think or as an exlamation of surpise

Ok let's let zas continue :)

👍

We had another problem after docker upgrade (this time minor version one) resulting in broken net stats

it happens last time (major version upgrade), so we thought it was because of that, but apparently something changed in docker that causes live-restore to not work well

that's very annoying, because the only way I found to fix it is to restart containers after the upgrade, and basically that nullifies the live-restore feature

this is something we need to investigate more deeply, I keep the focus on consul upgrade for now

apart that, I detected weird things about SOLR that yvanzo have (partially) addressed, concerning missing or duplicate replicas in our SOLR cluster

reosarevok: can you please add pixelpenguin, vscode, shivam_, arsh to the review list?