#dev-contrib | Python | Page 34

molten bough Oct 31, 2019, 6:26 PM

#

it's supposedly possible to get docker working fine under WSL but we tried and it's a real pain

mellow hare Oct 31, 2019, 6:26 PM

#

So far that's the only OS that shows issues with Docker. Possibly 8 as well, not sure

molten bough Oct 31, 2019, 6:26 PM

#

windows <10 will probably work OK with toolbox

mellow hare Oct 31, 2019, 6:26 PM

#

I don't like probably, though

#

I don't want to give a guide that "might" work

molten bough Oct 31, 2019, 6:26 PM

#

it's still worth remembering that toolbox is unsupported though, maybe at some point we'll need a feature that isn't part of that docker version

#

well, I don't have a machine to test on unfortunately

mellow hare Oct 31, 2019, 6:27 PM

#

And yeah, the fact that it's unsupported also makes me not want to suggest it

#

I don't want to suggest a hack, I want a solution for our users

molten bough Oct 31, 2019, 6:27 PM

#

there is basically a compromise to be made for working on the projects on w10 home right now

#

try to set up docker under WSL

#

or download and manually set up postgres

mellow hare Oct 31, 2019, 6:27 PM

#

Which we already have a functioning guide for

molten bough Oct 31, 2019, 6:27 PM

#

tbh I think the latter is less painful

mellow hare Oct 31, 2019, 6:27 PM

#

Exactly

molten bough Oct 31, 2019, 6:27 PM

#

right, okay

mellow hare Oct 31, 2019, 6:28 PM

#

I think we're all on the same page regarding W 10 Home's status in this. @hardy gorge I also think there's been a general consensus to keep the "Use Docker" option as the primary for the guide, especially since there's only one OS (and in fact only one version of that OS) that throws a wrench in it

molten bough Oct 31, 2019, 6:29 PM

#

the latter might be automatable as well tbh

#

Yeah, it does have an unattended mode

#

download installer, run unattended, connect as superuser and create the pydis role

mellow hare Oct 31, 2019, 6:31 PM

#

I'm not attempting to dismiss the others experiences here, as I know that Ak did have issues getting everything set up. However I feel that we will be benefiting more people by keeping Docker as the primary suggestion with a "Using Win 10 Home? Do this instead!" option

molten bough Oct 31, 2019, 6:31 PM

#

I agree, I think

#

the entire stack uses docker so it's more useful for people to be able to use docker if they can

brazen charm Oct 31, 2019, 6:31 PM

#

something about the project having to be on C: by default would be nice too, that kept me for a bit with the site not finding its files

molten bough Oct 31, 2019, 6:31 PM

#

It.. does?

#

That's odd

mellow hare Oct 31, 2019, 6:32 PM

#

I was under the impression that it was drive agnostic

molten bough Oct 31, 2019, 6:32 PM

#

It should be, it's Python

brazen charm Oct 31, 2019, 6:32 PM

#

Well you mentioned it to me gdude

molten bough Oct 31, 2019, 6:32 PM

#

Huh? I don't recall

#

Maybe I did

brazen charm Oct 31, 2019, 6:33 PM

#

under D: it ended with manage.py not found

#

moved it do C: and it works flawlessly

molten bough Oct 31, 2019, 6:33 PM

#

I don't remember that at all

mellow hare Oct 31, 2019, 6:33 PM

#

Is it that the project has to be on C or that the site/database instance has to be on the same drive, regardless of the letter?

#

I could certainly see that causing an issue

brazen charm Oct 31, 2019, 6:34 PM

#

it also worked when the site was ran under the bot, but then the bot didn't want to run

mellow hare Oct 31, 2019, 6:35 PM

#

That's bizarre. I think we'd have to see the errors or logs in order to work that out

#

If you can still replicate it

brazen charm Oct 31, 2019, 6:36 PM

#

my main problem with docker was that I couldn't find anything about the problems I was having, and if I did find something it was linux

#

could try but I wiped all the containers

mellow hare Oct 31, 2019, 6:36 PM

#

Ah, no worries then

hollow lichen Oct 31, 2019, 6:36 PM

#

A lazier option would be to point to somewhere you can download a unix/gnu based-OS :p

woeful thorn Oct 31, 2019, 6:36 PM

#

Or not

mellow hare Oct 31, 2019, 6:36 PM

#

I severely doubt that installing a whole new OS would make it better.

hollow lichen Oct 31, 2019, 6:37 PM

#

joke

mellow hare Oct 31, 2019, 6:37 PM

#

Regardless, I think we can make a troubleshooting guide as we find and squash problems

#

Or just correct it in the code as we receive the problems

#

I'll totally admit, when I first tried docker it was a chore getting it to work, but it's been incredibly simplified with all the commands we have in the projects

green oriole Oct 31, 2019, 6:41 PM

#

On docker you can't use pgAdmin to inspect the database?

molten bough Oct 31, 2019, 6:43 PM

#

You can

#

It runs on localhost:7777 I think

green oriole Oct 31, 2019, 6:48 PM

#

But how do you start it then?

#

Since it run on a container you don't have access to

molten bough Oct 31, 2019, 6:48 PM

#

Start what?

#

just run the container as normal

#

add it as a connection in pgadmin

green oriole Oct 31, 2019, 6:50 PM

#

pgAdmin start with the container?

molten bough Oct 31, 2019, 6:50 PM

#

No, just.. start it yourself

#

it's not in the container

mellow hare Oct 31, 2019, 6:53 PM

#

It being in a container doesn't mean you can't access it

#

Same way with a website. Just because it's on a different machine entirely doesn't mean you can't access it. It has ways built in for you to access it from the outside

molten bough Oct 31, 2019, 6:57 PM

#

pgAdmin isn't part of postgres, if that's what you're thinking

#

it's just a client

#

in the same way the site is a client

hardy gorge Oct 31, 2019, 7:15 PM

#

I think this will need to be discussed further first. This is an important part of our projects, so I want more input from our core developers group as well before taking a decision. We can obviously "blame Windows" for being the odd one out where it does not work, but it is one of the most used operating systems and probably the operating system that is used the most by new contributors, both for bot/site and for seasonalbot.

From that perspective, I think a guide that fits more people and works cross-platform (which is not impossible, since dockerless isn't an issue on linux/macos) should probably be the default guide. There should preferably be single, good, and general way of doing things.

mellow hare Oct 31, 2019, 7:15 PM

#

I guess I also don't understand why there has to be a dominant choice

last patio Oct 31, 2019, 7:16 PM

#

how hard is installing postgres on windows anyways

hardy gorge Oct 31, 2019, 7:16 PM

#

Much easier than installing docker

last patio Oct 31, 2019, 7:16 PM

#

yeah

#

I use pyenv for managing venvs & postgres on system and it works flawlessly, i'm on macos though

hardy gorge Oct 31, 2019, 7:17 PM

#

The ideal situation for me would be a single, good, main guide that most people are familiar with, instead of multiple, different guides less people are familiar with and have to be maintained separately. We already basically have just two or three people who can even help people set up our repositories on Windows.

molten bough Oct 31, 2019, 7:18 PM

#

It's pretty easy and you could actually just do it with a Python script

#

Since you can specify everything needed on the command line for an unattended install

last patio Oct 31, 2019, 7:19 PM

#

I wouldn't mess around with system dependencies from an unattended script

#

also would avoid too much magic

molten bough Oct 31, 2019, 7:19 PM

#

The windows installer actually has tons of questions and stuff, and unnecessary addons, etc

hardy gorge Oct 31, 2019, 7:19 PM

#

Postgres on windows, the last time I did, was basically download installer, run installer, check installation

mellow hare Oct 31, 2019, 7:21 PM

#

Yep

#

And I don't think we're suggesting multiple guides, just 2. And they're ones that we essentially already have

#

Our current guide literally covers both use cases

hardy gorge Oct 31, 2019, 7:22 PM

#

Our current guide only mentions the docker way: https://pythondiscord.com/pages/contributing/bot/ & https://pythondiscord.com/pages/contributing/site/

mellow hare Oct 31, 2019, 7:23 PM

#

Ah, I'm referring to the readme in the site repo

#

My mistake then

hardy gorge Oct 31, 2019, 7:24 PM

#

As I've said before, I'm not saying do away with docker; I'm saying write a good, main guide that should work on all platforms; have a secondary guide for user who like to use docker.

#

I'm probably going to continue using docker, I like it

mellow hare Oct 31, 2019, 7:24 PM

#

Whatever works

molten bough Oct 31, 2019, 7:25 PM

#

Did anyone figure out that docker logging output problem?

#

I basically just use docker for postgres right now because of it

mellow hare Oct 31, 2019, 7:26 PM

#

I was unaware there was one

molten bough Oct 31, 2019, 7:26 PM

#

Yeah, I noticed it with the site

#

Running it under docker loses half of the log output in dev

#

It's come up a couple times here

woeful thorn Oct 31, 2019, 7:26 PM

#

Is there an issue somewhere?

molten bough Oct 31, 2019, 7:26 PM

#

I'm not sure

mellow hare Oct 31, 2019, 7:27 PM

#

Yeah it's hard to keep track here, unfortunately

hardy gorge Oct 31, 2019, 7:27 PM

#

Yes, it doesn't happen anymore

#

The problem was the secret key

mellow hare Oct 31, 2019, 7:27 PM

#

Oh right yeah

molten bough Oct 31, 2019, 7:27 PM

#

The secret key?

#

Huh

hardy gorge Oct 31, 2019, 7:27 PM

#

yes, docker uses a SECRET_KEY; if that changes, your session is invalidated

molten bough Oct 31, 2019, 7:28 PM

#

Oh, that was a separate problem

hardy gorge Oct 31, 2019, 7:28 PM

#

If you have the dev server, it reloads on every change and the SECRET_KEY changed every time

#

Oh, then I probably don't know what you're talking about

molten bough Oct 31, 2019, 7:28 PM

#

Just loss of log output under docker

#

Some lines never make it to the terminal

hardy gorge Oct 31, 2019, 7:29 PM

#

Hmm, oh, right, I read your sentence wrong

molten bough Oct 31, 2019, 7:29 PM

#

I probably said I'd make an issue and forgot tbh

#

On mobile right now

#

Also, there are parts of .idea that are safe to put in the repo

#

Has the possibility of putting run configs in the repo for PyCharm been talked about?

crude gyro Oct 31, 2019, 7:43 PM

#

seems like bloat. far from everyone uses pycharm.

mellow hare Oct 31, 2019, 7:56 PM

#

Far from everyone even in our contributors use it

green oriole Oct 31, 2019, 7:57 PM

#

Talking about run config, can we add to the site pipfile the commands

run = "python manage.py run" 
rundebug = "python manage.py run --debug" ```?

molten bough Oct 31, 2019, 8:08 PM

#

Fair enough

#

I wouldn't call it run

#

pipenv run run?

#

Maybe just prod/debug

green oriole Oct 31, 2019, 9:01 PM

#

Yeah maybe we can name them differently

#

I'm tired of typing pipenv run python manage.py run --debug 😄

mellow hare Oct 31, 2019, 9:03 PM

#

I mean it should just be pipenv run start --debug

#

So I'm not 100% sure how pipenv run rundebug would be that much of an improvement or necessity

hollow lichen Nov 1, 2019, 8:36 AM

#

Instead of using DEBUG_MODE why don’t we just use __debug__? It’s build-in.

molten bough Nov 1, 2019, 8:39 AM

#

How's it work?

hardy gorge Nov 1, 2019, 8:39 AM

#

it depends on the run mode of python (the -O flags)

#

You can't change it yourself

molten bough Nov 1, 2019, 8:40 AM

#

Ah, so like a debug build of Python

#

That wouldn't work yeah

hardy gorge Nov 1, 2019, 8:40 AM

#

No

#

When you regularly run Python, __debug__ is True; if you run it with - O, or -OO, you run it in optimized mode

molten bough Nov 1, 2019, 8:40 AM

#

Ahh

hardy gorge Nov 1, 2019, 8:41 AM

#

One O will discard things like assert statements, with two, it will also discard docstrings

hollow lichen Nov 1, 2019, 8:42 AM

#

^

#

In production it’ll run using python -o instead of python

past falcon Nov 1, 2019, 9:04 AM

#

So what's the story behind the site-bot dependency? Seems a little bulky to need the site up to run the bot, at least for a dev environment. If it's just about the database, would it be possible to locally switcharoo the postgres with sqlite, to decouple things a little? Or is that just a terrible idea?

molten bough Nov 1, 2019, 9:07 AM

#

You don't need the site to run the bot

#

But it does need it for a lot of functionality

past falcon Nov 1, 2019, 9:07 AM

#

Oh, I was told I did

molten bough Nov 1, 2019, 9:07 AM

#

Part of the point is that the site can then be used as a management tool as well

#

Bot tags? Easily created in the Django admin

#

Quite a lot of functionality does require the site

hardy gorge Nov 1, 2019, 9:08 AM

#

Large parts of the bot depend on the API, both for getting information and sending information off. Running bot without site currently gets you a lot of exceptions.

molten bough Nov 1, 2019, 9:08 AM

#

A few basic things don't, I think

crude gyro Nov 1, 2019, 9:09 AM

#

we've gone a few rounds in the past about decoupling.

#

I fought pretty hard for it some time back.

#

but we were unable to reach a consensus for a number of reasons.

molten bough Nov 1, 2019, 9:10 AM

#

Oh yeah, I remember that!

crude gyro Nov 1, 2019, 9:10 AM

#

right now, as it stands, the coupling presents less of a problem than it ever has before

hardy gorge Nov 1, 2019, 9:11 AM

#

unless you're on Windows 10 Home ;)

crude gyro Nov 1, 2019, 9:11 AM

#

...unless you're on Windows 10 Home.

past falcon Nov 1, 2019, 9:11 AM

#

😄

molten bough Nov 1, 2019, 9:11 AM

#

Haha

crude gyro Nov 1, 2019, 9:11 AM

#

but if you're not, setting up the services to talk to each other is really easy.

molten bough Nov 1, 2019, 9:11 AM

#

Yeah

crude gyro Nov 1, 2019, 9:11 AM

#

and also, many argued that it's good design to have the bot make API calls.

#

having one database that's being accessed by two services is messier.

#

and has more potential for fuckup

molten bough Nov 1, 2019, 9:12 AM

#

Yeah, I think that was the argument I made

hardy gorge Nov 1, 2019, 9:12 AM

#

That's the general advice given to people who consider that for Django in general

#

Let Django take care of the database and communicate with Django

crude gyro Nov 1, 2019, 9:12 AM

#

yeah

hardy gorge Nov 1, 2019, 9:12 AM

#

However, I do think there's an option for decoupling, but it's not perfect.

#

It doesn't allow you to test insertions, so it's useless for database modifying things

crude gyro Nov 1, 2019, 9:13 AM

#

having the bot use a second database (like sqlite) also comes with downsides, particularly with the data that it makes sense to have both the site and the bot get access to.

green oriole Nov 1, 2019, 9:13 AM

#

We can’t just extract the api app and distribute it somewhere, so people doesn’t need to run the site entierly?

molten bough Nov 1, 2019, 9:13 AM

#

My main concern is, where do you do migrations? Do you use an ORM on the bot too? How do you share models?

crude gyro Nov 1, 2019, 9:13 AM

#

we have actually considered setting up a staging site

#

and allowing bot devs to just talk to that

hardy gorge Nov 1, 2019, 9:14 AM

#

We did say we did not want to make that open to all, since, well, people will abuse it

crude gyro Nov 1, 2019, 9:14 AM

#

mmyeah.

#

so if we do it, it might be accessible to contributors only, or even to core devs only. security policy still tbd

hardy gorge Nov 1, 2019, 9:15 AM

#

yes

green oriole Nov 1, 2019, 9:15 AM

#

We cannot just make an archive, host it on a server, and when you want to run the bot without the site, it just pullit from the server, unpack it, and run it?

hardy gorge Nov 1, 2019, 9:15 AM

#

How would you run an archive?

molten bough Nov 1, 2019, 9:16 AM

#

We call that Docker I believe

hardy gorge Nov 1, 2019, 9:16 AM

#

Yeah

#

That's basically what it does: A container that contains the stuff you want to deploy

#

Anyway, I hope WSL2 comes out soon

#

and this will not be an issue anymore

molten bough Nov 1, 2019, 9:16 AM

#

WSL2 is what breaks things actually

#

Appreciate the optimism though

hardy gorge Nov 1, 2019, 9:17 AM

#

In what way?

#

Microsoft has stated that it will be available for all versions of Windows and docker is building a Windows version of docker for it

molten bough Nov 1, 2019, 9:17 AM

#

It includes enough of hyper-v to make the virtualbox in docker toolbox not work

#

And we tried to use docker under it already

#

It does run, but you can't actually get at it from windows

crude gyro Nov 1, 2019, 9:18 AM

#

docker is building

hardy gorge Nov 1, 2019, 9:18 AM

#

Docker is building an actual new version of Docker for Windows specifically to be run using WSL2

crude gyro Nov 1, 2019, 9:18 AM

#

not has built

molten bough Nov 1, 2019, 9:18 AM

#

Ah okay

#

Well, I guess we'll see then

crude gyro Nov 1, 2019, 9:18 AM

#

https://docs.docker.com/docker-for-windows/wsl-tech-preview/

Docker Documentation

Docker Desktop WSL 2 Tech Preview

Overview Welcome to Docker Desktop WSL 2 Tech Preview. This Tech Preview introduces support to run Docker Desktop with WSL 2. We really appreciate you trialing this Tech Preview. Your...

#

can be tested already, looks like

molten bough Nov 1, 2019, 9:19 AM

#

I think @green oriole is on the insider preview so it might be worth a try then

crude gyro Nov 1, 2019, 9:19 AM

#

haang on

#

https://engineering.docker.com/2019/10/new-docker-desktop-wsl2-backend/

Docker Engineering Blog

Simon Ferquel

Introducing the Docker Desktop WSL 2 Backend - Docker Engineering Blog

What’s Changed Since the Tech Preview Earlier this year, we released a technical preview of our vision for the future of Docker development on Windows using WSL 2. We received lots of feedback from Windows Insiders via different channels, and collated common failure cases. ...

#

oh shit

green oriole Nov 1, 2019, 9:19 AM

#

Yeah i could try

crude gyro Nov 1, 2019, 9:19 AM

#

24th of october this was posted

molten bough Nov 1, 2019, 9:20 AM

#

Oh, super recent

crude gyro Nov 1, 2019, 9:20 AM

#

it looks like this is out, guys

past falcon Nov 1, 2019, 9:20 AM

#

Woo!

#

I think!

green oriole Nov 1, 2019, 9:20 AM

#

I’m going to try this afternoon

molten bough Nov 1, 2019, 9:20 AM

#

Well, WSL2 is still insider preview I think

green oriole Nov 1, 2019, 9:20 AM

#

Yeah it is

crude gyro Nov 1, 2019, 9:20 AM

#

oh is it?

#

I see

green oriole Nov 1, 2019, 9:20 AM

#

Yep

#

i installed the preview for that

#

(and it broken fortnite)

molten bough Nov 1, 2019, 9:21 AM

#

So that's why you've been so productive

#

:>

green oriole Nov 1, 2019, 9:21 AM

#

Yeah i guess that was for the best

hardy gorge Nov 1, 2019, 9:21 AM

#

Yes it is in the preview

#

Further down in the article, you can read that it's still in progress

#

There are changes still planned to solve issues that our, as of now, still there

#

but, it's looking promising

crude gyro Nov 1, 2019, 9:22 AM

#

WSL2 will be in 20H1

#

so like between january and june 2020

molten bough Nov 1, 2019, 9:22 AM

#

Damn, that's a big time window, haha

crude gyro Nov 1, 2019, 9:22 AM

#

yeah they do two releases a year

#

H1 and H2

hardy gorge Nov 1, 2019, 9:23 AM

#

As it appears, it requires some major updates to Windows as well

crude gyro Nov 1, 2019, 9:23 AM

#

but if the preview is relatively easy to install, and if it works well with the new docker experimental wsl2 mode, then that might be a legit alternative for those poor windows 10 home users

molten bough Nov 1, 2019, 9:23 AM

#

Yup, I agree

crude gyro Nov 1, 2019, 9:23 AM

#

because all of these things are out

#

they're just not out

molten bough Nov 1, 2019, 9:24 AM

#

Preview can be pretty breaky though

#

Especially on fast ring

hardy gorge Nov 1, 2019, 9:24 AM

#

What effect does signing up for the insiders program have?

crude gyro Nov 1, 2019, 9:24 AM

#

apparently it breaks fortnite

#

so that's one effect

molten bough Nov 1, 2019, 9:24 AM

#

Windows Update starts installing preview builds, ves

hardy gorge Nov 1, 2019, 9:24 AM

#

I asked for disadventages

molten bough Nov 1, 2019, 9:24 AM

#

That's about it

crude gyro Nov 1, 2019, 9:24 AM

#

hahaha

hardy gorge Nov 1, 2019, 9:24 AM

#

Oh, I did not

#

but I meant it

crude gyro Nov 1, 2019, 9:24 AM

#

you said EFFECTS

hardy gorge Nov 1, 2019, 9:24 AM

#

yeah

crude gyro Nov 1, 2019, 9:24 AM

#

but that's a good joke

#

terning5

molten bough Nov 1, 2019, 9:25 AM

#

it's a testbed basically

hardy gorge Nov 1, 2019, 9:25 AM

#

but not blessings

molten bough Nov 1, 2019, 9:25 AM

#

You're supposed to give feedback in the feedback app

#

I don't know anyone that does though

#

It also enables a shitton of telemetry

#

More than you'd ever see in a stable build

#

You can't turn it off either

hardy gorge Nov 1, 2019, 9:25 AM

#

Anyway, if it means that they'd get beta-updates or alpha-updates by default, I'm not sure if we can recommend it. We can suggest it, though

#

Ah, telemetry

molten bough Nov 1, 2019, 9:26 AM

#

Windows update enforces the telemetry setting on preview builds

#

If you don't update, the builds expire

#

So it's a fair commitment

green oriole Nov 1, 2019, 10:13 AM

#

The main downside of insider is that once you activated it, you can’t roll back to main release without reinstalling windows, even if it break stuff

molten bough Nov 1, 2019, 10:13 AM

#

well you can, but you have to wait for the next stable

#

and turn off insider preview before it comes around

glass pecan Nov 1, 2019, 10:13 AM

#

yeah it reverts eventually through updates

#

i turned it on once, but removed that windows install

#

now i get nonstop insider emails that i can't turn off

#

fun

green oriole Nov 1, 2019, 10:14 AM

#

You can turn it off, i did it

molten bough Nov 1, 2019, 10:15 AM

#

I used insider preview for years because I installed it literally as soon as it existed

#

like, w10 prerelease

#

it was relatively stable back then, but today.. eh

glass pecan Nov 1, 2019, 10:15 AM

#

pretty sure emails stop only when you turn off insider

#

which requires a win10 install

#

ill try again though, who knows it might be possible online now

#

📎 unknown.png

#

the pains of not installing corefonts

molten bough Nov 1, 2019, 10:17 AM

#

I, uh

#

wow

glass pecan Nov 1, 2019, 10:18 AM

#

yeah it doesn't fall back gracefully, does it

molten bough Nov 1, 2019, 10:18 AM

#

yeah, no, haha

hollow lichen Nov 1, 2019, 10:45 AM

#

any thought on using __debug__? It could avoid setting DEBUG_MODE in __init__ in a inelegant way.

molten bough Nov 1, 2019, 10:46 AM

#

It's not really a simple option I don't think

#

most people will never run with -O or -OO

#

it's just not very obvious and it can be handy to run debug mode code when running in optimised mode anyway

hollow lichen Nov 1, 2019, 10:52 AM

#

But it's also handy to run the bot without debug mode when testing it, which is not possible now, without changing the code. And I'm pretty sure no one did run the bot repo using -oo when testing it.

#

Anyway it was just a proposition, since __debug__ variable is very often forgotten

molten bough Nov 1, 2019, 10:54 AM

#

it's definitely possible

hollow lichen Nov 1, 2019, 11:01 AM

#

Maybe the best way to implement __debug__ would simply to assign DEBUG_MODE = __debug__without replacing it every time in the code-source.

#

This way it could be still possible to run debug mode and optimised mode, by manually setting the value of D_M to True. But since it's not a common practice, changing the code to achieve it seems fine to me.

green oriole Nov 1, 2019, 11:05 AM

#

But.. I don’t want the debug mode

#

i want to run as if it was in production

#

Configuration is good

molten bough Nov 1, 2019, 11:06 AM

#

which you won't be able to do with either pipenv or docker-compose if you have to use -O, yeah

hollow lichen Nov 1, 2019, 11:06 AM

#

Debug mode is useful due to debug logs

green oriole Nov 1, 2019, 11:06 AM

#

Leave the choice to people

hollow lichen Nov 1, 2019, 11:08 AM

#

Well currently there is no choice, debug mode is activated when not on the server

green oriole Nov 1, 2019, 11:09 AM

#

You can still set the env var to run in debug mode, i don’t see the point of using __default__

#

I have it disabled

molten bough Nov 1, 2019, 11:09 AM

#

DEBUG_MODE = True if 'local' in os.environ.get("SITE_URL", "local") else False

#

weird default value there

hollow lichen Nov 1, 2019, 11:09 AM

#

exactly

green oriole Nov 1, 2019, 11:12 AM

#

But.. the SITE_URL is not set to local isn’t?

molten bough Nov 1, 2019, 11:12 AM

#

that's not what the conditional says

#

although this whole line is too long

#

it could just be "local" in os.environ.get(...)

green oriole Nov 1, 2019, 11:12 AM

#

This line is wierd

hollow lichen Nov 1, 2019, 11:13 AM

#

I don't really understand this default value, that's why I propose a cleaner set of the variable

molten bough Nov 1, 2019, 11:13 AM

#

It's probably there just to make things easier for contribs

#

I don't personally see a big reason to change it

glass pecan Nov 1, 2019, 11:13 AM

#

is this bot

molten bough Nov 1, 2019, 11:14 AM

#

yeah

hollow lichen Nov 1, 2019, 11:14 AM

#

yes

molten bough Nov 1, 2019, 11:14 AM

#

yup

glass pecan Nov 1, 2019, 11:14 AM

#

right, config adjustments are going to happen soon as peeps know. suggestions about how to improve configuration is always welcome especially with it coming up.

#

specifically on this topic though, we are making it so debug is able to be chosen to be enabled or disabled by configuration

molten bough Nov 1, 2019, 11:15 AM

#

I was gonna say, default to this, override in config

glass pecan Nov 1, 2019, 11:15 AM

#

yep

green oriole Nov 1, 2019, 11:15 AM

#

Yeah there a config file, why we don’t use it?

hollow lichen Nov 1, 2019, 11:15 AM

#

in config.yml?

green oriole Nov 1, 2019, 11:16 AM

#

Yep

glass pecan Nov 1, 2019, 11:16 AM

#

both config.yml and env vars will be supported

molten bough Nov 1, 2019, 11:16 AM

#

well the entire config system will be overhauled anyway, as scragly said

#

will be an interesting issue/PR, that's for sure

hollow lichen Nov 1, 2019, 11:16 AM

#

we can't do relative import in a __init__.py

molten bough Nov 1, 2019, 11:16 AM

#

yeah you can

glass pecan Nov 1, 2019, 11:16 AM

#

what's a relative import got to do with this?

hollow lichen Nov 1, 2019, 11:17 AM

#

# We can't import this yet, so we have to define it ourselves
DEBUG_MODE = True if 'local' in os.environ.get("SITE_URL", "local") else False

#

Idk I just read the comment

molten bough Nov 1, 2019, 11:18 AM

#

well yeah, see, there is the debug mode paradox

brazen charm Nov 1, 2019, 11:18 AM

#

With debug mode, what's the purpose of all the extra logging? Some are useful but things like typing start could be controlled elsewhere as they're not needed most of the time and clutter things up a lot

molten bough Nov 1, 2019, 11:18 AM

#

you want to do debug logging when loading the config

#

but you need to load the config to set debug mode

glass pecan Nov 1, 2019, 11:19 AM

#

it's fine, i'm sure we'll get a good solution to get what we need

molten bough Nov 1, 2019, 11:19 AM

#

Yeah, I believe in the team

glass pecan Nov 1, 2019, 11:20 AM

#

what's the typing start reference numerlor?

hollow lichen Nov 1, 2019, 11:20 AM

#

My point was to fully exploit built-in variables

green oriole Nov 1, 2019, 11:22 AM

#

__debug__ is too special to be used imo

brazen charm Nov 1, 2019, 11:22 AM

#

There are debut logging events when someone starts typing

#

Or at least they appear like it

glass pecan Nov 1, 2019, 11:23 AM

#

is there? seems odd

#

where abouts is that

brazen charm Nov 1, 2019, 11:27 AM

#

https://i.imgur.com/Bb4tIoH.png for every message that is sent

molten bough Nov 1, 2019, 11:28 AM

#

yeah, that's discord.py logging

#

not the bot code

past falcon Nov 1, 2019, 11:28 AM

#

Good things come in threes.

brazen charm Nov 1, 2019, 11:28 AM

#

is it set in the bot? If it can be I think it should be moved from the normal debug

hollow lichen Nov 1, 2019, 11:29 AM

#

it's normally silenced @brazen charm

#

https://github.com/python-discord/bot/blob/35648fd85deca421783cbaf871ded048ebd3d37b/bot/__init__.py#L90

GitHub

python-discord/bot

Source code for our Discord bot. Contribute to python-discord/bot development by creating an account on GitHub.

glass pecan Nov 1, 2019, 11:29 AM

#

yeah i don't usually see dpy logs

#

at least, not their debug ones

brazen charm Nov 1, 2019, 11:30 AM

#

I've got it by default and that's all I care about, disabled it but it was getting in the way of everything before that. So it should be changed if possible imo

#

about the doc command I'm working on, should "New in version x.y." be removed from the top of modules or stay there? https://i.imgur.com/znczmbx.png

Imgur

glass pecan Nov 1, 2019, 11:34 AM

#

i kinda like it

molten bough Nov 1, 2019, 11:34 AM

#

Yeah, not a bad thing to keep around

#

especially for latest changes

glass pecan Nov 1, 2019, 11:34 AM

#

yeah

#

most especially since dependancies are a bloody mess for some projects

molten bough Nov 1, 2019, 11:34 AM

#

ain't that the truth

glass pecan Nov 1, 2019, 11:35 AM

#

we can't upgrade our python because it will cause catastrophic failure of the staff coffee machine

molten bough Nov 1, 2019, 11:39 AM

#

Any thoughts on blacking the code once the PRs have died down a bit by the way?

#

I wasn't a fan of it initially but I've started doing it to my projects

crude gyro Nov 1, 2019, 11:40 AM

#

we've tried a few times

#

people are never happy with the results.

glass pecan Nov 1, 2019, 11:40 AM

#

is this literally just using black on the code

hardy gorge Nov 1, 2019, 11:40 AM

#

yes

crude gyro Nov 1, 2019, 11:41 AM

#

I don't think there's much of a point in trying a third time because I just can't see the consensus going differently than it has

glass pecan Nov 1, 2019, 11:41 AM

#

is black's style the same as ours

hardy gorge Nov 1, 2019, 11:41 AM

#

no

glass pecan Nov 1, 2019, 11:41 AM

#

then i'm against it

#

lol

molten bough Nov 1, 2019, 11:41 AM

#

haha, fair

#

I didn't realise it had been discussed

hardy gorge Nov 1, 2019, 11:41 AM

#

but not that different; we've had a discussion with comparisons before

crude gyro Nov 1, 2019, 11:41 AM

#

I'm not sure black would conform to some of the rules we set in the linter, like flake8-docstrings.

glass pecan Nov 1, 2019, 11:41 AM

#

i think we've settled pretty well with a particular set of consistent styles

woeful thorn Nov 1, 2019, 11:42 AM

#

It does, I use it for flake8-annotations

glass pecan Nov 1, 2019, 11:42 AM

#

and it wasn't that long ago we actually applied all the linting rules

crude gyro Nov 1, 2019, 11:42 AM

#

mkay.

woeful thorn Nov 1, 2019, 11:42 AM

#

(Not explicitly)

crude gyro Nov 1, 2019, 11:42 AM

#

black isn't a linter, it's just a code reformatter. you'd run a single pass through all the code to make it all consistent.

molten bough Nov 1, 2019, 11:42 AM

#

I mean the advantage is that you can code in your own style, right?

#

and then black "fixes" it

glass pecan Nov 1, 2019, 11:42 AM

#

besides, everytime there's a style change, we've gotta review a massive amount of code changes and it destroys blame

molten bough Nov 1, 2019, 11:42 AM

#

you still need a linter though, yeah

crude gyro Nov 1, 2019, 11:43 AM

#

yeah it would really annihilate blame

molten bough Nov 1, 2019, 11:43 AM

#

yeah, it would, that's true

hollow lichen Nov 1, 2019, 11:44 AM

#

blame?

crude gyro Nov 1, 2019, 11:44 AM

#

some people will add it to a precommit though and then most linting is no longer necessary.

glass pecan Nov 1, 2019, 11:44 AM

#

blame is commit history

#

last edited by

molten bough Nov 1, 2019, 11:44 AM

#

it allows you to see who changed what line most recently, yeah

hollow lichen Nov 1, 2019, 11:44 AM

#

okay

woeful thorn Nov 1, 2019, 11:45 AM

#

I wouldn’t mind it for new things but there’s no will to do it for the existing codebase

crude gyro Nov 1, 2019, 11:45 AM

#

yeah I agree.

molten bough Nov 1, 2019, 11:45 AM

#

I do wonder just how many changes it'd make though

woeful thorn Nov 1, 2019, 11:45 AM

#

The Django migration was probably the best opportunity

#

A lot

crude gyro Nov 1, 2019, 11:45 AM

#

last time we tried it did a lot.

molten bough Nov 1, 2019, 11:45 AM

#

I mean my projects were mostly conformant already and I basically follow pydis style

crude gyro Nov 1, 2019, 11:45 AM

#

pydis style

molten bough Nov 1, 2019, 11:45 AM

#

it just outright skipped some files

crude gyro Nov 1, 2019, 11:45 AM

#

that makes me shudder a little

molten bough Nov 1, 2019, 11:45 AM

#

haha, sorry

hardy gorge Nov 1, 2019, 11:46 AM

#

We did try black the previous time we discussed this

#

Let me see if I can dig up the examples

woeful thorn Nov 1, 2019, 11:46 AM

#

I should have kept the branch

crude gyro Nov 1, 2019, 11:46 AM

#

yeah and then the core devs subsequently shot it down

molten bough Nov 1, 2019, 11:46 AM

#

I mean hey, if it's a bad idea, it's a bad idea

glass pecan Nov 1, 2019, 11:47 AM

#

@hollow lichen here's what i see as an example when i code

📎 unknown.png

woeful thorn Nov 1, 2019, 11:47 AM

#

I think it’s a great idea 😬

molten bough Nov 1, 2019, 11:47 AM

#

it's definitely a marmite tool either way

crude gyro Nov 1, 2019, 11:47 AM

#

I don't mind either way but I do agree with some of the arguments that were made against it.

glass pecan Nov 1, 2019, 11:47 AM

#

my cursor shows blame for the current line

molten bough Nov 1, 2019, 11:47 AM

#

what the heck plugin is that

#

that's neat

glass pecan Nov 1, 2019, 11:47 AM

#

git toolbox i think

#

yeah GitToolBox it is

molten bough Nov 1, 2019, 11:48 AM

#

yoink

glass pecan Nov 1, 2019, 11:48 AM

#

it does a lot more than that

#

but that specific thing is definitely nice

hollow lichen Nov 1, 2019, 11:48 AM

#

@glass pecan Oh ok, since I don't use and IDE I don't see them outside Github

glass pecan Nov 1, 2019, 11:48 AM

#

we use it in gh too, yeah

crude gyro Nov 1, 2019, 11:49 AM

#

can still see blame with git blame

hardy gorge Nov 1, 2019, 11:49 AM

#

I sometimes browse the git blame history to find the commit message of a certain decision

crude gyro Nov 1, 2019, 11:49 AM

#

same

hardy gorge Nov 1, 2019, 11:49 AM

#

but not all historical commit messages are, well, informative

crude gyro Nov 1, 2019, 11:49 AM

#

sadly.

glass pecan Nov 1, 2019, 11:49 AM

#

thats why commit messages are important

hardy gorge Nov 1, 2019, 11:49 AM

#

yes

molten bough Nov 1, 2019, 11:50 AM

#

https://xkcd.com/1296/

xkcd: Git Commit

glass pecan Nov 1, 2019, 11:50 AM

#

and also why addressing linting late is bad whip

crude gyro Nov 1, 2019, 11:50 AM

#

this xkcd makes me upset

molten bough Nov 1, 2019, 11:50 AM

#

:>

crude gyro Nov 1, 2019, 11:50 AM

#

because it makes it look like this is okay.

#

this isn't okay

#

not even the top one

molten bough Nov 1, 2019, 11:50 AM

#

It's totally not okay

hardy gorge Nov 1, 2019, 11:51 AM

#

Yes, but, luckily, usually the git blame around the linting errors stays intact even with linting corrections and/or you can view the previous blame. It's just more work.

glass pecan Nov 1, 2019, 11:51 AM

#

some people also don't realise you can put lengthy descriptions in commit descriptions after separating the summary with the body with newlines

crude gyro Nov 1, 2019, 11:51 AM

#

Created main loop & timing control is almost a worse commit message than misc bugfixes because the former could really use some more info while the latter may or may not be that important.

molten bough Nov 1, 2019, 11:51 AM

#

to be fair, it might have a summary

#

haha

hardy gorge Nov 1, 2019, 11:51 AM

#

I hope this is just the oneline log

#

but, randall munroe is not a dev, so probably not

molten bough Nov 1, 2019, 11:52 AM

#

yeah, true

#

well anyway

#

like I blacked django-simple-bulma in my django-simple-buefy fork and most files were untouched

#

but that's a fairly simple project I guess

past falcon Nov 1, 2019, 11:54 AM

#

I'm mentally preparing to be yelled at by lemon whenever my first commit drops. I'm sure I'll be breaking dozens of best-practices rules

glass pecan Nov 1, 2019, 11:54 AM

#

you're not going to find your expectation coming to pass

#

lemon doesn't yell

crude gyro Nov 1, 2019, 11:55 AM

#

inb4 lemon only murders people with his fistpunchers.

brazen charm Nov 1, 2019, 11:55 AM

#

The reviews are there for a reason :P

glass pecan Nov 1, 2019, 11:56 AM

#

he lulls you into a false sense of security, gently guiding you into doing things exactly how it's meant to be done, only for you to realise eventually that you're suddently spending way too much time contributing for pydis projects in not only your free time but your sleeping time, working time and eating time. you inevitably become the pydis slave you were meant to be

woeful thorn Nov 1, 2019, 11:56 AM

#

Unless you don’t lint your code

#

Then there’s no lull

glass pecan Nov 1, 2019, 11:56 AM

#

well yeah, people who don't lint aren't even human

crude gyro Nov 1, 2019, 11:56 AM

#

you sound bitter, @glass pecan

glass pecan Nov 1, 2019, 11:56 AM

#

not bitter, resigned

past falcon Nov 1, 2019, 11:56 AM

#

Lol. I'll invest in some body armor.

glass pecan Nov 1, 2019, 11:56 AM

#

resigned to my fate

crude gyro Nov 1, 2019, 11:56 AM

#

I'm not very yelly, that much is true.

woeful thorn Nov 1, 2019, 11:56 AM

#

I wonder how much GitHub would charge us for an enterprise server

exotic ember Nov 1, 2019, 11:57 AM

#

make sure to already have a PR open and merge master like, 6 bajillion times before it's ready

glass pecan Nov 1, 2019, 11:57 AM

#

probs a lot

molten bough Nov 1, 2019, 11:57 AM

#

oh, they'd charge you plenty

crude gyro Nov 1, 2019, 11:57 AM

#

@past falcon you got precommit enabled?

woeful thorn Nov 1, 2019, 11:57 AM

#

Then we could have hooks to bounce pushes if they fail linting

glass pecan Nov 1, 2019, 11:57 AM

#

that's a thing eh

molten bough Nov 1, 2019, 11:57 AM

#

yeah, true

glass pecan Nov 1, 2019, 11:57 AM

#

still not worth the money

molten bough Nov 1, 2019, 11:57 AM

#

you can do hooks server-side if you control the environment

past falcon Nov 1, 2019, 11:57 AM

#

I think so 🙂 The pipenv run precommit?

crude gyro Nov 1, 2019, 11:57 AM

#

yeah. serverside post-push hooks are a thing.

#

yeah sed

brazen charm Nov 1, 2019, 11:58 AM

#

I mean, I had precommit enabled but still failed the build because annotations didn't get installed properly

molten bough Nov 1, 2019, 11:58 AM

#

Gotta run them tests too

crude gyro Nov 1, 2019, 11:58 AM

#

yeah tests will also fail the build

#

I make that mistake myself, though.

glass pecan Nov 1, 2019, 11:58 AM

#

i did too

#

because coverage

crude gyro Nov 1, 2019, 11:58 AM

#

still not used to tests in bot

molten bough Nov 1, 2019, 11:59 AM

#

yeah

crude gyro Nov 1, 2019, 11:59 AM

#

I've failed a few builds because I made changes to constants.

#

and the test got fucked up

molten bough Nov 1, 2019, 11:59 AM

#

I'm still not happy with that monolithic test file on site

glass pecan Nov 1, 2019, 11:59 AM

#

literally first time i touched django personally with a pr i ended up forgetting about 100% cov

molten bough Nov 1, 2019, 11:59 AM

#

but still have no better ideas

woeful thorn Nov 1, 2019, 11:59 AM

#

@brazen charm keeping issues to yourself doesn’t help us fix them

crude gyro Nov 1, 2019, 11:59 AM

#

@molten bough eh? don't we have a bunch of test files on site?

molten bough Nov 1, 2019, 12:00 PM

#

oh sure, but you saw the allauth test file I wrote

glass pecan Nov 1, 2019, 12:00 PM

#

i think he's referencing the one he worked on specifically

#

yeah

molten bough Nov 1, 2019, 12:00 PM

#

like 200 lines of setup and then 50 lines of actual tests

brazen charm Nov 1, 2019, 12:00 PM

#

It just didn't install properly what issues are you on about?

crude gyro Nov 1, 2019, 12:00 PM

#

that one was fine.

#

isn't "it didn't install properly" an issue?

glass pecan Nov 1, 2019, 12:00 PM

#

only 250 lines

#

😛

crude gyro Nov 1, 2019, 12:00 PM

#

dude we have some tests that are thousands of lines where I work :(

#

your 250 are a dream compared to those

molten bough Nov 1, 2019, 12:00 PM

#

yeah but I bet most of that is actually tests

crude gyro Nov 1, 2019, 12:01 PM

#

not really.

glass pecan Nov 1, 2019, 12:01 PM

#

my bot has a module with 1k lines

#

📎 unknown.png

#

and it doesn't even do anything

brazen charm Nov 1, 2019, 12:01 PM

#

Not an issue you could influence afaik

crude gyro Nov 1, 2019, 12:01 PM

#

well let's talk about it anyway

#

why did it fail?

molten bough Nov 1, 2019, 12:01 PM

#

is that the pipenv hash check failure you get at random?

crude gyro Nov 1, 2019, 12:01 PM

#

could other people run into this problem?

#

is there anything we can do to prevent future contributors from experiencing that issue?

brazen charm Nov 1, 2019, 12:04 PM

#

My setup is different from the guide and had a few issues at the start so I really don't think it's something you could fix

crude gyro Nov 1, 2019, 12:04 PM

#

mkay. good to know.

#

thanks

past falcon Nov 1, 2019, 12:07 PM

#

One thing that bit me setting up the bot was that the contributing guide suggested to put
site: &DOMAIN "web:8000" in the yaml file, but what it turned out to want was
"pythondiscord.local:8000"

crude gyro Nov 1, 2019, 12:08 PM

#

yeah, this is one of the things that annoys me - it's because we have multiple approaches to how the containers should speak with each other.

#

the bot can set up a site in the docker-compose which you need to access via the docker container name

#

or you could be setting up site seperately with the other guide

#

in which case you'd want to use pythondiscord.local

#

we should just have a single way to do this, and imo, it should be the latter way..

#

but the former is undeniably convenient if you have no intention of ever working on site

past falcon Nov 1, 2019, 12:10 PM

#

Ah, ok. Yeah I'm one of those dockerless win10 homers. @green oriole saved the day though and helped me soldier through it, I have something functional now ^_^

crude gyro Nov 1, 2019, 12:12 PM

#

but rest assured that we're aware that the setup isn't perfect on every platform and we're thinking about how to improve it.

glass pecan Nov 1, 2019, 12:12 PM

#

yeah i didn't think w10 home would be such an issue, as even if you're working on having bot and site separately worked on, you can easily run both together using the project name env var

crude gyro Nov 1, 2019, 12:12 PM

#

it'll probably change a lot over the next year, in various ways.

exotic ember Nov 1, 2019, 2:24 PM

#

@hardy gorge I'm working on testing the pep command among other related commands, how would I mock the http get? I can't seem to monkeypatch bot.http_session

molten bough Nov 1, 2019, 2:24 PM

#

It depends how exactly it's being used

hardy gorge Nov 1, 2019, 2:24 PM

#

Why can't you mock it?

molten bough Nov 1, 2019, 2:24 PM

#

If you're importing it somehow you can mock out the import

#

rather than the definition

hardy gorge Nov 1, 2019, 2:25 PM

#

Do you pass a MockBot?

exotic ember Nov 1, 2019, 2:25 PM

#

yep

hardy gorge Nov 1, 2019, 2:25 PM

#

Because that one has a mocked http_session attribute ready to go, if I'm not mistaken

#

Let me check

exotic ember Nov 1, 2019, 2:25 PM

#

apparently not

hardy gorge Nov 1, 2019, 2:26 PM

#

Are you using the latest version?

#

>>> from tests.helpers import MockBot
>>> bot = MockBot()
>>> bot.http_session
<MagicMock name='mock.http_session' id='140019569944336'>

exotic ember Nov 1, 2019, 2:27 PM

#

yap, that was it

#

found it after merging master into my branch

hardy gorge Nov 1, 2019, 2:27 PM

#

Ah, nice

#

Just a question, what went wrong when trying to mock it?

exotic ember Nov 1, 2019, 2:27 PM

#

AttributeError

hardy gorge Nov 1, 2019, 2:27 PM

#

Right, so

#

The problem is that http_client is a custom attribute we set on our bot, not one that discord.py provides.

#

That means that when trying to access it, it doesn't follow the right specifications and it will throw that error to indicate "You're trying to access an attribute that should not exist"

#

You can get around that by explicitly setting that attribute on that mock

#

if it ever happens again

#

>>> bot.something
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib/python3.7/unittest/mock.py", line 593, in __getattr__
    raise AttributeError("Mock object has no attribute %r" % name)
AttributeError: Mock object has no attribute 'something'
>>> from unittest.mock import MagicMock
>>> bot.something = MagicMock()
>>> bot.something
<MagicMock name='mock.something' id='140019542812112'>

#

I think the README has some more info about that

exotic ember Nov 1, 2019, 2:30 PM

#

give me a second to rewind and check again, could've sworn I had the attribute error trying to assign a mock 👀

hardy gorge Nov 1, 2019, 2:30 PM

#

Did you try assigning an attribute of an attribute?

#

bot.something.something?

exotic ember Nov 1, 2019, 2:32 PM

#

hmm, that could be it

brazen charm Nov 1, 2019, 2:33 PM

#

One more thing I wanted to ask, a lot of (stdlib) module descriptions end like this https://i.imgur.com/dTczax1.png should the last line be removed if it ends with a :?

Imgur

exotic ember Nov 1, 2019, 2:35 PM

#

don't see any problems with that

brazen charm Nov 1, 2019, 6:25 PM

#

How do you restart docker itself? Had to restart my pc after putting it to sleep

molten bough Nov 1, 2019, 6:26 PM

#

on linux, sudo systemctl restart docker - at least on most distros

tawdry vapor Nov 1, 2019, 6:26 PM

#

Yup I have to do that command every time I resume my VM otherwise containers cannot connect to the internet 🤷

last patio Nov 1, 2019, 7:32 PM

#

whats this mean?

📎 unknown.png

woeful thorn Nov 1, 2019, 7:33 PM

#

113 passed, 10 other

#

Usually other is skipped

last patio Nov 1, 2019, 7:35 PM

#

thanks

#

i have no skips locally 🤔

📎 unknown.png

#

oh, maybe we need to set a fake token

#

we do, weird

#

oh now I see why

hardy gorge Nov 1, 2019, 7:48 PM

#

The skips are the 10 lists/dicts in constants we do not yet validate

last patio Nov 1, 2019, 7:49 PM

#

mhm

brazen charm Nov 1, 2019, 8:37 PM

#

should all the constants be global level or would it be okay to have a class attribute? For example I have a tuple of around half a dozen strings I end a search on, where would I put the declaration of that if it's pretty much only used in one method? Did it on the top so far but the class attribute makes more sense to me here

molten bough Nov 1, 2019, 8:56 PM

#

I prefer to have them at the top since they can all be in one place in all possible cases then

mellow hare Nov 1, 2019, 9:04 PM

#

Yeah I think consistency needs to win out on that

tough imp Nov 1, 2019, 9:14 PM

#

Can I please get assigned knackles https://github.com/python-discord/bot/issues/600

GitHub

Write unit tests for `bot/rules/links.py` · Issue #600 · python-...

Write unit tests for bot/rules/links.py. Implementation details Please make sure to read the general information in the meta issue and the testing README. We are aiming for a 100% branch coverage f...

mellow hare Nov 1, 2019, 9:16 PM

#

Done

tough imp Nov 1, 2019, 9:17 PM

#

thank

brazen charm Nov 1, 2019, 9:30 PM

#

I'm having a bit of an disagreement with PEP 8 E131 and flake8, it wants hanging indents like this

            search_end_tag = search_start_tag.find_next(
                lambda tag: any(attr in tag.get("class", ()) for attr in SEARCH_END_TAG_ATTRS)
                or tag.name == "table"
            )

but I think this makes more sense and is clearer on the intention

            search_end_tag = search_start_tag.find_next(
                lambda tag: any(attr in tag.get("class", ()) for attr in SEARCH_END_TAG_ATTRS)
                            or tag.name == "table"
            )

Should I just go with the one that passes the linter? Or reformat it even more

crude gyro Nov 1, 2019, 9:41 PM

#

maybe just refactor your multiline lambda into a function and pass that?

tawdry vapor Nov 1, 2019, 10:06 PM

#

I agree

#

that code is just too much to handle for reading sweatcat

#

~~Also I disagree that the second is clearer.~~ I rescind this, I misread the code originally

crude gyro Nov 1, 2019, 10:08 PM

#

@sleek mason are you sure the bot token is the issue? maybe you're making requests to the prod API or something

sleek mason Nov 1, 2019, 10:08 PM

#

hmm it's possible yeah

crude gyro Nov 1, 2019, 10:09 PM

#

did you follow the new guides?

brazen charm Nov 1, 2019, 10:09 PM

#

Should the any also be broken up? Thinking how it could be done but can't think of a clearer way that doesn't loop over it and checks itself instead of using the builtin

sleek mason Nov 1, 2019, 10:09 PM

#

i did all of the config changes in the contrib guide, i'm pretty sure

#

lemme double check

crude gyro Nov 1, 2019, 10:09 PM

#

the guide on our website, right?

sleek mason Nov 1, 2019, 10:10 PM

#

yeah

#

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

Python Discord | Contributing to Bot

We're a large, friendly community focused around the Python programming language. Our community is open to those who wish to learn the language, as well as those looking to help others.

crude gyro Nov 1, 2019, 10:10 PM

#

yeah good.

#

and how are you running the site?

sleek mason Nov 1, 2019, 10:11 PM

#

docker-compose up

#

do i need any debug stuff in there?

#

i feel like i do

tawdry vapor Nov 1, 2019, 10:11 PM

#

@brazen charm for example ```py
def predicate(tag):
x = any(attr in tag.get("class", ()) for attr in SEARCH_END_TAG_ATTRS)

return x or tag.name == "table"

Though give `x` a better name

crude gyro Nov 1, 2019, 10:12 PM

#

@sleek mason you're docker compose uping in the site repo?

tawdry vapor Nov 1, 2019, 10:12 PM

#

Or just turn it into a for loop

crude gyro Nov 1, 2019, 10:12 PM

#

or just letting the bot run both

sleek mason Nov 1, 2019, 10:12 PM

#

i'm doing exactly as it says here.

📎 unknown.png

#

is there any particular way i need to do it, perhaps?

crude gyro Nov 1, 2019, 10:13 PM

#

good. okay.

#

and how does the part of your config with the API stuff look?

brazen charm Nov 1, 2019, 10:13 PM

#

I made this to avoid the empty tuple check, or should I just use the any

    @staticmethod
    def match_end_tag(tag: Tag) -> bool:
        if tag.name == "table":
            return True
        tag_attr = tag.get("class", None)
        if tag_attr:
            for search_attr in SEARCH_END_TAG_ATTRS:
                if search_attr in tag_attr:
                    return True
        return False

sleek mason Nov 1, 2019, 10:14 PM

#

ah fuck i need to make it "pythondiscord.local" in there don't i

crude gyro Nov 1, 2019, 10:14 PM

#

yeah

molten bough Nov 1, 2019, 10:15 PM

#

that won't resolve if the bot is running under docker

sleek mason Nov 1, 2019, 10:15 PM

#

i definitely changed that according to this before but apparently it just didn't save like that? fuck knows

📎 unknown.png

crude gyro Nov 1, 2019, 10:15 PM

#

that'll resolve just fine

#

so did that work @sleek mason?

sleek mason Nov 1, 2019, 10:16 PM

#

if you get a mention in the other server again, then that's a no haha. give me a moment

#

yup, that's worked.

crude gyro Nov 1, 2019, 10:18 PM

#

good!

sleek mason Nov 1, 2019, 10:18 PM

#

i swear my previous issues with setting up the bot were because of a misconfiguration of that

#

it must just be a me problem

crude gyro Nov 1, 2019, 10:18 PM

#

happens.

sleek mason Nov 1, 2019, 10:18 PM

#

just hope i'm not going senile :D

#

okay well, thanks a bunch. now i can get working on whatever bot/site stuff i can without harassing everybody with @everyone mentions

#

(but i swear i definitely changed that to "web:8000" before. i specifically remember doing it. oh well.)

tawdry vapor Nov 1, 2019, 10:24 PM

#

@brazen charm Hmm that looks too convoluted

#

This may be the best IMO ```py
def predicate(tag):
for attr in SEARCH_END_TAG_ATTRS:
if attr in tag.get("class", ()):
return True

return tag.name == "table"


The `any()` would still be too long I think, if its variable was given a proper name.

#

Downside is it doesn't return early on the name comparison

#

If it's a marginal performance gain then it's better to sacrifice it for readability

brazen charm Nov 1, 2019, 10:27 PM

#

the default tuple feels a bit hacky to me for some reason

tawdry vapor Nov 1, 2019, 10:30 PM

#

I think it's OK

sleek mason Nov 2, 2019, 12:05 AM

#

when i try to git push --set-upstream origin zen-command i get myself a jolly good ```
ERROR: Permission to python-discord/bot.git denied to kingdom5500.
fatal: Could not read from remote repository.

Please make sure you have the correct access rights
and the repository exists.

am i forgetting anything or has something changed? :D

tawdry vapor Nov 2, 2019, 12:06 AM

#

That's trying to push to our repo rather than a fork

#

Maybe you were not given appropriate permissions on github

#

I cannot check for you though. Maybe @patent pivot can

patent pivot Nov 2, 2019, 12:07 AM

#

ay up

#

will have a look

#

nope I don't see anything @sleek mason

#

you are in all the required teams

#

plus it's a read error, not a write one

sleek mason Nov 2, 2019, 12:09 AM

#

how queer.

patent pivot Nov 2, 2019, 12:09 AM

#

this isn't a permissions error, it's a key thing presumably

molten bough Nov 2, 2019, 12:09 AM

#

You shouldn't need the set-upstream I'd have thought

#

Ah, yeah, check ssh keys

tawdry vapor Nov 2, 2019, 12:10 AM

#

You need it when you create a new branch

patent pivot Nov 2, 2019, 12:10 AM

#

keys are valid though

#

since it has ident'd you as kingdom5500

molten bough Nov 2, 2019, 12:10 AM

#

2fa?

patent pivot Nov 2, 2019, 12:10 AM

#

keys bypass it

tawdry vapor Nov 2, 2019, 12:10 AM

#

I thought ssh disregards 2fa

patent pivot Nov 2, 2019, 12:11 AM

#

yeah it does

molten bough Nov 2, 2019, 12:11 AM

#

I mean you need it on your account right?

sleek mason Nov 2, 2019, 12:11 AM

#

i have it.

molten bough Nov 2, 2019, 12:11 AM

#

Ah okay

sleek mason Nov 2, 2019, 12:11 AM

#

ssh git@github.com approved: Hi kingdom5500! You've successfully authenticated, but GitHub does not provide shell access.

patent pivot Nov 2, 2019, 12:11 AM

#

📎 unknown.png

#

hm

#

not sure what is going on

sleek mason Nov 2, 2019, 12:12 AM

#

i have no idea either.

tawdry vapor Nov 2, 2019, 12:12 AM

#

can you do

#

git remote -v

sleek mason Nov 2, 2019, 12:12 AM

#

origin  git@github.com:python-discord/bot.git (fetch)
origin  git@github.com:python-discord/bot.git (push)

#

looks fine and spicy to me

molten bough Nov 2, 2019, 12:12 AM

#

Yup

tawdry vapor Nov 2, 2019, 12:12 AM

#

Yeah... it does

#

does git fetch origin work?

sleek mason Nov 2, 2019, 12:13 AM

#

yup.

tawdry vapor Nov 2, 2019, 12:14 AM

#

I'm not sure

#

Try creating a new ssh key and using that?

patent pivot Nov 2, 2019, 12:16 AM

#

hold on

#

you arne't in the staff team

#

what

#

what tteam are you in then

#

you are only on the esoteric challenges team

#

hm

molten bough Nov 2, 2019, 12:17 AM

#

He should be a contrib I think?

#

Unless helpers go on the staff team now

patent pivot Nov 2, 2019, 12:17 AM

#

should be staff

#

they do

molten bough Nov 2, 2019, 12:17 AM

#

Aha

patent pivot Nov 2, 2019, 12:17 AM

#

fixed @sleek mason

sleek mason Nov 2, 2019, 12:17 AM

#

oh right okay cool

#

i'm just glad it wasn't my fault this time

patent pivot Nov 2, 2019, 12:18 AM

#

not sure why that happened, could've borked during transition to staff team i guess

#

lol

sleek mason Nov 2, 2019, 12:18 AM

#

oh well haha

patent pivot Nov 2, 2019, 12:19 AM

#

¯_(ツ)_/¯

📎 unknown.png

molten bough Nov 2, 2019, 12:23 AM

#

Big tangent in #python-discussion if anyone is around

#

It's not technically rule breaking I think? So don't wanna ping mods

green oriole Nov 2, 2019, 8:39 AM

#

I don't know about you but it looks like a part of the tag is missing here

#

!f-strings

stable mountainBOT Nov 2, 2019, 8:39 AM

#

In Python, there are several ways to do string interpolation, including using %s's and by using the + operator to concatenate strings together. However, because some of these methods offer poor readability and require typecasting to prevent errors, you should for the most part be using a feature called format strings.

In Python 3.6 or later, we can use f-strings like this:

snake = "Pythons"
print(f"{snake} are some of the largest snakes in the world")

In earlier versions of Python or in projects where backwards compatibility is very important, use str.format() like this:

snake = "Pythons"

# With str.format() you can either use indexes
print("{0} are some of the largest snakes in the world".format(snake))

# Or keyword arguments
print("{family} are some of the largest snakes in the world".format(family=snake))

green oriole Nov 2, 2019, 8:40 AM

#

Oh well

#

It didn't do it this time

sleek mason Nov 2, 2019, 5:56 PM

#

thanks for the review btw @exotic ember, changes have been made :D

hollow lichen Nov 2, 2019, 5:59 PM

#

What's the difference between and standard endpoint and an extended one?

hardy gorge Nov 2, 2019, 6:01 PM

#

For infractions? The data it includes in the response

#

If you look in the viewset for infractions, you can see that a normal request only gives you back the user ids for actor and user

#

    #### Response format
    >>> [
    ...     {
    ...         'id': 5,
    ...         'inserted_at': '2018-11-22T07:24:06.132307Z',
    ...         'expires_at': '5018-11-20T15:52:00Z',
    ...         'active': False,
    ...         'user': 172395097705414656,
    ...         'actor': 125435062127820800,
    ...         'type': 'ban',
    ...         'reason': 'He terk my jerb!',
    ...         'hidden': True
    ...     }
    ... ]

hollow lichen Nov 2, 2019, 6:03 PM

#

yes, I would like some help with https://github.com/python-discord/bot/pull/508 because I'm drowning. I looked at the post_infraction code, but don't understand where set an extended endpoint

GitHub

Add the current nickname of a user in infraction search by kraktus...

If the user has not changed its nickname, display None, otherwise display its current nickname.
closes #337

hardy gorge Nov 2, 2019, 6:03 PM

#

In the extended endpoint, you'd get back the full user information for the actor and user field

#

So, the 'user' field now gives you this information directly:

    #### Response format
    >>> {
    ...     'id': 409107086526644234,
    ...     'avatar': "3ba3c1acce584c20b1e96fc04bbe80eb",
    ...     'name': "Python",
    ...     'discriminator': 4329,
    ...     'roles': [
    ...         352427296948486144,
    ...         270988689419665409,
    ...         277546923144249364,
    ...         458226699344019457
    ...     ],
    ...     'in_guild': True
    ... }

#

That means you don't have to make a second query if you're interested in more than just the user's id

#

You can access expanded routes by modifying the endpoint url:

    ### Expanded routes
    All routes support expansion of `user` and `actor` in responses. To use an expanded route,
    append `/expanded` to the end of the route e.g. `GET /bot/infractions/expanded`.

hollow lichen Nov 2, 2019, 6:06 PM

#

instead of

await self.bot.api_client.get(
            'bot/infractions',
            params={'user__id': str(user.id)}
        )
``` I should put `expanded` somewhere?

#

ty, I was close then 🙂

hardy gorge Nov 2, 2019, 6:06 PM

#

The endpoints are documented in the viewset files over at the site repo

#

This is the one for infractions: https://github.com/python-discord/site/blob/master/pydis_site/apps/api/viewsets/bot/infraction.py

hollow lichen Nov 2, 2019, 6:08 PM

#

ok

#

I added expanded and it looks like it break management.py

#

TypeError: unhashable type: 'dict'

green oriole Nov 2, 2019, 6:30 PM

#

Can you paste the full traceback?

hollow lichen Nov 2, 2019, 6:33 PM

#

it's the full one

#

unfortunately

green oriole Nov 2, 2019, 6:34 PM

#

Oh

#

No line number?

hollow lichen Nov 2, 2019, 6:38 PM

#

nop, the error_handler does not help 😦

hardy gorge Nov 2, 2019, 6:57 PM

#

Did you change the endpoint for the whole cog? If so, you'd need to change all parts that rely on the response. My suspicion is that various parts now expect a bare ID, not a dict.

#

Now, @tawdry vapor is working on two moderation related PRs that change quite a bit, so changes throughout the cog may lead to problematic merge conflicts.

tawdry vapor Nov 2, 2019, 7:02 PM

#

Well they're actually done; just need review

#

In any case I do not mind resolving conflicts if it comes down to that

#

Though I did not touch the management cog IIRC

hollow lichen Nov 2, 2019, 7:03 PM

#

I have to say I'll gonna run out of time for finishing https://github.com/python-discord/bot/pull/508

GitHub

Add the current nickname of a user in infraction search by kraktus...

If the user has not changed its nickname, display None, otherwise display its current nickname.
closes #337

tawdry vapor Nov 2, 2019, 7:03 PM

#

What do you mean?

hollow lichen Nov 2, 2019, 7:03 PM

#

I'll no more have time to deal with it

#

Next week I'll be very busy IRL, and I also got other projects in python.

#

I just want to warn I may not be able to finish this PR

brazen charm Nov 2, 2019, 8:05 PM

#

How should exceptions be handled? Did this https://paste.fuelrats.com/narekeramu.py for https://github.com/python-discord/bot/issues/381 but not sure if it's correct, with the scope of the handling and how it's logged

jade tiger Nov 3, 2019, 3:36 AM

#

Hey hey, not being pushy at all, I know its a large-ish PR, but wondering if there was any word on https://github.com/python-discord/bot/pull/519? 😃

GitHub

Refactor the !help command. by mathsman5133 · Pull Request #519 ...

redirect_output has been adjusted to run the delete_invocation inside a task as the help command will wait for that to run before sending the help or doing anything else.

pagination has been adju...

exotic ember Nov 3, 2019, 3:41 AM

#

not sure I can do a full review, but I'll point out stuff I see

jade tiger Nov 3, 2019, 3:41 AM

#

👌 cheers

#

also, assuming the deps automatically use d.py 1.2.4 (~=1.2 should use the latest version, 1.2.4?),

https://github.com/python-discord/bot/blob/master/bot/patches/message_edited_at.py

patch got fixed in commit

https://github.com/Rapptz/discord.py/commit/6ed5a21a78182bfa31e3757866bf3975af4b606c

tawdry vapor Nov 3, 2019, 3:47 AM

#

If we were to rely on that we'd have to change the pinned version anyway

#

Otherwise, I think existing venvs won't be updated

#

since 1.2.3 would technically still match the criteria

#

but I don't know the details of how pipenv locks

#

Actually yeah cause pipenv locks, it will lock to 1.2.4 anyway

#

I think what I am saying is only relevant in like a requirements.txt, which we don't use

#

But would still be a good idea to make it >=1.2.4, ==2.*

hardy gorge Nov 3, 2019, 7:37 AM

#

The way we pin should allow for patch upgrades, but we do need to relock the environment. Regardless, people will have to update their local env with the new lockfile.

#

So, we should not have to update the dependency pinning in the Pipfile.

jade tiger Nov 3, 2019, 9:10 AM

#

thanks kosa, points all make sense, couple I'm not sure how I didn't pick up 😂

#

I'll wait til it's all in so I can do a bigger commit rather than dribs and drabs

molten bough Nov 3, 2019, 12:28 PM

#

This isn't really the right channel @mental edge, I think you want #303934982764625920

mental edge Nov 3, 2019, 12:28 PM

#

oh, I see!

molten bough Nov 3, 2019, 12:29 PM

#

This channel is for discussion of projects related to this specific community

mental edge Nov 3, 2019, 12:29 PM

#

my apologies and thanks for letting me know!

molten bough Nov 3, 2019, 12:29 PM

#

All good, all good

hollow lichen Nov 3, 2019, 12:52 PM

#

DeprecationWarning: The object should be created from async function resolver=AsyncResolver(),

#

DeprecationWarning: The object should be created from async function
  family=socket.AF_INET,

#

This is from __main__.py

clever wraith Nov 3, 2019, 2:12 PM

#

So, sort of related... A while back I wrote a Chrome/Vivaldi plugin for Hastebin (what Python-Discord's Code Uploader is powered by). Not sure if anyone wants to use it or maybe modify it to be more specific to this community, but if anyone's interested here's the link:
https://chrome.google.com/webstore/detail/hastebin/jnjbjegfekjhdodgjblogcemdmdajocl
And it's open source:
https://bitbucket.org/GeoDoX/hastebin/src/master/

Hastebin

This extension automatically generates and opens a Hastebin link.

crude gyro Nov 3, 2019, 2:47 PM

#

...related to what?

#

feels off topic for this channel. maybe you meant to put it in #303934982764625920

#

as for the paste service, we're writing our own version of that already, @patent pivot is more or less done.

patent pivot Nov 3, 2019, 2:49 PM

#

yep, working on CI stuff now

molten bough Nov 3, 2019, 2:55 PM

#

Oh nice, will that be under django as well?

patent pivot Nov 3, 2019, 3:27 PM

#

No, flask

#

Separate to pydis site

#

I hadn't learnt django when I started writing it

exotic ember Nov 3, 2019, 3:27 PM

#

open source when 😠 😠 😠

patent pivot Nov 3, 2019, 3:27 PM

#

plus it's just very simple

#

soon tm

exotic ember Nov 3, 2019, 3:28 PM

#

!sooncounter

molten bough Nov 3, 2019, 3:29 PM

#

Aha okay

tawdry vapor Nov 3, 2019, 6:30 PM

#

@hardy gorge Yeah, not a concern with pipenv perhaps. But what I was saying that generally, not pinning the patch version I think means that if pip (?) needed to, it would downgrade/not use the latest discord.py patch version since technically it'd be allowed by the version specification. 🤔

hardy gorge Nov 3, 2019, 8:03 PM

#

Yeah, that's true, I guess. It will obviously only do that during a re-lock, since everyone should sync --dev from the lock file, but it's something that could happen when we relock for something else.

More generally, though, isn't this the same thing that could happen with any of the things we've pinned to a patch version?

molten bough Nov 3, 2019, 8:06 PM

#

It's true, but the point of the specifier is that the versions should be compatible

#

Although we've had problems with misbehaving devs before..

hardy gorge Nov 3, 2019, 8:08 PM

#

I guess the point is that we'd updating the patch version specifically for a bug fix we'd then rely on. If we'd then resolve down again without noticing it, we'd go back to a version that has the bug we used to have a work-around for.

tawdry vapor Nov 3, 2019, 8:13 PM

#

Yes but generally we don't care about patch versions

#

In this case we'd be directly relying on a fix in that version

crude gyro Nov 3, 2019, 8:41 PM

#

@flat marsh
@jade tiger
@dusky bloom
@lapis hull

Hey guys. For your contributions during the Hacktoberfest event, you've been awarded the <@&295488872404484098> role! That means you've been added to the Contributors team in our GitHub organisation and can enjoy some new permissions in some of our repos.

You now have write permissions on bot, branding, site, seasonalbot and flake8-annotations. This doesn't give you the right to commit to the master branch and pull requests still need approval from core developers, but you don't have to work in a fork anymore.

You'll also find you have a colorful new role. Please make sure you read through the Contribution Guidelines, which can be found here: https://github.com/python-discord/meta/blob/master/CONTRIBUTING.md

Thanks a million for contributing to our repos during the event and congratulations.

GitHub

python-discord/meta

Issue tracker for suggestions and other questions relating to our community - python-discord/meta

jade tiger Nov 3, 2019, 8:50 PM

#

Cheers 👌 lots of fun, learnt a lot

lapis hull Nov 3, 2019, 9:14 PM

#

Woohoo, thanks! It was really satisfying to be able to contribute in even a real small way to something that's helped me pystrong

hollow lichen Nov 4, 2019, 4:35 PM

#

Hey, sorry to ask, but could someone review my PR? https://github.com/python-discord/bot/pull/500 I've been waiting 15 days 😉

GitHub

Token and bad code by kraktus · Pull Request #500 · python-disco...

This PR fix #389
Split the token_remover in two functions, one that checks that there is a token in the message (is_token_in_message), and the other takes the actions (deleting the message, etc.)
T...

molten bough Nov 4, 2019, 4:36 PM

#

I've been waiting a long time too, but don't worry, they'll get around to it :>

#

They have tons to do

mellow hare Nov 4, 2019, 4:37 PM

#

Looking at the PR it still looks like they're waiting on some requested changes

#

You need to make sure you mark those as complete if they are

hollow lichen Nov 4, 2019, 4:38 PM

#

Ofc! Just saying in case developers are also waiting for me to change things ^^

#

I forgot to resvolve one requested change

mellow hare Nov 4, 2019, 4:43 PM

#

I'm going to update the branch real quick so it's in sync with the master branch, so you might want to pull after I do that. I'll try and find time to look at it soon after

brazen charm Nov 4, 2019, 5:23 PM

#

looks like https://github.com/python-discord/bot/issues/636 can be closed

mellow hare Nov 4, 2019, 5:24 PM

#

Good spotting it

green oriole Nov 4, 2019, 5:50 PM

#

Everyone is waiting for reviewing 😄 We did not make it easy 😉

brazen charm Nov 4, 2019, 6:02 PM

#

you can also review 😛

#

Been looking on one issue that mentions a proxy user object but I've found 2 functions that create them that differ a bit in behaviour, is that intentional? Feels like they could share single function for creating them
one in moderation/utils https://github.com/python-discord/bot/blob/master/bot/cogs/moderation/utils.py#L33-L48
and the second one in watchchannels/watchchannel https://github.com/python-discord/bot/blob/master/bot/cogs/watchchannels/watchchannel.py#L27-L40

woeful thorn Nov 4, 2019, 6:19 PM

#

Probably not intentional, no

nocturne hare Nov 5, 2019, 3:22 PM

#

Tried logging in on pydis.com, putting this here

📎 unknown.png

mellow hare Nov 5, 2019, 3:22 PM

#

That's odd

#

Yeah that's really strange, I'm unsure why you're getting that. Working fine on mine

nocturne hare Nov 5, 2019, 3:24 PM

#

https://discordapp.com/oauth2/authorize?client_id=409107086526644234&redirect_uri=https%3A%2F%2Fpydis.com%2Faccounts%2Fdiscord%2Flogin%2Fcallback%2F&scope=identify+email&response_type=code&state=N37TCTVst2WE

hardy gorge Nov 5, 2019, 3:26 PM

#

@nocturne hare Okay, apparently it's an issue wit hpydis.com versus pythondiscord.com

#

You can log in using the latter, but not the former

#

Can you make an issue on the site repo?

brazen charm Nov 5, 2019, 3:28 PM

#

if I'm copying the behavior of this in a new class

    user.avatar_url_as = lambda static_format: None

would this be fine or should I specify the argument like above or the same ones as User.avatar_url_as takes?

    @staticmethod
    def avatar_url_as(*_, **__) -> None:
        return None

nocturne hare Nov 5, 2019, 3:33 PM

#

Was originally looking for stats.pydis.com is that gone now?

#

Also made a poorly formatted issue, but at least documented

mellow hare Nov 5, 2019, 3:41 PM

#

It's weird to me that some classes inherit from parent classes but don't invoke the super().__init__()

woeful thorn Nov 5, 2019, 3:44 PM

#

Why?

mellow hare Nov 5, 2019, 4:31 PM

#

I don't know. I think it's because I feel like there are some necessary bits that aren't being properly initialized if it isn't called.

#

I mean I know that's not actually the case but I just kind of associated it as just being part of the process

molten bough Nov 5, 2019, 5:00 PM

#

Oh, I never thought about the domain thing

#

Yeah whoever has access to the oauth app will need to add another url

#

but uh

#

maybe that domain should redirect

#

because mixing domains will give you two sets of cookies

mellow hare Nov 5, 2019, 5:05 PM

#

stats.pythondiscord.com redirects to pydis.com for some reason

molten bough Nov 5, 2019, 5:05 PM

#

Does the webhook no longer post issue comments or is it just slow?

#

ah okay, there it is

mellow hare Nov 5, 2019, 6:35 PM

#

Logistical question

#

I'm working on the issue where we're going to suggest someone uses the hastebin instead of muting them for huge blocks of code. My question is, does it make more sense to put it in the anti-spam cog itself or in the rule/newlines.py file?

#

Oh well crap, I just thought of something

#

I don't know if I can do what I wanted to do

molten bough Nov 5, 2019, 6:37 PM

#

I believe!

woeful thorn Nov 5, 2019, 6:38 PM

#

It shouldn't go in the rule

mellow hare Nov 5, 2019, 6:38 PM

#

I was going to use the codeblock_stripping() method from the bot/cogs/bot.py file but I guess I can't do that if it's in the class and not a class or static method? And yeah, fair point

#

I'll make it fit in the anti-spam cog

#

Likely just add an additional check to see if the rule that's being called out is the newlines one and then toss it to a function that handles the check for a codeblock

#

But now I am kind of trying to work out about the accessibility of the method

woeful thorn Nov 5, 2019, 6:40 PM

#

If there's no reason they can't be static methods, then you can make them static methods

mellow hare Nov 5, 2019, 6:42 PM

#

I'll have to make 3 of them into static methods, but that doesn't seem like it'll cause any particular conflicts from what I'm seeing...

#

Okay, yeah I think that's what I'll have to do. Thanks, I guess I just get scared to change the hard work that everyone has put into this thing. Need to Hem up and do this

#

Actually one last question: What would make more sense here, a class method or a static method? Because my hesitation is due to it referencing other methods in the class. I would think in this case a class method makes more sense...

#

Yeah, class method

#

Okay, got it

#

Sorry, just had to work it out

brazen charm Nov 5, 2019, 7:04 PM

#

to make sure I don't mess up anything, how do I sync my branch (with a pr open) to the master? The things I need to edit moved so the current conflict can be thrown away

mellow hare Nov 5, 2019, 7:05 PM

#

With the git command line tool?

#

Or what are you using

brazen charm Nov 5, 2019, 7:06 PM

#

got the git gui but never worked with is so probably command line

mellow hare Nov 5, 2019, 7:07 PM

#

git checkout master
git pull
git checkout [branch_name_here]
git merge master

brazen charm Nov 5, 2019, 7:08 PM

#

Do I make a commit to throw away the conflict or how does it work there? (Gotta reset my pc so hang on a minute)

mellow hare Nov 5, 2019, 7:09 PM

#

I'm not sure. I do most of my git stuff through PyCharm and have been spoiled by it

brazen charm Nov 5, 2019, 7:11 PM

#

Have pycharm so could try that too, only used it to commit so far though

mellow hare Nov 5, 2019, 7:12 PM

#

In PyCharm, in the top right hand corner you'll have some git icons.

📎 unknown.png

#

Hold on, making sure I'm going to say all this correctly

#

Oh okay, forget that. Bottom right of the screen.

#

You'll have Git: Your Branch Here

#

If you select that, you'll be able to select the origin/master branch, and then you can select Merge with Current

#

That'll update your current branch as well as prompt you to resolve any conflicts

brazen charm Nov 5, 2019, 7:15 PM

#

oh that's neat

mellow hare Nov 5, 2019, 7:15 PM

#

Like I said, I have been spoiled to death by this

brazen charm Nov 5, 2019, 7:16 PM

#

Thanks

mellow hare Nov 5, 2019, 7:16 PM

#

Happy to help

brazen charm Nov 5, 2019, 7:16 PM

#

Guess I'll also look into pycharm git tools too

mellow hare Nov 5, 2019, 7:18 PM

#

The most useful thing I found (and I'm not sure if this is in the community version or not), but if you select VCS > Git > View Pull Requests, it makes it stupid easy to compare and checkout current PRs

green oriole Nov 5, 2019, 7:21 PM

#

@brazen charm if you have a conflict git will abort the merge, go in a conflict state and ask you to fix the merge in those file. If you open the file you are going to see something like

<<<<<<< HEAD 
lemon = lie
======= 
lemon = ai
>>>>>>> lemon-true-origin```You need to delete all those lines and replace it with the most relevant things for you (you can even mix them like `lemon = (lie, ai)`). Then you commit your changes and redo the comment that triggered the conflict.

#

Or you can cheat and use a GUI 😄

brazen charm Nov 5, 2019, 7:26 PM

#

git is nice until I have to actually do something and then it's scary because I'm thinking I'll break something 😄

green oriole Nov 5, 2019, 7:26 PM

#

If you break something that's okay, git history is rewritable

last patio Nov 5, 2019, 7:27 PM

#

yeah but don't do thatl ol

#

rewriting history on a repository that others use is really unfun

green oriole Nov 5, 2019, 7:27 PM

#

Yeah avoid breaking git as most as possible

last patio Nov 5, 2019, 7:27 PM

#

but you can solve pretty much everything without rewriting history

green oriole Nov 5, 2019, 7:27 PM

#

If you haven't pushed yet rewriting is okay

brazen charm Nov 5, 2019, 7:27 PM

#

at least I can hide in my fork

green oriole Nov 5, 2019, 7:28 PM

#

Well, it will appear on the PR and on the pydis repo after 😄

#

Rewriting is useful sometime, like it avoid having to download UI-toolkit everytime you clone a repo 🙃 😄

mellow hare Nov 5, 2019, 7:40 PM

#

Please don't suggest rewriting git history as a practical solution. It's really really not

green oriole Nov 5, 2019, 7:55 PM

#

I meant.. Only rewrite if you haven't pushed and there is no other way around

#

If not.. Don't do it

#

Technically, we suggest using amend in the contribution guidelines, but it rewrite the history

hardy gorge Nov 5, 2019, 8:27 PM

#

The main point, like you say, is to get your history tidy before you push (or open a PR from a fork).

#

I use -amend all the time, almost on every PR, but not after I've pushed

mellow hare Nov 5, 2019, 9:00 PM

#

Fair enough

mellow hare Nov 6, 2019, 4:24 PM

#

Questioning my idea on how to tackle this issue. The methods in the bot.py file are overkill for what is needed for this. I'm trying to figure out a reasonable way to pare it down

glass pecan Nov 6, 2019, 4:24 PM

#

what issue?

mellow hare Nov 6, 2019, 4:24 PM

#

Oh ah.... one sec, snagging

#

https://github.com/python-discord/bot/issues/439

GitHub

Auto suggest the use of pydis hastebin on longer code messages · ...

When an user posts a codeblock that takes up a majority of the discord message view, it can become difficult to view the code while also communicating with the helped user which usually results in ...

molten bough Nov 6, 2019, 4:25 PM

#

I wonder if it'd be nice if one of the bots could post issue info if you said eg bot#439 in this channel

mellow hare Nov 6, 2019, 4:25 PM

#

It's the one where I'm having it suggest using the hastebin rather than muting the user, yeah

#

Well the issue is that I'd never remember what the issue number I'm working on is

molten bough Nov 6, 2019, 4:26 PM

#

PyCharm can track it

exotic ember Nov 6, 2019, 4:26 PM

#

.issue 439 bot

mellow hare Nov 6, 2019, 4:26 PM

#

So my initial idea was to just use the methods from the bot.py file. However those methods are for specifically detecting ill formatted code, not any python code in general

dusky shoreBOT Nov 6, 2019, 4:26 PM

#

GitHub

IssueOpen [bot] #439 Auto suggest the use of pydis hastebin on longer code messages

exotic ember Nov 6, 2019, 4:26 PM

#

although another command with extended info would be good

#

sorry hem

mellow hare Nov 6, 2019, 4:27 PM

#

You're good

hardy gorge Nov 6, 2019, 4:30 PM

#

@green oriole How about detecting if a mention was used to trigger the embed and only if it was invoked for another user and a mention was not used, but an id, it will add a ping outside of the embed?

green oriole Nov 6, 2019, 4:32 PM

#

I was thinking of something like that, I was just searching if you can easily detect if d.py converted it from an ID or a mention. I guess we can just use a string input and handle the rest ourselves?

hardy gorge Nov 6, 2019, 4:32 PM

#

I think it's easier to just look at the ctx.message.mentions list and see if it's empty or not if a target was used

#

if user and not ctx.message.mentions

green oriole Nov 6, 2019, 4:37 PM

#

Yeah that make more sense

mellow hare Nov 6, 2019, 4:37 PM

#

I feel like I'm on the cusp of an idea

#

log.trace() is only written when you're in debug mode right?

green oriole Nov 6, 2019, 4:41 PM

#

Don't think it is

hollow lichen Nov 6, 2019, 4:41 PM

#

.bot init

#

.bot __init__

#

sad it doesn't give a link

green oriole Nov 6, 2019, 4:42 PM

#

I think I don't run my bot in debug mode but I see the trace @mellow hare

hollow lichen Nov 6, 2019, 4:42 PM

#

trace level is 5, so I think it's above debug

mellow hare Nov 6, 2019, 4:42 PM

#

Gotcha gotcha

exotic ember Nov 6, 2019, 4:42 PM

#

@hollow lichen what do you mean by it doesn't give a link?

hardy gorge Nov 6, 2019, 4:43 PM

#

TRACE is below DEBUG

#

We define TRACE to have level 5 and DEBUG is 10.

hollow lichen Nov 6, 2019, 4:43 PM

#

It could give a link to the file named __init__ in the bot repo

hardy gorge Nov 6, 2019, 4:43 PM

#

These are the built-in levels: https://docs.python.org/3/library/logging.html#levels

#

TRACE is the lowest level we use, used to trace things through the code

green oriole Nov 6, 2019, 4:44 PM

#

BTW can we make (I don't thing we can but..) a type hint for trace so IDEs stop complaining about trace not being defined?

#

And for bot.api_client?

hollow lichen Nov 6, 2019, 4:45 PM

#

Thanks for the clarification, I thought each log level was separated by 1 level

hardy gorge Nov 6, 2019, 4:46 PM

#

IDEs complaining? Sublime and PyCharm both don't complain for me

#

I'm not sure that a type hint would help and we don't use type hints outside of function annotations

hollow lichen Nov 6, 2019, 4:46 PM

#

Sublime doesn't complain for me either, but I don't have a lot of plugins

green oriole Nov 6, 2019, 4:47 PM

#

Hmm PyCharm complain for me

#

It say something like no member api_client for Bot

#

Anyway

#

Should I edit the issue?

hardy gorge Nov 6, 2019, 4:49 PM

#

Yes, that's fine

mellow hare Nov 6, 2019, 4:53 PM

#

So I think I know what's partially throwing me off.

#

We have a try block that spans just under 100 lines

molten bough Nov 6, 2019, 4:55 PM

#

that is a long try block

mellow hare Nov 6, 2019, 4:55 PM

#

And I think it only needs to cover like

#

1 specific line

#

https://github.com/python-discord/bot/blob/master/bot/cogs/bot.py#L243 Then alllll the way down to line 341 before we see the except

GitHub

python-discord/bot

Source code for our Discord bot. Contribute to python-discord/bot development by creating an account on GitHub.

#

If I'm understanding correctly, it really only needs to be on line 287

green oriole Nov 6, 2019, 5:00 PM

#

Yeah that's wierd. You can try logging which line the SyntaxError occurred and do some testing if you want to make sure only that line should be covered

mellow hare Nov 6, 2019, 5:02 PM

#

Yeah might make a note about it but honestly, I'm mainly trying to distract myself from not being sure how to fix this.

#

The solution I have in mind really only works if their code is formatted correctly, i.e. using the backticks, having py or python to make discord's markup trigger, that stuff

#

I guess my real question/call for help is, what part of the on_message() listener in the bot.py file tells it to actually consider it valid Python code

hollow lichen Nov 6, 2019, 5:08 PM

#

This line actually tells if it's python or not: py tree = ast.parse(content[0])

mellow hare Nov 6, 2019, 5:09 PM

#

That's what I thought

green oriole Nov 6, 2019, 5:09 PM

#

The try except block isn't responsible to detect if it is valid code?

hollow lichen Nov 6, 2019, 5:09 PM

#

However I don't think it's the good way to test if the code is python, since any SyntaxError cancel it.

green oriole Nov 6, 2019, 5:10 PM

#

If an exception is raised it mean that it is not python code

#

So if no exception is raised, it mean it is a pythonic code

#

So just add an else after the try except?

#

(and move the if before the try except)

hollow lichen Nov 6, 2019, 5:13 PM

#

I also don't understand what does DEBUG_MODE has to do with this part of the code.

mellow hare Nov 6, 2019, 5:13 PM

#

No idea, it's beyond the scope of what I'm trying to do

hollow lichen Nov 6, 2019, 5:14 PM

#

My bad, it's always trigged when in debug mode, the not is not for the two conditions.

#

this block would need to be flatted anyway.

green oriole Nov 6, 2019, 5:26 PM

#

The DEBUG_MODE is messad up, it deactivate some feature, enforce others

#

We should have a better debug config

hardy gorge Nov 6, 2019, 5:28 PM

#

Define "messed up", since I don't have issues with it

#

We're already planning on removing the conditional loading of docs and verification

green oriole Nov 6, 2019, 5:29 PM

#

Like, it disables all filters but enforce the codeblock warning

#

There isn't any logic in what the debug mode changes

hardy gorge Nov 6, 2019, 5:30 PM

#

it does not disable filters, it restricts them to one channel

#

There was a logic to that, but recent changes removed the necessarity for those things

#

I think we have an issue asking for those to be removed

#

The reason was that we used to have a special testing channel on this server and wanted to make sure the testing bot did not activate things it shouldn't active

#

That channel was removed two weeks ago

green oriole Nov 6, 2019, 5:33 PM

#

Actually I opened a PR removing this

#

I think it has already been merged

hardy gorge Nov 6, 2019, 5:35 PM

#

The filters, right?

#

Not the other things

green oriole Nov 6, 2019, 5:35 PM

#

Yeah only on the filters

#

About the debug mode, like, it disable some feature, enforce some, restrict others..

#

I think we could have something more customizable as it isn't very user-friendly (when I started contributing I spend half an hour trying to understand what was going on)

hardy gorge Nov 6, 2019, 5:36 PM

#

Everything related to the old dev-test channel can be removed, since we don't use a staging bot on this server anymore

mellow hare Nov 6, 2019, 5:40 PM

#

@woeful thorn I asked yesterday about whether my changes to the newline spam thing should go in the rules or in the antispam thing itself, and I was wondering if you could explain why it shouldn't be in the rules. Because to me, this is a modification of the rules itself, and this could be most easily handled there

hardy gorge Nov 6, 2019, 5:41 PM

#

What are you changing?

mellow hare Nov 6, 2019, 5:41 PM

#

If there's a "too many newlines" spam detected, but it's Python code, then instead of muting them it tells them to use the hastebin

#

With how we have the antispam cog, all of those kinds of checks about what kind of rule is being violated is being delegated out to the individual rule files

#

And that's where it's easiest to control whether a given one should be counted as a rule violation or not

woeful thorn Nov 6, 2019, 5:47 PM

#

The behavior of sending it to pastebin should not be in the rules, which is what I assumed you were asking about

tawdry vapor Nov 6, 2019, 5:52 PM

#

Hi I have a PR that's pretty fast to review https://github.com/python-discord/bot/pull/619

GitHub

Moderation logging by MarkKoz · Pull Request #619 · python-disco...

#618 must be merged first.
The moderation modules were a bit bare in the logging. Most of the new logging is trace logging, but there are some debugs and warnings too.

mellow hare Nov 6, 2019, 5:52 PM

#

Not sending it there, merely telling them to use it instead

#

But you're still probably right, as we still want it to go through the deleting process... I think I just under-estimated what all would be needed in it along with what I knew how to do

gusty sonnet Nov 7, 2019, 4:09 AM

#

Hmm, question about log.info, my pylint has been screaming about not using f-string for it

#

it screams at me for thispy log.info(f"Re-applied {infraction['type']} to user {infraction['user']} upon rejoining.")and want me to do thispy log.info("Re-applied %s to user %s upon rejoining.", infraction['type'], infraction['user'])is there a convention / consensus on this?

#

I found this SO https://stackoverflow.com/a/34634301 and to quote from it

Stack Overflow

PyLint message: logging-format-interpolation

For the following code:

logger.debug('message: {}'.format('test'))
pylint produces the following warning:
logging-format-interpolation (W1202):

Use % formatting in logging functions and p...

#

Formatting of message arguments is deferred until it cannot be avoided. However, computing the arguments passed to the logging method can also be expensive, and you may want to avoid doing it if the logger will just throw away your event.```

#

Does that mean the latter is better?

molten bough Nov 7, 2019, 10:31 AM

#

@gusty sonnet Basically what it's saying is that if you let the logger do the formatting, then log messages that will be discarded (eg the level is debug bur you just want info and up) will not actually be formatted, saving a little time

gusty sonnet Nov 7, 2019, 10:32 AM

#

I see, so it's like of an internal thing for logging, yes?

#

Do we have an overall consensus for this however, it's safe to disregard it right?

molten bough Nov 7, 2019, 10:34 AM

#

It's safe, yeah, and if we don't already lint for it then I guess you can just format ahead of time

#

But I guess it depends how the others feel

hardy gorge Nov 7, 2019, 10:36 AM

#

We typically prefer the former over the latter, since the increase in readability weighs more strongly than the small increase in efficiency the latter has for us.

#

I had the exact same question a couple of months ago and that was the consensus back then, anyway

green oriole Nov 7, 2019, 5:43 PM

#

If I commit a suggestion, it is going to trigger a CI build right?

crude gyro Nov 7, 2019, 5:44 PM

#

we do build all pull requests, yeah

#

if that's what you're asking

green oriole Nov 7, 2019, 5:44 PM

#

And what if I know committing the suggestion will fail the linting because I'm going to get an unused import?

brazen charm Nov 7, 2019, 5:45 PM

#

suggestion as in a review?

green oriole Nov 7, 2019, 5:45 PM

#

Yep

#

This one to be precise https://github.com/python-discord/site/pull/300#discussion_r342435083

GitHub

Create offensive message model for python-discord/bot#617 by Akary...

Hi, this PR add a new offensive message model needed for python-discord/bot#617.
This PR closes #222.

brazen charm Nov 7, 2019, 5:46 PM

#

ah, commiting and pushing for that will trigger it afaik

green oriole Nov 7, 2019, 5:48 PM

#

The problem is that the push is done by github, so I can't remove the import before

#

Should I apply the suggestion and commit after, or should I make the changes myself?

mellow hare Nov 7, 2019, 5:59 PM

#

I'd make the commit yourself.

#

Then it'll either detect that the proper change has been made and the suggestion will close or you close it yourself. Either way, its solved, and that's what matters

hardy gorge Nov 7, 2019, 6:03 PM

#

Yes. If you're able to use commit the suggestion using the GitHub-feature you should (to give the author of the suggestion some credit), but if that's not really possible, just make the changes locally.

mellow hare Nov 7, 2019, 6:04 PM

#

You should only use a suggestion if you feel it's an acceptable change and won't cause any issues.

#

At least in my eyes

woeful thorn Nov 7, 2019, 6:11 PM

#

Just add them as a co-author to your commit

#

That’s all GH does

#

https://help.github.com/en/github/committing-changes-to-your-project/creating-a-commit-with-multiple-authors

Creating a commit with multiple authors - GitHub Help

You can attribute a commit to more than one author by adding one or more Co-authored-by trailers to the commit's message. Co-authored commits are visible on GitHub.

mellow hare Nov 7, 2019, 6:12 PM

#

Oh huh, neat

hollow cedar Nov 8, 2019, 1:47 AM

#

@hardy gorge

We typically prefer the former over the latter, since the increase in readability weighs more strongly than the small increase in efficiency the latter has for us.
If it helps, we use the C/printf style string formatting for logging within CPython, but for 99% of practical purposes, I think it's fine to use f-strings. Particularly in cases where it affects the readability.

mellow hare Nov 8, 2019, 2:21 PM

#

I feel like I just have a mental block for the C style

#

The % stuff always just kind of messed with me

crude gyro Nov 8, 2019, 2:25 PM

#

I hate c-style interpolation.

#

it's gross.

molten bough Nov 8, 2019, 2:26 PM

#

I don't like it either

woeful thorn Nov 8, 2019, 2:26 PM

#

I think it’s great

#

And they should add sprintf and fprintf to stlib

crude gyro Nov 8, 2019, 2:29 PM

#

I think they should just conditionally allow print to have any letter prefixed or suffixed.

#

tprintx? why not.

molten bough Nov 8, 2019, 2:30 PM

#

haha

crude gyro Nov 8, 2019, 2:30 PM

#

how about some kprinty

molten bough Nov 8, 2019, 2:30 PM

#

pprintt

crude gyro Nov 8, 2019, 2:30 PM

#

hahaha

#

prounounced the same as "print" but in slow motion

molten bough Nov 8, 2019, 2:30 PM

#

iprintd

crude gyro Nov 8, 2019, 2:31 PM

#

did you now.