#metabrainz

ruaok: i think its ready to be deployed. alastairp: sanity check please

My suggestion would be that you have a simple boolean field to flag an image as potentially offensive. Then you can have an profile option to enable/disable display of such images, with the default to not display for anonymous users.

looks good to me ruaok _lucifer

do we know what version of flask was used previously?

if it was on 0.9mumble then there _may_ be some incompatible upgrades, but I don't know the app enough to know if this was the case

not that old. I recently upgraded pacakges.

yeah, I see that the BU dependency is recent

mmm

yeah, looks like flask didn't change since v0.15, so looks good to me 👍

i have a dev instance up here http://bono.metabrainz.org:6500/ without login enabled if it helps

I'm more interested in what happens _after_ you log in

I'm just remembering back to places where they changed the flag on flask-login to know if a user was anonymous or not from a function to a property

that caused us no end of problems with various versions of BU floating around all of the projects

i can add the keys and we can test here before deploying in that case

it's OK, I just checked and we didn't upgrade any dependencies between 0.15 and 0.15

and MeB was running on 0.15 before. so it's fine

👍

[musicbrainz-docker] release v-2021-02-16 has been published by reosarevok: https://github.com/metabrainz/musicbrainz-docke...

ruaok: alastairp: MEB-117

MEB-117: Upgrade NodeJS https://tickets.metabrainz.org/browse/MEB-117

yeah, I know. I'm not the right person to dig into this. Its been on my list for a while now.

do you know how to do this?

it looks like the only dependency installed by node is lessc

and there's no build tool

i think upgrading is the easy part

so I'm 99% sure that it's going to be a matter of upgrading to 14 (latest LTS), and possibly upgrading the version of lessc

meb.org deployed and working fine.

testing to make sure everything works is harder

alastairp: so upgrade nodejs and the dependency. build image, run tests and visual inspection of the site?

yeah

I would go as far as to say if the css compiles, then everything is going to be fine

cool, i can probably do that.

alastairp: also, AB seems to be running an old version of PG.

what do you mean by "old"

i mean older than 12, which almost all other projects are using

I think we match it to the production version

it's a huge task to upgrade it, and we don't get anything useful out of a newer version, so it's fine for now

makes sense

I have tested sentry upgrade with AB, CB and LB. it works.

for meb website, there's this comment https://github.com/metabrainz/metabrainz.org/bl...

should i replace it with logging.info or the equivalent to capture_message in new sentry-sdk

ruaok: ^

if logging.info gets sent to sentry, then it's fine to change it to that

but if it doesn't then it should stay as capture_message

it depends on the configuration

if the logging level for sentry is configured to info, in that case this message won't be sent i think.

The default behaviour is to capture logging.info, but we configure it in LOG_SENTRY = { ..., level="OUR LEVEL"}.

I would say if we currently log all logger.info() messages to sentry then we can do that. if we don't, we should send the message directly

[musicbrainz-server] reosarevok opened pull request #1913 (master…MBS-11268-pt-2): MBS-11268 / MBS-11349: Improvements related to medium duration https://github.com/metabrainz/musicbrainz-serve...

alastairp: https://ci.metabrainz.org/job/listenbrainz-inte... the test to call our production labs api endpoint for recording lookup times out during testing. I personally think we should not be doing that in running integration tests. we should have a test for fetching the metadata with mocks and all that, but we should disable the metadata fetching during testing since it adds nothing to the tests and hitting a

production API for testing is questionable in the first place. Thoughts? How would you disable this?

oh, I see. the logical way. :)

ruaok: this is LB-782

LB-782: Docker networks in listenbrainz/jenkins unit tests don't allow external network requests https://tickets.metabrainz.org/browse/LB-782

I got around it in a few places by mocking

in this case I'm adding a flag to get_paylist with fetch_metadata, which allows the tests to pass a False flag in order to skip this issue for the one test we have failing now.

ruaok: did you get a chance to look at stripe.com/climate ? I'd love to get your opinion on us participating.

oh, I wanted to figure out what that number was for 2020, but then got angry at QuickBooks for never allowing the sorts of simple queries I want. let me see if I can do that after I submit this fix.

yeah, I think we could even mention that x% of donations via stripe go to carbon capture. wonder if that'd increase donations :)

[brainzutils-python] amCap1712 opened pull request #56 (master…sentry): Upgrade raven to sentry_sdk https://github.com/metabrainz/brainzutils-pytho...

mention on https://metabrainz.org/donate, i mean.

I also need to move to the new version of checkout. EU cc payments are likely failing right now.

oh duh. I should ask stripe, not quickbooks.

:D

i see we save a lot of payments info in the metabrainz.org DB.

like addresses etc.

if we start doing checkout, i guess we'll have to write webhooks to fill that in, not sure.

yes, addresses are needed in case the IRS wants to come snoop.

oh joy.

we could also save customers in stripe and run a script to call the API and fill the data in every day or so, i guess.

looking at it, the webhook is probably less work https://stripe.com/docs/payments/checkout/fulfi...

$20,859 in stripe transactions last year. 1% of that would be $208.

+++

nice.

ruaok: alastairp: i think that listenbrainz_spark is using flask (and BU) only for logging and loading config values. but the sentry_sdk comes with a PysparkIntegration so its easy to use it directly instead. that way we can remove the BU dependency from spark. do you think its worthwhile to make the shift?

big yes

Flask there makes no sense, so please!

👍

sounds good

we should upgrade our base images

where do we accept float input? in AB I guess, but that is still on py 2.x

[listenbrainz-server] mayhem opened pull request #1288 (master…playlist_collaborators_test_fixes): Playlist collaborators test fixes https://github.com/metabrainz/listenbrainz-serv...

Mr_Monkey: alastairp ^^ see what you think, but I hope that should improve testing and fix the failing test.

ruaok: and even the floating point data in AB isn't a single user-provided input

that's a fun bug

woo. all tests for 1288 pass. thoughts on the fix, alastairp ?

If we test the metadata fetching in another way, as you suggest, then this does look good. Thanks for looking at it, I was scratching my brain at those test failures yesterday.

well, the previous method didn't test the metadata fetching either. :)

let me merge that branch into yours and then add a test for the metadata fetch.

then hopefully we can merge this finally.

Sweet !

[listenbrainz-server] mayhem closed pull request #1288 (master…playlist_collaborators_test_fixes): Playlist collaborators test fixes https://github.com/metabrainz/listenbrainz-serv...

🤞 at tests

gah. not sure I care to write this test. this test will soon be thrown away when we add the more nuanced service layer that is missing. we're not losing a test, so I'm all for merging and moving on.

Mr_Monkey: there are open conversations on PR 1276. have a look, please?

OK

All converstions closed, pushed a small docstrings change.

it would be nice to add another couple of tests...

wait, I take that back. what I was thinking would already be covered by other tests.

integration tests passed. woo

[listenbrainz-server] MonkeyDo merged pull request #1276 (master…playlist_collaborators): Allow playlist collaborators to access private playlist https://github.com/metabrainz/listenbrainz-serv...

_lucifer: update your DB connect strings to match what is in config.py.sample

I am having trouble with timescaledb initialization. How can I remove the existing vdata?

ruaok: i have done that already

docker volume ls

docker volume rm <volumename>

yvanzo: is there a new timeline for rabbitmq deployment?

removing the volume worked. thanks!

np

ruaok: I just started merging PRs, migrating sir-test first.

my production image just passed all its tests, let me beta deploy. I should theoretically be ready to migrate in 15 minutes or so.

Mr_Monkey: woo, I finally was able to add a track to your collaborative playlist on beta. woo

ok, beta looks good.

Nice !

ruaok: I completed sir-test migration, getting ready for listenbrainz

I dont have a lot of time left today. :( how long do think it will take?

ok, sorry, about ten minutes... if nothing goes wrong.

lol, ok.

pushing the prod image very soon.

what is the env var value I need to set>

?

is `export VAR=value` good enough or does that need to happen in services.sh?

there is one more thing to do: configure vhost/exchanges/queues at the new rabbitmq instance.

export RABBITMQ_SERVICE="..." yes?

yvanzo: ok, what needs doing?

no, is listenbrainz doing it automatically? for example sir has https://sir.readthedocs.io/en/stable/setup/inde... https://github.com/metabrainz/sir/blob/master/s...

that looks right, I think. but I am not a rabbitmq expert.

there is https://github.com/metabrainz/listenbrainz-serv... too

These are current LB queues: https://gist.github.com/yvanzo/0baf0103ebb64646...

hmmm.

ten minutes have passed and we've not started doing anything yet.