dev-contrib

[clever wraith]

And don't hesitate to ask any questions in the issue/here

[gritty wind]

If you see an unassigned issue with a linked PR, please let us know. We either forgot and should fix it, someone skipped a step, or it is up for grabs

[trim cradle]

(bot-py3.11) λ poetry run pre-commit
check for merge conflicts................................................Passed
check toml...........................................(no files to check)Skipped
check yaml...........................................(no files to check)Skipped
fix end of files.........................................................Passed
trim trailing whitespace.................................................Passed
isort (python)...........................................................Passed
ruff.....................................................................Failed
- hook id: ruff
- exit code: 2 

error: TOML parse error at line 89, column 15
   |
89 |     "RUF005", "RUF012",
   |               ^^^^^^^^
Unknown rule selector: `RUF012`

is there something I need to update? idk what I'm even looking at here.

[clever wraith]

What ruff version do you have installed ?

Did you run poetry install ?

If your depencies file is new, but the ones in your environment don't match, you might run into that (that's my guess)

It could be that the RUF012 rule was added in a new version

[trim cradle]

great, that fixed it.
I swear, every time I contribute to the bot, so much has changed since last time that I have to relearn everything.

[clever wraith]

Ahahahha

All good ! 😄

[cursive relic]

Is this supposed to be like this? Or is it possibly supposed to be postgres?

[atomic ivy]

that's a typo

[cursive relic]

Does there happen to be an existing issue for it, I don't think I saw one

[atomic ivy]

typos don't need issue approval, I think 😅

(someone correct me, if I'm wrong)

[static canyon]

For something as small as this certainly not

[static canyon]

Feel free to just open a PR on site (I'm assuming that's where this is)

[cursive relic]

I don't think I'm able to make one

[static canyon]

You can fork the repository, make the changes on the fork, and then PR to the original

[cursive relic]

Ah, right. I don't need to pull the repo to local

[static canyon]

For such a small change you can use the GitHub site editor, yeah

[cursive relic]

Just need to pinpoint where the typo is, I'll take a look and try to see where it is and if I have problems finding it I'll ask here

[static canyon]

If you press . (you need to be signed in, and maybe on the home page) then it brings up the web ide which lets you search for text (so you can search for the typo)

[cursive relic]

Thx for the tip! I have never seen this before, wow. Why did I not know of this before

[static canyon]

It's a relatively new feature (few months iirc)

[cursive relic]

ahh, that explains it. But glad I discovered this, it will make working with some of my projects much more easier

what about for commit messages? How much for that level of detail should I explain what i did? Would something like this work? ```md
Typo fix.

Fix typo in bot.md file. Change the typo from psotgres to postgres.

Maybe this?

[vale ibex]

The first sentence of your extended description should be the commit message instead

"Update bot.md" isn't a useful commit message

github web defaults it to that since it can't generate a useful one for you

[cursive relic]

oh, right.. Would this be better instead: Fix a typo in bot.md leaving the rest as is

[clever wraith]

Or using conventional commits Like
fix: postgres typo in bot.md

[vale ibex]

Yea, that's fine

[cursive relic]

Done, thank you everyone!

[tall laurel]

hey

can anyone help me with something please?

like I have a simple project that I need someone to code with me

really simple don't worry

[sharp crag]

hi - this channel is for questions/discussions about our projects. There are a bunch of issues on the various repos. If you want to work on one of them, you can leave a comment and someone will assign it to you

[left nebula]

@stable mountain devs
the bot got problems with converting hyperlinks in the docs command (sometimes)

!d setattr

[stable mountain]

setattr

setattr(object, name, value)```
This is the counterpart of [`getattr()`](https://docs.python.org/3/library/functions.html#getattr "getattr"). The arguments are an object, a string, and an arbitrary value. The string may name an existing attribute or a new attribute. The function assigns the value to the attribute, provided the object allows it. For example, `setattr(x, 'foobar', 123)` is equivalent to `x.foobar = 123`.

*name* need not be a Python identifier as defined in [Identifiers and keywords](https://docs.python.org/3/reference/lexical_analysis.html#identifiers) unless the object chooses to enforce that, for example in a custom [`__getattribute__()`](https://docs.python.org/3/reference/datamodel.html#object.__getattribute__ "object.__getattribute__") or via [`__slots__`](https://docs.python.org/3/reference/datamodel.html#object.__slots__ "object.__slots__"). An attribute whose name is not an identifier will not be accessible using the dot notation, but is accessible through [`getattr()`](https://docs.python.org/3/library/functions.html#getattr "getattr") etc..

[left nebula]

for example

[vale ibex]

This is probably related to the changes Discord made recently to markdown

We get docs content directly from the source, in this case python.org docs

[left nebula]

I couldn't imagine why it would work sometimes then

[vale ibex]

They added "security" feature to markdown links to ensure text that looks like a link doesn't link to a different url

but I think it's quite rudimental how it flags that, so it's causing false positives when the text has underscores it seems

or maybe it's not recognising the url side of the markdown link as a valid url for some reason

[left nebula]

so weird looking link (dunder links) wont get hyperlinked then?

that would make sense somewhere

discord is weird

[vale ibex]

Yea, it's likely a benefit overall, IE you can't hyper link something like https://google.com

but the way they've implemented it causes some false positives

[left nebula]

yeah true

I was struggling with it myself and wanted to take a look at the python bot here so I noticed that

wouldn't know how to fix it while I dont know the flags that cause that

[vale ibex]

the only way to fix it would be to:

1. intercept the docs text and alter it to work
2. Wait for Discord to implement a fix

1 is a lot of work, since we'd need to correct all the places where it happens, and also make sure new docs/changes to docs are also fixed

and number 2 may or may not ever happen

so imo we just leave it as it is, and maybe Discord will fix it at some point

[left nebula]

yeah hopefully

I think you could change something in the markdown converter to remove/change some of these hyperlinks as long discord wont change something

check if href contains for example underscores (regular expressions or the link)

[cold island]

And it's not even obvious how you would alter it since changing characters can change the meaning of the text

[vale ibex]

Yea

[brisk brook]

What is this weird http://o.hi? https://github.com/python-discord/bot/blob/82d0ded4bc634c9fc81a010b2fb10bdacf57b823/bot/exts/moderation/modlog.py#L653

[stable mountain]

bot/exts/moderation/modlog.py line 653

content_before.append(f"[{sub}](http://o.hi)")```

[brisk brook]

It looks to be just for formatting, right?

[vocal prairie]

it's just to make things blue, yeah

it's not a real website

[clever wraith]

https://discord.com/channels/267624335836053506/1138495553278333039

closed after warn

[vale ibex]

yea, if the original message gets deleted and no other non-bot messages are sent, help channels close after 5 minutes.

[clever wraith]

[tawdry vapor]

@gritty wind For this PR did you figure out how we would install the additional packages (numpy, scipy, etc)? https://github.com/python-discord/snekbox/pull/175 I think we would need to install them separately for each version, right?

[gritty wind]

Arl suggested installing them separately to ensure we can have different versions for each and such

Someone had a look over the requirements.txt specification and we can specify the version right in there, keep it all as one file

All the python installs are complete, fully-functional installs so pip should work

[outer oasis]

Wait really? How? Like--

numpy==62;python-version==3.11
numpy==61;python-version==3.10

[gritty wind]

Seems like a good use-case for try-it-and-see

I believe that should work however

Might need to make that py ver a string

Or maybe not

Use quotes around specifiers in the shell when using >, <, or when using environment markers.
Do not use quotes in requirement files. There is only one exception: pip v7.0 and v7.0.1 (from May 2015) required quotes around specifiers containing environment markers in requirement files.

[clever wraith]

may i ask when you introduce the 3.12 playground to the bot. I remember there was something like that for 3.11 just dont remember at what phase

[clever wraith]

I don't think we'll be doing that this time.
There has been an internal discussion about this, and 3.11 was added because we had some big changes that we wanted people to test in an early stage. It's not the same case for 3.12.

But, when mark gets to that pull request we're trying to close on snekbox, it will become easier to add so we might eventually do it.

[clever wraith]

ah okay sad news then

hope it will eventually come

It's not definitive, of course.
It's just that we didn't come to an agreement of doing it 🙂

yeah i just shouldnt assume you will do it

Exactly

[thorny obsidian]

@timid sentinel thanks for the quick review! Incorporated your comments with the latest commit

[timid sentinel]

No problem, just noticed one more thing, then all good as far as I'm concerned

[outer oasis]

https://github.com/python-discord/bot/blob/main/bot/exts/info/information.py#L588-L590
Why doesn't this delete itself?

[stable mountain]

bot/exts/info/information.py lines 588 to 590

if invalid:
    await ctx.send(shorten(":x: Invalid rule indices: " + invalid, 75, placeholder=" ..."))
    return None```

[clever wraith]

I'd say no particular reason

[outer oasis]

What are peoples thoughts on adding the timed deletion here?

Mostly for consistency

[clever wraith]

An issue would be great.
It's better to point out that we don't have consistent behavior all across our bot responses.
So we can identify these places as I don't think all responses should be deleted.

[timid sentinel]

I'm not really a fan of timed deletion in general tbh, either chat has gone on so there's no real benefit in deletion, or it's still visible and therefore confusing as it looks like the bot didn't work. Not sure what others think though.

[clever wraith]

I don't see the benefit of deleting the bot embeds or whatever in the first place. It's not like we pay for message storage, and it's not spammy or clutters the space. Help embeds could be a good appliance, but still

[clever wraith]

I'm of the same opinion

[cold island]

If we want consistency I'd drop the timed deletion in other places as well lol, not a fan of them either

[clever wraith]

You could as well do it like in docs command

When its improper symbol then you delete embed with command

[gritty wind]

Tbh in a real-time chat platform, deleting messages “for space” or whatever else doesn’t make much sense

A delete button is better than automatic. No deleting is best

[thorny obsidian]

So I'm hitting a weird error with altering a docker compose.
I'm altering the dockerfile of snekbox and rebuilding. It builds and runs fine on its own. When I change the reference to the snekbox image in the forms-backend docker-compose.yml, suddenly I get the following error trying to start up snekbox. All I did was change the image name to the new one I built, so I'm not sure what could be going wrong

EDIT: resolved. I needed to add a build: . to my forms docker-compose for snekbox

[vale ibex]

@clever wraith I'm going to change colourst oa normal class

I just saw the validator and ignored it, but now I see it was just there to convert

[clever wraith]

Yeah, I added it myself

[vale ibex]

https://github.com/python-discord/bot/pull/2695/commits/238c9b49da979f14b46a017cad766221b4e203dc is the diff, I'm going to squash a few commits and merge unless you tell me this diff is horrible

[clever wraith]

But we can eventually just short-circuit the conversion

You forgot to inherit BaseModel

[vale ibex]

it doesn't need to

[clever wraith]

(Do we even need it ?)

yeah

Ah no

No no, never mind

[vale ibex]

Same as https://github.com/python-discord/bot/pull/2695/commits/3a279b3d3faa679684a5b1bb5abda925abf16c38

[clever wraith]

I misunderstood, sorry

But now no conversion happens, does discordpy take hex values ?

Nvm

This is not my day

[vale ibex]

hah no worries

alright, commits be cleaned

https://github.com/python-discord/bot/pull/2695/commits

[clever wraith]

Sweet!

[plush birch]

I'm trying to add a new alias for the rule about paid work, where should I create the issue? And is this the correct file to edit? (line 172) https://github.com/python-discord/site/blob/main/pydis_site/apps/api/views.py

[still comet]

where should I create the issue?
Seems like someone else already made an issue bot#2716

is this the correct file to edit?
That should be correct AFAIK

[dusky shore]

GitHub

[bot] #2716 Make !pay bring up Rule #9

[plush birch]

It was very complicated but I think I pulled it off ;)

- ["paid", "work", "money"]
+ ["pay", "paid", "work", "money"]

[timid sentinel]

lol, my test @stable mountain bot contains "discord" in the username and "discord" in the application name. Discord seem to have decided that isn't allowed anymore and I get errors trying to start the bot.

Going into the developer portal, first try changing the application name:

Validation errors:
name: Username cannot contain "discord"

Alright, so i'll change the username first... :

Application name cannot contain "discord"

Now i'm stuck, great...

can probably fix it by making the request manually I guess...

hmm no I think they're different endpoints 😐

[still comet]

I am trying to get my @stable mountain fork running with redis, but it can't seem to be able to connect to the redis server (redis.exceptions.ConnectionError: Error 111 connecting to redis:6379.).
Redis seems to be running fine, except there are no logs whatsoever and running ls in the terminal just hangs without any output.

Also, I'm not sure if it is relevant, but running docker compose up for the first time outputs 'postgres is unhealthy' even though docker says it's healthy. The issue disappears when running docker compose up again.

output (first run and bot error on subsequent runs): https://paste.pythondiscord.com/VC3Q

[vale ibex]

can you do docker compose pull and then docker compose up --build

[still comet]

I'm still getting the same error

[clever wraith]

For both redis and postgres?

The postgres unhealthy should mean that it doesn't recognize that user

If you bash into the container and try to connect with those credentials, does it work?

[clever wraith]

Also, just making sure

1. Did you make any changes to the compose file?
2. Do you have a .env file that might be overriding the redis credentials?

And if you don't want to block yourself for now, you can use fake redis

[still comet]

I haven't touched the compose file, and .env only has BOT_TOKEN and GUILD_ID

[clever wraith]

can you do docker compose down, then the 2 commands Chris gave you?

Which os are you on?

And in the create redis connection function , can you check if the host, user and pwd are correct?

Wait, the redis service in the compose file doesn't specify any credentials, so in unsure at what stage that's setup

I can have a look later

[clever wraith]

(which means that you might not even need any)

[still comet]

Win10

[still comet]

I'll try that again I guess

[still comet]

Nope, still getting both postgres and redis errors

The bot starts with fakeredis, but throws this when I try to load reminders:

❌ Failed to load extension bot.exts.utils.reminders:

TypeError: module() takes at most 2 arguments (3 given)```

[clever wraith]

Do you have the latest changes ?

Did you sync your fork, pull latest main changes, and checkout from that ?

[still comet]

I synced it yesterday

[clever wraith]

Is there a full stack trace of this ?

[still comet]

Here's the full bot log: https://paste.pythondiscord.com/E6IA

[clever wraith]

Can you tell me what the output of git log --oneline -n 1 --format=%H is ?

[still comet]

wait I think that might be something to do with my code

[still comet]

436b6dd12aedca818e3cf3f5a33e9ff8b8b492f6

[clever wraith]

Did you make any changes ? Or do you get that on a clean branch ?

[still comet]

Yeah I made some changes

I stashed them and it works now 🙂

[clever wraith]

Great

[still comet]

Thanks for the help!

[clever wraith]

Sure

Let's try solving the other one now

Is the postgres container running ?

[still comet]

Yes

https://paste.pythondiscord.com/YGLA Here's the log when docker compose up is run the first time.
Everything after 16:16:28 happened after the command exited

[clever wraith]

Everything after 16:16:28 happened after the command exited
Not sure I get this, are the logs saying that the service isn't healthy, but then it fixes itself ?

[still comet]

Yeah

[clever wraith]

Then, there's no problem, right ?

What does docker logs {postgres-container-id} give ?

You can get the id by running docker container ls (or from docker desktop)

[still comet]

https://paste.pythondiscord.com/V6EA

[clever wraith]

Does the bot run & connect to the api ?

and persist data ?

I think it might throw off that not ready error as the service is still spawning.
And that by then, pg isn't ready to accept any connections just yet.

But i'm only speculating here

I don't see anything strange in the logs you gave me

[still comet]

I gtg now, hopefully we can fix this next time.

[clever wraith]

in reference to bot#2722 the pep 0 embed already applies this functionality

[dusky shore]

GitHub

[bot] #2722 Move PEP link to title in pep command

[clever wraith]

!pep 8

[stable mountain]

**PEP 8 - Style Guide for Python Code**

Link

Status

Active

Created

05-Jul-2001

Type

Process

[clever wraith]

!pep 0

[stable mountain]

**PEP 0 - Index of Python Enhancement Proposals (PEPs)**

Status

Active

Created

13-Jul-2000

Type

Informational

[clever wraith]

[static canyon]

If you look at the source code that's because PEP0 has it's own function for building the embed (it's hardcoded, rather than generated from the API)

[stable mountain]

bot/exts/info/pep.py lines 63 to 74

def get_pep_zero_embed() -> Embed:
    """Get information embed about PEP 0."""
    pep_embed = Embed(
        title="**PEP 0 - Index of Python Enhancement Proposals (PEPs)**",
        url="https://peps.python.org/"
    )
    pep_embed.set_thumbnail(url=ICON_URL)
    pep_embed.add_field(name="Status", value="Active")
    pep_embed.add_field(name="Created", value="13-Jul-2000")
    pep_embed.add_field(name="Type", value="Informational")

    return pep_embed```

[clever wraith]

yeah i know i was just scrolling through its source and found that

just another reason to implement it in normal peps

[clever wraith]

is is still possible to work on bot#2692?
is so how should i approach sending other message via user being banned? do i create another CONSTANT in _channel.py ?

[dusky shore]

GitHub

[bot] #2692 Automatically close help thread on ban

[static canyon]

That sounds about right, and then you'd have a on_member_ban event listener in _cog.py that calls a new close_channel_from_ban function or something like that which will take the banned user, get their channel, send a custom message, and close it

[clever wraith]

Ah wasnt quite sure if i should create a new function for that

[static canyon]

Could probably even edit _close_help_post to take a custom message argument, which if not provided will just send the current message

[clever wraith]

Also should i create a ClosingReason for USER_BANNED or something Like that

In _stats

There is the enum

[static canyon]

That seems smart, yes

[clever wraith]

Okay sure ill claim the issue then

[static canyon]

I'd probably go for OP_BANNED just to be explicit about which user got banned

Or CLAIMANT_BANNED actually to use consistent terminology

[clever wraith]

Will do

[last trout]

Hi everyone!
I am Harsh Kumar, a python enthusiast. I want to start open source contribution to my favourite language of all Python. Can someone share any resource for the same?

[clever wraith]

@full fractal Can you try rebasing onto main ?

@last patio Looking at the history, there's one commit from MArch 18 that's messing things up

If you could, locally, do a hard reset to the commit before the merge one

git reset --hard {commit-hash-before-merge}

Then a git rebase main

You might run into conflicts, if that's the case, you'll need to resolve them then run git rebase --continue

[clever wraith]

Hey !

We have a couple of projects you could contribute to

!contribute

[stable mountain]

Contribute to Python Discord's open source projects

Looking to contribute to Open Source Projects for the first time? Want to add a feature or fix a bug on the bots on this server? We have on-going projects that people can contribute to, even if you've never contributed to open source before!

Projects to Contribute to
• Sir Lancebot - our fun, beginner-friendly bot
• Python - our utility & moderation bot
• Site - resources, guides, and more

Where to start

1. Read our contribution guide
2. Chat with us in #dev-contrib if you're ready to jump in or have any questions
3. Open an issue or ask to be assigned to an issue to work on

[clever wraith]

We've documented everything there

Have a look, and you can ask us anything you'd like related to our projects/contributing in this channel

[clever wraith]

@still comet I know what's wrong with the redis connection

When you're inside docker, you can't resolve 127.0.0.1

Inside docker, the host that will allow you to communicate with redis is the service name, in this case redis

So, if you set redis_host=redis in your .env file, and run again, it should do it

It's weird that it's not mentioned in sir-lance's docs

I'll make sure to add it later

[still comet]

Oh okay, I'll try that tomorrow

[clever wraith]

You'll need to remove the use fake redis env var

[tawdry vapor]

I think this makes sense to do. Packages may not be compatible across multiple versions, especially c extensions and such.

Would like to get some feedback on my current plan:

Use pyenv to install all interpreter versions we want into the image. There are ways to debloat the installs to get them down to ~70 MB per version (I looked at how the official Python image does this). Even if our base image is Python 3.11, still install a separate Python 3.11 interpreter, for example.

This lets us decouple snekbox's interpreter from those NsJail uses. It also means we can consistently use a normal pip install for each version without requiring a PYTHONUSERBASE workaround for one of them.

[clever wraith]

Using pyenv sounds reasonable.

The 70 MB was for how many versions of python?

We could support the ones that are still officially supported, and making the addition / decommission of versions easy.

[clever wraith]

Why would we want to decouple the one snekbox uses from NSJail's?

I thought we'd want to pass the version through snekbox, then relay to NSJail

[tawdry vapor]

There's no benefit to sharing them other than to save on build times and image sizes. Having them use separate interpreters lets us update them on separate cadences. It also offers more proper isolation - currently there are some quirks with exposing pip setuptools and wheel.

It's 70 MB per interpreter.

1 version = 70MB, 2 versions = 140 MB, etc

[clever wraith]

Sounds good to me

[tawdry vapor]

Ugh I updated the base image to Debian bookworm and now tests are failing - something to do with multiprocessing. But I cannot figure out why

https://paste.pythondiscord.com/ZINQ

[vale ibex]

Could you try setting multiprocessing start method to fork?

from multiprocessing import set_start_method
set_start_method('fork')

Iirc this is going to be the default in a few versions

[tawdry vapor]

Cool, that fixed it. Thanks!

Any context on why we have to do this with Debian 12?

[vale ibex]

Nice

Honestly it's been so many months since I looked into it I can't remember

But whenever something weird happens with mp this is quite often the fix

Its the fix to a number of windows issues too

[tawdry vapor]

I spoke too early - the test_multiprocessing_shared_memory_disabled is still broken

Oh I just misread the test results. Actually everything is still broken

[vale ibex]

Huh alright, back to the drawing board then lol

It's getting late here, i can look into it more tomorrow though

[tawdry vapor]

Thanks. I think I will just drop this change and worry about updating debian some other time.

[vale ibex]

Sounds good

Ooi is it on debian 11 now?

[tawdry vapor]

12

Oh currently it is on 11. I am trying to update to 12

[vale ibex]

Ah yea alright, wasn't sure if you were going alpine->bookworm or plain debian bump

[tawdry vapor]

here is the dockerfile if anyone wants to mess with it https://paste.pythondiscord.com/XZYA

[tawdry vapor]

I've put out a draft here of my approach to installing multiple Python versions in the snekbox image. Would appreciate feedback before I proceed with API changes https://github.com/python-discord/snekbox/pull/181

[clever wraith]

yo is any of those considered applying
bot#2646
bot#2679
meta#217

[dusky shore]

GitHub

[bot] #2646 Use new markdown features in messages
[bot] #2679 Code snippets for paste service links

GitHub

[meta] #217 Add .jsondb tag

[gritty wind]

Approved both bot issues, feel free to comment if you'd like to be assigned

You should bump the meta issue in #community-meta as the more relevant audience

[clever wraith]

@timid sentinel you mind helping me why does it still show the commits from the old PR? i already pulled them so idk

[timid sentinel]

I think the issue is cause by you working on your main branch, ideally you should create a new branch for each PR you want to do. I'm not a git expert so since the new changes are fairly small I would probably just copy the new markdown changes to a file somewhere, reset your main branch with the upstream main branch, create a new branch, and recreate the markdown feature commit on that branch.

[clever wraith]

managed to create a pr and reset the commits to the upstream

Like wookie said, the main branch is made so that we merge into it, we never commit directly.
Looking at your previous PR (attached image), I can see that you're working on the main branch
You should always make a new branch, make your changes there, and then create a pull request from that branch.

If you want to fix what you have now, you can do this

git reset --soft HEAD~2
git checkout -b use-new-markdown-in-help-channel-embed
git commit -am "{{your commit message here}}"

[clever wraith]

yeah thanks now i know that even on forks i still need branches
i solved it other way arround i fetched the upstream and then i did git reset --hard upstream/main and force pushed to fork repo so the commits are discarded

Yeah, but a hard reset makes you lose your changes

it wasnt big of a deal this time

i guess it can be merged right away

[vale ibex]

git is great, there's liek 4 ways to do things. this one keeps the 2 seperate commits, rather than squashing into one ```
git checkout -b use-new-markdown-in-help-channel-embed
git cherry-pick main~2
git checkout main
git reset --hard origin/main

[clever wraith]

not big of a change

[clever wraith]

cherry-pick i never knew how to use that

[vale ibex]

basically it's a way to apply a commit to the current branch

so you can give it a git sha, or a commit range and those commits will be added to your current branch

main~2 means the top 2 commits from the branch called main

[clever wraith]

ah nice

what does that do

[clever wraith]

Yeah, cherry picking is great.
I just never knew i could specify the commits to Cheryl pick like that, awesome

[vale ibex]

you'll likely not have access to it even if you do authorize

it's the thing that adds the Review requirements check to the actions section

clicking details forces it to refresh, but only staff can do that

[outer oasis]

Oh it does?

[vale ibex]

Yea

[outer oasis]

I thought it just showed you

[vale ibex]

nah page load triggers a recalc

[outer oasis]

huh

TIL

[clever wraith]

mhm

bot#2679

thinking about how to approach it
is there any way you share some API to fetch such information or it would need to be scraped?

[dusky shore]

GitHub

[bot] #2679 Code snippets for paste service links

[clever wraith]

scrolling through pinnwand docs and only found how to upload via api

[sharp crag]

You just need to send a GET to the raw paste

For instance, https://paste.pythondiscord.com/raw/YT6A

Note the /raw/

[clever wraith]

ah okay i see

C:\Users\PC>curl https://paste.pythondiscord.com/raw/M45Q
print("Hello World")
``` seems to be working

[clever wraith]

i also asked about that on pinnwand repo and thats what i got in response supakeen/pinnwand#193

[dusky shore]

GitHub

[pinnwand] #193 Is there a way to retrieve via API paste code from link?

[clever wraith]

i guess ill wait until they add that to api?

[sharp crag]

Looks like the author already has a PR for that so yeah

[clever wraith]

https://bpa.st/json/show/LOSA this way i also get the lexer

[clever wraith]

i feel like yall hate me for finding such small things
talking bout bot#2728 now

[dusky shore]

GitHub

[bot] #2728 Syntax error in !dotenv command

[clever wraith]

funny how it was there for 2 years now

[rapid swallow]

banned

!dotenv

[clever wraith]

[static canyon]

The example is somewhat wrong too

It should really just be load_dotenv() and not load_dotenv(".env"). It'll automatically look for the env file

You only need to manually specify the path when it's an odd file name or in a different dir

[clever wraith]

Yeah

[zealous spoke]

Should we add support to !eval for running stuff with Python 3.12, now that it's a release candidate, like we did for 3.11 last year? https://github.com/python-discord/bot/pull/2218

[vale ibex]

Yea, the intention is there, albeit via native mutli-version support in snekbox https://github.com/python-discord/snekbox/pull/181

[clever wraith]

You need to decide if you are doing it or not

[vale ibex]

read the last sentence :D

[clever wraith]

Glad to hear that's it might be available

[vale ibex]

found the problem for 3 of the 4 issues

/dev/shm doesn't exist in the container, which mp uses for it's SemLock

diff --git a/docker-compose.yml b/docker-compose.yml
index 0613abc..20f7c60 100644
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -26,6 +26,8 @@ services:
     volumes:
       - .:/snekbox
       - user-base:/snekbox/user_base
+      - snekbox-shm:/dev/shm

 volumes:
   user-base:
+  snekbox-shm:

This is the only remaining error https://paste.pythondiscord.com/637Q

fwiw this is my full diff (including dockerfile changes) on top of your multi-verison PR https://paste.pythondiscord.com/ESSA

The last error seems to be because it can't create the dir /test on the container for whatever reason

or it's not creating the dir by default anymore? 🤷

is it possible that Function not implemented just isn't the expected failure response when running under bookworm?

[tawdry vapor]

Thanks for investigating. Presumably shm is used even in Buster, so I wonder why it's no longer available. Were we relying on the base image to mount it for us?

[tawdry vapor]

This is probably the case. There may have been some kernel update or some config change that changed the behaviour of shm_open. Previously we were effectively asserting that it errored out with ENOSYS, but now it fails with ENOENT.

[tawdry vapor]

I don't see anything in the kernel suggesting a change, nor has its man page changed the documented error codes it returns

Oh it might be a glibc thing that changed

[tawdry vapor]

https://github.com/python-discord/snekbox/issues/182

[vale ibex]

ah nice, this does make sense now

[vale ibex]

Yea, I think we were, I'm also not sure if shm has specific properties or permisisons that mounting it int he way I suggested would mean it doesn't work. Might need further investigation

[clever wraith]

@vale ibex talking about bot#2726 i see almost all tags use the • symbol for listing and you also specify that in parser i guess thats deliberate action

[dusky shore]

GitHub

[bot] #2726 feat: use new markdown feature in new help channel message

[stable mountain]

bot/exts/info/doc/_parsing.py line 226

markdown_converter=DocMarkdownConverter(bullets="•", page_url=url),```

[clever wraith]

oh i guess im being stupid and this parser is not being used for tags

[vale ibex]

it's used for the docs command

[clever wraith]

i see

so, should i replace all • in tags to -

let me show you difference

before - after

[vale ibex]

yea seems fine, I think there'a few command doc strings that use that bullet too

!help mute

[stable mountain]

Command Help

!timeout <user> [duration=None] [reason=None]
Can also use: mute, tempmute

Timeout a user for the given reason and duration.

A unit of time should be appended to the duration. Units (∗case-sensitive):
 y - years
 m - months∗
 w - weeks
 d - days
 h - hours
 M - minutes∗
 s - seconds

Alternatively, an ISO 8601 timestamp can be provided for the duration.

If no duration is given, a one-hour duration is used by default.

[vale ibex]

!src mute

[stable mountain]

Command: timeout

Timeout a user for the given reason and duration.

Source Code

Go to GitHub

bot/exts/moderation/infraction/infractions.py:200

[clever wraith]

the !patreon supporters does use it

#bot-commands message

[vale ibex]

this uses an indent character for the effect, but a md list might be better, would need ot see what it looks liek

[vale ibex]

yea that can change

[clever wraith]

you mean this part ```md
 y - years
 m - months∗
 w - weeks
 d - days
 h - hours
 M - minutes∗
 s - seconds

[vale ibex]

Yea

[clever wraith]

one second

well didnt go as planned

ill see what i can do about it later on

now let me modify all tags

[tawdry vapor]

I think something else is going on, but I don't know what. I tried docker run with debian:bookworm-slim and python:3.11-slim and both had /dev/shm mounted.

[clever wraith]

bot#2730
my mistake wrong commit name its fix(logs) should be fix(tags) want me to close this pr and create new one with corrected syntax

[dusky shore]

GitHub

[bot] #2730 fix(tags): resolve SyntaxError in !dotenv command

[clever wraith]

It seems there are 2 open PRs to solve this issue

?

bot#2729

[dusky shore]

GitHub

[bot] #2729 Fix Dotenv Tag's Syntax

[clever wraith]

i see

but he didnt solve everything what was supposed to be

Let me handle that

[clever wraith]

?

https://github.com/python-discord/bot/pull/2730/files#diff-a05e714d3c2153c8d9f928e5617c0a11f27568a4d09a3b1a3c37cf9cac838f5dR17

I asked for feedback here and we came to the conclusion that the ".env" is the default value so no need to pass it

Yes, I wasn't talking about that

I didn't see any of the PRs

The issue here is that the other person didn't get assigned to the issue

ah okay

And neither were you, at least not explicitly since wookie told you to feel free to open a PR

well i was told to open a pr about it

i guess thats close to being assigned

Read my sentence 🙂

That's what I'm saying

[static canyon]

git commit --amend -m "new commit message"

git push --force

You can redo the commit message ^

[clever wraith]

oh let me try that

[clever wraith]

thanks! worked

[static canyon]

Not quite sure what's going on with the checks/actions though.. CC @clever wraith
https://github.com/python-discord/bot/actions/runs/5961212292/job/16170023845?pr=2730

Looks like the command is wrong (doesn't handle forks properly or something)

[clever wraith]

That's strange, I tested it on a fork from my github alt account

I'll look into it later

[vale ibex]

Ah yea, you're right. It's because we have ipc: none in the docker compose which is a private ipc ns without /dev/shm mounted

It worked before since it tried to use /tmp if /dev/shm wasn't avialable

but in glibc change you mentioned it no longer does that

So I think the "fix" it to change what return code we test for, rather than mounting /dev/shm

[clever wraith]

I didn't have to wait long for it to come in handy

forgot i already have a branch for markdown updates and had to move some so used this git cherry-pick update-markdown~2..update-markdown

[tawdry vapor]

Got it, will try to adjust the tests

[vale ibex]

Very nice

[vale ibex]

Nice sounds good, I'll keep an eye out for a PR

[clever wraith]

@vale ibex so pretty much changed everything i saw could be using new markdown

there is only one thing that got me https://github.com/python-discord/bot/blob/82301f2447e2f1df87433cd1d8bc5cffd42931a4/bot/exts/moderation/clean.py#L480-L489

i dont really see a way to use - so discord can format it

[stable mountain]

bot/exts/moderation/clean.py lines 480 to 489

\u2003• `​users`​: A series of user mentions, ID's, or names.
\u2003• `​first_limit`​ and `​second_limit`​: A message, a duration delta, or an ISO datetime.
At least one limit is required. The limits are *exclusive*.
If a message is provided, cleaning will happen in that channel, and channels cannot be provided.
If only one of them is provided, acts as `​clean until`​. If both are provided, acts as `​clean between`​.
\u2003• `​regex`​: A regex pattern the message must contain to be deleted.
The pattern must be provided enclosed in backticks.
If the pattern contains spaces, it still needs to be enclosed in double quotes on top of that.
\u2003• `​bots_only`​: Whether to delete only bots. If specified, users cannot be specified.
\u2003• `​channels`​: A series of channels to delete in, or an asterisk to delete from all public channels.```

[clever wraith]

when i replace all \u2003• with dash i get this and i dont really see why its not on new line

[vale ibex]

might need to put a whitespace unicode character at the start, since we trim docstrings in the custom help command iirc

[static canyon]

I guess replace \u2003• with \u2003-

[clever wraith]

Nah tried that too, then it doesnt replace it with markdown

[static canyon]

Hmm, probably because it has to start with a -

It might not be possible then

[clever wraith]

yeah but when it does start with - it results with: #dev-contrib message

[static canyon]

What happens if you put an explicit \n in the docstring on the line above?

"""
foo\n
- bar\n
- baz
"""```or something

I imagine it won't make a difference, but worth trying

[vale ibex]

https://github.com/python-discord/bot/blob/main/bot/exts/info/help.py#L311-L313

[stable mountain]

bot/exts/info/help.py lines 311 to 313

# Remove line breaks from docstrings, if not used to separate paragraphs.
# Allow overriding this behaviour via putting \u2003 at the start of a line.
formatted_doc = re.sub("(?<!\n)\n(?![\n\u2003])", " ", command.help) if command.help else None```

[vale ibex]

this is the logic we use to format the help command

[clever wraith]

seem like this is working fine

should i leave it?

[cold island]

!help clean

[stable mountain]

Command Help

!clean [users=None]... [first_limit=None] [second_limit=None] [regex=None] [bots_only=False] [channels=None]
Can also use: clear

Commands for cleaning messages in channels.

If arguments are provided, will act as a master command from which all subcommands can be derived.

• users: A series of user mentions, ID's, or names.
 • first_limit and second_limit: A message, a duration delta, or an ISO datetime. At least one limit is required. The limits are exclusive. If a message is provided, cleaning will happen in that channel, and channels cannot be provided. If only one of them is provided, acts as clean until. If both are provided, acts as clean between.
 • regex: A regex pattern the message must contain to be deleted. The pattern must be provided enclosed in backticks. If the pattern contains spaces, it still needs to be enclosed in double quotes on top of that.
 • bots_only: Whether to delete only bots. If specified, users cannot be specified.
 • channels: A series of channels to delete in, or an asterisk to delete from all public channels.

Subcommands:

!clean between <first_limit> <second_limit> [channel=None]
Delete all messages within range.
!clean bots <message_or_time> [channels=None]
Delete all messages posted by a bot, stop cleaning after reaching message_or_time.
!clean regex <regex> <message_or_time> [channels=None]
Delete all messages that match a certain regex, stop cleaning after reaching message_or_time.
!clean stop
If there is an ongoing cleaning process, attempt to immediately cancel it.
!clean until <until> [channel=None]
Delete all messages until a certain limit.
!clean users [users]... <message_or_time> [channels=None]
Delete messages posted by the provided users, stop cleaning after reaching message_or_time.

[cold island]

Seems like there are empty lines between the bullet points here

[cold island]

This might need some changing for it to work properly

[clever wraith]

Well you see the help docstring parser does it Like that

When i dont put new lines its like that #dev-contrib message

[static canyon]

Tbh I prefer that anyway

Makes it easier to read

[cold island]

yes, because of the regex above

!embed - ​users​: A series of user mentions, ID's, or names.

• ​first_limit​ and ​second_limit​: A message, a duration delta, or an ISO datetime. At least one limit is required. The limits are exclusive. If a message is provided, cleaning will happen in that channel, and channels cannot be provided. If only one of them is provided, acts as ​clean until​. If both are provided, acts as ​clean between​.
• ​regex​: A regex pattern the message must contain to be deleted. The pattern must be provided enclosed in backticks. If the pattern contains spaces, it still needs to be enclosed in double quotes on top of that.
• ​bots_only​: Whether to delete only bots. If specified, users cannot be specified.
• ​channels​: A series of channels to delete in, or an asterisk to delete from all public channels.

[stable mountain]

• ​users​: A series of user mentions, ID's, or names.
• ​first_limit​ and ​second_limit​: A message, a duration delta, or an ISO datetime. At least one limit is required. The limits are exclusive. If a message is provided, cleaning will happen in that channel, and channels cannot be provided. If only one of them is provided, acts as ​clean until​. If both are provided, acts as ​clean between​.
• ​regex​: A regex pattern the message must contain to be deleted. The pattern must be provided enclosed in backticks. If the pattern contains spaces, it still needs to be enclosed in double quotes on top of that.
• ​bots_only​: Whether to delete only bots. If specified, users cannot be specified.
• ​channels​: A series of channels to delete in, or an asterisk to delete from all public channels.

[cold island]

There's a bit of spacing with the native markdown

[static canyon]

Ah yes, nice

[static canyon]

We should probably go for this output then

[cold island]

@clever wraith basically this needs another lookahead that the line doesn't start with a dash

[clever wraith]

changing it to formatted_doc = re.sub("(?<!\n)\n(?![\n\u2003])", "\n", command.help) if command.help else None seems to solve the issue but i guess it might break something for others

import re

def f():
    """
    Commands for cleaning messages in channels.

    If arguments are provided, will act as a master command from which all subcommands can be derived.

    - `users`: A series of user mentions, ID's, or names.
    - `first_limit` and `second_limit`: A message, a duration delta, or an ISO datetime.
    At least one limit is required. The limits are *exclusive*.
    If a message is provided, cleaning will happen in that channel, and channels cannot be provided.
    If only one of them is provided, acts as `clean until`. If both are provided, acts as `clean between`.
    - `regex`: A regex pattern the message must contain to be deleted.
    The pattern must be provided enclosed in backticks.
    If the pattern contains spaces, it still needs to be enclosed in double quotes on top of that.
    - `bots_only`: Whether to delete only bots. If specified, users cannot be specified.
    - `channels`: A series of channels to delete in, or an asterisk to delete from all public channels.
    """
    ...
    
    
print(re.sub("(?<!\n)\n(?![\n\u2003])", "\n", f.__doc__))
``` results with:


```text
    Commands for cleaning messages in channels.

    If arguments are provided, will act as a master command from which all subcommands can be derived.

    - `users`: A series of user mentions, ID's, or names.
    - `first_limit` and `second_limit`: A message, a duration delta, or an ISO datetime.
    At least one limit is required. The limits are *exclusive*.
    If a message is provided, cleaning will happen in that channel, and channels cannot be provided.
    If only one of them is provided, acts as `clean until`. If both are provided, acts as `clean between`.
    - `regex`: A regex pattern the message must contain to be deleted.
    The pattern must be provided enclosed in backticks.
    If the pattern contains spaces, it still needs to be enclosed in double quotes on top of that.
    - `bots_only`: Whether to delete only bots. If specified, users cannot be specified.
    - `channels`: A series of channels to delete in, or an asterisk to delete from all public channels.

but not sure if this regex is pointless now

[cold island]

This replaces a newline with a newline

[clever wraith]

yep seeing it now

it changes nothing

[vale ibex]

\u2003 is an emspace, we likely don't want to use that anymore. We likely want to look at another way to have docstrings that are multi-line in code but then not in-help

probably something for another unicode char, maybe a non-breaking space or zero-width space

[clever wraith]

!e print("\u00A0")

[stable mountain]

@clever wraith :warning: Your 3.11 eval job has completed with return code 0.

[clever wraith]

thats no break space

!charinfo

[stable mountain]

Missing required argument

characters

Command Help

!charinfo <characters>
Shows you information on up to 50 unicode characters.

[clever wraith]

well it doesnt even see that

[vale ibex]

some whitespace gets stripped by Discord, so you might need to use a different zero-width space char

[clever wraith]

!charinfo x y

[stable mountain]

Character Info

\u0078 : LATIN SMALL LETTER X - x
\u0020 : SPACE -
\u0079 : LATIN SMALL LETTER Y - y

Full Raw Text

\u0078\u0020\u0079

[vale ibex]

will need some experimentation

[clever wraith]

hmm just a space

i know for a fact that this one ᅟ is not being stripped

!charinfo ᅟ

[stable mountain]

Character Info

\u115f : HANGUL CHOSEONG FILLER - ᅟ

[vale ibex]

this will likely be better served with another issue & PR

since we want to discuss implementation

which you can work on if you want, or not if you didn't want to

[cold island]

hmm does it really matter? it's not a character we use anywhere except here

Like, what prompted the idea we should change it?

[vale ibex]

I was guessing it gets output in discord, which messes with the md

we could use the same character, we'd just need to make sure it gets replaced with a new line

[cold island]

ah, you think there's a character that'll work with the bullet points?

[vale ibex]

Could be anything really, we'd just need to replace it with a newline in the help formatter

so I imagine we could even get it working with the current emspace, would just need regex changing I think

This is all guessing btw, haven't actually messed with it

all we need is a way to indicate in-code that a newline in doc string = newline in help docs

but by default strip newlines

[clever wraith]

cant you create a regex that adds new line when on the next line there is a - (im regex noob)

[cold island]

I guess we can scratch our heads on it more properly, but the most immediate solution seems to be to add a (?!-) at the end of the pattern here

[clever wraith]

okay then bot#2726 is already marked as ready

[dusky shore]

GitHub

[bot] #2726 feat: use new markdown feature in new help channel message

[static canyon]

@clever wraith for bot#2730 we should probably also update the "further reading" link to https://tutorial.vco.sh/tips/tokens/ (the old domain is now for sale)

[dusky shore]

GitHub

[bot] #2730 fix(tags): resolve SyntaxError in !dotenv command

[static canyon]

And since we're making a change anyway, there's inconsistent casing of python vs Python and programs in your programs environment is posessive so should be program's

[clever wraith]

https://web.archive.org/web/20210507171627/https://vcokltfre.dev/tips/tokens

[static canyon]

It's actually just at https://tutorial.vco.sh/tips/tokens/ now. I.e. vcokltfre purchased a different domain

[clever wraith]

ah okay i see

yeah sure i can change those, you wish to co-author and suggest a change?

[static canyon]

I can't sign in to my GH account atm, so if you're able to do it I'll just leave to you :)

Feel free to mention that it was suggested by me if you wish

[clever wraith]

sure thing will do, so all blame goes on you haha

[static canyon]

Looks good, thanks @clever wraith :)

Now you just need to wait for a core-dev to come around and merge it

[clever wraith]

yeah

dont know if i should re-request review though

oh nvm

I just merged it

already merged

[static canyon]

It's so small that I wouldn't've worried about it anyway tbh

!dotenv

[stable mountain]

Using .env files in Python

.env (dotenv) files are a type of file commonly used for storing application secrets and variables, for example API tokens and URLs, although they may also be used for storing other configurable values. While they are commonly used for storing secrets, at a high level their purpose is to load environment variables into a program.

Dotenv files are especially suited for storing secrets as they are a key-value store in a file, which can be easily loaded in most programming languages and ignored by version control systems like Git with a single entry in a .gitignore file.

In python you can use dotenv files with the python-dotenv module from PyPI, which can be installed with pip install python-dotenv. To use dotenv files you'll first need a file called .env, with content such as the following:

TOKEN=a00418c85bff087b49f23923efe40aa5

Next, in your main Python file, you need to load the environment variables from the dotenv file you just created:

from dotenv import load_dotenv()

load_dotenv(".env")

The variables from the file have now been loaded into your programs environment, and you can access them using os.getenv() anywhere in your program, like this:

from os import getenv

my_token = getenv("TOKEN")

For further reading about tokens and secrets, please read this explanation.

[clever wraith]

not here yet

[static canyon]

Ah, need to wait for bot to restart

Wasn't sure whether that was the case for tags or not

[clever wraith]

it is deploying just now

[static canyon]

!dotenv

[stable mountain]

Using .env files in Python

.env (dotenv) files are a type of file commonly used for storing application secrets and variables, for example API tokens and URLs, although they may also be used for storing other configurable values. While they are commonly used for storing secrets, at a high level their purpose is to load environment variables into a program.

Dotenv files are especially suited for storing secrets as they are a key-value store in a file, which can be easily loaded in most programming languages and ignored by version control systems like Git with a single entry in a .gitignore file.

In python you can use dotenv files with the python-dotenv module from PyPI, which can be installed with pip install python-dotenv. To use dotenv files you'll first need a file called .env, with content such as the following:

TOKEN=a00418c85bff087b49f23923efe40aa5

Next, in your main Python file, you need to load the environment variables from the dotenv file you just created:

from dotenv import load_dotenv()

load_dotenv(".env")

The variables from the file have now been loaded into your programs environment, and you can access them using os.getenv() anywhere in your program, like this:

from os import getenv

my_token = getenv("TOKEN")

For further reading about tokens and secrets, please read this explanation.

Using .env files in Python

.env (dotenv) files are a type of file commonly used for storing application secrets and variables, for example API tokens and URLs, although they may also be used for storing other configurable values. While they are commonly used for storing secrets, at a high level their purpose is to load environment variables into a program.

Dotenv files are especially suited for storing secrets as they are a key-value store in a file, which can be easily loaded in most programming languages and ignored by version control systems like Git with a single entry in a .gitignore file.

In Python you can use dotenv files with the python-dotenv module from PyPI, which can be installed with pip install python-dotenv. To use dotenv files you'll first need a file called .env, with content such as the following:

TOKEN=a00418c85bff087b49f23923efe40aa5

Next, in your main Python file, you need to load the environment variables from the dotenv file you just created:

from dotenv import load_dotenv

load_dotenv()

The variables from the file have now been loaded into your program's environment, and you can access them using os.getenv() anywhere in your program, like this:

from os import getenv

my_token = getenv("TOKEN")

For further reading about tokens and secrets, please read this explanation.

[clever wraith]

!dotenv

[static canyon]

There we go :)

[clever wraith]

whop

[static canyon]

@clever wraith By chance do you have time to give the required 2nd review for bot#2520?

[dusky shore]

GitHub

[bot] #2520 Handle off-topic channel names that are invalid for servers in Server Discovery

[static canyon]

It's been waiting for quite some time 😅

[clever wraith]

Sorry, I have to leave now.
I'll try to take a look when I'm back ! 😄

[static canyon]

No worries, would appreciate if you could take a look later :)

[clever wraith]

@still comet I made a small oopsie, I tried to merge your PR, so when I rebased, it appears as if I co-authred

Can you force push your branch ? That'll clear it, and we'll merge it later

[still comet]

Sure

@clever wraith ping

[clever wraith]

i was wondering why it is like that on mine #2730

[still comet]

How do I do that again?

[clever wraith]

git push --force i would say

[still comet]

So I can't do it on mobile?

I'll do it when I get home then

[clever wraith]

Yep

Sorry again

[still comet]

(bot-py3.11) danielgu@1 python-bot % git checkout confirm-reminder-edit
Switched to branch 'confirm-reminder-edit'
Your branch is up to date with 'origin/confirm-reminder-edit'.
(bot-py3.11) danielgu@1 python-bot % git push --force                  
Everything up-to-date
``` Am I doing something wrong?

[static canyon]

What does glo output?

*git log --oneline

[still comet]

(bot-py3.11) danielgu@1 python-bot % git log --oneline
25ab875f (HEAD -> confirm-reminder-edit, origin/confirm-reminder-edit) Prevent ghost pings when reminder modification is cancelled
bd7274f4 Ensure view gets deleted if view times out
75202828 Specify which action is being performed in confirmation message
c1829572 Apply changes from code review
34d6a4d3 Mention reminder owner in confirmation message
2a05c7bb Apply name changes from code review everywhere
ab930e68 Reply to original message when sending confirmation prompt
da48f9a1 Apply suggestions from code review
8ad236c0 Apply suggestions from code review
015b9956 Update docstring to better match function behavior
b9eae655 Ask for confirmation when an admin modifies another user's reminders
9d424c09 Create ModifyConfirmationView
c18abf8a Merge pull request #2730 from DownDev/fix-2728
9d2e5b20 feat(tags): update link and syntax in dotenv tag

[static canyon]

Which PR is this again?

[still comet]

bot#2719

[dusky shore]

GitHub

[bot] #2719 Make admins confirm edits of other users' reminders

[static canyon]

Hmm, sec

[static canyon]

Are you on your forked repo?

[still comet]

yeah

Should I be on the actual repo?

[static canyon]

I forget what half of my aliases are set to, gimme a sec

[static canyon]

You made the changes on your fork, so no, you want to be on your fork

[still comet]

Okay

[static canyon]

If you git checkout main && git pull does it say "up to date" or pull commits?

[still comet]

Already up to date.

[static canyon]

Actually

I think I'm overcomplicating this

git checkout confirm-reminder-edit && git pull does this say "up to date"?

[still comet]

Also up to date

[static canyon]

Huh

What does git remote -v output?

[still comet]

(bot-py3.11) danielgu@1 python-bot % git remote -v
origin  https://github.com/n0Oo0Oo0b/python-bot.git (fetch)
origin  https://github.com/n0Oo0Oo0b/python-bot.git (push)
upstream        https://github.com/python-discord/bot.git (fetch)
upstream        https://github.com/python-discord/bot.git (push)
(bot-py3.11) danielgu@1 python-bot % 

[static canyon]

What about git pull origin/confirm-reminder-edit?

[still comet]

(bot-py3.11) danielgu@1 python-bot % git pull origin/confirm-reminder-edit
fatal: 'origin/confirm-reminder-edit' does not appear to be a git repository
fatal: Could not read from remote repository.

Please make sure you have the correct access rights
and the repository exists.
(bot-py3.11) danielgu@1 python-bot % 

[static canyon]

Err

Maybe a space instead of the /

[still comet]

(bot-py3.11) danielgu@1 python-bot % git pull origin confirm-reminder-edit
From https://github.com/n0Oo0Oo0b/python-bot
 * branch              confirm-reminder-edit -> FETCH_HEAD
Already up to date.
(bot-py3.11) danielgu@1 python-bot % 

I'm not on the computer where I originally made the commits, does that matter?

[static canyon]

I'm trying to remember how to do this

For some reason your local branch differs to the branch on github (https://github.com/n0Oo0Oo0b/python-bot/tree/confirm-reminder-edit)

git pull should fix that

What happens if you do git checkout a8e9c14 && git push?

(@still comet)

[still comet]

(bot-py3.11) danielgu@1 python-bot % git checkout a8e9c14 && git push
error: pathspec 'a8e9c14' did not match any file(s) known to git

[vale ibex]

I'm guessing your local branch matches the remote because you've ran git pull

are you just trying to remove bella as a comitter?

[static canyon]

But their git log doesn't match the remote log (commit hashes are different)

[static canyon]

Yes

[vale ibex]

what's the pr?

[still comet]

Yeah, since I didn't have the changes on my laptop

[static canyon]

^

[still comet]

bot#2719

[dusky shore]

GitHub

[bot] #2719 Make admins confirm edits of other users' reminders

[vale ibex]

this git log looks the same as remote to me

if you run git rebase --no-ff 9d424c098c0151a1051d2d86ff4349274b4250a0

then git push -f it should remove the committer info

[static canyon]

Right, I was looking at the hashes on the PR rather than the remote

[vale ibex]

yea, that's where i was looking

the hashes are the same

[static canyon]

Those hashes aren't the same..

Oh wait

It's inversed

[vale ibex]

yea

[still comet]

That seems to have worked ```
(bot-py3.11) danielgu@1 python-bot % git rebase --no-ff 9d424c098c0151a1051d2d86ff4349274b4250a0
Current branch confirm-reminder-edit is up to date, rebase forced.
Successfully rebased and updated refs/heads/confirm-reminder-edit.
(bot-py3.11) danielgu@1 python-bot % git push -f
Enumerating objects: 76, done.
Counting objects: 100% (76/76), done.
Delta compression using up to 8 threads
Compressing objects: 100% (11/11), done.
Writing objects: 100% (66/66), 6.30 KiB | 239.00 KiB/s, done.
Total 66 (delta 56), reused 55 (delta 55), pack-reused 0
remote: Resolving deltas: 100% (56/56), completed with 5 local objects.
To https://github.com/n0Oo0Oo0b/python-bot.git

• 25ab875f...ee194289 confirm-reminder-edit -> confirm-reminder-edit (forced update)
  (bot-py3.11) danielgu@1 python-bot %

[static canyon]

Mb lmao

[vale ibex]

ah yea, just need to do 1 more acutally

it's exclusive, so doesn't inclujde the commit hash you specify

if you run git rebase --no-ff 9d424c098c0151a1051d2d86ff4349274b4250a0~1

~1 means 1 before this

[static canyon]

Yeah

All git stuff is exclusive on commits iirc

[vale ibex]

then git push -f again

[clever wraith]

is -f short for --force?

[vale ibex]

ye

[still comet]

(bot-py3.11) danielgu@1 python-bot % git rebase --no-ff 9d424c098c0151a1051d2d86ff4349274b4250a0~1
Current branch confirm-reminder-edit is up to date, rebase forced.
Successfully rebased and updated refs/heads/confirm-reminder-edit.
(bot-py3.11) danielgu@1 python-bot % git push -f                                                  
Enumerating objects: 77, done.
Counting objects: 100% (77/77), done.
Delta compression using up to 8 threads
Compressing objects: 100% (12/12), done.
Writing objects: 100% (72/72), 6.81 KiB | 232.00 KiB/s, done.
Total 72 (delta 61), reused 60 (delta 60), pack-reused 0
remote: Resolving deltas: 100% (61/61), completed with 5 local objects.
To https://github.com/n0Oo0Oo0b/python-bot.git
 + ee194289...2bdd46a4 confirm-reminder-edit -> confirm-reminder-edit (forced update)
(bot-py3.11) danielgu@1 python-bot % 
``` like this?

[vale ibex]

yup, that's done it

all your's now

[static canyon]

You can see here: https://github.com/python-discord/bot/pull/2719/commits

[vale ibex]

other than the co-authored ones ofc

[still comet]

Great, thanks for the help!

[vale ibex]

np

will merge once lint passes :)

[tawdry vapor]

Taking this snekbox PR out of draft. I think it's better to split the work into separate PRs. I don't see a reason to wait for the API to merge this. https://github.com/python-discord/snekbox/pull/181

[vale ibex]

Ah alright cool, I'll bump my comment on the approach to an approval then :)

[tawdry vapor]

Thanks. Will leave it open for a bit longer to give others a chance to review

[static canyon]

Can I please get a second review on bot#2520?

[dusky shore]

GitHub

[bot] #2520 Handle off-topic channel names that are invalid for servers in Server Discovery

[clever wraith]

@woeful yew We can merge the RFC PR if you update the branch 😄

[woeful yew]

i just merged the fork with the upstream

[clever wraith]

Nice, I've merged the PR.
Thanks !

[woeful yew]

🎉 first PR! :D

[pulsar elk]

Does anyone need contributions?

[clever wraith]

there is always something to do on python discord repos

you can check issues on any project https://github.com/python-discord

[static canyon]

@clever wraith thanks for the review on this; fyi I've just finished addressing your comments and rewriting the core logic. Haven't been able to test it yet but feel free to leave comments :).

Fwiw it helps if you ping me here on Discord when you leave reviews, since I'm not all-that-active on GitHub and can miss things

[clever wraith]

Sounds good, I'll have another look as soon as I can

[clever wraith]

Man, why does it post my comments when I cancel the review

Anyway, I was trying to get it but something came up, so i'll try to do it tomorrow

[static canyon]

Yeah, no worries, thanks 👍

[vale ibex]

@static canyon @clever wraith just glancing at the OT name PR, haven't looked at it fully to leave a review, but keep in mind that fetching a name from the site api marks it as used

So you will need to make more requests to site afterwards to enable the ones you didn't use

you could also delete the ones that failed too.

[clever wraith]

I remeber reading a discussing where hemlock and mark said we don't want to do that as it's a shame to lose all the ones we've accumulated over time

[clever wraith]

That should be fine, we only don't use the ones that are prohibited, and as soon as discord prohibits us from doing the update, we mark the name as disabled.

Aaaah

I see what you mean

[vale ibex]

no, i mean you're fetching a lot of names from site up front, which disables them

[clever wraith]

Yes yes, good point

[woeful yew]

should the .rfc command be available in pygen or just in the places where ppl have perms to use general lancebot cmds?

[static canyon]

By default the commands are limited to #sir-lancebot-playground (maybe a couple other channels) to help prevent spam in main channels, but I think it'd be okay to allow in #python-discussion (and the help channels if it isn't already) since it's not used that often and isn't a particularly spammy commands. It's basically the same as !pep and !pypi in size, which are already both allowed

Sounds worth creating an issue on the lancebot repo

[static canyon]

Thanks for pointing that out; I have redone the logic a touch in 3b7364c -- I figured since it's so unlikely for a name to be invalid that it makes more sense to just fetch the extra channel name each time one fails (up to the MAX_RENAME_ATTEMPTS).

And just realised that if the final retry fails it'll still add a replacement (even though it won't be used), so added handling for that too

Unfortunately still don't have the time to properly test the rewrite

[static canyon]

Are you sure that fetching a channel name marks it as inactive? I'm testing now and it doesn't seem to...

[vale ibex]

yea, if the 'random_items' kwarg is passed, then they will be marked as used

how many off topic names do you have in your test site?

[static canyon]

    async def _fetch_ot_names(self, count: int) -> list[str]:
        try:
            return await self.bot.api_client.get(
                "bot/off-topic-channel-names", params={"random_items": count}
            )
        except ResponseCodeError as e:
            log.error(f"Failed to get new off-topic channel names: code {e.response.status}")
            raise```
Is `used` the same as `active`?

[vale ibex]

no, they're two different fields

[thorny obsidian]

No, those are distinct

active is it it'll ever come up again, even in future cycles

[vale ibex]

.

[static canyon]

Only 3 atm, have only just gotten time to test stuff so am testing various amounts

[vale ibex]

how many are you requesting?

[static canyon]

3

Actually, that's another point

[vale ibex]

yea, on the first request, they should be marked at used

on the second call they will be made active again

[static canyon]

🤔

[vale ibex]

https://github.com/python-discord/site/blob/main/pydis_site/apps/api/viewsets/bot/off_topic_channel_name.py#L114-L129

[static canyon]

1st call -> succeeds, marks as used
2nd call -> ???

[vale ibex]

first call you'll get 3

actually yea, looking at the code this should work ust fine

it'll always return those 3, and it'll always set used to False

[static canyon]

Yeah, I think it'll automatically reset everything on the 2nd call and then just return them again, and mark as used

Yeah

Yep

Oops

[vale ibex]

yup, channel edits have a tight rate limit

[static canyon]

Do we care about multiple channels being renamed to the same thing being considered valid?

(It's an easy fix if we want that to be considered invalid)

[vale ibex]

how would that be possible?

[static canyon]

Pool of names: ['a', 'b', 'c']

fetch 3 names -> ['a', 'b', 'c']
rename ot0 to 'a'

• 'a' isn't valid, so invalidate and fetch another name
• will fetch either 'b' or 'c', since all names are used (so sets to unused)
• will try again. It (hypothetically) fetches 'b', which is valid, successfully renames

rename ot1 to 'b'

• 'b' is valid, so succeeds

ot0 and ot1 are now both called 'b'

I'm 90% convinced we don't want to allow that

[vale ibex]

So it relies on the pool of names being equal to the number being fetched?

if so, then ignore it, since we won't hit that scenario and it's not worth maintaining extra code

[static canyon]

Nope. It's just the idea that the same name could be picked originally AND in a re-roll

[vale ibex]

how I thought we were getting site to send us all the names up front?

[static canyon]

Nah, I changed that

[static canyon]

^

[vale ibex]

so to hit this scenario then, we'd need to be close to the end of the available pool

fetch one that can't be set

[static canyon]

We just need one name to be invalid

[vale ibex]

and then hit the 1 in 1000+ chance to get the same name after refreshing the pool?

[static canyon]

If the name for ot0 is invalid, then it'll reroll and potentially fetch the same name as was selected for ot1/ot2

This is literally like a 3 line change

Because we just need to pass params={used: false}

[vale ibex]

only if the used flag was reset for all anmes, and then site miraculously returns a name that was fetched int he rpevioous call

since we fetch with random items, it will always fetch used=false

[static canyon]

No, even without that it would

[static canyon]

Oh, if it always does that then it's fine

[vale ibex]

yea, that's kinda the whole point of the flag

[static canyon]

Hmm

But does that persist across separate calls?

[vale ibex]

if you call with random_items, then it will fetch unused names

unless it has none left

then it will reset all names to unused and fetch how ever many you need

[static canyon]

Right

[static canyon]

So yeah, unless this happens it'll never hit a duplicate

[vale ibex]

yea, so just ignore it

[static canyon]

Okay, cool

Right, found a few issues in testing that I've now fixed: c8705f8

[static canyon]

So fyi @clever wraith bot#2520 should be ready for another review whenever you get a chance :)

[tawdry vapor]

After a few years this PR is good to go for final reviews I think https://github.com/python-discord/snekbox/pull/113

[hoary haven]

@still comet thanks for that reminder check PR!

[tawdry vapor]

@vale ibex nice test for the installation of deps! Would it be possible to move the find command into an external shell script and reference it in both the test and deployment.yml?

[vale ibex]

Yea good point that would be better, shiuld be trivial since both CI and the init container have access to the repo files

Can look at it later this evening when I'm home

[tawdry vapor]

Thanks I can review later today

[vale ibex]

snekbox#190

[dusky shore]

GitHub

[snekbox] #190 Make CI and deployment.yaml share the same dependency install script

[vale ibex]

Unsure what's causing the permission issue

docker compose run --rm -T --entrypoint scripts/install_eval_deps.sh snekbox runs fine locally on a fresh image

alright sorted it, likely an issue with the -x bit not being set but somehow working locally

[clever wraith]

since when is the development category above bot commands 😳

[cold island]

🤔

[vale ibex]

docker compose baited me into thinking that the script would be included the actual image, raised snekbox#191

[dusky shore]

GitHub

[snekbox] #191 Include install eval deps in final image

[tawdry vapor]

Can we modify dockerignore instead?

Or do you think it makes more sense to explicitly copy it earlier in the build

[vale ibex]

it already has an entry in docker ignore, we just don't copy that file into the final build stage

[tawdry vapor]

Don't we do copy .?

[vale ibex]

not in the final build step, no

snekbox is pip installed though

I wonder if the files are somewhere in the box

ah, no /usr/local/lib/python3.11/site-packages/snekbox just has the files from the snekbox dir itself (which makes sense thinking on it)

This is why we need to copy in requirements and config manually in the last stage too

[full fractal]

please sir robin, just work 🥺

(I think my old robin configuration from 2022 is outdated)

uhhhh, okay updating the category ID fixed the first crash, but now this

[woeful yew]

making .rfc work in multiple channels is turning out... harder than expected

even @whitelist_override(bypass_defaults=True) doesnt work??

[woeful yew]

update: i added @commands.guild_only() and it works

why

[vale ibex]

it's going to be an issue with one of the roles the cj cog needs being missing

I am planning to move to pydantic which will warn on boot if these are missing, rather than giving unreadable errors like this

I just went throught he same pain on the test server, Add ```
CODE_JAM_API=http://code_jam_management:8000
CATEGORY_SUMMER_CODE_JAM=999762973612589187
CODE_JAM_PARTICIPANTS=476190341566038027
BOT_SIR_ROBIN=804448198642892850
BOT_SIR_LANCEBOT=804448198642892850
BOT_PYTHON=804448198642892850

804448198642892850 is my bot's role, feel free to change it to your own bot's role

[vale ibex]

I've ran into a few dev annoyances around sir-robin, planning to fix each of them in a few PRs

[cold island]

I wonder if the bot could just infer that it's whatever category that has "code jam" in it. Worst case the channel is created in the wrong category

The participants role should be easy to infer as well

[vale ibex]

it does that already

it also creates one if one doesn't exist

[cold island]

then how did the error happen

[vale ibex]

the error was for a role id not being set in config, rather than the category

the error in the screenshot is complaining about an overwrite target being None

[full fractal]

Could Sir Robin get the extension reload command?

I'll pay you with my eternal gratitude if you could add that

[full fractal]

yeah it took me like half an hour to fix my Sir Robin configuration last night...

[timid sentinel]

would be good to get reviews on bot#2436 and bot#2439 if anyone has a chance, they've both been open for a while

[dusky shore]

GitHub

[bot] #2436 Centralise error handling for commands invoked from error handler
[bot] #2439 Make help showable through button on command error message.

[vale ibex]

reviewed 👌

[coarse salmon]

hey i have one idea

like, maybe a bot which is assistant? like the feature in stackoverflow where it asks "is question related to these?" we can use it in python-help

or we can add it to sirlance bot

as a feature

[kind flower]

I dont think being more like stackoverflow is a goal we should have.
besides that our help forum is meant to be a live conversation, meaning people are seeking help with specific problems, so I personally dont think this would fit in well here.

how would it even work? would we keep a cache of all help posts ever posted, or use discord search (can bots do that?) to find similar posts?

[coarse salmon]

yes

it can make the users find the help a bit fast

if it's quite same

[clever wraith]

since you are on help channels topic i had an idea for a while to add dwm (did we manage) tag (can be named better just saw this name on few servers) that would simply tell user that if he doesnt need help anymore to run !close :)

ill make this an issue on meta prolly soon

[clever wraith]

I don't think this is really needed

We used to want people to close posts since we had a limit with the old system

Now that each person gets their own post, the timeout itself will do.

[clever wraith]

Like vivax said, the entire purpose is to get the conversational part

2 people might have the same question

But they won't necessarily ask the same followup questions

Or have the same base knowledge

The entire purpose of what we're trying to do is to give full liberty to the helpee to ask whatever they want and respond in real time

Plus, the complexity of what you're proposing is a bit not clear as we'll need more details

How will you manage the old posts and index them

And how do you see the content comparison being determined?

[clever wraith]

fair

[patent pivot]

#dev-log message

PR closure speedrun? thoughts?

[coarse salmon]

Maybe your right

[clever wraith]

Teach me senpai

Great work!

[static canyon]

@clever wraith do you think you'll be able to take another look at bot#2520 soon™️?

[dusky shore]

GitHub

[bot] #2520 Handle off-topic channel names that are invalid for servers in Server Discovery

[clever wraith]

Hey,

I don't think I'll be able to have a look this weekend, I'll try to get to it during the week

[static canyon]

Cool, thanks 😄

[clever wraith]

hey, having trouble running site with docker
full error: https://paste.pythondiscord.com/SB7Q

by investigating a bit i found that it comes from libsass cause they use https://github.com/sass/libsass-python/blob/main/setup.cfg#L9 outdated param it should be license_files quite not sure what to do from now

[stable mountain]

setup.cfg line 9

license_file = LICENSE```

[gritty wind]

Try installing the library manually with pip, and opening an issue upstream if it does not already exist

[clever wraith]

yeah i managed to get it working locally apparently doesnt work when docker tries to install it

[vale ibex]

we use the latest version of poetry in our dockerfile, so it's likely that new versions of peotry are using a newer version of setup tools to install packages

[rapid igloo]

hi all, what's the status of pydis and hacktoberfest in relation to the hacktoberfest commands for sir lancebot?

according to this
#dev-branding message
we're still keeping the commands. but afaik hackstats is broken (possibly for almost 2 years). plus, we have a stale PR from sir-lancebot#919, which includes suggestions for improvements of logic in the cog, not to mention hacktoberfest accepting gitlab MRs now. I feel like if we're keeping the cog it should be fixed and not be left to grow spiderwebs (pun intended). since I wrote much of the unlinted code () for hackstats' 2020 logic, I feel responsible for picking up the work for refactoring the cog.

would this be in scope for sir lancebot? or are there newer updates internally to what we're going to do with this cog?

speaking of updates, the reason I'm unsure is that there were no public announcements in #announcements or #changelog AFAICT in 2021 (nor 2022) about this decision given that pydis participated in hacktoberfest since like 2018. the branding was archived and from a non-staff's perspective suddenly python bot starts announcing the halloween event rather than hacktoberfest. the only explanations visible were in #dev-contrib and #dev-branding

[dusky shore]

GitHub

[sir-lancebot] #919 Hacktoberfest Re-organisation

[thorny obsidian]

There was an internal decision made about how we (as a server) participate in Hacktoberfest. We didn't feel it necessary to announce we weren't making it a whole event because it wasn't widely participated in in this server and it's not an event we run either.

If someone would like to get the Hacktoberfest cog up and running they're certainly welcome to, but for me it's not at all a priority.

[atomic ivy]

We're not a recruitment board.

[subtle kraken]

if am not mistaken
python/peps#3418 might have broken our !pep command from the looks of it

[dusky shore]

GitHub

[peps] #3418 All PEPs: Move to peps/ folder

[subtle kraken]

since they have been moved from the root folder to peps/ folder our cache cant seem to build anything out of the request it gets

[sharp timber]

looks like we probably just have to update the pep listing api url

[subtle kraken]

if am not mistaken it could be just 1 loc change?

yeah yeah

[sharp timber]

currently hitting https://api.github.com/repos/python/peps/contents?ref=main which doesn't have the peps in it anymore

looks like https://api.github.com/repos/python/peps/contents/peps?ref=main is the updated url

[subtle kraken]

if am not mistaken https://api.github.com/repos/python/peps/contents/peps?ref=main would fix it

thanks bast

[sharp timber]

lmfao

Woulda been faster but github really didnt' want to link me to the actual api docs for some reason

[subtle kraken]

hah

[vale ibex]

So this? ```diff
diff --git a/bot/exts/info/pep.py b/bot/exts/info/pep.py
index 08ccbddc..2b552dc4 100644
--- a/bot/exts/info/pep.py
+++ b/bot/exts/info/pep.py
@@ -14,7 +14,7 @@ log = get_logger(name)

ICON_URL = "https://www.python.org/static/opengraph-icon-200x200.png"
BASE_PEP_URL = "https://peps.python.org/pep-"
-PEPS_LISTING_API_URL = "https://api.github.com/repos/python/peps/contents?ref=main"
+PEPS_LISTING_API_URL = "https://api.github.com/repos/python/peps/contents/peps?ref=main"

pep_cache = AsyncCache()

[subtle kraken]

sgtm

but dont put me on the blamelist thanks

[vale ibex]

lol np just running locally now

[subtle kraken]

do put me on authors list though

[vale ibex]

looks good

[atomic ivy]

Good/bad news! They're the same list

[subtle kraken]

just put NO BLAME next to my name

thanks

[atomic ivy]

for a moment, you had me thinking that's an actual feature

[vale ibex]

pelase give co-authored-by name & email

ty

also Bast if you want to be on commit

[subtle kraken]

what name

[vale ibex]

github Co-authored-by

need your gh username and email

[subtle kraken]

sowwy
never used that feat, lemme find my email

the gh public one innit

[vale ibex]

yea

git config --get user.name && git config --get user.email

will do it

[subtle kraken]

Senjan21
[e-mail]```

thanks

[vale ibex]

bot#2758

[dusky shore]

GitHub

[bot] #2758 Update pep listings URL

[subtle kraken]

is it even appropriate for me to review a PR I coauthored

[vale ibex]

Yea

The point of reviewing is to get multiple sets of eyes on the PR, so if multiple people worked on something, so long as they actually review their own code it achieves the same goal

Not to say that sometimes having fresh eyes on something isn't good

[atomic ivy]

how does that work for one-line changes?

[subtle kraken]

I reviewed it then 😩

[vale ibex]

Hah yea, I'm talking more generally

One liners like this are better served by just testing the code works

[past epoch]

^ @outer oasis

[outer oasis]

[tawdry vapor]

@fossil veldt Hey, any ideas why this test starts failing when both of the following are done?

1. Upgrade the image to Debian Bookworm
2. Enable /dev/shm in the container (by way of ipc=private)

If only one of those is done, then the test passes (though on Bookworm, integ tests fail because shm is needed but Bookworm's newer version of glibc no longer falls back to /tmp when /dev/shm is missing)

[stable mountain]

tests/test_nsjail.py line 228

def test_file_parsing_timeout(self):```

[fossil veldt]

hm... what's the failing assertion?

[tawdry vapor]

Oh sorry, meant to include that. The return code is 0 instead of None

Which I guess means it's not timing out at all.

The tests for the timer utility still pass though

[fossil veldt]

strange, I wonder if it just read the files really quickly, or the symlinks just didn't work in the test script

perhaps add a test for result.files being empty? Or if you break there and see if it returned all the files

[tawdry vapor]

Haven't set up remote debugging so won't be able to set breakpoint.

But yeah I could just try increasing the number of files

[fossil veldt]

also by "enable dev/shm" you mean on the docker running snekbox or the nsjail container?

[tawdry vapor]

On the container. But we also merged a change recently to enable it in nsjail.

Though nsjail just uses a new tmpfs mount - it doesn't rely on /dev/shm

Well, I ran the test again with no changes and now it's fine?

IDK what's up

[vale ibex]

Ah, missed this conversation here and added some comments on the PR.

[vale ibex]

Found the issue after some further digging

the test is failing due to a timing issue, there's a race between the timeout and the max file size. If your PC is fast enough, you'll parse the number of files needed to hit the max file size limit before the timeout of 1s

so CI passes because it's (relatively) slow

@fossil veldt @tawdry vapor

the only way I've found to make this pass on my PC is to increase memfs_instance_size on that test

I had to double it to make it just go above 1s

It's easier to see this if you change the coverage command to target that test specifically so you get stdout IE ```diff
-'coverage run -m unittest; e=$?; chown --reference=. .coverage; exit $e'
+'coverage run -m unittest tests.test_nsjail.NsJailTests.test_file_parsing_timeout; e=$?; chown --reference=. .coverage; exit $e'

I'm guessing bookworm got faster at whatever that codepath is doing

[tawdry vapor]

Why would increasing the memfs size increase processing time? I would expect that to be correlated to the size of the files or the amount of files instead.

[vale ibex]

it seems to process one file at a time, summing the size of files as they're processed

so it will kepp processing files until either the timeout is hit, or the sum of processed files exceeds the memfs limit given

[tawdry vapor]

So increasing the memfs size lets it process more files which means more time spent

[vale ibex]

yea

and signal.alarm() only takes an integer amount of seconds, so we can't reduce timeout below 1 for the test

[tawdry vapor]

I'm confused why there are two different places where it tracks the time

One is with the with time_limit(self.files_timeout):. The other is in fs.files_list which uses time.monotonic. Do we really need both?

https://github.com/python-discord/snekbox/commit/46dd83036e28dd2b0aa0179afafd9e3867bff37d

[vale ibex]

I shouldn't think so, the context manager should catch it in all cases since its sending an inturrupt

There was at least a 3rd place I saw us tracking time too, not sure if it's also under that context manager or not

Making some dinner atm

[tawdry vapor]

I'd rather not have our tests require a minimum of 1 GiB of RAM. I'm looking into using signal.setitimer instead of signal.alarm, since the former can accept a float

[vale ibex]

Hah yea seems like a reasonable call

[tawdry vapor]

How does this work? Why does it alarm again in the finally block?

[stable mountain]

snekbox/utils/timed.py line 34

signal.alarm(0)```

[tawdry vapor]

Does this mean it always raises a timeout exception when the context manager is exited?

It's supposed to "disable the alarm" according to one of the examples in the docs

That makes sense - if the code finishes early it shouldn't let the alarm still go through

[patent pivot]

lol i love snekbox

[vale ibex]

Haha yea I went through the same process earlier

Annoyingly windows doesn't support sigalrm so I don't get any intellisense feedback on signal.SIGALRM nor signal.alarm

[fossil veldt]

Looking at it, there was a previous change that skips file processing when the the total size exceeds the instance size, so if the memfs instance size is too low it just skips further files and doesn't trigger the timeout - hence the test fail https://github.com/python-discord/snekbox/commit/16b1a13e206fa34bfc7af05363d5e78742e26e40

It might be better to change the test to mock patch the file read / base64 encode to do a time.sleep(2) with a parsing time limit of 1 or something

[tawdry vapor]

Was debating that but I was able to change the timer to support floats for timeouts, so we can set the timeout to be so low that pretty much any machine would time out

Why is this doing -5 by the way? https://github.com/python-discord/snekbox/blob/main/tests/test_nsjail.py#L236

[stable mountain]

tests/test_nsjail.py line 236

for _ in range((size // 1024) - 5):```

[fossil veldt]

Yeah I guess that also works, essentially it just needs to test that it can actually interrupt when the time limit is up

[fossil veldt]

I think it was exceeding the instance size without it or something? Don't really remember, the per file metadata / symlink also takes up some space or something

[tawdry vapor]

The memfs size was being set to 512 MiB while the code was only writing 32 MiB - 5 if I understand correctly

I see symlinks are followed, so the total size includes the underlying file's size for each symlink. But that would mean the total size used by that test code would be like 2.7 GB

[vale ibex]

wouldn't each symlink be tiny?

[tawdry vapor]

Yeah but we keep a running total of the size of all encountered files. And we follow symlinks so we count the size of the underlying file, not the size of the symlink

[vale ibex]

ahhh right yea

So yea in that case memfs size might be 512MiB but in fact we don't actually use that much ram

[tawdry vapor]

But it's a running total so we don't need to set the memfs size to be over the total - it just needs to be large enough that we will timeout before we reach the size limit

I've updated the PR with a fix for that test now

[vale ibex]

Nice one

I love being bleeding edge

[cursive relic]

Rule command seems to be down for some reason (and maybe !d as well?)

!rule ad

[stable mountain]

Sorry, an unexpected error occurred. Please let us know!

ClientConnectorError: Cannot connect to host site.default.svc.cluster.local:80 ssl:default [Connect call failed ('10.128.90.73', 80)]

[cursive relic]

^ happens

[stable mountain]

Sorry, an unexpected error occurred. Please let us know!

ClientConnectorError: Cannot connect to host site.default.svc.cluster.local:80 ssl:default [Connect call failed ('10.128.90.73', 80)]

[static canyon]

Yep, we're aware -- it's being investigated internally

[vale ibex]

Yea, it's all resolved now

[grim prairie]

I have an idea... what if (optionally) sharing code in help threads could work as a repository like GitHub? I.e. when a thread is created you get a repo link to which you could push your code to. And helpers could clone the repo to run code and suggest changes? Pushed code would show up as a bot message.

[clever wraith]

I'm not sure of how much of a benefit that will carry.

Being able to simplify the problem and explaining it is, IMO, a very good exercice for the helpee.

As you might know, we appreciate it when people demonstrate effort in implementing solutions and trying things, since that clearly says '' I've tried this, but it doesn't work and I don't know why '' instead of '' hey, do the work for me ''

Plus, all the implementation that this carries isn't trivial, how would you hook that repo to the appropriate channel? Who owns the repo? Etc.

To resume, I don't belive this will be of value to the system we already have.

Let's not mention people who might put malicious code, so that an innocent people trying to help get caught in it

[grim prairie]

All fair points. I was envisioning the help thread would still be the central point of communication.

And I guess for 95+% of help threads the user may not be familiar enough with a vcs to effectively use it

And for the 5% that are; they're probably already using one

[clever wraith]

Yes, that too.

Don't get me wrong, i didn't mean to kill your idea because it could indeed make some things easier.

It's just that when weighing out the pros and cons, it doesn't balance out.

[vocal wolf]

yo @dull swan is there an issue for https://github.com/python-discord/bot/pull/2749?

[dull swan]

I don’t think so, we were talking about it in PyGen and I just pushed the fix up real quick. I can make an issue for it.

[vocal wolf]

nah, was just wondering if there were other examples of the problem

it's all good

[vocal wolf]

@dull swan henlo it me again

I tested the PR, works as intended, very nice. Once another person approves, it'll be merged.

[dull swan]

Thanks Xith!

[vocal wolf]

@clever wraith Hello, I shall be testing this PR out in the very near future. In the meantime, could you fix the conflicts? Thanks!

sir-lancebot#1197

[dusky shore]

GitHub

[sir-lancebot] #1197 Make deleting a bookmark a context menu command

[vocal wolf]

I have reviewed

[clever wraith]

Hellew,

I will try to get to it when I can, my pc has been broken for a while so i can't do much

Thanks for reviewing

[vocal wolf]

rip compooper

[clever wraith]

Indeed, indeed

[devout sand]

2 post didn't auto close in #1035199133436354600 after more than 1 hours

[static canyon]

Without linking to the posts in question, there's not really anything we can do

(Also keep in mind it's an hour from last message, not an hour from opening)

[devout sand]

Yeah

It was closed by a mod manually yesterday

Like one that exist for like 5 hours after last message

And another one have multiple section that have more than 1 hours of inactivity and didn't automatically closed

[static canyon]

Again, without links there's nothing we can do

Most of the time though the issue is on Discord's end, not ours

[devout sand]

https://discord.com/channels/267624335836053506/1157653197305958463

https://discord.com/channels/267624335836053506/1157637341280751636

@static canyon

[static canyon]

Yeah, iirc those are both around the time discord was having issues yesterday, so I don't think it's anything too noteworthy in this case

But thanks for letting us know :)

[devout sand]

Oh ok

Np

[signal saffron]

1

[sharp crag]

I found this interesting:
https://github.com/python-discord/bot/blob/main/bot/exts/help_channels/_cog.py#L111
Since has_any_role throws an error instead of returning False, does that mean if someone that's not helper tries to use !title in a help channel, it'll throw an exception?

[stable mountain]

bot/exts/help_channels/_cog.py line 111

if not await commands.has_any_role(constants.Roles.helpers).predicate(ctx):```

[static canyon]

If it does indeed throw an error, then yeah

But I imagine it's being handled somewhere

Specifically, here: https://github.com/python-discord/bot/blob/main/bot/exts/backend/error_handler.py#L88-L90

[stable mountain]

bot/exts/backend/error_handler.py lines 88 to 90

elif isinstance(e, errors.CheckFailure):
    log.debug(debug_message)
    await self.handle_check_failure(ctx, e)```

[static canyon]

Altho it actually just gets silently suppressed in handle_check_failure

So even if it does error, it just gets silently suppressed, meaning you'd never notice it

Actually, think it's classified as a ContextCheckFailure, which means it'll just send a message saying you can't use the command in that channel

[static canyon]

Tl;Dr; it errors, but it's handled via a message telling the invoker they can't use the command in that channel

[sharp crag]

Ah, interesting. I would've thought they'd just have done something like constants.Roles.helpers in {role.id for role in ctx.roles}

[static canyon]

I guess it should technically instead use this util that we created: https://github.com/python-discord/bot/blob/main/bot/utils/checks.py#L97-L107

[stable mountain]

bot/utils/checks.py lines 97 to 107

async def has_any_role_check(ctx: Context, *roles: str | int) -> bool:
    """
    Returns True if the context's author has any of the specified roles.

    `​roles`​ are the names or IDs of the roles for which to check.
    False is always returns if the context is outside a guild.
    """
    try:
        return await has_any_role(*roles).predicate(ctx)
    except CheckFailure:
        return False```

[sharp crag]

or that, yeah

I think it's more explicit

Left a review. @trim cradle cc^

More of a suggestion than requiring changes

[static canyon]

@trim cradle ^ would be nice to keep the comment, but not a big thing

[gleaming jay]

Heya fellas, coming through with some concerns involving the configure task in poetry
I'm setting up, plaster the guild id right into my .env, give 'er a whirl, and I'm getting a 404 back.
The route looks fine; I see nothing wrong with it at all. I've quadruple-checked to make sure I'm not pasting in the guild ID wrong or some other shenanigans, but no, everything is as it should be.

Am I missing something, or can someone replicate this behavior?

[sharp crag]

Which route is this? Also can you paste the traceback, if you have any?

[gleaming jay]

Sure can, just a moment

here's the traceback,

it's on the guilds/{guild_id}/roles route

(also cool, I didn't know we switched to httpx internally )

[sharp crag]

That route does seem correct. And since you say you've checked the guild ID...

[gleaming jay]

Hmmm, I'm a tad curious now, I'm going to mod the script a bit to see if a hunch does something

mmm, no

[sharp crag]

Perhaps it's transient? Discord did recently just die about half an hour ago. My bots went offline with 503. Maybe it's just that?

[gleaming jay]

Mmm, no, I was having this issue last evening as well.

the earliest occurence being... ~17:00?

[sharp crag]

I'm assuming you've checked that the bot is already in the server which you're bootstrapping?

[gleaming jay]

Yep yep.

Odd, it worked this time.

The hell?

[sharp crag]

Discord is drunk again it seems

[gleaming jay]

:3

well, now I can start working on making !raw a context menu thingie

gracias, robino

[sharp crag]

I didn't do anything but yeah, haha

[gleaming jay]

you gave some mean side-eye

put the api back in its place

[sharp crag]

Noted next time the API is acting up again. Just look at it funny.

[sharp crag]

Oh, and feel free to request a review from me on that one. I'm planning on taking on a similar PR (allow !raw to be used while replying to a message) so some insight on how you do it might be helpful

[gleaming jay]

I was actually planning to squash both sequentially

but I will give you a shout for sure

[sharp crag]

thanks c:

[devout sand]

Would be funny if the stable is corrupted

[vale ibex]

snekbox#195 is up for 3.12

[dusky shore]

GitHub

[snekbox] #195 Python 3.12

[vale ibex]

Since the above PR has been merged, bot#2777 is up to update the bot's output

[dusky shore]

GitHub

[bot] #2777 Update all references of 3.11 in eval to 3.12

[vale ibex]

I've also added support for 3.12 in our poetry base image, but a few packages we use, such as mutlidict and yarl, don't support it quite yet. Might be a few weeks until we can bump the version in our projects

[clever wraith]

!eval command needs updating

#bot-commands message
3.11 eval job completed even though its 3.12

also its not possible to switch to 3.11 either way using !eval 3.11 <code> syntax

and i remember making pr about changing the description to this Currently only 3.11 version is supported. - should be changed also

[vale ibex]

look 2 messages up

[clever wraith]

oh. so you have a pr about it

[vale ibex]

Yea :D

[clever wraith]

👍

[median ferry]

hi, so this is my first time contributing wanna get into open sourcing, have identified a small the issue and commented on it, how does it go from here?
#dev-log message

eventually wanna get to working on bigger bugs and bots issues as well

[clever wraith]

you wait for the core developer to comment on it and if its a real bug you can be assigned to it and start working to fix it

[median ferry]

Oh okay thanks

[atomic ivy]

!rule 6

[stable mountain]

Rules

6. Do not post unapproved advertising.