#metabrainz

i branched from oauth logout2 branch

but if both were merged to master, github should “hide” those commits which are already in master

i too wondered that

nothing really important, but it might make the work easier

yup

anyway, good feature. I run it but I still have to check that it works as expected

in future if i have an unmerged feature and start working on another one, should i branch from it or the master

zas: done

ruaok: artists which have actor names included would have 0 similar artists, no?

it depends: if Second feature depends on First one, you might branch from First. But try to avoid it as to keep PRs as simple as possible (and adding a Description with what you wanted to achieve would help, too)

often yes. but they can appear on a different compilation with the same artist credit and the count could be more than 0.

ok i'll keep that in mind

i think zbar would work well for barcode

i'll start work on it one preferences pr is merged

Umm....Okay. i will stick to eliminating te lines which have 0 similar artists

great

the ones*

good, I'll review right now then. So you can start pristine

And should find out the reason why pink floyd shows up so many times. And many other artists too

An even distribution would be better.

zas: still have to update database triggers, and to rebuild some indexes: artist, cdstub, label, release

zas: added label isni triggers, stopping sir-prod and rebuilding search indexes

ok, good

I think 7.7 is better at caching, it seems cpu load decreased a bit while more (caching) memory is used. Also document caches hit rates seems to have improved. We'll see in next days if it confirms though.

that sounds promising!

cpu load decrease might also be due to https://issues.apache.org/jira/browse/SOLR-13349

yes, very likely

I have to run a few errands, afk for one hour or so.

ruaok: Did you consider using a relative compilation appearance instead of absolute? E.g., {number of appearances together}/{sum of compilations that either artist appear on}?

yes. the scaling of the result is intentionally not done in an effort to not reduce the detail (range) of the data on purpose.

the consumer (in this case pristine) should decide how to do that.

Does the consumer have access to the {sum of compilations that either artist appear on} if they want to try using a relative weight?

that would need to be computed still.

amCap1712: I've reviewed with a comment for you to think about

gr0uch0mars: i saw it

i think that is acually not actually an issue

when a user logs in from the browser, it is asked whether to remember it or not

and also the login is in browser, we cannot control the cookies of browser

however for safety purposes we could add a warning to logout from browser as well. what do you think on this gr0uch0mars

no no, it might have been perfectly the browser. In that case we could think about how to alert the user about that (in case someone tries something like what I did)

But I don't think it's prioritary. Usually people will login with their account only

ok gr0uch0mars

lol. iliekcomputers are you going to present that paper that you submitted? accepted 4 days before the conference?

done reindexing artist, label, cdstub, and release (other search indexes did not change)

hi alastairp, after we discussed yesterday I decided it would be easiest to just put all the work in one PR for now - that way you can view it and try running it yourself without jumping between PRs, and I can just continue adding to the branch?

aidanlw17: the problem with that is that it may be hard to follow such a large PR.

I would urge you to find ways to submit smaller PRs and as frequently as you can.

Mineo: Can you please add me to maintainers of sir.readthedocs.io so I can update it?

letting things ball up only creates problems down the road.

ruaok: noted, I split it up if that will be easier for us to follow and work with.

*I will

yes.

and if you get a smaller PR accepted and merged before the mid-terms, you have *something* to show for your efforts.

Okay! so it's important that there is finished work submitted prior to our first evaluation

I'd say so yes. otherwise we still don't know if you're a bullshit artist. 😝

not that we have anything against BS artists, we have a few here as well. Right, reosarevok ?

Fair enough 😅 lol I swear I'm not a BS artist, I'll make some smaller PRs to prove it

good plan. we'll leave the BS to the professionals.

exactly!

I guess by now I am a BS professional, but I wouldn't call myself an artist to be fair

BS craftsman at most

BS enthusiast.

zas: https://stats.metabrainz.org/d/000000078/ should be updated to monitor sir-prod that replaced sir-beta

I'll do it tomorrow, containers whitelist needs to be updated and redeployed

ruaok: Can you please upgrade my BS test vm up to 4 vCPUs? It should probably be the setup for the next MB VM.

zas: OK, thanks :)

I'm out for the evening... Zas can do it...

Cyna: we can create a separate branch for your gsoc to merge things into

yvanzo: which VM ? at hetzner or gcloud ?

I guess that's yvanzo-musicbrainz-docker-test,upgrading to 4 vcpus