ON DELETE SET NULL; + the check on the review table means that if we delete a license (for whatever reason?!) the check constraint is going to fail on all of the reviews with that license. is that right?
2022-04-06 09621, 2022
lucifer
alastairp: just close and reopen or create a new branch and new PR?
ah well but we'd keep existing reviews of that license so not deleting it from table.
2022-04-06 09659, 2022
alastairp
yeah right, and we should consider upgrading to CC4.0 licenses anyway - which means that we should mark all of the other ones as "hidden" or something anyway
2022-04-06 09613, 2022
lucifer
there's probably another detailed ticket around which i forgot to link to this.
2022-04-06 09617, 2022
lucifer
yes makes sense
2022-04-06 09625, 2022
alastairp
it'd be nice to have some further explanation/motivation for that ticket :)
2022-04-06 09624, 2022
yellowhatpro
Hello akshaaatt , for the BP implementation what are your views on the Media3 libraries??
2022-04-06 09624, 2022
yellowhatpro
It includes Exoplayer media player that seems to be useful for our purpose
maybe 1 minute faster? but with just one run, not enough data to reach a conclusion.
2022-04-06 09643, 2022
lucifer
not much benefit for full case i see.
2022-04-06 09636, 2022
lucifer
for the incremental case, amount of data needed at the end is limited so order of joins probably matters more. in full case, almost everything is going to be pulled so probably not much difference anyway.
I'll have the call in for even the full DB case, just in case.
2022-04-06 09612, 2022
mayhem
I'll push that commit and then merge.
2022-04-06 09615, 2022
lucifer
👍
2022-04-06 09622, 2022
alastairp
any idea why this manually sets form.errors instead of using the form validation (this code predates your update, so it's a general question)
2022-04-06 09641, 2022
alastairp
I guess we need both the form and the review which we migt not have in the form instance?
2022-04-06 09617, 2022
lucifer
i think because we want to only enforce this conditionally. and the review and current_user maybe unavailble in form validator.
2022-04-06 09635, 2022
alastairp
yeah. I see that we could pass in the review to the form constructor, save it in self, then pop it off of kwargs before calling super.init
2022-04-06 09654, 2022
alastairp
but that much work for simply moving a check from one file to another, not sure that's useful
2022-04-06 09620, 2022
lucifer
i'd leave it as is for now, we can improve that in future.
2022-04-06 09641, 2022
alastairp
yeah
2022-04-06 09627, 2022
lucifer
i have been having hard time with #327 with this similar issue actually, the validator doesn't add error to form due to some reason. i had to update to latest wtforms as well to fix other issues.
2022-04-06 09642, 2022
lucifer
once, these PRs are merged i'll rebase and try again.
alastairp: yes assuming random distribution of characters, we could model this as probability of collision of uuids of say length 12. i'll need to look up how to calculate it exactly but knowing about the birthday problem my guess is its likely the number would be lower than one would expect.
2022-04-06 09626, 2022
lucifer
thanks you too for doing the much needed CB work! :D
2022-04-06 09610, 2022
lucifer
the LB PR, i am unsure if we can finish it tomorrow. my vague recollection is that only tests are needed but could be wrong.
2022-04-06 09655, 2022
monkey
Happy to help with the merge conflicts on #1605 if it's needed
2022-04-06 09634, 2022
bitmap
reosarevok: yes MBS-11755 should be solved too, I'll link the tickets
alastairp: ah i found someone who did the calculation, https://en.wikipedia.org/wiki/Birthday_problem (Probability section) so for length 12 theres 75% chance of collision in 2.8 * 10 ^ 7 elements which is also approximately the number of recording mbids we have.
2022-04-06 09608, 2022
bitmap
reosarevok: re: "I guess that updating of the dodgy existing data will actually have some effect on mirrors?" the upgrade script runs on mirrors, so they will get fixed tag counts too, if that's what you mean. I'm pretty sure we specified that in the blog
2022-04-06 09635, 2022
bitmap
reosarevok: should we/is there a way to mark MBS-12141 as "release pending"? I remember seeing that option...
alastairp: interesting read. it goes upto even storing the TZDB version in each entry!
2022-04-06 09611, 2022
lucifer
but those challenges have to deal with future times so thankfully one we can do without.
2022-04-06 09619, 2022
reosarevok
bitmap: why not just "in development branch" if it has been merged to a development branch?
2022-04-06 09632, 2022
reosarevok
Oh, I guess it might accidentally be moved when releasing or something
2022-04-06 09610, 2022
yvanzo
zas: For some reason /var/opt/solrcloud-collections-backup belongs to zookeeper on mb-solr-7. It should belong to solr instead, as on other Solr hosts. Changing its ownership as root failed with: Permission denied. Opening a shell as zookeeper failed too. Any way out?
2022-04-06 09653, 2022
zas
let me check
2022-04-06 09600, 2022
zas
weird, cannot cd as root...
2022-04-06 09635, 2022
mayhem
uh oh
2022-04-06 09654, 2022
BrainzGit
[musicbrainz-server] 14reosarevok opened pull request #2478 (03schema-change-2022-q2…mbs-11760-delete-triggers-before-creating): Delete new triggers if exist before creating https://github.com/metabrainz/musicbrainz-server/…
2022-04-06 09621, 2022
reosarevok
yvanzo, bitmap ^ I think we need that actually, my bad
2022-04-06 09603, 2022
zas
yvanzo: that's because those are nfs mounts, and uid/gid aren't the same on all machines
2022-04-06 09604, 2022
PrathameshG[m]
Hey mayhem
2022-04-06 09635, 2022
PrathameshG[m]
CC alastairp lucifer
2022-04-06 09643, 2022
PrathameshG[m]
You've Gotta be very busy with the high influx of GSoC proposals & candidates incoming your way, but if you could find some time, could you give some feedback on my proposal?
2022-04-06 09643, 2022
PrathameshG[m]
I think there's still a lot that I could improve upon. Just need your specific criticism to address the most important issues first
2022-04-06 09656, 2022
yvanzo
zas: So we should have the install scripts to specify uid/gid when creating solr/zookeeper users?
2022-04-06 09622, 2022
zas
usually this is handled by id mapping
2022-04-06 09642, 2022
zas
but for some reason it doesn't work right now, not sure why
2022-04-06 09651, 2022
zas
ok, it will need few more steps to work properly
2022-04-06 09625, 2022
yvanzo
There is no hurry, it won’t run again before Sat at best.
tl,dr; we think the project's goals (not just the proposal) need to redefined, it probably won't be possible to get a production ready app by the end of summer. we want to lay more emphasis on experimenting and finding flaws or lack threrof in the data.
it was a long discussion and we agreed on some issues but not on all. we can probably discuss it again tomorrow and then either me, mayhem or alastairp can post a coherent comment on how the proposal/project should be redefined.
2022-04-06 09655, 2022
zas
but apparently there's another issue, still get perm denied
2022-04-06 09654, 2022
yvanzo
yes, even as solr
2022-04-06 09659, 2022
zas
when those shares were set up ? did they work before?
2022-04-06 09624, 2022
zas
or you were trying to set those up right now?
2022-04-06 09625, 2022
Sophist_UK has quit
2022-04-06 09608, 2022
Sophist-UK joined the channel
2022-04-06 09643, 2022
yvanzo
zas: It did work on March 26th at least.
2022-04-06 09625, 2022
yvanzo
zas: It has been set up one year ago and it was working until then.
2022-04-06 09645, 2022
zas
very weird then
2022-04-06 09656, 2022
yvanzo
Backup did fail for some reason (might be unrelated) on March 30th.
2022-04-06 09620, 2022
agatzk has quit
2022-04-06 09623, 2022
yvanzo
This share is used twice a week: backup is written, then downloaded and removed. the whole operation takes a few hours at most.
2022-04-06 09602, 2022
aerozol
moin!
2022-04-06 09612, 2022
yvanzo
hi aerozol!
2022-04-06 09656, 2022
akshaaatt
Kia ora aerozol!
2022-04-06 09619, 2022
aerozol
:D
2022-04-06 09634, 2022
zas
yvanzo: I'll have a look tomorrow morning, at this point no idea. nfs is an heavy source of headaches...
<lucifer> "it was a long discussion and..." <- Thanks, I took a look at the discussion you guys had earlier tomorrow, and it was really helpful. Looking forward to your comments tomorrow on the updated proposal. :))
