🪅-progaming

[valid jetty]

yeah it's crazy

i wanna see a real dragon one day

all 27 wing software users will cry

[frosty obsidian]

zeet doesn't realize the world runs on frogs

[valid jetty]

wing you would love this graphics programming server im in

they're kinda obsessed with frogs

[deep mulch]

@valid jetty https://tetris.wiki/Super_Rotation_System i need to implement this

you might like

[frosty obsidian]

im not actually that into frogs, my pfp just stuck

[deep mulch]

wing has a framed portrait of his pfp on the wall

@frosty obsidian you wear a shirt with your pfp on it

[valid jetty]

i know what this is one of my friends made tetris for a school project

the same project that im making a compiler for a japanese language for lol

[frosty obsidian]

zeet loves that tetris is very strictly standardized

[deep mulch]

yop

[frosty obsidian]

you have so much work to do if you want compliance

[deep mulch]

you will do

[frosty obsidian]

nop

[deep mulch]

you will be play tester

[frosty obsidian]

haven't played tetris in forever

[deep mulch]

you have played minecraft too much

@frosty obsidian plays minecraft and doomscrolls twitter

[valid jetty]

@lucid trail did you figure out if it works with a lower triangle count?

the dragons lower in triangle count as you go down the order

because i started with dragon.obj and then kept adding more triangles once i saw "oh, it's working with this many just fine.. can it handle more??"

until i got to the full dragon at dragon4

iirc dragon 1 is 0.005, dragon 2 is 0.05, dragon 3 is 0.1, dragon 4 is 1

in terms of decimated scaling

[frosty obsidian]

@deep mulch the compose pdf rewrite looks good

nothing stuck out to me

[lucid trail]

yeah it does

[valid jetty]

ok cool

[lucid trail]

works with cube

works with penger too

[valid jetty]

yeah that's expected those have like barely any triangles

[valid jetty]

finally

[valid jetty]

https://media.discordapp.net/attachments/390264006402899979/1433647811232268348/render_srcnn_photo_noise3.png?ex=69057412&is=69042292&hm=93028fb6df417785f5c618faee188280ab1924257c2adc9611bbad85b11a8719&=&format=png&quality=lossless&width=3123&height=1757

4.4m triangles!!!

[jade stone]

how much fps

[valid jetty]

like 0.4

[deep mulch]

@valid jetty make it realtime when

[lucid trail]

@valid jetty it was a raylib issue

very cool

[stoic helm]

cant give discord bots display name styles

[ornate quiver]

spooky

[valid jetty]

what was the issue?

[nimble bone]

@valid jetty

rewrite Elle in Elle

[valid jetty]

soon

[icy geode]

Is there a CSS snippet to remove custom status of users from the server list if so what is it called? (Sorry if it's the wrong channel, I have no idea where to ask)

[deep mulch]

@valid jetty rewrite Elle in typescript type system

[lucid trail]

i don’t actually really know

[deep mulch]

@valid jetty make inverse kinematics in elle

would be cool

[icy geode]

Never mind, I figured it out

[deep mulch]

@valid jetty https://youtu.be/0X4TVwS60Qg?si=AB6ACjH9bAatwtNv

yop

[deep mulch]

has anyone here done OCR

or used tesseract

[dense sand]

Is this x86?

[shrewd canopy]

Yes

[deep mulch]

yes

[dense sand]

Osdev so cool

[pseudo sierra]

I have a bit yes but tessie is literally named after tesseract so ask lewi

I think

I really hope I'm not misremembering that cuz that'd be awkward

[deep mulch]

@pseudo sierra bwa

[pseudo sierra]

@deep mulch bwa

[solemn ravine]

FINALLY 😭

[valid jetty]

i think i have the same issue on linux so any tips would be helpful :3

[lucid trail]

i rebuilt raylib from head

[valid jetty]

oh

okie i'll try that

[deep mulch]

how use OCR to read handwritten text in a table

oh

is there something I can use

[shrewd canopy]

Train your ai

[spark tiger]

how is it None if i literally ensure before that there are five args

let mut args = env::args();

[lucid trail]

what is the type of args? is it an iter?

[spark tiger]

it’s Args from std::something

[spark tiger]

^

[lucid trail]

nth might be consuming

yeah Args is an iter

[spark tiger]

oh that might explain why the compiler wanted args to be mutable i guess?

[lucid trail]

yay! here it is

[fleet cedar]

.nth(n) means calling next() n+1 times

Try collecting to a vec first

[spark tiger]

yay it works thanks

now i just need to do error handling... sigh

[hoary sluice]

args is an iterator not an array, you are trying to index an iterator

[fleet cedar]

Which is pretty dumb ngl, because args() always creates a Vec<String> up front and then wraps a vec::IntoIter

To avoid issues if something changes argv while iterating

[crude star]

damn they really do that

[valid jetty]

@lucid trail can you pull and try to build this scene to see if it works

you may need to lower the resolution from src/lib/enums.le, its rendering at 4k rn

but i just wanna see if the 4.4m triangle scene works for you

[lucid trail]

will do

i’m going to watch the reze arc movie first tho

[deep mulch]

@valid jetty hiii

[valid jetty]

hiii

[lucid trail]

it works

[valid jetty]

yay

[solemn ravine]

@woven mesa how likely is it someone would register 1000 app ids

I mean ngl my current account has a few

[deep mulch]

@valid jetty https://blog.shashanktomar.com/posts/strange-attractors

[woven mesa]

unlikely

[solemn ravine]

@woven mesa I dont understand this api it gives me like a trillion lines of json data

[woven mesa]

youtube is worse

[solemn ravine]

cant be as bad as this

this api is gross why is there so much junk I dont need

why are these prepended with its id

[woven mesa]

put all of ur json in a file and check its filesize

[solemn ravine]

@woven mesa I figured out the endpoint for changing capabilites

heres what I would need to send

I would need to send like the entire thing

[woven mesa]

baller

[solemn ravine]

I really hope that the capabilities array is like simple and not overly complicated

[woven mesa]

true

[solemn ravine]

I got another app id with every single thging possible

85 entries

[woven mesa]

insane

[solemn ravine]

88 now

okay it seems like only these are different from the other ones

its the ones with more than one singular option

which is fine but I'm not entirely sure how I would know when to give these

I might just ignore these

@woven mesa bro

why do people do this

[pseudo sierra]

horror

[woven mesa]

horror

[pearl stag]

AppIDViewModel.swift: Line 57

request.httpBody = "{\"data\":{\"relationships\":{\"bundleIdCapabilities\":{\"data\":[{\"relationships\":{\"capability\":{\"data\":{\"id\":\"INCREASED_MEMORY_LIMIT\",\"type\":\"capabilities\"}}},\"type\":\"bundleIdCapabilities\",\"attributes\":{\"settings\":[],\"enabled\":true}}]}},\"id\":\"\(appID.identifier)\",\"attributes\":{\"hasExclusiveManagedCapabilities\":false,\"teamId\":\"\(team.identifier)\",\"bundleType\":\"bundle\",\"identifier\":\"\(appID.bundleIdentifier)\",\"seedId\":\"\(team.identifier)\",\"name\":\"\(appID.name)\"},\"type\":\"bundleIds\"}}".data(using: .utf8)

[solemn ravine]

boooo

[signal oak]

{
  "data": {
    "relationships": {
      "bundleIdCapabilities": {
        "data": [
          {
            "relationships": {
              "capability": {
                "data": {
                  "id": "INCREASED_MEMORY_LIMIT",
                  "type": "capabilities"
                }
              }
            },
            "type": "bundleIdCapabilities",
            "attributes": {
              "settings": [],
              "enabled": true
            }
          }
        ]
      }
    },
    "id": "\(appID.identifier)",
    "attributes": {
      "hasExclusiveManagedCapabilities": false,
      "teamId": "\(team.identifier)",
      "bundleType": "bundle",
      "identifier": "\(appID.bundleIdentifier)",
      "seedId": "\(team.identifier)",
      "name": "\(appID.name)"
    },
    "type": "bundleIds"
  }
}

Message URL

#🪅-progaming message

Question

make this json readable

[solemn ravine]

thank you

why does it give it so much junk data

is that really needed

[worldly sigil]

free junk data to flood your types

[dense sand]

https://www.minecraft.net/en-us/article/removing-obfuscation-in-java-edition

Yay!

[solemn ravine]

this might be goated

[winged mantle]

is it just a skill issue or do circular module dependencies often make things cleaner

go packages are a one sided relationship

go packages are parasocial

[fleet cedar]

Circular module dependencies usually imply they shouldn't be different modules

[winged mantle]

in my case merging all the modules would be insane

they're separate plugins

[granite moat]

I made a small java agent crate for rust

#[java_agent_onload]
fn onload(args: SafeAgentArgs) -> AgentStatus {
    println!("hi agent world!");

    AgentStatus::Success
}

I'm testing out if it's possible to write a modloader in rust

which it should be with jvmti

[winged mantle]

idk maybe i9t's not necessary

but i'm too emotionally attached to my current typescript codebase to rewrite it in go

[granite moat]

wait how does that work

[solemn ravine]

https://docs.rs/readonly/latest/readonly/

[granite moat]

ooo wait

that could be super useful for crates which does unsafe stuff under the hood while still allowing access to fields

[winged mantle]

idk i'm gonna go back to writing C like god indended for now

c is the perfect recreational programming language, it makes things hard in a fun way

[fleet cedar]

A language without generics isn't worth using tbh

[winged mantle]

c just makes generics more fun

[shrewd canopy]

Assembly doesn't have generics
Doesn't mean it shouldn't be used :)

[fleet cedar]

You should most certainly not write programs in asm what are you talking about

[winged mantle]

i mean most of the time you don't need to use it directly

[fleet cedar]

Individual functions, sometimes if thr situation calls for it

[shrewd canopy]

What if my program is for creating virtual machines

[granite moat]

well that makes more sense

for bare metal or extreme performance requirements (small parts) asm can make sense

for example ffmpeg

uses assembly for media encoding/decoding in hot paths

[fleet cedar]

Then some small, delineated parts may need asm

While the majority can be written in a higher level language

[winged mantle]

sometimes impractical things are fun/interesting

life is short, write some asm

[solemn ravine]

@woven mesa all capabilities available on free accounts

[woven mesa]

WHAT

[solemn ravine]

wdym what

[woven mesa]

i misunderstood without context from notification

disregard my overreaction

[solemn ravine]

wtf could this be

never heard of a marzipan before

[fleet cedar]

Isn't that the colorful sugar paste you have on cakes and stuff

[solemn ravine]

"Capability to migrate apps to macOS platform"

yeah lets give it a name of a food

ill need to start work on developing the submitting csr part

then I can get to fun stuff

luckily someone has already made code for it

[royal nymph]

I'm crying how can you be this retarded

I would get if you accidentally open an issue cause you don't know better

but then he was told to disclose it privately and his issue got deleted yet he doubles down on making an issue 😭

[fleet cedar]

Someone hasn't heard of responsible disclosure

[night sphinx]

the kind of person to not ask why there are fences in a barn and just breaks in

[royal nymph]

Also triples down by immediately publishing an exploit poc while the issue has only been fixed for 2 days

(I also read his poc and it's so cringy lmao)

[solemn ravine]

it exists for a reason yes

if you 100% know that optional value will never be null then use it

[fleet cedar]

Never. .expect() is fine sometimes, though

Try camino crate if you don't want to bother with non-utf8 (aka invalid) file paths

I do understand that the stdlib doesn't want to impose restrictions on what systems it supports (such as using utf8), but in this instance I kinda wish it did

[wooden drift]

https://youtube.com/shorts/5aFDj54RWQw?si=phgfxNG9PgvpCSB-

[hoary sluice]

@fleet cedar what type to use for a file received from web frontend

u8 slice?

i dont need content type

[fleet cedar]

Depends on size, streaming, and whether your web framework has anything for it (it probably does)

[valid jetty]

@hoary sluice i added spreading

its not like js spreading, its position in the struct init is irrelevant

it sets default values for the struct rather than override values

it was kinda easy

[hoary sluice]

i use axum multipart for file uploads to the api but the thing that needs the file data is unrelated to the api and i want it to be independent

it only needs the file content

[hoary sluice]

awesome

[lucid trail]

@valid jetty elle -> (patched) raylib -> MGL -> metal

[hoary sluice]

can i use a string if the file can be ascii, ansel or utf8?

[valid jetty]

whaaaaa thats awesome

how does it work

[hoary sluice]

ok ansel is too niche

[lucid trail]

i modify raylib to use glsl 430 with location specifiers for the SPIRV compliation, then build raylib without glfw, link it against libmgl and libglflw when compiling the elle

i dont have to modify elle's raylib at all, but it's still very broken

i spent several days on this

lol

drawing any text breaks everything, also absolutely horrible experience overall

[hoary sluice]

@valid jetty should i write a gedcom parser in elle

are there any projects using elle in prod

can elle interop with rust

only via c ffi right

[deep mulch]

@valid jetty hiii

elle will soon be official github language

[valid jetty]

i made like 3 relatively big projects

[hoary sluice]

raytracer and ichigo

and what else

[valid jetty]

pomee

my rhythm game

[hoary sluice]

but none of these are prod

[valid jetty]

well yeah but its not a prod language

im nowhere near 1.0

backwards incompatible changes happen almost every commit

[hoary sluice]

i can fix compiler version

if i were to write the parser in elle i would just connect them with a rest api right

[valid jetty]

theres no rest fetching yet in the stdlib lol

just tcp sockets which you can use for http

[hoary sluice]

altho a gedcom parser crate would probably be more useful

theres 3 rn, one of them defines the datastructures for gedcom x which is different, one only supports the 1996 version and one is a fork of the old parser that has a "goal" to parse gedcom 7.0 but doesnt yet

[valid jetty]

why do you need a gedcom parser anyway

[hoary sluice]

im making a family tree viewer

there are no good ones out there

theres like 50 options and all of them are garbage

or bloated

[fleet cedar]

Will yours be not garbage?

[hoary sluice]

yes

its not that hard

its like an aoc day 24 basically

the tree part

all the versions ive seen are either

1. extremely bloated, require signup, 50$ a month, very laggy (ancestry and myheritage)
2. simple gedcom to png converters
3. have 1980 ui
4. are (sometimes paid) windows apps

4 usually includes 3

[valid jetty]

eagely

in my raytracer for traversing my bounding volume hierarchy i had to use a dfs

[deep mulch]

elle discord client when @valid jetty

needs to compete with vencord

[valid jetty]

first time in like 4 months that i used a traversal algorithm that isnt for aoc or school

[hoary sluice]

reference!!

[hoary sluice]

me too

in the tree viewer

i had to do a bfs

to determine the generation

of a person

[hoary sluice]

you did bfs in school??

[valid jetty]

yeah like, recently ish

we also did dijkstra but not A*

[hoary sluice]

i think we had 1 lesson of algos in school

80 page pdf on graphs in uni

[valid jetty]

we were tought all abstract data structures too

like stacks, queues, hashmaps, etc

but those are easy to use anyway

[hoary sluice]

each of these had 1 powerpoint slide and 5 seconds dedicated to them

[valid jetty]

we were tought ""trees"" but they didnt really go that in depth

like they didnt mention red black trees or avl trees

or really anything more advanced than base level knowledge

[hoary sluice]

i might just make my own format instead

and use the gedcom 5.5.1 parser from cargo

to convert

[deep mulch]

djistraka @valid jetty

[valid jetty]

i'm curious, since i don't have xcode i can't build opengl newer than 4.1

could you try, with raylib and opengl 4.3, to change the code in the raytracer to use an ssbo instead of a texture buffer and record the performance improvement

you seem to understand the codebase so i think you'd know how to do it

if possible, split the triangles and bvh into 2 separate ssbos

that would also make the full 7.2m triangle viable to render since you don't need to push the triangles and bvh into the same buffer

[deep mulch]

that's such an insane limitation

[solemn ravine]

is that actually true

I dont think it is

xcode is not required for anything

[deep mulch]

oh

rosie crazy

[solemn ravine]

just saying, xcode just comes with a bunch of tools used for development, is it necessary to install? no. is it convenient? yes.

but also opengl on mac is crazy

[valid jetty]

i will try build without xcode

i do have clt

[lucid trail]

i would, theres much to fix with mgl/raylib first tho

also zed now has a settings menu!

[valid jetty]

yep i saw

tbh i prefer the json format

[lucid trail]

yeah if i'm trying to find an option though

[valid jetty]

am i doing something wrong building MGL?

theres no BUILD.md so i dont know how im supposed to build it

[lucid trail]

are you looking at my repo?

[valid jetty]

i did

sh ./clone_external_sources
sh ./build_external_sources

then i commented out the =NULL from the thing MGL added and built everything
but i cant get the main MGL to build, various things are missing and i dont know why

like the python script

[lucid trail]

Ah that

one sec

[valid jetty]

non existent

[lucid trail]

yeah the one i'm working on is only using cmake now

try https://github.com/tnixc/mgl

[valid jetty]

lmao the models are making it take so long to clone

[lucid trail]

i also have raylib in there

[valid jetty]

husk

[lucid trail]

i think the build for raylib is messed up rn

[jade stone]

paige so insane

[deep mulch]

collect my paiges

[pseudo sierra]

paige so sane

[valid jetty]

ok what am i missing

i cloned with --recursive but the spirv stuff isnt here

blehhhh

blehhhhhhh

[deep mulch]

love

[jade stone]

nop

esbuild is sooo good

[lucid trail]

yeah i've ran into this too

uhhh how did i fix it... hmm

[valid jetty]

ok i just cloned them manually

now to fix the Findglslang.cmake

[jade stone]

@pseudo sierra how

esbuild slander

[lucid trail]

i dont have SPIRV-Cross here, idk how this is even actually building

[pseudo sierra]

[lucid trail]

ah it was cached !!!! i was just using the shit in my build folder

[deep mulch]

thought this was a typo

Findglefjksjgkejf

[lucid trail]

i had spriv-cross at some point and then removed it

[valid jetty]

well its weird because its in .gitmodules

[lucid trail]

yeah

i know

[valid jetty]

just for some reason git doesnt clone it when i clone the repo

not sure why

[lucid trail]

subprojects is in gitignore

oh i know why

in the original it would've been something the get_external got

fixed, pushed

i wonder how cmake never cared that spirv was not actually checked out

[valid jetty]

ok well that somewhat worked but i still get this

[lucid trail]

uhh try brew install glslang

[valid jetty]

at one point i should like do this

[lucid trail]

no hell nah

it's horrid

[valid jetty]

hm

does 4.6 work correctly with text and stuff?

[lucid trail]

wdym by 4.6

[valid jetty]

opengl 4.6

[lucid trail]

what has opengl to do with it

[valid jetty]

nono its irrelevant

im just asking

[lucid trail]

i mean the raylib text drawing is not working but that's because of some issue in MGL

i think

[valid jetty]

i meant does 4.6 work correctly with text in raylib

ah

uhbbb ```sh
rlgl.h:887:14: fatal error: 'GL/glcorearb.h' file not found
#include <GL/glcorearb.h>
^~~~~~~~~~~~~~~~

[lucid trail]

i'm also running into this

i'm trying to find a commit where this worked

it really makes no sense to me

why the hell is trying to build the examples if i'm in /src

[valid jetty]

i think i got it to build

nvm

[lucid trail]

wow running make changes the makefile

this is actually cursed

@valid jetty i got it

just go into raylib and run cmake --build build, should work

[valid jetty]

i went through a bigger thing than that

[lucid trail]

pull, i fixed this by making raylib build without glfw

[valid jetty]

i downloaded rlcorearb.h locally and moved some includes around and it worked

[valid jetty]

oh ok

[lucid trail]

yeah i also moved raylib into examples

[valid jetty]

i think you forgot something

raylib is gone

[lucid trail]

oh yeah that's really strange

oops

[valid jetty]

lmao

[lucid trail]

i see what i did

ok wow

ok sure just revert that

really, i should just remove build from gitignore

@valid jetty does this work

1. cmake -B build -S . -DUSE_EXTERNAL_GLFW=ON -DBUILD_SHARED_LIBS=OFF
2. cmake --build build --target raylib

[valid jetty]

is this the correct result

[lucid trail]

yes

amazing

[valid jetty]

ok cool

[lucid trail]

i dont quite understand why drawing text is cooking it, since all its doing should be DrawTexturePro, but we are already doing that

[little bluff]

i want a plugin which appends text at the end of messages but i know its probably not made, but i need a starting point since im a absoulute noob, is there a plugin which modifies the sent text after SENT but before its sent to the servers? that way i can just modify it

[fleet cedar]

vp textreplace

[little bluff]

thank you, goat

[coral mountain]

absolutely not

i downgraded to sequoia

new metal sucks ass

you can't even have 10 apps open with it

[little bluff]

thanks to this goated plugin, i can now be takopi-pi.

[fleet cedar]

Just keep in mind that homestuck typing quirks are mega cringe

[crude star]

are you really takopi if you can't say pi after every sentence yourself

[little bluff]

takopi doesn't say himself either-pi, its just his habit-pi

-pi:P

[crude star]

😭

[little bluff]

oh no back to coding 🥀 its affecting emoticons-pi

fml-pi

:P

yayyy-pi

[royal nymph]

you have psychosis

[winged mantle]

has anyone made a typing quirk plugin

[mellow plinth]

testing a plugin I made where it makes the avatars and server icons square, because I like it looking like that.

[winged mantle]

can't you justu se css

[gilded surge]

out = out.replaceAll(/\`.(.+?)`/g,"$1").split("\n").flatMap((entry) => entry.split("     | ")).filter((entry) => ["rv","wv","wl","st"].some((register) => entry.includes(register))).filter((entry) => !entry.includes("⌽")).map((entry) => entry.replaceAll(" ","")).map((entry) => entry.split("┃┃").map((instr) => instr.replaceAll("┃","")));
    const chunk = (array, chunk_size) => Array(Math.ceil(array.length / chunk_size)).fill().map((_, index) => index * chunk_size).map(begin => array.slice(begin, begin + chunk_size));
    let left = [];
    let right = [];
    out.forEach((pair) => { left.push(pair[0]); right.push(pair[1]); })
    let leftChunks = chunk(left, 4)
    let rightChunks = chunk(right, 4)
    let instructions = (leftChunks.length >= rightChunks.length ? leftChunks : rightChunks).flatMap((_,i) => [leftChunks[i],rightChunks[i]]).flatMap(_ => _).filter((entry) => entry != null).map((entry) => entry.split(",")).map((entry) => [[entry[0].slice(0,3),entry[0].slice(3)],entry[1]]).map((entry) => entry.filter((instr) => instr != null)).map((entry) => typeof entry[1] !== "string" ? entry[0] : entry[0].concat(entry[1]))

javascript, truly the most beautiful language

you know it's working when you have more code than empty space

[shrewd canopy]

wtf is this

is this some pseudo asm

[gilded surge]

basically yeah

it's for a puzzle

[shrewd canopy]

drop

[gilded surge]

you would have to buy a game

[shrewd canopy]

why

[gilded surge]

cause the puzzle is in the game

you would also be forced to use javascript all of the time

[shrewd canopy]

is the game about reverse engineering basically?

[gilded surge]

nah

this is just a huge ass puzzle somebody dropped

actually yeah the game is kinda about reverse engineering

the games tagline is literally "before you is a system about which you know nothing"

[shrewd canopy]

interesting

[gilded surge]

it's a MUD, if you've ever heard of those

[valid jetty]

@solid gazelle using elle as a scripting language

#!/usr/bin/env -S ellec --silent --run
use std/prelude;

namespace DIR;

enum DirentType @repr(u8) {
    Dir = 0b00000100,
    Reg = 0b00001000,
}

struct Dirent {
    u64        ino,
    u64        seekoff,
    u16        reclen,
    u16        namlen,
    DirentType type,
    char       name // static array if you take the addr at this field :)
}

external fn opendir(string path) @alias(DIR::open) -> DIR *;
external fn readdir(DIR *dir) @alias(DIR::read) -> Dirent *;
external fn closedir(DIR *dir) @alias(DIR::close);

fn collect_files(string path, string[] out) {
    dir := DIR::open(path) ?: $die();
    defer dir.close();

    while dirent := dir.read() {
        if [".", ".."].contains(&dirent.name) { continue; }
        out_path := "{}/{}".format(path, &dirent.name);

        if dirent.type == DirentType::Dir {
            collect_files(out_path, out);
        } else if dirent.type == DirentType::Reg {
            out.push(out_path);
        }
    }
}

fn generate_sources() {
    path := "src";
    $printf("Generating files at {}...", path.color("green").reset());
    collect_files(path, files := []string);
    io::write_to_file("docs/srcs.txt", files.join("\n"));
    $printf(
        "Generated {} files at {}!",
        "{}".format(files.len()).color("green").reset(),
        path.color("green").reset()
    );
}

fn compile_docs() {
    path := "docs/main.typ";
    $printf("Compiling typst at {}...", path.color("green").reset());
    libc::system("typst c {} --root ..".format(path));
    $printf("Compiled {}!", path.color("green").reset());
}

fn main() {
    generate_sources();
    compile_docs();
}

[solid gazelle]

@valid jetty Elle package manager

[fleet cedar]

Why is elle named elle?

Just because it's cute?

[valid jetty]

yeah basically

and this is mia https://github.com/acquitelol/elle/raw/rewrite/images/mia_512x.png?raw=true

[fleet cedar]

Who is she?

[solid gazelle]

what's her pronouns

[valid jetty]

a butterfly named after one of my friends

[deep mulch]

1 eye

@valid jetty what if i use ai to make it hd

[valid jetty]

ITS SIMPLE BY DESIGN

[fleet cedar]

What of I use ai to make you hd

[fleet cedar]

What of I use ai to make you hd

I only posted that once

[deep mulch]

OOP when

[valid jetty]

never

[deep mulch]

soonn

elle will be java compatible

fully supported interop

@valid jetty wyd

[valid jetty]

inheritance is horrible, polymorphism already exists in some form, encapsulation treats the developer like a baby (i dont trust that you can use my AdderFactorySolutionProviderFactoryStrategy correctly so im gonna spoonfeed you)

[deep mulch]

@valid jetty inheritance is perfection

use

[valid jetty]

the polymorphism

[pseudo sierra]

zoot master rage baiter

[valid jetty]

zoot im not mad just dissapointed

[deep mulch]

rosie so mad

[ornate quiver]

this is so cute

[deep mulch]

i need to make rosie rage

[valid jetty]

https://tenor.com/view/cat-gif-15732275568813898827

[deep mulch]

i feed minky butterflies @valid jetty

[valid jetty]

so mean

[valid jetty]

https://fxtwitter.com/jarksob_/status/1984500204729745641

jacksob (@jarksob_)

went to the .. the chippy .. and opened up for a real yumtastic day .. and theres a .. there .. fuck .. my god .. it's moving .. oh my god .. chip .. chip cat ?? a cat ?? in my chips ??

**💬 44 🔁 1.5K ❤️ 10.7K 👁️ 115.9K **

FxTwitter · Nov 1, 2025, 5:58 AM

[frosty obsidian]

oh this is programming

[deep mulch]

wing moment

[valid jetty]

https://fxtwitter.com/royaltybnb/status/1984855972175597795

royalty ™ (@royaltybnb)

I just bought this meme

**❤️ 1 👁️ 177 **

FxTwitter · Nov 2, 2025, 5:31 AM

[frosty obsidian]

uhh pretend thats a java spring reference

[deep mulch]

rosie is stalking wing

[crude star]

❤️ 1

[frosty obsidian]

actually i got the image before that was tweeted

[deep mulch]

@valid jetty remove all functions from elle

add only classes

[valid jetty]

it can basically be functional if i implement closures

[frosty obsidian]

make elle both functional and object oriented

[valid jetty]

@lyric latch love?

[frosty obsidian]

wrong

constructors only

[pseudo sierra]

do

[fleet cedar]

Constructors only, no classes

[deep mulch]

make ellescript

[valid jetty]

https://github.com/acquitelol/ellecord

[deep mulch]

rewrite elle in C

[fleet cedar]

Rewrite c in elle

[hoary sluice]

@valid jetty rewrite elle in icypeas

[valid jetty]

see now that is a great idea

[deep mulch]

@valid jetty make elle 2

[nimble bone]

implement https://music.apple.com/ca/album/closure/1547315522?i=1547316002

[fleet cedar]

Have you done Bad Apple in elle yet

[valid jetty]

no lol

can't be that hard can it

get an mp4 of bad apple in 960x540, ffmpeg it to a bin, read it in elle, display a raylib window and swap to a new frame in the bin every frame

then get a bad apple audio and make it also play in sync with the bin data

you can just read it in chunks of 960*540*#size(i32) and directly set that as the color array in the image to display via raylib

[winged mantle]

recreate the actual original stage

[deep mulch]

how is rosie so smart

[valid jetty]

i was basically right

[solid gazelle]

ELLE JVM

[winged mantle]

I removed something from a file and github copilot suggests the commit message as adding the thing

[ivory heath]

Raw rgb my beloved

Could actually do gray for bad apple

[valid jetty]

yep i did

oh wrong reply

yep i did

surprisingly simple

this is with full resolution

it still runs at 30fps (the framerate of bad apple) but i made it 10% res in the code afterwards just so the bin isnt so big when i commit

even with 10% the bin is 42mb

i may just gitignore the bin and mp3 since the video produces those when you run the program

[valid jetty]

i cleaned it up a bit

but yeah its not that hard

[deep mulch]

rosie the worlds smartest person

[valid jetty]

this is like not that complicated

[deep mulch]

wrong

[valid jetty]

the only downside of this approach is that there is a 1gb temp file during playback

[shrewd canopy]

horror

optimize when

[valid jetty]

optimize how

thats just how much raw data there is in 200 seconds of footage at 480x360 at 30fps at 1 byte per color

[deep mulch]

optimize

[pseudo sierra]

shrimply don't use raw data

[valid jetty]

raylib cant really load anything other than raw data it can but videos arent comprised of images per se

its not like i can output a bunch of pngs and render them in sequence

like i could do that but it wouldnt work very well

[deep mulch]

rosie has never heard of frames

[valid jetty]

if anything i would just use libavcodec and at that point i just have a shitter mpv

[deep mulch]

@valid jetty

[little bluff]

no-pi, she's just my cutie-pi :)

[little bluff]

its okay if its cats ❤️‍🩹-pi

meow

[little bluff]

my text appender is appending after emojis 🥀 back to coding-pi.-pi.-pi.-pi.

[little bluff]

hooorayy-pi!!!! it works now-pi✊

[fleet cedar]

I hope you know this is cringe af

[royal nymph]

@deep mulch Rate the setup

[royal nymph]

thanks i hate it

[deep mulch]

insanity

[cinder egret]

so cute-pi

[deep mulch]

@valid jetty explain what impedenance is

[valid jetty]

?????

[pseudo sierra]

I think zoot might just be stupid

[deep mulch]

rosie doesnt know

i am the smart one

@valid jetty explain reactance

explain how to read a smith chart

explain SWR

[valid jetty]

im not doing anything to do with electricity

[deep mulch]

explain dbm vs dbfs

@valid jetty explain how a yagi uda antenna works

explain how a dipole works

explain how an endfed halfwave works

explain everything NOW

[valid jetty]

two equal but opposite electric charges are spaced apart

in an electric field the positive side feels like an attractive force while the negative side feels like a repelling force

[valid jetty]

idk

[valid jetty]

idk

[pseudo sierra]

@deep mulch explain how to talk to a woman

[deep mulch]

nop

@blazing quartz knows

[valid jetty]

different reference points to measure signal levels

[valid jetty]

idk

[blazing quartz]

hello there gorgeous, how are you today @royal nymph

[valid jetty]

zoot you forget im doing computer science not electronics

why the heck would i know any of this

[deep mulch]

cause youre a genius

[valid jetty]

its equivalent to asking a marine biologist on historical events from the 1200s

[pseudo sierra]

zootie a bit tupid but that's okay we love him anyway

[deep mulch]

@valid jetty calculate the wire length needed for a resonant 40m antenna

display: none;

[pseudo sierra]

display: zoot;

[deep mulch]

bwa

[frosty obsidian]

you ask in a discord server

[winged mantle]

/run

#include <stdio.h>
#include <stdarg.h>

void putchars(int ch, ...) {
    va_list list;
    va_start(list, ch);

    if (ch == '\0')
        return;

    putchar(ch);

    for (;;) {
        ch = va_arg(list, int);

        if (ch == '\0')
            break;

        putchar(ch);
    }

    va_end(list);
}

int main() {
    putchars('H', 'e', 'l', 'l', 'o', ' ', 'w', 'o', 'r', 'l', 'd', '\n', '\0');
}

[rugged berry]

Here is your c(10.2.0) output @winged mantle

Hello world

[winged mantle]

for some reason i wrote this instead of the normal hello world to test build setup

[winged mantle]

the only good way to develop a website: backend in C with mongoose and sqlite

wasm based frontend using sdl with clay

[valid jetty]

..why

[winged mantle]

good

what if c compilers supported
#pragma using prefix_

and that removed the prefix on all symbols

[winged mantle]

*i meant to say sqlite, not mysql

don't husk

[deep mulch]

rehusked

[little bluff]

cringe is fun-pi

[little bluff]

really-pi? its my favorite one (took alot of time-pi)

[little bluff]

this man knows-pi!!️❤️

[deep mulch]

@valid jetty

[crude star]

i don't think you can fix the reddit mascot being so ugly

[little bluff]

it looks like a cute mascot to me-pi

[winged mantle]

everyone loves sqlite

nobody likes mysql

[pseudo sierra]

true

[dense sand]

anyone here familiar with bun workspaces? no matter what i try i get this

hmm weird

still happening, the current setup is like this:

• Root(dependencies: @repo/portalbush)
  • @repo/portalbush
  • @repo/portalfish (dependencies: @repo/portalbush)

i want portalfish to depend on portalbush

should be working right?

smh bun is not bunning

[jade stone]

Something something just use pnpm

[dense sand]

Bun stopped working so im rewriting the entire project

[deep mulch]

@jade stone

[jade stone]

I have somewhat mixed feelings on pnpm, but they're overall positive

[deep mulch]

stinky sadan

[jade stone]

(plus at least you can just edit and debug the code if something doesn't work)

[nimble bone]

i await bun discontinuation

await Bun.discontinue();

[deep mulch]

Bun will have first party grok integration

ai in stdlib

[valid jetty]

what do i make

• never (!) type in the type system
• static arrays in the type system
• elle intrinsics to #env like PLATFORM, VERSION, TARGET, ARCH, accessible at runtime, optionally disableable at compile-time (--nointrinsics)

these are the features that would be most interesting to make i think

[royal nymph]

all three

[valid jetty]

well yeah but which one do i make right now

[royal nymph]

all three

[valid jetty]

i feel like never type would be quite easy to make maybe

[royal nymph]

never type > static arrays > the other option imo

[lucid trail]

are there tuples of any length rn

[valid jetty]

nop

[royal nymph]

never type should be easy and is very useful

[valid jetty]

yeah but so are static arrays in the type system

useful for any real work with C libraries

[valid jetty]

i dont wanna make this rn because it sounds like a lot of work

@royal nymph

[deep mulch]

jvm support

[lucid trail]

my vote is for static arrays

[deep mulch]

inheritance

[deep mulch]

elle build script

[valid jetty]

Makefile

[deep mulch]

no semicolons

[deep mulch]

do something like cargo build.rs

oh rosie my Rosie

[valid jetty]

look

[deep mulch]

eh

[valid jetty]

you can make to build or make run to also run

and it comes built in with incremental building and stuff (well i guess if you make the raytracer recipe depend on all source files)

[deep mulch]

id only use make for like C and C++

[valid jetty]

its great

[deep mulch]

does Elle have external dependencies yet

[valid jetty]

nop

[deep mulch]

dependencies would be a dedicated build system task

[royal nymph]

only if it's enforced that a never return type actually never returns

and even then idk

see how typescript handles this and replicate

typescript never is pretty sane

[deep mulch]

is that like C++ noreturn

[valid jetty]

well i want to use the never type to run defers

like i want to call all deferrers before calling a function that returns never

[deep mulch]

add attributes to Elle @valid jetty

[valid jetty]

so yes, it has to be enforced that the never type actually means "this never returns"

[deep mulch]

add proc macros

add borrow checker

[deep mulch]

roinga

add Elle inline assembly

[valid jetty]

zoot is just spewing words

[deep mulch]

does qbe allow that

[valid jetty]

time for 50000 refactors

[lucid trail]

it was a feature that got removed

[deep mulch]

@valid jetty make compiler auto unbox

[valid jetty]

what was a feature, inline assembly?

[lucid trail]

oh that was IR

[valid jetty]

i had inline IR at one point yes

well

its still here

[lucid trail]

just extern C for inline asm

[deep mulch]

elle will soon be ported to java

[valid jetty]

or.. not

[royal nymph]

feat: get rid of rosie

[lucid trail]

add vectorization

[deep mulch]

add avx512

https://github.com/st-tech/ppf-contact-solver @valid jetty

sooo cool

[valid jetty]

i saw this in a 2 minute papers video

[deep mulch]

good

me too

roie probably jotts down notes watching technical videos

@valid jetty you write like 5 pages of notes

Jotting this down

[valid jetty]

@lucid trail should i allow contravariance for static array types?

[lucid trail]

what is contravariance

[valid jetty]

as in, should i32[41] count as a subtype of i32[40]

[deep mulch]

that's weird I think

[valid jetty]

or should the sizes need to match exactly for it to work

[lucid trail]

by subtype you mean it can be "promoted" to i32[40]?

[valid jetty]

yes

[lucid trail]

i feel like that could cause unintuitive behavior

[deep mulch]

for array sizes that seems kinda confusing

yea

[valid jetty]

fn foo(fn(i32[4]) cb) {
    cb(#[1, 2, 3, 4, 5]);
}

[deep mulch]

does Elle promote integer types @valid jetty

[valid jetty]

ok so just when theyre exactly equal, got it

[valid jetty]

yes

[lucid trail]

what is xs here

[valid jetty]

also, should pointers be equal to static arrays as long as they have the same underlying type

[valid jetty]

a function which acccepts a static array of 4 i32s

oh wait

ignore that lol

[deep mulch]

uhh

explain

[valid jetty]

a static array is just a pointer internally

if i have an i32* and i pass it where a i32[4] is expected should type checking pass

[lucid trail]

where is the length held though

[valid jetty]

its not

[lucid trail]

it's gone yeah

[valid jetty]

youre just giving responsibility to the developer that their buffer at that pointer is big enough

[lucid trail]

maybe i'm thinking too much about safety

[deep mulch]

probably should require explicitly casting

[valid jetty]

i can do that

[shrewd canopy]

Old CPU support dropped

[deep mulch]

Elle doesn't support x86_64 so wouldn't lose much

right

@valid jetty

I can't remember

my mind is frayed

[valid jetty]

this is so cursed

[valid jetty]

??? yes it does

[deep mulch]

I thought it didnt

[valid jetty]

x86_64 is 64 bit

[deep mulch]

I know that

[deep mulch]

x86_64 should be default

[valid jetty]

??? no its not just set to macos and aarch64

[deep mulch]

nvm the host architecture should be default

[valid jetty]

its based on the os youre running the compiler on

yes thats what happens

[deep mulch]

"Defaults to host architecture" would be a better message

[deep mulch]

maybe it was windows support I was thinking of

[sharp girder]

True

[deep mulch]

evil order

what's () => { in rust

[valid jetty]

thats just a macro body

[deep mulch]

@valid jetty @valid jetty

[shrewd canopy]

Nah thats husk

[valid jetty]

it does

zoot is just spreading misinformation

[deep mulch]

hush

I thought it didnt

hush now rooe

rooe

@valid jetty Elle needs reflection

make Elle faster

[lucid trail]

cs is so boring and tedious

so good

[valid jetty]

hmm well

@lucid trail this is an interesting predicament

the size of a buffer may be a runtime value

[lucid trail]

okay

[valid jetty]

so therefore the ergonomics of this kind of code arent nice

fn foo(i32[4] xs) {
    ...
}

fn main() {
    i32 xs[4]; // you cant type this as a static array because the size is a runtime value, so it must be a pointer
    foo(xs); // either you implicitly cast the pointer to the static array type, giving the bounds access responsiblity to the developer
    foo(#cast(i32[4], xs)); // or you have to cast, which is kinda not ergonomic to use and annoying
}

[lucid trail]

i dont really see the issue here

[valid jetty]

which one is better

[lucid trail]

why is the size a runtime value? like the 4 will be unknown at compile time?

foo(xs);

[valid jetty]

you can put any value there

use std/rand;

fn main() {
    i32 xs[rand::random(0, 100)];
}

thats why it cant be typed as a static array

[lucid trail]

oh, i was under the impression you couldnt do that like in C

[valid jetty]

you can very much do that in C

in C it also decays to a pointer afaik

[lucid trail]

yeah that's pretty bad actually

[deep mulch]

NOOO

NOT THAT SYNTAX

elle has fallen

[valid jetty]

?

[deep mulch]

array size with the name

[valid jetty]

that syntax has literally existed for over a year

[deep mulch]

should be with the type

[valid jetty]

well i guess you can also do that now

but for that case the size must be an integer literal

[deep mulch]

hm

[valid jetty]

im adding static arrays to the type system

which means this code will be valid

fn main() {
    i32[40] xs;
}

but the 40 cannot be a runtime value anymore

i guess i can remove the buffer syntax then? but its useful in some cases

if i remove that syntax then the only way to allocate a runtime-value-based static buffer is to allocate it on the heap and i would rather not lock the developer to do that if i can help it

[lucid trail]

should you not err on passing an array of unknown size to a function that requires a specific size

[valid jetty]

sigh

should this be convertible implicitly

[lucid trail]

add .into() already

[valid jetty]

traits dont exist, that wouldnt work very well

ok well

[lucid trail]

surely you can implicitly convert char[2] to string

[valid jetty]

well

[valid jetty]

well how is this any different from allowing converting any static array type to any pointer of the same type

string is just a unique char* type

[lucid trail]

can you not do this

its just converting the other way is unsafe

[valid jetty]

oh true actually i think i actually allow that case

ok well

this is so cursed

[lucid trail]

?? is this for casting

[winged mantle]

it's actually logical

it just doesn't scale very well when you have an array of function pointers

the declaration reflects the usage

int ints[5];
int i = ints[0];

[fleet cedar]

So you say, but yet ints[5] is ub

[valid jetty]

yes

[winged mantle]

😭

[valid jetty]

the behavior of #cast is just a big long function at compile time that defines what can cast to what

[winged mantle]

i don't think this has ever been a problem, the real issue is that it gets confusing when it's more complex

[valid jetty]

in fact it's quite awful https://github.com/acquitelol/elle/blob/rewrite/src/compiler/lib/convert.rs

but it works

[winged mantle]

als o in c++ it's just a mess

[deep mulch]

as when

[winged mantle]

though it's already is strange with stuff like void

void *myPtr;
void v = *myPtr;

[valid jetty]

it won't let you deref that

[winged mantle]

that's the point

[valid jetty]

why is that strange

[lucid trail]

better than a match fr

[winged mantle]

the declaration doesn't reflect the usage in the same way

[deep mulch]

roinga

[fleet cedar]

Declaration matches usage is and has always been nonsense

[winged mantle]

that doesn't stop it from being sort of logical some of the time

[fleet cedar]

int * a declares a variable a of type int*, not a variable *a of type int

[valid jetty]

well no but semantically void has no size, therefore dereferencing is just nonsense because in assembly the type you're dereferencing refers to the size of the register to load from

like you can't load a buffer of (*≧∀≦*) sized stack memory

[fleet cedar]

It's not that it has no size, it's that it's not a type

[valid jetty]

well it very much is a type, it's the unit type

[fleet cedar]

Not in c++

[valid jetty]

hm

[deep mulch]

I prefer the pointer as *type name = value

[winged mantle]

AWFUL

[valid jetty]

nono this is good

much easier to parse in fact

[winged mantle]

it's still ambigious

what if it's unary mutliplication

[valid jetty]

postfix pointers are a nightmare to parse

what the fuck is a unary multiplication

[fleet cedar]

Type after is objectively superior in a world with type inference

[winged mantle]

multiplication equivilent of virgin birth

ok nah i made that up

[fleet cedar]

Why are you parsing a multiplication in my type

[valid jetty]

lol true

ok well i think i'm done implementing the static arrays

i just need to test all cases

there are a lot of edge cases where this could go wrong

[deep mulch]

@valid jetty you love C++ callback type syntax

[winged mantle]

what even is that

[deep mulch]

typedef void (*cb)(int);

[winged mantle]

that's function pointer

🙄

is that what you kotlin devs call it

[fleet cedar]

I prefer typedef void cb(int); and then using cb*

[valid jetty]

• static array of structs, setting a struct's field
• struct with a static array, taking the address of an index in that array in the struct
• multidimensional static arrays
• dynamic array of static arrays
• static array of dynamic arrays

etc.

[deep mulch]

pointer to a callback then

[valid jetty]

pointer to any function

and anyway this is the C syntax

[deep mulch]

guh does it really matter I was just pointing out the syntax

yea

[valid jetty]

c++ has int(int, int)

iirc

and std::Function<int, int, int> for closures

something like that i'm not a c++ dev

[deep mulch]

roieee😭

[winged mantle]

til this works...

[fleet cedar]

No, std::function<int(int, int)> for dyn dispatched functions

[deep mulch]

yop

[valid jetty]

close enough

oh i'm thinking of the c# syntax lmao

[deep mulch]

@valid jetty rewrite Elle in C#

[valid jetty]

i guess this is what happens when you use 500 languages

[fleet cedar]

int(int, int) is kinda like rust's [T] in that you can't have values of it directly, only behind pointers

[valid jetty]

oh hm

idk in elle you can only have function pointers

fn(i32, i32) -> i32

dyn dispatch is a myth

[deep mulch]

@valid jetty can you make alias to a function pointer in elle

[valid jetty]

no not really

[deep mulch]

do

[dense sand]

@lyric latch any idea what could be a good way of getting an npm dep in docker from private registry

[valid jetty]

one at a time

[jade stone]

i swear to god i am going to lobotomize typescript-eslint

[valid jetty]

WHAT

[maiden iris]

i LOVE eslint but i hate it actually

[valid jetty]

@lucid trail should this be allowed

[lucid trail]

i think it should be explicit and fail on the length not being 16

[valid jetty]

hmm ok

[valid jetty]

seems like its working

[lucid trail]

is there anything like i32[2].join(i32[2]) => i32[4]

that would be cool

[valid jetty]

you cant set the size in generics, otherwise yes

fn __arr__::join<T, A, B>(T[A] self, T[B] other) -> T[A + B] {
    T[A + B] res;
    
    for i in 0..A + B {
        if i < A {
            res[i] = self[i];
        } else {
            res[i] = other[i - A];
        }
    }

    return res;
}
``` something like that

this has a lottttt of holes however

this is probably about 10 refactors away from being real syntax

[lucid trail]

T[A][B].transpose() -> T[B][A]

[nimble bone]

good idea

node should add window.ai unironically

[valid jetty]

i just dont know how i would make the static array bounds generic

hmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm

i have.. an idea

however whether it works uhhh..

[lucid trail]

interesting

[runic sundial]

okok

images

html

css

do you like

always nest them in divs?

NECOBOB

[jade stone]

My website is >50% divs

[runic sundial]

teach

[jade stone]

https://react.dev

[runic sundial]

ok

but what if im not javascript

[jade stone]

Idk

[valid jetty]

@lucid trail

[lucid trail]

woaw how did you do it

[valid jetty]

i made the size an internal type that just holds a number and then monomorphized it like any other type

this type is not representable with code but it can be monomorphized

[deep mulch]

@valid jetty hii

roie

[vapid prairie]

if want_to_make_good_code then
  make_good_code()
end

[valid jetty]

this is sooooo silly thank you for this idea

[lucid trail]

YAY

[deep mulch]

roie

[lucid trail]

could you do T[A][B]...[F].rank() => 6 (has no reason to exist but would be cool)

although I would assume no variadic generics yet

that would simply be insane

[valid jetty]

if there were conditionals based on type conditions, that wouldnt be too hard

also i feel like this is the wrong way round?

well no actually

you access it correctly by a[row][col] but the type is defined as T[cols][rows]

[lucid trail]

could you do specific values like restricting .cross_product() to 3 and 7

[lucid trail]

isn't it actually ambiguous

[valid jetty]

you could do ```rs
fn foo<A, B>(i32[A][3], i32[7][B])

[valid jetty]

hm wdym

[lucid trail]

oh wait that was a stupid question

[lucid trail]

it's not clear if its 4x2 or 2x4 when i look at it

[valid jetty]

well i think most languages would type it as T[4][2] (row major) which is a problem because i type it as column major

[lucid trail]

hm

[valid jetty]

im about to make it row major

[lucid trail]

does T[A][i] ensure i<A, array indexing

[valid jetty]

why would it need to

[lucid trail]

kind of like a tuple

[valid jetty]

ok there we go

row major

[lucid trail]

instead of asserting a.len() == b[0].len() couldn't you just do T[B][C] b

[valid jetty]

true!

[valid jetty]

yep it works

fn __arr__::mul<T, A, B, C>(T[A][B] a, T[B][C] b) {
    T[A][C] res;

    for i in 0..a.len() {
        for j in 0..b[0].len() {
            sum := 0;

            for k in 0..a[0].len() {
                sum += a[i][k] * b[k][j];
            }

            res[i][j] = sum;
        }
    }

    return res;
}

hehe~2

[lucid trail]

that’s sick

[deep mulch]

smh crab couldve been Rust

[deep mulch]

oh okay

[frosty obsidian]

got tile revealing working

[deep mulch]

sleep

[hoary sluice]

@valid jetty im doing aoc in rust this year

[hoary sluice]

is this minesweeper or gambling?

[valid jetty]

i've decided

i may try it in elle

but i won't be competing for the fastest time i'll just do it whenever i get to it

i need my beauty sleep i can't sleep for 3h a night every night and wake up at 4:45am the whole december again

or well until the 12th

[fleet cedar]

I'll probably go with golf python again

[hoary sluice]

i will be in Kazakhstan starting day 7

so it unlocks at 10 am 😄

[hoary sluice]

you know you can just go to sleep earlier

[nimble bone]

im doing AOC in JS this year

[hoary sluice]

does this mean i wont have anyone to compete with

[nimble bone]

TypeScript

or i'll redo the one lang a day thing

i think Satan did it in ts types

so insane

[hoary sluice]

@valid jetty should i test raytracer

on 9070xt

[shrewd canopy]

i wanna to do it in c++ templates

[hoary sluice]

ur elle raytracer

or is it on the cpu

[valid jetty]

gpu

[valid jetty]

yeah

[deep mulch]

add 2 hours delay

so I have a chance

[crude star]

I will do aoc in uiua

while also teaching a friend how to code

[valid jetty]

it will be like a >12h delay, if i don't wake up to do it i will just wake up at a normal time and then i have school

at school i have lessons and people to be with, so i'll just do it when i get home

[royal nymph]

css syntax is crazy @crude star

[crude star]

of???

when did they add that

[royal nymph]

https://developer.mozilla.org/en-US/docs/Web/CSS/Reference/Selectors/:nth-child#syntax