dev-contrib

[cold moon]

Their docs say gunicorn -k uvicorn.workers.UvicornWorker, I think this worker part is important.

[cold moon]

@eternal owl You can access to command arguments from Context, what is available in checks.

[eternal owl]

ooh the context.args, right, I forgot

lemme try

[patent pivot]

ahhh you're right, that is how it should be done. good research!

[eternal owl]

looks like ctx.args and ctx.kwargs dont really work with check(), I have to parse them from ctx.message.content, is it still worth using check?

@cold moon

[cold moon]

Huh, then not @eternal owl

[cold moon]

Why https://github.com/python-discord/public-stats/blob/master/.github/workflows/docker-publish.yml checkout code twice?

[cold moon]

@patent pivot I'm stuck with https://github.com/python-discord/forms-backend/pull/1 : Linting (flake8) have some parsing issues what raise errors, but in local device I don't get this error. Can you help to solve this issue? I tried to use specific actions and other ways, but they don't run flake8 extensions.

[patent pivot]

I'll have a gander, though I'm pretty busy today

[cold moon]

Okay

[cold moon]

@patent pivot Solved issue in staff lounge. Problem was that this linted venv. Now only thing that I have to do tomorrow is fix linting errors.

[patent pivot]

Ah perfect

[short snow]

https://github.com/python-discord/bot/issues/1300, hmmm no content issue

[subtle kraken]

taken care of

[short snow]

[short snow]

probably like the way ks123 suggested for bounty board

u could also add helpers and helpers + to that too

[eternal owl]

I don't think it makes a lot of difference

[obsidian patio]

I agree about that @eternal owl, though I’ll elaborate on my reasoning on the issue later on

[cold moon]

I think our Contributors is enough wise that they don't make flood PRs and reviews

[hardy gorge]

I think that, too

[green oriole]

Yeah, that's a pretty good point

[hardy gorge]

It's a tool to make contributions visible and the threshold of having the contributors role should be enough to make it work

Anyway, @mellow hare is taking a lead on this project and we discussed it yesterday during the admin meeting as well. I think we're going to trial it in one form or another. I don't really think that it will lead to so many problems. The staff-only reviewfest worked just fine as well.

[green oriole]

Well, this sounds more like the Hacktoberfest 2019 than 2020

[mellow hare]

I'll elaborate on it when I get to the office

[obsidian patio]

@hardy gorge could you leave your thoughts on the issue as well, so we have the thoughts on a single place?

[hardy gorge]

Huh?

[obsidian patio]

This

[hardy gorge]

This was something that was already extensively discussed; these issues are just about implementation, right?

[obsidian patio]

I think that’s very much about implementation, as well as if it’s even a good idea at all, to some extent. We haven’t really reached a consensus on either of the topics imo

[eternal owl]

do we want the leaderboard to be displayed in a channel which the bot edits as time goes on or will it be displayed at the end of each month in the #779073065560703007 channel or maybe even a command?

I was thinking, maybe we could display it at the top of #779073065560703007

and the bot constantly updates the message

[hardy gorge]

We haven’t really reached a consensus on either of the topics imo

We've already extensively discussed the idea within the admin meeting. We're going forward with it, although the exact shape may change.

I trust our contributors and staff enough to not act to the detriment of our projects. If they can't, they should not have either role.

[green oriole]

Honestly it would have been nice if core devs could have been kept in the loop, we are pretty much affected by those. The original proposal was about including everyone, wasn't it?

[hardy gorge]

This is just a playful way of both highlighting areas that need attention and highlighting the contributions people make.

There were various variants we discussed, yeah.

[obsidian patio]

I agree about the idea of trying to reward people for contributing fundamentally, but these kinds of systems often lead to unintended consequences

[hardy gorge]

Yes, they could. That's why we decided to restrict it to contributors and staff.

[obsidian patio]

I don’t quite understand why you’re so determined on the idea that you won’t reconsider it though

[green oriole]

This wasn't very clear haha, I thought it was open to everyone

[hardy gorge]

Yeah, it's good to discuss those concerns. There's no harm in voicing them.

But, ultimately, we also discuss things within the admin group and decide that we want to trial something in an area

We're not deaf for concerns, but we also want to move forward and trial some projects, especially if we'd otherwise shut everything down and/or get stuck in consensus hell

[obsidian patio]

I hear what you’re saying, but right now it has no consensus outside of the admin group at least. I fundamentally like the idea of rewarding people for contributions, but this one has major flaws that need to be addressed. We won’t get stuck in consensus hell by taking in the input from other contributors, staff members and core developers. Including people and making them feel like an important part of the system is one of the key pieces to keep people engaged long term

Once again, I’ll mention the importance of long term stability of the contributions. We’re not looking for people to peak with a high-activity period where they reach the top of the leaderboard. We’re looking for long term contributors who view others as collaborators and team mates — not competitors

[hardy gorge]

I think I expressed myself wrongly: We do appreciate input, but sometimes we do make decisions as the admin team regarding policy points we think are important.

That's a valid concern and we will take note of it. I don't think it's a given, as we are talking about a group of people who are known to us: contributors and staff members. We simply don't know if this is going to be a problem and how much of an issue it's going to be. I don't really believe it's going to be an issue at all, but if it is going to be an issue, a trial should show that.

[obsidian patio]

I think that’s a fair point. The issue is that what I’m referring to isn’t a one month thing. It may work in such a short time frame. The issue is if we consider the long term. If you vision a year into the future, do you think that we will have more great and consistent contributors? Will this leaderboard contribute to that?

[mellow hare]

You're right, and this is a failure on my end to properly communicate with the Core folks ahead of time. I do want you to know that I'm going to make sure this doesn't harm any critical or important parts of the repos. Anything that must be handled carefully will be marked as such

[hardy gorge]

Personally, I don't think it will change a lot; people who are motivated to contribute will still do so. I do think it will make frequent contributors more visible. It's not easy to see who's a frequent contributor to our projects without having a close look at the GitHub activity while it's very easy those who are very active in policy discussions in staff channels.

I do think the bounty board is a good tool to highlight issues and PRs that are currently of great interest to our projects.

[green oriole]

alright, I appreciate that hem

[hardy gorge]

Not all features have an equal priority and handpicking a few that need exposure is a good way of showcasing them here, within the community

(as opposed to relying people to visit GitHub, selects labels, and browse like that)

[mellow hare]

The main thing I wanted to get out of this was both calling attention to things that need to get done AND trying to get more members of the community involved with the process.

As I say all the time (probably to the point of being annoying), this community is first and foremost an education server. If I can get the opportunity to get people interested not only in PRs but also Code Reviews, then I think it's a step worth at least trying

Some people need that push to get going. Hopefully once they get a taste they'll be interested in the process and stick with it

[obsidian patio]

The issue I find with this logic is assuming that “motivation” is something constant or at least something that you can’t really affect. This entire discussion is about motivating people to contribute. Going by the “motivated contributors will continue to contribute” won’t work. We need to focus on motivating

[mellow hare]

Before (if) this goes public, I'm going to explicitly say that high scores do not mean you'll get contributor.

[thorny obsidian]

On that note, and I'll make a comment on the issue too, I think code reviews should be weighted higher than producing code.

[mellow hare]

Noted

The current scoring system is just a rough idea I had at the moment

I'm more than willing to have it shuffled around until we find something that feels right

[green oriole]

I think making it public will end up in an endless Hacktoberfest 2019

[hardy gorge]

Yeah, we discussed that yesterday as well

[mellow hare]

@green oriole yes and no.

If we start getting code reviews into the mix, it won't be nearly as tedious

[thorny obsidian]

I just had a good convo with my best friend over the weekend about overall strats for encouraging code reviews, it happens to be very applicable here!

[green oriole]

But how can be sure that someone would left a review did its job correctly, and isn't just looking for point? We will have to do the work twice.

[mellow hare]

Correction, I will

[thorny obsidian]

Are you concerned that staff and contributors won't actually review code?

[mellow hare]

And that's something I'm willing to take up

He's saying whenever it gets opened to public

[green oriole]

Gosh, that's a lot of work for just one man as incredible as he might be

[mellow hare]

Considering I'm going to be looking through the repos daily, it should be fine

And I could ask for assistance from time to time

[thorny obsidian]

I think that's putting the cart in front of the horse right now. I think the code review guidelines will help as well. Maybe have a checklist people can follow.

[mellow hare]

Oh trust me, I don't plan on being the only poor sod going through them all

But something to consider, ALL of our PRs require a core dev's signoff anyway

[green oriole]

Yeah, but you do need at least 2 good code reviews to be done before merging

[mellow hare]

If the core dev notices that the PR got a lgtm but they managed to find tons of stuff, we'll know that the other person isn't putting in their best

[green oriole]

As cool and well done the code review guidelines might be, we can't force people to follow them

But I agree on that point, it sounds like a problem for later

We are working on a lot of "if" right now

[mellow hare]

Exactly

For now, it's internal

And like Kat said, the guide needs to be done first

..... and I'd like help with that

I've hit a wall on how to articulate it

It shouldn't get stuck behind me just because I've got writer's block

[hardy gorge]

I'm confused about one aspect: Did we not decide to keep it contrib+staff yesterday?

[mellow hare]

For now we did, yes

But I feel like the board won't serve its purpose until it gets the community involved

I love our contribs, but they're, what, 0.001% of the total server pop?

[green oriole]

!role 295488872404484098

[stable mountain]

Contributors info

ID

295488872404484098

Colour (RGB)

#55cc6c

Colour (HSV)

0.37 0.58 204

Member count

35

Position

17

Permission code

0

[mellow hare]

And most have already become staff

[green oriole]

0.035%

[mellow hare]

That's still teeny-weeny

[hardy gorge]

What we discussed yesterday is that the board is a way of highlighting contributions. Most of us won't see what happens on GitHub at all, as it happens outside of the community.

[mellow hare]

I agree. As I said, this is a trial and it's locked down for staff and contribs

I'll try to keep more focused on the current issue at hand then

[hardy gorge]

Anyway, we can cross the public bridge once we get to that point

[mellow hare]

Yep

[cold moon]

I think this should be contributors and staff + invited normal members.

[mellow hare]

If we're thinking someone is worthy of being on the board, would we not have them become a contrib anyway?

I don't want this to feel like it's an exclusive club

Everyone on this server should have a stake in the game, as it were

[eternal owl]

https://pythondiscord.com/pages/contributing/sir-lancebot/#development-environment
We should also include USE_FAKEREDIS in the additional env variables section
As sometimes, we won't be running the project with docker

[mellow hare]

I'm in no way trying to dismiss anyone's thoughts or opinions on the matter. I want to hear everyone's thoughts on this, either publicly, privately or on the repos

Yeah we should

[green oriole]

There's a meta issue for those

https://github.com/python-discord/meta/issues/60

[eternal owl]

I will comment in that

[green oriole]

If you can comment on it, that'd be nice

[eternal owl]

done

[green oriole]

thanks!

[eternal owl]

also, what the 2 reddit commands we have on our bots?
I guess we want to have it on only 1 bot, on @dusky shore and remove it from @stable mountain

[mellow hare]

Ideally

[green oriole]

We would need to migrate #reddit to lancebot too, but yeah

[obsidian patio]

This is something @green oriole has talked about a lot hahaha

[eternal owl]

isn't #reddit powered by webhook?

[cold moon]

This is sending it with webhook yes

[eternal owl]

yea, so what is there to migrate to @dusky shore from @stable mountain

Unless we really like the reddit command on python compared to @dusky shore's reddit command

[green oriole]

But @stable mountain is sending webhook messages, isn't it?

[cold moon]

Yes, @stable mountain sends these webhooks

[stable mountain]

Gah, so many pings!

[green oriole]

so @dusky shore will have to send th- ... lol

[eternal owl]

someone just woke up @stable mountain

[eternal owl]

oh i thought those were coming straight from reddit

Maybe we can open an issue for the reddit command and migration?

[green oriole]

Yeah, I'll do that

!remind 1h30M do that

[stable mountain]

No problem.

Your reminder will arrive in 1 hour and 30 minutes!

ID: 915, Due: 2020-11-23T16:54:12

[green oriole]

@thorny obsidian why do you think we shouldn't just give one point for both?

[thorny obsidian]

Because I want to specifically incentivize code reviews as they can sometimes be seen as less "fun" or what have you.

[eternal owl]

also, can you also include https://github.com/python-discord/sir-lancebot/issues/348
this issue while you are creating it. So, 1 PR can clear 2 Issues at once

@green oriole

[green oriole]

Alright, sounds good

[eternal owl]

ty

[cold moon]

When any core dev have to time, @dusky shore 524 PR should get core dev review, as this already have 2 approvals.

[cold moon]

Will we have summary page in site about this year Hacktoberfest?

[green oriole]

I didn't know we had one for 2019 lol

[cold moon]

I don't know do we have for 2019, but I think we should have summaries for our events.

[green oriole]

Aaah

Maybe @thorny obsidian could be interested about writing one?

[thorny obsidian]

haha, I was just about to ask what the process is for putting up something on the site since that sounds like my domain and a good idea.

[cold moon]

I think this should be possible to have draft posts?

[green oriole]

Haha, we are currently in the process of migrating to a markdown based workflow

[cold moon]

Event pages will have HTML

I'm migrating them right now

[green oriole]

You'd simply have to commit markdown to our site repo

Well, once the migration is complete

[thorny obsidian]

I'll just create markdown that I hold in reserve until the migration is complete

[green oriole]

Sounds good

[thorny obsidian]

I guess this might be more of an @wild ridge question, but will it support html + markdown or strict markdown?

[green oriole]

If you need to pull any stats about the hacktoberfest or whatever, feel free to ask us

[cold moon]

Like I said, staff decided to have full HTML event pages that is in repo.

[thorny obsidian]

oh whoops, I can't read. Thanks!

[green oriole]

Oooh, I thought we were also going to have markdown

Alright thanks

[cold moon]

Like this https://github.com/python-discord/site/blob/dewikification/pydis_site/templates/events/pages/code-jams/7/_index.html

You can also open PR now to dewikification branch, then this go live same time when dewikification is completed.

[thorny obsidian]

Oh neat, I will give that a go. I'm still a bit of a git noob, so I might need core dev assistance if I fuck it up

[green oriole]

As I always say, no worry, you can always get rid of the proofs if you mess up haha

[cold moon]

And you can't make any disaster, because we have protected branches 😆

[stable mountain]

@green oriole

It has arrived!

Here's your reminder: do that.
[Jump back to when you created the reminder](#dev-contrib message)

[green oriole]

Smh just when I shut down my computer

[rapid swallow]

shut down your computer and don't tell us 😉

[patent pivot]

@cold moon review left on forms-backend#1

[tawdry vapor]

This PR needs one more review. The refactor will make it easier to work on the future changes to the code, some of which I think are upcoming. https://github.com/python-discord/bot/pull/1294

[hardy gorge]

!remind 8h Review

[stable mountain]

I got you.

Your reminder will arrive in 8 hours!

ID: 917, Due: 2020-11-24T07:28:43

[late wolf]

For the bot, I see lots of logging, are all of those logs in a file?
Or are they sent to the specific channels?

[hardy gorge]

The bot does write its logs to a file within the docker container, but we can also view the logs from the command line using kubectl

The logging level is set to INFO in production

Which means there are not as many log messages in production as you see in development mode

[late wolf]

ah so it writes the logs to a file, i see thx

[hardy gorge]

It does, but that's not how we view the logs

The logs outputted are also capture by docker (you could view them with docker logs container) and those are relayed by kubernetes

It would still work without writing them to file

[late wolf]

those are relayed by kubernetes
where are they relayed?
to the site or the server

[hardy gorge]

Someone with access can access them from the command line with a tool called kubectl

We're also currently setting up a dashboard

[late wolf]

oh i see

[stable mountain]

@hardy gorge

It has arrived!

Here's your reminder: Review ![this](https://cdn.discordapp.com/emojis/470903994118832130.webp?size=128 "this").
[Jump back to when you created the reminder](#dev-contrib message)

[hardy gorge]

!remind 2h

[stable mountain]

Affirmative!

Your reminder will arrive in 2 hours!

ID: 920, Due: 2020-11-24T09:29:49

[valid glade]

!remind 2h

[stable mountain]

Missing required argument

content

Command Help

!remind [mentions]... <expiration> <content>
Can also use: reminder, reminders, remindme

Commands for managing your reminders.

Subcommands:

!remind delete <id_>
Delete one of your active reminders.
!remind edit
Commands for modifying your current reminders.
!remind list
View a paginated embed of all reminders for your user.
!remind new [mentions]... <expiration> <content>
Set yourself a simple reminder.

[eternal owl]

@green oriole you can assign the reddit issues to me

[green oriole]

Sure thing

Have you commented on them?

[eternal owl]

done

[green oriole]

Done!

[eternal owl]

ty

[stable mountain]

@hardy gorge

It has arrived!

Here's your reminder: ![this](https://cdn.discordapp.com/emojis/470903994118832130.webp?size=128 "this").
[Jump back to when you created the reminder](#dev-contrib message)

[green oriole]

The reminder chain haha

[surreal venture]

@hollow cedar Sorry if I am disturbing you , but can you please drop a review on my PR?

[dusky shore]

GitHub

[sir-lancebot] #509 Spook Name Rate Game

[eternal owl]

I will take a look soon

[cold moon]

@surreal venture About 509, persisting data to JSON doesn't work anymore. You have to use Redis now.

[cold moon]

Is there any way to get Postgres or MongoDB to @dusky shore ? I'm currently thinking about AoC implemention, but Redis don't support dicts and lists and is very limited.

[green oriole]

Weeeeellll

I would be against that because it will add a pretty major complexity to the bot

I think seeing if async-redis-cache could support those type instead sounds like a better idea (CC @hardy gorge)

[brazen charm]

worst case scenario you should be able to pickle/json it

[cold moon]

I don't want to store JSON.

[hardy gorge]

Could you describe the type you're after? A RedisCache is already pretty close to a dictionary

[cold moon]

I don't want to create namespace for every user

I want user -> user info

[brazen charm]

What's the data layout you need to store?

user info would be a list for example?

[cold moon]

No, simple example is such dict: {score: x, days: [{day: 1, completed: time, star2: bool}]}

[brazen charm]

serializing it before sending it off to redis sounds like the simplest solution to me if we're avoiding proper databases because of ease of use

[green oriole]

I agree

[cold moon]

Is there no way like fake DB?

Like fakeredis?

[hardy gorge]

fakeredis does not give persistence

It's faking redis without actually having a redis backend

[brazen charm]

I don't think there's anything readily available for postgres; redis is much less complicated to replicate the behaviour of

[hardy gorge]

Regardless, serialization sounds like the way to go here, although you need to be mindful of race conditions

The hash sets that we're using basically have two keys and one value: (namespace->key->value)

That's the deepest nesting of native redis types you'll get

[cold moon]

Let's see what I can do with JSON. I will make not persistent list to cog that store all entities that is being edited and this check for them.

[brazen charm]

I'd lean towards pickle here personally (and used it it for the doc redis cache), as it's more performant and nobody will be viewing the raw data in redis anyway

[cold moon]

@patent pivot About forms-backend, do you really want type hints for self and cls?

[surreal venture]

What's Redis now?

[patent pivot]

ah, no I don't feel free to add back the ignore for that

[cold moon]

@patent pivot PR is now ready for another review.

[patent pivot]

approved @cold moon. @hardy gorge if you get a chance could you give it a onceover from a CI PoV

[late wolf]

What's the ratelimit of snekbox?

[green oriole]

We don't have any rate limit AFAIK

[late wolf]

oh

[cold moon]

@late wolf I think this have max 2 evals running at same time, but no actual ratelimit.

What is going on in #dev-log ?

[lament hedge]

sry it's me. implementing requested changes and every time i merge a requested change it auto runs the tests

idk how to stop it

[cold moon]

Oh, this is in fork

[grave summit]

how can I start contributing to the bot?

[cold moon]

How to store Pickle object to Redis and then get it again? Redis need argument to be string, but pickle wants bytes.

[brazen charm]

redis strings should be able to store binary data

[cold moon]

I mean I need to provide some encoding when converting Redis string to bytes (I use str converting when I store to Redis).

What encoding I should provide?

[green oriole]

Actually you can only have one eval running at the same time

[green oriole]

And, well, actions is going

[cold moon]

I thought that this is main repo master

[cold moon]

@brazen charm I think I'm still going to use JSON. This is just much easier than pickling and depickling object.

[brazen charm]

Ah right I guess rediscache will try to turn everything into strings

[hardy gorge]

There's a max of two eval workers

Just one job per person at the same time enforced by @stable mountain I think

Other jobs just queue up and we process a max of two at a time

Maybe not. I think that's what Mark recently said when we were thinking about the container memory limit

Ah, yes

There are two workers available (restricted on the gunicorn level)

[brazen charm]

Any reason rediscache doesn't have a bytes datatype? Feels like a useful type to have @hardy gorge

[hardy gorge]

It's never come up before

I think we could, but we'd have to refactor some bits to bypass the encoding/decoding steps

[cold moon]

I already solved issue with JSON, but I think this might really be useful.

[brazen charm]

Just looked at the code, is it not using pickle intentionally to simplify what it can do or for the size gains? Doing dumps/loads directly instead of the typestrings would work nicely and expand the possible types to a wide range

[rapid swallow]

Perhaps it has something to do with race conditions or something

[hardy gorge]

I'm not sure if I want to pickle/unpickle everything. I'd rather give the bytes option to the end user and let them handle it themselves.

Including making sure that nothing "bad" ever gets pickled and then unpickled

[brazen charm]

Even then, using pickle in the background with restricted types sound simpler to me

[trim cradle]

I want to make sure that my fix for the !close command is right but I'm not sure exactly what testing to do before I put in the PR

I'm pretty sure it's working

[cold island]

What is the fix

Also you can submit a PR draft if you don't want to say you're done yet but want to allow other people to take a look

[trim cradle]

I'm realizing that docker isn't using my virtual environment

[cold island]

Why would it

It's using whatever is in the image

[trim cradle]

I see

my image is missing emoji, then

[hardy gorge]

Is it part of the Pipfile.lock?

If so, you should be able to rebuild your container

[cold island]

Your test bot just doesn't have access to those emojis

If that's what you mean

[trim cradle]

I think docker-compose pull is what I need. The bot would no longer start once I pulled from master.

[hardy gorge]

you could try a docker-compose build

Assuming that you're using the docker-compose in the bot repository

[trim cradle]

I'm trying that now

[hardy gorge]

It does not pull an image from somewhere; rather, it builds it from the dockerfile using your current files

[trim cradle]

This is what I get only having experience programming in an academic setting.

[trim cradle]

I'm noticing that commits in the bot repo typically have a main commit message and then a longer, secondary message. What is this called?

[cold island]

@trim cradle you mean the commit subject and body?

[trim cradle]

Sure

[cold island]

I'm just not sure if I'm answering to the right thing

[trim cradle]

I guess I'll have to figure out how to get that in the commit tool in PyCharm

[cold island]

You basically write why you're making the commit in short, and then after a blank line explain in more detail if you think it's necessary

[sullen phoenix]

the blank line is key

[cold island]

If you're committing from the command line, you can do that by using the -m flag twice

[trim cradle]

I'm learning

[sullen phoenix]

i'd suggest to not use the -m flag at all

[cold island]

😮

Let it prompt you mean?

[sullen phoenix]

yeah, git will open your editor of choice instead

[cold island]

considering I have to write every commit twice I'd rather just have it all right away 😄

[sullen phoenix]

i just use nano since it's quick and easy

lol

[trim cradle]

I only do git commit -m if I'm working in a terminal-only setting

ie my uni's mainframes

[cold moon]

@hardy gorge WARNING "::error file=%(path)s,line=%(row)d,col=%(col)d::[flake8] %(code)s: %(text)s" is an unknown formatter. Falling back to default. I get this error when I try to use same formatting like other repos use for annotations. (verbose mode)

[gritty wind]

I get the same error too, yet it still formats using that string. I think it might either be a bug, or bad wording (unknown formatter I.E not a builtin one, default = parsing)

[cold moon]

Yep, I found this in docs

I will make changes when I'm back from school

[lament hedge]

I can't find formatter for Expiry object to send expiry date in message, like in !tempmute it sends message with until 2020-11-25 19:32 (59 minutes and 58 seconds).

can someone point to me where this formatter is?

[cold island]

You mean something like this? https://github.com/python-discord/bot/blob/0a49cb61085c2ceb8974decc69b200905bcf14e7/bot/utils/time.py#L118

[lament hedge]

I think that's it, thx

[cold moon]

@hardy gorge made changes to forms-backend GH Actions and deployment PR.

[hardy gorge]

I was thinking about your matcher. Do you know if it shortcuts after the first match in the pattern array or does it try to match all entries, regardless of previous matching ones?

The matcher seems to be a bit underdocumented

[cold moon]

I don't know. I used Regex that pre made actions used. This doesn't have official docs yet, only third party posts

[hardy gorge]

Yeah, I was looking into them as well during my initial investigation

But the documentation is lacking and I decided that the --format option was really straight forward

However, if it only matches the first, you could still have the "warning" vs "error" level thing

[cold moon]

This use now again format too

Pre-made actions don't use extensions

[hardy gorge]

By adding a matcher for the warning first, then a "matches all error codes" for the rest with an error level after

[hardy gorge]

Some tried to allow for it and it broke the matchers they were using initially

That's what made me look for a different solution

[cold moon]

Yes, I reached same after trying multiple times

But I will leave from WiFi connection area now.

[hardy gorge]

I'll have a look later. There wasn't much to change, so it's probably good to go anyway

The files already looked good

[grave summit]

do I really need to install all this stuff to contribute to the bot?

[green oriole]

If the stuff is all the dependency at the top of the contributing guide, yes

[grave summit]

I meant the test server and account

[green oriole]

Yes

Hey @clever wraith, have you seen our contrib guides here?https://pythondiscord.com/pages/contributing/

[grave summit]

how do I activate the virtual envirnoment creaated by pipenv?

[green oriole]

You don't really need to activate it

What do you want to do?

[grave summit]

I mean later I'd to run the scripts

shoudln't I do that within the ve?

[green oriole]

To run pipenv run you don't need to enable the venv

Pipenv will sort it out itself

[grave summit]

ah great, I always used virtualenv so far, so I was confused

how about docker-compose shoudln't I install that within the ve?

[green oriole]

You'd usually do a system wide install of docker-compose

[clever wraith]

how i setup @stable mountain bot @green oriole ?

[grave summit]

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

[clever wraith]

i download it in the github repo

[green oriole]

You should take a look at the link Kingboi posted

[clever wraith]

how i install pipenv on vs code?

[grave summit]

you have to use cmd for that

[clever wraith]

but which cmd

[cold moon]

@patent pivot forms backend PR is now ready for merge

[grave summit]

did you ever use pip install before,

[clever wraith]

no

[grave summit]

ok then you have a bit of reading to do before contributing in my opinion

it's a bit unclear what I should do with all these ids?
https://pythondiscord.com/pages/contributing/setting-test-server-and-bot-account/#obtain-the-ids

[green oriole]

Paste them in your config.yml

I believe it is described later in the guide

[grave summit]

this is found in the cloned repo?

[green oriole]

No, you need to make a copy of config-default.yml

[clever wraith]

send dm to minty codess

[grave summit]

ok I think I got it now, sorry for all the questions

[green oriole]

No worries, I don't mind

[grave summit]

do I have to cpy the ids of all the channels?

[brazen charm]

You don't need to create all of them but some are required for the general setup and then for whatever you'd be working on to test it out. Could point them all to one or two channels if you're being lazy

[cold moon]

@patent pivot Sorry about ping again but I started thinking that when my PR in forms-backend get merged, this is first time when this actually go to production?

[patent pivot]

itttttttt is

yeah

[cold moon]

Nice!

[patent pivot]

that is a good point actually

we'll need to add a secret to contain the application environment

[cold moon]

Yes, KUBECONFIG and GHCR user and password

[patent pivot]

not those

like actual env vars for the container

[cold moon]

Oh, these. Yeah

[patent pivot]

those secrets are on the org already

[cold moon]

DB URL and secret key

[patent pivot]

and oauth creds

[cold moon]

I think after this is done, next thing I'm going to do in this repo is to write nice README.

[patent pivot]

okay suggestion added

[cold moon]

Commited

[trim cradle]

Trying to figure out the flake8 situation

[patent pivot]

address will beeeeeeee

hmm

api.forms.pythondiscord.com

[cold moon]

Amazing!

[patent pivot]

eh, can't figure out two levels of SSL, for now it'll be forms-api

merged

[cold moon]

@patent pivot Deployment failed

[patent pivot]

ah

I see the problem

fixed

https://forms-api.pythondiscord.com/

[cold moon]

Nice!! I love to see when new things go live!

[gritty wind]

What's up?

[trim cradle]

sorry, I didn't notice the ping for some reason. I was having trouble setting up the pre-commit checks in part because gitbash sometimes behaves unpredictably. I switched to powershell: https://paste.pythondiscord.com/walukivimi.coffeescript

I think this output means that I lost my chance to run flake8 because I already committed the changes?

[gritty wind]

You need to add the files to git before running the pre-commit, otherwise it'll stash them all, then try to run pre-commit on 0 files, thus the skip

And uhh, I didn't ping you whoops

[trim cradle]

you replied and I thought that would give me some kind of alert.

[gritty wind]

It does, but only if I keep the mention option active

[gritty wind]

Like so

I turned it off because of muscle memory 😅

[trim cradle]

I see. The changes I made are already committed and pushed.

what does "add to git" mean, in this case? The files in question are tracked by git.

[gritty wind]

They are tracked, thus without git add they were left in the staging area

That functionality is what allows for commiting only certain changes from your change log

See here:

before running the commit command, you must use the add command to add any new or modified files to the index.
https://git-scm.com/docs/git-add

[trim cradle]

@gritty wind the changes are already committed, so what is there to add?

[hardy gorge]

You can also run flake8 directly.

[gritty wind]

In theory, if you try to commit with unstaged changes, you'll see something like

[hardy gorge]

You can also just list the status

[gritty wind]

Some of your changes must've been added, and the rest were stored away by the precommit

[WARNING] Unstaged files detected.
[INFO] Stashing unstaged files to C:\Users\Steele.cache\pre-commit\patch1606418810.

[trim cradle]

I'll have to look at this after my family does our Thanksgiving stuff. Thanks so far @hardy gorge @gritty wind

[gritty wind]

Have fun, and happy thanksgiving

[cold moon]

I found that forms-frontend also don't use Docker. Should I create Docker + Kubernetes there too?

[lament hedge]

for some unholy reason Pipfile.lock committed with some other files. Do I need to revert that or don't worry bout it?

[surreal plaza]

Can you guys help me out here?

[green oriole]

Hey @surreal plaza, see #❓｜how-to-get-help, although this look like an assigment, we probably can't help you with that

[surreal plaza]

Thank you! <3

[cold island]

Well, did you change the dependencies?

[lament hedge]

I don't think I did, but can't guarantee it didn't happen on accident

[cold island]

If you think you might have done something by mistake then it's probably better to revert it.

[green oriole]

If almost the whole file changed, it probably means that you locked with the wrong python version

[cold moon]

.issue 531

[dusky shore]

GitHub

[sir-lancebot] #531 Advent of Code 2020

[cold moon]

Can any Core Dev give review to this?

[patent pivot]

I'll request the core dev team now

lol of course it just requests mark that's helpful

[cold moon]

@patent pivot Created PR to forms-backend for README

[patent pivot]

good structure, I'll fix some phrasing stuff and then it'll be good to go

[cold moon]

How will be new forms created and answers collected? Custom interface in forms-frontend, with plain API requests or even another like dashboard?

[patent pivot]

custom interface

[cold moon]

So there will be total 3 apps for forms: forms-frontend, forms-backend and forms-dashboard?

[patent pivot]

nah, custom interface inside forms-frontend

[green oriole]

ks is working on the forms but isn't in the event team smh

[cold moon]

Oh

[patent pivot]

I think the backend logic is in place already

ks maybe you should like

be in the insights team

[cold moon]

What team this is?

[thorny obsidian]

lol

[green oriole]

well, you didn't really had a choice, did you

[patent pivot]

the team working on forms and eventually insights (our version of DO currents or other dev surveys)

no one gets a choice!

my team my rules

[green oriole]

Hahaha

"You are mine now!"

[thorny obsidian]

The next code jam will most likely try to use whatever comes out of Insights/Forms

[cold moon]

😆

[thorny obsidian]

So yes, please more people on Insights

[patent pivot]

ks I think the logic for the admin stuff is already there

        user_details["admin"] = request.state.db.admins.find_one(
            {"_id": user_details["id"]}
        ) is not None

so if there is an entry in the admin table for a uid then they get an admin key in the assigned JWT

and the frontend will read that and display the creation UI

as with most react apps everyone will get access to the frontend UI if they play around with their console, so we'll validate it on the admin side (we can verify that we signed the JWT they submit with)

[cold moon]

I have seen one Vue app where all logic was in frontend and no validation in backend. Data that was finally submitted...

[patent pivot]

lol yikes

yeah we'll probably just write a decorator for it

[green oriole]

"Trust me backend, my will is good"

[cold moon]

Should I make PR to use to use GH Actions, Docker and Kubernetes for forms-frontend?

[patent pivot]

hmmmmm

not right now

we may still use netlify for production

[cold moon]

Oh, you will not use Kubernetes for this?

[patent pivot]

probably not

it feels a bit overkill, y'know?

it's a fully static thing

we'll use netlify or S3 or maybe even github pages

[cold moon]

Committed README PR suggestions

[patent pivot]

last review comments left, then it's ready

[green oriole]

We don't have them in #dev-log, do we

[cold moon]

But is there currently something more to do?

[patent pivot]

backend is not there yet

I'll add it after dinner

[green oriole]

Fair enough

[patent pivot]

not yet, I can write up some issues after dinner

[cold moon]

👍

[cold island]

So I have an Expiry argument I want to make optional, and then give it a default value inside the function. Would I set it explicitly, or would I use the appropriate converter for it? doing it explicitly is not ideal in case the implementation of the Expiry object changes. But I don't know if it's acceptable to use a converter explicitly

Setting it explicitly would be duration = datetime.datetime.utcnow() + datetime.timedelta(hours=1), which works technically

[brazen charm]

Nothing wrong with calling the converter yourself

[cold moon]

Finally all code jams pages are migrated to HTML...

Only game jam pages left

Or actually, there is only 1 page

[cold moon]

@wild ridge Changes in content app is made (with bigger refactor). Also new PR https://github.com/python-discord/site/pull/427

[crude gyro]

if someone wants a very easy issue to solve on our site - https://github.com/python-discord/site/issues/428

[gritty wind]

Running into a bit of a conundrum here. I have a file importing classes from discord, and importing classes from constants. now the constants imports include guild, and the import I want is discord.Guild.

What is my course of action here:

1. Make both explicit imports, and use constants.xy, discord.xy, etc
2. Make just the discord imports explicit
3. Leave the existing imports, and double up the discord imports so I can use discord.Guild

[green oriole]

Hmm, you could use as imports, something like a

import discord.Guild as DiscordGuild
import constants.Guild as ConstantGuild```or use dotted path for both

[brazen charm]

Both of those feel like they benefit from having the module name in front of them to me

[gritty wind]

Most of constants really could use that

I guess I'll do Akarys' suggestion and leave it up to reviewers

[brazen charm]

I don't like that too much, feels like discord.Guild and constants.Guild but with extra steps

[sullen phoenix]

i’d rather namespace it

[gritty wind]

It is, but the if we do go with discord.Guild and constants.Guild, do we do that for all the imports?

[brazen charm]

Probably, depends on what those imports are and how they are used

[green oriole]

The small issue is that this will feel disorganized depending on how the code look like

TBH, do whatever feels right

[trim cradle]

It looks like the help_channels.py file was deleted in the help channel refactor, even though _cog.py appears to be the spiritual successor to that file. What's the most git-thonic way to integrate my changes into that file?

[brazen charm]

So you have changes in help channels and need to merge master which removed it?

[trim cradle]

I have a few commits in help_channels.py that I can no longer merge into master because that file was deleted, yes

so I'm not sure what to do except figure out how different _cog.py is from what help_channels.py was before I made those commits and copy/paste them in.

doesn't git support continuity for a file if you move or rename it?

[gritty wind]

One idea would be merging into the head before deletion, updating the branch, and adding another commit to migrate to the new file

At least from a history perspective, things make more sense

[trim cradle]

I'm worried I'd make the history even more difficult to follow if I did that.

I might just have to make a new branch and try again.

[green oriole]

If the file has been properly tagged as a rename, yes, otherwise no

[brazen charm]

I'd probably just introduce the changes to the new file as a part of the merge

[green oriole]

You can try something like this https://stackoverflow.com/questions/16526321/how-to-apply-a-git-patch-to-a-file-with-a-different-name-and-path

[gritty wind]

The way I understood it is, rename+major refractor, which would mean having to update the code anyways, so merge resolve will fail

[trim cradle]

the first commit on _cog.py is recent, so it doesn't appear that that's what was done.

[green oriole]

@gritty wind FYI we don't assign GitHub authors to PRs

[gritty wind]

oh

I assumed we did because the bot adds a need author tag

[green oriole]

We sometimes do if someone will walk the author through the whole thing

the status: waiting for author tag?

[gritty wind]

Yeah

[green oriole]

It is to communicate that changes need to be done by the author

[gritty wind]

Huh, the way I originally read it was "Author needed to take over"

Good to know

[green oriole]

That's pretty fair, I think the tags could deserve some little descriptions

[gritty wind]

Though, if the tag doesn't mean what I originally thought it meant, how does adding one for that purpose sound?

[green oriole]

Hmm, that's pretty sad, can you try the patch method I linked?

[gritty wind]

I don't have statistics on how many stalled PRs there are, so I don't know how useful it would be

[green oriole]

Author needed to take over in what sense? Someone else should take over the PR?

[gritty wind]

As in the original author is no longer working on the PR, but their changes can still be useful

So a new author can continue fixing the PR for merging

[trim cradle]

so this is a way of retroactively saying "treat the first commit on this file as coming directly after the last commit on this file"?

[green oriole]

Right, this is so rare that I'm not sure if it would really be worth it

Yes Stel

Hmm actually it will only appear as one commit

[tawdry vapor]

It's not possible to "tag" a file as renamed in git afaik

It's just something git will automatically figure out based on how similar the file is

I remember having frustrations years ago when refactoring stuff

If the file is explicitly moved with git mv, and then edited significantly, git will still think it's a new file

It's just a fundamental way of how git was designed is what I read

[trim cradle]

I see

I've almost finished re-introducing the changes @tawdry vapor

though I have to leave now, so I'll have to figure out what the situation is with the automated checks later.

[sullen phoenix]

hey @gritty wind, could you also line up the spacing for the channel IDs?

i don't want to add another comment lol

[patent pivot]

git calls it a "rewrite", so a rename is a 100% rewrite to a new file (that's why you'll sometimes see like 74% similar etc. etc.)

[gritty wind]

Ouch, that was a bit of an oversight

[sullen phoenix]

it's all good

oh yeah

they should also be alphabetized @gritty wind

sorry lol

[gritty wind]

I'll do that, but perhaps that should be a comment on the thread too?

[sullen phoenix]

sure, i'll comment it

[gritty wind]

Its just staff voice chat and voice chat that need to be switched, correct?

or do the categories themselves need to be switched too

[sullen phoenix]

for the voice category, yeah

also the voice channels category, and the Channels class

[gritty wind]

Not sure what you mean by that. This is what I have right now

Channels:
        ....

        # Voice Channels
        code_help_voice_1:                  751592231726481530
        ...

        # Voice Chat
        code_help_chat_1:                   755154969761677312
        ....```

(With the staff and voice chat switched)

[sullen phoenix]

give me a bit, currently eating lunch

[patent pivot]

what's for lunch

[sullen phoenix]

i’m having a breakfast sandwich actually

eggs, sausage, bacon, etc. in it

[patent pivot]

nice

[sullen phoenix]

@gritty wind so for example, this would be the Channels class in alphabetical order https://paste.pythondiscord.com/acazobupeb.py
and this'd be the yaml: https://paste.pythondiscord.com/agiqosurud.yaml****

[patent pivot]

replies with python would be cool

https://images-ext-2.discordapp.net/external/9KzjtPc9G42UjlTpCFomN6NPgyLiObDI2GIMH8aBMag/https/u.dejay.dev/%25F0%259F%258F%258E%25F0%259F%258D%258B%25E2%259B%2584%25EF%25B8%258F.png

[sullen phoenix]

that would be cool

[green oriole]

But... that's just noise?

[patent pivot]

it's pretty noise

[green oriole]

I mean, we don't have some insane 10 second delays

It will probably always reply to the last message

[patent pivot]

this is true

[gritty wind]

Oh, i get you know

I should open an issue for refactoring both files after this then

Maybe later

[trim cradle]

Are all async redis cache operations expensive? I'm thinking it would be useful for the increment and decrement methods to return the new value.

Which would have the consequence of making them asymmetric with a[b] += c

[sullen phoenix]

it's only the additions you've made that need alphabetizing

all the others should already be alphabetized

oh nice, you already made the changes

[gritty wind]

In theory, yes, but:

change_log
announcements
python_news
...

like right above it

how_to_get_help
cooldown```

don't want to flood this channel though

[sullen phoenix]

oh lol, i didn't look there

[gritty wind]

It can be tackled later

[sullen phoenix]

sorry

yeah, that'd be a good idea then

sounds good

[gritty wind]

I think that git hook got sent before I pushed the submit button lol

[cold moon]

For any core dev: https://github.com/python-discord/bot/pull/1310 is ready for merge.

[trim cradle]

It's taking docker desktop a long time to restart. is that normal?

[gritty wind]

How long have you been waiting?

[trim cradle]

ten minutes maybe?

[gritty wind]

And is it the backend, or the container

Ok that’s way too long

[trim cradle]

and if I do docker-compose up:

    raise DockerException(
docker.errors.DockerException: Error while fetching server API version: (2, 'CreateFile', 'The system cannot find the file specified.')```

might just reboot my whole computer

[gritty wind]

Was there a recent update or big change on your system?

[trim cradle]

not that I know of

but I have environment issues with pretty much anything I try to do

which I realize is a vague category of problems

not just for the bot, but in general.

[gritty wind]

Try rebooting then, and take it from there

[trim cradle]

alright

[gritty wind]

It’s a honestly a miracle docker even supports windows

[trim cradle]

I run all my work stuff via an ssh connection because I'm not willing to waste any more of my life trying to set up ML/AI packages on Windows.

I want the "run these apt get commands and do the pip install again, problem solved" experience all the time.

[gritty wind]

I’m sure there are more than enough people here that would love to get you dualbooting off their grandmas secret os recipe

[trim cradle]

I sat down to dual boot one day after wiping my laptop and reinstalling Windows, which in itself solved a lot of problems

but as it turns out, I still had to defragment the hard drive to create room for a new partition

not sure why that was still needed--I guess because I didn't use the secure wipe so there's some way in which that memory is still allocated, or something.

in either case, it probably would have taken over 24 hours to defragment at the rate it was going and I can't be without my computer for even a third of that time with school and such.

[gritty wind]

I feel ya, my laptop has practically become a desktop at this point because I refuse to go to the repair center

Did docker get its crap together?

[trim cradle]

@gritty wind yeah, looks like my changes are working now. I just need to figure out the automated checks and then I can fix the PR.

my changes to the bot, that is

[gritty wind]

Checks as in precommit hooks?

[trim cradle]

I think so

[gritty wind]

Ah, well I wish you luck

Feel free to ask if you run into any issues

[trim cradle]

"mixed line ending" is the only one that failed.

[gritty wind]

Yeah that usually fails

Run again

And make no changes

[trim cradle]

it passed that time

weird

[gritty wind]

It changed it when it failed

[trim cradle]

Check Yaml...........................................(no files to check)Skipped```

I assume this doesn't matter

[gritty wind]

The thing in, after that it should in theory mark it as fixed and pass

Yeah that’s fine unless you were modifying something you expect to run on those files

[gritty wind]

I don’t have a fix for this yet, but GitHub desktop seems to be able to handle it for some reason. Idk why

[trim cradle]

what did the line ending thing even mean? is that the issue with Windows doing line breaks differently?

[gritty wind]

I always assumed so

[cold island]

yeah

Windows uses CRLF, unix uses LF

[gritty wind]

It’s not a problem unless you ask the precommit to run on your entire project, I.e if you try to run lint

[cold island]

yeah I always run the commit command twice when working on the bot

[gritty wind]

There’s gotta be a way around that lol

Ig I can manually turn off that check

[cold island]

probably, but I'm too lazy to do the required research lol

[trim cradle]

        claimant = self.bot.get_guild(constants.Guild.id).get_member(claimant_id)
        if claimant is None:
            log.info(f'{claimant_id} left the guild during their help session; the cooldown role won\'t be removed')
            return

is this useful to log?

[green oriole]

Yeah, it is

[cold island]

sure, why not

[gritty wind]

At some point we had also come to the conclusion that you can clone the repo using LF so it wouldn’t even face any issues

[green oriole]

It won't be triggered very often anyway

[gritty wind]

It does mean I have to commit twice like zig said, and from my editor that means running code analysis and limiting twice

Also it doesn’t work if you have staged and unstated changes

Also an issue on this:

https://github.com/pre-commit/pre-commit-hooks/issues/436

[trim cradle]

sigh. now the precommit check won't pass and running the check doesn't fix the line ending

[brazen charm]

You can set git to use lf for commits

[trim cradle]

I'll have to see if I can make pycharm use LF

[gritty wind]

Git already automatically uses lf, it’s the checkout and your editor that cause the problem

[trim cradle]

turns out pycharm has a toggle for that on the screen at all times

weird allocation of screen space

[gritty wind]

Funny enough, most editors I can think of do have it in the corner too

[trim cradle]

I'm not sure why running the check isn't fixing the line endings anymore

[gritty wind]

Do you have unstated changes?

(You’d a get a message about stashing if you do)

[brazen charm]

git will use native by default, setting autocrlf to input and eol to lf fixed the behaviour for me

Not sure if you need the eol set

[trim cradle]

I see; doing git add again fixes that

[gritty wind]

The default behavior is to commit with lf, but the issue arises when the editors automatically use crlf, which some do

[gritty wind]

Yeah it sucks

[trim cradle]

In my workflow, git add is usually a formality

[brazen charm]

Try doing git config core.autocrlf input and see if it fixes it, having to rely on the hook to fix it is a pain with merges (and in general)

[gritty wind]

i see commits

[trim cradle]

Error: [flake8] D205: 1 blank line required between summary line and description
Error: [flake8] D415: First line should end with a period, question mark, or exclamation point
Error: Process completed with exit code 1.``` welp

[gritty wind]

Hm, why didn't that get flagged?

[trim cradle]

is it an objection to the commit message? or perhaps an earlier commit message?

[gritty wind]

oh right, the precommit hooks only run on the current staged files

[trim cradle]

these are all the implicated commits: https://github.com/python-discord/bot/pull/1306/commits

[gritty wind]

Would be handy if the path was included

Try running pipenv run flake8 locally

Though it did fail flake on the last commit

[trim cradle]

PS C:\Users\Steele\development\projects\pydis\bot> pipenv run flake8
Loading .env environment variables...
``` Frozen on this step for some reason

[gritty wind]

There is no output while processing (unless you add -V). Did it take you back to terminal yet, or are you still on that step?

[trim cradle]

still on that step

[gritty wind]

For how much time approximately?

[trim cradle]

Presumably right around when you said "Try running pipenv run flake8 locally"

not sure how much of a delay there was

[gritty wind]

it only takes a few seconds for me, can you run it with verbose then?

[trim cradle]

sure thing

a lot of text happened and now it's idle but hasn't returned

[gritty wind]

If you can, dump it into the pastebin

[trim cradle]

!paste

https://paste.pythondiscord.com/desepiduzi.js

[gritty wind]

I'm not necessarily seeing anything that should make it break

Wait yes I do

bot only has a 150 odd files

Why Checking 2832 files

[trim cradle]

should I be working in a subdirectory of where I am currently?

[gritty wind]

You should be fine running the flake command directly in your project root

[trim cradle]

that's where I am

[gritty wind]

I'm just not sure why there are thousands of files there

Perhaps logs or something?

[trim cradle]

let's see

I only have one log file

[cold moon]

For some reason, this don't load extensions

Maybe this don't load config too and lint like venv or such things

[trim cradle]

wow a lot of text is happening now

[gritty wind]

If you find where you have ~2k files within your project dir, you'll be golden

[trim cradle]

it's searching the virtual environment.

[gritty wind]

yikes

[cold moon]

This means that this don't load config

[gritty wind]

On an unrelated note, I went ahead and tried to commit your commit (the one that failed the action), and it passed both manual invocation of the precommit, and the actual precommit. It failed a manual lint

[trim cradle]

what does that mean?

[gritty wind]

So I checkout out your branch, and reset to the second to last commit. I then ran:

1. pipenv run lint [passed]
2. git commit [passed]
   (rollback)
3. pipenv run flake8 [failed]

The workflow does step 3

So I'm guessing the precommit and the manual invocation of flake8 might be running two separate tasks?

[trim cradle]

I have no idea

I'll be right back

[cold moon]

pipenv run lint runs flake8 inside pre-commit

[gritty wind]

Yes, which is why its weird that produces a different result from running flake8 using pipenv run flake8 or activating then flake8

Ok, nevermind it does find the error, but the git hook still passes

and now it doesnt and it does, alright i give up

[trim cradle]

I'm not sure what to do

[gritty wind]

So I've done more testing, the exact same commit passes in:

• My IDE (pycharm)
• Github Desktop

It fails in:

• git bash (mingw)
• regular terminal git

I haven't been this confused in my life

[trim cradle]

looks like I may have more line break issues

so I'll have to run that fix recursively for the project

must be from pulling stuff from origin

[cold moon]

Can anybody link PR? I try to run this myself too.

[trim cradle]

https://github.com/python-discord/bot/pull/1306

[brazen charm]

Do you have a global flake8 install?

[trim cradle]

I don't believe that I do @brazen charm

[gritty wind]

Yes, but shouldn't it be running pipenv anyways (and we don't have extra rules, so even a system wide flake8 would fail)

[cold moon]

https://github.com/python-discord/bot/actions/runs/390215777 Check this page, this give you exact locations with annotations

[gritty wind]

Finding them isn't the hard part

[brazen charm]

pycharm has a bug where its precommit doesn't activate the environment and just runs flake8 outside of it

[gritty wind]

Alright, I'll uninstall and test

[brazen charm]

I'd check that to make sure it isn't because of that

[gritty wind]

And now it can't find flake8

Well, it was that apparently

How do you get around it?

[brazen charm]

You have to manually patch the path

[trim cradle]

    async def unclaim_channel(self, channel: discord.TextChannel) -> None:
        """
        Remove the claimant from the claimant cache and remove the cooldown role
        if it was their last open help channel.
        """

It objects to the second line of this, namely the initial """

[brazen charm]

#dev-contrib message

[trim cradle]

but it might be an objection to the docstring itself

[gritty wind]

Yeah that docstring should be one line, or three lines

[trim cradle]

Interesting

[gritty wind]

First line is always a short description, followed by a break, followed by paragraphs

[trim cradle]

now 217 files are modified because of the line breaks

sigh

I'll just have to delete my clone when this PR is done and try that again

[brazen charm]

Have you tried setting the autocrlf config?

[trim cradle]

for git itself? no

[brazen charm]

Try #dev-contrib message, it fixed the newline issues for me. I also set eol to lf but the docs say that's ignored hen autocrlf is not false

[trim cradle]

alright

I ran Numerlor's command, and I fixed the docstring

PS C:\Users\Steele\development\projects\pydis\bot> pipenv run pre-commit
Loading .env environment variables...
Check for merge conflicts................................................Passed
Check Toml...........................................(no files to check)Skipped
Check Yaml...........................................(no files to check)Skipped
Fix End of Files.........................................................Passed
Mixed line ending........................................................Passed
Trim Trailing Whitespace.................................................Passed
check blanket noqa.......................................................Passed
Flake8...................................................................Passed

does this mean I can commit then push now?

[gritty wind]

Did you fix the other annotations?

[trim cradle]

right. I need to make the linter exclude the venv directory

[cold moon]

When I run pipenv run lint in PyCharm terminal, I get both these errors

[trim cradle]

the only issue there is that it tries to fix line breaks in 200+ files

[gritty wind]

Oh thats safe to commit

[trim cradle]

but I don't want one commit to change 200+ files 😦

[gritty wind]

Numerlor's fix would require you recheck out the project

It won't, under the hood git converts everything to LF anyways, so it recognizes no changes need to be made

Just check the commit before pushing

[trim cradle]

so if I just go ahead and commit all those files (except the one that I actually changed), nothing will actually happen?

[gritty wind]

Assuming all the changes are line endings, then yes

[trim cradle]

alright

[gritty wind]

If you delete the local repo, then set global autocrlf, then reclone, this won't be an issue

[trim cradle]

I'll do that when this is done

[gritty wind]

Yeah, wise choice

[trim cradle]

pushed. Now we see if I get to not cry

[gritty wind]

I mean.. probably, but at least the line endings didn't get changed :D

flake8 passed

[short snow]

how does the aoc subscribe thing work, i saw the source but couldn't figure it out, like for every new challenge it pings.

[gritty wind]

The old one, or the PR?

In the PR, if you check the cog, line 88, it

1. gets the time left
2. sleeps that time + 1 second
3. checks for the new puzzle up to five times
4. send a role ping if it was found
5. repeat for 25 days

The subscribe function itself just adds people to a list

[short snow]

thanks!

[gritty wind]

OS: Windows 10
Command: aoc leaderboard
Config: 631135 Last year's pydis leaderboard. If you need my cookie, lmk

Exception:
https://paste.pythondiscord.com/alafejusus.rust

Some of the unicode characters:
\u2b50, \u0142

This seems to be on the redis end, as it doesn't stop the program, and I can not catch an exception there

@hardy gorge

(in reference to lance#531)

oh right, fakeredis is True

Doesnt actually break the program, so lower priority issue

[sullen phoenix]

...it's a logging error? that's new

[hardy gorge]

Weird. It may be due to the encoding used for logging in Windows. Maybe it's using that windows encoding, not unicode.

Ah yeah, it's using cp1252

Hmm, worth looking into, but it won't hurt the feature in deployment

[gritty wind]

👍

[hardy gorge]

My guess is that it's going to occur for any logging statements that try to log unicode characters

[gritty wind]

Oh, good catch

[cold moon]

I will soon run this locally

[gritty wind]

I'm writing a conclusion on the issue, then closing

[hardy gorge]

I'll make some final changes. I have a few hours to do change stuff, but I also have to do a bit of work.

I only slept 2 hours last night 🙃

[gritty wind]

Ouch thats rough

[cold moon]

[hardy gorge]

Ah, yeah, coding is fun

@cold moon the comment is correct; it does check range(1, 25) (excluding the 25th, because we don't need countdowns on the 25th for the 26th)

The docstring and the comment are just not entirely consistent

[cold moon]

But I think then docstring should be changed

[brazen charm]

The file handler that's created for the logging will need utf8 set

Should be a simple fix as a kaizen

[hardy gorge]

@gritty wind I can't check it myself here on Linux, but could you check if the error is now gone?

I've added an encoding to the file handler

[gritty wind]

Yeah, it’s working as expected

I don’t think it should’ve necessarily been part of this PR but I’ll approve it for time’s sake

[brazen charm]

No need to create an issue and a pr for every little change

[gritty wind]

Agreed, but this is unrelated from the actual PR

[green oriole]

At worst, a core dev can push to master

[hardy gorge]

It's not really unrelated, as merging this PR will cause everyone testing sir-lancebot's AoC features on Windows run into this warning

It's also a small change that falls within the kaizen allowance, I'd say

[gritty wind]

Either way, it’s in now

Who’s needed for review btw?

[hardy gorge]

I'm not sure. I think @sullen phoenix made himself code owner of everything, which means that he must approve or I have to merge using my admin permissions

There needs to be a second approval as well, I think?

[sullen phoenix]

lol

yeah, there should be

nope wrong channel

[green oriole]

We can change that

[sullen phoenix]

is the PR for the code owner review already merged?

ah nope

[gritty wind]

This is for files that don’t have a code owner afaik

[green oriole]

We can just merge ves’ PR

[mellow hare]

I need to bribe someone to help me finish the code review guide

I just haven't had any time due to work

(and my self imposed hiatus last week)

[thorny obsidian]

What needs to be done for it?

[mellow hare]

Pretty much just the procedure part of it. Or the "How" as I have it listed

I've already did the who, the what, and the why

Started a little bit into the how and just got stuck

I've been using Akarys' template he gave me to work from (it's on the issue, I'll snag that real quick)

[gritty wind]

How long do you expect it to be

[mellow hare]

Length isn't the goal. I just want to make sure that it's clear and helpful for people brand new to the concept of code review

It can be as long or short as it needs to be as long as it gives someone a good idea of how to do it

https://github.com/python-discord/organisation/issues/299

[green oriole]

For the non-staffer here https://paste.pythondiscord.com/dugolijigu.md

[mellow hare]

Oh derp right

[green oriole]

I could polish my comment, if that can help

[mellow hare]

Sorry

It'd help immensely. If you did that I could just rip it whole sale

Because what you have is incredibly solid

It just needs some polish

[green oriole]

Good lord, there are a lot of spelling mistakes in there

[mellow hare]

And I was also trying to give it the proper Hemlockisms

[green oriole]

Alright, I'll try to do that pretty soon

In fact

[mellow hare]

I really appreciate it. I feel bad that I'm having to push it off to someone

[green oriole]

!remind 16h30M polish the thing

[stable mountain]

Yeah okay.

Your reminder will arrive in 16 hours and 30 minutes!

ID: 968, Due: 2020-12-02T08:17:10

[green oriole]

Taking your own words

Life gets in the way

That's alright

[mellow hare]

You are my hero

[karmic frost]

Hi is there a stats command like to do!stats the bot will answer all the uptime CPU storage

[gritty wind]

I think I misunderstood your original phrasing. If you are trying to implement it for your own bot, you'd have to implement it from scratch. There is no uptime, but that should be easy to implement using the time or datetime module. For hardware usage, you can use a library such as https://pypi.org/project/psutil/

Either way, now that I understand what you are asking, that question is indeed better suited for the help channels, sorry

[gritty wind]

https://github.com/python-discord/sir-lancebot/compare/d664c68ecd81...ddb3dcf00573#diff-e5bace3e9f8f891dbdcc81752a2b2104a54662304da3b068d64df115ce598aecR66
Why's that not just environ.get("...", "").split(",") 👀

[hardy gorge]

because I'm stupid

(the actual reason is that I initially converted them to int and did not think about it after removing that)

[gritty wind]

I was guessing it was something along those lines, anyways it isn't a big deal :D

[hardy gorge]

If anyone has a bit of time: https://github.com/python-discord/sir-lancebot/pull/535

One of the background tasks has definitely failed (the status updater) and the other one, the new puzzle announcer, may have failed (although I hope it's still running). Regardless, both were a bit broken due to the deseasonification of @dusky shore. This PR should fix them.

[gritty wind]

Hmm can someone investigate this:
<#782715290437943306 message>

Current leaderboard:
https://paste.pythondiscord.com/raw/oyonurijuz

Entry on the pydis leaderboard:

23. 307 ** Lucas Lindström
    (both are gold stars)
    https://adventofcode.com/2020/leaderboard/private/view/631135

[thorny obsidian]

Oh that's curious

[gritty wind]

Issue persisted, looking into it
https://paste.pythondiscord.com/raw/jakacebivo

[obtuse arrow]

That's me, btw. 🙂

[thorny obsidian]

You're just special @obtuse arrow

[obtuse arrow]

I'm number zero

That's better than one

[sullen phoenix]

lol, you have all of them solved, right?

[obtuse arrow]

Yep

[sullen phoenix]

yeah, iiinteresting

[obtuse arrow]

Could it be the nordic character in my name?

I don't see anyone else with non-ascii in the table, I think.

[sullen phoenix]

hmm

[thorny obsidian]

That's what I'm thinking

[gritty wind]

Who has access to the logs

[hardy gorge]

I have

[gritty wind]

Anything interesting in there?

[hardy gorge]

No

sadly

[gritty wind]

Hmm, I'll do some testing locally

[sullen phoenix]

i suspect it's redis having issues with that character

[gritty wind]

Welp, got a json decoding error

[sullen phoenix]

oh

[gritty wind]

It is not the ö

Reason still unclear

Issue narrowed down: it happens in lines 80-95 of _helpers

Its the loop that parses the raw leaderboards and assigns the stars

It parses correctly at first, but fails afterwards

Issue identified: repeated names

The second one has no stars

We have to store ids instead of names

[cold moon]

Looks like Ves changed it from IDs to names. This should be changed to IDs back yes.

[gritty wind]

I was half way through opening an issue, should I keep writing it, or will it just be patched

[cold moon]

Let's write it, but attach critical label I think?

[gritty wind]

👍

[hardy gorge]

We can patch it immediately

[gritty wind]

Welp, alright

Alright, I think I have a patch ready

[jagged forge]

Hey Admin, is it allowed here to share python GitHub projects?

[gritty wind]

This is specifically for the python-discord projects

If you want to share your own projects, check out the general and offtopic channels

[jagged forge]

cool 👍

[gritty wind]

Would it make sense to include the ID on the leaderboard too?

[hardy gorge]

I think that would just be noise

[gritty wind]

Alright... what do I do with my code now lol

I seem to possibly have write access to master

[cold moon]

You are not core dev

[gritty wind]

Which is why I'm confused

I might just make a paste and let one of you handle it from here lol

[obtuse arrow]

Ah, ok, so it was because I'm also in the old PyDis leaderboard?

[gritty wind]

No, you have two entries in the old leaderboard

[obtuse arrow]

On another account?

[gritty wind]

Yeah

[obtuse arrow]

Ah, right.

[gritty wind]

https://paste.pythondiscord.com/godosaxezi.py

[hardy gorge]

@gritty wind Thanks. I'll create a patch from that. I'll change a few thinks to make the naming consistent (like name to member_id in that star result)

[stable mountain]

@green oriole

It has arrived!

Here's your reminder: polish the thing.
[Jump back to when you created the reminder](#dev-contrib message)

[green oriole]

Yes siiir

[gentle harness]

Chrome crashed when sending the new version, I am really sad

[mellow trench]

any specific reason snekbox requires Python 3.9?

[gentle harness]

Well, it allows us to run 3.9 code

[mellow trench]

ah

I didn't think of that lol

[thorny obsidian]

So I'm trying to create a custom cog-level error handler for a change to the AoC commands. I have it so that I can send the new error message, but the error still propagates through to the other bot-level error handlers. Is there a way to prevent that?

I only want the first one to appear, but I don't know how to stop the error from propagating through the handler

[tawdry vapor]

Try setting an attribute named "handled" on the exception

Never mind, that isn't checked by this bot

But this is https://github.com/python-discord/sir-lancebot/blob/master/bot/exts/evergreen/error_handler.py#L45

You can edit that to check for error handlers on the cog too

Or a custom attribute like @stable mountain does, which is more fine gained

[gritty wind]

That check specifically seems to be implemented on the global bot level using d.py, and there doesn't seem to be a direct override. I think you might be able to add an optional ignore parameter to the check itself, then decorate any functions you want to ignore

The check in question
https://github.com/python-discord/sir-lancebot/blob/master/bot/utils/decorators.py#L189

[thorny obsidian]

So I can't find an example in seasonal lancebot of the "on_error" being used to have a cog-level error handler, and my attempts so far to use it have not been successful. I feel like I'm missing something basic.

[gritty wind]

That on error seems to be specifically for the evergreen cog possibly?

[thorny obsidian]

For your potential proposed solution where you add an optional ignore parameter and then decorate functions, what exactly would that look like? I'm still a relative newbie to d.py

[gritty wind]

So a lot of the commands in that aoc cog have the @override_in_channel deco. What you could do is change the deco signature to accept a skip bool, which if true returns true, otherwise carries on to the rest of the predicate

I'm testing that right now

Any commands you'd want exempt from that check, would then get the skip parameter

[tawdry vapor]

I disagree

Seems reversed to how it should be

[cold moon]

@tawdry vapor Is https://github.com/python-discord/bot/pull/1016#discussion_r531343809 status code really 403? 404 seems more logic for not found.

[tawdry vapor]

Yeah I meant 404

[cold moon]

Okay, gonna make changes now

[gritty wind]

Well, the current way to override the global check is by manually using override_in_channel to add the channels you want to the allowed list. There would have to be a change on the in_channel_check level if you want to implement your own bypass functionality

The current iteration checks if the override_in_channel is part of the callback, so you could technically add a has error checker attr there

[tawdry vapor]

I thought it's just the message that needs to be changed

[gritty wind]

I thought the goal was to suppress the error propagation?

But I do agree about not using a skip parameter, I think a separate skip decorator might be better

[tawdry vapor]

Yeah, so a custom message can be used

[gritty wind]

Well, you could also have more logic than just a message

[tawdry vapor]

Maybe not for a custom message, but either way, the error can be stopped from propagating by adding a custom attribute that the global error handler will check for and skip if present. So it doesn't really stop it from propagating, it just makes it get skipped.

[gritty wind]

But I think I see what you're getting at

[tawdry vapor]

By reversed I mean that I feel that the handler should be modified, not the check.

[gritty wind]

Yeah that already exists with the hasattr on line 214

[thorny obsidian]

@eternal owl helped! It turns out I was missing something basic with how I was implementing it.

[tawdry vapor]

It doesn't work for cog error handlers, only per-command. Which I proposed to change

It's not really good anyway cause if a local error handler exists then it means it'd be responsible for handling all errors.

[eternal owl]

@tawdry vapor should we have the same way to check on both the bots if their is an error handler for that command/cog?

on @dusky shore we check if the command has the attr on_error but in @stable mountain we check if the error has the attribute handled

could we make it same on the both of bots?

[tawdry vapor]

Yes

Checking for handled is better

[eternal owl]

yep

Should I open an issue and a PR?

[tawdry vapor]

Sure

[eternal owl]

okay

https://github.com/python-discord/sir-lancebot/issues/538

[gritty wind]

Here's a counterargument: you don't want that global check to look for the error handlers, as they can be handling other errors, while that check is specifically responsible for one type of error.

Unless you get the error handler to propagate that error like normal anyways

[eternal owl]

The global handler will only handle errors which were not handled by the local error handler if done properly

[tawdry vapor]

Yes, that's why I am suggesting to check for an attribute on the error rather than looking if any error handler exists

[gritty wind]

Except the global handler as far as I can tell doesn't send the message about the channel being incorrect

Ah, I see

[eternal owl]

mark, can you look at the issue and approve it if right

[tawdry vapor]

The global handler doesn't control the message but it does propagate it

It has a condition to check for the InChannelCheckFailure exception

The check doesn't directly send a message; it just raises an exception containing a custom message

[thorny obsidian]

*grumble grumble* pre-commit hook is annoyingly helpful

[eternal owl]

its a pain on windows tho, it changes all my line endings in alll the files, thats why I just do flake8, lol

[gritty wind]

The hook itself should only run on changed files?

run lint would go over all the files, but not the pre-commit hook itself

[eternal owl]

oh yea, I meant run lint

[gritty wind]

Ah, yup

If you don't have any work saved right now, you could try setting autoctrlf

[tawdry vapor]

Change your editor to use LF and then you won't have problems

[eternal owl]

git still says that I have modified those files and have to commit them, lol

[gritty wind]

Yeah that would have to change in the autoctrlf setting

[tawdry vapor]

Your editor needs to be using LF and git needs to be checking out files as LF too

[gritty wind]

Git on windows by default checksout ctrlf, and commits LF

thus it tricks itself into thinking the remote repo uses crlf

[eternal owl]

lemme see

[tawdry vapor]

Just set it to false to disable it

[eternal owl]

where do I do that?

[tawdry vapor]

git config --global core.autocrlf false