dev-contrib

[fallen patrol]

still.... I am wondering about a cli that makes like

all of the bot configurations

like

[outer oasis]

do it then

[fallen patrol]

botstrap makes pybot easier to set up than lance now

I have to set seven variables myself????? 🤮 /s

[patent pivot]

we should figure out the emotes thing

[fallen patrol]

app emojis

[patent pivot]

app emojis is true

[fallen patrol]

I'll let you know when I've figured this problem out myself

trying to solve it for Monty

ok bot#3417 is now ready for review

[dusky shore]

GitHub

[bot] #3417 refactor: update botstrap to be more beginner-friendly

[fallen patrol]

hmmm i could add some basic config to it to add config for like 🟢 and stuff where it makes sense

https://github.com/python-discord/bot/issues/3374 can be closed btw

crazy

[patent pivot]

my thought a while ago (and this was probably pre-app emojis was):

• A) we host a bot that people can add, it adds all the emotes from the emote servers and then leaves
• B) we host a web app that has people OAuth2 in, paste a bot token, confirm they own the bot token (or are in the team that owns it), confirm they have perms in the target guild and then add the emotes using their app

i am unsure if either of those are good ideas, but they'd be fun

[fallen patrol]

they would be fun

mainly my idea is thinking what's the best way to support integrity checking without

• crashing
• making startups way too long

at least with bot we have botstrap and if we turn it into a cli we can add a sync app emojis command

[patent pivot]

yea

the main thing is fetching those emotes

[fallen patrol]

for monty I've been considering a sort of complicated thing using github and last_changed

you have the current hash in the bot environment, right?

https://api.github.com/repos/OWNER/REPO/contents/assets/emojis?ref=commithash

and comparing the ids of all of the emojis to api.github.com/repos/{owner}/{repo}/commits/?path=assets/emojis

[thorny obsidian]

Is there an example and filled out .env.server?

I don't know what format pydantic wants the webhooks envvars in and the contrib guidelines are vague about it

[fallen patrol]

kat is this for the internal test guild?

#Webhooks
webhooks_big_brother__id=1430330254509211691
webhooks_big_brother__channel=1430304585095970951
webhooks_dev_log__id=1430330295152017549
webhooks_dev_log__channel=1430304584198525126
webhooks_duck_pond__id=1430330306732359820
webhooks_duck_pond__channel=1430304584626077710
webhooks_incidents__id=1430330317696139305
webhooks_incidents__channel=1430304584382939261
webhooks_incidents_archive__id=1430330335199232052
webhooks_incidents_archive__channel=1430304584382939262
webhooks_python_news__id=1430330340710416496
webhooks_python_news__channel=1430304583996932110

[thorny obsidian]

I have my own test server, I jsut want to know the format

[fallen patrol]

use botstrap, it will create these webhooks for you and set that up

this is why i was wondering #dev-contrib message

[thorny obsidian]

I'd rather not. I'm aware of botstrap.

[fallen patrol]

fair

if you have a reason you'd like to share, I'm curently working on both botstrap and the contrib guidelines about this very thing, so if you have more to add, I'd appreciate it

[thorny obsidian]

I have had my server forever. It's setup how I like it. I would rather stab a fork into my eye than have to migrate.

I just want to know the webhook config format that pydantic won't bitch about

[outer oasis]

Note to self: keep forks away from Kat

[fallen patrol]

I just sent it above

[outer oasis]

Yeah that should be right

[thorny obsidian]

Yeah, that's still not enough information for pydantic looking at the docker logs

[outer oasis]

[fallen patrol]

wild. That's generated with botstrap

[patent pivot]

what does pydantic want?

[outer oasis]

(what's it upset about?)

[fallen patrol]

Put it in .env if using docker

[thorny obsidian]

and it keeps going. I have the channels defined above and I figured pydantic would be able to pull from that

[thorny obsidian]

Following the contrib guidelines tells me to build the .env.server manually
I'm going to follow the guidelines

[fallen patrol]

    volumes:
      - .:/bot:ro

in theory this should mount .env

might say that, but I have a PR in the works that changes it to .env

I also have a 3 year old test guild

[fallen patrol]

can you provide your webhook config here?

[thorny obsidian]

It's what you have posted above but with my own webhooks for my personal testing server

[fallen patrol]

Okay, can you post what you have anyways?

[thorny obsidian]

It's only the values that change, so I don't see the need

[fallen patrol]

Then I don't know what to tell you. it works for me

[thorny obsidian]

I'm posting here to ask the core devs as well, not necessarily specifically you

[fallen patrol]

you do need to make sure that webhooks_big_brother__channel is also set

the double __ is important

[thorny obsidian]

So that part is what I did not know and what the contrib guide does not detail.

[patent pivot]

ahh yeah pydantic attributes

we should call that out with an admonition

it is not intuitive

[thorny obsidian]

Good lord I miss our old constants and config.yml setup compared to this

Can pydantic not reference to me setting it up earlier? Do I need to list the channel ids twice in 2 different places?

[fallen patrol]

I'll see if I can add it joe

[outer oasis]

It's this
https://github.com/python-discord/bot/blob/main/bot/constants.py#L19
Because Webhook is a nested class here, you have to use the delimiter to move from _Webhooks into the Webhook

[stable mountain]

bot/constants.py line 19

env_nested_delimiter = "__",```

[fallen patrol]

ugh

[patent pivot]

if they're in the same env file you can reference later with $variable but the model I would guess does not automatically look elsewhere at the channel constants

i do think the fact we're using __ is kind of a sign that maybe we've gone a little down the wrong path with configuration

[fallen patrol]

class _Webhooks(EnvConfig, env_prefix="webhooks_"):

    big_brother: Webhook = Webhook(id=569133704568373283, channel=Channels.big_brother)

this code requires channel to also be provided

given we only seem to use channel to create it if it doesn't already exist, and that is already mapped to something.......

[fallen patrol]

we could have supported toml or even the old yaml files 😔 https://docs.pydantic.dev/latest/api/pydantic_settings/#pydantic_settings.TomlConfigSettingsSource

[patent pivot]

yea this is kind of what I think

and I think maybe that's the direction botstrap should go -- I think something like ```toml
[webhooks]
big_brother = { id = "1430330254509211691", channel = "1430304585095970951" }

is obviously cleaner than the `__` stuff

[fallen patrol]

fixed

https://docs.pydantic.dev/latest/concepts/pydantic_settings/#nested-model-default-partial-updates

we only need to set the ID now

want me to seperate this from the other PR?

[patent pivot]

nice one

nah if you're changing botstrap this is fine to go in -- just make sure it has a decent commit message

[fallen patrol]

all right

I also made an issue about cli but I didn't want to feature creep this one

[patent pivot]

noice

[fallen patrol]

yay

its still a nested model for now because I don't want to break existing configuration

[patent pivot]

i think nested is fine -- i just think it needs to be something that's not an env file

which makes maintaining back compat down the line easier

[fallen patrol]

yeah

I won't add support atm but we can consider moving to toml files

Problem with that is if we want to support

• autogen file
• env file
• user set toml file

which is a lot of configuration files lol

so botstrap can still be a first class configurable way to set the configuration

but then on that side, if we do switch to toml we can probably update that one configuration in place with tomlkit

and put autogen in the same file...

though it is nice to have "configuration file that my autogen will never touch" and seperation that way

[fallen patrol]

bot#3420

[dusky shore]

GitHub

[bot] #3420 Upgrade configuration to use a toml file over .env.server

[fallen patrol]

what's the failed_file emoji look like?

Emojis.failed_file

[patent pivot]

[fallen patrol]

and is nomination_discussion also a thread?

[fallen patrol]

thanks, chose some twemojis to use as replacements:

that should be enough for now to provide some coverage

it sure fixes this finally

[patent pivot]

no that's a channel

[fallen patrol]

its missing in the template

are you able to add?

also, is the template of the test server or is it a different server which literally exists for the template?

[patent pivot]

yea can do one sec

should be sorted

i believe it's just the one of the internal test server -- which is what I just updated lol

so if it's not that

then yikes

[fallen patrol]

lol

https://discord.new/zmHtscpYN9E3

discord moment

it apparently caches templates forever

[patent pivot]

it's in that template when i click at least

[fallen patrol]

ye i see it too

i had to reload my client though lmao

it literally caches it forever

[fallen patrol]

@patent pivot yay it worked!

[patent pivot]

hell yeah

that's very cool

[fallen patrol]

off by one 😭

...that was an easy fix

[fallen patrol]

okay so it may create roles in the wrong order

[fallen patrol]

Discord pls

[fallen patrol]

!ot this channel is for contributing to @pydis projects, not for other projects. If you're wanting to share a project with others, I recommend an off topic channel

[stable mountain]

Off-topic channel

#ot2-never-nester’s-nightmare

Please read our off-topic etiquette before participating in conversations.

[harsh hill]

im so soryy

i thought this was the off topic

thank you senpai

[fallen patrol]

I have just realised that with an emoji regex I can actually add app emojis to botcore

except why would i bother doing that when i can just.... upload the emojis to the guild

lol

[fallen patrol]

darn pep api

!pep 401

[stable mountain]

**PEP 401 - BDFL Retirement**

Status

Rejected

Created

01-Apr-2009

Type

Process

[fallen patrol]

by using the peps.json api we lose the easter egg 😔

on the other hand, maybe its a feature that it doesn't show it in the embed

[fallen patrol]

Should display names be counted in the name filter, even if the user has a nickname set?

[patent pivot]

hmm probably not

well

hmmmm

I will ask internally

[patent pivot]

the current behaviour is that they're not counted?

[fallen patrol]

ye not included #ot0-psvm’s-eternal-disapproval message

whups didnot mean display name, meant..... whatever that user field is again

...whichever field is the User Display Name, not to be confused with the server/guild Nickname or the Username

[patent pivot]

nice we don't have a short link just for contributing

[fallen patrol]

https://pydis.com/contributing.md

huh, you're right

(arguably contributing.md should be it)

maybe there's a reason why it doesn't...

https://github.com/python-discord/site/blob/91767256be6729355c78d08ddcde8233c8a5bf3e/CONTRIBUTING.md?plain=1#L1-L3

[stable mountain]

CONTRIBUTING.md?plain=1 lines 1 to 3

# Contributing Guidelines

The Contributing Guidelines for Python Discord projects can be found [on our website](https://pydis.com/contributing.md).```

[fallen patrol]

eh fair

seems that might benefit if it pointed to https://www.pythondiscord.com/pages/guides/pydis-guides/contributing/site/ or https://pydis.com/contributing-site

[patent pivot]

yea I think we should probably link to both the guidelines and guidance from within those files

[toxic wren]

#dev-announcements message congrats!

[toxic wren]

btw the slash commands issue is on hold for now(I'm very busy this week), I will provide updates later. Thanks!

[fallen patrol]

!remind 3h open issue / discuss about "Click here to delete recent paste" having a context button or link

[stable mountain]

You're the boss!

Your reminder will arrive on <t:1761194770:F>!

ID: 8076

[fallen patrol]

TLDR:

[stable mountain]

It has arrived!

Here's your reminder: open issue / discuss about "Click here to delete recent paste" having a context button or link
[Jump back to when you created the reminder](#dev-contrib message)

[fallen patrol]

...how do renovate and dependabot rename a pull without causing an event to show up?

https://github.com/python-discord/forms-backend/pull/307

I've never seen that before

[patent pivot]

what do you mean

there are rename events in that log

rename events might only be shown to repo maintainers?

[fallen patrol]

My eyes are just tired ><

• mobile contrast

[patent pivot]

hahaha

[fallen patrol]

I wish GitHub exposed discussions on the rest API 😔

And organization level discussions

meta#93 😔

...tbh I think lancebot has a GitHub token? We could implement discussions using the graphql API

[undone nimbus]

I just noticed an issue within pybot on https://discord.com/channels/267624335836053506/1431166143883640934 where it didn't automatically close the forum post when OP left the server, eivl had to close it himself

[patent pivot]

that's not an issue, it's intentional behaviour

the thread would still timeout after the regular wait period (1 hour from last post).

in the event a user leaves we put the message in the channel to make folks helping aware a user has left, but we purposefully don't close the channel in case people are still having a discussion / a wider conversation has started there.

[undone nimbus]

well fuck, thought I noticed a bug

[vale ibex]

Yeah sorry, we don't write bugs around here

[toxic wren]

https://github.com/python-discord/bot/issues?q=is%3Aissue

[celest charm]

404

[green oriole]

I believe a functionality was added to snekbox where you can write local files

Is it enabled by default in the docker image?

I'm tempted to use snekbox in a different project to evaluate Blockly scripts

[tame pebble]

Those files are written to a tmpfs IIRC

[tame pebble]

I think so yes

[green oriole]

Hmm I'll have to look at how to inject a custom config then

[fallen patrol]

@patent pivot May I get the hacktoberfest-accepted label on sir-lancebot#1702, site#1567, bot#3413?

[dusky shore]

GitHub

[sir-lancebot] #1702 chore: assert uv lock is locked in ci and pre-commit
[site] #1567 chore: assert uv lock is locked in ci and pre-commit
[bot] #3413 fix: check for unknown/None content_type for pastebin uploader

[patent pivot]

done 👍

[fallen patrol]

Thanks!

I'll fix 3413 soon and mark it as ready but covid vaccine took my energy the past few days and now I'm behind on my classes

I really do want to add pyright to bot.... So much so that I'll probably add it to my local files at minimum lmao

Unfortunately, fixing typing issues can sometimes result in new bugs

[celest charm]

what do we actually get from adding type checking to CI?

[patent pivot]

if we do it i think we do it in a non-mandatory way

supporting github annotations is a nice feature that would highlight a decent few things especially if we are baselining it

especially if we generate that baseline based on the main branch, to very specifically only highlight new typing issues raised by a PR

[celest charm]

maybe we should use mypy instead and just diff the results or something like that?

[patent pivot]

basically seperate it out of the lint workflow into it's own thing that:

• runs, generates a baseline on current main
• checks out PR
• runs again to highlight issues and add them as annotations
• does not fail the PR if problems are reported

[celest charm]

(there's a mypy_baseline package, haven't used it)

[patent pivot]

main reason i was thinking basedpyright is it's probabyl closer to the pylance that a lot of folks already have in VS Code, I think most of our contributions probably come from people using that

but also basedpyrgiht supports github annotations natively

[fallen patrol]

I would use mypy I don't really care

[celest charm]

oh, mypy doesn't have this?

[patent pivot]

not natively afaik

[fallen patrol]

This is so minor (to add support for) and so cool too

(genuinely it's just GITHUB_ACTIONS==true use a different output mechanism)

[patent pivot]

https://github.com/auchenberg/volkswagen

[fallen patrol]

Lmao

Oh no.... There's no contributing guide for sir-robin, is there

[patent pivot]

probably not

though it's probably somewhat similar to lance i'd imagine

[celest charm]

pyright/basedpyright will add on the order of a minute to the CI time

assuming it's only ran once and not also on main

actually no, it's significantly faster 🤔 I must be thinking of a different project

yeah, it's about 10s on my computer

[fallen patrol]

When can we move botstrap to its own repo? 👀

I genuinely want to unify all of the configurations (backwards compatible) and then make botstrap able to generate all of the configs lol

[patent pivot]

yeah -- but remember we'd run it in parallel anyway so it'd be negligible compared to the docker build step

[thorny obsidian]

Sir Robin operates a bit differently and looser compared to other bots because it supports a fairly specific and time-dependent need

[patent pivot]

could put something in core to try move towards unifying it.

i think our crazy env files are getting a bit silly

[fallen patrol]

Real

[thorny obsidian]

Also I am aware of the Advent of Code changes. We're handling the adjustments to that internally and when we get a bit closer.

[fallen patrol]

I made an issue so it would be publicly tracked, as I wasn't sure it was known. I saw a lot of advent of code issues from 2024 or before that I also want to see about tackling if you'd like help

[thorny obsidian]

There are other things associated with Advent of Code that extend beyond just the bot, so it's being tracked and handled elsewhere until we get closer and have a confirmed plan on how we want to approach it.

[fallen patrol]

Even issues like sir-robin#107?

[dusky shore]

GitHub

[sir-robin] #107 AoC: put specified user within the top 10 leaderboard if user is in the top 10

[thorny obsidian]

? I'm not sure what you're asking. I'm just letting you know that we are aware, have been aware, but are tracking the aoc change separately because there are other parts of advent of code that are not associated with Sir Robin that we also need to deal with.

[fallen patrol]

I'm asking if I can make that change above, not the other ones.

There's a few AOC issues I've been interested in implementing: sir-robin#65 sir-robin#107

[dusky shore]

GitHub

[sir-robin] #65 AoC: Use a forum channel for per-day spoiler discussion
[sir-robin] #107 AoC: put specified user within the top 10 leaderboard if user is in the top 10

[thorny obsidian]

That was not clear since the issue hadn't been commented on. I'll take a look at the issues in a week or so and triage what's still relevant, what's not, and what our approach is going to be.

[celest charm]

Idea @patent pivot: if we do adopt basedpyright in a non-intrusive way, we could adopt some of the rules in an intrusive way because violating them is Just Wrong

• error on reportGeneralTypeIssues which would catch nonsense annotations like:

def incorrectly_annotated(banana: 69 = 420) -> any:
    ...

• error on deprecated typing aliases like typing.Callable. Because ruff's UP006 rule is buggy and does not catch all of them.
• error on reportMissingTypeArgument to require all generics to have type parameters, like list[Whatever] or list[Any] instead of just list, and discord.Context[OurBot] instead of discord.Context
• reportUnboundVariable and reportPossiblyUnboundVariable: yes, this thing

[stable mountain]

bot/exts/help_channels/_channel.py lines 194 to 199

try:
    # Fetch the post again, to ensure we have the latest info
    post = await get_or_fetch_channel(bot.instance, post_id)
except discord.HTTPException:
    log.trace(f"Not closing missing post #{post} ({post.id}).")
    return```

[fallen patrol]

Hadn't commented because I didn't see a label approving the issue, I normally wait until they're approved for that

But ye I'll go comment now

[thorny obsidian]

Like I said above, Sir Robin operates differently than our other bots because of it's highly specific usage and close ties to the Events Team. Pretty much all issues created are good for people to claim, comment with their approach, and things move forward from there.

If there's an issue that I think is a no-go, I'll be pro-active about it.

[fallen patrol]

Sounds good

[fallen patrol]

Yeah fs

Do we want to implement type checking on lance first before we implement on bot?

Or bot before lance

[celest charm]

power move: abandon basedpyright, launch proprietary VSCode with Pylance in it in CI

If the screenshot is more than x% red color, fail the pipeline

[fallen patrol]

-# and we set the theme to hot dog stand

[celest charm]

enqueue red alert theme

[fallen patrol]

Overcomplicate the configuration by putting it all in pydis core, and allow bots to extend it further... Hmmmmmmm. .... There has to be a good way to do this. Though yeah, for the most part it IS duplicated code on three or more repositories.

The issue with putting it all in one place is now there's no easy way to be tracking the minimum number of config variables for a bot

And I think that latter detail is significant

[patent pivot]

ehh

[fallen patrol]

Eg if we have a channels object that contains all of the channels of pydis, there's nothing stopping lancebot from using more than the few variables it uses now and leaving contributing out of sync again

Even if there's just a diagnosis tool that can be used to see every config variable in use by lance/robin/pybot

[patent pivot]

the actual config needs to remain in the bots obviously, the parsing can be farmed out to the TOML and env modes of pydantic settings, the thing that we can try generalise is seeding a discord environment based on that config

I suspect we might be able to do some smart subclassing of pydantic field types to do that

something that appears transparently as an integer when fetched as config but can be read as a channel or role or something by botstrap

and then we just have dev_required=True on those things and botstrap only creaets those it needs

or even some sort of typing.Annotated thing

[fallen patrol]

Does it need to remain in the bots itself? Defaults are provided for every role and channel variable that match the pydis server. Bot and lance and Robin have their own minimal roles and channels that are defined in all three bots, with bot having the most extensive amount.

Pydis core could have that complete configuration in it

[patent pivot]

we aren't going to do that because then updating config means updating a dependency on the bots

[fallen patrol]

Ahhhh

Which may be undesired true true

[patent pivot]

it is already a pain sometimes when we make a change to core we want on all bots related to a constant or osomething and have to then go and bump everything

[fallen patrol]

I mean you could override it with a subclass but that becomes annoying too

[patent pivot]

then there's no real point in putting it in core

[fallen patrol]

I was just thinking for channels & roles, not for everything

Cause then you can move a lot of the redirect output helpers and stuff to pydis core

[patent pivot]

yea -- it's not a bad idea but my thoughts are just that we have enough bespoke stuff that we're going to end up with a situation where our config is even more fragmented than it currently is

having config pulled from two places, one of which is a dependency, is worse than just having a local version that's got duplicated values

[fallen patrol]

That's fair

[fallen patrol]

But this is worth considering I think

[patent pivot]

yea the question there is that we don't need/use those utilities on other bots

there are things that we have copied between but generally we've been quite good at moving stuff to pydis_core now

[fallen patrol]

Don't we use it on bot and lance?

[patent pivot]

no

not that I can see

[fallen patrol]

Right I'm confusing identical commands having different checks

There was a check I thought on more than one, hmm

[fallen patrol]

Ah well.

Theorized procedure:

• update botstrap on bot with existing PR
• draft updated constants to have dev_required fields and whatnot. Use init subclass and subclass from a custom base model and field class defined in bot_core which use a global registry to save all fields created: this allows botcore to create a configuration file from the constants defined in the consumer of botcore

Eg init_subclass checks all fields in the subclass to get the fields with dev_required=True so it can set those variables

We can probably move botstrap to pydis core and add a cli entrypoint

[patent pivot]

priority before that is probably adding some sort of toml support again to stop the env nesting madness

[fallen patrol]

yeah

bot#3417 cleans up the botstrap code but I won't be making other changes to botstrap until its merged

[dusky shore]

GitHub

[bot] #3417 refactor: update botstrap to be more beginner-friendly

[fallen patrol]

since rewriting to use toml requires updating each line that writes to config_str

[fallen patrol]

TIL Discord added clickable email syntax: [e-mail]

<[e-mail]>

Email and phone number markdown uses mailto: and tel: URI schemes respectively that can optionally be prefixed (e.g. mailto:[e-mail]). Email markdown supports headers, values must be URL Encoded (e.g. <[e-mail]?subject=Message%20Title&body=Message%20Content>).

🎉

[fallen patrol]

dear god

[fallen patrol]

(I appreciate it was f-strings before which is not correct, but if you're adding some f-strings and some %s strings we may as well unify).
@floral lion what does this mean? You'd rather have %s or f in logging?

[patent pivot]

wrong ping -- but basically if the file is already using f-strings then probably stick with that, else move everything to %s formatting

the proposed changes appears to have a mix

on main it's currently just f-strings

[fallen patrol]

thoughts on enabling https://docs.astral.sh/ruff/rules/#flake8-logging-format-g

[patent pivot]

potentially -- for now i'm more bothered about mixing multiple styles within one file

[fallen patrol]

Aight will make them all %

[patent pivot]

there are still f-string statements in that file if you were trying to update all of them

I am also not sure if you've addressed all the review feedback I've left yet so just ping me once you have done -- but the webhook one needs addressing as well as the general over complexity of that with statement (needs breaking into functions)

& there are still a couple of audit log reasons which are unnecessarily verbose (e.g. including the name on emoji create is not very useful, it's always going to appear alongside the emoji name)

[fallen patrol]

I haven't, I will rerequest when done

Been breaking it into a botstrapper class (bikeshed the name after I commit it smh)

[fallen patrol]

I don't think I saw any but I also may have not pushed that yet

Ah yep
https://github.com/python-discord/bot/blob/57c92b0cdb3807dff58a87f39b57132b2f8f1001/botstrap.py#L171
I haven't pushed that

[stable mountain]

botstrap.py line 171

log.info(f"Server {self.guild_id} has been successfully updated to a community.")```

[patent pivot]

no worries -- just saw some force pushing and wasn't sure

[fallen patrol]

@patent pivot that's change fr

big ol' "just change everything"

rather than functions just went with this cause it makes bot#3420 easier to manage in the long term

[dusky shore]

GitHub

[bot] #3420 Upgrade configuration to use a toml file over .env.server

[fallen patrol]

I will probably modify it again tbh

cause that's still noisy

[fallen patrol]

Lmao
#dev-log message
Oops

[vale ibex]

All good, I like rage baiting Joe too

[patent pivot]

https://tenor.com/view/erm-dog-nerd-nerd-emoji-glasses-gif-9712427073917878441

[safe oyster]

hi, there

I am looking for a developer who can implement to login with FB, apple account

[fallen patrol]

Still..

Just fixed it

-# in a parallel universe

[fallen patrol]

holdon

all of this testing I've done and that code is useless?

in a new environment and wow

get this with no config

these lines aren't even used

because pydantic parses the configuration immediately when bot.constants is imported

[patent pivot]

lol

that hacktoberfest "no maintainer" filtering doesn't seem to work very well

[patent pivot]

🤣

[fallen patrol]

welp guess I'm going the long route and filtering the pydantic validation exception

otherwise i need to move load_dotenv above the import from bot

[fallen patrol]

And I'm planning ahead for toml: if I have to change something related to dotenv handling then I'm making it not care about the loading method

....now I remember that when I was making user-facing code using pydantic I went with attrs and marshmallow

Data validation with useful errors for the end user is tedious

Yay fixing all validation errors with custom
https://docs.pydantic.dev/latest/errors/errors/ and some constants.py refactorings

@patent pivot I think implementing this to solve #dev-contrib message is probably large enough scope to be a second pull, as it will affect constants.py so I'm just gonna leave a comment documenting why we depend on load dotenv and so on so on

[patent pivot]

didn't we solve that by having the partial model thing

that is why there is a webhook change in your PR right?

[fallen patrol]

Solve the underlying error yes, the cryptic error messages no

[fallen patrol]

it also errors now if the guild id is set to this server lol

[lucid orbit]

If I want to fix a single regex pattern, should I open an issue for it or can I just immediately make a PR?

[patent pivot]

uhhh if it's a bug just open a PR

[fallen patrol]

where is this?

oh good god i broke discord again

just accidently bypassed the 10 webhooks per channel limit

[patent pivot]

spoo||est

🤣

[fallen patrol]

spoollest

[patent pivot]

spoo|y

[patent pivot]

pipe

[fallen patrol]

I think don't fix. Leave the easter egg

[lucid orbit]

spoo||aaaaa

[lucid orbit]

yk, I've been thinking of this as well

[patent pivot]

@lyric temple https://github.com/python-discord/sir-lancebot/commit/cacb8a287846cb3485114cfc176e66815d7e7bd9

named and shamed

[fallen patrol]

hmmm does it only happen once per halloween themed spooky message?

-# hype

[patent pivot]

lmfao great relevant issues matiiss

[lucid orbit]

thank you thank you

[patent pivot]

i think it's a funny bug but the chances of it being discovered as an easter egg are low and we should make more fun easter eggs

[fallen patrol]

doot! batwomen are a spooky danger to skeleton. Have you made a jack-o-lantern out of a pumpkin this halloween?

[patent pivot]

easter eggs that only work during october are THUMBS DOWN!

[fallen patrol]

we should also add plural to some of them

and maybe use emoji to set these

cause this is unreadable \U0001F3F3\uFE0F\u200D\U0001F308

when we already depend on emoji

!pypi emoji

[stable mountain]

emoji v2.15.0

Emoji for Python

Released on <t:1758456782:D>.

[patent pivot]

forgor we don't have ff-bot on lancebot

[fallen patrol]

ff?

oh also i updated the botstrap pr again

[lucid orbit]

tbf, the comment was pretty intimidating

[fallen patrol]

this is why i use the github cli for my issues and pulls

it lets me open ones without the template 🤡

[patent pivot]

we should just use character names

i was thinking this about the botstrap PR also tbf

i kind of hate unicode in source code

[fallen patrol]

I don't use any unicode do I?

ah right cause I refactored that out

[lucid orbit]

I'll make sure to reference that same screenshot and discussion for all my future regex PRs

[patent pivot]

!e ```py
print("\N{JACK-O-LANTERN}")

[stable mountain]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

🎃

[patent pivot]

lmfao

[fallen patrol]

huh

!e print('\U0001F3F3\uFE0F\u200D\U0001F308')

[stable mountain]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

🏳️‍🌈

[fallen patrol]

or just

:rainbow_flag:

[patent pivot]

!e ```py
print("\N{WAVING WHITE FLAG}\N{VARIATION SELECTOR-16}\N{ZERO WIDTH JOINER}\N{RAINBOW}")

[stable mountain]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

🏳️‍🌈

[patent pivot]

🤣

[patent pivot]

https://github.com/jb3/ff_bot

[fallen patrol]

yeah its a bit.... long

[patent pivot]

devops repos have it

[fallen patrol]

hype

[lucid orbit]

but it's readable

[patent pivot]

can do /merge and it'll do a fast-forward merge so your commit signing gets retained

[fallen patrol]

ah nice

[celest charm]

!e

print("\N{BANANA}\N{DUCK}")

[stable mountain]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

🍌🦆

[fallen patrol]

oh yeah that reminds me that python 3.14 and 3.15 don't have commit signing on their releases

[patent pivot]

long for that specific instance but generally I much prefer it to unicode code escapes and actual unicode symbols in source files

[fallen patrol]

agreed

[celest charm]

non-ASCII unicode symbols*

[fallen patrol]

we should be unifiying all of them anyhow

like in constants or config

[patent pivot]

https://tenor.com/view/slow-internet-slow-internet-cast-wizard-gif-15526244104305474494

[fallen patrol]

no you don't

not another rickroll

this is perfectly fine code

ignore the red, I do use it

[celest charm]

what does the str do?

[patent pivot]

there are still outstanding review comments

[fallen patrol]

yeah I never said i finished them just that I updated it :^)

[patent pivot]

lol

[fallen patrol]

also do you prefer the reviewer or the pr author to mark conversations as resolved when they've made the changes

[patent pivot]

ehh I don't mind really -- if it's been addressed exactly as suggseted then mark reviewed

[lucid orbit]

btw, I wanted to create an issue on the repo, but ig I might as well ask here first, is lancebot not supposed to include embeds in bookmarks?

or I guess specifically images

[celest charm]

This is my personality. An annoying nitpicking mosquito. Unfortunately I am not blessed with above average intelligence intelligence, so I'm not eligible to work in projects that desperately want this role, like nuclear reactors and foundational software.

[patent pivot]

lmfao

this botstrap is so dense

[celest charm]

You say something like "ord returns the ASCII value of the provided 1-length string" and you will hear the high-pitched buzzing sound even though I am 2000 km away

[patent pivot]

!duck 1432906604432326756

!duckify 1432906604432326756

thanks man

that one made me laugh

slowly but surely i will turn duck-pond into sean-in-devoops

[celest charm]

Nice try. However, this is a continuation message that requires the crucial context that I identify as a mosquito.

[patent pivot]

some greats in this channel

[fallen patrol]

what about it

[fallen patrol]

arguably lance can use message forwards for bookmarks now and lessen a lot of the complexity

[patent pivot]

nothing specifically just a lot of dense methods

[fallen patrol]

what about dense

[patent pivot]

but i've been writing fastapi apps for the last few months so maybe have got used to tiny methods

[fallen patrol]

ah complex

[patent pivot]

just things like get_channels etc have a lot going on

also i will forever leave a review comment where comments don't have a capital letter

comments should always be fully formed sentences

[celest charm]

if only we could get computer to point out formatting mistakes

[patent pivot]

https://tenor.com/view/1984-george-orwell-george-orwell-1984-1984-chips-1984-ruffles-gif-7144052804264359069

[lucid orbit]

that makes sense

[fallen patrol]

thanks, i appreciate this

i Really Do

[celest charm]

stupid question: why are we depending on httpx and aiohttp

[fallen patrol]

my guess is sync cocde

[patent pivot]

hmmm

[fallen patrol]

aiohttp is used by the discord api wrapper y'all use

[patent pivot]

yeah it's literally just botstrap

that uses httpx

[fallen patrol]

yeah

[patent pivot]

we should Do Something About That someday

[fallen patrol]

no i'm not refactoring it to use aiohttp this pr

[patent pivot]

yea i'll do it someday

[fallen patrol]

httpx is awesome and I actually use it in my bot now lol

!pypi githubkit

[stable mountain]

githubkit v0.13.5

GitHub SDK for Python

Released on <t:1761408762:D>.

[fallen patrol]

genuinely a great piece of software

[patent pivot]

yeah it's nice

[fallen patrol]

but I pass a custom transport in order to use aiohttp over httpcore

so I have httpx as my caller and then I just use my aiohttp session I use for other things too

its so useful

can't believe the httpx dev wants to split it into sync and async projects

[celest charm]

people in 1995: we gotta prefer composition over inheritance you know...
botstrap: we inherit from httpx.Client

[fallen patrol]

honestly....

sigh

[celest charm]

sorry, I'm just adding unnecessary comments

[fallen patrol]

joe i'm not rebasing and changing my history any more at this point

[fallen patrol]

no we shouldn't though

[patent pivot]

i think i'm going to squash it excluding the change to constants anyway so don't worry about it

[fallen patrol]

yeah fair

if you look at botstrap its actually somewhat close to toml support but not in this pull (because toml support requires lockstep with contributing guide which needs a pre-update before toml support)

dict !!

means its already ready for tomli_w.dump(s)

though possibly tomlkit for comments and supporting user provided config and preservation of format

[fallen patrol]

@celest charm @patent pivot can i yeet the special handling of forum channels? I haven't managed to make a server without forum channels enabled

[celest charm]

I think so, yes

[patent pivot]

sounds reasonable

[celest charm]

now it rolled out everywhere

[fallen patrol]

that cleans up a lot of the other special handling code

[patent pivot]

i suppose if you add that upgrade script in the future it might be worth handling the case where the expected channel types are different

however

i think that's a fair case where deleting the channel/telling the user to delete the channel and re-run works

[fallen patrol]

upgrade script in the future will be entirely reconfirmation on everything

but upgrade script is after additional CLI refactor -> migration to pydis_core

[patent pivot]

yea

[fallen patrol]

also will probably want to use click for pydis_core

given we only need the cli in dev it can be an extra

[patent pivot]

yea i think we need to probably sit down and specify how that all works a long time before any PRs start

specifically what ends up living where at the end of things

[fallen patrol]

tool.pydis_core in pyproject.toml moment

[lyric temple]

I see nothing but provided value.

[fallen patrol]

spoo||y

hmmmpf

[fallen patrol]

yay

thx joe

[outer oasis]

Only six for a shirt?

[fallen patrol]

yea

I mean I did a lil more than that

[toxic wren]

what is this in?

[fallen patrol]

Hacktoberfest!

[toxic wren]

https://hacktoberfest.com/

?

I thought pythondiscord discontinued this

[fallen patrol]

I asked Joe if I could have the tag lol

I just wanted the shirt (it's a very comfy shirt)

[patent pivot]

it's not a python discord project

we didn't explicitly tag issues with hacktoberfest

but merged and approved PRs count to hacktoberfest

[toxic wren]

so 6 PRS from a singular project or combined between a few

lol I have 3 PRs for pythondiscord

lemme see if I do another one for the / commands issue.

in like a few days

[patent pivot]

has to be in the next 2 days if you want it to count hahaha

[toxic wren]

yes 💀

imma try anyway

[toxic wren]

?

I prbably have PR's in other OSS projects

I have another PR in Kimi K2

so total of 4

[patent pivot]

any project on github

you can just login to hacktoberfest.com and it'll tell you

[toxic wren]

Oh the PR's should be between oct 1 to 31

the old PRS are in september

oh lol

@patent pivot Isn't the change as simple as

- @commands.group(name="docs", aliases=("doc", "d"), invoke_without_command=True)
+ @commands.hybrid_group(name="docs", aliases=("doc", "d"), invoke_without_command=True)

seems to not work. Is there something else that's causing it not to work?

[patent pivot]

no because we talked about how we should support autocomplete for symbols etc.

we are adding slash commands to make the experience better, not to have it be the same

i am unsure why it may or may not be working -- do you have an error?

[toxic wren]

nope, it just doesn't show up

[patent pivot]

hmm

that seems more likely a problem with syncing slash commands then

i can't remember how/if we do that reliably in dev

[toxic wren]

ideas?

[patent pivot]

not sure exactly -- we should sync commands to the guild that you are using for development https://github.com/python-discord/bot-core/blob/74072b3709d87dc811a4972aec1ef172101e49bf/pydis_core/_bot.py#L174-L178

[stable mountain]

pydis_core/_bot.py lines 174 to 178

async def _sync_app_commands(self) -> None:
    """Sync global & guild specific application commands after extensions are loaded."""
    await self._extension_loading_task
    await self.tree.sync()
    await self.tree.sync(guild=discord.Object(self.guild_id))```

[patent pivot]

make sure you are restarting your discord client between running the bot and once it finishes syncing because this is probably a local issue

[fallen patrol]

you have to add the label cause the topic isn't added to the repo
but when you add the label its automatically accepted even if not already merged

in other words you just need to open + get label in next day and a half

[toxic wren]

ah

btw I have 4 PR's open in a few hours somehow

I can open like 4 more tmr

and make a maintainer give me the tag

[toxic wren]

can you add it to a draft?

maybe I could make a [WIP] Docs slash command PR?
@patent pivot would you add the tag?

[patent pivot]

I mean don't open it if it's not ready and is just going to be commit spam getting it to a point where it's ready

[toxic wren]

I mean I don't have a working demo, Though I could commit [this](#dev-contrib message) ? what do you think?

[patent pivot]

uhh, maybe? if you've got nothing that works and that's not guaranteed to be the way to implement it I'm unsure, we might end up not being able to use hybrid commands here etc

for example !d is a group but we don't want it to be for the purposes of slash commands probably

[toxic wren]

Ill make aPR tmr, and we can go over it later. ok?

[patent pivot]

sure -- but don't open it for the sake of opening it, because your change above is not correct

[fallen patrol]

ok listen-- botstrap was ready I just added more features each day

smh

-# and then a few more
-# and then another refactor the next time I looked at it

I mean tbf I thought it was done until I reviewed my own code after joe's comments

[toxic wren]

@patent pivot bot#3428 Can you add the tag?

[dusky shore]

GitHub

[bot] #3428 [WIP]Maka the docs command a hybrid slash command

[patent pivot]

I'd rather see some sort of actual attempt at the solution? As mentioned above, we cannot use hybrid_group here and at the very least it'd be good to see something functional.

We stopped participating heavily in Hacktoberfest because of the number of PRs like this at a volume that was unmanageable.

I will have a think about it but I'd rather you take a lower hanging issue with a quicker solution and actually PR a fix than PRing a line that's not the solution to the issue at all.

As I said above, don't open this for the sake of opening it. This PR isn't even passing the tests.

& if it's a WIP then open it as a draft

at the very least before adding the label I'd be looking for something functional and that actually passes tests, not a one line change that's invalid

[toxic wren]

Understandable, Ill look for other issues.

[fallen patrol]

@toxic wren how good are you with regex?

bot#3415 is an easy-ish regex one that just needs to remove (?:C\d+)? from the provided urls, see the issue content

[dusky shore]

GitHub

[bot] #3415 code snippet additional character support

[fallen patrol]

bot#3425 likely requires internal deliberation around how to allow attachments more than 6.

[dusky shore]

GitHub

[bot] #3425 Attachment spam limits to 6 attachments but messages can contain 10 attachments

[wind ruin]

testing in another server, it seems like a single message with 10 attachments can take the entire window at once?
not sure if we want to allow that tbh

[toxic wren]

not that good, but I can try to take on bot#3415

[dusky shore]

GitHub

[bot] #3415 code snippet additional character support

[fervent sage]

Depends on the attachment type, images should always be coerced into a gallery view now (and the bot can tell whether they're images from the file type)

[wind ruin]

weird, might be bugged or something? (just sent a bunch of images from discord + a random game I had saved - that is all one message)

[fervent sage]

good lord thats cursed

[fervent sage]

lmfao

the wall of joe

[patent pivot]

hmmm

that is noisy

how does it handle uhhh

[fervent sage]

how big is just one of them? presumably the difference between the 3x3 grid and one isn't as big as the 3x3+1 above

[stable mountain]

Please react with ✅ to upload your file(s) to our paste bin, which is more accessible for some users.

[patent pivot]

sweet

[patent pivot]

the 3x3+1 is over double just one image

[fallen patrol]

either way there's no flaw in 10 in one IMO. At the very least, it probably shouldn't time a user out for 10 attachments the first go, but maybe the second?

like the @here filter

[patent pivot]

yeah we probably shouldn't time out

but i don't know that allowing all 10 attachments (image or text) is a good idea

[fallen patrol]

hm

[patent pivot]

so

the uv lock enforcement changes

(for good reason) now cause all dependabot PRs to fail lol

interesting

i take that back

apparently enabling beta ecosystems uses a version of dependabot uv support that does not support lockfiles

cc @vale ibex you may find that noteworthy for other dependabot repos we have using uv

[vale ibex]

Ah nice

at some point you had to enable it for uv support at all

and now it breaks with it, great stuff

[patent pivot]

excellent features from github as per usual

[toxic wren]

4

another got merged and maintainer forgot to add tag or did not read the msg

[patent pivot]

that's fine

they count if they get merged

iirc

oh wait no yeah it does need the label

interesting

seems a bit unclear from this

time for bot on 3.14

[fallen patrol]

Wait for 3.14.1--------

[fallen patrol]

It needs label if repo doesn't have hacktoberfest topic

[patent pivot]

there's not a huge reason to hold off is there

[fallen patrol]

Hold off for Patience?

[patent pivot]

i haven't seen any major bugs reported in 3.14.0 that seem to be affecting us (going off the local build)

yea but i want to do a wider bump of dependencies and uv version etc. which doesn't really need 3.14.1 to come along, and when it does it should be a one or two line change

we automatically adopt patch versions anyway

we pin to minor

[fallen patrol]

Dependabot pip support is different from uv ecosystem

[patent pivot]

so 3.14.1 will just happen when it is released

[patent pivot]

that's not what's happened here, having uv in the beta ecosystem supports less than the uv in the regular dependabot ecosystem

[fallen patrol]

Pip bumps project.dependencies

uv ecosystem bumps uv.lock versions

Ah

Because beta was never bumped lmao nice

[patent pivot]

yea

uv beta env never touched uv.lock

it wasn't supported originally

[fallen patrol]

Times I've broken CI: 4

-# Can I increase this number

[patent pivot]

btw @fallen patrol

Traceback (most recent call last):
  File "<frozen runpy>", line 198, in _run_module_as_main
  File "<frozen runpy>", line 88, in _run_code
  File "/Users/joe/Projects/python-discord/bot/botstrap.py", line 476, in <module>
    changes_made = botstrap.run()
  File "/Users/joe/Projects/python-discord/bot/botstrap.py", line 469, in run
    changes |= self.write_config_env(config)
               ~~~~~~~~~~~~~~~~~~~~~^^^^^^^^
  File "/Users/joe/Projects/python-discord/bot/botstrap.py", line 425, in write_config_env
    with self.env_file.open("r+") as file:
         ~~~~~~~~~~~~~~~~~~^^^^^^
  File "/Users/joe/.local/share/uv/python/cpython-3.13.7-macos-aarch64-none/lib/python3.13/pathlib/_local.py", line 537, in open
    return io.open(self, mode, buffering, encoding, errors, newline)
           ~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
FileNotFoundError: [Errno 2] No such file or directory: '.env.server'

testing your PR

obv creating .env.server as an empty file fixes it, but doesn't check for existence right now

[fallen patrol]

Whatttt it worked for me lol

Will fix 😔

I'd even deleted it lol

[patent pivot]

maybe a macos thing lol

[fallen patrol]

cheeese

httpx.HTTPStatusError: Client error '401 Unauthorized' for url 'https://discord.com/api/v10/applications/@me'

hm

i should really update raise_for_status

[fallen patrol]

pushed

[patent pivot]

noice

[fallen patrol]

remove this

just remove it. It uses project.requires-python if this isn't defined

or make the pr and let me comment that for the coauthorship credit 😄

[patent pivot]

hahaha it's removed in my local version

[fallen patrol]

then I'm saving the other comment for when the PR is made

have to figure out not caching in a property.... hm

[patent pivot]

why not just use a method?

[fallen patrol]

yeah it was easier than i thought

but functools.cache would add complexity because I need to manually update the cache

[patent pivot]

hmm

not necessarily

oh i guess if you want to not refetch next time

but you can just .cache_clear() or whatever

interestingly i found a genuine problem between 3.13 and 3.14

cool

[fallen patrol]

Fixed

[fallen patrol]

This is why the delay

Lol

[patent pivot]

what lol

which thing did you find

[patent pivot]

cool so pytest-xdist now hates our mocks

[celest charm]

Side question: when is a truly unbounded cache a good idea?

as in: @lru_cache(maxsize=None) instead of e.g. @lru_cache(maxsize=1_000_000)

[patent pivot]

cool so it's actually pytest-subtests

0.15.0 breaks with the DumpErrors

https://github.com/pytest-dev/pytest-subtests/issues/202

[timid sentinel]

Good to see the markdown tests doing their job, markdownify loves their breaking changes lol

[patent pivot]

🤣 yeah what a great change

such a simple thing to add backcompat for as well

[fallen patrol]

Or .cache which is inherently unbounded

[fallen patrol]

no i mean in some cases I need to set the cache object through a different method

[patent pivot]

yea i figured that after you said it -- when you do the patching etc. and refresh the object

[fallen patrol]

patching to app info returns the updated app info payload which should replace the cached object
and creating channels in botstrap upgrade could be added to the cached channels but I've opted to not do that on account of depending on the order of cached channels

why do you pin every dependency in pyproject.toml and not use the lock file as.... a lock file? @patent pivot

[patent pivot]

dependabot do that for me

but basically because we almost never want to automatically adopt any new versions, even patch versions, and so it's easier to have dependabot handle the upgrades and then look for stuff with uv tree --outdated --depth 1 every now and then (like now) for wider upgrades

[celest charm]

Yeah that was what I meant. What's the use case for an unbounded cache? Seems like a bad idea

[toxic wren]

yes they all have the hactober label, none of the maintainer put a "accepted" label, they just merged it

[patent pivot]

oh well that's fine then

[toxic wren]

lets gooo

[patent pivot]

if it's in a hacktoberfest labelled repo and merged then it's accepted

[toxic wren]

so I have to wait a week for an email for the t-shirt?

so they finish reviewing

[patent pivot]

something like that yea

assuming they haven't gone over 10,000 people that have claimed a shirt

[toxic wren]

I heard in the official dicord that they "don't check the PR's", they just wait for a week and if the repo is not disqualified then it is accepted

[toxic wren]

Im keeping my hopes up

whew that was a lot of work for 2 days

atleast I can enjoy Halloween

[fallen patrol]

THIS EXISTED?

[toxic wren]

yes?

[fallen patrol]

i got a code but i haven't redeemed it hm

[patent pivot]

yeah it's the only reason we pin versions like that lol

[fallen patrol]

maybe I should do that

[patent pivot]

yeah i claimed mine

thanks to snekbox for accepting my contributions 👍

[fallen patrol]

i-

lmao

[celest charm]

joebama awards joebama

.jpeg

thank you lancebot

[fallen patrol]

see I actually got about 50 from https://gh.arielle.codes/monty but i wanted to do it right and get my 6 from an org i'm not affliated with

[patent pivot]

yeah true

[celest charm]

we need to photoshop your face onto this

[patent pivot]

i decided hacktoberfest have done enough overall harm to OSS repos in the past that i didn't care to nab a t-shirt

[toxic wren]

oh lol 3 of mine are from my own repo, I made the PRS like on mid oct without realising it.

[fallen patrol]

that's fair

yeah hacktoberfest is interesting lol

the tshirt is all i care about really

[celest charm]

I thought it's opt-in for repositories now?

[fallen patrol]

but also i probably owed something to pydis now

[celest charm]

Or did they revert that?

[fallen patrol]

its opt-in

[toxic wren]

it still is opt-in

[fallen patrol]

also https://github.com/python-discord/bot/pull/3417#discussion_r2471020763

[toxic wren]

Nice repo! I might contrib next week or smth

[fallen patrol]

and https://github.com/python-discord/bot/pull/3417#discussion_r2471890448

[toxic wren]

When does hacktober end in EDT

Would OCR help?
#community-meta message

like very barebones

[patent pivot]

yeah but there are steps to take before we start running users images through OCR

[toxic wren]

a simple python OCR lib that checks for keywords in the image and if it matches a banned word it will alert Mods

[toxic wren]

like?

[patent pivot]

it subverts user expectations and for the most part I doubt it'd return a very good / consistent result

[patent pivot]

well, spammers almost always hit the same channels, same named images and even the same links, we need to become better at detecting it without looking at the image contents

[toxic wren]

I mean the bot is OSS and the server already has logging

[toxic wren]

ok

[patent pivot]

all looking at image contents does is add a silly amount of processing onto every image sent

it's not an inexpensive operation to run at scale

[toxic wren]

example channel: #pyweek-game-jam

[patent pivot]

5,200 images sent this month, 51,000 sent this year, which is fairly expensive to run on

[patent pivot]

yea

well

kind of

#pyweek-game-jam gets hit a lot but our filters are picking that up (hence why you are seeing the bot messages)

[toxic wren]

yeah, i get the concern

[toxic wren]

yes I saw a mod talking about this

kinda sad

maybe you should just delete the channel and make a new one so ID gets reset and all the spam just does not work, though the spammer probably has logs

[patent pivot]

meh i mean during actual pyweek it is fine

it's just because the channel sees very little activity outside of pyweek

i was thinking about reducing the time before the bot messages get auto-deleted

[toxic wren]

Will it happen?

[patent pivot]

i've not put it in mod-tools yet so hasn't been discussed

[toxic wren]

probably a 1 line change

[patent pivot]

yea i wrote the feature

[toxic wren]

i mean you wrote most of the bot

[patent pivot]

nah team effort

we crossed 10k commits recently on bot

that's cool

https://github.com/python-discord/bot/commit/f686da5c09c98f02ae8f802057359dd1d9f117d7

10,000th commit

[toxic wren]

First commit https://github.com/python-discord/bot/commit/68da80efc00d9932a209d5cccd8d344cec0f09ea

On Feb 3 2018

[patent pivot]

crazy

7 years

[toxic wren]

yeah

I joined github like a few months ago

I was still a kid then

[patent pivot]

i joined july 2016 seemingly

[fallen patrol]

i almost reviewed this lmao

[toxic wren]

February 26, 2025 was when I joined

[fallen patrol]

...and i bugged it

[timid sentinel]

damn, I joined GitHub when I was 11

I'm impressed that I didn't somehow forget my account/lose my login at some point

[toxic wren]

which was?

like year

Im totally not gonna predict your age

[patent pivot]

1992

i always remember wookie's birthday because he was born 10 days after the fall of the soviet union

[royal prawn]

cap
you got me for about 10 minutes though

[toxic wren]

if he was born in 1992 and he made his github when he was 11 he made his acc 2003 but github started in 2008

[timid sentinel]

yeah man, my due date was two weeks earlier but I figured I'd let things settle down a bit before being born

[patent pivot]

lmfao

[fervent sage]

I joined at 12 apparently lol

[toxic wren]

Someone just sold be on uv and I love it

[celest charm]

...that was in 1991, no?

[patent pivot]

december 26th 1991

10 days later

1992

[celest charm]

oh right

datetime hard

[patent pivot]

lmfao

[toxic wren]

huh

December 23

[royal prawn]

uh whoops
meant to forward that one

accidental staff channel leak

[toxic wren]

ah ths why I couldn't find it

good thing you forwarded it instead of linking it

[royal prawn]

yeah if there's no channel name and timestamp beneath the text it's from a channel you can't access

[timid sentinel]

:P

[toxic wren]

ah

[toxic wren]

https://www.holopin.io/hacktoberfest2025 yay

6k have the badge

do 4k left

I think I can make it

[fallen patrol]

@toxic wren if you're looking for an issue to implement
https://github.com/python-discord/sir-lancebot/issues/1703

[toxic wren]

thx

[toxic wren]

bot#3431 is a random issue, @patent pivot

bot#3431

[dusky shore]

GitHub

[bot] #3431 pocketoption-bot.

[fallen patrol]

they'll deal with it

#dev-contrib message

[patent pivot]

removed, but please don't comment on them yourself, all it does is generate more notification spam for people in their email inboxes, we get to them when we can and it's not the end of the day if one stays up whilst I'm asleep or something 👍

[fallen patrol]

Every time I'm bored I do the boring thing of finding an invite bypass on Python

Think I found another one

Well that was easy

[toxic wren]

wdym?

[toxic wren]

oh i get it now

[fallen patrol]

Why not use a select menu for roles rather than buttons?

It do be big and fat on mobile and yeah

I would show a select menu example but I can't since I deleted that code in favour of

I...... Don't think we need to do this for commands which don't have any arguments?

It's subscribe. There's nothing special about it

Makes sense for eval but if the ctx.args is falsey then uploading is useless IMO

[patent pivot]

we have this as well

which is why nothing got changed

[patent pivot]

seems reasonable, but I don't think it's that harmful

pastes delete after 30 days, so not fussed really

@toxic wren make sure you're not trying to claim a load of issues without actually working the ones you're currently assigned to (e.g. the /docs command)

[toxic wren]

I am facing hiccups in the implementation of the /docs. I am still working on it though, and sir-lancebot#1703 seems like a simple url change, or maybe a permanent image.

[dusky shore]

GitHub

[sir-lancebot] #1703 .monster show goofy doesn't show an image

[fallen patrol]

It doesn't have love fest, revival of code, or advent of code though

Stuck message?
#voice-verification message

[patent pivot]

yea bot restart probably

[patent pivot]

yeah, but you have commented on bot#3405, bot#3422 and bot#3425 as well asking for those to be assigned

[dusky shore]

GitHub

[bot] #3405 Configured channels do not exist in server: [('rule_alerts', 1386490058739290182)].
[bot] #3422 Provide context when dming delete link from paste uploader
[bot] #3425 Attachment spam limits to 6 attachments but messages can contain 10 attachments

[toxic wren]

bot#3422 and bot#3425 are pretty simple, considering bot#3425 gets approved

[dusky shore]

GitHub

[bot] #3422 Provide context when dming delete link from paste uploader
[bot] #3425 Attachment spam limits to 6 attachments but messages can contain 10 attachments

[patent pivot]

right, I guess I'm just trying to figure out your priorities, I will assign all of them to you though

I'm changing the scope of 3425 though

[toxic wren]

You should probably not assign bot#3405 though, I should probably work on that after /docs

[dusky shore]

GitHub

[bot] #3405 Configured channels do not exist in server: [('rule_alerts', 1386490058739290182)].

[fallen patrol]

I think joes concern is mainly don't start a bunch of PRs until you have some experience with the project, aka have a handful of merged pulls where you understand the style, the project, and how everything works together

Eg https://github.com/python-discord/bot/commits?author=DragonSenseiGuy

Vs https://github.com/python-discord/bot/commits?author=jb3

[patent pivot]

assigned to both of the former

[toxic wren]

ok thanks Ill keep what @fallen patrol said in mind.

[fallen patrol]

Though to be clear I speak as myself and my views don't reflect pydis or joe... etcetera

[patent pivot]

i don't really mean anything at all around style and how things work together, if there's things that aren't understood there it gets solved in the PR review stage and things don't merge until reviewers are happy. I just mean as a matter of time management and trying to help guide people it's easier if they want to work on and commit to 1 thing at a time compared to several.

[fallen patrol]

Hhhhhuh?

[toxic wren]

@fallen patrol to clarify in bot#3422 you want to change the message that the bot sends to DM's to include where I uploaded the file?

[dusky shore]

GitHub

[bot] #3422 Provide context when dming delete link from paste uploader

[patent pivot]

https://github.com/python-discord/bot/blob/2640aaa5665426e5350b3ce81a743a9f4c3898aa/bot/exts/utils/attachment_pastebin_uploader.py#L131 this line needs changing just to include message.jump_url

[stable mountain]

bot/exts/utils/attachment_pastebin_uploader.py line 131

await message.author.send(content=f"[Click here](<{paste_response.removal}>) to delete your recent paste.")```

[toxic wren]

can I get where the user uploaded the file?

like the channel

[patent pivot]

yes, using jump_url

that gives you the message link

like [#dev-contrib message](/guild/267624335836053506/channel/635950537262759947/)

"[Click here](<{paste_response.removal}>) to delete the pasted code from [your message](<{message.jump_url}>)."

[toxic wren]

yes currently testing the changes

why does my local bot say:
Hey @toxic wren!
It looks like you tried to attach a Python file - please use a code-pasting service such as https://paste.pythondiscord.com/

while @stable mountain says:

Click here to delete your recent paste.

is it because this bot has access to pydis pastebin backend?

[patent pivot]

no

your local bot is showing the anti-spam response

https://github.com/python-discord/bot/blob/8fdc3708b28667ce40240dc7221eb1d2194ac6d4/bot/exts/filtering/_filter_lists/extension.py#L18-L21

[stable mountain]

bot/exts/filtering/_filter_lists/extension.py lines 18 to 21

PY_EMBED_DESCRIPTION = (
    "It looks like you tried to attach a Python file - "
    f"please use a code-pasting service such as {PASTE_URL}"
)```

[patent pivot]

i don't know why that'd necessarily trip differently based on local dev vs prod

[toxic wren]

I tried in a help post and other channels, still get same response

[patent pivot]

interesting

ahh

!filter list extension

[stable mountain]

List of allowed extensions (21 total)

652. .json
2743. .py
2742. .txt
2744. .js
2764. .sql
2769. .log

Page 2/2

[fallen patrol]

It does. Yeah.

[patent pivot]

@toxic wren add py to your allowed files

i forget extensions is a dynamic filter

[fallen patrol]

Ahhhhh okay

[toxic wren]

ah

[fallen patrol]

Wonder about a dev cog that doesn't load on prod

[toxic wren]

btw bot#3432 ready for review

[dusky shore]

GitHub

[bot] #3432 Provide context when dming delete link from paste uploader

[patent pivot]

for what purpose

[patent pivot]

failed linting

[toxic wren]

spoke too soon 😅

[fallen patrol]

true... that can be fixed with a conributing guide snippet

[patent pivot]

yea

it does feel we rely on this behaviour a bit

so should probably just have it as a filter list seeded default

https://github.com/python-discord/site/blob/5f4cf8f5e79530099b4f53af6c52f4225704ebc8/pydis_site/apps/api/migrations/0059_populate_filterlists.py#L94-L122

just needs a new migration probably

if we want it to always be there then we add a new migration to add if not exists

[turbid pilot]

Hello guys

Is anyone online now ?

[toxic wren]

yeah, why?

[turbid pilot]

I'm new here

[toxic wren]

do you want to contribute to pythondiscord?

[turbid pilot]

And I am looking to make new friends

[fallen patrol]

hm

[toxic wren]

um...

!off-topic

[stable mountain]

Off-topic channels

There are three off-topic channels:

• #ot0-psvm’s-eternal-disapproval
• #ot1-perplexing-regexing
• #ot2-never-nester’s-nightmare

The channel names change every night at midnight UTC and are often fun meta references to jokes or conversations that happened on the server.

See our off-topic etiquette page for more guidance on how the channels should be used.

[turbid pilot]

Yeah but it depends on the kinds contributions

[patent pivot]

@toxic wren the way you've solved that leaves a newline in the message sent to users

and still fails lint

[toxic wren]

All contibutions are welcome

[patent pivot]

you can run all the CI tests locally before you commit using uv run task lint btw

[fallen patrol]

I see no problem with 10 attachments

https://private-user-images.githubusercontent.com/71233171/508606027-6b2863e0-7d68-4b9f-8a07-fe2e710591fc.png?jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.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.Jp961aSLe3Fm3x7BYxP2YrJFDOpzrP2UDvsXEHSSpl4

its smaller than 3 github embeds

[toxic wren]

bot#3432 ready for review

[dusky shore]

GitHub

[bot] #3432 Provide context when dming delete link from paste uploader

[patent pivot]

f-strings don't work multiline like that

!e ```py
something = "hello!"

print(
f"line 1: {something} "
"line 2: {something} "
)

[stable mountain]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

line 1: hello! line 2: {something} 

[patent pivot]

that's what you've done

you have to have an f-string on every line where you're using interpolated values

!e ```py
something = "hello!"

print(
f"line 1: {something} "
f"line 2: {something} "
)

[stable mountain]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

line 1: hello! line 2: hello! 

[toxic wren]

fixed

[toxic wren]

bot#3432 ready for review

[dusky shore]

GitHub

[bot] #3432 Provide context when dming delete link from paste uploader

[fallen patrol]

looked to review bot#3434 since it didn't seem to fit existing styles and now.... I have thoughts about python.... https://github.com/python-discord/bot/blob/8fdc3708b28667ce40240dc7221eb1d2194ac6d4/bot/bot.py#L65-L70

Returning within a context manager that suppresses the exception meaning this statement of ```py
with suppress(Exception):
maybe_raise(Exception)
return

will sometimes return there and sometimes will continue past and I just.... python what is this

that above statement is functionally equivalent to the following:
```py
try:
  maybe_raise(Exception)
except Exception:
  pass
else:
  return

[dusky shore]

GitHub

[bot] #3434 Handle closed DMs in paste uploader

[celest charm]

Yep, context managers can swallow exceptions

[patent pivot]

what do you mean by "didn't seem to fit existing styles"

we use except discord.Forbidden: or except Forbidden: in 6 places on bot and 3 of them are what this is, just a logging statement and nothing else

❯ rg 'except (discord\.)?Forbidden' --after-context=3
bot/decorators.py
185:                except discord.Forbidden:
186-                    log.info(
187-                        "Failed to DM %s with redirected command paste removal link, user has bot DMs disabled",
188-                        ctx.author.name

bot/exts/moderation/defcon.py
124:                except Forbidden:
125-                    log.debug(f"Cannot send DEFCON rejection DM to {member}: DMs disabled")
126-                except Exception:
127-                    # Broadly catch exceptions because DM isn't critical, but it's imperative to kick them.

bot/exts/backend/error_handler.py
136:                except Forbidden:
137-                    await self.handle_unexpected_error(ctx, e.original)
138-            else:
139-                await self.handle_unexpected_error(ctx, e.original)

bot/exts/moderation/infraction/_scheduler.py
125:        except discord.Forbidden:
126-            log.info(f"Bot lacks permissions to delete message {message_id} in channel {channel_id}.")
127-        except discord.HTTPException as e:
128-            if e.code == ARCHIVED_THREAD_ERROR:
--
519:        except discord.Forbidden:
520-            log.warning(f"Failed to deactivate infraction #{id_} ({type_}): bot lacks permissions.")
521-            log_text["Failure"] = "The bot lacks permissions to do this (role hierarchy?)"
522-            log_content = mod_role.mention

bot/exts/filtering/_settings_types/actions/infraction_and_notification.py
183:            except Forbidden:
184-                ctx.action_descriptions.append("failed to notify")
185-
186-    async def action(self, ctx: FilterContext) -> None:

[fallen patrol]

yeah. I couldn't remember the from import vs import vs contextlib vs try/except styles

There's a mix-match of each style

[patent pivot]

yea

me and chris were just playing around with copilot seeing whether it can do anything decent with issues on pydis repos

just given it a trickier issue

[fallen patrol]

for fun, give it .monster needs to be migrated to redis

[patent pivot]

wait why does it need to be migrated to redis

[fallen patrol]

cause it doesn't save anything lol

.src monster leaderboard

[dusky shore]

Command: monster leaderboard

Shows the current standings.

Source Code

Go to GitHub

bot/exts/holidays/halloween/monstersurvey.py:171

[patent pivot]

oh

lol

[fallen patrol]

https://github.com/python-discord/sir-lancebot/blob/cb14fcf7e96defe829ac022c6568269e0e3042e7/bot/exts/holidays/halloween/monstersurvey.py#L31-L34

[patent pivot]

funny

[stable mountain]

bot/exts/holidays/halloween/monstersurvey.py lines 31 to 34

def json_write(self) -> None:
    """Write voting results to a local JSON file."""
    log.info("Saved Monster Survey Results")
    self.registry_path.write_text(json.dumps(self.voter_registry, indent=2))```

[patent pivot]

yea

[fallen patrol]

this gets nuked every deploy

[patent pivot]

no one changed it when it got copied over from salt i guess

[fallen patrol]

salt?

salt bot?

[patent pivot]

no

saltstack

[fallen patrol]

hm

meta#202

[dusky shore]

GitHub

[meta] #202 Keyword Alerts

[fallen patrol]

organisation#202

🤔

mmmm validation

[patent pivot]

what has organisation#202 got to do with things

[fallen patrol]

https://github.com/python-discord/sir-lancebot/pull/329#issue-533518294

[patent pivot]

oh

yeah

that's not meta

[fallen patrol]

was git blaming

it has been renamed a few times.

a few.

[fallen patrol]

private repo?

[patent pivot]

yeah

the all-staff internal meta repo

[fallen patrol]

interesting

brb finding security bug

actually i need to check if monty is still vulnerable after i rewrote all of the github handling

[patent pivot]

lol

finding organisation issues is not the end of the world

[fallen patrol]

yeah

[patent pivot]

finding the other repos internally is more problematic

[fallen patrol]

no exactly

having knowledge of The Organisation means a point that can be tested without finding confidential information

The Organisation

wait a moment

git are you drunk

how did a json file become a python file

[patent pivot]

the bot doesn't pull issues with a token that can access organisation though anyway iirc

[fallen patrol]

ye

if it did it would have replied already

maybe.