Biggest item was sorting out a few YIM illustrations. monkey’s picked a fav so I will go into production and draw the rest now
Go jasje!
jasje[m]
Hi everyone!
MonkeyPython
looking forward to see how the YiM will look this year!
jasje[m]
This month is exams again ;(
But still got free time since they are spaced out
Got to review some prs by Shreyas:
He recently started contributing to LB Android and so far its amazing!
Thats it for me
Shreyas: want to participate?
reosarevok[m]
Still on my list for today: yellowhatpro, akshaaatt, bitmap, derat, zas, kellnerd, mayhem, yvanzo, reosarevok, lucifer
jasje[m]
Looks like he is not around
lucifer: go
lucifer[m]
Hi all!
Last week, I worked on improving the import of lastfm loved tracks, fixing sitewide statistics and submitting listens using oauth.
That's it for me. [@mayhem:chatbrainz.org](https://matrix.to/#/@mayhem:chatbrainz.org) next?
mayhem[m] joined the channel
mayhem[m]
hiya!
last week was some work on improving cron jobs using sentry which looks quite promising.
lots of Meb background tasks as per usual
but the most important mission since friday has been: where do you find the best Feuerzangenbowle in Hamburg's Weihnachtsmaerkte?
tux0r
when in hamburg, try the whisky plaza
fuck feuerzangenbowle!
mayhem[m]
bis jezt, ist die am Rathausmarkt die beste, Gerhard Hauptmann Platz is die zweitbeste. yum.
we'll report complete findings next week.
mayhem[m] notes that we has whiskey on fleetstrasse
fin. more serious updates next week, while a bit less tipsy!
zas: go!
zas[m] joined the channel
zas[m]
Hey
reosarevok[m]
Still on my list for today: yellowhatpro, akshaaatt, bitmap, derat, kellnerd, yvanzo, reosarevok
zas[m]
Last week I focused on finding why solr cluster was getting unstable, found out it was happening when one node was lacking memory, increased JVM max memory a bit, it didn't go well, unexplained very high disk reads, which disappeared after a reboot of all nodes...
Cluster seems now stabler, and a bit faster.
Apart that, I spent some time cleaning up influxdb, we had a lot of metrics data from retired servers.
And I upgraded grafana to last 11.x version, I had to fix few dashboards and alerts.
Plus usual supervision, minor admin tasks, system upgrades, user support, Picard PR reviews and MB edits. fin. bitmap?
bitmap[m] joined the channel
bitmap[m]
hey
last week I mostly did code review, deployed some configuration changes for the new oauth endpoints, and looked into some mb-solr issues with yvanzo
also made progress on fixing github actions test failures (but more cropped up 😬) and writing edit class tests for the alt. tracklists branch
that's about it, go yvanzo!
yvanzo[m]
Hi!
Last week, I deployed a security update of Jira for MetaBrainz Tickets.
I nudged people that were at the Solr meeting during the summit.
With bitmap and lucifer, we worked on SolrCloud 9 deployment to mirrors.
This is needed to get rid of the SolrCloud 7 cluster in production we pay for.
SEARCH-733: Initialize locally with Solr 9 in SolrCloud mode
yvanzo[m]
With reosarevok, we worked on creating a testbed for the search indexer.
This is needed to make performance tests on the servers (hip & trille) we pay for.
Plus MBS dependency security alert, ticket triage, PR review.
Fin, go reosarevok?
reosarevok[m]
Hi!
Last week I spent some time working on making prometheus and grafana work with mb-docker (with some results)
I also finished and got merged some code from 2021 to convert a form to React 🫣
And to update the code to filter unofficial and VA release groups (about that old)
Which is almost ready now, I think
Other than that, the usual small tickets, and some classical editing
What about you, derat ?
Still on my list for today: yellowhatpro, akshaaatt, kellnerd
derat[m] joined the channel
derat[m]
sure, and hello!
i realized yesterday that it might be possible to make yambs support merging recordings across multiple releases, so i spent some time implementing that. it seems to be working, so i'll probably expose it through the web interface soon. i assume that it's less streamlined than the mass-merge userscript, but it's nice to have a non-userscript option
other than that, i spent a decent amount of time pointing editors toward style guidelines and cleaning up old releases from editors who didn't respond to edit notes :-P
that's it for me. yellowhatpro, next?
aerozol[m]
yellowhatno
yvanzo[m]
Only kellnerd seems to be around.
kellnerd[m] joined the channel
kellnerd[m]
Hi, I can go :)
Last week I've started to refactor the data flow of the BB import producer and consumer.
There were a few redundant and confusing properties which made it annoying to recursively find and potentially import entities for use in relationships.
Looking better now, I also found another bunch of unused code that could be deleted following SoC.
Other than that, the usual MB edits and a Harmony update to seed edit notes for ISRC submission with MagicISRC.
akshaaatt: Your chance to go, if you are around
reosarevok[m]
Guessing not
I think that's it for our reviews today then!
Thanks everyone
aerozol[m]
That’s 3 missed for akshaatt, do I remove from regulars or does it not count for employees?
reosarevok[m]
He isn't really :)
aerozol[m]
Can discuss after, soz
Ah okay, I might be out of date with my info
reosarevok[m]
But I would not remove without asking, maybe mail? :)
aerozol[m]
removered
reosarevok[m]
(he's a very active volunteer but he has a day job somewhere fancy IIRC)
There is one extra topic: beginner edit limits
yvanzo[m]
He was nice enough to send a review by mail usually.
Just have to remind it. :)
reosarevok[m]
This came up again recently because derat has been struggling a bit with users adding a ton of releases with importers before anyone can teach them to do it better :)
We had had some talk in the past of restricting beginners to a number of edits per day
(and maybe make them all go to vote)
aerozol[m]
imo - *everyone has been struggling a bit, but derat actually reports them
reosarevok[m]
It's important to limit them to some degree especially if they all go to a vote, otherwise they'll be overwhelming
derat[m]
heh. i don't really want to vote no on beginner edits that are adding releases with e.g. wrong capitalization, but it'd be nice if new editors were slowed down a bit until they've had time to internalize the guidelines
reosarevok[m]
I don't like the idea of a blanket restriction of editing because, well, I was doing classical again this week and you can get 100+ edits by trying to do one release semi-properly
yvanzo[m]
Which kind of edits more particularly?
reosarevok[m]
(but having 100 or 500 add release edits as a limit makes no sense either)
derat[m]
tangentially, i'm not sure what the onboarding flow is currently like for new users. is there e.g. a welcome page that points at the guidelines?
reosarevok[m]
So I was thinking, should we look at what are the three or four more problematic edits for beginners, and limit those?
yvanzo[m]
Would you want to limit beginners to adding X releases per day?
reosarevok[m]
Add release and Add cover (maybe event) art seem obvious
aerozol[m]
It’s never fun to find that a new editor has borked a release, check their recent edits, and they have added 10s or 100s of edits in the first few days
reosarevok[m]
(from what people have complained about)
yvanzo: yeah, probably X releases, X pieces of cover art, at least - dunno if something else :)
aerozol[m]
I would do all edits tbh. Even if they are fixing caps it’s not the worst if they do a few releases and then take a break. We risk making the user onboarding even more complex with various specific rules
yvanzo[m]
It would be totally feasible to have a limit set for each type of edit.
reosarevok[m]
aerozol: I'm pretty sure I would have left if I had been blocked after a few edits, then forgotten to come back, so I don't like that idea much :p
Well, a few hundred edits is a problem if they import a few hundred releases 😅
I guess we could have a general few-hundred-edit limit
And then a lower one for releases and artwork
derat[m]
how would the limits be enforced? for e.g. editing a release, i'm not sure that it's clear what edits would be created until they've gone to the trouble of entering them. blocking off the add-release form if they're at the limit is probably easier
aerozol[m]
Honestly, anything’s a start. Once we have it in we can tweak it. But specific limits are fine too. It’s not fine if it’s badly explained to the user though, because then we deal with more confusion. So I think that would have to be implemented carefully
Blocking the add release form with a nice message sounds quite clean though!
reosarevok[m]
I think blocking the add release form might actually be the easiest, yes, in that sense
yvanzo[m]
10 releases, 20 images, and 1,000 edits a day seems reasonable to me
reosarevok[m]
I'd be happy with that
If the edits go to a vote, that means they're also more likely to actually stop being a beginner after the 10 day limit, so it should not be a problem for too long
monkey[m]
We could start at 20 release/day and see if the situation has improved sufficiently, if we wanted a bit more leeway, but otherwise that sounds fine
yvanzo[m]
That would be a good start. We could then try to further limit the number of destructive edits too.
monkey[m]
+1 for heavily limiting destructive edits
reosarevok[m]
Yeah, I wouldn't necessarily mind limiting merges for beginners, but that's a smaller problem in that less of them probably find the merge button :)
(although I don't doubt the ones who do might make a mess)
kellnerd[m]
20 releases is too much IMO, personally I think I added one or two releases on day one
More releases than that is a sign of lazy mass imports to be honest
aerozol[m]
Gets tricky if someone is doing a lot of work on a big release and then gets hit by the limiter on the “submit” page. Maybe it’s a matter of blocking edit pages, but still letting edits go over if they’re all at once.
reosarevok[m]
Well, I don't expect most beginners to add relationships or anything to that effect - adding 10 releases with imports properly takes 30 minutes :)
(at most)
If you just add the release without extras
So 10 might be fine?
But yeah, we should make sure we don't lock someone who already has the page open, I guess
yvanzo[m]
1,000 edits should allow them to add a lot of relationships if they are up to.
derat[m]
i'd say that approximately all of the messy edits that i've been seeing lately are userscript release imports from deezer or discogs
reosarevok[m]
Yeah, I guess you could block the release editor and be like "but you could add some relationships to your releases!" :p
aerozol[m]
p.s. I also like the idea of blocking edit pages because it gives us a full screen of real estate to leave a celebratory image thanking them for their work, and text and links to forums, voting etc 🥳
reosarevok[m]
Ok, I think we have some reasonable ideas of what to look into
Can someone document it on the ticket? I'm too tired 😅
