also git config --global core.eol lf
#dev-contrib
tawdry vapor
shadow elbow
😂
tawdry vapor
I don't have time to review right now
tawdry vapor
I don't know if anything was relying on the previous behaviour
eternal owl
okay
Let me cross check again
the surprising part is, a cog was using error.handled = True
cold moon
Core Devs: Ready for merge: https://github.com/python-discord/bot/pull/1214
gritty wind
Created a script that would allow for identification of incorrect AoC leaderboard calculations:
https://paste.pythondiscord.com/hozunuwaxo.py
3 Relevant variables at the top have to be filled in
late wolf
Is there any reason, why the bot keeps on typing even after when the command is already displayed
gritty wind
Seems like a quirk of d.py
I tested with this, and it just held for a while after sending the message
@bot.command()
async def test(ctx):
async with ctx.channel.typing():
await ctx.send("Test")```ah i see
green oriole
Typing events will last for 5 seconds, because of the way discord work
manic echo
Hello, In this file, https://github.com/python-discord/bot/blob/master/tests/bot/test_constants.py
What is the use of line 14? - origin = typing.get_origin(annotation)
IMO the functionality will remain same i.e the assertion passes for non-normalised types too.
green oriole
We need to use normalized types in order do a proper isinstance check though
manic echo
`import typing as t
from collections import deque
t.get_origin(t.Deque)
<class 'collections.deque'>
isinstance(deque(), deque)
True
isinstance(deque(), t.Deque)
True`
Similar for list too, output for both is same.
green oriole
!e The issue is with
from typing import List
print(isinstance([1, 2, 3], List[int]))```
stable mountainBOT
@green oriole :x: Your eval job has completed with return code 1.
001 | Traceback (most recent call last):
002 | File "<string>", line 3, in <module>
003 | File "/usr/local/lib/python3.9/typing.py", line 657, in __instancecheck__
004 | return self.__subclasscheck__(type(obj))
005 | File "/usr/local/lib/python3.9/typing.py", line 660, in __subclasscheck__
006 | raise TypeError("Subscripted generics cannot be used with"
007 | TypeError: Subscripted generics cannot be used with class and instance checks
But it will work with get_origin
manic echo
!e The issue is with
```py
from typing import List
print(isinstance([1, 2, 3], ...
Thanks!
mellow hare
@green oriole It looks awesome! I really appreciate it
@eternal owl It'll commit them but when you try to push them it'll only send the ones you modified, not all the line ending ones
At least in my experience
eternal owl
okay
surreal venture
@nocturne hare Can you please drop a review on my pull request?
dusky shoreBOT
cold moon
@surreal venture You haven't still addressed my comment about Redis. You can see how Redis can be used in https://github.com/python-discord/sir-lancebot/blob/master/bot/exts/halloween/candy_collection.py
cold moon
Should I close https://github.com/python-discord/bot/pull/1275 because I just don't know what to do with it? I'd leave this to anybody smarter.
tawdry vapor
Yes, you can close it if you do not plan to continue working on it.
I think it's better to live with false positives and wait for a fix rather than fix that and introduce an easy way to circumvent the filter.
green oriole
... Oops
We tried to keep PRs open in the past to keep the work already done, I feel like we should do the same here?
Sorry, I didn't checked this channel first, my bad
tawdry vapor
That's situational. In this case, I don't think there's much to salvage from the current diff since a more comprehensive approach is required, which means something more complex.
green oriole
you never seen that
surreal venture
I do not know why, but I am unable to run seasonalbot. Every time I run pipenv run start, it gives me an error:
socket.gaierror: [Errno -2] Name or service not known
I have installed redis (from source) and have run pipenv sync --dev
oh wait do I need to start the redis-server?
At what port does the bot expect redis to be running?
Nevermind I got the port
Nope, it still isn't working
I am running Ubuntu 20.04.1 (AND I AM EXTREMELY EXCITED) with python version 3.8.5
Please do ping me if you know why it isn't working
sullen phoenix
@surreal venture you can enable fakeredis in the config
surreal venture
/etc/redis/6379.conf?
sullen phoenix
no, the bot's config
surreal venture
how do I do that?
sullen phoenix
so you won't need a redis server running at all
surreal venture
USE_FAKEREDIS in .env?
sullen phoenix
set use_fakeredis to true in your config
surreal venture
ok thank you!
sullen phoenix
`USE_FAKEREDIS` in `.env`?
no, that won't work
cold moon
SeasonalBot have this in environment variable USE_FAKEREDIS=true
sullen phoenix
oh, i missed that this was seasonalbot
i thought you were talking about @stable mountain, sorry
@surreal venture 
cold moon
Yeah, JSON have to be migrated to Redis for #509
surreal venture
Thanks!
surreal venture
Just one more question, when we initiate RedisCache, do we need to specify the namespace? Also, will it automatically load the data stored for that namespace (if there is some data)?
cold moon
When you define this inside class (like in example I linked), no.
surreal venture
Ok, since the namespace would be class_name.var_name by default?
hardy gorge
A RedisCache needs a namespace, as that's what tries to prevent us from having collisions in redis
Yeah, but you don't have to worry about that
RedisCache will automatically set that namespace and you never have to manually provide it
surreal venture
Ok thanks!
hardy gorge
Just make sure you assign a class attribute to it
surreal venture
And it automatically loads the data from the namespace?
hardy gorge
class MyClass:
some_attribute = RedisCache()
It does not load data automatically, you have to use the get and set methods (or any of the other methods)
Basically, we don't want to have a local copy of the data that could get out of sync with the state in the Redis database
Each time you want a value, you just ask for it and it will get you the value from Redis
surreal venture
Okay thank you!
surreal venture
I can't make my __init__ an async function, right?
How do I add some code in which I get something I need to await?
late wolf
use property decorater
surreal venture
isn't that for like getter and setter?
late wolf
u can use it as an attribute, so you could basically do
class A:
@property
async def something(self):
return "hello there"
async def somefunction(self):
await self.something #raises some error, but this is the basic idea u can use it as an attribute
I am not sure I understand
So like this is my code:
def __init__(self, bot: Bot) -> None:
self.bot = bot
...
self.some_attr = await something
How would I put that in my __init__?
late wolf
u don't,u can use self.something as an attribute.
__init__ can't have async and await
surreal venture
Oh ok I see what you mean now, thanks!
brazen charm
I don't really see how that's relevant to running async in an init
eternal owl
you can set self.some_attr to None and then load it using self.bot.loop.create_task(func)
@surreal venture
gritty wind
Other parts of the codebase schedule the tasks for separate execution, but I don't know if that is viable in your case
surreal venture
you can set `self.some_attr` to None and then load it using `self.bot.loop.creat...
That looks promising. I'll look into it thanks!
brazen charm
You can use a separate coro that you call manually or schedule it with the create task like above
surreal venture
is there HGET and HSET in async-rediscache?
(and sorry for asking so many questions)
hardy gorge
That's what a RedisCache uses
The namespace is the main key, the key you provide is the second key
surreal venture
What if I want to store a dictionary in a RedisCache instance? How would I go about doing that? Normally in redis-py you could do redis.Redis().hset() and redis.Redis().hget().
brazen charm
json if you need it for more keys
surreal venture
I was using json, but ks123 asked me to use redis instead
hardy gorge
You typically represent a dictionary by a RedisCache directly
Since that's what a RedisCache is more or less a stand-in for
A RedisCache is the hashset you're after
surreal venture
so I define another RedisCache for each nested dictionary?
Wait can I convert it to a string and load it back again using json.dumps and json.loads?
hardy gorge
Right, nested dictionaries is another story. You can't really do that with Redis either.
there's only key -> key -> value, which in Python translates to "name of the dictionary" -> "key" -> "value"
A single, non-nested dictionary would already be a hashset
A non-hashset "set" is like assigning a single name to a single value (no container other than the entire database)
surreal venture
So basically, I need my RedisCache to be like int: Dict[int, int, str] or something
surreal venture
Wait can I convert it to a string and load it back again using `json.dumps` and ...
Can I do this?
green oriole
Sure
hardy gorge
That should work (but: it could create potential race conditions)
The problem with serializing and deserializing JSON is that if you get this sequence of tasks, you lose data:
green oriole
That is true, json operations aren’t atomic
Could it be possible to make them atomic at the library level? Maybe using a lock or something?
hardy gorge
-
task 1 fetches data
-
task 1 deserializes and modifies local dict
-
task 2 fetches data (OLD data)
-
task 1 serializes as JSON
-
task 1 stores new JSON to Redis
-
task 2 makes changes to OLD dict
-
task 2 serializes
-
task 2 overwrites changes made by task 1
surreal venture
oh... that would be bad
hardy gorge
Could it be possible to make them atomic at the library level? Maybe using a loc...
There is a lock you could use to create mutually exclusive callables, but you need to be careful
It's a decorator
surreal venture
but this would apply if task 1 and task 2 are running in parallel or near parallel right?
green oriole
That’s cool
surreal venture
but this would apply if `task 1` and `task 2` are running in parallel or near pa...
Because if task 2 runs only after task 1 finishes writing, then it wouldn't create any problems?
hardy gorge
Well, our bots use asyncio and most of our functions await at intermediate points, meaning that, yeah, multiple tasks could run and alternate
hardy gorge
Because if `task 2` runs only after `task 1` finishes writing, then it wouldn't ...
That's correct, but it's up to you to make sure that they
surreal venture
Ok, thanks for your help, I'll see what I can do!
hardy gorge
The race condition is possible with non-JSON as well, but, with JSON, there's typically more going on (more simultaneous values that rely on a single key update)
surreal venture
Can I just use JSON to make the data persistent instead of using redis? What advantages does redis have over JSON?
glass pecan
redis is a hosted service that already exists and doesn't require us to assign a specific persistent storage volume to the bot service
because of our infrastructure, if we had a file-based persistance, we'd need to add a persistent volume to the hosted container that runs the bot
that costs us money
surreal venture
oh oh ok
glass pecan
since working with it is practically the same feel as json, it shouldn't be a hassle
surreal venture
Except, you can't store nested dictionaries
which is a really big down for me
since I have nested dictionaries
glass pecan
that's just a case of reorganising your data schema tbh
glass pecan
you can still store id references to cross reference other data if needed
cold moon
Otherwise data will get lost
surreal venture
Using JSON to persist isn't option anymore
But I can use JSON for one-time data right?
glass pecan
yep
cold moon
Yes, but this will be overwritten every time when bot redeploys
glass pecan
it's better to use a dict tho in that case
since it's going to live/die the same way
surreal venture
but it's a really long file
surreal venture
2207 lines
surreal venture
what
glass pecan
was thinking it would be GBs worth of data
gritty wind
I'm trying to figure out how the verified attribute gets set on the API, as I can't find it in the bot's code, nor can I get it to be set when a user verifies
green oriole
Hmm, the verified attribute on the discord objects, or on the site database @gritty wind?
gritty wind
Site DB
The users endpoint
it just returns null on all users
The rest of the verification process works normally and adds the role
tawdry vapor
What verified attribute?
brazen charm
Doesn' that just work through roles?
tawdry vapor
Do you mean the role?
gritty wind
api.site/users/{user.id}/metricity_data
Huh I wonder if metricity sets it
I mean, I have metricity running too
green oriole
I was going to check django admin, but I cannot login with discord
tawdry vapor
Yes that is a metricity-only thing
green oriole
Yeah, we don't have that on the site db
gritty wind
So... metricity sets it?
tawdry vapor
Well it is on the site DB. I don't think there's a separate postgres instance.
Yes metricity sets it
gritty wind
Oh, if that's what you're referring to, I have them set to the same DB
I'm just trying to get it to update though
green oriole
It is on the postgres instance, but not on the site db, metricity is using a separate database
gritty wind
I'm reviewing a PR that's using it
Which endpoint would that be?
https://github.com/python-discord/site/issues/398
This specifies that this endpoint should return it
tawdry vapor
Probably the one you sent earlier
gritty wind
But.. wait how do I connect the site API to the metricity DB?
Sorry, I'm still figuring out my setup for the projects
tawdry vapor
It automatically connects
gritty wind
Huh, well it returns different values from what is stored in the DB
I'm not seeing any errors anywhere either
tawdry vapor
How are you checking what's in the DB
gritty wind
I was selecting the columns using SQL directly
And there is some connection I would think, because the message counts is reported
tawdry vapor
Are you using the default db connection string?
gritty wind
For the site itself, I'm just running docker-compose up, no changes, and for the metricity I'm not changing anything (the only change I have it getting it to connect through host network)
Metricity also has the correct channel IDs
tawdry vapor
What change exactly?
Are you sure the bot is connecting to the same db as the site?
gritty wind
Well, same postgres instance, but metricity uses metricity db, the site uses the site
Let me grab them
tawdry vapor
Yeah that's what i mean
gritty wind
Should it not be setup like that?
tawdry vapor
It should
I was asking if the site connects to the same metricity db that the bot connects to
gritty wind
Ah, in that case yeah
I can test to confirm
Ok update
It is definitly looking in the right place but
It selects the verified_at, which is never updated for me, even when the actual verification status is set
tawdry vapor
Did you set the right IDs in the config?
gritty wind
The metricity IDs?
tawdry vapor
The role ID in metricity's config
gritty wind
Yeah, that one is set
I can try to set it again and verify again
But I figure if it did detect the verrification...
tawdry vapor
What do you mean detect verification?
gritty wind
It set the verified status to true for my test account
tawdry vapor
You mean when you manually query the db?
gritty wind
Yes
tawdry vapor
But it doesn't show that through the api?
gritty wind
Is it working for you?
There is is_verified and a verified_at
the api uses the verified_at
Well...
I changed nothing
tawdry vapor
I've never tried running it so I don't know
gritty wind
Just reverified and
it works
I must've updated it at some point
Thanks for helping me :D
surreal venture
What is dump.rdb?
gritty wind
A Redis dump of your active DataBase cache
Here is the redis page on data persistence https://redis.io/topics/persistence
Do you have specific questions about its connection with the projects?
surreal venture
it shouldn't be commited right?
gritty wind
No, it shouldn't
surreal venture
Okay, that's all I wanted to know. Thanks!
surreal venture
@cold moon I have changed my code to use RedisCache to make the data persist, can you check if the code is good to go or if it needs some more improvements?
@nocturne hare Can you also review my pull request?
dusky shoreBOT
clever wraith
anyone help me....how to run this...?\
brazen charm
You'll want to follow the guide at https://pythondiscord.com/pages/contributing/bot/
Mind that running the bot locally is really only useful for development on it. If you need specific features out of it for your own use it'll be easier to just take out those features and make them work without the bot's environment
cold moon
Why Core Dev approvals use checks? This isn't really nice when so much commits show ❌ even when linting success, but core dev approvals not. This should be label instead I think.
gritty wind
We could just have a dummy account on the whole repo to prevent prs not approved by core devs, without failing all the actions
green oriole
That was one of the proposal, using the github bot for that, but we don't actually require changes
gritty wind
Well the check was implemented to allow core-devs to remove themselves from getting messages for the whole repo, but if we add just one account thats required for all prs, we can stick with the old system, but not have to fail commits
patent pivot
we're actively changing it right now
the way it's going to work is that the PR will display pending
similar to this
gritty wind
Does the commit status show as failed or pending when there are failed and pending tasks?
I image it would resolve lazily?
patent pivot
well, it's as you say there
failed shows failed, pending shows pending
oh you mean if there is still a pending task when there is a failure
good Q, I think it displays failed, but at the bottom of the PR it will always display
new system is running on forms backend now
I approved now, but that check was pending
green oriole
So it is working?
patent pivot
forms frontend and backend have it now
if you visit https://policy-bot.pythondiscord.com/details/python-discord/forms-backend/13 you can see the status of PRs
looks like this if pending
this if approved
green oriole
Hmm, the apps has some scopes I'm not sure it really needs
patent pivot
(these links appear on the PR checks)
green oriole
green oriole
Well, it would need it only for the account it is configured on, not for everyone who visit the website
crude gyro
A check is the correct thing to use.
a failing check doesn't mean "linting error"
that's just what you're personally used to
a failing check means "this isn't ready to merge"
a PR without a core dev approval is not ready to merge.
green oriole
it might just not be specifying the scope and that is the defaults
That is some pretty aggressive defaults then haha
patent pivot
lol yeah
gritty wind
As a counter, if I look at a PR and see failing tests, I will assume the author is working on updating the PR and wasn't using a draft
If all PRs will be marked with a fail, what use is it
patent pivot
policy-bot will continue to use checks, but stay in the "pending" stage (as shown in #dev-contrib message)
so you won't get the X, but you will not get the green for go
crude gyro
I think pending is infinitely worse and I hate this change.
gritty wind
so you won't get the X, but you will not get the green for go
If everything is setup correctly, yellow would be the new green
I get what you're trying to say, but having everything red would make everything useless for the most part
patent pivot
I don't really see how pending is much worse, I think that a description saying the rules are not met and then linking to a site where you can see a visual representation of why makes sense
patent pivot
and this allows us to have more complex rules, like "if it touches a CI file do not pass until a devops member reviews"
crude gyro
pending, to me, means "checks are running and haven't completed yet, but this will soon change". Having everything in pending really triggers me a million times harder than having failing checks - which is semantically correct.
a core dev has not approved. we're checking if the core dev has approved. It's a failed check.
gritty wind
If you are really set against pending, we can compromise and only have it run on PRs that don't have a core-dev by default
cold moon
I have more seen pending than failing for such cases
gritty wind
That way what is pending is the assignment of a core dev to the pr
patent pivot
yeah, I think pending is the github norm, it's what I've used at work and on other repos where checks like this are enforced
crude gyro
I've never seen pending used like this
but okay
if anyone in the world has used it like this, I guess it's marginally less gross to me
What happens if the core dev requests changes?
gritty wind
I suppose, another alternative alternative would be to give only core devs write perms
crude gyro
does it then fail?
cold moon
I don't think giving write permissions only to Core Devs is good idea
patent pivot
hmmm, I believe that by default that is a success since it has received a core developer review.
@cold moon can you open a dummy PR to forms-backend
just change a file or something
crude gyro
so, if it failed the check on changes requested, passed it on approval, pending if nothing, then I can sort of see how that makes semantic sense
cold moon
I will change gitignore
patent pivot
cheers ks
actually before you do
let me just update the org review policy
to set this disapproval thing
crude gyro
pending what. the core dev review has happened. It was not an approval. There is nothing pending. You have to make changes.
patent pivot
okay @cold moon policy updated
crude gyro
it sends an email? good, because you should be taking action.
patent pivot
okay so we're in a pending stage on ks's PR now
crude gyro
I've totally missed that this was a thing
patent pivot
🎉
crude gyro
nice
patent pivot
crude gyro
that works for me. does the check really have to be called policy-bot main though
let's give it a very descriptive name
patent pivot
we can rename the first bit, I think the latter is just saying it's the review requirements for main
crude gyro
yeah, something like that
patent pivot
one second
cold moon
Core Dev Police
patent pivot
so Core Dev Approval: main is how it will look
patent pivot
yeah
crude gyro
but yeah I can live with this. thanks for changes.
you're very accommodating
I'll send the money to the usual account
patent pivot
hahahha
and then that Details link will link to the page I was screenshotting above
if we ever make more complex rules about CI/devops or whatnot then that'll appear in the flow
patent pivot
actually @green oriole it won't generate an extra email for requested review
it'll send the usual one when someone requests changes
but since it isn't github actions you don't get a vague "run failed" email
crude gyro
yeah. because that won't trigger it
patent pivot
yep
the actual policy for this: https://github.com/python-discord/.github/blob/main/review-policies/core-developers.yml
contrib approval on this would be funky fresh https://github.com/python-discord/site/pull/431
green oriole
Alright cool
surreal venture
Whenever I run a command, it is running twice for me, is it happening with anyone else?
gritty wind
The mod bot?
gritty wind
Huh, could you have an extra instance running without knowing?
surreal venture
Hmm... maybe that's the case
gritty wind
Try disabling one, and see if the commands still get processed
If you want to force stop all instances, reset your token
nocturne hare
<@!188735535169273858> Can you also review my pull request?
Can look at it in a few, not home yet
surreal venture
Sure, take your time!
nocturne hare
@crude gyro Can you iterate keys in the Async RedisCache?
Or do you have to call .items, and index/unpack the key from index 0
brazen charm
You'll have to iterate items
crude gyro
we used to have a keys but I think it was abandoned in a later iteration
we only model the actual redis stuff now.
when it was called RedisDict, I did have a .keys for it.
gritty wind
@trim cradle is bot/1319 ready for review or does it still need discussion?
green oriole
We are discussing it atm 
trim cradle
@gritty wind we're discussing it in a mod channel. Thanks for asking!
gritty wind
Good stuff
green oriole
You can still review it though
gritty mortar
Hi devs, I wanted to start contributing, mostly reviews to the community, can someone help to identify where and what to pick up. Apologies if this is the wrong channel.
gritty wind
To start reviewing, I'd start by following the guides on the website (https://pythondiscord.com/pages/contributing/) to set up all the projects (+ metricity). From there, pick one of the source repos (https://github.com/python-discord), and pick a PR. There isn't a reviewing guide yet, but keep your eyes open.
And, this is the right channel for such discussions.
gritty mortar
Thanks I will start with the setup
gritty wind
Once you do, you can start looking through the PRs for all three till you find something you feel capable of contributing to (any comments, suggestions, or modifications, even a 👍 can be helpful).
sullen phoenix
the reviewing guide is coming very soon™️
i’m think hem is finished with it actually
if that’s the case, it just needs to be released
green oriole
Hey @eternal owl, on sir-lancebot#540 you have a 2.2k line changes because of line endings, can you revert that please?
eternal owl
wait, I thought line ends are ignored :/
gritty wind
Only if you didnt change your default settings
eternal owl
ahh
gritty wind
You should be able to fix that by changing the autocrlf setting, and redownloading your project
patent pivot
sir lancebot, site, bot and forms frontend/backend are now all using policy bot
obsidian patio
sir lancebot, site, bot and forms frontend/backend are now all using policy bot
What's policy bot?
patent pivot
gritty wind
It forces all PRs to need a core-dev approval to be merged. It is used so we get pending status instead of a failed check
obsidian patio
Ahh, okay. Was this discussed before I arrived at the staff meeting?
gritty wind
it was not discussed in a helper staff meeting
scroll up in this channel for discussion
patent pivot
nah I did mention it in the news section
gritty wind
Why Core Dev approvals use checks? This isn't really nice when so much commits s...
here
nevermind then
patent pivot
What's policy bot?
it has taken action on https://github.com/python-discord/site/pull/432 now
quasi crypt
Anyone know where I might find an expert on snekbox? I have an issue... Seems like if I send the server two requests at approximately the same time one or both gets an out-of-memory error (return code 137).
gritty wind
Sebastiaan was working on OOM issues recently, they might be able to work on this. I'll point him towards here if I see him
But out of curiosity, can you include some specific info like os, and snekbot version?
patent pivot
Anyone know where I might find an expert on snekbox? I have an issue... Seems li...
cc @tawdry vapor
tawdry vapor
That error may be misleading since it just assumed that SIGKILL = killed due to OOM but really it could be coming from anywhere.
NsJail doesn't log OOM termination because it's actually handled by the kernel via cgroups. NsJail only directly handles killing for an exceeded time limit and therefore you'd see a log for that.
It may be an issue with how cgroups are set up where all nsjail processes share the same cgroup. I thought it would apply separately to each process but maybe it's more like a pool.
Can you open an issue with reproduction steps so I can remember look into it later?
quasi crypt
@tawdry vapor definitely. Thanks for being willing to look into it.
quasi crypt
@shadow elbow @gritty wind https://github.com/python-discord/snekbox/issues/83
Seems like it is a memory issue based on my testing... Hm.
shadow elbow
<@!743745980633186387> <@!386599446571384843> https://github.com/python-discor...
May be you've tagged the wrong mark.
trim cradle
If I run pipenv run pre-commit in powershell, it says everything is fine. But PyCharm says that flake8 fails when I click "commit".
brazen charm
fails in what way?
trim cradle
Flake8...................................................................Failed
- hook id: flake8
- exit code: 1
Courtesy Notice: Pipenv found itself running within a virtual environment, so it will automatically use that environment, instead of cre... (show balloon)
``` @brazen charmI suppose I could try committing it using powershell
does git commit -m recognize \n?
brazen charm
I believe you can chain multiple -m for the subject and body
Do you have a global flake8 that pycharm could be picking up?
trim cradle
so git commit -m "main message" -m "secondary message"?
trim cradle
Do you have a global flake8 that pycharm could be picking up?
I don't know, I haven't used anything like flake8 outside this project.
I was able to make commits earlier today
sullen phoenix
so `git commit -m "main message" -m "secondary message"`?
you can just run git commit and it'll open your default editor set for git
once you type in the commit message and close the editor, it'll commit
i just use nano for git
trim cradle
For some reason that didn't work either
PS C:\Users\Steele\development\projects\pydis\bot> git commit
Check for merge conflicts................................................Passed
Check Toml...........................................(no files to check)Skipped
Check Yaml...........................................(no files to check)Skipped
Fix End of Files.........................................................Passed
Mixed line ending........................................................Passed
Trim Trailing Whitespace.................................................Passed
check blanket noqa.......................................................Passed
Flake8...................................................................Failed
- hook id: flake8
- exit code: 1
Courtesy Notice: Pipenv found itself running within a virtual environment, so it will automatically use that environment, instead of creating its own for any project. You can set PIPENV_IGNORE_VIRTUALENVS=1 to force pipenv to ignore that environment and create its own instead. You can set PIPENV_VERBOSITY=-1 to suppress this warning.
Loading .env environment variables...
'flake8' is not recognized as an internal or external command,
operable program or batch file.
I'll look into it some more tomorrow
brazen charm
Run it under the pipenv environment
sharp timber
^ it looks like pycharm is not using the virtual environment set in the directory for some reason (?)
gritty wind
Someone told me that was a bug before
(Windows specifically)
You'd either update the hook to make sure it actually uses the environment, or you change the pre-commit settings to run on system which would then use pipenv
The second solution won't work though, because you'll have to commit the modified pre-commit
So you're left with modifying the file
surreal venture
@trim cradle
'flake8' is not recognized as an internal or external command,
operable program or batch file.
Are you sure you have it installed?
gritty wind
Its probably installed in the pipenv, but the precommits don't use the env correctly on windows
(and possibly other oses)
surreal venture
Well, you can always try installing it...
because pre-commit is working fine for me on Ubuntu
I uninstalled my global flake8 and ran pipenv run lint and it worked well (using the virtual environment)
gritty wind
The thing about the global flake8
is that its caused problems for me and a couple other people, by reporting false negatives (returning no errors when the workflow fails on github)
Might've even been on one of steler's commits, i don't remember
gritty wind
Uhh @patent pivot https://github.com/python-discord/bot/pull/1319
It seems to just be using any disapproval, is that intended?
cold moon
I think this is meant that all users with write access that require changes count to there
But only Core Devs approvals count
gritty wind
The check should be renamed then?
gritty wind
There can be cases of disapprovals, with approving core devs
cold moon
Not sure how this work in such cases.
gritty wind
It'll still fail I'd assume
patent pivot
It seems to just be using *any* disapproval, is that intended?
It's using a disapproval from org members. The check name is slightly misleading I guess, I'm open to any alternatives. Maybe just "Review policy"
hardy gorge
@patent pivot Does it take core dev membership into account in the check?
What happens if a regular org member approves, but a core dev has not yet approved yet?
gritty wind
They are two independent checks on the website (so it'll still be pending), but any disapproval would mean failure apparently
patent pivot
Yes, core developer approval passes the check, disapproval from anyone in the org fails the check
crude gyro
that's fine
hardy gorge
That looks like the right setup to me
hardy gorge
Shiny: #dev-log message
obsidian patio
Nice 😄
surreal venture
Just curious, do changes in the .github folder not show as a commit? Because I just updated my branch and git pulled, and it does show changes, but it doesn't show as any commits in my forked repo:
(I forked it before the name change, if it helps)
trim cradle
@gritty wind I hadn't yet pushed what I have locally when you gave your feedback on github, so I converted the PR to a draft to make that more clear.
gritty wind
It wasn’t a super detailed review anyways, I’ll check back on it when it’s ready 😄
trim cradle
Is it possible that flake8 is objecting to my commit message, and that's why I can't commit?
hardy gorge
What kind of output are you getting, @trim cradle ?
trim cradle
...
Flake8...................................................................Failed - hook id: flake8 - exit code: 1 Courtesy Notice: Pipenv found itself running within a virtual environment, so it will automatically use that environment, instead of creating its own for any project. You can set PIPENV_IGNORE_VIRTUALENVS=1 to force pipenv to ignore that environment and create its own instead. You can set PIPENV_VERBOSITY=-1 to suppress this warning. Loading .env environment variables... 'flake8' is not recognized as an internal or external command, operable program or batch file.```PyCharm is using the correct virtual environment, but my guess is that it's not using the pipenv that's in that environment.
if that even makes sense
Yes, I'm using a "Virtualenv environment" and not a "Pipenv environment"
hardy gorge
pre-commit uses pipenv run to run flake8 to make sure it selects the appropriate environment
I assume that's what is not working for you
What I do is create an environment with pipenv sync --dev from the command line and then use the "existing environment" option in PyCharm, selecting the pipenv venv, as I don't like how PyCharm manages pipenv environments.
However, what happens if you try to run flake8 manually with the venv activated?
trim cradle
However, what happens if you try to run `flake8` manually with the venv activate...
it runs fine
hardy gorge
what does it give you for flake8 --version?
gritty wind
Theres a pycharm bug (possibly just on windows)
trim cradle
what does it give you for `flake8 --version`?
that command doesn't work, with or without the venv activated. However it might be because pycharm is doing something with that environment currently to recognize it as a pipenv.
gritty wind
Someone walked me through fixing it a while ago
and no, a system install doesnt work as intended
green oriole
What does pipenv run lint gives you? Or you don't use pipenv at all?
gritty wind
trim cradle
What does `pipenv run lint` gives you? Or you don't use pipenv at all?
everything passes
hardy gorge
It could still be that it bypasses your environment, uses a flake8 version installed systemwide, without our plugins
trim cradle
I think it might be working now. One moment and we'll see
clever wraith
how do i make a dockerfile for snekbox
shut aspen
how do i make a dockerfile for snekbox
you dont, it already has one
clever wraith
how can i use it?
shut aspen
but its also on docker hub
gritty wind
The code review guide is done!
"sw" + "e" * 231 + "t"
mellow hare
HA
clever wraith
yeah
shut aspen
this is my entire snekbox code for my bot
mellow hare
`"sw" + "e" * 231 + "t"`
I just hope it reads well. And that people benefit from it
clever wraith
now im getting this
https://paste.pythondiscord.com/qusoboxesu.rb
gritty wind
I just hope it reads well. And that people benefit from it
There was someone asking about how to get into reviewing PyDis code a couple of days ago, could have them dry run it
shut aspen
what is that?
clever wraith
container list
gritty wind
Can you upload the output of running dir as well?
(btw, you don't need to build the image yourself if you are just using it)
gritty wind
Yeah
shut aspen
no its on docker hub
clever wraith
yeah
gritty wind
You only need to build it if you're modifying it yourself
clever wraith
Hey guys I am new
gritty wind
Hello
clever wraith
shut aspen
i use compose
clever wraith
I wanna learn python!
shut aspen
but its the same thing basically
gritty wind
It was erroring out because the copy instruction failed
clever wraith
please help me out
gritty wind
There should be a config folder there too
clever wraith
```docker run --ipc=none --privileged -p 8060:8060 ghcr.io/python-discord/snekbo...
already ran that
shut aspen
what did it do
gritty wind
@clever wraith please check out #❓｜how-to-get-help for general python help
clever wraith
thank you!!!!
clever wraith
thank you
gritty wind
But this is only relevant if you still want to build yourself
clever wraith
why do people scroll down to the very bottom, and go to #dev-contrib instead of #❓｜how-to-get-help
gritty wind
Should be, yeah
shut aspen
snekbox:
image: pythondiscord/snekbox:latest
init: true
ipc: none
networks:
- ClemBotNet
ports:
- "8060:8060"
privileged: true```this is literally the only time snekbox is mentioned in my botcompose pulls the image, runs it on the docker network and assigns those ports
gritty wind
You can get the address and port through docker desktop
shut aspen
ideally you run BOTH yourbot and snekbox as containers
clever wraith
so how do i make requests to it? as in what do i include in the request
shut aspen
then you can network them with docker
clever wraith
and do i have to make requests with a specific module or
clever wraith
oo
shut aspen
have you ever worked with an api before?
like http stuff
or is this your first time
gritty wind
(Also don't forget its a POST endpoint)
clever wraith
i can make requests with aiohttp, right?
shut aspen
yes
clever wraith
good
shut aspen
this is how i did it
go wild
as you can see its dirt simple
async def _post_eval(self, code) -> t.Union[str, None]:
data = {
"input" : code
}
json_data = json.dumps(data)
async with aiohttp.ClientSession() as s:
async with s.post(BotSecrets.get_instance().repl_url,
data=json_data,
headers=HEADERS) as resp:
if resp.status == 200:
return json.loads(await resp.text())
so it returns text with the eval result?
shut aspen
it returns that json that you saw in the comment i just linked
shut aspen
its in the comment lol
clever wraith
what comment
shut aspen
np
clever wraith
@shut aspen it says localhost:8060 is an invalid url
should i replace it with my ip address
shut aspen
that is an ip
shut aspen
well no cuz its not running on your public ip
"ReplUrl": "http://localhost:8060/eval",
this is in my config
clever wraith
damn i even forgot the /eval oops one sec
one sec i forgot the headers
im getting http code 400
gritty wind
What does your code look like right now?
shut aspen
async def _post_eval(self, code) -> t.Union[str, None]:
data = {
"input" : code
}
json_data = json.dumps(data)
async with aiohttp.ClientSession() as s:
async with s.post(BotSecrets.get_instance().repl_url,
data=json_data,
headers=HEADERS) as resp:
if resp.status == 200:
return json.loads(await resp.text())
hm? @shut aspen
shut aspen
how does yours differ from mine
clever wraith
you used json.dumps?
shut aspen
yee
clever wraith
Ignoring exception in command eval:
Traceback (most recent call last):
File "C:\Users\vndev\AppData\Roaming\Python\Python39\site-packages\discord\ext\commands\core.py", line 85, in wrapped
ret = await coro(*args, **kwargs)
File "C:\onyx\onyx\main.py", line 54, in eval
async with s.post("http://localhost:8060/eval", data = {"input": code}, headers = HEADERS) as resp:
File "C:\Users\vndev\AppData\Roaming\Python\Python39\site-packages\aiohttp\client.py", line 1012, in __aenter__
self._resp = await self._coro
File "C:\Users\vndev\AppData\Roaming\Python\Python39\site-packages\aiohttp\client.py", line 376, in _request
headers = self._prepare_headers(headers)
File "C:\Users\vndev\AppData\Roaming\Python\Python39\site-packages\aiohttp\client.py", line 836, in _prepare_headers
headers = CIMultiDict(headers)
TypeError: CIMultiDict takes either dict or list of (key, value) pairs
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "C:\Users\vndev\AppData\Roaming\Python\Python39\site-packages\discord\ext\commands\bot.py", line 903, in invoke
await ctx.command.invoke(ctx)
File "C:\Users\vndev\AppData\Roaming\Python\Python39\site-packages\discord\ext\commands\core.py", line 859, in invoke
await injected(*ctx.args, **ctx.kwargs)
File "C:\Users\vndev\AppData\Roaming\Python\Python39\site-packages\discord\ext\commands\core.py", line 94, in wrapped
raise CommandInvokeError(exc) from exc
discord.ext.commands.errors.CommandInvokeError: Command raised an exception: TypeError: CIMultiDict takes either dict or list of (key, value) pairs
i get this traceback
i did HEADERS = json.dumps(HEADERS)
shut aspen
the headers can be a normal dict
shut aspen
i think you can do json={dict}
but i did data=json.dumps
probably the same thing
try both /shrug
gritty wind
code = None
clever wraith
lmao
gritty wind
You can actually just delete lines 8 & 9
They don't do anything useful
(and as more of a sidenote, don't name your variables after builtin python ones, such as input)
clever wraith
(and as more of a sidenote, don't name your variables after builtin python ones,...
yeah i know, but i thought it wouldnt affect it as im not using the input function and it's only in the scope of the command
clever wraith
You can actually just delete lines 8 & 9
wait but wouldn't that cause scoping issues since im defining that variable in a with/if statement?
gritty wind
It is, but why introduce unneeded complexity ig
^ That was to your first comment on input
Your else currently has a return
clever wraith
oh
gritty wind
Which means your variables are only used if you go into the if
gritty wind
Yeah
clever wraith
ok
gritty wind
!e
if True:
a = "hello world"
else:
exit()
print(a)```@gritty wind :white_check_mark: Your eval job has completed with return code 0.
hello world
how do i install a package from github instead of pypi
gritty wind
A github python package? It would depend on the package, but there should be instructions on the package's repo
trim cradle
A github python package? It would depend on the package, but there should be ins...
when is it ever not pip install git+...?
gritty wind
When you have a terrible programmer such as myself publishing it
late wolf
Is there an issue open on making the setup for contributors easier?
If so could someone link it to me, i would like to take a look,
happen to ping me
green oriole
Hey @late wolf, we are currently working on a project, smartconfig, in order to simplify the whole setup process. Once it is done, you can be sure that I'll work 'em! In the meantime, if you have any idea or suggestion for the guides, feel free to comment on meta#60
.issue 60 meta
dusky shoreBOT
late wolf
ohh
slim widget
Hey, I was wondering what people think about adding a .tldr command to Sir Lancebot.
It would retrieve the tldr-page for the given shell command: https://github.com/tldr-pages/tldr
I can vouch for the quality of these pages. I use them regularly.
green oriole
I didn't know about this website, I'd 100% support this, it looks very cool
eternal owl
isn't that similar to man pages?
patent pivot
the readme sums up how it isn't manpages
slim widget
Yep, but they're shorter and emphasise common examples.
eternal owl
alright
slim widget
This might make a good first-contribution for someone?
Since I think it should be relatively straightforward.
eternal owl
I think, yes
I think using a line paginator will also be helpful
during the implementation
just in case
patent pivot
yep
green oriole
Hmm, I am not sure, I usually label as first good contrib anything that would help you familiarize with the code base, that isn't really the case here
That could work, I dunno
Yeah, if it is a well written issue that should be fine
sorry thinking out loud again
trim cradle
self.guild = self.bot.get_guild(Guild.id) -- The expression on the right returns None when this is in the init method for a cog, I guess because it runs before the bot is fully online. I checked the docs to see if there's an async method for cogs that runs once the bot is fully online and didn't see one. Is there something I'm missing?
trim cradle
`wait_until_guild_available`
🔥 👌🏻
sullen phoenix
just await that method before getting the guild, and you'll be fine
trim cradle
just await that method before getting the guild, and you'll be fine
I don't believe you can await in an init method?
sullen phoenix
no, i'd await in on_ready
or
you can create an async method and then create a task for it in __init__
let me find an example
trim cradle
def __init__(self, bot: Bot) -> None:
self.bot = bot
@Cog.event
async def on_ready(self) -> None:
await self.bot.wait_until_guild_available()
self.guild = self.bot.get_guild(Guild.id)
self.developer_role = self.guild.get_role(Roles.verified)
so I can't just do this?
sullen phoenix
class SomeCog(Cog):
def __init__(self, bot: Bot) -> None:
self.bot = bot
self.guild = None
self.bot.loop.create_task(self.get_guild())
async def get_guild(self) -> None:
await self.bot.wait_until_guild_available()
self.guild = self.bot.get_guild(Guild.id)
no, you have to wait for the guild to be available
either way works
trim cradle
how about now?
sullen phoenix
i'm unsure if Cog.event exists, i've only seen Cog.listener()
but other than that, yeah
it's also good practice to initialize those variables in __init__ as None
trim cradle
PyCharm was insisting that I do that, but I figured this wasn't the edge case they were discouraging.
brazen charm
imo it's not really necessary when you have a clear initializer coro like that, just adds some noise to the init
Make sure that the coro finished before going to the parts of the code that use the vars and you should be good
late wolf
Hey so I had an idea.
Instead of us having to replace all the id's manually, why not make a setup command, which creates the config-yaml file, gets their id of the channels using their names, and adds that to the config-yaml file.
short snow
https://cht.sh/ a similar resource to tldr pages. but more better i feel.
late wolf
ping me if u have any thoughts on #dev-contrib message
sullen phoenix
@late wolf we’re already working on a project called smartconfig
late wolf
oh ok
cold moon
I think https://github.com/python-discord/site/issues/406 can be closed now. Linked PR is merged.
I think one part of site is missed in dewikification: Contributing section. This should be in different app. Should I create issue?
patent pivot
I thought we settled on using the content application for that
cold moon
Content app doesn't support bulma widgets like contributing pages have
surreal venture
Just curious, do changes in the .github folder not show as a commit? Because I j...
Does anyone know why this happens?
hardy gorge
I'm not sure.
short snow
https://cht.sh/ what do u guys think about this compared to tldr pages. it is easier to get data from this in a nice way. i have made code also for it
hardy gorge
Hey @eternal owl, could you give me an update on the error handler PR for @dusky shore? Maybe we can work out a way to get it to a mergeable state soon so that we can merge the Advent of Code changes soon as well.
gritty wind
As a counter argument, tldr can be self hosted, and has a python library, while cht.sh would require extra http requests and parsing to remove the non-tldr section.
The page mentions a github page, which isn't linked, so I haven't searched too thoroughly
thorny obsidian
The tldr makes it clearer what's available and how to use it. cheat.sh doesn't make it obvious what exactly I'm supposed to do. Their github makes it slightly clearer, but I don't think I'd remember the syntax for it
gritty wind
There is a TLDR header on some pages which is pretty much what the tldr content is
But not all pages have it, and some have really bad formatting
Hey @hardy gorge mind sharing what might be wrong with the webhook?
Is it just authentication or something
green oriole
Which webhook?
hardy gorge
It is
It's what I feared and kinda hoped
Some secrets are not available in a pull_request run from a fork
gritty wind
That seems like a mostly good thing
hardy gorge
yes
but there was no easy way for me to test it
Anyway, I have a solution ready, which also ensures that we're actually running our workflow file instead of a potentially modified one in a fork
green oriole
workflow_run?
hardy gorge
That could work, but that will make getting information about the run into the webhook embed more difficult later (which is something that's planned for later)
At the same time, it could be the safest
green oriole
Hmm
hardy gorge
You know what, I'll figure out the passing of output later. There are other ways of doing that.
green oriole
Can't get everything you need from the API, or GraphQL?
hardy gorge
Let's just fix this
eternal owl
@hardy gorge I think it will be better to use error.original in the local error handler (snakes cog) as the FileNotFoundError (aka OSError) does occur if the file is not found (example: https://github.com/python-discord/sir-lancebot/blob/master/bot/exts/evergreen/snakes/_snakes_cog.py#L178)
just error will be CommandInvokeError
I am not sure how the .snake video and .snake get commands would raise an OSError
but the .snake card command can raise one(check the example link above)
hardy gorge
I'm not that intimately familiar with the code and not looking at it right now; do any of those commands use Pillow's Image?
hardy gorge
Because Image.open raises an OSError when an empty stream of bytes is passed to it
I suspect that's why they added the handler
eternal owl
not empty, but even a file which is not found
hardy gorge
We create a BytesIO from downloaded bytes, right?
If that BytesIO stream does not yield bytes, Image.open does raise an OSError
green oriole
That's a weird behavior
eternal owl
We create a BytesIO from downloaded bytes, right?
yes
then that stream is passed into Image.open
green oriole
All in all, do we even need the special case? I haven't looked into it too much, it just change the error message a little bit, right?
eternal owl
yes, just gotta change it use error.original instead of just error
in the isinstance
hardy gorge
That's a weird behavior
It's because OSError is raised for when I/O fails. This is usually when the file stream you're trying to read does not give you anything (or is inaccessible), but we're using a virtual in-memory stream.
green oriole
I mean, OSError isn't very adapted here, nothing goes through the OS
hardy gorge
It's because you'd normally pass it an actual file stream.
It's not that Pillow tries to work out whether or not you're using an actual file or faking it using an in-memory file object-like stream
It just tries to read from your (potentially pseudo) I/O stream
And if that fails, it raises the appropriate error
eternal owl
I just tested with an incorrect path and it gives me FileNotFoundError
green oriole
I'm not sure if the FileIO abc group intends you to use OSError here, since the actual methods can take the actual data from anywhere
well, that's not really the point anyway
hardy gorge
It kinda is the appropriate error, as OSError is the exception defined by the docs to be the one to be raised for I/O-related errors.
green oriole
when a system function
eternal owl
I just tested with an empty buffer, raises OSError
green oriole
And here you are just reading from memory using a non-system method
Also, you don't have any errno to fill in
eternal owl
yes, just gotta change it use `error.original` instead of just `error`
should I go ahead and do this?
btw, the actual image error is UnidentifiedImageError
which comes under OSError
also, I think I can remove the error handler for get_command and video_command, I don't see how these 2 commands can raise an OS error.
green oriole
Let's keep this for another PR that isn't time critical 
Yeah, I think you can do that, what do you think @hardy gorge ?
trim cradle
Who's bot is running on the test server, out of curiosity?
green oriole
Senjan's and Hem's bots are running
trim cradle
Just out of curiosity, how expensive is it to have the same kinds of event listeners in multiple cogs? I assume there's one component of the bot that's actually doing the listening, and when that happens it calls the event listener in each cog.
gritty wind
Yeah pretty much, it calls each handler until it runs out. The overhead isn't huge (you can see the dispatcher here: https://github.com/Rapptz/discord.py/blob/master/discord/client.py#L361), but the price would depend on how long the handler itself takes.
copper pawn
Hey I am looking to learn the pytest library. Does anybody need some testing for a python project so that I can practice? Add me/message me if you do
gritty wind
Hey, this channel is specifically for discussion about the python-discord projects (github.com/python-discord/), if you'd like to talk about python testing in general, check out #unit-testing if you want general discussions or #❓｜how-to-get-help for specific questions
trim cradle
How much of the bot's computation time is spent checking messages against the regular expressions?
tawdry vapor
It is undoubtedly you, if anyone, that would be the first to profile that.
rapid swallow
I've always wondered about that
cold moon
@cold island About snowflake command, do you mean creating custom converter? discord.py don't have in-built snowflake converter.
cold island
We have one in converters.py
cold moon
Oh, I didn't know that
rapid swallow
what's the snowflake command?
cold island
rapid swallow
How much of the bot's computation time is spent checking messages against the re...
does someone know the answer to this?
gritty wind
I doubt that’s tracked
rapid swallow
do you guys use some fancy tree thing when checking against the blacklisted words in the database
hmm, alright
surreal venture
Isn't the hastebin highlighting a little off? It happens when I copy paste and save the code very quickly: https://paste.pythondiscord.com/jevorukace.md
oh wait nevermind
hardy gorge
It's just really, really, really bad at guessing the right language
change the extension manually and it'll be better
trim cradle
do you guys use some fancy tree thing when checking against the blacklisted word...
we have a bunch of regular expressions for terrible things people might say and as far as I know most messages just get checked against every single one.
I believe messages from staff members aren't checked.
trim cradle
I will go check and report back
There are currently 60 patterns.
Some patterns might be intended to capture multiple words depending on which similarly-spelled things you think of as the same word
brazen charm
@patent pivot didn't the saving on the hastebin default to .py?
patent pivot
it did — though that disappeared with the migration to pydis infra since people complained about it
rapid swallow
Thank you Stelercus
obsidian patio
@patent pivot do you think adding an optional option to change language, defaulting to python, could be added?
patent pivot
well, we don't maintain hastebin, so we'd have to maintain a fork if we did that
brazen charm
Huh, wouldn't think that'd be a problem with how it's being used, saw more people confused why they were getting coffeescript
patent pivot
we'd either fork it or raise an issue upstream
trim cradle
we'd either fork it or raise an issue upstream
unlimited forks
patent pivot
lol
obsidian patio
Hmmm, yeah
patent pivot
the way we deploy it already is rather hacky
hardy gorge
did people have an issue with automatic py?
I'd say people now have to change it to py more often than people having to change it from py to something else
patent pivot
yeah, I'd agree
I mean we can add the defaulting back to py, but it's just something else to maintain
obsidian patio
It’s probably better to default to py and not maintain a fork for some features that most people probably don’t use. If you’d be interested in opening an issue in the hastebin repo about allowing users to pass the language in some sense, that would be cool though
patent pivot
hastebin upstream is kinda dead
I don't have the time to write up an issue, but if someone wants to open something brief on https://github.com/seejohnrun/haste-server then we'd definitely set that option if implemented
quasi crypt
Anyone know the best way to get snekbox running on GCP? I've tried...
- Cloud Run (fails because it can't find nsjail after booting up -- maybe because no privileged mode).
- Compute Engine (adds a ton of authentication overhead / non-severless setup seems expensive. Not sure if it works, deleted the container once I read their auth docs)
- App Engine (currently trying it now but I suspect it'll fail like Cloud Run since I don't see a privileged option here either)
tawdry vapor
Compute engine should be able to run privileged containers. I'm not sure about the others.
GKE also supports privileged containers
Maybe you can find a way to configure the k8s service account cloud run uses so you can enabled privileged containers
sharp timber
!warn 533618406767722499 please do not post memes randomly into the server
stable mountainBOT
:incoming_envelope: :ok_hand: applied warning to @ivory drift.
quasi crypt
I wonder if I'm overcomplicating things by trying to use snekbox for my needs. I need something that runs serverless in the cloud that can run arbitrary code... And I like the memory limits snekbox provides... But maybe I could just run exec in cloud functions with a 50mb memory limit per request and call it a day? I mean, if it's severless I'm not sure there's a ton of risk there. Tried to Google it and the jury seems to be out -- someone mentioned protecting my source code would be a factor, but my source code in that case would just be exec
tawdry vapor
Not sure. I thought that setting a limit on the container will cause it to kill the container.
Sorry, you said cloud functions. I don't know about those.
Keep in mind that NsJail is built on linux kernel features like cgroups and namespaces. You can think of it as a nice wrapper for those features. There may be a feasible way to use those features with a container directly.
After all, containers are implemented using those same features
ivory drift
!warn 533618406767722499 please do not post memes randomly into the server
wdym?
gritty wind
@patent pivot
It's a PR from before the migration that had the migration recently merged into it
Shows up fine on the site though
gritty wind
Take your time
patent pivot
looking into things
well
I see the log
"error":"failed to compute approval status: failed to list commits: 3 commits were not found while loading pushed dates
I forgot you were looking into this when I approved 👀
vocal wolf
uh oh, did something break?
green oriole
Yup
vocal wolf
D:
sullen phoenix
site ded
green oriole
Well, cloudflare is probably dead
patent pivot
oh
green oriole
Yup
vocal wolf
o noes site is ded
green oriole
Look at your pings haha
gritty wind
oh no
vocal wolf
lol confirmed from at least 3 coasts
patent pivot
yeah looks healthy here
gritty wind
Its not loading for me, possibly just my shitty internet
sullen phoenix
site's dead for me
vocal wolf
smae
green oriole
Probably cloudflare's shitty internet
gritty wind
lol
vocal wolf
heck cloudflare
sullen phoenix
weird, my sites haven't had any issues
vocal wolf
hmmm
sullen phoenix
it's not a cloudflare issue
gritty wind
Even if its a cloudflare issue, its unlikely all off cloudflare fails at once
gritty wind
Now we can't blame CF :P
cold island
Getting this when trying to start up @dusky shore. Anyone familiar?
gritty wind
Check your site config
Either that, or redis
Actually, from that stack trace, definitely redis
cold island
Where would I find it? I don't see anything about it in https://pythondiscord.com/pages/contributing/sir-lancebot/
gritty wind
It isn't on that page (yet)
but set the fake-redis env var
USE_FAKEREDIS=True
If you want to, you could setup redis, but it isn't easy on a windows machine
brazen charm
Fairly easy through docker
cold island
excellent, that did it. I'll update the page
gritty wind
Actually, if you hold off for a minute
I might be able to dig through the config for all the env variables
gritty wind
.
I can add them myself, but I first have to figure out which ones are worth adding
A bunch of these are not so useful and would pad out the page quite a bit like the AoC ones
I think I will add the fake_redis and Tokens list for now
cold island
https://pythondiscord.com/pages/contributing/configure-environment-variables/ It's a pretty short page at the moment. You can separate them by which ones are mandatory and which ones are optional
I think having a record of them would be useful
gritty wind
Alright, I'll see what I can do
cold island
Although if you're actually using redis then you don't need fake-redis right? so it's mandatory, but it's either or
gritty wind
I've added the tokens, look over them if you have a sec
I'd also appreciate someone looking into the docs for wolfram's api
cold island
Awesome, thank you
This might even work as its own subpage where we can list all variables
gritty wind
Let me finish this list first
gritty wind
Ok, I've gone ahead and done that
Feel free to copy it to another page if you want to
I might do the same for the bot and the site tomorrow
clever wraith
do we have a constant for the #❓｜how-to-get-help channel?
cold island
I don't think we're using it anywhere in the bot atm
clever wraith
I don't think we're using it anywhere in the bot atm
do you mean you have it, but it's not in use? will i have to create one?
cold island
If we're not using it then it's probably not in the config. I can take a look
brazen charm
No, the id is not stored anywhere atm as it's just a static message
patent pivot
I doubt we'd make that a constant
fetching the last message sent by the bot in the channel sounds wiser
clever wraith
making the channel a constant is what i mean
cold island
fetching the last message sent by the bot in the channel sounds wiser
wdym, you need to know which channel
clever wraith
fetching the last message sent by the bot in the channel sounds wiser
and i already am doing it that way
patent pivot
oh right, yeah
cold island
I assigned you to the issue@clever wraith
clever wraith
thank you
how should i get the latest message? i was thinking of pulling it from channel.history but i don't know if there's a better way you guys use
cold island
or is it "I assigned the issue to you"? 🤔
gritty wind
He’s handling the issue, so “he’s assigned to it”
clever wraith
how should i get the latest message? i was thinking of pulling it from `channel....
?
and should i just create the message if the length of #❓｜how-to-get-help is under 2?
cold island
looking at the docs, TextChannel seems to have a last_message_id property https://discordpy.readthedocs.io/en/v1.0.0/api.html#discord.TextChannel.last_message_id
clever wraith
hm