python-discussion

[spice hill]

well, it's a singular book, so it can't depend that much

[raven urchin]

I mean, it depends on the book and author.

[charred tusk]

I think you need to upgrade your understanding to the 4.5e edition of books

Someone please know where that number comes from

[cerulean ravine]

"2nd edition" always means, "use this one instead of the 1st edition". If it was a sequel, it would be called something else.

[spice hill]

alright, so my understanding of the world is still somewhat intact

it looks like Kleppmann revised the structure of the book quite a bit

[raven urchin]

Writing Data Intensive Apps 2: The Phantom Menace

Oh wait 2 was the attack of the clones lol

[inland karma]

im actually gooint to the capital tomorrow, ill stop by and see if they indeed have 1 copy left of the second edition

[kind monolith]

Hey Guys, is there any Website where you can complete some small python tasks/scripts/fix small bugs and get a small Payback?

[solid estuary]

hi

whats up

[pallid garden]

you do realize people do it for free here?

[charred tusk]

They want to earn money methinks

[pallid garden]

why would people use that platform instead of just asking for help online?

with the consideration that it's just a small task

[kind monolith]

I mean i want to help with small tasks and get paid a small amount

[charred tusk]

You must be new to the internet

[wise imp]

here's an idea, you make an "AI" service that fixes bugs... in reality it's just this

[charred tusk]

Too late
That idea has already been stolen
https://ia.acs.org.au/article/2025/the-company-whose--ai--was-actually-700-humans-in-india.html

[pallid garden]

Actually Internet chatters

[wise imp]

yk, ideas can be stolen more than once, they don't have a borrow checker unlike rust

[spice hill]

unsafe {
    copyright.violate();
}

[pallid garden]

you can borrow something multiple times...

...as immutable references

[wise imp]

listen, idek what rust is

some metal oxidation or something

[pallid garden]

"reading does not conflict"

[bright mauve]

immutable references can't stop me because i can't read

[wise imp]

hmm, ideas are immutable, aren't they?

[bright mauve]

you'd better hope not, or there is little hope for humanity

[knotty ice]

[pallid garden]

the real question is: is the Ship of Theseus mutable

[craggy willow]

humanity is immutable, aren't they?

[proud escarp]

would there happen to be a cleaner way to generate index-ish IDs for entries in a sqlmodel.SQLModel table?

def create_item(session: Session, item_body: ItemBody):
    last_id = len(session.exec(select(Item)).all())
    item = Item(
        id=last_id,
        name=item_body.name,
        price=item_body.price,
        expiry_date=item_body.expiry_date,
    )
    session.add(item)
    session.commit()
    session.refresh(item)
    return item

I was using itertools.count but this fails some tests of mine because the counter state persists across different tests

# counter = count()

class Item(SQLModel, table=True):
    # id: int = Field(primary_key=True, default_factory=lambda: next(counter))
    id: int = Field(primary_key=True)
    name: str = Field()
    price: int = Field()
    expiry_date: date | None = Field(default=None)

[wise imp]

no no, every time you come up with like an improvement or whatever, it's a new idea, it's like versioning but for ideas

damn, gotta write an RFC for versioning of ideas

[raven urchin]

No advertisement is allowed here

At most, you can advertise your open source projects at #1468524576479641744

[wise imp]

what?

[nova breach]

how can I install Python 3.13.13 in Debian?

[steel whale]

you use uv

[forest elm]

anyone studying in LUND university sweden?

[proud escarp]

you can use uv.. but if you want a global python i recommend pyenv

[raven urchin]

The bestbian lesbian 🙏

[steel whale]

you still use uv

you can override the PATH

and make uv's pip install user-wide instead of system-wide

[proud escarp]

it supports setting a global python (well, it doesnt "support" it but you can override the PATH yeah), but not global dependencies very well

[nova breach]

can these override python installed by apt?

[kind monolith]

Thats not an advertisement

[steel whale]

or just use uv pip, it launches a user-wide venv

[raven urchin]

It is, you're advertising yourself to be hired.

[steady rain]

uv won't mess up the system python

[raven urchin]

These things are virtual environments, it's the recommended way to dev in Linux overall.

[proud escarp]

no

[kind monolith]

Im not i asked if people know a website to find small tasks to do for money

[nova breach]

and should I uninstall apt python after installed uv?

[steady rain]

Did you see my answer in career discussion?

[raven urchin]

Oh, then sorry, I thought you asked if we had those tasks for you.

[steady rain]

Not necessary

[proud escarp]

#editors-ides message
i was facing this issue in the past

[kind monolith]

Yes

[spice hill]

no, your system likely depends on python for a lot of things

[proud escarp]

i use uv for my projects, pyenv for managing my pythons

[nova breach]

i see

[raven urchin]

Then you'd break anything that uses python in your system.

[wise imp]

I was using itertools.count but this fails some tests of mine because the counter state persists across different tests
can't you just patch counter to be a new instance for each test or set it in setup?
though a counter like that would also be a danger for multithreaded or multiprocessed environments

[steady rain]

uv can do both

[raven urchin]

If you want something more up-to-date you should use Debian Sid or another rolling-release distro like Open SUSE tumbleweed.

[proud escarp]

i was facing issues in the past with uv and a global python
#python-discussion message

[spice hill]

you know about autoincrement?

[nova breach]

so...... there are linux apps run on python?

[proud escarp]

no

[raven urchin]

Of course lol

[nova breach]

trixie is not the newest now?

[proud escarp]

though a counter like that would also be a danger for multithreaded or multiprocessed environments
could make it a contextvar

[raven urchin]

That's the unstable branch of Debian, it's more up-to-date.

It's like Arch, rolling release.

[proud escarp]

index-ish IDs are weird anyways

[raven urchin]

The adduser script runs on Python, for example.

[bright mauve]

arch does some testing before releasing software though. debian sid is where you test software before releasing, so it breaks. sid is more comparable to the arch upstream, if anything

[craggy willow]

can you use uuid

[raven urchin]

Well, yeah

[nova breach]

trixie is just the newest stable?

[raven urchin]

Yes

[bright mauve]

yes. just keep in mind that debian keeps older versions of software in order to be stable

[wise imp]

ehhh, maybe, but I'd follow fix error's suggestion of using an auto-increment (pretty sure it's just an SQL-ish feature (presumably implemented by the DBMS))

[spice hill]

See this: https://sqlite.org/autoinc.html#summary (or the equivalent documentation if you're using a different database engine)
If this is your table: sql create table fruits ( uid integer primary key autoincrement, name text not null, count integer not null ); then uid becomes an optional field that increments on each insertion. So you can do insert into fruits (name, count) values ('banana', 42); and it will pick the next available uid automatically.
I would recommend interacting with an SQL database directly before jumping to ORMs and other tools that wrap them. But it looks like SQLModel does support automatic ID generation: https://sqlmodel.tiangolo.com/tutorial/automatic-id-none-refresh/

[raven urchin]

Debian releases a new version every x amount of time, I don't know the exact time, but it makes it pretty outdated compared to other distros

But also makes it insanely stable.

[proud escarp]

mm thank you

[wise imp]

can you imagine, a language that doesn't support trailing commas (also Java, hate that about it, too)

[proud escarp]

yeah, but again that would be persist across the tests.. i could patch them with a fixture i guess, not that it would matter if it were persistent in the case of uuids
but anyways, auto-increment is the way

[cerulean ravine]

what are you checking in your tests? Maybe they need to change

[proud escarp]

haha the colorscheme is different for the database tutorials

[nova breach]

seems apt don't allow me for installing uv......

[proud escarp]

well, i was testing if the index-ish IDs were being generated properly. of course, i wouldnt be doing that with uuids
why was i trying to make the IDs index-ish? i dont know, im just playing around

def test_item_id_generation(client: TestClient):
    for i in range(10):
        response = client.post("/items/", json=test_item_body)
        assert response.status_code == 200
        assert response.json()["id"] == i

    response = client.get("/items/")
    assert len(response.json()) == 10

[spice hill]

Is this an actual requirement for the project? Or an implementation detail?

[bright mauve]

you can probably do the "standalone install" by downloading the installer from the uv website and executing it in the terminal

you can give this a shot https://docs.astral.sh/uv/getting-started/installation/

[cerulean ravine]

instead you might want to test that the 10 responses have 10 different ids, and not care what they are

[proud escarp]

im working on a hypothetical inventory api for the means of playing around with fastapi as i was curious to try it out, the requirements are vague, apart from "by trying to add this i should learn something"

anyways, the automatic id generation works

[cerulean ravine]

also, when testing these kinds of things, it can be very helpful to reset the state of the world between tests

[proud escarp]

yes, i do that with fixtures

@pytest.fixture
def client():
    engine = create_engine(
        "sqlite://",
        connect_args={"check_same_thread": False},
        poolclass=StaticPool,
    )
    SQLModel.metadata.create_all(engine)

    def get_test_session():
        with Session(engine) as s:
            yield s

    router.dependency_overrides[get_session] = get_test_session
    yield TestClient(router)
    router.dependency_overrides.clear()

[cerulean ravine]

where is the data stored? You said the tests were affecting each other.

[ashen cipher]

hello peoples

[proud escarp]

yes well, if i were using uuids. ill probably switch to that now for a different table, to try those out as well

[raven urchin]

You have to run their installer on Debian

curl -LsSf https://astral.sh/uv/install.sh | sh

[cerulean ravine]

even if you are using numbers: you don't care what the numbers are. The test is too specific.

[raven urchin]

Or add the community repo

[proud escarp]

no, that was because i was using a global counter before for the ID generation which the above fixture doesnt deal with:

counter = count()


class Item(SQLModel, table=True):
    id: int = Field(primary_key=True, default_factory=lambda: next(counter))

i could patch this too with a fixture but.. yeah, anyways. auto-increment is the way to go

[sudden basin]

https://discord.com/channels/267624335836053506/1498296381918679110

[proud escarp]

well you would if you were making index-ish IDs, i suppose
but i wonder if there's any actual use case in using index-ish IDs
i just did it because... well, that's where my mind went. id probably use uuids for all practical purposes

even if there were a use case of index-ish IDs, it should probably be a different field, and the actual id be a uuid

[floral terrace]

uuids for most apps is kinda overkilled

[proud escarp]

is it? seems kinda basic to me

[spice hill]

It's a good idea to avoid relying on implementation details when testing. That way, you won't have to rewrite the tests when refactoring the system.
Identify the actual requirements for the feature and test them. For example, here's what I'd expect from a basic system that supports adding and listing some generic items:

1. the generated items have unique identifiers (the client of the API doesn't need to know how they are generated)
2. after receiving the ID for the item, you can then grab the item by that ID
3. after adding an item, it will be present when asking for a list of all the items
4. IDs are unique, meaning they are not repeated ever again, even if other actions take place in your system.
   So you might create a test for each of those. Testing number 4 will be difficult and/or impossible

[robust ledge]

You could force number 4 there but there's little value in it from what I can see. You'd be testing the failure route of an insert which should be covered in a different test.

[ashen cipher]

c developers /j

[proud escarp]

yes, these are better

testing number 4 well... ill leave it up to the auto-increment technique author

[floral terrace]

wdym basic? if you are using any kinda orm that works with any database, the uuids are automatically generated by the orm so I dont see the benefit of uuids over int, except more taking up more memories

[spice hill]

non-incrementing IDs can work as a defense in depth to protect against enumeration attacks

[raven urchin]

So, guys, I can make pretty much the same stuff in C and in Rust but Rust is easier? I think that I still don't understand the relation properly.

[floral terrace]

the maximum value of bigintfield in Django is around 9 billions iirc

[spice hill]

I suppose you can generate random 64-bit numbers as opposed to 128-bit UUIDs

[proud escarp]

that is true

[steady rain]

C requires you to uphold a lot of invariants yourself, whereas Rust makes them part of the language.

[floral terrace]

do you think your app need to generate this many instances ?

[raven urchin]

In dumb friendly terms?

Or an example?

[jade robin]

C requires you to think more and lets your code fuck up, rust does the thinking for you and forces you to address issues

[azure pine]

Made something practical today , Image Resizer hehe .. using Pillow.

[proud escarp]

but it's not about number of cases, if you want a unique identifier a uuid is simply better than using integers
if you want to identify rows by some integer value, like rank, that should not be the id it should be some other field like "rank"

unless im not getting what you mean

[spice hill]

You could repeatedly add and remove items and see if the IDs repeat. Here's an example footgun I'm thinking of:
https://sqlite.org/autoinc.html#summary

If the AUTOINCREMENT keyword appears after INTEGER PRIMARY KEY, that changes the automatic ROWID assignment algorithm to prevent the reuse of ROWIDs over the lifetime of the database. In other words, the purpose of AUTOINCREMENT is to prevent the reuse of ROWIDs from previously deleted rows.
so if you don't specify AUTOINCREMENT, the IDs will look like they're unique since they're just incrementing forever, but they're not

[raven urchin]

Rust sounds better for dumb people like me then

What does C do better than Rust?

[steady rain]

Don't tell yourself that you're dumb.

[steady rain]

If you need to work at a low level, like implementing an operating system.

[jade robin]

idk, other stuff is fairly compatible with C

[gleaming knoll]

it compiles faster (because the compiler doesnt do as much) [different from compiling to faster code]

[raven urchin]

And even like that it's used in the Linux Kernel, interesting. I wonder what they use it for

[ashen cipher]

rust has C FFI

[wise imp]

mmm, isn't that exactly what Rust also supports

[robust ledge]

Feels like an overly aggressive test that is reaching into a space outside the scope of the implementation. Maybe I'm thinking of it from the wrong angle. I suppose it could cover the CREATE queries.

[jade robin]

that sounds like a win for C then

[floral terrace]

I don't think I understand you, what do you need the id for, how is uuid beter than int in terms of being unique?

[ashen cipher]

C FFI means you can also go Rust -> C FFI -> language of your choice which supports C FFI (basically all of them)

[jade robin]

yeah and you're dealing with C anyways

[raven urchin]

Well, if I believe I'm dumb, I don't get mad at myself when I make dumb mistakes, but also force myself to do more because if I'm not talented, at least I'll be hardworking!

[steady rain]

Afaik, rust is "low level for the sake of high performance", whereas C is "low level for the sake of operating at low levels of abstraction".

[ashen cipher]

not directly

[gleaming knoll]

rust is whatever "level" you want. the source code of most userspace rust apps is pretty high-level,
but you can drop down in case you need to

[raven urchin]

Oh, this sounds pretty cool ;0

[steady rain]

In your usage, is talent something that can be cultivated, or only something that's innate?

[frigid trout]

I don't know whether I should make a portfolio website or learn how to make games, which one should I do?

[steady rain]

You can put games on your portfolio website, no?

[frigid trout]

Yes, but I haven't made the portfolio website...

So should I make things that can go on the portfolio or should I make the portfolio itself?

[jade robin]

sure i mean, it still counts as a point for "why you should implement some things in C" because of "not directly"

[raven urchin]

Well, talent usually means innate. For me, it's the skill of being able to learn both fast and properly at the same time. I just focus on learning properly, try to not rush it.

[gleaming knoll]

i would choose rust for a lot of projects because it simply has better tools for abstraction and good compiler errors
the ecosystem is also pretty good at this point, and there is a sane package manager that everyone uses

[ashen cipher]

there are libraries that abstract all that away for you (see: pyo3)

[steady rain]

Make whatever you want and put it on GitHub.
Static websites are trivial to make, so I wouldn't put too much stock into it.

[frigid trout]

okay.

[proud escarp]

what do you need the id for
if you have the id of an item, and you want all the other details so you use the id to get it
non-incremental random integer IDs would do just as fine as UUIDs, i was talking about index-ish integer IDs. i suppose it works out as an ID but id rather the ID be something else and that be "rank" or such

[wise imp]

pro tip: put the portfolio website on the portfolio website as one of your projects, thus making the portfolio website automatically populate itself

[frigid trout]

I was thinking about that.

[spice hill]

You can also blog about creating the portfolio website, and list the blog as one of the projects

[floral terrace]

no, if you don't treat the id like a rank : ). You can do query using int just fine

[ashen cipher]

recursive portfolio

[frigid trout]

What is a blog?

[spice hill]

[raven urchin]

Nom, nom, I'm excited about web dev, I want to start soooonnnn

[frigid trout]

It's just another website, right a blog.

[wise imp]

binary log? what does the b actually stand for

[gleaming knoll]

it was weblog but got shortened to blog
so the "b" is "web"

[raven urchin]

Guys, what do you think it's a good project to "graduate" from a gui framework (Pyside) and start learning something else?

[ashen cipher]

we blog

still doesnt explain it /j

[swift sparrow]

already done with pyside?

[floral terrace]

I can't see why you need to treat int id differently than uuid in your case. See it as a name of instance.

[raven urchin]

No, I want to make something big and more complex before jumping to my next topic

[wise imp]

a todo app

[swift sparrow]

yeah, I think a todo is a nice jump from what you just did

[nova breach]

can forky run python newer than trixie without uv?

[raven urchin]

What's a todo?

[swift sparrow]

like a task list

[raven urchin]

I can only think of "all" in Spanish lol

[raven urchin]

Ohhh, a to-do list app?

[swift sparrow]

yeah

QListWidget will be your friend

or maybe even QTreeWidget if you want to make some categories

[raven urchin]

Mhmmm, that sounds good

I could even make kind of like a project roadmap-thingy

Instead of just a to-do list, make it related to what I'm making

[ashen cipher]

kanban board

[raven urchin]

Oohhh, this sounds great

[wise imp]

but you are making a to-do list app
are you gonna make a to-do-list-app-project app? 😄

[raven urchin]

Yes :D

[swift sparrow]

todo to-do

[frigid trout]

Make a game.

[sturdy cosmos]

Hi chat

[raven urchin]

I'm thinking of a lot of functionality to add to it, this sound incredibly could in my mind

[sturdy cosmos]

I’m new here

[wise imp]

not in pyside

[frigid trout]

!slorb

[edgy kraken]

Go-to beginner resources

Here are the top free resources we recommend for people who are new to programming:

• Automate the Boring Stuff — an online book (also available to purchase as a physical book)
• Harvard’s CS50P course — video lectures (slides and notes provided) with exercises
• Python Programming MOOC 2026 course — text-based lessons with exercises
• Corey Schafer's YouTube playlist

For a full, curated list of educational resources we recommend, please see our resources page!

[raven urchin]

I already made a simple game, but Pyside isn't really to make games, it's to make apps.

[swift sparrow]

there's lots of simple GUI games you could make in pyside

[frigid trout]

I never heard of that one cool.

[swift sparrow]

battleship, hangman, blackjack

[wise imp]

ah, fair, minesweeper comes to mind

[raven urchin]

Oohhh, making those would also be cool

I still have a ton of projects that I can make in Pyside hehe

[swift sparrow]

I've been using pyside for 6+ years and still learn new things all the time

[raven urchin]

It's a great framework

Mhmm, I think I will make my project-todo-manager app

[knotty ice]

[raven urchin]

It can be either pretty simple or insanely big

[ashen cipher]

minesweeper players terrify me

[raven urchin]

I learnt how to play and it was like...

[frigid trout]

Hello.

[raven urchin]

Now I forgor

[frigid trout]

I still don't understand how to play Minesweeper.

[gleaming knoll]

minesweeper players are neural networks larping as humans

[ashen cipher]

yea basicaly

[frigid trout]

Hi, I am new.

[gleaming knoll]

hi new, i am lambda

[ashen cipher]

dammit

[raven urchin]

Well, the numbers mean how many bombs are adjacent to the button

[ashen cipher]

hi lambda, im @ashen cipher

[frigid trout]

Let's take this discussion into the ot0?

[ashen cipher]

i need to change my server display name one of these days

[raven urchin]

Hii

[gleaming knoll]

bro said that like it takes more than 5 seconds
/nick x

[ashen cipher]

issue is i need to think of one

[raven urchin]

The best thing about having an android phone is being able to install Termux and run pretty much anything I make in my PC

[gleaming knoll]

android wins by not locking its users :::

the bar is that low

[ashen cipher]

not locking its users too much*

hm

yeah i dont like this one

[ocean ridge]

suggest a name for my terminal coloring python library

[gleaming knoll]

vec null is peano naturals

[ocean ridge]

its a library like colorama

[solar ravine]

Hello, I'm new too

[ocean ridge]

not app, python library

[spark obsidian]

yatcpl :^)

[ashen cipher]

not really

[gleaming knoll]

yet another terminal coloring python library

[ocean ridge]

i was thinking like pastelcolors or spectrum or vibgyor or smth like that

[raven urchin]

Oh, I need to make a TUI app some day

[ocean ridge]

its my first python lib 🥀

[raven urchin]

I really love TUI

[ashen cipher]

poxel

[gleaming knoll]

zero = []
succ(n) = [*n, null]
all vec<null>s represent a given n, equal to the vec's size

[ocean ridge]

i was thinking maybe even include my name 🙃

[autumn forge]

variagattyd

[ocean ridge]

i need an easy t oremember name 🙃

[ashen cipher]

well yeah, that would be more like ```rs
enum Number {
Nonzero(Box<Number>),
Zero(Vec<>)
}

Vec<Number>```

[ocean ridge]

i found a really good name!!!
VIBRANT

[mellow hill]

what is the use of indexing in python

[ashen cipher]

vexel

[autumn forge]

vibe rant

[ashen cipher]

how has that not bene taken

[ocean ridge]

letsgo :D

[ocean ridge]

IKR SUCH A STEAL

[raven urchin]

That already exists actually

[ashen cipher]

i still like vexel more

[ashen cipher]

!pypi vexel

[edgy kraken]

Not in my house!

Package could not be found.

[ocean ridge]

vibrant is more related to colors

[gleaming knoll]

its when you have an ordered bunch of items and want to access a specific one by its numeric order

[ashen cipher]

arguable

[raven urchin]

Not a python pkg tho, it's a website

[ashen cipher]

https://en.wikipedia.org/wiki/Vexel taken by a wikipedia page 💀

[mellow hill]

also what is the use for using code line in modules like variable= x

[gleaming knoll]

you mean "what are the double underscores on both sides for in stuff like __name__?"

[raven urchin]

___variable___= __x__

[autumn forge]

vercel?

[crisp jay]

Secret variables

[ashen cipher]

that's __variable

[crisp jay]

Thats private

[gleaming knoll]

thats mangled
private is single _x

[mellow hill]

exactly like i came upon a topic where im learning python it used the underscored variable explaing what it shows something in module in another python code whereas some other in main code

[ashen cipher]

🥴

[gleaming knoll]

thats specifically about if __name__ == "__main__":

[mellow hill]

yes

yeah thets the exact code

[gleaming knoll]

in general the double underscores on both sides of a name are for names "reserved" by python
the __name__ global variable is one of those, python sets it to the name of the module. for the entry point module, which the program started running from, its the string "__main__"

[mellow hill]

so how does it help in anything in general as a function

[ashen cipher]

!name

[edgy kraken]

Did you mean...

» if-name-main
» names
» off-topic-names

[gleaming knoll]

it makes sure it wont collide with names people would use for their actual variables
the assumption is that people wont normally name their things __x__, so its only used for special things

[ashen cipher]

!if-name-main

[edgy kraken]

`if __name__ == "__main__"`

This is a convention for code that should run if the file is the main file of your program:

def main():
    ...

if __name__ == "__main__":
    main()

If the file is run directly, then the main() function will be run. If the file is imported, it will not run.

For more about why you would do this and how it works, see if __name__ == "__main__".

[ashen cipher]

well that isnt very helpful

[mellow hill]

so secret variable sort of

[ashen cipher]

is it just me or is #dev-contrib locked?

[gleaming knoll]

yeah but you just need to grab the contributors role in #roles
its spam prevention i think

[ashen cipher]

o

[pallid frigate]

hey

[wise imp]

it's not really meant to be secret at all, it's just formed in a way to avoid name collisions

[gleaming knoll]

"theenglishcoder2026" with a russian display name 😭

[ashen cipher]

yea bc he codes in english

[proud escarp]

exactly

[ashen cipher]

yk i like this more

[crisp jay]

Arent they called dunder methods? (Or like magic)

[gleaming knoll]

__name__ is not a method. yes its a dunder, double underscore (on both sides, idk what __x would be called though)

[crisp jay]

Private

[gleaming knoll]

its not just private though, its mangled. and this doesnt represent the fact that its, well, a double underscore

[robust ledge]

A name mangled attribute.

[pallid frigate]

plus i am english

[crisp jay]

Its like,protected with the name of the class (idk about it outside tho)

[tame vapor]

mono dunder

[ashen cipher]

question - is english an ethnicity, nation, country, or state

[crisp jay]

Yes

[tame vapor]

its a misunderstanding

[robust ledge]

It's not protected at all though. I use the term "internal" to indicate it's not meant for outside access. But I also use that for single leading underscores too.

[ashen cipher]

thank

[pallid frigate]

what to code..?

[crisp jay]

Try to solve problems from projecteuler.net

[tame vapor]

boring ngl. But again , i dont have alternative other than "mkae projects that u find interesting" anyway

[crisp jay]

"Boring",bro it took me 10 hours to do a percentage of problem 20 in C

[tame vapor]

percentage of problem 20 ? whats that

[crisp jay]

I need to find the sum of the digits from the number 100!

[peak relic]

did you do this yet? #python-discussion message

[crisp jay]

So im working with strings and took me hours to make a simple addition function

[gleaming knoll]

import sys as _система
from typing import TextIO as ТекстовыйВводВывод

объект = object
строка = str
булев = bool
Ничего = None
Нет = False

получить_аттрибут = getattr

@lambda класс: класс()
class система:
  псевдонимы = {"стандартный_вывод": "stdout"}
  стандартный_вывод: ТекстовыйВводВывод
  def __getattr__(я, имя):
    return получить_аттрибут(_система, я.псевдонимы.get(имя, имя))

def вывести(*объекты: объект, разделитель: строка = " ", конец: строка = "\n", файл: ТекстовыйВводВывод | Ничего = Ничего, смыть: булев = Нет) -> Ничего:
  if файл is Ничего: # :( cant translate keywords
    файл = система.стандартный_вывод
  assert файл is not None # stupidass typecheckers
  форматированная_строка = разделитель.join(строка(объект) for объект in объекты) + конец
  # translating methods would suck too, would need fishhook or something
  файл.write(форматированная_строка)
  if смыть:
    файл.flush()

вывести("Привет, Мир!")

😭 😭

[tame vapor]

🤷

[tame vapor]

privet

[crisp jay]

Is this GTA VI?

[gleaming knoll]

actual programming crime

[raven urchin]

The first thing you have to learn to program is English 🥀

[runic flower]

there's a shortcut for this isn't there? you don't actually want to count them all.

[raven urchin]

That's Half Life 3

[ashen cipher]

cs3

[floral terrace]

programming war crimes

[gleaming knoll]

everything except the empty set

[raven urchin]

CS2 is actually turning good recently.

[ashen cipher]

this implies a programming war

[steady rain]

@ashen cipher your nickname is the universal set?

[spice hill]

import typing as печатает

[ashen cipher]

-1

[crisp jay]

Is there a way to calculate only the mantissa of a big number? Because I know that 100! will have alot of useless zeros

[gleaming knoll]

except the empty set

[ashen cipher]

ah

yk what sure

[steady rain]

Fun

[gleaming knoll]

ction
al
relationship
between sets

[floral terrace]

isn't that an actual syntax proposal for the initialization of python empty set

[raven urchin]

[gleaming knoll]

!pep 802

[edgy kraken]

**PEP 802 - Display Syntax for the Empty Set**

Status

Draft

Python Version

3.15

Created

08-Aug-2025

Type

Standards Track

[gleaming knoll]

{/}

[ashen cipher]

should've made it o/ instead so it would've looked like someone waving

[gleaming knoll]

should've made some pancakes

[runic flower]

quite a few, but not too many.

93,326,215,443,944,152,681,699,238,856,266,700,490,715,968,264,381,621,468,592,963,895,217,599,993,229,915,608,941,463,976,156,518,286,253,697,920,827,223,758,251,185,210,916,864,000,000,000,000,000,000,000,000

[floral terrace]

I just hate you

[ashen cipher]

thanks

[gleaming knoll]

dw the set of everything except the empty set i love you so it cancels out

[ashen cipher]

skull

[velvet trout]

Moscow in my python code?

[spice hill]

5. {*()}, perhaps optimising to compile this to the BUILD_SET opcode
   prediction: this is going to win

(in a few years)

[mellow hill]

checked the python docs now clarified thanks

[ashen cipher]

its more likely than you think /j

[velvet trout]

[runic flower]

lets just break everything, {} = empty set {:} = empty dict

[bronze dragon]

Standards Track means accepted, right?

[crisp jay]

(,)

[ashen cipher]

(/)

[gleaming knoll]

this is how it shouyld've ben

[velvet trout]

Isn't there status?

[bronze dragon]

ahh right

still Draft

[runic flower]

probably would have been, if sets weren't added years after dicts.

[velvet trout]

Standards Track is one of those which also includes Informative and other types

[bronze dragon]

right, right

[velvet trout]

left, left

[crisp jay]

Sets has their own hash function iirc

[spice hill]

!pep 832

[edgy kraken]

**PEP 832 - Virtual environment discovery**

Status

Draft

Python Version

3.15

Created

19-Jan-2026

Type

Standards Track

[pallid frigate]

hey what should i code? im at the level of file manage ment

[gleaming knoll]

manage ment

[gleaming knoll]

set.__hash__ is None

[ashen cipher]

frozenset

[gleaming knoll]

who froze my sets

[ashen cipher]

me, mario

[craggy willow]

never seen this before

[gleaming knoll]

i wanted to eat some warm sets

[craggy willow]

kinda cool

[velvet trout]

Disney

[gleaming knoll]

dis.ney? whats that

[velvet trout]

Disney froze your sets

Disney's 2013 frozen i mean

not disassembly module lmao

[raven urchin]

Wdym "level of file management". I didn't know we had levels. 😭

[pallid frigate]

idrk like learning

[ashen cipher]

yea im level 7

[gleaming knoll]

we had levels for a bit after april 1

[velvet trout]

do you know with open(): ...? Since you said file management

[ashen cipher]

6.5 +- 0.5

[spice hill]

https://github.com/rust-lang/hashbrown

[raven urchin]

This is the first time I've ever heard this phrase. What a horrible thing to say.

[pallid frigate]

yep

[pallid frigate]

mb

[wheat mica]

Yall quickly, is this valid?

type Noneable[T] = T | None

[velvet trout]

Explain to me what it does

[raven urchin]

Learning is amazing, don't say you don't like it 😔

[spice hill]

Yes, but not very useful

[gleaming knoll]

!pep 695 in versions after and including 3.12 yes

[edgy kraken]

**PEP 695 - Type Parameter Syntax**

Status

Final

Python Version

3.12

Created

15-Jun-2022

Type

Standards Track

[velvet trout]

Isn't this tryna typing.Optional[T] ?

[wheat mica]

I rly just meant like "type using generic type" but easier for me to give code ngl

[velvet trout]

Isn't it the same thing

[wheat mica]

It was jst an example DW

[velvet trout]

ooh

[wise imp]

yes

[spice hill]

Yes, that's the right syntax for defining a generic type alias. See the PEP above and the page in the typing docs (https://typing.python.org/en/latest/spec/aliases.html)

[velvet trout]

Noice

[wheat mica]

Thx

[pallid frigate]

It lets you to write directly in the file instead of print(variable, file=filename), you can do f.write("Hello World!") if you add as f at the end

[velvet trout]

[pallid frigate]

and other stuff but thats what i mainly do with it

[velvet trout]

I never used it like that. Good point

[wheat mica]

So basically fprintf vs fwrite is what I’m getting from this

[velvet trout]

Holy C

[gleaming knoll]

the file argument to print has to be an actual file object with a write method, not a file name

[velvet trout]

I hate C, i need namespace not prefixes

[pallid frigate]

• you can do variables in f.write(), but it only takes ONE argument so you can't go:
  f.write("Hello! ", "World!")

[wheat mica]

My brain is used to C yk

[velvet trout]

Is file= has to be full file protocol or is a typing.Protocol with write() method?

[wheat mica]

You could just do "Hello! " "World!"

[velvet trout]

Not like i know anything about file protocol if it exists.

[gleaming knoll]

just .write(str)
and if you do flush=True then also .flush()

[velvet trout]

So it's structural typing ?

[gleaming knoll]

yes, or really its duck typing. the stdlib itself doesnt even define it

[velvet trout]

Ooh

[pallid frigate]

like

print("Hello World!", file=filename)```
That would be using it in print

[raven urchin]

with open("test.txt", "w") as f:
    f.write("Sana sana colita de rana")```

?

[pallid frigate]

Yeah

[velvet trout]

Aren't they the same in python or i missed something

[crisp jay]

I hate that C doesnt have a 1024 bit integer by default

[velvet trout]

Structual typing vs ducking typing

[gleaming knoll]

its weird to call this a filename. its not a filename, its a file.

[wheat mica]

Thats not rly a filename

[autumn forge]

"filename" isn't a file name

[wheat mica]

Thats an open file

[pallid frigate]

ik but im using it as an example

[velvet trout]

Y'all count your int bits? Hold my int()

[wheat mica]

Ngl IVe never had a use for that big an int but understandable

[oblique spindle]

I hate your nickname

[velvet trout]

Is it true you people gotta pass your container's length count everywhere if one wants to iterate over it? Idk C

[wheat mica]

Hatersll hate yk

[wheat mica]

Yes

[velvet trout]

[wheat mica]

Or you end it with a special value

[velvet trout]

I love OOP.

[wheat mica]

Cuz the arrays decay to pointers

[velvet trout]

Ooh like a sentinel?

[floral terrace]

haskell guy gonna have a stroke bruh

[velvet trout]

Holy physics

[pallid frigate]

with open(userinput, "w") as f:
  f.write("Hello World!")

[dry yacht]

Well. Sorta. Not always

[pallid frigate]

or you can do "a" for it not to overwrite it

[gleaming knoll]

use C23, then you can use _BitInt upto BITINT_MAXWIDTH
on x86-64 gcc its 65535

[wheat mica]

Well, when you pass an array to a function it decays

[velvet trout]

Why not filename as varname this time?

[velvet trout]

XD

[pallid frigate]

cuz its input from the user, also last time it was just to declare filename

[dry yacht]

Yeah, but only the "outermost layer" decays. If you have a 2d array, the inner arrays don't decay

[wheat mica]

Filename and extention implies smth like how fat stores them, basically this: NAME EXT

[velvet trout]

Isn't it more appropriate to name it filename this time? Unless you really want to differentiate or validate user_input vs final filename.

[pallid frigate]

extension is (e.g) .py or .cpp

[pallid frigate]

differiantiate plus its input from the user

[velvet trout]

pathlib.Path can help you in partification of given file string i guess

[wheat mica]

Like FILE PY

[velvet trout]

Then what you did was right

[pallid frigate]

🙂

[wheat mica]

Thats how it’s implemented in fat

Filename is 11bytes padded with spaces

Then 3 for extention

Also padded with spaces

[pallid frigate]

im talking abt python

[velvet trout]

What if user inputed more than that?

[gleaming knoll]

this is real

[velvet trout]

No, that's irrational!

[wheat mica]

Im saying thats what your prompt implies

[gleaming knoll]

ints cant be irational, they are (x:int)/1

[pallid frigate]

oh

[velvet trout]

Damn

[gleaming knoll]

and it supports arithmetic operations
no printf specifier for them though so need to mod10 loop yourself

[wheat mica]

so it’s x/1/1/1/1/1/1…

[leaden plinth]

guys im heading into the algorithm part of python, wish me luck 🥲

[wheat mica]

cuz if x is of type int then x is x/1

And so on

Forever and ever

[soft coral]

I am using ruff, what setting should I change for ruff to auto-explode when there's a trailing comma?

def func(x, y,):
  ...

-->

def func(
  x,
  y,
):
  pass

Rn it removes the comma

[crisp jay]

Ints are basically a number from the integer set,because they are a subset of the rational,it can be written as n / m

[wheat mica]

Im jst playing with the syntax fam

Technically any int x can be written as n/10 as long as n = 10x btw

[floral terrace]

line-length = 1

[dry pike]

it should do that already

https://docs.astral.sh/ruff/settings/#format_skip-magic-trailing-comma this is false by default, so not skipped

[soft coral]

Yeah It seems like it's not ruff, maybe some other formatter running as well

[pallid frigate]

with open(user_input, "w") as file:
    file.write("Hello World\n"[0])
    file.write("Hello World!\n"[1])
    file.write("Hello World\n"[2])
    file.write("Hello World\n"[3])
    file.write("Hello World\n"[4])
    file.write("Hello World\n"[5])
    file.write("Hello World\n"[6])
    file.write("Hello World\n"[7])
    file.write("Hello World\n"[8])
    file.write("Hello World\n"[9])
    file.write("Hello World\n"[10])

I wouldn't do this but you know

[edgy kraken]

Hey @pallid frigate!

Please edit your message to use a code block

Add a py after the three backticks.

```py
print('Hello, world!')
```

This will result in the following:

print('Hello, world!')```

[pallid frigate]

user_input = input("Enter File.EXTENSION: ")
with open(user_input, "w") as file:
    file.write("Hello World\n"[0])
    file.write("Hello World!\n"[1])
    file.write("Hello World\n"[2])
    file.write("Hello World\n"[3])
    file.write("Hello World\n"[4])
    file.write("Hello World\n"[5])
    file.write("Hello World\n"[6])
    file.write("Hello World\n"[7])
    file.write("Hello World\n"[8])
    file.write("Hello World\n"[9])
    file.write("Hello World\n"[10])

I wouldn't do this but you know

[soft coral]

It seems that it first auto-fixes it, as a trailing comma is not allowed, and then formats it. But it should be other way around in vscode..

[pallid frigate]

didn't work

py print("")

[dry pike]

you have a space after the py

it needs to be the last thing on the line

[pallid frigate]

oh

[knotty ice]

will python save the world from A.I ?

[pallid frigate]

user_input = input("Enter File.EXTENSION: ")
with open(user_input, "w") as file:
    file.write("Hello World\n"[0])
    file.write("Hello World!\n"[1])
    file.write("Hello World\n"[2])
    file.write("Hello World\n"[3])
    file.write("Hello World\n"[4])
    file.write("Hello World\n"[5])
    file.write("Hello World\n"[6])
    file.write("Hello World\n"[7])
    file.write("Hello World\n"[8])
    file.write("Hello World\n"[9])
    file.write("Hello World\n"[10])

I wouldn't do this but you know

maybe

[craggy willow]

should post this a few more times in case we didn't see it yet

[pallid frigate]

working on python thing

[gleaming knoll]

the interpreter seeing you write car by car ::
the kernel doing the write syscalls car by car ::
😭 😭 💀 🥀

[dry pike]

you can just edit the original msg bro

[dry pike]

car by car?

[gleaming knoll]

teh h in char is deaf
and mute

its probally dead tbh

caracter

[craggy willow]

chairacter

[dry pike]

car actor

also known as Vin Diesel

[craggy willow]

actor based language is really just string based language

[pallid frigate]

user_input = input("Enter File.EXTENSION: ")
with open(user_input, "w") as file:
    x = 0
    while x < 20:
        file.write("hello worldaaaaaaaaaaaaaaaaaaaaa\n"[x])
        x += 1

Use a loop

[raven urchin]

[crisp jay]

dry: do repeat yourself

[raven urchin]

LMAO

[pallid frigate]

nah

[golden mortar]

Sometimes repeating yourself is the best choice

[charred tusk]

Yeah two similar functions with a different key detail can be cleaner than some big abstraction to handle both cases

[wheat mica]

I love how writing the 1st 20 chars of a string is less loc in C than Python

Or u jst use some annoying string slicing

[crisp jay]

sometimes sed is the solution

[wheat mica]

This is the code btw ```c
(void) printf("Enter filename: ");
char user_input[1024];
(void) scanf("%s", user_input);
FILE* file = fopen(user_input, "w");
fwrite("hello worldaaaaaaaaaaaaaaaaaaaaaa\n, 20, sizeof(char), file);

[dry pike]

"annoying"

[wheat mica]

String slicing is kinda annoying imo

[dry pike]

that's a funny joke

[gleaming knoll]

i dont think thats a valid usage of scanf
this also doesnt check for errors, and doesnt close the file

[wheat mica]

Fk, im stupid

My brain is cooked rn

[crisp jay]

also syntax error

[gleaming knoll]

eh, thats obvious, no need to mention it

[tame vapor]

scanf("%s"); seeing this gives me chills

[wheat mica]

It’s a segfault waiting to happen

lol

[gleaming knoll]

its a compile error waiting to happen, assigning it to char* that is
and yeah reading an arbitrary sized string is not a good idea

[wheat mica]

Ignore the warnings and it’ll have 3 segfaultable locations

[crisp jay]

i love writting C and my code doesnt work on one compiler but works on the other

very cool

[tame vapor]

dont write bad code

[gleaming knoll]

you probally wrote some really weird C then

[tame vapor]

or used some compiler extensions, or non standard C stuff

[wheat mica]

My brain is kinda fried rn ngl

[nova breach]

can forky install python newer than trixie without uv?

[crisp jay]

maybe its alloca

[wheat mica]

im usually not this bad when it comes to writing C

[tame vapor]

best to ask the C questions in C server instead of python discussion xd

[crisp jay]

im going to get cooked

[tame vapor]

why ?

[wheat mica]

we all write bad C sometimes DW

Like this: #python-discussion message

[gleaming knoll]

forky and trixie being the debian releases?

[nova breach]

yep

[wheat mica]

I see 2 segfaults waiting to happen

[gleaming knoll]

you dont need uv on any platform to install python versions. its just one of the version managers
you could like, compile from source

[pseudo sorrel]

hi

how do i get pic perms

[lament veldt]

is there any difference b/w typing.NamedTuple and @dataclass with frozen=True?

[gleaming knoll]

a namedtuple is a tuple, a dataclass isnt. so a namedtuple can be like, unpacked, or passed to something which checks for isinstance(_, tuple)
in terms of memory usage : each tuple instance has to store the amount of fields, the instance of a (slotted) dataclass doesnt

both are "immutable" in the same way: cant reassign what the fields point to, but if the fields are mutable - can be mutated
(though bypassing this is easier for frozen dataclasses, just object.__setattr__)

[nova breach]

anyway, apt can only install a relatively old version python in trixie

[terse mauve]

yes. a named tuple is much more lightweight. but dataclasses has more functionality

[gleaming knoll]

add slots=True to the dataclass (should be the default tbh) and now the dataclass is 1 size_t smaller than the tuple

[lament veldt]

i was thinking that but i use dataclasses everywhere in the code, i was wondering if its bad for consistency

[lament veldt]

how's the tuple is more lightweight if it has more 1 size_t?

[gleaming knoll]

the tuple is only more lightweight if the dataclass doesnt use slots

[terse mauve]

You can use both. Use whatever fits your needs.

Using dataclasses everywhere "just because" can actually slow your program by a lot (for big applications)

[gleaming knoll]

i wonder if you can propose a program where either of the options is significantly slower/faster, or uses significantly more/less memory

[hybrid nebula]

I don't see how dataclasses can be faster

[crisp jay]

yea it was alloca

[hybrid nebula]

alloca

[gleaming knoll]

it doesnt really make sense to measure speed here, its nanoseconds for access in all cases. memory - sure, and there is an objective answer: an instance of a class with slots is smaller by atleast 1 size_t than an equivalent tuple because it doesnt need to store its size

[hybrid nebula]

do dataclasses have slots?

[gleaming knoll]

with @dataclass(slots=True), yes

[hybrid nebula]

ah

[spice hill]

actually, dataclass might take up less memory

[hybrid nebula]

really? how?

[tawny gale]

Hi all

[hybrid nebula]

iirc tuples don't overallocate

[proud escarp]

where did i hear something about slots not mattering as much from 3.11 onwards

[terse mauve]

dataclasses are just code generators, you can define slots, ordering, frozen (and many more), dataclasses are great when you have a class that you'd want to implement those things for you anyway, or when there's a need for that class instance to be converted to tuples/dicts

[spice hill]

!e

from typing import NamedTuple
from dataclasses import dataclass
import sys

@dataclass(slots=True, frozen=True)
class DcPoint:
    x: int
    y: int

class NtPoint(NamedTuple):
    x: int
    y: int

print(sys.getsizeof(DcPoint(420, 69)))
print(sys.getsizeof(NtPoint(420, 69)))
print(sys.getsizeof((420, 69)))

[edgy kraken]

:white_check_mark: Your 3.14 eval job has completed with return code 0.

001 | 48
002 | 64
003 | 64

[spice hill]

Tuples store their length as part of their layout

i think that's why

[hybrid nebula]

ah

[terse mauve]

imo using dataclasses like you'd use namedtuples or SimpleNamespace is an overkill

[hybrid nebula]

so just what we said, slots

[proud escarp]

oh, october is the month the python versions die
rip 3.10... 10 being october... fitting..

[hybrid nebula]

october is where pythons go to die
...wait what

[spice hill]

wake me up when october ends

[proud escarp]

i cant believe 3.11 will be the lowest living version soon... to me it still feels like it just came out..

[steady rain]

you know what's gonna happen in four days?

[spice hill]

may be

[terse mauve]

in my head 3.6 just came out with fstrings

[proud escarp]

i think python was 3.6 when i started using it

or 3.7

i was inspired by this person in my scratch.mit.edu class who was making games with pygame

[harsh swallow]

I remember when I had to use 3.4 💔 and when I used 3.10 for some stuff 3.10 as current newest. Damn

[gleaming knoll]

https://github.com/python/cpython/pull/28802
but slots are still a non-zero improvement

[spice hill]

Feel old yet? 2015 was 20 years ago.

[proud escarp]

nice one

[proud escarp]

noice

[terse mauve]

😁

[crisp jay]

for some reason uv defaults to 3.10

[crisp jay]

1% faster

[raven urchin]

Hello

[harsh swallow]

I love that you make me feel even older than I really am...

[spice hill]

I was under the impression slots are for memory, not speed

[harsh swallow]

Why not both?

😁

[dull dune]

definitely both, you avoid the key lookup in teh dict

[spice hill]

true

[wet sail]

Hello, I'm sorry but am I allowed to send an OPTIONAL to join discord server message?

I'll wait for a staffs response

[steady rain]

you can't post any guild invites that aren't in our allowlist, if that's whata you're asking.

guild invites that aren't allowlisted will be instantly deleted.

[bronze dragon]

Why emphasize the "OPTIONAL to join"? All server invites are inherently optional

[steady rain]

discord.gg/python is actually mandatory

[wet sail]

It's like a double barrier but good thinking

[velvet trout]

Guys i need help

[steady rain]

well, don't bury the lead.

[wet sail]

If you want can I give you some insights about the server so you can decide further? Because I personally feel like it would be worth it

It's up to you here

[steady rain]

you can send a message to @pseudo carbon about it.

[wet sail]

Sure

[brisk gazelle]

The EPA might want to have a word, elsewise.

[gleaming knoll]

yeah its not supposed to be about speed, its a memory usage optimization
though all performance related changes do speed benchmarks too

[velvet trout]

def lex(lexemes: Iterable[str]) -> Generator[Token]:
    end_of_options = False

    for lexeme in lexemes:
        if not end_of_options:
            if lexeme == END_OF_OPTIONS:
                end_of_options = True
                continue

            if lexeme.startswith(OptionPrefix.LONG):
                yield _lex_long_option(lexeme)
                continue
            elif lexeme.startswith(OptionPrefix.SHORT):
                yield _lex_short_option_or_group(lexeme)
                continue

        yield ArgumentToken(lexeme)

I know what subclasses of the Token class (base dataclass) will be yielded.
And for type checkers, it's hard for them to know what subclasses of Token are available. Hence, matching subclasses of Token exhaustively will require including the base Token class itself which is unnecessary and vague.

I have a type LexerToken = OptionToken | blah blah, should i type annotate the YieldType as LexerToken rather than generic?

I have always yielded a generic type (indoctrination and now sudden change makes me feel concerned) until the day i found that it's hard for the type system to know without sealing types (which isn't present in python currently).

So do I yield by LexerToken or do i not (go generic Token)?

Or is there something else that i am not aware of that is perhaps more appropriate over these two options?

[wet sail]

I just did, thank you

[cerulean ravine]

"lede"

[steady rain]

they just deicded to spell it that way so it's not the same as Pb.

[cerulean ravine]

I know 🙂

[cerulean ravine]

their workplace has a lot of Pb and things named for Pb

[brisk gazelle]

Pb&J.

[velvet trout]

Hmmm... So this kind of thing is called as casual-chatting? where you don't stay efficient by putting the help question directly with your 'i need help', which is a tradeoff for casual chatting? And lots of people on internet are casual in chatting which is why they don't put help question in same message and sometimes rather wait for someone to attend for helping them hmm good discovery

[gleaming knoll]

class Token: ...
class OptionToken(Token): ...
class SomeOtherToken(Token): ...

def f() -> Token:
  return OptionToken() if random.choice([False, True]) else SomeOtherToken()

#
def g():
  t1 = f() # just some abstract Token
  if isinstance(t1, OptionToken): # t1: OptionToken in this branch
    ...
  else: # even though in practice this will always be `SomeOtherToken`, the `Token` type does not tell us that
    ...
# (OptionToken | SomeOtherToken) would provide this information though, so in the `else` `t1: SomeOtherToken`

thats the main difference

[velvet trout]

What is Pb and J

[gleaming knoll]

peanut butter and jelly

[main swan]

Lead and Jam.

(in a sandwhich of course)

[velvet trout]

Loll

[raven urchin]

[gleaming knoll]

so if you can specify the exact union of types rather than just the base class, it simply provides a bit more information. some programs will need to do less narrowing
its not a big deal

[raven urchin]

After I eat, I'll code. Maybe I should make blackjack

[fathom beacon]

What do I do in a for loop to format a list nicely?

[main swan]

Classic programming assignment. Also one I've never done lmao.

[fathom beacon]

for string in colors? is string valid

[raven urchin]

Oh

[gleaming knoll]

you can have any assignment target (an assignment target is something that can appear on the lhs of an assignment, e.g. variable name) after for
string is a valid assignment target, as in, you can do string = something, so yes
though usually if you have colors, you'd call the singular variable color

[main swan]

Maybe I should make blackjack

[fathom beacon]

doin that rn haha

[fathom beacon]

so for string in color: is good?

[raven urchin]

Oohhh, how much have you done so far?

I have to investigate about the Blackjack rules

[gleaming knoll]

its valid. whether its good depends on what is color

[fathom beacon]

I wanted to get familiar with the random module but ended up just using some tut code and improvising from that, kind of like a mini gam,bling

[velvet trout]

t1 knows is OptionToken | SomeOtherToken ?

How come else knows its SomeOtherToken since t1 is a base Token?

[fathom beacon]

color is a list of 6 different strings

[spice hill]

Check out our latest Pb&J snack. Indulgent and sweet, it boasts life-shortening and intelligence-reducing properties while being quite heavy in the hand!

[gleaming knoll]

it doesnt

[velvet trout]

Ooh

[fathom beacon]

I just don't get what I can use in for foobar in color

like number works

but it doesn't give me a list of stuff i can use

[gleaming knoll]

the thing after for is what the values from the iterable will be assigned to, so you're inventing a new name, not using an existing one
same as x = 1 just makes a new variable x out of nothing, for x in something: will be assigning the values from something to x one by one and running the inside of the for loop

[fathom beacon]

why do u edit so much lol

[gleaming knoll]

idk, its a habit
i dont like writing the message out fully before sending it
so instead i write a bit, and then iteratively edit

[fathom beacon]

I do

[velvet trout]

So i guess its better to Explicitly type annotate a Union of what we would return in f() , right?

[brittle merlin]

it feels more like human talk

[gleaming knoll]

it provides more information, whether that is enough to be better is questionable

[main swan]

Very git-esque

[drifting hollow]

does anyone know how i can include homebrew packages in a bundled app using pyinstaller?

[velvet trout]

Yeah true

[raven urchin]

numbers = [1, 2, 3] 

for number in numbers:
    ...

for i in numbers:
    ...```

They're the same.

[drifting hollow]

i can't find any answer onlişne

[fathom beacon]

OHHHHHH

[brittle merlin]

yeah

[fathom beacon]

Bruh that's crazy

[gleaming knoll]

with a union of specific types, if you then add yield SomeNewTokenType() in the function, you'd need to change the signature
if you annotate it as simply Token - you dont, it will work for all token types

[drifting hollow]

a programming rule : don't take names serious

[fathom beacon]

so i can put for skibiditoiletgang in colors

[runic flower]

mostly. i is generally used as an index variable, so it could be confusing in the second usecase there

[fathom beacon]

lmao got it

[velvet trout]

That is what helpful for typecheckers, so i think its appropriate to yield LexerToken

[raven urchin]

It's pretty useful. But you can name anything anything

[raven urchin]

I just wanted to demonstrate that the name doesn't matter

[fathom beacon]

Riiiiight thanks guys

[raven urchin]

Well, it does, but in this case you assign whatever name you want.

[runic flower]

which is why I replied with "sometimes the name matters" 🙂

[main swan]

😛

for i, number in enumerate(numbers):
    ...

[drifting hollow]

but don't take it too unserious after learning (atleast to make other peoples and possibly your life easier)

[raven urchin]

Glad to help. :D

[velvet trout]

Tho i would need to maintain the LexerToken union, but i maintain all the tokens so i guess these not a bad thing...

[fathom beacon]

Oh mine kidna steered into the direction of like case opening

[brisk gazelle]

My standard is this: Descriptive and concise.

[raven urchin]

Don't understand what you're referring to... CS case opening
...?

VASS GOLD GOLD GOLD GOLD

[velvet trout]

I do not see any other third option either. Ima lock in yield LexerToken union

[gleaming knoll]

😭 ohnepixel

[fathom beacon]

Okay I thought it had to be a data type or something

[raven urchin]

Bless this guy 🙏

[brisk gazelle]

The idea being the enhancement of legibility.

[raven urchin]

100k in skins, -100k in skill.

[brittle merlin]

bro he really ate packing peanuts

thats the funniest clip ever

[main swan]

don't have to worry about naming variables if you write pointfree...

[brisk gazelle]

They can be made out of cellulose, now.

[brittle merlin]

him telling its fine and gagging in the end

[brittle merlin]

i guess

[raven urchin]

Funny exercise I like is re-creating built-in methods, functions and other stuff

[fathom beacon]

Yeah I'll name it something appropriate

[main swan]

That's a pretty valid exercise.

[fathom beacon]

CS2 yuh

I've spent 1600€ on that

[crisp jay]

i: int = 0
while (i < len(numbers)):
    ...
    i += 1

[main swan]

Sure, but mine gives extra redundancy.

So in case i fails, you still got number.

[crisp jay]

more LOC == more productivity

[raven urchin]

Holy God

I spent the 15$ of Prime and that's it

[steady rain]

why the self-evident type hint and the () for the while?

[fathom beacon]

Yeah I love gambling and I have disposable 'income'

[gleaming knoll]

i = 0
iterator = iter(numbers)
while True:
  try:
    next(iterator)
  except StopIteration:
    break  
  ...
  i += 1
del iterator

[fathom beacon]

are there active vcs in this server

[steady rain]

Python will never be statically typed, but each release seems to have more support for type annotations.

[raven urchin]

I need some of that disposable income, because I dislike gambling, I would do amazing stuff with it... Home labbing would go brrrr.

[crisp jay]

i like static type

its fun to tell the types of variables

[brittle merlin]

they should just

do it man its getting hella confusing now

[fathom beacon]

5 days ago I won a 320€ AK47 - B The Monster

0.13%

[raven urchin]

Oh my god.

[brittle merlin]

fine

[raven urchin]

#ot0-psvm’s-eternal-disapproval

Come here

[steady rain]

dynamic typing is incredibly baked in to the language. switching to static typing would be creating a new language, basically.

[fathom beacon]

I love gambling because It's fun, It's a waste of money if you don't enjoy it tho

[runic flower]

I mean, other languages do exist. You don't have to confine yourself to just one.

[steady rain]

right in front of my salad?

[spice hill]

In addition to changing the language, the existing type annotation system and type checkers would have to be completely scrapped too

[gleaming knoll]

(as they should be tbh)

[main swan]

I'm working on a gradually typed language like python that becomes more efficient with additional typing.

[brittle merlin]

if hypothetically it did happen, what would happen to all the packages and the community?

[spice hill]

99.9% of popular libraries will stop working and will have to be rewritten

[steady rain]

the python community probably wouldn't adopt that new language, and it wouldn't have a package ecoystem unless a lot of people basically started from scratch to create it.

[brittle merlin]

Im deep in python tutorial hell icant get out help aaaghhhh

[main swan]

Imo "don't have to" should be replaced with "shouldn't"

[runic flower]

python is easy. programming is hard. be glad you're not learning programming using a harder language. 🙂
when you get stuck ask questions. 🙂

[crisp jay]

i: int = 0
iterator: Iterator[int] = iter(numbers)
while (1):
  opcode: int = next(iterator,-1)
  if opcode == -1:
      break
  ...
  i += 1
del iterator

[raw bramble]

How would it work?

[main swan]

you should warp your brain with other languages like Haskell, Prolog, etc.

[gleaming knoll]

the way it works in actually gradually typed languages like C#

[brittle merlin]

why is that? ik its a crazy thing to think but couldnt there be a way, where python becomes static and then all the developers just have to go and add type annotation to the exsiting dynamic code

[raw bramble]

I’m not remember how it work :(

[main swan]

and then anger the community when you use weird ideas in the python community you brought from those communities.

[gleaming knoll]

numbers famously never include -1

[steady rain]

you can like verbose type annotations, but I don't get the parentheses around loop statements.

[crisp jay]

i like C

[runic flower]

generally when fundamental changes of that scope are introduced it just becomes a new language.

[gleaming knoll]

add them in if then too

[raw bramble]

Reform vs Revolution. Either it would take forever, or it would be a different language

[brittle merlin]

im ngl, i really really enjoyed my time learning java for college because how concerete that language felt. im thinking if i should persure this path or not haha

[raw bramble]

And even then, 3.14 is a different language from 3.15

[runic flower]

Java's not my favorite language. But at least it makes Kotlin easy to pick up. 🙂

[raw bramble]

Is there a version of Python that’s green

[brittle merlin]

ohh theres so many kotlin developers in india i think

[raw bramble]

Like it’s a bright green coloured Python rather than a blue and yellow Python

[spice hill]

all the developers just have to go and add type annotation to the existing dynamic code
It's not how programming works. It's not like Python users are children and Python core devs are their parents who can tell them what to do.
If Python 3.X+1 introduces a change that means manually changing millions of lines of people's code, Python users will just stick to Python 3.X, and they will stop financially supporting the Python Software Foundation because it clearly went off the rails.

[crisp jay]

the perl 6 incident

[spice hill]

Python's type annotations are not foolproof, they have a lot of "holes". It's not like Java or C or C++. Any would be an example of such hole: ```py
def f(x: Any) -> None:
x.foo().bar() # type checks fine

f("banana") # also type checks fine

[pastel sluice]

one of the ways devs can go is to choose a linting tool that checks type annotations, and to set it for maximum strictness, so that any code without consistent annotations gets flagged. but that's entirely opt-in and is not enforced at runtime.

[brittle merlin]

understandable

its just how python is

[main swan]

I'm not sure Java, C, or C++ are sound. I thought things like null ruin soundness iirc.

(insert Rust my beloved)

[spice hill]

maybe C and C++ were a bad example, with how easily you can screw things up

[main swan]

nah it's all good. I get what you mean. I'm just trying to help if you get some type theorist that says "Well acsuallly..."

[spice hill]

I probably wouldn't use the word soundness myself, that's just the name of the repo (:

For instance, if you write this function in Rust: rs fn banana(x: u32) { } there's no way to make it so that x is a string. It wouldn't even make sense in the universe of Rust, and banana won't accidentally use string methods on x even if you do screw this up in some major way

(well, there are some bugs in rust that sometimes lead to unsoundness iirc)

[rancid fractal]

Tbh Kotlin resolved most of Java issues with its stuff like String vs String?, but it still has issues with Java interlop types and unchecked casts

[gleaming knoll]

mostly lifetime related

[spice hill]

yeah

[main swan]

I'm really unsound personally. I'm just used to it by now.

to me everything is just a set. ZFC supremacy.

[gleaming knoll]

im trying to use the coalton mine IDE but it has no vim motions so im dying
its interesting that they tried to make an IDE

[main swan]

why are you forsaking the goated editor?

[pseudo sorrel]

i made my first ever python tool today

[gleaming knoll]

i would be an emacs user if i learned about emacs before i learned about vim

[main swan]

Honestly, I just need my editor to have easymotion variant for it then I'll use it.

[spice hill]

It's also about tradeoffs. For example, this is a narrowing that type checkers support: ```py
class X:
prop: object

def foo(x: X) -> None:
reveal_type(x.prop) # object
if isinstance(x.prop, int):
reveal_type(x.prop) # int
``` this is clearly wrong because x.prop could change after it's checked (or it could be a @property, in which case this narrowing still works). But prohibiting this was decided to cause too much inconvenience

[main swan]

Being able to do:

>>> x = 3
>>> x
3
>>> x = "happy"
>>> x
'happy'

Is actually wild. I'm working on a compiled python like language, and getting this functionality is surprisingly annoying.

[steady rain]

why would you ever do that?

[cerulean ravine]

do you want to support that?

[spice hill]

I like this feature in Rust, but I would also understand if the language doesn't want to support it

[main swan]

I'm trying to keep the language as fun as possible. So my plan is: if there is no type then you can. if there is a specified type then you can't.

[gleaming knoll]

a compiled language would usually separate assignment from declaration
then this is easy to implement

[spice hill]

Is it statically or dynamically typed?

[main swan]

I'm aiming for gradually typed.

I originally tried re-implementing python, but there were too many issues. So, now it's mostly python like, gradually typed.

[spice hill]

Why gradual typing for a new language?

[main swan]

The idea is prototype fast, optimize later.

I want something that you can quickly spitball ideas like Python, but if you want to sit down and type it later, it ends up being pretty fast too.

I mean, maybe it sounds stupid.

but I kind of want it at least personally

[spice hill]

I think the thing that would be popular today or in 10 years is less in the direction of lawlessness, but rather a statically typed language that introduces types in a very non-bureaucratic way. Things like: type inference everywhere, structural records/tuples

[pastel sluice]

@main swan this was an idea I explored myself - anything untyped would be an "object" type that would require a specific runtime, anything typed would be native code. You could declare a "strict" mode so that the program would not compile if any "object" types remained. Didn't get very far as I was simply not good at language or compiler design

[main swan]

Well, if you get interested again, we can team up maybe? I have a repo for the language and stuff.

[spice hill]

(but I'm also not a language designer)

[gleaming knoll]

bro described roc

[spice hill]

yeah, something like that

[pastel sluice]

I have way too much else to do :D I'm leaving all this to people far more dedicated and knowledgeable about the field than I am

[gleaming knoll]

just use common lisp

[main swan]

I have no idea about language design lol. I need someone to guide me too haha. But alright. no worries. 😄

i've also been getting side tracked by other smaller projects i'm working on. but we'll get there eventually (I think?)

[spice hill]

-# 🪦🥀

The language does sound interesting though.

[main swan]

Alright. I'll keep working on it then! 🙂 (I think I've just been having esteem issues recently. My website idea I implemented was torn to shreds by the math community lol)

[spice hill]

-# that comment was about Roc

but, good luck to you as well 🙂

[main swan]

Ooh, it does look interesting indeed.

[spice hill]

unfortunately you won't be able to write it on paper

[pastel sluice]

what happened to your website idea (and what was it?)

[inner adder]

There are so many cool languages that I get curious about and then see that they're dead, too new to really enjoy playing with, or in just some sort of limbo.

Or you get that "Eh... why learn something new when it's not going to go anywhere" moment

[gleaming knoll]

real

[spice hill]

RIP Elm

[pastel sluice]

My interest in languages as such wanes the more I find I want to work on some specific project, because 99.99% of the time I can do what I want with Python, C (via Cython), or one of the other top few languages out there.

[main swan]

It was an apparently very spicy math site similar-ish to Project Euler: Read some of the Reddit comments about it

[pastel sluice]

And that says at least as much about my needs and interests as it does any of those languages as such

[inner adder]

This. I was SO stoked for Elm to be more than it ended up being

[inner adder]

I just have this habit of wanting to try new things for the sake of learning them. I think that's mainly because I do it as a hobby, so it doesn't matter a whole lot if it's useful.

But the moment I see that a language doesn't have string interpolation, I nope out of there

[pastel sluice]

I think they had a valid point: how do you calculate the closeness of a solution to a problem that doesn't have one yet?

[gleaming knoll]

imagine learning about dependent types to end up writing dynamically typed slop

[rose basin]

Hi there, I wish you all having a great day

[main swan]

It's mostly for fun: but the point is that you can frame each problem as an optimization problem.

[gleaming knoll]

the fact that even 1 new language became so used (rust) is already insane tbh

[pastel sluice]

I mean, if you want to build it for fun, build it

[main swan]

It exists. (I have built it and apparently lost all my credibility in the math field)

[spice hill]

tbh, I can live without string interpolation (if it buys the language something else)

[pastel sluice]

as I do so much work with text, I cannot

[prisma cloak]

hey can anyone help me with vs code theme change i tried but it won't work even though i have uninstalled extension

[pastel sluice]

one post that garnered some negative reactions equates to total, permanent, irrevocable loss of credibility?

[main swan]

I have had some irl interactions that have also been unsavory once mentioning that I made the website lol. But screw em anyway, i'm gonna keep working on it (but I have some other more immediate stuff to do atm)

[rose basin]

speaking of rust, I don't get why uv can't resume a download

[grave tree]

If you'd like some feedback, I think it would be better received if you answered some of the obvious questions up front when posting/talking about it.

Something like how you define closeness would be good to know upfront when first hearing about the project, since that is the obvious question. It shows you've put thought into it and how you're approaching things.

Also referencing its closest analogs so people can have a good comparison point.

I think it's a fun project and some adjustments on how it's communicated could make it be received better

[main swan]

I think I did tho unless I'm misunderstanding. my comment got downvoted to oblivion (check below the top comment)

[grave tree]

My point is that it's not in your post. Someone had to ask and then you responded. That's not being upfront

[main swan]

Alright. I see what you are saying.

[grave tree]

It would also help to answer some of this stuff in the repo README as well.

For better or for worse, how something is communicated can have a large impact on how people judge the underlying idea

[gleaming knoll]

what can it "buy"?

[rose basin]

true

[main swan]

well, lesson learned. Unfortunately, I can't really re-post back there again. I will update the README soon.

[spice hill]

Idk, simplicity? For example, Elm doesn't have traits.

[grave tree]

I mean, it's reddit. Sure you can just give it a few months and with some improvements

[gleaming knoll]

okay yeah that makes a bit of sense (though elm has stuff like comparable)

[spice hill]

yeah that's a bit undercooked

[gleaming knoll]

if a language has some form of implementing a generic operation for a given type then fstrings seem simple

[spice hill]

Well, you could have interpolation that only supports strings, like "Hello, ${name}!". But it'd be kinda undercooked too.

[gleaming knoll]

you'd end up with ${x.show()} everywhere

[spice hill]

That is also an option. Though Rust doesn't have this, only allowing simple expressions like variable names and attribute accesses in interpolated strings

(but it does have Display and Debug)

So you sometimes end up having to go back to println!("foo {0} bar {1} baz {2}", expr1, expr2, expr3); anyway

[frigid trout]

Hello

[ocean ridge]

vibrant is reserved by pypi

i thought i found a cool name for my project and it throws an error

[cerulean ravine]

what error?

[ocean ridge]

Uploading vibrant-0.1.0-py3-none-any.whl
100% ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 12.5/12.5 kB • 00:00 • ?     
INFO     Response from https://upload.pypi.org/legacy/:
         400 Bad Request
INFO     <html>
          <head>
           <title>400 The name 'vibrant' isn't allowed. See
         https://pypi.org/help/#project-name for more information.</title> 
          </head>
          <body>
           <h1>400 The name 'vibrant' isn't allowed. See
         https://pypi.org/help/#project-name for more information.</h1>    
           The server could not comply with the request since it is either 
         malformed or otherwise incorrect.<br/><br/>
         The name &#x27;vibrant&#x27; isn&#x27;t allowed. See
         https://pypi.org/help/#project-name for more information.


          </body>
         </html>
ERROR    HTTPError: 400 Bad Request from https://upload.pypi.org/legacy/   
         Bad Request

[spice hill]

maybe it's automated typosquatting protection

[ocean ridge]

yeah

[spice hill]

vibrance is a package that exist

[ocean ridge]

vibrance? that isnt as good as vibrant though

i've decided to go with vibrantcli

[cerulean ravine]

what does it do? Are you going to support it into the future?

[ocean ridge]

or could u suggest some other good names?

[spice hill]

I meant that vibrance is a thing that already exists on PyPI. So maybe PyPI automatically disallows taking a similar name.
I could be completely wrong though, I don't know the details of how anti-typosquatting works onPyPI

[ocean ridge]

no its a hobby project. Its something like colorama. colorful ansi text. it does have suport for windows cmd which only a few color text providers give

[gleaming knoll]

vibransi

[inner adder]

colors_are_cool

[ocean ridge]

too long isnt it

[spice hill]

Does it need to be on PyPI?

[ocean ridge]

if i do pip install <lib> it installs from pypi right

[dull dune]

did you go to the urL?

[gleaming knoll]

real
just install with git from some repo hosting

[dull dune]

The project name is too similar to an existing project and may be confusable.

[inner adder]

oh_hey_is_this_a_color_library_that_is_dope

[ocean ridge]

yes

[ocean ridge]

i want to have one package of my own on pypi 🥺

[spice hill]

You can also install things from a git repo https://stackoverflow.com/questions/20101834/pip-install-from-git-repo-branch

[dull dune]

then I'm not sure where the confusion is

[unborn lagoon]

The name vibrant conflicts with a popular JS project: https://github.com/Vibrant-Colors/node-vibrant

[ocean ridge]

kk

[dull dune]

sincerely, unless it's something useful that you intend to actually support in good faith goin gfowrard, don't.

[ocean ridge]

can you all suggest some good names?

[dull dune]

everyone does that

[ocean ridge]

sure il work on it a lot but theres only finite things that could be done here in this project

how about vibgyor?

[frosty oriole]

chatgpt could actually be good for this

[gleaming knoll]

im ngl, just package rather than author/package is making me sad

[fervent matrix]

a_super_duper_colour_lib

[ocean ridge]

chromapy?

[frigid trout]

You must be Veltrix? Do you use Python for your websites? I heard that you make websites from B3ttaFish

[ocean ridge]

you are bettafish stop the bluff

[gleaming knoll]

they are clearly alphafish, not beta
display names never lie. they cant be set to arbitrary things

[frigid trout]

No I am AlphaFish.

BettaFish's friend

[cerulean ravine]

how does this library differ from other libraries that help with terminal coloring?

[unborn lagoon]

I guess you could do author-package, but yeah, namespaces of that kind would be nice.

[charred tusk]

GitHub freaking out
I have work to do 🙁

[ocean ridge]

mine supports windows cmd ansi coloring like colorama and has rgb + hex support which colorama doesnt have. and its really easy to use

[ocean ridge]

what happened

[robust ledge]

It's telling you to take a break.

[inner adder]

Jesus christ, all the ones I'm looking for are already taken or are being name squatted

[frigid trout]

When I tried to use Github earlier today it gave me some weird error message and saying that I can't upload my file.

[charred tusk]

I think receiving 7 NDRs saying my now 44mb email chain is too big should have been my first cue

[ocean ridge]

rainbowcli?

[inner adder]

Looking

That looks up for grabs

[ocean ridge]

yeps

[gleaming knoll]

but your library is not a cli
its a library

[charred tusk]

https://www.githubstatus.com/

[ocean ridge]

sadge

but hey do i care?

[spice hill]

name it rainbowgui

[cerulean ravine]

you should

[dull dune]

(thus the urging to not release to pypi)

[ocean ridge]

its certainly not a gui

[spice hill]

well it's not CLI either

[chilly whale]

There is a process to request that a name squatted PyPI name be handed over to you. No guarantees, but it's possible

[charred tusk]

It sure does take a while

[ocean ridge]

ok wait il think some more

[gleaming knoll]

name it veltrix-ansi-lib and be done with it

[spice hill]

I just remember that I have an almost-empty PyPI package I used for testing purposes. Hopefully PyPI doesn't kill me

but the name is unlikely to be contested

Yeah, namespacing by putting your name sounds like a good move, if you want the project on PyPI

[ocean ridge]

kk

[frigid trout]

Name it Gilbert.

[ocean ridge]

il brainstorm a few more names and if i cant find a good one, il use my name

[charred tusk]

bro
How did you know my name?

[ocean ridge]

@inner adder managed to find any?

[inner adder]

Still thinking

[frigid trout]

What, that''s Gilbert's name. My profile picture's name

[inner adder]

I'm usually good at naming things but so many are already taken

[ocean ridge]

ikr me too

[frigid trout]

Gilbert-The Defecttive Fish.

[charred tusk]

Yeah -- all of my projects have my org name on them, $DAYJOB or otherwise

[inner adder]

Got it

mantis_shrimp

[gleaming knoll]

Gilbert's name is Gilbert

[ocean ridge]

rainbow-terminal?

[inner adder]

They can see a significantly larger number of colors than other animals

And they're really colorful as well

[ocean ridge]

how is that even remotely related to my project 🙃

[frigid trout]

Is it common practice to make different repositories for different files?

[charred tusk]

of all the animals in the world...

[frigid trout]

You can name it Sparrow, since they also can vibrant colors

[charred tusk]

We give colors to SHRIMP!?

[fervent matrix]

What would you call a recipe book app (I'm trying to come up with some some names)

[ocean ridge]

ima name it rainbow-terminal. sounds cool too

[gleaming knoll]

what do you really mean
a repository is supposed to be for a project. a project can include many files

[inner adder]

Look at this colorful little bastard

[frigid trout]

Like 1 repo for 1 .py file and etc

[spice hill]

smh image perms abuse

[frosty oriole]

lowkey a good idea

[ocean ridge]

many ppl wont know this colorful little bastard

[frigid trout]

Name it Manti

[frosty oriole]

mantis shrimps also can see a lot of color

[frigid trout]

mantI

[inner adder]

They will if you make it the logo for the package

[charred tusk]

you're a good idea

[ocean ridge]

mantis-colors

[frosty oriole]

thanks

[frigid trout]

What is this for?