#metabrainz

<bitmap[m]> "sent the latest SAML logs" <- could you please do one more attempt w/ trace? i wouldn't be surprised if there's still more work to do on my end before working state, but this round should help me tell if i'm on the right track

<derat[m]> "reosarevok, bitmap: i noticed..." <- I suspect that was a mistake since those rels are pretty much identical to each other. I'll let reosarevok look into it since he added them, but I agree with your assumption that `type0-type1-name` should be unique and something is wrong there

<julian45[m]> "could you please do one more..." <- unfortunately I get the same "Couldn’t sign you in" message from Google as the last run, sent your the latest trace

s/your/you/

<bitmap[m]> "unfortunately I get the same "..." <- thanks! getting closer -- one more shot (again w/ trace) if you don't mind?

julian45[m]: this time it didn't take me through kandim at all, just directly to the "Couldn’t sign you in" page :)

some more progress as shown in traces! made some more config changes -- once more for tonight please, then let's call it

<julian45[m]> "some more progress as shown in..." <- sorry for delay, got the same as last time again

np, thanks for still trying

<derat[m]> "reosarevok, bitmap: i noticed..." <- Ah fuck, I guess I entirely missed the existing one and added a new one rather than just moving it... no, very much not intentional, I'll look into deleting one

Phew, luckily one of them had no uses yet, so I could remove it safely

Thanks for letting us know, derat! If you find any other dumbs I've dumbed, scream at me :)

Upgrading ChatBrainz.

lucifer_: is it normal that I get a log in for chatbrainz now, but it rejects my creds? :D

reosarevok: yeah i am working on it, i upgraded the pg version and restored the backup but it didn't succeed.

Ok :) No worries, just making sure it wasn't my fault!

hi!

it seems the discord bridge didn't come back up. I'll look into that but if there are any other issues let me know.

lucifer: I'm getting "Unable to decrypt message" in old PMs even after verifying the session

reosarevok: can you try signing out and signing in again?

Sure

<11techmorningstar> hey again

huh weird, the old discord bridge was still running on the server but the new one is missing.

<bitmap[m]> "sorry for delay, got the same as..." <- aaaaand time for one more shot w/ trace please. (got in touch with the person who drafted the proxy config i worked from, who pointed out a possible issue w/ a config line for one of the tools the proxy uses)

tbh if we don't have much more luck soon (though i might try setting up a test account myself), maybe kanidm + this proxy won't be the best move for MeB. not much use for the potential niceties like ssh key distro + linux integrated auth if we can't get it to handle relatively basic UX cases well

still getting the "Couldn't sign you in" page :(

argh!

thx

testing - 3

not typing anymore

okay seems all bridges are linked again

Upgrade complete, let me know if you face any issues.

bitmap, reosarevok: I plan to release an Harmony update soon which updates the external links of already existing releases, for which MBS-13993 would be very useful to have.

MBS-13993: Automatic deduplication of seeded external links https://tickets.metabrainz.org/browse/MBS-13993

Thanks for volunteering! 😍

I know that I could implement deduplication in Harmony, but it would require a change in the MB API include parameters. That would be a breaking change for permalinks or a lot of work to mitigate that, which I would like to avoid 😅

I know bitmap was looking into refactoring that whole link code

Not sure how doable it would be to improve deduplication while at it - honestly, possibly less of a nightmare than the refactoring itself, but

I theoretically have some code fixing small issues (but not the main one ) in https://github.com/metabrainz/musicbrainz-serve... - how much of that is still relevant and even more, how much would be relevant post-refactoring, I dunno

(it might be best as something for bitmap to just consult in case it's useful during that process)

In any case, I guess my answer is "I know it sucks now and we should improve it, but unless bitmap does it, I'm not sure it'll happen soon"

(I'll let him update you on the progress of that refactoring if any)

Yes, I've seen your PR. Having at least that one released would be great, otherwise I would rather not release the Harmony feature and deal with the complaints about the seeded duplicates 🫣

I wish I had decided to include "too many things" (isrcs+url-rels) back when I implemented the MB provider for Harmony, then I would be able to detect existing ISRCs and URLs now 😞

Now I either have to implement something like a cut-off timestamp after which a different API URL format is generated for permalinks or I have to support fallback URLs in my caching library.