Thanks for the feedback on the homepage mayhem - easy enough changes, I’ll put them through
monkey: You’ll definitely get a doc or a ticket from me for the final design, but if you think it looks okay so far I will just focus on getting this done for now, and do the ticket for implementation at the end. If you want a more detailed in-progress look let me know and I’ll export to a doc/ticket any time
P.S. not sure if this is relevant info re, the missing LB tracklists, but users were reporting that submitting listens from an album could trigger tracklists to be added shortly. So whether or not the issue is transient might get muddied by the fact that eventually someone submits a listen for most albums
minimal has quit
bitmap[m]
<lucifer[m]> "bitmap: zas timescale writer..." <- is that the "listenbrainz-timescale-writer-prod" container on kiss? (because the last log entry for it is from 2024-11-03, and I was able to connect to v2.pgbouncer-master.service.consul fine from inside it)
d4rkie has quit
d4rkie joined the channel
d4rkie has quit
d4rkie joined the channel
Kladky has quit
Maxr1998 has quit
Maxr1998 joined the channel
Maxr1998 has quit
Maxr1998 joined the channel
pite joined the channel
pite has quit
d4rkie has quit
d4rkie joined the channel
d4rkie has quit
d4rkie joined the channel
Kladky joined the channel
mayhem[m]
<aerozol[m]1> "P.S. not sure if this is..." <- yeah, I understand the problem now. tricky thing to solve, but I'll have a think.
texke has quit
texke joined the channel
monkey[m]
aerozol: The mockup looks fine to me, I think indeed this is a good solution if we don't want to entirely rework the page.
yvanzo: any luck figuring out the translations test issues?
bitmap: I just noticed (again?) derat's comment on MBS-13561 - do you have any suggestions on how big an extra load this is and whether it would cause any issues?
I'd love your opinion on this - do you feel it makes sense to print plays / listeners data in MB? If so, what's the most friendly way of getting those? (I'm assuming hitting an endpoint and caching it for a day or so would work for general playlists, but not sure how that would work for plays by the current user tbh without hitting LB each time)
s/playlists/plays/
yvanzo[m]
reosarevok: I’m not investigating this issue.
reosarevok[m]
So should I or should I not translate further? 😅
You said we so I thought you were looking into it, if not I can try to figure something out I guess
minimal joined the channel
monkey[m]
reosarevok: I think it might make sense to surface some data from LB on MB pages, but I'm biased I guess. It might move some users to check out LB features, while also adding some relevant data for those users who might be interested in listening/popularity stats.
Not sure who, but I'm sure some users will oppose this addition. I'm curious to hear arguments against it (other than "I don't care about seeing them")
My main argument against it is that the play counts are not especially accurate and there are some issues with it. But those will be resolved in due time
monkey: reosarevok: Even without being a funnel to LB (if we’re being cynical), I think listening stats add value to MB in terms of giving an indication of artist popularity/how known an artist is. As a MB editor I would find that useful. Coincidentally some users were just recently asking for MB filters that would make use of those kinds of stats: https://community.metabrainz.org/t/what-do-you-...
SigHunter has quit
mayhem[m]
popularity is something often desired in MB -- at least it used to be.
similar artists also makes loads of sense
ajhalili2006 joined the channel
aerozol[m]1
I think there will be minimal pushback if it is implemented in the way in the mockup (basic text), as long as it doesn’t mash loading times or something. The current LB button is quite large - I suspect that if we added LB stats and as an exchange made the “play” button smaller then even the cynics would be happy? Jesus2099 isn’t on here is he? He made a “hide the LB link” script, after all
SigHunter joined the channel
minimal pushes back
I will assume that was in service of some excellent wordplay rather than not wanting LB stats anywhere in MB :P
minimal
;-)
aerozol[m]1
P.S. I’m going to go ping jesus2099 on the forums to see what he has to say
bitmap: hi! while trying to connect to critiquebrainz_test_db, i get the following error: `sqlalchemy.exc.OperationalError: (psycopg2.OperationalError) connection to server at "10.2.2.25", port 65436 failed: FATAL: no pg_hba.conf entry for host "172.17.0.1", user "critiquebrainz_test", database "critiquebrainz_test_db", no encryption`. i do see it in the config.json but not in pg_conf.hba on postgres-hendrix, is the new config
updated on production?
bitmap[m]
<lucifer[m]> "bitmap: hi! while trying to..." <- oops, there was step missing from the syswiki instructions, can you try now?
Kladky has quit
lucifer[m]
bitmap: works now, thanks!
leonardo- joined the channel
saumon has quit
HSOWA has quit
leonardo has quit
murr4y has quit
HSOWA joined the channel
mayhem[m]
lucifer: I was thinking about the following solution for the tracklists problem: Create a new function that calculates *only* the canonical releases table and nothing else. This normally executes in about 1 minute. We can rotate only that table into place and then run that maybe 4-6 times a day. As a stopgap measure, this might be ok. I still need to check to make sure that there is no code that would break if the id sequences change,
but I think we search that table by MBIDs, so it should be fine. What do you think?
lucifer[m]
mayhem: i think that might work but i need to take a look at the tables and their interactions. because i am not currently sure if it only that table or more tables that need to be update with higher frequency.
mayhem[m]
agreed, we need to look at those details.
saumon joined the channel
murr4y joined the channel
lucifer[m]
aerozol, monkey, reosarevok , mayhem : test.critiquebrainz.org is now running on test.metabrainz.org oauth provider now, please test and let me know if anything needs to be fixed or changed.
mayhem[m] uploaded an image: (34KiB) < https://matrix.chatbrainz.org/_matrix/media/v3/download/chatbrainz.org/PtEyGUESWBNIqXtYYPqhVnOf/image.png >
mayhem[m]
should this now say metabrainz?
it does log in with musicbrainz. is that intended?
so this oauth workflow is "backend only"? meaning the UI hasn't changed?
lucifer[m]
i think for now, letting it say musicbrainz is right because users have to register on musicbrainz, when we move user registeration to meb.org then it would make sense to change the text here.
mayhem[m]
ok, user registration is what I mean by user bits. ok, sounds good. for now clearly no changes needed.
lucifer[m]
the UI has changed for approving access to apps, revoking access to apps, creating new oauth apps. etc. but not for login or registeration.
mayhem[m]
logging in and out works as expected. what else needs testing?
lucifer[m]
that's mostly it i think.
mayhem[m]
well done, seems to work fine!
lucifer[m]
there's a couple of more things to be tested but i think we need to do LB for that. CB doesn;t use those features.
rating stuff in MB using the tokens issued by MeB OAuth provider for instance.
this is likely as much as we can test before putting the oauth provider in production.
