Android apps can easily be decompiled. Any user could get our refresh token, which is a concerning thing
jasje
akshaaatt: no idea what refresh tokens do :(
akshaaatt
As the name suggests, refresh tokens allow us to refresh the access tokens for a user
access tokens live for a short period of time
Hence, we use refresh tokens to keep the user signed in to the app at regular intervals of time
jasje
when does mb token expire?
akshaaatt
For example, access tokens can live for 1hr but a refresh token can live for 30 days
monkey
lucifer, mayhem : Looks like our LB API docs cannot build since LB#2360 due to an incompatibility between the newer versions of werkzeug (> 2.1.2) and sphinxcontrib.autohttp
I have no idea if that is a good way to resolve this sort of python dependency issues; if you don't mind I'll leave that with you. I did however create a ticket for it with all the details I could find: LB-1239
Hi nexus ! We already use Enzyme in our test suite
However, we will be moving away from using it, as it is pretty much abandoned
For now though, that's how we're testing react components
nexus
Oh right.. okay then..i will try to not use it as you are anyways not using it in future
akshaaatt
I haven’t calculated the exact math yet jasje
Do it as suits you
monkey
In preparation for our future move from enzyme to react-testing-library, if possible, try to avoid setting or asserting the state or props directly, and instead emulate user actions and then assert that the UI is doing what is expected
nexus: I think you'll have a very hard time without Enzyme ! Do use it, but prefer testing UI and user interaction rather than internal state of the component :)
nexus
Sure ..i'll try
monkey
The entire codebase is nothing but tests of the internal state of the component, so it's not a great example to follow
We do have a couple of tests I can point you to however. Give me one minute
Since it tests a functional component that uses hooks, we cannot touch the internals like the state.
nexus
thanks
that sure would be a lot helpful
jasje
akshaaatt: alrighty
zas
yvanzo: in 4 minutes I'll apply changes and reboot
ok?
yvanzo
zas: no :D
16:45
zas
ah UTC....
yvanzo
yes
zas
:)
yvanzo
I checked every container, it seems fine to have brainzbot down for a few minutes, picard *beta* website too (prod website has a second container on zappa), mbspotify too.
For MB, I will just stop cron after it completed both on trille and aretha (for hourly json dumps incremental).
jivte joined the channel
d4rkie joined the channel
d4rk has quit
zas
k
yvanzo
alastairp, lucifer: crontab -l says that critiquebrainz-cron-prod has currently no crontab (for root and critiquebrainz users).
jivte has quit
The file /etc/cron.d/critiquebrainz is present but it is probably not parsed.
nvm, log files show that it ran daily as expected.
I thought that 'crontab -l' would list jobs from this file too.
nexus has quit
lucifer
yvanzo: so all fine?
yvanzo
yes :)
atj
yvanzo: why would you think that? it's far too logical!
yvanzo
:D
jasje
monkey: how do i use `/profile/music-services/critiquebrainz/refresh/` :P
Checked that cron services are back too. (They just started with containers.)
jasje
akshaaatt: alright adding to the proposal
yvanzo
Cgroup Driver: systemd / Cgroup Version: 2
lucifer
akshaaatt: there's lots of stuff to do about it. alastairp was going to take it up but probably a few weeks in that. but he may available for discussion sometime sooner, we can sketch out what work is pending and divideit accordingly.
(almost forgot restarting json dump hourly cron on aretha)
reosarevok
ROpdebee: we got the following exception in Sentry: "Error: cannot call methods on artworkViewer prior to initialization; attempted to call method 'close'"