I'm just mildly annoyed utcnow doesn't set the tz on it's own
#dev-contrib
gritty wind
vale ibex
legacy
molten perch
Wow that was fast, thank you! π
green oriole
my pleasure doing business with you π΅οΈ
molten perch
I had to do a bit of a trick with the cache, but it works
fallen patrol
bump, can I submit pr to apply this to Lance?
doesn't seem worthy of an issue, but if no one says yeah I'll make one lol
green oriole
I don't understand what you want to change and why
fallen patrol
I don't understand what you want to change and why
the paginator on Lance edits every message twice?
OK I will make an issue lol
the paginator in Lance was taken from bot years ago. the bug was fixed on bot, but not on Lance. this patch fixed it on bot last year. it was not ported to Lance.
green oriole
I see, you can PR it
gritty wind
Hey @vale ibex did you have to manually cross-off the task in bot#1961 or was it closed automatically with the PR?
dusky shoreBOT
vale ibex
I did it π
fallen patrol
Aww, would've been cool
write workflow for it /s
gritty wind
Just tell joe, he'll find a way to make review bot do it
patent pivot
interesting idea
vale ibex
Yo, whoever want's to be a part of the termination of @brisk belfry feel free to check out bot#1928
dusky shoreBOT
green oriole
vale ibex
heh
patent pivot
wrong
trail pilot
I was about to say the same thing haha
and that lead to my question, would you want logic suggestions or nitpicks?
green oriole
Hey @last patio, could you have a short peek at https://github.com/python-discord/bot/pull/1928 please ?
trail pilot
What do you all think of sir-lancebot#948 which migrates the quiz command into threads and also adds a few questions onto the most common categories for the quiz command?
dusky shoreBOT
trail pilot
Actually about that last part, I'm not sure if this is just deja vu but was there an issue on adding more questions to the quiz command?
vocal prairie
sir-lancebot#815 maybe?
dusky shoreBOT
trail pilot
Yep it was that, thank you
so for now it might just have to be migrating to threads since that issue is for Objectivitix i think
cold island
What do you all think of sir-lancebot#948 which migrates the `quiz` command into...
Seems fine, we just don't use threads publicly yet
trail pilot
Seems fine, we just don't use threads publicly yet
Ahh, alright
fallen patrol
I see, you can PR it
seasonalbot#950 π
dusky shoreBOT
austere hornet
Never heard of seasonalbot before
green oriole
It was sir lancebot like 10 months ago
thorny obsidian
seasonalbot is what lancebot used to be called
dusky shoreBOT
fallen patrol
say...
@clever wraith does the appeals plugin have a way to allow mods and up into the appeals server without kicking them automatically?
austere hornet
It was sir lancebot like 10 months ago
Ah ok
gritty wind
Anyone with the dev-ops role does not get kicked
fallen patrol
but what if there's a new mod who needs to be added π
clever wraith
<@456226577798135808> does the appeals plugin have a way to allow mods and up in...
yeah it does, devops and owners are whitelisted by defualt and staff can be given roles to bypass it
gritty wind
No real reason to be on that server
fallen patrol
how can staff be given roles to bypass if they get kicked lol
gritty wind
Given the roles here
fallen patrol
ah
green oriole
Given the roles here
There is a staff role you can have on the server to not get kicked
But it is a mystery how you assign it
fallen patrol
yeah
fallen patrol
But it is a mystery how you assign it
solution: assign devops here, join, assign staff role there and remove the devops role here
green oriole
I mean, we would just restart modmail and do it while the bot is restarting lol
fallen patrol
lmao
green oriole
We won't assign devops just for this
It does have some perms, it isn't a vanity role
fallen patrol
ah
patent pivot
We won't assign devops just for this
lol i think we did
green oriole
With who though
patent pivot
scale got in somehow
green oriole
Scale got yeeted didn't he
fallen patrol
well tbf scale has admin so
patent pivot
what a hacker
fallen patrol
not really like scale is getting extra perms
green oriole
Doesn't matter for scale anyway, he's admin lol
patent pivot
probably won't be possible at some point down the line
when we do
zero trust all the thingsβ’οΈ
green oriole
You should really do the thing
fallen patrol
zero trust all the thingsβ’οΈ
imagine not trusting your admins /s
green oriole
Just because it is so technically interesting
patent pivot
imagine not trusting your admins /s
don't trust myself tbh
it's not actually perm removal, it's just requiring things like 2fa to use perms
patent pivot
nothing is changing?
think like sudo
austere hornet
I'm so confused
fallen patrol
think like sudo
don't understand lol
patent pivot
escalate temporarily to have perms
patent pivot
bot command + site somewhere
fallen patrol
yes I want the juicy implementation details
fallen patrol
bot command + site somewhere
okay so it effectively would not do anything
patent pivot
huh?
green oriole
I was thinking about command + DM to the bot
green oriole
Having a web challenge run on the CF edge could be interesting
fallen patrol
and how would perms be granted?
fallen patrol
with a role that has admin, below the bots highest role
patent pivot
above all roles
fallen patrol
no, it would be below the bots highest role
green oriole
Web challenge would allow stateless bot which would be sick
patent pivot
no, it would be below the bots highest role
ah, right yeah
it is something that i am tying into
snow don
fallen patrol
which now means the bot is in control
patent pivot
my cool project name
green oriole
Ah yes
patent pivot
but snowdon will do more than thta
fallen patrol
so some of the current admin user commands would not be accessible to them unless superuser
eg eval
because anyone who had access to the eval command could bypass the whole system
and just yeet themselves an admin role
patent pivot
eval would be something we'd look at yeah, it would probably be a separate bot to python but we'd need to see how we can hook it in
fallen patrol
yes I do think of the discord permissions system daily
patent pivot
presumably have snowdon expose an API where python can check if the user has sudo status
thorny obsidian
My big concern is redundancy in case there are issues with the bot's connection or the api gateway for the bot
fallen patrol
would this apply to admins or owners too?
okay I'm basically asking are you gonna make lemon and Sebastiaan have to do the 2fa
patent pivot
My big concern is redundancy in case there are issues with the bot's connection ...
mentioned this in mods actually, we'd distribute a doomsday script to mods that can restore their mod perms and the same for admins, requires some form of quorum (e.g. 2 mods or whatever), completely off our infrastructure
patent pivot
would this apply to admins or owners too?
from what we've discussed they want that, yep
thorny obsidian
(With me no longer being an admin but still needing to do things like move channels, role adjustments, and etc. It's essentially zero trust and it's such a pain to get things done tbh)
fallen patrol
so what happens to your permanently superuser perms Joe π
thorny obsidian
mentioned this in mods actually, we'd distribute a doomsday script to mods that ...
what would this script look like? How would it be run? Because if it's relying on connections with the gateway but for whatever reason discord is limiting it for bots how would that work?
patent pivot
so what happens to your permanently superuser perms Joe π
they remain as is, ideally they wouldn't but at least it means the vector is greatly reduced
fallen patrol
(With me no longer being an admin but still needing to do things like move chann...
I'm surprised the events lead role doesnt give you more perms
patent pivot
what would this script look like? How would it be run? Because if it's relying o...
a good question, and one to figure out, running it would probably use something like Shamir secret sharing to decrypt a token and use that to restore perms, not entirely sure of the implementation yet, almost all of this is just ideas in my mind. As for a situation where bots cannot use the REST API, I can't think of one where discord is online and that is the case, so it's one to think about but not one I expect we'd have to deal with
thorny obsidian
a good question, and one to figure out, running it would probably use something ...
but at some point it would still use a bot token, right? I could easily imagine a scenario where the python bot gets accidentally rate limited or something happens with it.
patent pivot
it wouldn't use python, all of this would be on a separate team on separate bots
fallen patrol
btw what's the status of botcore?
since botcore is supposed to dedupe code between bots
patent pivot
think it's in Zig's hands, not sure where we're at with it
fallen patrol
wish i could help lol
gritty wind
We merged the first commit a couple days ago
The repo itself is public though, you can see that
thorny obsidian
I guess I'm just ultra hesitant of the zero-trust thing, cause with my events lead role it has been Not The Most Fun β’οΈ and if there's a hiccup you still have to determine if it's serious enough for the Doomsday script or not. It just sounds like a lot of friction to getting stuff done. Buuut I'm no longer an admin so *shrug*
patent pivot
I'm not sure it's similar to your situation
this isn't removal of permissions, admins will still have all the permissions they need, it's just audited and needs 2FA, I'm not even sure if we'll restrict Administrator
but yeah it's something that still needs a lot of specifying
brisk brook
Zero-trust o_O?
My Dad has done work for a bank, and they need zero-trust for all microservices... and when I say, trust me, it does not sound fun.
thorny obsidian
this isn't removal of permissions, admins will still have all the permissions th...
but from what you described earlier, you don't have the perms until you request them. So they don't necessarily have the perms off the bat or with the "usual" role
brisk brook
All sockets between the microservices are encrypted, audit logs are done on an encrypted Blockchain, etc.
thorny obsidian
Because right now, anytime I want to do something with moving channels, or roles, or whatever it's me requesting to either have those perms or it be done. Even if an admin is immediately around to give that to me, it's still a delay and additional friction to getting things done when you're on a roll with getting things done.
patent pivot
but from what you described earlier, you don't have the perms until you request ...
yeah, but the request is entirely automated and required a 2FA code, might even use attributes like IP or geo so we only ask for one 2FA/day or whatever
but yeah it's something we've yet to specify fully, UX will be something that we balance out with security
patent pivot
Zero-trust o_O?
My Dad has done work for a bank, and they need zero-trust for a...
lol were not going to go full ZT
but as much as is feasible
and auditing permission uses feels like something that is fairly alright
gritty wind
yeah, but the request is entirely automated and required a 2FA code, might even ...
Could you just draw a big circle on the map for my geo lmao
brisk brook
and auditing permission uses feels like something that is fairly alright
An Admin doesn't need to be an admin all the time yeah.
cold island
Could you just draw a big circle on the map for my geo lmao
Scale the adventurer
austere hornet
Lol
gritty wind
What's the unlabeled zone between scaleios and not scaleios
What's the government hiding
What do they not want us to know
fallen patrol
The repo itself is public though, you can see that
actually i thought all of the issues for it were internal lol
gritty wind
There aren't any issues written down afaik
patent pivot
lol, yea, there aren't
fallen patrol
or was that smartconfig
gritty wind
forms probably
fallen patrol
ah, that's your project π€‘
gritty wind
That one does have some issues written down, just not public
fallen patrol
yeah
thorny obsidian
@molten perch what's the command structure for .aoc daystar?
Is it still something like: .aoc daystar 5-2
?
thorny obsidian
oooh, nice nice!
molten perch
#changelog message
It's in the changelogs π
And the leaderboard command works like: .aoc lb [name] if it is not found, then it'll send back an error.
cursive relic
Somehow .aoc join ain't working :/
patent pivot
... the answer is nothing <:lemon_sweat:754441881718620281>
yes it does
it doesn't raise from there, but it logs an error to sentry
thorny obsidian
oh how does it know to log to sentry?
patent pivot
not all leaderboard failures notify, transient HTTP doesn't, but if it gets a bad response (e.g. bad cookie) then it'll log
stable mountainBOT
bot/exts/events/advent_of_code/_helpers.py lines 261 to 262
log.error("It seems like the fallback cookie has expired!")
raise FetchingLeaderboardFailedError from None```bot/exts/events/advent_of_code/_helpers.py lines 278 to 279
log.error(f"reached 'unreachable' state while fetching board `β{leaderboard.id}`β.")
raise FetchingLeaderboardFailedError```aaaaah, okie
patent pivot
but we could make it more visible yea
vale ibex
We should resolve those in sentry so that when they happen again they get shoved in #dev-alerts
patent pivot
lol
no
alerts are sensitive, include user data
a lot
lol
if mod commands fail, they get relayed. if site fails, IPs and browsers get relayed, so on
gritty wind
Somewhere, in the depth of that channel, you can find my IP, browser, location, probably my phone number and mother's maiden name too
I've opened PRs on lance and bot to action #dev-announcements
patent pivot
lmfao
austere hornet
Lol
gritty wind
sir-lancebot#951
bot#1966
dusky shoreBOT
gritty wind
They are tiny, would appreciate a review
Are there any more projects that need this
maaaaybe site
nop
fallen patrol
I've opened PRs on lance and bot to action <#765137969753096202>
maybe edit the announcement to say, we're dumping it?
gritty wind
I'll put out a new announcment
But seeing as like 1 person actually uses it, it really does not matter
For most people, you'll have to do nothing
I could probably ping them individually too
fallen patrol
But seeing as like 1 person actually uses it, it really does not matter
π I reacted to it too
gritty wind
I know, I saw
fallen patrol
oh i forgot
fallen patrol
oh i forgot
Please make sure to ping me with an explanation of what you use it for if you vote B
@gritty windTldr I use file logging for seeing differences between runs, or after a test run and they fail.
patent pivot
are our logs deterministic π
gritty wind
- We don't log tests
- There are definitely better ways to see what tests failed
- What joe just said
- It doesn't really matter to be able to see what changed between runs tbh
patent pivot
1. We don't log tests
2. There are definitely better ways to see what tests fail...
even when Iβm a troll Iβm helpful
gritty wind
It's why we keep you around
austere hornet
Lmfao
vale ibex
Once bot#1868 is merged, we should probably update/remove .aoc subscribe
dusky shoreBOT
fallen patrol
Once bot#1868 is merged, we should probably update/remove .aoc subscribe
given how close december is, probably keep it but direct to the new command
vale ibex
can also remove role management perms from lance then too
fallen patrol
neat
green oriole
even when Iβm a troll Iβm helpful
that's the whole reasoning behind your devops daddy role
austere hornet
Lmao
molten perch
Yep, we'll definitely need to implement parallel testing on the API in the future π€
dusky shoreBOT
molten perch
we have it on bot if you want a reference
We might have to introduce some changes on the database creation.
(Create databases for each worker, I guess?)
patent pivot
xdist my beloved
patent pivot
We might have to introduce some changes on the database creation.
(Create databa...
maybe pg virtualenv can do that
vale ibex
We might have to introduce some changes on the database creation.
(Create databa...
We could use pg venv, or look at a way to make it so that it's threadsafe
molten perch
Okay, let's just discuss it before doing anything. π
molten perch
maybe pg virtualenv can do that
How's it different from creating a database and tearing it down at the end of the test? We could use worker ids, and maybe it's easier while using docker?
See: https://github.com/pytest-dev/pytest-xdist#identifying-the-worker-process-during-a-test
austere hornet
Lol
thorny obsidian
Once bot#1868 is merged, we should probably update/remove .aoc subscribe
How close is your PR to being merged?
molten perch
Will you do that, or should I do it?
dusky shoreBOT
vale ibex
How close is your PR to being merged?
pretty close, jsut needs reviews
has had some feedback implemented already
dusky shoreBOT
fallen patrol
Yep, we'll definitely need to implement parallel testing on the API in the futur...
lol 45 seconds isn't that long
molten perch
That's one endpoint
fallen patrol
depending on how the tests are distributed, it won't help
molten perch
What do you mean? What will not help?
Oh, paralel testing. No, I'm almost certain it will.
I mean 45 seconds is not long, but if we aren't gonna implement paralel testing, it's gonna take much longer.
That's what I want to avoid.
clever wraith
Does @stable mountain have slash commands and buttons?
static canyon
Does <@409107086526644234> have slash commands and buttons?
Not slash commands but it does have buttons
clever wraith
using dpy 2.0?
clever wraith
Do you guys plan on adding your own implementation of slash commands or via a fork?
oh
static canyon
Do you guys plan on adding your own implementation of slash commands or via a fo...
Neither will happen any time soon (within sort of 3-6months) but we'll most likely pick a fork
But we probably won't be using slash commands
fallen patrol
api doesn't go to <#622895325144940554> ?
okay so doesn't look like pr comments go to #dev-log
some form of comment lol
trail pilot
Should .aoc [command] be limited to #aoc-bot-commands?
thorny obsidian
Should `.aoc [command]` be limited to <#897932607545823342>?
only join, subscribe, and countdown are allowed there. The other ones are restricted
trail pilot
ah alright
itβs kind of taking over #advent-of-code though and Iβm sure it isnβt a worry now but in the future it might be
specifically .aoc join
thorny obsidian
I'm okay with it taking over for now because there isn't much else to talk about with AoC
clever wraith
!remind 166h dusk time.
stable mountainBOT
clever wraith
Fuck wrong channel
austere hornet
@clever wraith you can run !remind delete 3522 if you want to delete that
clever wraith
No I want the reminder I just sent it in the wrong place
austere hornet
Ah ok
austere hornet
No I want the reminder I just sent it in the wrong place
Well I meant you could delete this one and send it in the right place
molten perch
I noticed a "minor" error in the AoC dayandstar feature,
could someone take a look at sir-lancebot#954 ?
It's 3 liner basically, shouldn't take more than a minute to review.
dusky shoreBOT
molten perch
Where?
Nope, if you run postgres with docker it has to be in correspondence with the url specified in docker.
For now it's enough if you specify a bogus value, but as soon as my PR gets merged(if it ever gets merged lol) you'll have to specify an actual db url.
But if you did not touch any of the API part, the tests should pass. π
fallen patrol
For now it's enough if you specify a bogus value, but as soon as my PR gets merg...
is that really a good idea?
molten perch
Yes, it is.
It's under development. I can't really do anything else.
The current tests don't use the database at all, so you can just provide a bogus value. It does not matter.
But my PR introduces tests that actually require a database to be set up.
fallen patrol
what's your pr number?
molten perch
api#30
fallen patrol
api#30
will leave review
kind of seems overscoped, both a migration and setting up the test suite...
molten perch
That's how it is.
I worked a lot on that PR.
In order to set up an endpoint I had to create a testing suite.
And actually, tests are the core parts of the endpoints.
The only additional things are the fixtures.
fallen patrol
if one builds on the other then it can be done with two prs, pointing the migration pr towards the testing suite pr and leaving as draft but specifying do not merge
because it's overscoped and large that's likely why there's no reviews yet π
and I know that from experience, having made some 2000 line diff prs lol
molten perch
It is not large at all.
I just modified a few models so that it's more convenient. (we're talking about one-liner changes)
Secondly, I pushed that PR yesterday
fallen patrol
For now it's enough if you specify a bogus value, but as soon as my PR gets merg...
ah, I was explaining towards the (if it ever gets merged lol) part here
molten perch
To be totally honest at this point, I was trying to be humble.
fallen patrol
ah
fallen patrol
hey could someone look at bot#1967 lol might want to close it
dusky shoreBOT
austere hornet
Lol what the hell
austere hornet
Thanks @cold island
molten perch
Hey! Would someone mind to take a look at sir-lancebot#954 ? It only needs one review.
(It's a three-liner basically)
dusky shoreBOT
molten perch
Thanks a lot.
austere hornet
Hey! I would appreciate a few reviews on sir-lancebot#901. Thanks so much!
austere hornet
I've noticed that .halloweenify and .spookyfact aren't locked to October. Should they be?
Also noticed a grammatical mistake here: https://github.com/python-discord/sir-lancebot/blob/main/bot/exts/avatar_modification/avatar_modify.py#L289
stable mountainBOT
bot/exts/avatar_modification/avatar_modify.py line 289
brief="Spookify an user's avatar."```May I PR a fix?
austere hornet
Also if I could get assigned to sir-lancebot#937, that would be great. Thanks!
dusky shoreBOT
full fractal
bot_1 | 2021-11-22 02:29:42 | bot.utils.scheduling | ERROR | Error in task Task-16 140701060127440!
bot_1 | Traceback (most recent call last):
bot_1 | File "/bot/bot/exts/help_channels/_cog.py", line 316, in init_cog
bot_1 | await self.move_idle_channel(channel, has_task=False)
bot_1 | File "/bot/bot/exts/help_channels/_cog.py", line 352, in move_idle_channel
bot_1 | await self.unclaim_channel(channel, closed_on=closed_on)
bot_1 | File "/bot/bot/utils/lock.py", line 108, in wrapper
bot_1 | return await func(*args, **kwargs)
bot_1 | File "/bot/bot/exts/help_channels/_cog.py", line 436, in unclaim_channel
bot_1 | return await _unclaim_channel(channel, claimant_id, closed_on)
bot_1 | File "/bot/bot/exts/help_channels/_cog.py", line 448, in _unclaim_channel
bot_1 | claimant = await members.get_or_fetch_member(self.bot.get_guild(constants.Guild.id), claimant_id)
bot_1 | File "/bot/bot/utils/members.py", line 20, in get_or_fetch_member
bot_1 | member = await guild.fetch_member(member_id)
bot_1 | File "/usr/local/lib/python3.9/site-packages/discord/guild.py", line 1714, in fetch_member
bot_1 | data = await self._state.http.get_member(self.id, member_id)
bot_1 | File "/usr/local/lib/python3.9/site-packages/discord/http.py", line 335, in request
bot_1 | raise HTTPException(response, data)
bot_1 | discord.errors.HTTPException: 400 Bad Request (error code: 50035): Invalid Form Body
bot_1 | In user_id: Value "None" is not snowflake.
whyyyy
did someone change the help channels in the test server 
tawdry vapor
did someone change the help channels in the test server <:pithink:65224755990927...
it's cause of this #dev-contrib message
full fractal
Makes sense, how am I supposed to fix it?
hmm, !dormant didn't work as apparently it's disabled?
tawdry vapor
Manually move the channels to dormant and reload the extension I guess
full fractal
seems to have fixed it, thank you!
static canyon
@green oriole fyi your issue (bot#1968) is a dup. of one I made a few weeks ago (bot#1930)
dusky shoreBOT
static canyon
Not sure what the process for duplicates is
clever wraith
:BaguetteSwing: duplicate :BaguetteSwing:
Thanks discord, appreciate it
Well, in most cases you can copy the information from the newer issue to the older and close the newer one
static canyon
I guess I'll do that then, thanks π
clever wraith
I think it is fine personally
static canyon
I hope so :p
static canyon
π
austere hornet
@static canyon Just as an FYI so you don't yell at me for committing failed lint code like last time lol. I committed Bluenix's suggestion which probably caused a line spacing issue, which I'll try to fix later today. Thanks!
static canyon
Sounds good π
Although to be clear, I wasn't meaning to "yell" at you last time; was just making sure you had pre-commit etc.. on because it makes life way easier
I'll probably review your PR over the weekend or something
austere hornet
Although to be clear, I wasn't meaning to "yell" at you last time; was just maki...
Ah ok
static canyon
!remind 5d Review sir-lancebot#901
stable mountainBOT
austere hornet
Thanks
fallen patrol
heyyy could a admin or owner add api to #dev-log π₯Ί
I use dev-log to keep track of stuff over github notifications lmao
short snow
@vale ibex I will just ping you
short snow
welp π
vale ibex
seems owner only
short snow
okay
@patent pivot please do, (I am pinging cuz this is like the 4th time someone is asking to do that)
fallen patrol
4th time someone
that was me lol
short snow
i did once, rest you lol
fallen patrol
not all comments are
patent pivot
okay, enabled the PR comments that weren't on
fallen patrol
thanks!
short snow
thanks!
molten perch
pytest-xdist is a real wonder π
Managed to cut API test times to one-fifth of the original times.
green oriole
there is no real reason to not add it
molten perch
Yeah, the issue is more about the hows, because of the databases.
But I managed to get it work with worker ids.
Just waiting for an approval or similar on api#31
dusky shoreBOT
molten perch
Done-zo.
It took a bit longer because of package installs.
But the test only took 11 secs, in contrast to that yesterday it took 24 secs.
vale ibex
Nice
fallen patrol
Yeah, the issue is more about the hows, because of the databases.
But I managed ...
there's another plugin and many hooks to do that lol
lmfao https://nvd.nist.gov/vuln/detail/CVE-2021-41250 medium severity
molten perch
there's another plugin and many hooks to do that lol
That was literally two lines of code. I wouldn't bother with hooks and plugins if I can use a built in feature that works perfectly fine (it seems so)
Feel free to check it out.
fallen patrol
built in feature
the hooks are built in
molten perch
Okay.
π
I still believe, that we shouldn't bother with hooks and plugins, when you can do the same thing with worker ids.
And I find it more convenient. But, we'll see π
fallen patrol
I still believe, that we shouldn't bother with hooks and plugins, when you can d...
well, this looks like it runs on the runner, sure. But how do I run it locally?
molten perch
Just the way you would run it normally. pytest
green oriole
lmfao https://nvd.nist.gov/vuln/detail/CVE-2021-41250 medium severity
I feel like I recognize that ID
fallen patrol
so the database is already running, everything is already set up?
molten perch
As of now, you have to have a database up and running, and provide a db url in the env.
Then, pytest.
But you can also run in docker.(not sure though)
We might create a task, not sure.
fallen patrol
As of now, you have to have a database up and running, and provide a db url in t...
gotcha
so it'll give me a very cryptic error if there isn't a database running π
molten perch
Most certainly yes.
vale ibex
We should just support it in docker
molten perch
Yeah, like the migrations
vale ibex
if you want to run psql locally then that's up to you
fallen patrol
Most certainly yes.
use a pytest hook to block the test run if there isn't a database, or mark all tests that rely on a database and run them if it exists
vale ibex
If you have the knowledge to have a psql server running locally, then you likely have the knowledge to run pytest against it
molten perch
Seems overcomplicated to me.
vale ibex
There will always be a database, since we will have it running in docker
molten perch
Okay, but I'm not sure why we would torture ourselves with hooks and such.
fallen patrol
tests/conftest.py
import pytest
def pytest_configure():
if not database:
pytest.exit('A database must be running to test')
and database just needs to validate it can connect
molten perch
We're gonna support it only in docker. As Chris mentioned
fallen patrol
could also use pytest_sessionstart
not sure why you think hooks are torture lol
but docker is another fine solution
molten perch
Because those are integrations tests, I'm not sure why would someone go ahead and run them just out of the blue.
But sure, we're gonna support it in docker only, but still..
fallen patrol
I'm not sure why would someone go ahead and run them just out of the blue.
I read both readme.md and contributing.md and neither mention a database is required to run the tests
While it is implied, it is never mentioned. So that's my review, I guess. I'll convert it to a github review
molten perch
Yeah, since...
We don't have the tests yet. π
It's a PR yet to be merged, I'll open PR to make the README better later on.
But I'm a bit busy with the endpoints already, I'm terribly sorry.
fallen patrol
It's a PR yet to be merged, I'll open PR to make the README better later on.
I was reading that pr. Since it requires it, there should be a sentence that says it requires a running database, and tests must be run in docker. Doesn't need a whole pr to add it, a sentence should do. π
thorny obsidian
> I'm not sure why would someone go ahead and run them just out of the blue.
I ...
The API repo is still very much under development and the readme and contributing are by no means finalized. We're not even using the API yet. So it should be fine to only support docker as d0rs4n and Chris have so far indicated that they'll do.
fallen patrol
Yes, I know its under development, that's why I'm reviewing it.
fallen patrol
I feel like I recognize that ID
you may. lol
molten perch
@vale ibex It seems like that I cannot run pytest in docker, because the dev packages won't be installed when building the container. (--no-dev)
Do I just add a poetry install to the development "target"?
vale ibex
Yea, you could
but I'd suggest swapping production and development around if that is the case
since you always want to "build up" a docker image
molten perch
Yea, you could
Yeah, I can't really, since poetry is not installed in the development stage.
vale ibex
ahh
molten perch
And If I were to swap the stages around it would mean that dev packages are installed even in the prod. stage.
So, I'm clueless π
vale ibex
Nah, you'd do the install with --no-dev first
then in the dev stage install again without the flag
if production is before dev, then it would be stopped before that second install
molten perch
Yeah, but poetry is installed in the builder stage
Then, even if I were to put production before dev, I still wouldn't be able to use poetry. (I guess, I tested it out.)
vale ibex
Yea, you'd need to do FROM builder as development
rather than FROM base as development
Is there a particular reason we use assertEqual in site tests rather than assertCountEqual when comparing json? I don't think we really care about ordering, just that the actual key:value pairs match.
austere hornet
Bump #dev-contrib message
molten perch
Yea, you'd need to do `FROM builder as development`
Managed to get it work. (Did not swap it in the end, so the order is: builder, development(from builder), prod(from base))
Also, is there a particular reason as to why the app volume is read only? Pytest wanted to write to some cache file, so I had to remove it. Now it works fine.
vale ibex
It's read only so that code changed in teh container for whatever reason doesn't mess up your local copy
We should keep that volume read only and add specific writeable volumes for what we need
molten perch
Oh, yes. Makes sense, sorry.
vale ibex
No worries π
fallen patrol
pytest writes to cache to be able to help debugging test runs
austere hornet
Yea sure, seems fine to me
Is this referring to the mistake I found?
vale ibex
it's referring to the message you linked.
austere hornet
Thanks!
molten perch
We should keep that volume read only and add specific writeable volumes for what...
So, If I do something like that:
- .:/app:ro
- ./.pytest_cache:/app/.pytest_cache
- ./.hypothesis:/app/.hypothesis
It should be fine, right? π
(The tests work, I'm just curious if it seems fine to you)
molten perch
Sure! π
austere hornet
the grammatic error fix seems fine too
PR is ready, sir-lancebot#955. Could you please review?
austere hornet
Thanks @molten perch !
fallen patrol
So, If I do something like that:
```yaml
- .:/app:ro
- ./.pytest_cac...
you can also move pytest's cache, if you desire (not saying this is a good solution nor better than the above, it's not, it's just an option)
what is hypothesis?
trail pilot
Hypothesis is a testing framework in Python iirc
clever wraith
what is hypothesis?
An educated theory \s
austere hornet
Lmao
cursive relic
are link shorters blacklisted?
static canyon
are link shorters blacklisted?
For the most part, yes
Soonβ’οΈ they won't need to be but for now they are
cursive relic
but some of them (at least 1) seems to be available
static canyon
It's not something where they all have to be blacklisted
It's more that bit.ly for example has been used loads by people sending the nitro scam
cursive relic
oh I see
static canyon
And so we blacklisted the entire website until we get url unfurling
Which will be soonβ’οΈ (there is a PR for it somewhere on bot)
cursive relic
oo, nice
gritty wind
Soonβ’οΈ they won't need to be but for now they are
No they are still blacklisted
My PR deletes the shortened URL before unfurling it
We unfurl for the purposes of applying further filters
There is no good reason to use a shortener here
static canyon
Right, fair enough
green oriole
lmao
short snow
also, I remember an issue for this already existing
green oriole
So I cannot remove the do not merge tag on https://github.com/python-discord/site/pull/627#pullrequestreview-813563184
Chrome just crashes if I try to do that
patent pivot
removed it
green oriole
Thanks
fallen patrol
Chrome just crashes if I try to do that
use firefox π€‘
green oriole
no
brisk brook
@austere hornet I am reviewing your PR and one of your comments gave me a great idea.
You asked about the author check - and yes we can use it!
Add a self.checks = set() attribute inside __init__().
Inside the command after defining the author_check() add it to this set using self.checks.add(author_check) (and remove it after you're done).
Now inside the listener you can have this code: ```python
for check in self.checks:
if check(after):
break
else:
# This means that no check returned True
return
Now put the code that adds the message here:
...
You can do if not any(check(after) for check in self.checks):
green oriole
damnit Zig
cold island
brisk brook
Ah, right
austere hornet
Now inside the listener you can have this code: ```python
for check in self.chec...
!remind 5H this
stable mountainBOT
austere hornet
Thanks!
molten perch
Hey if you have some time, would you mind to take a look at api#30?
(It supports running tests in Docker, so It should be fairly easy to set it up, but you can read the README, it has been modified as well)
cold island
Hmm the otn endpoint was already migrated right?
molten perch
No, it has not been.
It built upon my previous PR that was close due to recent discussions regarding SQLAlchemy async, and 2.0 Syntax.
So that's the first endpoint effectively
But I can almost certainly say that this is the last heavyweight PR, since we've set up everything we need(including this PR):
Tests, ORM, Migrations.
cold island
I see
I think once we have one endpoint migrated we should look into using the new API in prod so that we can test that it's working properly, and then continue to the other endpoints
stable mountainBOT
!remind 5H this
It has arrived!
austere hornet
!remind 3H trying to do homework smh
stable mountainBOT
molten perch
I think once we have one endpoint migrated we should look into using the new API...
Yeah, sure! π That would be really cool!
stable mountainBOT
!remind 3H trying to do homework smh
It has arrived!
green oriole
setup-python supports caching for both pip and pipenv projects.
RIP
gritty wind
Use the composite action smhhhhhh
Iβve already scammed my way into botcore, and soon, every other repo
fallen patrol
> setup-python supports caching for both pip and pipenv projects.
RIP
use pip. poetry invokes pip underneath.
so it should cache the dependencies for poetry too
static canyon
!ban 816637783233527839 7d You were told literally yesterday to not self-advertise. Should you decide to rejoin, make sure you DON'T do this again.
stable mountainBOT
:incoming_envelope: :ok_hand: applied ban to @tardy wraith until <t:1638362201:f> (6 days and 23 hours).
molten perch
Hey! I've been looking at the site and bot code for a while, but I don't seem to get, when the active field of a Reminder will be set to False, I haven't really thought of that before.
It seems like that when a user is reminded, the reminder will just simply be deleted π
(or maybe that's is only for backwards-compatibility? )
molten perch
I have a PR open on bot to fix that
Oh I see.
I was just confused since the Patch Request Body format looks like that:
>>> {
... 'mentions': list[int],
... 'content': str,
... 'expiration': str, # ISO-formatted datetime
... 'failures': int
... }
So, activate is not accepted in that case, in the FastAPI version.
I'll go ahead and add it.
austere hornet
sir-lancebot#901 needs core dev reviews! (non-core dev reviews are also appreciated ofc)
fallen patrol
stable mountainBOT
.github/workflows/lint-test.yaml lines 84 to 91
# Run flake8 and have it format the linting errors in the format of
# the GitHub Workflow command to register error annotations. This
# means that our flake8 output is automatically added as an error
# annotation to both the run result and in the "Files" tab of a
# pull request.
#
# Format used:
# ::error file={filename},line={line},col={col}::{message}```The page is called formatting output inside github actions iirc
austere hornet
Hey does anyone know when PyDis will incorporate threads on the server?
cold island
We've been discussing it. I think it's probably a matter of an admin having time to pay proper attention to it.
austere hornet
We've been discussing it. I think it's probably a matter of an admin having time...
Ah I see. Thanks
green oriole
I have a bunch of things ready for that admin 
austere hornet
I have a bunch of things ready for that admin <:hyperlemon:435707922563989524>
Alright
molten perch
Hey! If any of you could look at api#30 it would be greatly appreciated. The PR might seem a lot, but actually it's just a few files, the rest of them are one-liner tiny model improvements, and Docker - or - Github Actions - related things.
(It should also be pretty straightforward to set the project up and run the tests!)
Thanks! π
austere hornet
What do people think about me using the nltk module for validation in sir-lancebot#901 (as stated in Shom770's latest comment)?
austere hornet
!pypi nltk
stable mountainBOT
trail pilot
Yeah, I think nltk.corpus.wordnet has support for all the english nouns and more
static canyon
Is there a way to get all the issues/PRs I've commented on for a specific repo? Presumably it's stored somewhere so that GitHub knows to email you when a new comment/commit/etc. is made
green oriole
What do people think about me using the `nltk` module for validation in sir-lanc...
Sure
green oriole
Is there a way to get all the issues/PRs I've commented on for a specific repo? ...
commenter:TizzySaurus org:python-discord
You can add some is: tags to search for issues or PRs, and set their status
static canyon
Perfect, thanks π
dim topaz
Just looking for some assistance with snekbox. In particular the following config
2021-11-25 09:57:01,374 | 12 | snekbox.nsjail | WARNING | Failed to set the memory swap limit for the cgroup. This is probably because CONFIG_MEMCG_SWAP or CONFIG_MEMCG_SWAP_ENABLED is unset. Please ensure swap memory is disabled on the system.```Tryna sus the correct config
Also please @ me on a response, Im off soon
vale ibex
stable mountainBOT
snekbox/nsjail.py lines 98 to 107
try:
# Swap limit is specified as the sum of the memory and swap limits.
# Therefore, setting it equal to the memory limit effectively disables swapping.
(mem / "memory.memsw.limit_in_bytes").write_text(mem_max, encoding="utf-8")
except PermissionError:
log.warning(
"Failed to set the memory swap limit for the cgroup. "
"This is probably because CONFIG_MEMCG_SWAP or CONFIG_MEMCG_SWAP_ENABLED is unset. "
"Please ensure swap memory is disabled on the system."
)```from googling around, those two seem to be kernel flags, are you running this in docker or locally?
green oriole
Isn't it because those env vars are set and not unset?
The problem is that you have a swap on your system
dim topaz
Its using the docker container
green oriole
Try running sudo swapon
dim topaz
Correct, 2G
I'm not sure of the flags in snekbox and if I can disable it without removing swap
green oriole
Then you'd want to turn it off if you want the memory limit to work
dim topaz
Although I could just remove it, not a big deal
green oriole
We could add an option to not turn on the memory limit
Due to how the swap system works, we cannot set a limit when you have some swap set
dim topaz
π
molten perch
Hey! If someone could take a look at api#30 it would be greatly appreciated. The PR might seem a lot, but actually it's just a few files, the rest of them are one-liner tiny model improvements, and Docker - or - Github Actions - related things.
(It should also be pretty straightforward to set the project up and run the tests!)
Thanks! If you have any questions regarding how to set it up, feel free to ask! π
static canyon
site#621 is failing because of a merge conflict but when I do what it says to resolve it says no conflicts were found?
Error:CommandError: Conflicting migrations detected; multiple leaf nodes in the migration graph: (0075_infraction_dm_sent, 0075_add_redirects_filter in api). To fix them run 'python manage.py makemigrations --merge' Error: Process completed with exit code 1.When I do the command:```
C:\Users\tizzy\site>python manage.py makemigrations --merge
No conflicts detected to merge.
It seems like a migration has been merged since you generated your migration, so you have to modify the βdependenciesβ part in your migration to be up to date with the current migration βHEADβ(?) Iβm on phone so I canβt send you a snippet but you can follow the syntax of the previous migrations.
(In case I havenβt wrote enough βmigrationβ: migration)
brazen charm
or just delete your migration during the merge, create a new one, then add it and commit
fallen patrol
what.
cold island
yeah
mellow hare
Cheers
cold island
python-dotenv
mellow hare
Solid. Re-writing my little input getter for advent of code
Figured I should do it correctly this time
fallen patrol
Solid. Re-writing my little input getter for advent of code
python-dotenv, importable as dotenv
however, given what you're writing, I would look at using environs :D
!pip environs
stable mountainBOT
fallen patrol
will serialize the types for you
mellow hare
I'm happy and familiar enough with dotenv. You can feed it a file path to the .env file, right?
static canyon
I'm happy and familiar enough with dotenv. You can feed it a file path to the `...
By default it searches the work directory
static canyon
But I believe you can also specify
mellow hare
I just realized I was thinking the wrong way anyway
I'm good now, just had a derp
For some reason my brain was like "Yeah put it in the project root" despite me putting the input getter in a sub project folder, so I'll just put it there with that
Ignore the crazy man
vale ibex
I might be wrong, but I think it pulls it from sys.path
mellow hare
I know it loads the .env stuff alongside that
vale ibex
so if it's in that or a sub dir then it should be fine π
mellow hare
Structure is:
advent_of_code/
advent_of_code/
input_getter.py
So yeah, they see me derpin
vale ibex
hah
fallen patrol
But I believe you can also specify
yes, you can also customize what it searches, too
fallen patrol
@patent pivot how [can it be harmful]?
patent pivot
<@!165023948638126080> how [can it be harmful]?
how what
oh i'm mainly just trolling
but force pushes can be harmful because they change history
especially on a PR whihc multiple people are working on
but Volcyy is a genius, an expert, the top of his field, I know if anyone is careful about force pushes, it's him
fallen patrol
ah
i've been force pushing a lot recently lol
mostly because i've been debugging github actions
gritty wind
This is why you have a private repo with an easy to remember name
This is just the main branch π€‘
Canβt actually see the action count # on the mobile app tho
fallen patrol
This is why you have a private repo with an easy to remember name
I suppose... the problem was it was related to the existing workflows and extra integrations.
specifically was setting up coveralls and ensuring it was set up in a way to not cause workflow failure if coveralls failed
fallen patrol
lol @last patio how often do you update your about me with Joe quotes?
last patio
every time the queen retreats to enter a "new phase"
patent pivot
green oriole
But is mongodb Webscale?
austere hornet
<:lizshunk:775413631222939650>
Lmao
How would I install the dependencies for nltk with poetry install? Is it simply poetry install nltk?
!pypi nltk
stable mountainBOT
static canyon
iirc when i was messing around with nltk earlier it's a bit of a weird one
You install nltk and then you install the bit that you actually want (i.e. corpus iirc)
static canyon
nltk.corpus.words is what we'd want (a from nltk.corpus import words)
>>> import nltk
>>> nltk.download()
This is what they show on the website
Then it brings up a GUI and you select corpus or whatever it is
timid sentinel
Hmm I don't really understand the need for nltk
static canyon
Here we go, "words" is shown there (you actually want wordnet but same process)
timid sentinel
If it's just validating user input, I don't really see the need
static canyon
If it's just validating user input, I don't really see the need
The main idea iirc is that it gives a list of words
Which I guess could also be used in .hangman
timid sentinel
Looking at the issue in question that doesn't seem to be the use case here
Unless I'm misunderstanding
static canyon
Relevant comment on the PR: https://github.com/python-discord/sir-lancebot/pull/901#pullrequestreview-815493133
timid sentinel
Yep, seems sort of out of scope and unnecessary
I think it's alright to let people enter nonsense, part of the fun
austere hornet
I think it's alright to let people enter nonsense, part of the fun
I was thinking the same thing and so was @brisk brook , cc @trail pilot since I DM'd you about this too
trail pilot
Hm, I didn't know you had to do all this jank to download the corpus itself
so it's fine as is
patent pivot
just rolled out a couple of security headers to all python discord web properties
if anyone notices something weird on pythondiscord.com please ping me, if any staff notice something weird on an internal service also ping me
full fractal
if anyone notices something weird on pythondiscord.com please ping me, if any st...
Is it just me or is the icons not loading at all (not sure if this is pre-existing issue but I just noticed it)
patent pivot
oh
huh
that's
lol
ahhhhh
so
yeah that's related to this
hmmmmm
hmmmmmm
so the reason is that I stopped us sharing pythondiscord.com with sites that we link out to
thorny obsidian
How I know it's devops trickery hour:
patent pivot
and one such site is fontawesome, who uses our referrer to figure out if we are allowed to use pro icons
I think I can leak a bit more info, I'll just tell browsers to set the referrer to www.pythondiscord.com with no page details
@full fractal give that a shot
patent pivot
How I know it's devops trickery hour:
lmfao true true
full fractal
And it works, no 403 from fontawesome in the logs
patent pivot
awesome, cheers
full fractal
thanks!
tawdry vapor
I'm not sure of the flags in snekbox and if I can disable it without removing sw...
I don't think the answer you received was correct. Swap can be enabled on your system, but then your kernel must have support for managing swap memory resources I.e. It should support setting a swap limit in a group. This is what those kernel flags mentioned in the error are for. If swap is enabled but can't be controlled by the cgroup, then the process would effectively have no memory limit since it would be able to just use the swap space.
Alternative to those kernel flags is to completely disable swap on the system.
The message is warning you that snekbox can't limit swap memory so it's now your responsibility to disable swap
It shows that warning even when swap is already disabled too, in which case you can just ignore it.
short snow
Would love some reviews on:
dim topaz
I don't think the answer you received was correct. Swap can be enabled on your s...
Yea. My kernel doesn't support the c group flag so I had to disable swap. Although I managed to completely brick the server in the process somehow. Not sure how I managed to fuck up something as simple as disabling swap but yea
cursive relic
If I copy some systems or parts from @stable mountain I need to add licence from the repo to my code even if it is modified source, right?
green oriole
yes, you need our license to appear somewhere
cursive relic
good to know, I was about to make some commits and had to get this checked up, Adding it to on top of files where I use anything copied should be good?
green oriole
yeah it is fine
cursive relic
great, thx!
austere hornet
sir-lancebot#901 needs core dev reviews! (non-core dev reviews are also apprecia...
Bump
short snow
@austere hornet Could you add some example executions (screenshot of running the command) to your christmasify PR?
austere hornet
<@!853309207922212965> Could you add some example executions (screenshot of runn...
Sure!
Well actually it isn't really ready yet, but I will do that when it is. Sounds good?
short snow
π yep
I just wanted to see how the feature wasn't going, haven't planned on reviewing it yet btw
cursive relic
What is the url it's making a session to? A closed one that can't be seen or smth from discord itself? https://github.com/python-discord/bot/blob/1302632c39469a124e85e6b43c5278526874ce5e/bot/exts/filters/filtering.py#L561
stable mountainBOT
bot/exts/filters/filtering.py line 561
f"{URLs.discord_invite_api}/{invite}", params={"with_counts": "true"}```These constants are defined here https://github.com/python-discord/bot/blob/main/config-default.yml#L377-L379
stable mountainBOT
config-default.yml lines 377 to 379
# Discord API URLs
discord_api: &DISCORD_API "https://discordapp.com/api/v7/"
discord_invite_api: !JOIN [*DISCORD_API, "invites"]```If you have an invite code, you can see any guild, even in the discord client
Why are we still using v7 lol
Iβm not even sure what v7 is, itβs documented as βdoesnβt look like anything to meβ in the docs
cursive relic
These constants are defined here https://github.com/python-discord/bot/blob/main...
Oh I was just looking at the wrong place, thank you!
fallen patrol
Iβm not even sure what v7 is, itβs documented as βdoesnβt look like anything to ...
v7 is actually what dpy 1.7 was using for the longest time
now dpy 2.0 uses v8 for the REST api and v9 for the gateway
fallen patrol
v7 is actually what dpy 1.7 was using for the longest time
(it was like this because as i was told, ppl were using it before it was documented, so everything was documented as v8.)
brisk brook
Iβm not even sure what v7 is, itβs documented as βdoesnβt look like anything to ...
I remember being told that it's v6 but with some bugfixes/better errors/etc. and library devs went "Ooh, shiny" for it
thorny obsidian
@summer garden consider yourself assigned for the epoch command
cursive relic
is the http_session smth from dpy or not? I can't seem to find from anywhere https://github.com/python-discord/bot/blob/1302632c39469a124e85e6b43c5278526874ce5e/bot/exts/filters/filtering.py#L560
stable mountainBOT
bot/exts/filters/filtering.py line 560
response = await self.bot.http_session.get(```the .get is from python src but session is smth else?
fallen patrol
I remember being told that it's v6 but with some bugfixes/better errors/etc. and...
and I think it requires intents
fallen patrol
is the ``http_session`` smth from dpy or not? I can't seem to find from anywhere...
!d aiohttp.ClientSession
stable mountainBOT
cursive relic
;-;
fallen patrol
http_session is an instance of that ^
cursive relic
that explains stuff
fallen patrol
.get() is actually a GET request to the provided url endpoint
elder belfry
yes, you need our license to appear somewhere
will it be sufficient to just put the link of the repo in the readme/commit message ?
fallen patrol
will it be sufficient to just put the link of the repo in the readme/commit mess...
ianal, but no
depends on how much is copied
cursive relic
!d aiohttp.ClientSession
is this inbuild to python, yes?
stable mountainBOT
fallen patrol
Itβs included in d.py
that's a very confusing description imo
cursive relic
that's good to know, thx! I'll get back to work
fallen patrol
it's a dependency of discord.py, which means if you've installed discord.py it'll already be installed
gritty wind
will it be sufficient to just put the link of the repo in the readme/commit mess...
The terms of the license include adding the license to your code:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
gritty wind
it's a dependency of discord.py, which means if you've installed discord.py it'l...
Umm⦠okay sure
brazen charm
MIT is as simple as it gets, you need to include the license including the copyright license somewhere with your software. a separate license file would be the usual choice
or add the copyright notice if your project is also mit
clever wraith
Stop sending messages in dev log joe
cursive relic
MIT is as simple as it gets, you need to include the license including the copyr...
if I want to include it in seperate file, what would it be named like?
elder belfry
depends on how much is copied
ohh crap 
so is there any suggested/best-practice way to include them then?
like I usually copy template files ig? like github workflow, tox.ini ... etc
fallen patrol
IANAL, but it does say substantial and a configuration file hardly seems substantial to me IMO
brazen charm
if I want to include it in seperate file, what would it be named like?
license_3rd_party or something, it should just be clear, it'd also be good practice to include what it applies to
cursive relic
so far I have just added the license to every file that copies smth from the bot
frozen galleon
Hi, I found this in the code-base: https://github.com/python-discord/bot/blob/main/bot/exts/moderation/clean.py#L286-L290 (Introduced in PR #1793). As the comment below suggest, we can only allow up to 100 messages for bulk deletion, that's fine and all, but the condition itself forces the number to be equal (== ) to 100 messages, rather than forcing it to be less, that's most likely a typo and should probably be fixed
stable mountainBOT
bot/exts/moderation/clean.py lines 286 to 290
if len(to_delete) == 100:
# Only up to 100 messages can be deleted in a bulk
await channel.delete_messages(to_delete)
deleted.extend(to_delete)
to_delete.clear()```I assume it was missed in the review since the PR edited a lot of things and the commit that changed this was actually quite old (Nov 22, 2020) but I'm fairly sure it's wrong since I don't see any reason to enforce the length to be exactly 100, especially with that comment saying the number just needs to be less or equal to 100
thorny obsidian
It's not a typo. It's adding messages to a queue to be deleted, if it hits 100, then it'll delete it. If you look further down the code, you'll see that if there's anything leftover (i.e. >0) then it'll call a delete on those messages.
If you changed that line to be <=, it would delete messages one by one, not bulk
frozen galleon
oh, I see, my bad then, I just quickly glanced over it in a diff from another PR that didn't show the rest of the code before it and I was too lazy to read over it since it just immediately seemed wrong, in that case it's fine
thorny obsidian
Yeah, in general I recommend reading the context of the code and the code before and after a snippet.
molten perch
Hey! Sorry that I keep promoting my PR π
So, I would love to have some reviews on api#30.
It's not that bad, the core part is ~6 files in total, and there are also taskipy tasks that make things easier, like test , that will run tests inside docker, so that no-one has to bother with setting things up.(More about on that, in the README in the PR)
Thanks! π
last patio
I will try to take a look at it soon, I'm currently looking through other PRs which need a review
https://github.com/python-discord/site/pull/572 this is pretty important, since it means inconsistent data
molten perch
Indeed, would love to have some reviews on that as well. π
vale ibex
Cool just checked myself and it looks good
Did you want to rebase off main rather than add another merge commit? π
Auto merge should just merge it after ci then
molten perch
Yeah, I guess so.
I wanted to rebase but I might have misinterpreted it π
vale ibex
You did, but more commits have been merged to main since π
So it's out of date with main again
molten perch
Oh, so I managed to do it in the end, cool. π
So, do I rebase it again, or just auto-merge it? π
vale ibex
auto-merge will merge when all criteria are met, one of which is being up to date with the branch it's being merged into (main)
if you rebase onto main and forcepush, it will bring it up to date, so then auto-merge will merge it
sometimes we use the "update branch" feature in GitHub to bring a branch up to date with main, but that adds a merge commit, so is nice to avoid sometimes
molten perch
sometimes we use the "update branch" feature in GitHub to bring a branch up to d...
Yes, I've done that a five times, never ever again. π
molten perch
auto-merge will merge when all criteria are met, one of which is being up to dat...
I'll do that, thanks for the heads-up. I was always confused about that π
I have finally conquered all the main Pydis Open Source Projects π
summer garden
Are we using a wrapper for message components with lancebot?
summer garden
Has that implemented message components though?
fallen patrol
hmmm, given that dpy is going to break in only a few months (5 at max) what is the plan, if any, from pydis yet?
gritty wind
Still hasn't changed
fallen patrol
I believe that discord.py's implementation of message components are broken or destroyed with most forks if any other features interaction api features from the forks want to be used.
ah gotcha, okay
green oriole
And it isn't gonna break to my knowledge
gritty wind
Wait for the first library to come out as a strong contender for the main library
And what change are you referring to
Yeah what Akarys said
green oriole
We have post april 2022 support
vale ibex
It will, discord have confirmed that post April you need to supply the message content intent when connecting to the gateway if you want message contents
fallen patrol
And it isn't gonna break to my knowledge
in April you will need to define a different intent number, doable with a monkey patch
vale ibex
So we either monkeypatch intent flags (not too hard) or migrate
gritty wind
That much should be easy to patch in though
Not as difficult as migration if we don't have a good contender for sure lol
fallen patrol
but at that point unpatched dpy will be broken, so I'm not wrong
fallen patrol
it's not officially documented yet, nor was announced when dpy shutdown, so I'm not sure how that would be possible
that intent you're thinking of is a different intent
green oriole
Really? The intent number has been known for months, hasn't it?
fallen patrol
there's a message intent and a (TBA) message content intent
fallen patrol
it's likely 1 << 15 but again, not yet documented.
green oriole
Could we make our own fork for small things like that, so we don't have to monkypatch for now?
vale ibex
since I checked it
We could fork it, or monkey patch it
not too much difficulty either way
green oriole
Yeah
vale ibex
but at that point, I imagine a contender for migration will eb fairly obvious
gritty wind
Patch is my preference, but that's not very relevant right now
fallen patrol
given that there's... 7 bots for here now a fork might seem more useful
vale ibex
disnake seems like the winner right now, but 4 months can change that
green oriole
Yeah, we will see about it in a bit
That's an option we still have if we don't want to migrate
vale ibex
Yea
gritty wind
given that there's... 7 bots for here now a fork might seem more useful
We'll have bot core by then hopefully, and also its just a small snippet tbh.
vale ibex
So, to answer your question arl, we haven't really got a plan other than wait it out and decide later π
fallen patrol
disnake seems like the winner right now, but 4 months can change that
yeah also novus, possibly
vale ibex
but we have some things we can do to extend the time we can wait, such as patching
fallen patrol
I'm against py-cord and some other libs for some personal reasons
vale ibex
Since the most important thing to us in a lib is that has decent maintainers
vale ibex
and we can't really tell that right now, since most of these libs have only been around for 4-5 months
fallen patrol
since it's new, some of the forks ended up as some people's first big project
vale ibex
yup
vale ibex
I haven't used it really, nor kept up with any drama, if there is any
Not really too interested in it either π
fallen patrol
tldr py-cord is poorly maintained and basically rewriting the whole lib afaict
vale ibex
Ah
Well the main thing for us will be a lib that has minimal changes to the core implementation
summer garden
am I wrong in thinking buttons and dropdowns are not implemented in dpy?
fallen patrol
I've been participating on the disnake repo for a few days since I like it but not it itself
vale ibex
am I wrong in thinking buttons and dropdowns are not implemented in dpy?
They are implemented in d.py
only in the master branch alpha version
summer garden
Is that the one we're using? π
vale ibex
Yea
green oriole
Yup
vale ibex
Docs are here https://discordpy.readthedocs.io/en/master/
green oriole
Someone had fun migrating 
vale ibex
lol
vale ibex
Only things it doesn't have right now are context menus and slash commands
fallen patrol
Well the main thing for us will be a lib that has minimal changes to the core im...
disnake is drop-in rn. aaaand it implements slash commands and other features too so once a switch to it occurs its not too hard afterwards.
they do have a disnake discord shim, to not have to rewrite all of the imports, but they don't recommend using it
vale ibex
and anything new that discord adds
fallen patrol
and anything new that discord adds
like channel banners, scheduled events, text in voice, timeouts, etc
fallen patrol
I highly recommend looking at a fork sooner rather than later given upcoming timeouts which can replace muting and work much better
a discord side block on all server interaction which does not relate to permissions at all
green oriole
Right now we have an excellent infraction system that works perfectly, while it will surely be better to use the native impl there is no rush
It should also be a really small change
famous last words
patent pivot
I highly recommend looking at a fork sooner rather than later given upcoming tim...
yeah this isn't something we urgently need
our system will always be more granular, so we need to see if we want to integrate with this or stay as is
fervent sage
But it is something that will become more painful to implement the longer we go without it
cold island
Why will it become more painful?
fallen patrol
forks will deviate and seperate from the initial dpy implementation. Right now, most of them are mostly a drop in replacement. As time goes on, they will have their own features, and as discord gets more features, they will become breaking features and need to be implemented.
at least, i think that's what @fervent sage means
fervent sage
Why will it become more painful?
Kinda what arl said but also the more services we have using dpy the more places we will need to update
patent pivot
forks will deviate and seperate from the initial dpy implementation. Right now, ...
discord donβt tend to have many breaking changes - if they do itβll be on a new api version
patent pivot
Kinda what arl said but also the more services we have using dpy the more places...
thatβs true yeah - we should actually think about that for some upcoming projects
fallen patrol
discord donβt tend to have many breaking changes - if they do itβll be on a new ...
hmmm, i wonder if the next voice feature will be v10 then
patent pivot
text in voice channels?
fallen patrol
yeah
patent pivot
on current api version but doesnβt seem to break dpy from my last testing, not sure if thatβs still the case
fallen patrol
you have text in voice channels here?
patent pivot
not here
fervent sage
fortunately in terms of new channels discord.py just kinda ignores those events
djs and eris both outright crash 
at least they used to
had some great fun crashlooping zeppelin with stage channels back in the day lol
brisk brook
forks will deviate and seperate from the initial dpy implementation. Right now, ...
Exactly, that's a good thing
What if we pick one and it deviates in a way we disagree with?
fervent sage
Thats not really a valid point since any of them can at any point
Like sure it could happen but it could always happen
short of maintaining our own fork (which seems like a bad idea) you cant solve that
brisk brook
Right, that's true. But arl's point is that we should switch ASAP so that it will be easier to migrate because forks will eventually deviate - and I think the opposite. If we wait and see what general direction they're heading it is easier to make an educated decision.
fervent sage
It's been a couple of months, and we can already see which general direction they're heading
if any of the major forks were going to make drastic changes to the paths they're on it would have happened by now
patent pivot
Right, that's true. But arl's point is that we should switch ASAP so that it wil...
I think I'm here
brisk brook
Oh yeah, I don't think any of them will make any huge changes. They all seem to have the goal of continuing the work of discord.py
fallen patrol
IMO the drastic changes are about to happen, with text in voice, channel banners, and timeouts. That will be where implementations will be separate between libraries
brisk brook
It's been a couple of months, and we can already see which general direction the...
I disagree with this, I genuinely do not see the difference between the forks without jumping into the diffs.
fallen patrol
Right, that's true. But arl's point is that we should switch ASAP so that it wil...
I suggest a higher priority, but not asap fixes
vocal prairie
I may be missing information, but I don't think Discord has made any really major changes to the API since the deprecation, so I'm of the opinion that we should wait a while longer to see their general direction (so I'm with Blue).
fallen patrol
There's still until april, but by april there must be a solution--even if that's just a monkeypatch
brisk brook
IMO the drastic changes are about to happen, with text in voice, channel banners...
That's great, we can wait until that and then make a good decision when that has happened
fallen patrol
IMO the drastic changes are about to happen, with text in voice, channel banners...
!remind 3w these released yet?
stable mountainBOT
fallen patrol
slash commands are a huge difference between them
ye, and that's also where I'm curious-- which forks break the discord.ui.view module?
brisk brook
slash commands are a huge difference between them
That's a good point, disnake does the fastapi-style and the rest seem to try to have a good compatibility with discord.ext.commands
fallen patrol
Since as those features are adopted more on pydis bots, that namespace may be broken by forks if implementing slash commands is also wanted
fervent sage
That's a good point, disnake does the fastapi-style and the rest seem to try to ...
Im confused; nextcord doesnt yet have slash commands, pycord's dont work in cogs...?
hence neither can be compatible surely
fallen patrol
also pycord overall has not the best maintainers IMHO
brisk brook
hence neither can be compatible surely
I am saying that they're trying to go that route, as in be as similar to that system as possible.
fallen patrol
Since as those features are adopted more on pydis bots, that namespace may be br...
I haven't double checked recently, but a few weeks ago I tried a few forks with my bot having a few view components and they had issues with the discord.ui.View module, enough so that they need to add more support or remove it
fallen patrol
Im confused; nextcord doesnt yet have slash commands, pycord's dont work in cogs...
personally i don't think the maintainers of pycord or nextcord are super decent maintainers but that may just be me
fervent sage
fallen patrol
wait are you maintaining one of them?
fervent sage
I may own the nextcord discord
fallen patrol
oh, oops
fervent sage
dw
fervent sage
RPS contributes to like 4 of the major py libs
fallen patrol
interesting commits, but neat
fervent sage
My brain is currently split across like 6 libs
fallen patrol
what are all of them?
fervent sage
2 of my own, nextcord, pycord, disnake, and hikari 
dusky shoreBOT
fallen patrol
not sure what they're expecting, they've contacted the owner and the owner said no.
fervent sage
doesnt pypi make clear that requests like that wont happen?
fallen patrol
yes
and if you're making an issue that sounds more like you're trying to convince someone of solving a dispute.... π€¦ββοΈ
For what it's worth, the current project has just 57 lines of meaningful code, while my project has thousands.
also this line just makes me dislike it in all
anyways, that's why I dislike pycord, but not the right channel at this point lel
patent pivot
ultimately I think taking Blue's approach and holding off for now is the best bet, things are very much still up in the air with all forks but we've got plenty of time to think about things
it's not likely we're going to use the native muting functionality anytime soon, and we wouldn't change libraries just to support text-in-voice either
fallen patrol
it's not likely we're going to use the native muting functionality anytime soon,...
so, it'll be like threads and not publicly enabled for a few months+?
patent pivot
maybe, maybe not, we'll see if the feature has a place and discuss it, that's something for Hemlock's domain, but usage of it would be blocked until we could support it in @stable mountain
the current system works well enough
fallen patrol
π
brisk brook
pypa/pypi-support#1438
Holy shit they won't win this. Considering they even got a response from the owner explicitly declining that request.
Besides, pycord has been a thing for like a year now by kyb3r: https://github.com/kyb3r/pycord
I always found it rather shitty that they took that name, though I don't know if they've (the maintainers of the discord.py fork called pycord) got permissions to use the name by kyb3r.
fallen patrol
Holy shit they won't win this. Considering they even got a response from the own...
yeah, its pretty bad what they've done imo. Furthermore they apparently will release their actual library in January, as currently what is on pypi is an exact copy of discord.py 1.7.3.... and nothing else. Not even sure why the pypi name matters to them so much as they don't even use it (so far)
fallen patrol
Holy shit they won't win this. Considering they even got a response from the own...
but also that pycord by kyb3r mentions in the readme, its a dead project, with a most recent commit in December 2019 Β―_(γ)_/Β―
brisk brook
In response to this text - why name your package that way if it was already taken. That's on you!
brisk brook
but also that pycord by kyb3r mentions in the readme, its a dead project, with a...
I still consider that a shitty thing to do, imagine if someone comes out with a library called discord.py in two years
fallen patrol
wasn't there a fork called that?
brisk brook
wasn't there a fork called that?
fallen patrol
actually I shouldn't joke about that, someone will think I'm being serious
brisk brook
Oh I get it lmao
fervent sage
To weigh in on the pycord pypa thing: yeah that's a really shitty request imo
after being denied the name by the package owner they have no right to try and take it by force
trail pilot
I think it's also on them for not being explicit enough about py-cord versus pycord and as Bluenix said why choose that name if it was taken in the first place
brisk brook
In response to this text - why name your package that way if it was already take...
Like, what's happening here is that this ffmpeg wrapper is "getting hurt" (I don't know what the best way to put it is) because they (discord.py fork maintainers) chose a bad name. They are the ones who chose a bad name and they (discord.py fork maintainers) are to blame.
fallen patrol
Oh I get it lmao
except actually, discord.py is the second name for that project, it used to be called something else.
patent pivot
pydis rewrites to disco π
fallen patrol
used to be called pydiscord
brisk brook
As for the original owner's claims that "any change to the name would break all projects which relies on it," I reviewed the listed dependents on GitHub. From the code in each of these repositories it is clear that all of them depend on my project, not the current project. I could not find any evidence of any repositories that depend on the current project, across all of GitHub.
Tl;Dr "I don't think there's any damage made because the original owner's code is useless"
Did I get that right? I am confused
remote wigeon
reject asyncio revert to gevent
patent pivot
it is a weird one, it is off topic for this channel though
fervent sage
where is it not offtopic?
patent pivot
where is it not offtopic?
off topic, probably, or I guess #discord-bots
trail pilot
Did I get that right? I am confused
You got that right
patent pivot
but nitty gritty about package naming is not for here
brisk brook
it is a weird one, it is off topic for this channel though
Good point we've ranted enough lol
fervent sage
its not off topic nor purely discord related though?
fallen patrol
patent pivot
it's not dev-contrib related, either
brisk brook
its not off topic nor purely discord related though?
python-general? /s
patent pivot
libraries out there are relevant, but pypi name claims (and the validity) are not lol
fervent sage
I'd suggest a new channel but in 3 months when we agree on a name it will be too late π
remote wigeon
fallen patrol
I'd suggest a new channel but in 3 months when we agree on a name it will be too...
and it'll already be taken by joe last year
brisk brook
Ah welp to get back on topic then I guess.. tomorrow will be review day for me π
Got a few PRs that I need to test π€·ββοΈ
austere hornet
https://tenor.com/view/reject-modernity-return-to-monke-monke-gif-19167526
Lmfao
magic arch
Alright, in the main bot, constants.Colours.bright_green exists for some reason and is used exactly 3 times. Twice in the help channel system (DM to author and available message) and once for the !modpings off command. There's also constants.Colours.soft_green which is incredibly similar and used 19 times in 12 files. Can we get rid of bright_green and just use one green color?
Side-by-side comparison
For comparison, orange and soft_orange have a huge difference, and soft_red doesn't even have a red to compare with.
austere hornet
<@!853309207922212965> Could you add some example executions (screenshot of runn...
!remind 8D this
stable mountainBOT
austere hornet
!remind delete 3568
stable mountainBOT
austere hornet
<@!853309207922212965> Could you add some example executions (screenshot of runn...
!remind 7D this
stable mountainBOT
austere hornet
!remind edit duration 7D8H
stable mountainBOT
austere hornet
!remind edit duration 3569 7D8H
stable mountainBOT
cold island
Alright, in the main bot, `constants.Colours.bright_green` exists for some reaso...
The color choice seems pretty deliberate, and they seem distinct enough to me. Is there a technical reason to not have two greens?
The bright green embed color is used with an icon that has the same color, Using soft green with that icon won't really look good
stable mountainBOT
!remind 5d Review sir-lancebot#901
It has arrived!
magic arch
I guess there's no technical reason, it's more of just a visual annoyance here:
cold island
I guess there's no technical reason, it's more of just a visual annoyance here:
A visual annoyance as in, it's too bright?
magic arch
As in they're very slightly different, in the same place, and both related to the help system. It makes it seem like they should be consistent but they're not.
magic arch
Yeah that seems much more reasonable than what I did π
vale ibex
I'd quite like us to get away from named colours like this