Mr_Monkey: I am continuing with PR #408 making myself comfortable with end to end tests. I have even written some basic tests which I would show you once they are done
Mr_Monkey
đź‘Ť
Chinmay3199
I think I should continue with tests only. Thanks
ruaok
iliekcomputers: how does one deal with merge conflicts in package-lock.json in LB? its a generated file and I've got a massive conflict while trying to rebase the timescale branch
alastairp
what's the url for granfa again?
Mr_Monkey
ruaok: if you installed a new package in the branch, then use the branch's package-lock file. If not, then use the master file
stats. that's it. I tried statistics, logs, granfa
ruaok
Mr_Monkey: ok. thanks.
Mr_Monkey
Cyna: So you could do `const relationshipSet = await author.relationshipSet()` followed by `const relationships = relationshipSet.relationships()`, or…
alastairp
zas: hi. can you explain to me what kind of http log stats we can do on granfa?
Ideally I'd like to be able to see how many POSTs we get, as well as how many GETs to a specific URL prefix
not urgent at all, it's just the next thing on my list that I was looking at
shivam-kapila
iliekcomputers: Hi. Are you here?
alastairp
would you like an issue on mbstats, a ticket on jira....?
zas
create an MBH ticket on Jira
shivam-kapila
ruaok: Hi. There is still some timing issue between timescale_writer and RabbitMQ. When we run `develop.sh up` by the time it starts the timescale writer, the rabbitMQ pipeline is still not up. So the writer exits. Did you get me?
zas
basically that's adding the method field to special log format, and adding parsing code so we can output this to influx
ruaok
yes, I've been trying to figure out what is going on there, shivam-kapila.
it breaks when do I an import,but not when i record a single listen.
alastairp
Freso: MBH-292 looks fixed to me. can you confirm?
ruaok: Might be possible that just after writing 1st listen it exits.
ruaok
using the submit.py script that is in the docs, I can enter hundreds and the pipeline works.
I've been working to rebase the timescale branch on master. I stopped trying to use rebase and I've made a new branch.
the merge was much cleaner, but now there are two tests failing that need to be resolved.
but they are minor. let me look at timescale_writer exiting
shivam-kapila
Oh. I can have a look at the tests. I haven't tested much as that stupid brainzutils error took my whole weekend.
ruaok
ok, look. at `timescale-rebased`
shivam-kapila
Okay
iliekcomputers: Please ping me when you got some time. The brainzutils error is finally done. I need to know your views on my approach.
iliekcomputers
shivam-kapila: hey
can you open a PR and I'll take a look
?
shivam-kapila
I will
iliekcomputers
thanks
base off master, because from what I understand the issue should be in master too
(so as to not complicate the timescale branch with an unrelated issue)
shivam-kapila
Actually I just used pip install --upgrade instead of pip install in requirements.txt. Messybrainz is based on brainzutilsv1.10.0 which didnt have the methods in cache.py
Should I use --upgrade flag with all pip installs. I had this doubt.
iliekcomputers
I do not think that is a good idea
that would upgrade everything to the latest version which would lead to breakages
shivam-kapila
I am afraid it might break some code
Yes. Same
iliekcomputers
let's just upgrade brainzutils in messybrainz
shivam-kapila
Will that be able to solve issue? Bkz if someone is using the projects already, they will have older versions of brainzutils and the docker build will skip upgrading it in that case. It returns `Requirement already satisfied`
I saw that my LB is based on v1.8.0(I dont know how even when MsB was on v1.10.1) and it didnt upgrade that.
I searched on livegrep from where its getting v1.8.0 but no use
iliekcomputers
we're trying to fix the problem in production first, and i'll rebuild to fix it
alastairp
shivam-kapila: docker build doesn't work like that. If you update requirements.txt then the dependencies will be installed again from scratch
iliekcomputers
someone needs to make messybrainz a package in pip
:D
the github clone stuff is not ideal
shivam-kapila
+1
MsB and Brainzutils should be in pip.
alastairp: But why didnt I upgrade to 1.10.1 from 1.8.0? Did I do something wrong.
iliekcomputers
livegrep has been very behind
alastairp
I don't know the specific steps that you did, but that should be the case
iliekcomputers
there are no typescript files in there yet
we merged those like a week ago
shivam-kapila
Yeah. I noticed that too
alastairp
however, if you're talking about listenbrainz + messybrainz + dependencies, then there might be some additional complications
shivam-kapila
alastairp: There are so many dependency hierarchies
alastairp
iliekcomputers: it might be interesting to see how to put messybrainz in lb's requirements.txt, which would remove the git clone stuff
here's a radical idea: merge messybrainz into listenbrainz, it isn't used anywhere else, there's lots of duplicated code. there'd be one requirements.txt so no conflicts, no need to do the git clone OR pip install :D
alastairp
yeah, that's not a crazy idea
originally messybrainz was a separate webservice, but for reasons that was stopped
ruaok
can't really argue against it.
alastairp
so, put it as a separate package in lb and remove all the webservice stuff
Zastai
CatQuest: transclusion seems to be based on a specific table (https://beta.musicbrainz.org/admin/wikidoc) - your page is not in that list, which I think is why it has "This page has not been reviewed by our documentation team" at the top
iliekcomputers
+1
ruaok
I'm ok with that.
Zastai
CatQuest: I'm not sure how update propagation works for page not in that table; it probably requires a "transclusion editor" to push a button
BrainzGit
[musicbrainz-server] mwiencek opened pull request #1475 (master…csrf): MBS-10717, MBS-10778: Set SameSite=None on session cookie and add CSRF tokens to user/admin forms https://github.com/metabrainz/musicbrainz-serve...
shivam-kapila: huh, I just saw that - I don't know why we install messybrainz both from requiremets.txt and using git clone in the Dockerfile. that's definitely wrong
iliekcomputers
i haven't looked at that stuff in a long time, but i have a feeling the git clone is needed because we create the messybrainz schema in dev which needs the sql files
alastairp
ah, that might be the case. however, if that was the reason then an alternative solution would have been to install the sql files into the package when installing from pip
>It looks like RabbitMQ is closing your connection on you, perhaps because you're redeclaring an exchange that is already declared with different parameters?
could this be happening somewhere
ruaok
interesting. what is weird about it is that it happens after running fine for a while.
alastairp
zas: FYI I deleted 50GB of data from boingo, in case you have reporting and noticed something funny
Mr_Monkey: I have a suggestion for entity.js file.. We have to all tests for all the functions in entity.js to ensure they are working properly. It would help avoid errors and provide clarity
For example, getEntityByBbid() could surely use test case to verify cases like if entityBbid exist or not
chaban
!m bitmap for adding CSRF tokens
BrainzBot
You're doing good work, bitmap for adding CSRF tokens!
Mr_Monkey
If you're suggesting adding more tests where tests are missing, I 100% agree !
As for the length, try seeing what authorRelationships.toJSON() gives you
People up for reviews: yvanzo, bitmap, pristine__, ruaok, Mr_Monkey, Freso, zas, alastairp, reosarevok, shivam-kapila, _lucifer, Cyna,CatQuest, ishaanshah[m] – anyone else who wants to give review, please let me know ASAP. :)
yvanzo: Go!
yvanzo
Hi!
Last week, I released a new version of MBS and updated beta.
I finally merged, deployed, tested, fixed, and redeployed MBS-10546.