dev-contrib

[tawdry vapor]

I think the point of doing it is to make sure checks like tests still pass when merging

But it feels redundant to have 2 merge commits in a row

[subtle kraken]

suppose its good for bigger PR's then

[cold moon]

@subtle kraken I will look into it (and into all other my pending PRs) when my PrestaShop PR is done.

[subtle kraken]

sounds good to me

[crude gyro]

it's tempting to press that big juicy update branch button

but ideally we should just do it right before we're gonna merge

so a good rule of thumb is to not bother until you already got approvals.

[subtle kraken]

I only have been updating just before merging so far

[crude gyro]

then why would there be two in a row?

[subtle kraken]

To which PR are you referring ?

[crude gyro]

ask @tawdry vapor

[tawdry vapor]

2 in a row because the first merges master into feature, and the second merges feature into master

If the branch is out of date you cannot merge the PR without first updating the branch, unless you use admin overrides.

I'm saying to just use admin override most of the time

Worst that will happen is a test or lint will fail and it won't get deployed, which may take an extra commit to fix but you would have needed to do that commit anyway. Difference is you now have to do it after merging the PR rather than before.

So if you're willing to take on that responsibility instead of the PR author then force the merge 👿

[tough imp]

can merge & run tests locally

but I always thought the double merge was something that was agreed we'd do

because it's safer

[crude gyro]

yeah I don't really like your suggestion, @tawdry vapor. there's a risk the CI will fail on the merge then, and that's messier to clean up.

[tawdry vapor]

Messier in what sense?

[crude gyro]

you'd need to push your fix to master or open another PR

[tawdry vapor]

How is a push to master messy?

[crude gyro]

and master would be in an unstable state.

until it was fixed.

an unstable master is just bad.

[tawdry vapor]

Point is it shouldn't happen anyway for these small PRs where only 1 string is changed

[crude gyro]

probably not. but I think it's nice merge an updated branch as a rule of thumb, and honestly the commit history is not something I ever worry about.

commits are for blame and search, not for browsing

and a merge commit doesnt show up in blame annotations

so it never really enters my world.

[tawdry vapor]

I sometimes search by browsing the log

It's not a big deal

Just a suggestion that you can use your discretion to do it. if you don't want to that's fine.

[crude gyro]

yeah. I sometimes would for a oneliner.

[tawdry vapor]

Why is Dependabot unable to open PRs for JS dependencies on the site

It just shits itself

[patent pivot]

incompatibilities between deps we pin at and the new ones

JS deps are a shitstorm

[tawdry vapor]

Oh

It's cause we don't even have a package-lock.json

What the fuck?

What is it analysing then?

[cold moon]

When you don't want "merge master into" commit then use rebase + force push.

[woeful thorn]

I don't think we should be advocating for rebasing and force pushing as a rule of thumb

[brazen charm]

Would adding the pr # to commits be something conventional? Think I've seen it from someone here and realised it would've been handy the few times I was looking through the history

[tawdry vapor]

I don't mind either way

May be usefully locally, but GH already shows relevant PRs when looking at commits

[brazen charm]

Didn't notice that when looking through specific commits in the history, is that new or did I just miss it?

[tawdry vapor]

Been around for as long as I can remember

[brazen token]

how did yall create a timed mute system on the python bot

[tawdry vapor]

Through a custom scheduler https://github.com/python-discord/bot/blob/master/bot/utils/scheduling.py

Combined with storing infractions in a database for persistence. They are re-scheduled when the bot restarts.

[green mesa]

I have one question about modmail I mean when someone message modmail it creates threat I wnt to know when it will close threat

Like if any mod message user will user get the message through modmail

Can user can know who is messaging?

[indigo wedge]

where is everyone usually typing?

I'm new here and there are lots of channels!

[tawdry vapor]

#python-discussion I suppose

[indigo wedge]

ok thnx

[crude gyro]

@green mesa modmail isn't a pydis project, and I have no idea what you're asking anyway

[hardy gorge]

@green mesa We've configured it that threads are closed by moderators on our end. I think ModMail has an option to allow the other party to close the threads too, but we're not using it. (As @crude gyro mentioned, it's not one of our projects; we're just using it.)

As for the identity of the mod who replied, we typically use a reply command that includes username of the author in the reply. It means that people should typically be able to see which mod replied to them. There's also an anonymous reply option that does not reveal the identity of the mod but rather says something like "Python Discord Moderators". This means the answer is "it depends", I guess, as it depends on which option we use to reply.

[green mesa]

Ok ty ves for this answer I am looking for

[neon garnet]

mypy does have a way to be used programmatically https://mypy.readthedocs.io/en/stable/extending_mypy.html

[green oriole]

Yeah, but is that completely safe?

We could do the same thing as for the regex command, format some code and send it to snekbox

[neon garnet]

I'll take a look around first

[green oriole]

Here it is https://github.com/python-discord/bot/pull/952

[neon garnet]

Ohhh, that's fair

Think I'll set up the dev environment soon to check a few things

[green oriole]

Good luck!

[neon garnet]

Thanks 😄

[tawdry vapor]

@brazen charm If I address your requests in https://github.com/python-discord/bot/pull/957 would you mind looking it over again soon for approval?

[brazen charm]

not at all, should have the time

[tawdry vapor]

excellent

[brazen charm]

@patent pivot how much of a non issue was the memory usage? Poking around I found out that just importing intersphinx eats up 10-20MB of memory and then fetching inventories seems to keep some unecessary references alive. If we copied the code and made it more suitable for the bot's use case we could at least drop that usage while getting rid of the big dependency and adjusting it for async instead of running in executors

[patent pivot]

hmmmm

memory usage is a non-issue as long as it isn't leaking

we have enough

[brazen charm]

Also more on the side of getting rid of sphinx when only 5 definitions from of the whole package are used through the inventory fetching

[tawdry vapor]

I've pushed those changes btw

[cold moon]

Sorry about not addressing reviews, but I got extra requirements for paid PrestaShop module development and I have to finish this first.

[tawdry vapor]

No problem since it's mostly low priority

This just got bumped up since I need to add support for multiple code jam categories

And the jam is coming up

[brazen charm]

@tawdry vapor Why have you decided against the tests in the code block detection PR? Just wondering if it had something to do with parsing as I'm not quite sure how to tackle it for the doc html parsing

[tawdry vapor]

Because I don't feel I have an adequate understanding of software testing to write tests

Not really related to what was being tested

For testing HTML parsing you can have sample HTML documents and pass them to the parser, then assert the results

[brazen charm]

You've made some tests PRs to the bot before, what changed your mind about your abilities? I consider myself not very experienced with testing and sometimes it feels like I'm doing things the wrong way so would be nice to have some insight. Here it'd be useful because of all the checks I do after introducing a change and still miss things. The code is atomic enough to be tested with normal asserts at points but I feel like exact asserts, or without looking for some structure, on the end result would be too restrictive on changes to the code

[tawdry vapor]

I cannot reconcile the conundrum of precise tests vs. decoupled tests

[shut aspen]

just thought id throw this out there, i implemented the ability for my bot to print its own source and ive found it crazy useful on my server

[subtle kraken]

I think we had that idea but it would easily flood the chat
Especially if some command takes many lines

[shut aspen]

ya it could prob be abused, tho the ability to specify lines is nice

i have just found personally i use it non stop when developing and helping others contribute

cuz they can show me what theyve written in the test server with just a command

[subtle kraken]

source command of @stable mountain takes command/cog as input so you dont need to know files nor lines

[shut aspen]

ya mine defaults to the whole file if you dont give it lines

itll chunk it and send it all in the channel, which yes gets wordy lol

its more used for testing and contributing then as a real server command

idk ive found it useful, thought id suggest it

[celest charm]

You could use := in that method

if (event := getattr(value, __event_listener__)):
    messenger.subscribe(event, value)

And I have a suspicion you might be reinventing or overengineering something...

wrong channel, sorry

[shut aspen]

@celest charm I’m still on 3.7 or i would 😃

[celest charm]

oh

[shut aspen]

Ya Big F ik lol

[celest charm]

@shut aspen Sorry, my keyboard screwed up somehow, 3.7 is understandable

Anyway, you could use getattr(obj, name, None) to make it fall back to None

[shut aspen]

I would update but that requires the entire community to update and that’s a pain lol

[celest charm]

true

[shut aspen]

And ahh ok, that’s the default overload i assume

[celest charm]

anyway, this is off-topic here

[brazen charm]

I'm creating a second subclass now for doc parsing to change behaviour of the used modules, do you think it'd be worth it to place them in a separate module so only the relevant parsing parts are in the module that does parsing?

Taking around a 100 lines now because some of the code had to be copied over

[tawdry vapor]

Don't know what the code is exactly but separating them sounds good.

[brazen charm]

Any ideas on a good name for it? The subclasses are only marginally related, one is the MarkdownConverter subclass which is in the master code (+ some methods) and second is from one of bs4's objects to allow some extra behaviour

[tawdry vapor]

So one is for MD and one for HTML?

[brazen charm]

Yeah could look at it like that https://paste.fuelrats.com/bagevuvayo.py

[tawdry vapor]

So you could name one markdown and the other html

[tawdry vapor]

Is it feasible to make the Bot instance a global? Getting tired of having to pass it around or create classes to store it.

I'm imagining it to be pretty simple: create the instance in another module, then import that in __main__. That'll avoid circular imports.

[tough imp]

seasonalbot instantiates it in the module where it's defined

then the instance is imported into __main__

I think it's probably a feasible approach

[tawdry vapor]

My concern is that it's not explicit, so it may be a gotcha if someone tries to use code that relies on the bot. It's a gotcha since obviously the bot has to be running in like 95% of the cases it'll be used for.

So there'll be static method or functions in modules that are ostensibly fine to call before the bot is running, but they aren't.

But the convenience of it being global outweighs that risk, I think.

[tough imp]

the only time where I wanted to import the instance somewhere where I didn't have it already (so neither main nor within a cog instance) was when I was doing something that I probably shouldn't have been doing, it felt dirty and you talked me out of it iirc

and yeah you're right

hmm, what are you working on by the way?

[tawdry vapor]

Syncers 3.0

Rewriting it again cause I found the previous design too hard to work with

[glass pecan]

@tawdry vapor what you describe is why i originally had the seasonalbot instance bot within constants.py before we had a dedicated bot.py. Now that it's in it's own module, it can be imported from there instead, which should be fine enough of a setup in place of a global.

Wouldn't take much to set a similar setup in Python bot.

There's a couple of other methods that are slightly over-engineered for now that I've considered using in my own bots too, and was actively working on when I was more active on my own projects that allow for globally accessed commonly needed contextual variables via a proxy object.

No matter the setup though, I'm not a fan of plain globals

The IDE doesn't detect what it is without additional tweaking and typehints.

It's also not intuitive to new contributors

[crude gyro]

yeah I also think it's a bit implicit. but having the bot be some sort of singleton that you can import really easily seems alright. kind of like how flask does from flask import g

[glass pecan]

ye, that's pretty much what it would be

could make it top-level even

from bot import bot seems nicer than

from bot.bot import bot lol

so much bot

[crude gyro]

yeah, from bot import bot seems alright to me

[glass pecan]

it may require a proxy of sorts to do it, but i'm sure it'll be fine

mostly to avoid cyclic imports

[tawdry vapor]

I'd appreciate a review on this if anyone has time this weekend https://github.com/python-discord/bot/pull/835

If you're keen I also have 2 more PRs you can take a look at

[hardy gorge]

I'm working on it. There are some interesting things in this PR

[crude gyro]

if you're handing out reviews, I'd still love one for the bot filterlist PR

[brazen charm]

Speaking of reviews, the pr for truncating internal eval should be ready now

[tawdry vapor]

Yeah, we're getting a bit behind on reviews. But I suppose none are high priority

Anyway, thanks for having a look

[brazen charm]

I was skipping the more async heavy ones when reviewing before since I don't understand python async that well, but can look into it more later if it's needed

[obsidian patio]

Do I have to use pipenv for the bot? I haven't used pipenv before and it's acting in a weird way for me. Would something like venv work as well?

[sullen phoenix]

you shouldn't really be directly using pipenv unless you're working on something for site as well that the bot depends on

have you seen the setup guide?

[obsidian patio]

Yeah. I'm going through it right now

This part:
https://pythondiscord.com/pages/contributing/installing-project-dependencies/#installing-dependencies-with-the-command-line

It asks me to run the following two commands:
pipenv sync --dev
pipenv run precommit

[sullen phoenix]

ahh alright, what troubles is pipenv giving you?

[obsidian patio]

RuntimeError: location not created nor specified is probably the main part of it. The traceback starts with the following though:
Courtesy Notice: Pipenv found itself running within a virtual environment, so it will automatically use that environment, instead of creating its own for any project. You can set PIPENV_IGNORE_VIRTUALENVS=1 to force pipenv to ignore that environment and create its own instead. You can set PIPENV_VERBOSITY=-1 to suppress this warning.

[sullen phoenix]

you can ignore that

[obsidian patio]

I cloned the repo to an external SSD, which may have soemthing to do with it

[sullen phoenix]

what's the rest of it?

[obsidian patio]

Failed to load paths: /bin/sh: /Users/X/.local/share/virtualenvs/bot-6jSLwY9f/bin/python: No such file or directory

Output: 
Failed to load paths: /bin/sh: /Users/X/.local/share/virtualenvs/bot-6jSLwY9f/bin/python: No such file or directory

Output: 
Failed to load paths: /bin/sh: /Users/X/.local/share/virtualenvs/bot-6jSLwY9f/bin/python: No such file or directory

And a really long traceback that ends with the RuntimeError I showed you earlier

[sullen phoenix]

right

i usually get that when using docker, but it works anyways

i guess you could try

pip install -U "pipenv==2018.11.26"

[obsidian patio]

I'm not running docker though?

I'll see if that helps

[sullen phoenix]

this probably isn't the issue, but can you comfirm that this does exist?

/Users/X/.local/share/virtualenvs/bot-6jSLwY9f/bin/python

[obsidian patio]

I tried to run that pip install. It seems like it's not giving me any errors, yet

[sullen phoenix]

alright

[obsidian patio]

I think it worked. Should I run pipenv shell to activate it, as it tells me that I can do, or can I leave the pipenv from now on?

[sullen phoenix]

you shouldn't need to activate it

maybe try and run lint

pipenv run lint

[obsidian patio]

What will that do?

[sullen phoenix]

if that works, you should be set

that'll just lint the code

[obsidian patio]

Aah okay

[sullen phoenix]

or even better.. run the tests pipenv run test

[obsidian patio]

Already running the lint now

Seems like it passed the lint

Thanks for the help :)

[sullen phoenix]

alright, cool

no problem!

[obsidian patio]

I'm a bit confused about this part. What do you actually want me to do?https://pythondiscord.com/pages/contributing/bot/#server-setup

This part:

Setup categories, channels, emojis, roles, and webhooks in your server. To see what needs to be added, please refer to the following sections in the config-default.yml file:

style.emojis
guild.categories
guild.channels
guild.roles
guild.webhooks

We understand this is tedious and are working on a better solution for setting up test servers.

[sullen phoenix]

you need to setup a test server for the bot following the config file in those specific sections

when i say server, i mean discord server

[obsidian patio]

Should I setup everything and then update the IDs in the config file to match the server info?

[sullen phoenix]

yes, exactly

it is a bit tedious

[obsidian patio]

So, I should create every channel and add every emoji manually?

[sullen phoenix]

you don’t have to add the emojis

as for the channels, you’ll need to add the ones in the config

[obsidian patio]

It won't really be complete without emojis ahhah

But yeah, thanks again

[sullen phoenix]

no problem 👍

[obsidian patio]

This feels like a silly thing to ask, but how do I add the emojis?

I can't seem to find any image files in the repo

[sullen phoenix]

erm, most of the emojis aren’t public

[obsidian patio]

Is there a way to add them?

[sullen phoenix]

there is not

[obsidian patio]

Okay

[sullen phoenix]

they should work anyway

except maybe for the trashcan

[obsidian patio]

Oh, okay. What was the part about adding emojis about?

Why not the trash hahah?

[sullen phoenix]

because @stable mountain reacts using the trashcan

the other custom emojis are in embeds afaik

[obsidian patio]

Okay

[shut aspen]

so im looking into using snekbox for my own bots evaluation, does the container restart on every eval?

or is nsjail the primary sandboxing tool

[green oriole]

It is nsjail

[crude gyro]

@velvet ether hey dude, thanks so much for the work you've been doing on django-simple-bulma, I really appreciate it.

are you keen to take on the rest of the submodule tasks? I guess maybe it's only extensions. I think the approach you suggested sounds good.

I've also been thinking about how we can do the same sort of thing for FontAwesome - it'd be really cool if it used the latest version of this as well, but I'm a little unsure how we could achieve that.

still, "always use the latest version of everything" is a very cool feature.

[velvet ether]

No problem :) I was working on a django project and needed a stylesheet that wasn't bootstrap, so i figured i might as well help out

I looked around for a fontawesome cdn that just provided the latest version, but couldn't find any. In my project i just ended up serving it myself

[crude gyro]

yeah, that's the problem.

we'd need some cleverness to solve that, I think

the latest versions of fontawesome expect you to make a user and use a custom CDN link for only you

[velvet ether]

if we make people add that to bulma_settings it would basically be the same as adding a <link> manually in some template

[crude gyro]

the right approach might be to have the user put their info in BULMA_SETTINGS

yeah

I mean, you're right

the template tag makes things a little more readable though, maybe.

but I agree that the feature is a little dumb

I wonder how much info they'd have to put into BULMA_SETTINGS for us to generate the CDN link

and if we can support Pro as well

there may be a little bit of convenience there, if they just need to put a token and we can support both Pro and Regular

[velvet ether]

i guess

[crude gyro]

alternatively we could remove the feature, but that's not backwards compatible and so it might be a hassle for existing users.

maybe we could make the token optional, and just do what we're currently doing (provide an older version) if no token is provided

[velvet ether]

i can generate an api token on their website, but can't quite figure out how to use it

[crude gyro]

so they have an API?

hm.

iunno, it just feels like they are going in a different direction

they want everyone signing up

and getting in the way of that is kinda stupid

maybe we should just remove the support.

[velvet ether]

don't get me wrong, font awesome is awesome, but damn is it fragmented. There seems to be a different website for every release and then some stragglers like github pages and some cdns. Also i can find no documentation for their api 🙄

[crude gyro]

yeah, agreed.

[velvet ether]

and yeah, the font_awesome tag really isn't that usefulanymore

[crude gyro]

yeah maybe not.

I'll give it a think

[molten bough]

font-awesome expects you to do one of three things

1. Use their CDN and kits system (the user-specific links you mentioned)

2. Download it and serve it statically from the site itself, but you can't put the pro icons in a git repo

3. Use your credentials to get everything from an NPM repo at build time

that last one is probably your best bet for automation

but the issue then becomes, do you make django-simple-bulma/whatever django addon rely on npm or implement some custom download logic?

the API keys exist to allow one to configure kits programmatically btw, the wordpress plugin uses them for that

[green oriole]

You could pull them in your setup.py file perhaps?

[molten bough]

I mean it's certainly possible to configure a kit that way

in the settings.py

that's the absolute easiest option, but it doesn't solve the issue of people that want to serve the assets themselves

the API is documented, it's graphql, but it doesn't seem useful other than that

[crude gyro]

I'm leaning towards

• Allow users to specify their kit token or whatever in BULMA_SETTINGS, and use that if provided.
• Otherwise, use an older CDN like we're doing now.

[molten bough]

Yeah, I think probably a whole download system is probably out of scope

[crude gyro]

I would be leaning towards just removing the feature, but I don't wanna break backwards compat

[molten bough]

kits are going to be the most common use-case

[crude gyro]

but how much info would we need in the BULMA_SETTINGS to generate a kit CDN link?

is it just a single token?

and is it the same for pro and regular?

[molten bough]

There are no free kits

you just use their CDN

the point of kits is that you can customize them in the font-awesome site's ui

[crude gyro]

okay, you're getting hung up on terminology

the CDN link for the free stuff still has a token in it

at least in recent versions

and you have to like sign up with your email and become a free member to get that token, right?

[molten bough]

hmm

yeah, it does seem that way

[crude gyro]

what I'd like to do here is

• put as little information as possible into BULMA_SETTINGS - maybe a single token?
• generate either a Pro or a Regular CDN link using this token
• If token is not provided, just put an older version CDN like we're doing now

[molten bough]

alright, I just made a free kit and yeah it's the same

<script src="https://kit.fontawesome.com/IDENTIFIER.js" crossorigin="anonymous"></script>

[crude gyro]

hm, okay. so basically we just need that identifier

[molten bough]

that's all there is to it

[crude gyro]

that seems pretty easy to implement then

[molten bough]

since this is a free test kit, here's what the site shows

you just get that script tag

the settings are just things like, the version, whether to use web fonts/SVG, and a handful of misc options

they don't change the url

[crude gyro]

very cool

that suits us perfectly then

@velvet ether Hey dude, are you around?

just wanted to say, I've assigned you a Contributor role here on the server as thanks for the work you've put in. Enjoy the fancy green color. Also, if you'd like to handle this FontAwesome thing (see the conversation above) and the Bulma-Extensions submodulification, they're all yours. We'd be grateful for those changes. I think that would really modernize the whole app and make it more maintainable.

If not, I can probably handle them myself.

[molten bough]

There'll be an official django plugin when FA 6 rolls around as well, so no need to overengineer

[crude gyro]

huh?

what do you mean

are FA making one?

[molten bough]

yup

[crude gyro]

huh.

[molten bough]

so you won't need to worry about it after that

there's almost no info on it right now unfortunately, but

[crude gyro]

okay, cool.

guess we'll add some info about that in the readme at some point

when it comes out

[molten bough]

I already obtained a pro license for it so you guys will be able to use it as soon as the beta rolls around, if that's your prerogative

I'll let you know

[crude gyro]

cool.

thanks for that

[molten bough]

👍

oh right, I got an email the other day as well

apparently they're releasing a feature that will let you add your own icons to kits if you wanted a pydis icon or anything

should be.. "soon™️" ("very shortly")

[crude gyro]

well, I mean, it sorta already has the pydis icon

[molten bough]

Oh yeah, they do have the python logo don't they

[crude gyro]

yep

[velvet ether]

just wanted to say, I've assigned you a Contributor role here on the server as thanks for the work you've put in. Enjoy the fancy green color. Also, if you'd like to handle this FontAwesome thing (see the conversation above) and the Bulma-Extensions submodulification, they're all yours. We'd be grateful for those changes. I think that would really modernize the whole app and make it more maintainable.

If not, I can probably handle them myself.
@crude gyro Thank you, I appreciate it. I'm not at home right now, but if be happy to look at it later today.

[crude gyro]

there's no rush, @velvet ether, I'd be happy even if you looked at it a week from now.

I'm just looking for a clarification on whether I should put it on my own agenda or leave it to you.

I don't think these changes are urgent, but I'd love to see them at some point.

[velvet ether]

I know, but I'm gonna work on it later today anyway

[crude gyro]

that's fine of course

[crude gyro]

@brazen charm hmmm

what are you up to?

working on something fun?

[brazen charm]

No, just adding one validator and setting up a normal environment instead of a static clone of the repo

[brazen charm]

Any way I could run the site tests through docker? Want to avoid messing with postgres on the host machine

[glass pecan]

should be able to by specifying a different command on launching the container

are you using compose?

[brazen charm]

Yes

[glass pecan]

docker-compose run web
is the same as if you'd do
docker-compose run web run --debug
which is eq to if you'd done
python manage.py run --debug if you were locally running it.

i'm like 80% sure on this anyhow lol

using that baseline, you're able to pass anything you want to manage.py instead of run

with django, to run tests, you'd normally do manage.py test

so i assume it's docker-compose run web test to get it to run tests

let me know if that's wrong though, in which case i should be able to dig up how i set it up previously

[brazen charm]

Yeah I was doing something similiar before but at first it seemed to not apply migrations at first and then it wasn't started in debug. This is a lot simpler and only needs the debug env var added, thanks

[crude gyro]

I just run the tests in PyCharm with the docker-compose for the site running

required minimal setup, and no local postgres

[brazen charm]

Pycharm didn't seem to like them much when I tried through it, how does your run config look?

[crude gyro]

well, I changed it so that the template for Django tests contains some environment variables

the database URL I'm using is DATABASE_URL=postgres://pysite:pysite@localhost:7777/pysite

other than that, nothing spceial

then I basically just right click the tests folder and run

and it works like a charm

[brazen charm]

Complains about something with the db config when I try that

[brazen charm]

For migration names, is it just just ####_change ?

[crude gyro]

yes.

[patent pivot]

lol why does the github bot send the introduction twice

guess it is treating ready to review as PR opened

oh

  def handle_cast({:pull_request, :ready_for_review, payload}, state) do
    handle_cast({:pull_request, :opened, payload}, state)
  end

yeah okay

fixed

[cold moon]

Does GitHub bot collect some statistics too?

[patent pivot]

no stats on the github bot, no

it's completely stateless

It could collect stats, but most of what it is sent is collectable by the GitHub API as well

[cold moon]

Why this last revert?

[brazen charm]

<@&267629731250176001>

Help channels broke with it

[woeful thorn]

!tempmute 731475449162760214 1H shh

[stable mountain]

:incoming_envelope: :ok_hand: applied mute to @clever wraith until 2020-08-05 16:43 (59 minutes and 59 seconds).

[cold moon]

Oh

[tough imp]

There was a bug that I didn't catch in testing

it was causing channels to not move out of available sometimes when spoken in

looking into it now

[tawdry vapor]

I think I wanna do some work on this cog. Clean it up a bit and investigate the 1 available channel bug

Would appreciate getting these merged first https://github.com/python-discord/bot/pulls?q=is%3Apr+is%3Aopen+label%3A"a%3A+help+channels"

Ah also 886 since it moves out some funcs to utils, but whatever, I'll take what i can get

[obsidian patio]

When the code jam ends, I should be free to work on #dev-log message. Shouldn’t take too long

[tawdry vapor]

Review please https://github.com/python-discord/bot/pull/886

Like I said earlier, It'd be nice to have the PRs that touch help channels out of the way so I can do some refactoring there

[brazen charm]

Got a WIP but working solution for https://github.com/python-discord/bot/pull/1014#issuecomment-663992753 at https://github.com/Numerlor/bot/pull/1/files, but I'm unsure on the design because of how the objects are linked; would be great if someone could look at the comments I've made on the PR in cog.py( and https://github.com/python-discord/bot/issues/1067) so I can get them commited and then I'll try looking at using lxml directly for some performance gains

[velvet ether]

Hey @crude gyro, did you see I my pr on django-simple-bulma? I got lots of other stuff to do, so it's no rush, but I still just wanted to make sure it didn't get lost in your feed :)

[molten bough]

I think he's taking a break at the moment, at least from what I heard in the OT channels

[crude gyro]

I was on a bit of a vacation, yeah.

but I'm back now

thanks for the reminder, @velvet ether, my apologies, it got lost in my feed

coming back to work after a vacation turns out to be quite busy

[green oriole]

Vac.. ations ? Like the owners of this server sometimes sleep?

[crude gyro]

just took a vacation from working on Python Discord stuff and from my dayjob, and focused on a personal project for a bit.

even on my vacation, I still can't help but spend most days programming

https://lemonsaur.us

[tawdry vapor]

I wish our API had more specific error codes

Cause doing this is annoying py if e.status == 400 and "already exists" in next(e.response_json.get("id", []), ""):

It gets even worse. I have to do iter() before next()

[hardy gorge]

That does sound like a hassle

[tawdry vapor]

It may not be necessary, but IDK what I can and can't rely on from DRF's default error handler

[hardy gorge]

I'm trying to think of a neat way to solve this.

Some APIs return a unique error code in the JSON

[tawdry vapor]

I can abstract it into a custom exception that subclasses ResponseCodeError for now

Yes, Discord's API does that

[hardy gorge]

If I'm not mistaken, most of our current response will be {"field_name": "some reason in a long string"}

[tawdry vapor]

https://www.django-rest-framework.org/api-guide/exceptions/#exception-handling-in-rest-framework-views

It doesn't explicitly state it will always have a list as the value for validation errors

[hardy gorge]

"Any example validation error might look like this:" (emphasis mine)

Yeah

It's basically "it could be this, but, hey, no guarantees, and good luck"

Anyway, the custom exception part beneath that is indeed interesting

[tawdry vapor]

It'll probably be in a list so I'll simplify the code https://github.com/encode/django-rest-framework/blob/3d708ac7005e29ca75aa68b447edb8d9dc111637/rest_framework/exceptions.py#L152

[tawdry vapor]

What do you all think about my idea to prefix non-extension modules with an underscore? https://github.com/python-discord/bot/issues/160

This is so they can be distinguished from extensions by the !ext reload command and its friends

[hardy gorge]

Doesn't sound like a bad idea to me

[tawdry vapor]

Well, I suppose that's enough for me to move forward

People are welcome to complain in the PR if they dislike that

[thick rock]

the bot has complained about code block escape attempt twice to me

when I use !e

what's that about

[tawdry vapor]

It doesn't allow triple backticks or triple zero-width characters

I don't really know why the zero-width characters are disallowed but the backticks would obviously end the Markdown codeblock prematurely

[thick rock]

I don't get it

so don;t have anything after closing ```?

how can I avoid triple backticks

both times I pasted the same thing again, and it accepted it, at least I think I didn;t change anything

[tawdry vapor]

If your input is in a code block and you put any text before or after it, the command will not be able to extract the code from the code block and will therefore treat the entire message as the code.

[green oriole]

Why aren't we getting rid of the text before and after the codeblock btw?

It would allow us to place explanations before and after

And it is just a small regex change

[tawdry vapor]

No reason

[green oriole]

Time to open an issue I guess then

Will do that tomorrow

[little pecan]

need some help. Anybody Willing?

[exotic ember]

Is it related to our projects on github? If not, #❓｜how-to-get-help would be a better place.

[neon phoenix]

I'd like to contribute to the bot but I don't know how to set up the bot and the site. Is there a guide to get started?

I'm using Manjaro XFCE (a Linux distrubution)

[exotic ember]

https://pythondiscord.com/pages/contributing/

feel free to ask here if you get stuck at some point

[neon phoenix]

in that page it only says what are contributors and what's the contributors role

[exotic ember]

The Bot and Site headers are clickable
https://pythondiscord.com/pages/contributing/bot/
https://pythondiscord.com/pages/contributing/site/

[neon phoenix]

oh

well, thanks

i though it was going to lead me to github

[exotic ember]

that probably says something about the UX though

the link that says Github leads you to github

[neon phoenix]

i should start setting up the site, right?

[green oriole]

Yup, you'll need the site for the bot anyway

[neon phoenix]

okay

git clone go brrr

damn i'm so excited to contribute

hm, about the test server

there is a new feature called "server templates"

they're used to create a clone of the server, but they do not contain the members and messages

if anyone has already a test server they could share it here so people could set up their own one in just one click

[neon phoenix]

how do i set up the big brother and the talent pool webhooks?

also, i'm getting this error because i'm not going to connect to pythondiscord.com for obvious reasons

[molten bough]

You need to also run the site locally

[neon phoenix]

i did

it's running on 0.0.0.0:8000

[molten bough]

No, that's the listen address, you connect to it at 127.0.0.1

But you need to add a hosts file entry really

Like api.localhost

[neon phoenix]

hm

[molten bough]

Or you could just configure it to not use a subdomain

[neon phoenix]

i think this is okay now

[molten bough]

No, that api entry needs changing

So it's just *DOMAIN

[neon phoenix]

i see
i got 3 pings lol

[molten bough]

Rather than joining that list

[neon phoenix]

okay, so, what do i do??

[molten bough]

I'm not sure I follow - did you do what I said already?

[neon phoenix]

i didn't really understand you

i don't really know yaml

i think it doesn't work because api.0.0.0.0:8000 or api.127.0.0.1:8000 are not valid

[molten bough]

Well, yes

look at the site_api line

&API is giving that entry a name, which can be referred to using *API later in the document

[neon phoenix]

i see

[molten bough]

after that, we're joining api. and the entry labelled DOMAIN together

[neon phoenix]

okay

[molten bough]

when really all you want there is *DOMAIN

[neon phoenix]

oh

so

[molten bough]

however

now that I think about it

I think the django setup does require a subdomain

so, the DOMAIN should be set to localhost:8080

you can test whether that works, but if it doesn't, you need to edit your hosts file and add api.localhost

the hosts file is OS-specific and not part of the project

[neon phoenix]

oh sheet

[molten bough]

https://pythondiscord.com/pages/contributing/hosts-file/

[neon phoenix]

hm, but

im not using docker

[molten bough]

that doesn't really matter

you can follow this guide, it defaults to non-docker instructions and tells you how to modify them if you're using docker

then you'll change the DOMAIN entry in your config to pythondiscord.local:8000

[neon phoenix]

okay

[molten bough]

ensure that you've followed the rest of this too https://pythondiscord.com/pages/contributing/site/

[neon phoenix]

done

now why am i getting this?

[molten bough]

pretty much exactly what it says

[neon phoenix]

also, thanks :D

[molten bough]

you haven't synced yet, so

you'll need to confirm that

[neon phoenix]

now everything works :D

[molten bough]

No worries

[neon phoenix]

i managed to mute myself for 1 month

thanks for helping me @molten bough :D

[molten bough]

All good

[neon phoenix]

by the way i'd like to share my python discord server template :) https://discord.new/DtaKJqa7K2BA

note: verified role is developers role

[patent pivot]

where is the DUCK POND

[timid sentinel]

There was already this one https://discordapp.com/channels/267624335836053506/635950537262759947/694868859865530418 although I reckon it's quite out of date now

This one https://discord.new/zmHtscpYN9E3

[neon phoenix]

where is the DUCK POND
o shit

well

is it really necessary?

[patent pivot]

not really lol

[neon phoenix]

well can we pin one of the two templates?

because they can be useful

[neon phoenix]

btw, about this issue: https://github.com/python-discord/bot/issues/1047
is there anyone working on it? i could do it if there isn't people working on it

[tawdry vapor]

Ask Joseph, he wrote he was interested in implementing them

[neon phoenix]

@patent pivot read above please

[patent pivot]

oh yeah, I have some stuff locally for that

so assigned to me, I'll PR it either today or tomorrow

[neon phoenix]

okay

[neon phoenix]

how do i add docs so the docs command will work?

[brazen charm]

You need the inventories of the individual docs pages e.g. http://docs.python.org/objects.inv and then add that with the base url through the set command or the admin panel on the site

[neon phoenix]

hm, something's not right

[brazen charm]

!d

[stable mountain]

All inventories (`18` total)

• aiohttp
• arcade
• dateutil
• discordpy
• django
• flask
• kivy

Page 1/3

[neon phoenix]

:o

[brazen charm]

The base url for the package is here, and you need to grab discordpy's inventory, python.org's only has symbols that are on docs.python.org

[neon phoenix]

how would i do that?

[brazen charm]

the inventories are almost always behind the base url + /objects.inv

[neon phoenix]

huh, it seems that there's nothing on https://pypi.python.org/pypi/discord.py/objects.inv

maybe i'm not getting something or something's weird

[brazen charm]

The base url is what's hyperlinked in the inventories embed above

[neon phoenix]

oh

https://discordpy.readthedocs.io/objects.inv is a valid one?

[brazen charm]

It's then used to construct the absolute urls when you fetch a doc symbol so it's the part that's shared between all of them in the package

You're missing the en/stable/ path

[neon phoenix]

yay now there is one :D

btw, i finally made it, so i can finally test my changes for https://github.com/python-discord/bot/issues/1098

(can anyone assign me?)

[brazen charm]

You should wait for it to be discussed and approved, personally I don't think it's very impactful of a change but also don't see much use in deleting the doc embeds

[neon phoenix]

fair enough

i'll wait

i'll not pr until it's approved

[brazen charm]

(You can start that dicussion by weighing in to potentially speed it up)

[neon phoenix]

wdym?

[mellow hare]

As in like... adding any insights or comments you have about that particular issue

Something to help clarify why this particular issue should be carried out

[brisk tinsel]

guys

[crude gyro]

yes

[neon phoenix]

yes

[neon phoenix]

i have a question about git

i got assigned for the issue #1098 and i commited, but, how do i push?

is git push enough?

sorry, but this is my first "big" contribution and i don't want things break

btw i linted beforehand :)

[carmine sluice]

dev

local.clietn

local.client

[green oriole]

git push should be enough yeah, and it will probably ask you to run a longer command to set the upstream, you can do what gut says

Github will probably even respond with a link to directly open a PR

(all the lines starting with remote: are send by github)

[neon phoenix]

also, for some reason the lint is successful but the azure pipelines lint is not successful

bruh

[green oriole]

That's weird

[neon phoenix]

guess i'll need to commit again with some fixes to satisfy azure

[opal pebble]

is there an easy way I could copy the !doc command to my own bot as a util?

[sullen phoenix]

sure, have you seen the source?

you would just need to implement the database persistence on your bot instead of on a separate site (unless you have a separate site)

[opal pebble]

I have the source open atm

what's database persistence?

[brazen charm]

the bot requests inventory info from the site api for the bot here, but it's trivial to do that from a local source or an another way

[opal pebble]

and what do you mean by a local source?

I haven't been doing programming like this for very long, so even though my skills have been steadily increasing over the last few months since I started making a bot, I have yet to learn about things and terms like this

[neon phoenix]

local source -> something saved locally, for example a file

and, be aware that @stable mountain (bot) highly relies on https://pythondiscord.com/ (site)

Infractions, emails, tokens, docs... everything is saved there, in a database

(with emails and tokens I don't mean your discord ones, with emails I mean your github email that is got when you link your github there)

[opal pebble]

okay

and how should I implement the database persistence on my bot?

[tawdry vapor]

Saving the inventories to a file or to a local database

You'd replace the calls to the API with some file reading/writing operations, or some SQL if you're using a database.

[crude gyro]

hahaha @velvet ether, Remove blasphemous whitespace is a pretty good commit message.

[velvet ether]

Thank you @crude gyro 😁 . Also, I'll try your suggestion about passing the element, give me a couple of minutes

[crude gyro]

@celest charm is our JS guru. could you have a look at https://github.com/python-discord/django-simple-bulma/pull/59/files, @celest charm, and see if you can suggest a way that we can do this without this horrible scope bleeding stuff?

it's so ugly to have the show and hide functions defined inside the loop and then have the other function, defined earlier, make use of them

it just.. doesn't feel very Pythonic

[celest charm]

I think IIFEs can be replaced by simple scopes?

😄

oh, huh

no, functions bleed out

what a beautiful language

[crude gyro]

I know, but can't we solve this in a cleaner way? In my dream solution, we'd define all those functions outside the loop and then somehow still make it work on the right elements.

[celest charm]

Maybe you can use .bind

Oh, wait. You can just define the functions outside of the loop, right?

and use this

(disclaimer: I am not really a JS guru 👀)

[velvet ether]

What do you think about this? (untested) https://hastebin.com/feneqabuko.js

alright, if you pass the menu_element to the wrapper, then this also works

[celest charm]

Wait. How are you using show and hide in handle_event if they are defined outside of it?

[crude gyro]

scope bleed. that's the thing that sucks.

I think this is better, @velvet ether

what you showed there

if it works

[velvet ether]

it does

[crude gyro]

then I'd be happy with that

[velvet ether]

Wait. How are you using show and hide in handle_event if they are defined outside of it?
i have no clue

[crude gyro]

gotta run now but I'll have a look tomorrow and probably do a 2.0 release

[celest charm]

But handle_event will refer to the last defined function, right? Have you tested it with multiple elements?

[velvet ether]

as i also explained on the pr, i did not inted for this to happen

[crude gyro]

thank you so much for all the excellent work you've put towards 2.0

[velvet ether]

np

But handle_event will refer to the last defined function, right? Have you tested it with multiple elements?
yikes you are right

[celest charm]

I think it works exactly as it does in Python

If you don't want the functions to bleed out and you don't need to use/pass this inside, you can create an arrow function and bind it to a variable

oh, wait, you can bind normal functions to variables.

So instead of

function show() {
    element.classList.add("is-active");
}

do

const show = function() {
    element.classList.add("is-active");
}

or, if you don't need this inside the handler,

const show = () => {
    element.classList.add("is-active");
}

and then you don't need an IIFE

[crude gyro]

can you explain IIFE?

[celest charm]

Immediately-invoked function expression

[crude gyro]

ah yeah

[celest charm]

so yes, probably the best way is to create a higher-order function outside the loop

as JonasUJ suggested

[velvet ether]

but it's not working rn 😳

[celest charm]

hold on

why do you need menu_element?

You're attaching the event handler to element, not menu_element

[velvet ether]

mistake

I tried to clarify that in the msg right after

[celest charm]

oh, sorry

I somehow missed it

why are you removing the event listener after the element has been clicked?

[velvet ether]

i dont know actually

is just converted the original script away from using jquery

it works fine without so i guess there is no reason for it

no sorry, that is a wrong way to think about it

also, menu_element was actually just some leftover

I got it down to this now https://hastebin.com/genibebame.js. It work as far as i can tell, but I also said that the previous couple of times

[celest charm]

maybe move show to the named function and pass in the element instead?

[neon phoenix]

One question about bot/cogs/doc.py: is the doc_cleanup function I made OK or it's preferably to use the cleanup function from bot/utils/messages.py?

[cold moon]

@neon phoenix use premade cleanup function

[neon phoenix]

okay

new commit today afternoon if I can

for the pr #1106

BTW, the help cog does the same :p

It doesn't use the pre-made cleanup function

[short snow]

How does the docs work for the bot? do you write it or it is web-scraped from some place?

[molten bough]

The docs command?

[short snow]

yes

[molten bough]

I think it uses sphinx's indexing files from documentation sites generated with it

[short snow]

oh!

[brazen charm]

It uses those to get urls pointing to what you can request, then those are scraped for relevant contents and turned into markdown

[neon phoenix]

One question: how actually the Azure Pipelines test the bot?

i have no idea how a program can test a bot lol

[cold moon]

@neon phoenix This can be tested with Mocking

[neon phoenix]

mocking????

well, i did it

i used wait_for_deletion from /bot/cogs/messages.py instead of using the original doc_cleanup

now it's time to wait for approval :|

[neon phoenix]

@cold moon one thing about the comment you made

you also said to not specify client

wouldn't it break? in the wait_for_deletion's sources, the client parameter is: client: Optional[Client] = None

[cold moon]

https://github.com/python-discord/bot/blob/master/bot/utils/messages.py#L39 this is optional parameter.

[neon phoenix]

oh, i didn't see that

it seems i need to read all the bot's source code if i want to write cleaner code, at this point

uh, wait a moment

[cold moon]

This is in docstring too

[neon phoenix]

if i don't set the client parameter to self.bot, this exception AttributeError: 'Member' object has no attribute 'wait_for' is raised

because discord.Member doesn't have wait_for, but discord.Client/discord.ext.commands.Bot/bot.Bot does

hm, i think i'll open an issue

[cold moon]

Hmm

Yep

[neon phoenix]

@cold moon so, finally, what do you think it's cleaner?

await wait_for_deletion(await ctx.send(embed=doc_embed), (ctx.author.id,), client=self.bot)```
or
```py
doc_embed = await ctx.send(embed=doc_embed)
await wait_for_deletion(doc_embed, (ctx.author.id,), client=self.bot)```
IMO the first one is cleaner because there are no function calls inside the function call

¯_(ツ)_/¯

[thorny fossil]

hey soo im kinda new here. I took an online basic python class and i would like to move forward with learning python. Where or how would i do that?

[neon phoenix]

!resources @thorny fossil you can check this out

[stable mountain]

Resources

The Resources page on our website contains a list of hand-selected learning resources that we regularly recommend to both beginners and experts.

https://pythondiscord.com/pages/resources

[neon phoenix]

and... this is not the right channel

please read the topic

[crude gyro]

@velvet ether did you see the bug posted on django-simple-bulma today? https://github.com/python-discord/django-simple-bulma/issues/61

seems like the new release is pretty much unusable

[velvet ether]

wah

then i must have misunderstood how the MANIFEST.in works

[crude gyro]

the only thing I can think of is that sdist just doesn't have the submodules available to look inside of, it just sees, you know, a .git or whatever

but I don't know. it's a little hard to troubleshoot. I thought the release would've failed if something like that was true.

did you try to build it locally?

okay, yeah, the build does have some warnings but didn't fail

warning: no directories found matching 'django_simple_bulma/bulma/sass'
warning: no directories found matching 'django_simple_bulma/extensions/*/src'
no previously-included directories found matching 'django_simple_bulma/extensions/*/src/docs'
no previously-included directories found matching 'django_simple_bulma/extensions/*/src/demo'
warning: no directories found matching 'django_simple_bulma/extensions/bulma-coolcheckboxes/dist'
warning: no files found matching 'django_simple_bulma/extensions/bulma-megamenu/megamenu.css'

[velvet ether]

running python setup.py sdist bdist_wheel in my local repo gets me a build and dist folder. In the build folder i can see all the bulma files

did you run git submodule update?

[crude gyro]

probably not

[velvet ether]

it is required to pull them

[crude gyro]

might need to be added to the release workflow then.

https://github.com/python-discord/django-simple-bulma/blob/master/.github/workflows/release.yml

here's the current workflow

basically all it does is check out the master branch, install python, install dependencies, and build

so I guess we need a step to do that submodule update then

[velvet ether]

so like this? yaml - name: Install build dependencies & build run: | git submodule update python -m pip install --upgrade pip pip install setuptools wheel python setup.py sdist bdist_wheel

[crude gyro]

maybe, but I'm not sure git will be available at that stage.

so we miiight have to install it first

maybe we can just add this

https://github.com/marketplace/actions/checkout-submodules

then we can just add this on the step before the build

steps:
- name: Checkout submodules
  uses: textbook/git-checkout-submodule-action@master

[velvet ether]

I'm unfamiliar with github workflows so I wouldn't know

[crude gyro]

using marketplace actions is nice. someone else can go through the trouble of maintaining it.

anyway I can try to add this step

attempting a new release now.

[velvet ether]

it ain't there

[crude gyro]

the action failed

[velvet ether]

oh

[crude gyro]

hm.

[velvet ether]

is it trying to run the git command in .githib/workflows dir instead of the base dir?

[crude gyro]

hm, I found something

ah yes, okay

we should be able to use this for it https://github.com/actions/checkout

which, actually, we were already using. but submodules need to explicitly be set to True

[neon phoenix]

Sorry for interrupting you guys, but, do you think a repository map will be a good idea on the site and bot repos?

[tawdry vapor]

recursive is probably a better value, just to be safe

Unless it would never make sense for one of your submodules to have a submodule it depends on for normal functionality

@neon phoenix What's that?

[crude gyro]

yeah I'm setting it to recursive

[neon phoenix]

a markdown file, which inside there all folders and files are explained

[tawdry vapor]

We should probably do a better job of adding module docstrings rather than relying on a separate file then.

I suppose not every folder is a package though

[neon phoenix]

Yes, but, there's a lot of files

[tawdry vapor]

I dunno, I don't think it's necessary.

Their names should be self-documenting

[neon phoenix]

The repo map could help contributors that are beginners, so they don't need to really dive in
idk

Tbh I didn't know bot/utils/messages.py was a thing

[velvet ether]

as someone who haven't contributed to either, I would say the site is at least decently understandable

[crude gyro]

@velvet ether I think it was resolved with that last release, although I'm still getting this warning in the build log:

warning: no files found matching 'django_simple_bulma/extensions/bulma-megamenu/megamenu.css'

not sure why, but I don't think it has anything to do with the release workflow at least

[velvet ether]

i know

it's an scss file

🤦‍♂️

[crude gyro]

oh, lol.

so we just need to add an s in the MANIFEST file?

[velvet ether]

that should do it

[crude gyro]

okay, I'll do a quick release for that too

[velvet ether]

sorry about all this... It's quite a bummer of a v2 release

[crude gyro]

naw, it's fine. he was probably the only person who noticed :D

[hardy gorge]

Haha

I just DM'ed you the same thing about the s

[crude gyro]

man

[hardy gorge]

I guess that errors are going to be made

Version numbers are free, man!

[crude gyro]

yeah it's fine. it's natural even for quite popular packages to have little bugfix releases on the same day as a new major release

[velvet ether]

little bugfix releases

[crude gyro]

granted, they're usually not as serious as "this package is now unusable"

[hardy gorge]

I guess one way to fix these build issues is to have a CI that uses the test PyPI which you can test before tagging an official release

[crude gyro]

but the megamenu thing is a small bugfix

[hardy gorge]

Haha

[crude gyro]

I honestly don't think it's that major of a problem if it's broken for a few hours. it's nice that it has enough users that someone will report a bug

[velvet ether]

this is what you get when you don't specify a version in your requirements.txt xd

[crude gyro]

the only people who would've been affected by this are people who chose today as their day to update

[hardy gorge]

Nah, but if you want to try a fix to the CI, having an actions file that pushes to the that test PyPI repository that only keeps your package around for a short while may work. It allows you to test the entire thing without actually pushing to PyPI proper.

[velvet ether]

seems to work now

[crude gyro]

nice.

[green oriole]

Was that an issue that running a local pip install (pip install <path to the source>) wouldn't catch?

[hardy gorge]

Basically, yes

The issue was that we needed to include a submodule during the build step

So, the files need to be there in the CI

And if you do have the files locally (which is probably because you need it to run/dev the package), it will look like it works locally

but then fail in the CI

[green oriole]

I've just tried to local install the package without installing anything, using the v2.0.0 tag, and it seems to build without any warning, that's really weird

[crude gyro]

there are warnings in the build log

they just don't stop the build

I would've expected the build to fail

that's basically why this happened

but it's fixed now so it's all good

[brazen charm]

@green mesa inv files are fetched through the intersphinx fetch_inventory function, they have their own format

[green mesa]

Ooo how can I start

I don't know much about this topic¯_(ツ)_/¯

Can u link me some useful link😅 easy one which I can understand

@brazen charm

[brazen charm]

There isn't much to it. You just install sphinx and use the sphinx.ext.intersphinx.fetch_inventory function

[green mesa]

how will I know what all it takes?

[brazen charm]

you can look at its help() output or find it in the docs

[green mesa]

Not there

It's not showing at docs

😐

it says fetch_inventory did not match any query

[brazen charm]

You can find the function and its docstring here https://github.com/sphinx-doc/sphinx/blob/3.x/sphinx/ext/intersphinx.py#L152. I've also just uploaded an async implementation for it at https://gist.github.com/Numerlor/e57096cedd06dd536cdddff86be2661f, but that's adjusted for the behaviour that's needed for @stable mountain

You can find the usage in @stable mountain's source, but you need the mock app (you can pass the uri to get the full links as output) and inv is the link to the inventory

[green mesa]

ok ty i will look into this

😄

[green mesa]

    config=SimpleNamespace(
        intersphinx_timeout=3,
        tls_verify=True,
        user_agent="python3:python-discord/bot:1.0.0"```
this is mandatory

[green mesa]

examples and explanation is not given

[brazen charm]

The inventory files only get you the link to the symbol, you have to parse it yourself from the site

[green mesa]

ooooo

mean now bs4 comes into play

[tawdry vapor]

writings tests sure does take a long time, or I suck

Also spent all morning fixing the code in that bottom commit, but rebased so it doesn't show.

[green oriole]

It sure does

[tawdry vapor]

Apparently 22% of commits on bot's master are merge commits

[brazen charm]

That's a lot more than I'd expect

[crude gyro]

probably a good thing.

means we're good at keeping PRs small, basically.

it also probably means we're a bit too eager when it comes to clicking on Update Branch

[green mesa]

Hello
!ask command is not working 🙂

I tried thrice !ask but py bot is not responding other commands are working except !ask

[tawdry vapor]

It's been removed

I think it was announced

[green mesa]

Ahh y that was very useful😐

Can we some command which explains what helpers need when someone ask for help

I think it was announced
Sorry but I aren't able to find any announcement for that🙂

[tawdry vapor]

You can find the discussion in #community-meta then from a few days ago

[green mesa]

Ok ty mark🙂

[crude gyro]

https://github.com/python-discord/bot/issues/1114

very low hanging fruit issue here if someone wants an easy contribution.

[velvet ether]

@crude gyro About issue #6 https://github.com/python-discord/django-simple-bulma/issues/6

I agree that it should be kept at very low priority, but since it is a feature of Bulma (https://bulma.io/documentation/overview/modular/), I think we should at least keep it open.

[crude gyro]

hm, interesting.

didn't realize.

I agree then.

[patent pivot]

Hey Contribs, could anyone take this? https://github.com/python-discord/bot/issues/1115

quite urgent

[brazen charm]

Can take a look at it in a bit

[green mesa]

i wnt to particapte in hacktoberfest of pydis

how i need to participate 😄

i can work on issue with WIP tags but that issue also have approved tag

so its have been approved?

[subtle kraken]

WIP on an issue most likely means that someone is already working on it

[green mesa]

oo so on which issue i can work

[patent pivot]

If you don't see any activity in it though ask a core developer and we could reassign it

Otherwise the approved ones

[green mesa]

one more question i am using python 3.7.7 will i need to download 3.8 or i can work with 3.7.7

[brazen charm]

For the bot, you should use 3.8 as that is its specified environment

[green mesa]

ok

ty Numerlor

[green oriole]

Do we have a little guide somewhere about how to setup the SocialApp on a local website?

I'm seeing 3 fields in the social accounts tab of the admin panel

I somehow figured it out

.>

How do you store if a user has accepted the privacy police? I have a user entry with my id an the "In guild" checkbox checked

[crude gyro]

this just means you're not in the User table at all

so, if you're on the server you can usually just sync the users to solve it

although I'm not sure how well OAuth works locally?

[brazen charm]

For the reminders issue, should I expand the api to allow some fetching of reminders by id, or just figure out the result from the available author filter

[green oriole]

I think my bot is currently broken, and I'm a bit too lazy tbh

[crude gyro]

For the reminders issue, should I expand the api to allow some fetching of reminders by id, or just figure out the result from the available author filter
@brazen charm

you can't fetch reminders by ID already?

[green oriole]

But I am in the User table

You can delete them by id, but not fetch them by id iirc

[brazen charm]

filtering is restricted to active and author ids, and get is disallowed on the direct reminders/id path

[green mesa]

what configuration should i have with flake8 before PRing

[green oriole]

There's a configuration already in the repo

[green mesa]

ok

[green oriole]

pipenv run lint should do the trick

[green mesa]

wow ty

[green oriole]

So all auth now work, but I can't access the setting page

Access to fetch at 'http://pythondiscord.local:8000/accounts/settings' from origin 'http://0.0.0.0:8000' has been blocked by CORS policy: No 'Access-Control-Allow-Origin' header is present on the requested resource. If an opaque response serves your needs, set the request's mode to 'no-cors' to fetch the resource with CORS disabled.```

what

[green mesa]

where is how to setup bot is given

i arent able to find

i wnt to work on seasonal bot

[green oriole]

https://pythondiscord.com/pages/contributing/ they are all on the contributing page

[green mesa]

ty buddy

[green mesa]

MONTH_OVERRIDE whats the use of this?

[green oriole]

To make the bot think that we are in a particular month

[green mesa]

oo so i have to give 8?

[obsidian patio]

A crop like this would probably work

I would prefer not to cut off the hair on the side. That just makes it look weird

[crude gyro]

yes, that looks pretty good. would've been nice if we saw the whole drop maybe

could possibly make the drop a tiny bit smaller

[obsidian patio]

I agree about wanting to see the whole drop. Making it smaller could make it less distinct, though

[crude gyro]

well, how about you make the drop a tiny bit smaller and the crop a tiny bit larger and sorta meet at the middle

maybe even move the drop down a few pixels

use all of these in combination to solve it, instead of just adjusting one of those factors

[obsidian patio]

Yep, I just scaled the crop and moved it slightly down the face

[crude gyro]

that could work

we should test it

[obsidian patio]

Yeah. I'm quite happy with how that looks

[crude gyro]

can you give me a png that has no whitespace

[obsidian patio]

Sure

[green mesa]

nice

emoji

personal access token is github, api token? which bot needs??

[obsidian patio]

Here we go

[green mesa]

[obsidian patio]

@crude gyro

[crude gyro]

😅

[obsidian patio]

Hello

[crude gyro]

yeah that's pretty good

nice work

[obsidian patio]

Awesome, thanks

[crude gyro]

now make us a pull request

[green mesa]

👍

[obsidian patio]

Sure thing

[crude gyro]

and grats on your first lemoji

the first of many!

@green oriole

[green oriole]

Yay

[obsidian patio]

Any preferred dimentions for the png?

[crude gyro]

just do what we've done for the other ones I guess?

[obsidian patio]

Right, 128x128

[green mesa]

github api token is http://api.github.com/users/{username}?

[crude gyro]

@green mesa what are you doing?

I'm not sure why you need a github API token

[green mesa]

i wnt to contribute to seasonal bot so configuring env vars but i am stuck with github api

where to get

[molten bough]

It's in your account settings, under developer on the left side

[crude gyro]

sounds like an xy problem to me

[molten bough]

iirc

[crude gyro]

I don't see why you would need a github token to work on seasonalbot

just disable that extension if it's throwing up errors, or configure the env var to anything at all

unless you're working on a github API-related feature for seasonalbot, it's not relevant

[green mesa]

ok

[brazen charm]

It only raises ratelimits

[green mesa]

i am not doing anything related to github API

[obsidian patio]

First PR All feedback is appreciated

[green mesa]

Loading .env environment variables…
Traceback (most recent call last):
  File "__main__.py", line 6, in <module>
    from bot.bot import bot
  File "C:\Users\HP\Desktop\seasonalbot\bot\bot.py", line 13, in <module>
    from bot.constants import Channels, Client, MODERATION_ROLES
ModuleNotFoundError: No module named 'bot.constants'; 'bot' is not a package
PS C:\Users\HP\Desktop\seasonalbot\bot>```
getting this annoying error

[crude gyro]

that's not how you run the bot

[brazen charm]

use pipenv run start, or if you're using some kind of an ide for it you have tu run it as a module

[crude gyro]

if you wanted to run it with a regular command, you'd use python -m bot, and that would run the __main__.py as a module

[green mesa]

ooo

[crude gyro]

but, yeah, we have a run script so you should just do pipenv run start

[green mesa]

i am new to pipenv

never used it

[brazen charm]

How to run it etc. should all be in the contributing page for seasonalbot akarys linked

[crude gyro]

yes

[green mesa]

ok

wow ty

now its online

[tough imp]

github token is used by the branding manager

but it's not necessary, it only raises ratelimits for the API

unless you're working on the cog, you don't need it

[green mesa]

Ok

[brazen charm]

@tawdry vapor Just so I'm sure I didn't misunderstand how something works, how would I get them by ids without the api change? filter_fields is defined with ('active', 'author__id') for the reminder viewset so passing in id: id_val through params didn't have any effect and returned the whole reminders list

[tawdry vapor]

I was probably misled then

I only tried with a single reminder active and it looked like it worked cause it returned something

[neon phoenix]

can anyone close issue #1098?

btw, is anyone working on this? https://github.com/python-discord/bot/issues/796

[cold moon]

I'd like to work on it, but this is still not approved

[green oriole]

Shame that PyPi still doesn't have a search api

[patent pivot]

doooooes it not?

How does poetry search?

ah

https://github.com/python-poetry/poetry/blob/master/poetry/repositories/pypi_repository.py#L164-L194 lol

[tawdry vapor]

@hardy gorge why did you name this a sentinel? Is that really an appropriate name? https://github.com/python-discord/bot/pull/835#discussion_r463948061

That class basically looks like a mutex to me

Except it raises an error instead of blocking

I couldn't find any examples of that elsewhere

So I don't know what the proper name for it is

But surely it's some kind of mutex

Maybe it should just wait instead of raising an error. Chances are whatever operation happens 2nd will end up raising an error anyway.

Never mind. In the case of a scheduled task, we must abort it if e.g. the duration was modified, or the task itself would have to be responsible for checking if it's still valid somehow.

[hardy gorge]

Hmm, I just called it that because it acts as a gatekeeper

So, like a sentinel (the soldier at the gate) who either lets you through or not

It's probably not the correct term from a compsci/programming point of view

You're probably more aware of that terminology than I am

[tawdry vapor]

Perhaps, but only since today

Was trying to research a better name for it

The closest I got was PTHREAD_MUTEX_ERRORCHECK but that's only for raising an error if the same thread tries to lock a mutex it already owns

Nothing came up for a "non-blocking" mutex

[obsidian patio]

Any thoughts on this new Lemoji?

This is how it looks in a message (approximately):

Any feedback is appreciated

[crude gyro]

it's a little harder to identify as a lemon without the green hair, but

obviously the hair shouldn't be there

I don't hate it.

[hardy gorge]

I think it's recognizable enough in the context of our server

Since we have a bunch of lemojis, this one will work

It won't do as well used externally, I think

As far as being recognizable as a lemoji

[obsidian patio]

Yeah. I agree

Is there anything specific that you’d like to see changed with it before I create a pull request?

I don't hate it.
@crude gyro I must say that this really doesn’t sound too good

[crude gyro]

sorry

[obsidian patio]

It won't do as well used externally, I think
@hardy gorge are the lemojis actually used externally? Of course they would be used to some extent, but would anyone who isn’t from this server know what a lemoji even is?

No worries. I need honest critique in order to improve it

[crude gyro]

I'm a little preoccupied. it really is rather good, I love the cloud particularly.

I'm just worried that the head looks like a yellow strawberry

[obsidian patio]

Bahahah that’s really true

A chocolate dipped, yellow strawberry

[hardy gorge]

No, I'm not too worried about external use. It's just that in context with the other lemoji, it's recognizable; outside of that context, you may not recognize it as a lemon

[crude gyro]

I do really like it apart from that

[patent pivot]

No, I'm not too worried about external use. It's just that in context with the other lemoji, it's recognizable; outside of that context, you may not recognize it as a lemon
I think the recognition of lemojis outside the community is questionable anyway, since most lemons do not have beards

[crude gyro]

the really good ones do

[obsidian patio]

I’ll give it some minor tweaks tomorrow and push it to the branding repo

Sadly, it’ll have to be bald

[crude gyro]

we will just have to live with that

[obsidian patio]

Sadly, yes. At least the beard survived the explosion

[hardy gorge]

It's just lemon, but in 20 years.

[neon phoenix]

Do you think processing commands on edited messages is a good idea?

[bleak grail]

.http_cat 418 doesn't exist and that makes me sad

[green oriole]

It is because the bot is using the http module to validate the code and they didn't included 418

It will be fixed in 3.9

@neon phoenix the issue is that d.py doesn't have a native support for it, does it?