@yvanzo hi, I fetched these dumps : cover art, mbdata, in mirror mode, and then switched to dev mode to run server by changing the env var for compose file `COMPOSE_FILE=docker-compose.yml:compose/musicbrainz-dev.yml
`
I suppose this to work, I tried `sudo docker-compose exec db psql -U musicbrainz -d musicbrainz_db` in the db container spawned with the server, and the release group table is populated, but the server is not showing it.
huhridge has quit
Rotab has quit
huhridge joined the channel
huhridge has quit
huhridge joined the channel
huhridge has quit
yvanzo
yellowhatpro: Ok but it has only the cover art data, likely not the full RG data.
Why would this not be detected as a duplicate in prod, but be in my local setup
atj
my guess would be a weird charset issue
discordbrainz
<13Pratha-Fish> @reosarevok nothing special, I've just stopped getting notified by irccloud on PC as well as Android for some reason. At least the discord notifications are working tho.
mayhem
lucifer: do you happen to have some time to put in another rev on the lb radio tag endpoint today?
lucifer
mayhem: sure. can you remind me what changes are needed?
lucifer: now that I've finished the artist LB radio endpoint, I feel like I know what I really want from the tag endpoint.
monkey[m]: will do.
lucifer: the artist endpoint has a "max_num_recordings" option and then the API endpoint randomly selects that number of tracks given the other criteria.
that greatly reduces the amount of data fetched, transferred and ultimately wasted.
and then collapse all the entities into one list.
is that doable?
lucifer
sure, can do.
mayhem
thx
zas
lucifer: about hendrix service, that's http or uwsgi? Should new-oauth2/ removed from URL passed?
using the service name is a bit complicated at this point (it will require a consul template), so I'll use hardcoded IP & port for this test
lucifer
zas: uwsgi, new-oauth2 should be re-written to oauth2. sounds good.
Ah, thanks for looking at that failing test, it's driving me mad
mayhem
yeah, me too.
likely just a timing issue.
lucifer
reosarevok, yvanzo: does musicbrainz allow ?next= urls to redirect back to where the user started after log in?
yellowhatpro
@yvanzo I ran into some issue, but I think it was coz I was doing it all dirty way.
All I did was:
1. build db with caa schema (I was thinking to comment all the non-caa stuff from InitDb.pl but oh well, I forgot to do this)
2. Now I have caa schema, but additional schema as well (generated coz of InitDb).
3. Dropped unnecessary schema which will prevent createdb script failure coz of already existing schema and stuff.
3. Switched to dev mode, commented caa stuff from InitDb.pl, build the container, and created db.
Things worked fine, but there is slight (may be bigger?) issue. It seems the index_listing view is somehow dependent on musicbrainz schema, and it got deleted when I cascade dropped mb schema, coz it would create issue during createdb. So now I dont have index_listing view, and thus can't get release group (luckily this time it showed server error so I know things will work if I have index_listing).
Is there any cleaner approach? Or how should I improvise what I was doing.
yellowhatpro: Anything that allows you to debug the cover art priority issue for now is good enough. There is no clean approach anyway because the fetch script hasn’t been designed to download from both directories. I will look into improving this script to ease CAA dev setup.
yellowhatpro
np, I was thining to build the index_listing view after we have all the data.
If it fails, I do have other tricks too, will play with them then
yvanzo
Actually we should probably just include the CAA tables in sample data.
I mean just the rows with FKs related to the sample data of course.
bitmap, reosarevok: What do you think about this solution? ^
yellowhatpro
Yeah that was the last trick, when I first dumps, I tried similar with mbdump and others, but there was some URL error iirc. Didn't try with caa though
when I was exploring dumps*
building index_listing view worked. I now have the view, lemme spin the server and check if I can see page there
yvanzo: that sounds like it wouldn't hurt, in order to test CAA stuff better anyway
discordbrainz
<07munishk> @lucifer: could you please have a look at the PR for websockets tests, I have added the playing now test and assert statement for comparing the sent and received data https://github.com/metabrainz/listenbrainz-serv...
monkey: added the tooltip idea to my PR, take a look whenever you can!
monkey[m]
Will do
ansh (IRC): Hi! What did you do yesterday to resolve the problems you had with signing into apple music? I'm getting the same problem connecting + ERROR_FAILED_TO_VERIFY_JWT message running locally now
Got it :) I guess sign out and sign back in eh? Makes sense that it wouldn't like us signing into test.LB and localhost at the ssame time
ansh
monkey[m]: I created a new identifier and certificate from the apple developer portal.
monkey[m]
Ahhh, I see. Thanks
huhridge has quit
huhridge joined the channel
Satyaraj[m] joined the channel
Satyaraj[m]
Hey lucifer (IRC) , I'm getting this error on running my local env
Satyaraj[m] uploaded an image: (87KiB) < https://matrix.moviebrainz.org/_matrix/media/v3/download/matrix.org/MhYdiKORvIjCASAhmwYEeQgE/Screenshot%202024-04-16%20at%207.27.42%20PM.png >
lucifer
Satyaraj[m]: `./develop.sh psql` and run `ALTER TABLE user_setting ADD brainzplayer JSONB;`
monkey[m]
lucifer (IRC): Is there a way we could add a command to run the SQL update scripts sequentially ?
Sort of a "if you get an SQL error in local dev, run this command to update the DB" sort of solution
lucifer
monkey[m]: possibly yes but it would need some work to implement that.
monkey[m]
Fair enough
lucifer
can look into that later.
monkey[m]
No rush I waas just curious
bitmap
yvanzo: yes, dumping the CAA rows with the sample data makes sense to me
huhridge has quit
minimal joined the channel
huhridge joined the channel
mayhem
lucifer: did you start working on the tag search stuff?
mayhem is waffling about
reosarevok
Mmm, waffles
Arsen has quit
qookie has quit
mayhem
lucifer: now that I've dug in a bit more into the tag element, here is the improved result format that would be great: