🪅-progaming

[valid jetty]

??????????????????

[deep mulch]

sorry

[deep mulch]

ROSIE

spamming my feed

[lucid trail]

average swiftui experience

ts pmo

[shrewd canopy]

This would not happen if you just used Jetpack Compose

:)

[lucid trail]

swift ui is not a serious framework

[lucid trail]

i should look into this

[deep mulch]

yop

[royal nymph]

rosie forking her own repo

[lucid trail]

god swiftui actually sucks

rewriting parts of my app with UIKit because swiftui is too slow

[lucid trail]

true and factual

it was for a totally normal UI bug though

nothing complex

Nope, it was for a toolbar and I forgot to put Spacer inside a ToolbarItem

[nimble bone]

Vibe Coding Enthusiasts

[deep mulch]

@valid jetty

rosie.pi

rosie.314159

[hoary sluice]

@valid jetty

husk

transalted by chatgcpt

task is in german

u can translate it if u wanna red it

read

[deep mulch]

transalted

[valid jetty]

@deep mulch

[deep mulch]

NO

[limpid mica]

no

[deep mulch]

@87flowers hi ii

[limpid mica]

import math
print(chr(sum(range(int(''.join(reversed(str(int(math.sqrt(2025))))), 0x10)))))

[spark tiger]

/run

import math
print(chr(sum(range(int(''.join(reversed(str(int(math.sqrt(2025))))), 0x10)))))

[deep mulch]

amog

[spark tiger]

it literally said why 😭

❓

[shrewd canopy]

this

[spark tiger]

yes

[shrewd canopy]

many languages

[spark tiger]

any language u just gotta specify it after ``

you need three ` s

[pseudo sierra]

do
```py
code

[nimble bone]

#🤖-bot-commands for ur shitposting tyyy

you are clogging up the channel

[winged mantle]

https://www.youtube.com/watch?v=8M0QfLUDaaA great video which told me nothiung i didn't already know

[frosty obsidian]

you can't

its completely sandboxed

[nimble bone]

i don't care do it in #🤖-bot-commands or get banned

[nimble bone]

that too

[deep mulch]

@frosty obsidian wingyyy

[royal nymph]

vprune 100 by bots

vprune 100 by @glad moth

[gilded surge]

was trying to debug why iwlwifi (the wifi module for the linux kernel) was failing on startup and looked it up and eventually figured out that a WINDOWS 11 SETTING is to blame for broken linux drivers, apparently if you enter hibernate windows 11 interjects some stupid fast startup thing that prevents your wifi adapter from starting up correctly

and that happens on all future boots from then on until you disable the setting

[deep mulch]

you love

[shrewd canopy]

nearly all of desktop firmwares are designed and made for windows 11

No wonder why it explodes

[hoary sluice]

@valid jetty @deep mulch my uni uses Go

as the entry language

i was gonna ask if i can submit my homework in a diff language

but ig this is a good reason to learn go

[valid jetty]

lmao oh god

[winged mantle]

go good

[hoary sluice]

yet another reason to use nvim

[shrewd canopy]

It depends on context

Bruh

[fleet cedar]

Some people need to grow a spine

[ionic lake]

Is that jetbrains

[shrewd canopy]

And i dont think nvim works best on windows

[fleet cedar]

I do wonder why the politically corrects went for blocklist and allowlist rather than the obvious redlist and greenlist

[balmy lintel]

it doesn't

whitelist literally means being allowed to do something

[winged mantle]

blocklist and allowlist is more logic

i don't really understand what's meant to be wrong with blacklist/whitelist though

i've never seen an actual... explanation lol

[fleet cedar]

Some people thing everything involving monochrome colors is inherently racist

[winged mantle]

or like master branch? wheres the evidence to suggest there's something wrong with that?

[fleet cedar]

I do agree that master/slave terminology is pretty sus

[winged mantle]

feels like some things got caught in the crossfire of master/slave language

[fleet cedar]

Yeah

[winged mantle]

master/slave is , yes but sometimes there's just a master

[fleet cedar]

Main is a better name than master, but changing the default after several decades is just dumb

[winged mantle]

but who ever talks about the term "master recording" being offensive in music

its not master as in "in charge", it's more like master as in original

[fleet cedar]

Or remastering a game

[winged mantle]

yeah that doesn't mean... reownering

[winged mantle]

nobody really has any strong arguments its all really odd

[royal nymph]

jetbrains has another thing where it recommends that you use neutral words like Mx instead of Ms/Mr

[fleet cedar]

Ew

[winged mantle]

lmao wat

what if that... is somebody's title

[winged mantle]

i think its a little silly to make things always gender neutral tbh

like if you're talking about an actual policeman

[fleet cedar]

I think gender neutral terms are fine if you do -folk

Policefolk, firefolk, chairfolk

[winged mantle]

police officer

fire officer

[fleet cedar]

Chair officer

[spark tiger]

my code is always so inclusive! 😊

[winged mantle]

idk i think this is kind of performative activism

sorry...

[fleet cedar]

It is, yes

[royal nymph]

Also like 95% of firefighter are men

if you want to be inclusive maybe work on getting more women interested in the job than policing people's language

[fleet cedar]

My code isn't particularly inclusive, I have <string> and <vector> in pch so that I don't have to #include them

[winged mantle]

tables have turned

you will be most successful from creating a "non-woke, dei free" ide

[fleet cedar]

What's dei

[winged mantle]

[x] correct gender neutral/female language
change all pronouns to he/him

[royal nymph]

[X] Keep politics out of open source - Prefer he/him over woke pronouns like they/them

[winged mantle]

ctrl+alt+a Insert ableist language

probably the sort of thing people like these days...

[fleet cedar]

I don't mind they/them, that one's been used for centuries

[winged mantle]

i'm not "anti-woke"

i do find some stuff a bit silly though lol

sometimes

allowlist is a more logical word arguably

[nimble bone]

[X] Keep politics out of open source - Prefer nonwoke maintainers over woke maintainers like @royal nymph

[hoary sluice]

the wokes are trying to ruin our standards

[balmy lintel]

most useless thing

everything/everyone is male for me unless they specify otherwise

[hoary sluice]

can it stop fucking glazing me

[hoary sluice]

everything/everyone is male.

[royal nymph]

change personality in chatgpt settings

[hoary sluice]

ty

[jade stone]

Why would you ever use svelte though?

[winged mantle]

should i switch to sqlite

i guess probably not (because i'm too far into this project)... but is sqlite fine even if you have multiple processes accessing the same db

[hoary sluice]

@valid jetty is this actually bad

let a = &self.something;
do_stuff(a.read_stuff());
let mut b = self.borrow_mut();
b.do_stuff();

chatgpt told me this is ub because a is still alive while self is mutably borrowed in b

@fleet cedar

even tho a isnt used after b is mutably borrowed

[valid jetty]

thats fine as long as you dont use a after creating b

or well, rust wont let you

[hoary sluice]

chat says itll work now but it might break with a different compiler

[valid jetty]

???

why are you listening to chatgpt

Excellent! You're thinking like a true vibe coder now. Here is the correct solution:

[hoary sluice]

if i was listening to it i wouldnt have asked you

im just confused

as to why it thinks theres an issue here

[fleet cedar]

Because it doesn't think

It's autocomplete

[hoary sluice]

You can’t have a shared (&T) and mutable (&mut T) reference to the same data at the same time.
Even if you don’t use the shared one after creating the mutable one, both exist concurrently, which breaks Rust’s aliasing rules.

The compiler and optimizer assume a &mut T is the only reference to that memory.
If an old &T still exists, the optimizer may reorder or remove reads/writes, producing wrong results.

So code like:

let a = &self.field;
let b = self.borrow_mut(); // mutable borrow of all `self`

is undefined behavior because a and b could alias.
Rust forbids this to keep memory safety and compiler optimizations valid.

specifically when you borrow some field of self immutably before then mutably borrowing all of self

[fleet cedar]

No, NLL was added many years ago

[hoary sluice]

i will be arguing with it until one of us proves their point

[fleet cedar]

References are tracked by liveness, not just lexical scope

[hoary sluice]

i am aware

[spark tiger]

#1426968928554455222

[hoary sluice]

and so is he

[manic condor]

Hi all, quick question! I love using the pinned DMs plugin, but I was wondering if there's a way to hide this section here? I'm trying to clean up my view a bit. Thanks in advance. Also if anyone has a code for hiding the sicker button that be awesome too.

[fleet cedar]

Is pindms different from favoriting a chat in vanilla?

[manic condor]

this is what i am using if you are talking to me

[hoary sluice]

@fleet cedar can you help me understand this The update would be a compiler optimization that fully trusts the noalias guarantee on &mut references.
It would allow LLVM (or a MIR pass) to reuse or reorder memory reads and writes across function calls that take &mut, assuming no other reference can see the same data.
Once that optimization lands, code that still keeps an &T alive while calling a &mut self function could start reading stale values—a silent miscompile—because the compiler now skips reloading memory it believes cannot change.

or why it is wrong

cause i dont

[fleet cedar]

That sounds like complete poop

If you wanna discuss questions that's fine, but discussing ai poop answers is nonproductive

[hoary sluice]

it does, its saying because there is another reader it will skip the write

or something

[fleet cedar]

It's writing words in plausible-looking sequences

[hoary sluice]

Let’s be exact.

If your immutable reference is never used after the mutable borrow, then in practice there is no observable problem.
The compiled code will behave the same no matter how the optimizer reorders things, because nothing reads through that old reference again.

The formal UB label only exists because Rust’s aliasing model has to forbid all overlap so it can safely mark every &mut as noalias without analysing each case.
That’s a conservative rule for correctness proofs, not a statement that your specific code will miscompile today or tomorrow.

So:

• Real-world effect: nothing bad happens, because the old reference is never used.
• Language-lawyer view: still outside the defined model, because the compiler isn’t required to keep it working.
  Both statements are true. The “UB” flag is about what Rust guarantees, not about an actual runtime hazard in your exact program.

[fleet cedar]

If you want to discuss nuances of borrow checking, a human is much better than a word poop machine

[hoary sluice]

ok thats what i thought

[hoary sluice]

its writing 'your code is ub according to the spec but it will never actually matter' but in many complicated words

and i dont even know if its right about it being """ub"""

[fleet cedar]

And it's wrong, because if it was ub then borrowck wouldn't accept it

[hoary sluice]

that was my argument

it said the borrow checker cant detect all ub

[fleet cedar]

If you ask your questions here I can answer them

I do not waste my time arguing against word poop

[hoary sluice]

i was moreso trying to convince the llm that im right

[fleet cedar]

A futile endeavor

[hoary sluice]

well i did it

im right 😎

i shouldve just gone to sleep instead

[fleet cedar]

I'm sure it still snuck in an incorrect edge case

Like it always does

[hoary sluice]

yea kinda

but not really

[fleet cedar]

Yep it's still claiming there's ub

[hoary sluice]

its also saying the code is safe

[fleet cedar]

It's literally saying "the code is ub but compiler lets it slide"

[hoary sluice]

well

yes

lol

it feels like its instructed to make the code it generates as verbose as possible

so even if you use it as a typing assistant you still have to refactor what it spits out

[fleet cedar]

Most likely yes

[hoary sluice]

wtf its been an hour

im never debating an llm again

[frosty obsidian]

you can't debate something that literally has no capacity for reason

[deep mulch]

in this project i have theres a massive workflow with building and testing and release

would it be better to split the release into its own?

but then i would have to copy/paste like 200 lines of jobs

@jade stone help

[ionic lake]

build a matrix

[deep mulch]

i did

its over 300 lines total

[jade stone]

That's what I did

[deep mulch]

what about the duplicate jobs

[jade stone]

[ornate quiver]

does that workflow also run for non-release things?

[deep mulch]

yes

https://github.com/dawidd6/action-download-artifact i found this

could run build workflow, then release workflow would be able to take those artifacts

idk

[ornate quiver]

I would split it off
there's subtle differences that don't make putting them together good
like for example you shouldn't be using cache for release builds

[deep mulch]

oh really

i didnt think about that

[ornate quiver]

yeah

[deep mulch]

i willd o that

[royal nymph]

why not use the official one https://github.com/actions/download-artifact

[deep mulch]

Let's suppose you have a workflow with a job in it that at the end uploads an artifact using actions/upload-artifact action and you want to download this artifact in another workflow that is run after the first one. Official actions/download-artifact does not allow this.

vee blindd

[ornate quiver]

that's across multiple workflows, not jobs

why would you need cross-workflow artifacts

oh you sent that a long time ago

[supple whale]

gh so brokey recently

[pseudo sierra]

had this happen to me a few days ago

[supple whale]

HORRRRYYY

thats crazy

never seen that1

[spark tiger]

😭 https://x.com/lundukejournal/status/1977590202873901194

[hoary sluice]

should detec google chrome/ms teams/minecraft/oracle (so any java)/any cisco soft instead

who cares what vaxrys opinions are

nothing in the world changes because of vaxrys opinions

[shrewd canopy]

whats wrong with google chrome and minecraft/java

[shrewd canopy]

I support blocking omarchy/arch and hyprland

[pseudo sierra]

fell for the bait award

[hoary sluice]

they are usa software

and host genocide

minecraft specifically just because its microsoft owned

java because of oracle

[shrewd canopy]

can say same about windows, discord and literally most of x86 hardware you have

[hoary sluice]

yes

[shrewd canopy]

doesn't make sense at all

[hoary sluice]

neither does the tool from the post

[hoary sluice]

i just didnt feel like listing everything remotely related to fascism

[shrewd canopy]

i would bundle some code from it to my programs tbh

have to block linux users

[deep mulch]

🦅ly @hoary sluice

[hoary sluice]

😡

@valid jetty

i just did some progaming for like 15 minutes and then realized my mouse was disconnected the whole time

i think ive achieved perfection

[shrewd canopy]

I prefer supporting predictable and stable platforms (linux isnt one of them), sorry

[hoary sluice]

https://tenor.com/view/4lokogangemperor-rage-bait-ragebait-china-lore-gif-7803056600524123396

[shrewd canopy]

not ragebait unfortunately

[hoary sluice]

windows being predictable and stable is one of the things ive heard of all time

[deep mulch]

true

[stark timber]

ok am running into this issue now 😭 where can I find that proxy wrapper u were talking about

[serene elk]

just maka the target of the proxy a function

instead an object

new Proxy(function(){}, {....})

[stark timber]

olrite

[spark tiger]

i didn’t? i’m just posting because it’s funny especially the maintainer review

[serene elk]

and then if u run into an issue with making Object.keys or similar work let me know

there's another thing in particular of when you make the target a function

[stark timber]

[pseudo sierra]

I meant lunduke

[spark tiger]

oh mb

[valid jetty]

https://cdn.discordapp.com/attachments/542264318465671170/1427668253752168579/Screen_Recording_2025-10-14_at_15.39.31.mov?ex=68efb32c&is=68ee61ac&hm=39df94a519d765eb1d380eb18ac1a61fcf950ff8549cc04e7fd571e0b1319eb0& https://cdn.discordapp.com/attachments/542264318465671170/1427669054201790604/Screen_Recording_2025-10-14_at_15.36.34.mov?ex=68efb3eb&is=68ee626b&hm=9246ec83d3acbe46ef78df51c291c4d57824d37f103a6660c1e23bd67c171096&

@hoary sluice @deep mulch i plotted a julia set

basically you iteratively call f(z) = z^2 + C

and then shade the point based on whether the resulting curve diverges (goes off to infinity) or converges (approaches a limit point), and if diverges how fast it does so

and with my mouse i'm controlling the value of C

and if you render that you get this very beautiful fractal

[deep mulch]

is that in elle

[valid jetty]

yes

[deep mulch]

oh is it entirely cpu

no shader?

[valid jetty]

nono it is a shader

this is the shader

[deep mulch]

oh

[valid jetty]

this is the elle code that renders it

[deep mulch]

whats fs

[valid jetty]

fragment shader

glsl

[deep mulch]

oh

isn't the extension glsl

[valid jetty]

idk

i call it .vs and .fs usually

[deep mulch]

this is actually simpler than I expected

[deep mulch]

evil

[valid jetty]

well yeah

[valid jetty]

^^^

it's like surprisingly simple

[deep mulch]

I should do in my compose fractal thing

[valid jetty]

the test for divergence is basically whether it exits the unit circle lol

that's the cplx_abs(z) > 2

[deep mulch]

https://github.com/zt64/compose-scripts/blob/main/host/src/commonMain/resources/scripts/Fractal.composable.kts you love

I mostly just used AI cause I literally could not find anything that did mandelbrot programmatically that wasn't single letter variables

[hoary sluice]

you didnt have to explain what diverging and converging means, i dont understand what youre trying to say either way

[hoary sluice]

what is z

[deep mulch]

julia seems easier

[hoary sluice]

what makes this different from a Mandelbrot

[deep mulch]

@hoary sluice you love my compose scripts project

[royal nymph]

just kang from vencord

[crude star]

julia and mandelbrot sets are dual

[valid jetty]

in mandelbrot C is kept the same and you plot differently based on the starting position

in julia you keep starting position the same and change C

yes

[royal nymph]

https://github.com/Vendicated/Vencord/blob/main/src/utils/lazy.ts

[valid jetty]

the values of C that converge are actually defined by the mandelbrot set

[hoary sluice]

oh so z is constant

[valid jetty]

no lol

z is a variable in f(z)

[deep mulch]

oh Rosie my Rosie

there was some cool fractal video I saw

[hoary sluice]

you wrote f(x)

[valid jetty]

oh

[hoary sluice]

do u work in mobile dev

so what is z

[crude star]

https://youtu.be/Ed1gsyxxwM0 (not the julia set but he ends up making it at some point)

[valid jetty]

f(z) = z^2 + C

[deep mulch]

function!!

[valid jetty]

start with some z, typically the position of the pixel you're calculating

then you just apply f(z) = z^2 + C some amount of times

if the output point is within the unit circle (|z| < 2), then it converges, otherwise it diverges

and you control the value of C

[hoary sluice]

what is z

[valid jetty]

the more iterations, the higher resolution your fractal

[valid jetty]

usually the position of the pixel as a complex number

the initial z*

[deep mulch]

how many dimensions are julia sets

[hoary sluice]

i get it

you iterate f(z)

[deep mulch]

@valid jetty https://youtu.be/aLR4YpTnM6c

[hoary sluice]

the guy in the video explained it

[valid jetty]

that's what i said lol

[valid jetty]

depending on the value of C, between 1 and 2

[valid jetty]

yeah this was partially what started my obsession with fractals

[hoary sluice]

oh i didnt read this message

[deep mulch]

@valid jetty make it so you can step through the iterations

@valid jetty make elle compile to spir-v

[valid jetty]

well specifically i start with

float x = mix(x_min, x_max, fragTexCoord.x);
float y = mix(y_min, y_max, fragTexCoord.y);
vec2 z = vec2(x, y);
``` which basically maps the pixel coord into a value from x_min to x_max and the same for y

[valid jetty]

that's like definitely doable

[deep mulch]

world to screen or something

[valid jetty]

you can just control the value of max_iterations

[deep mulch]

@valid jetty hii

hop off silent typing

so bad

[valid jetty]

nop

now that i have client mods on my phone again i will take full advantage

[deep mulch]

evil

[winged mantle]

i have a good idea

[deep mulch]

kode tode

[valid jetty]

the other way

screen to world

[stark timber]

I could

[stark timber]

is implementing get enough

'cause I tried that but it always returns undefined

like this should be returning a rejected promise

[winged mantle]

i don't think my idea will work

i wanted to make iostream in kotlin :(

there's no <<

and you can't even overload shl

[serene elk]

no

you need apply too

[stark timber]

yea that's what I was thinking

ty

[deep mulch]

oop

make elle run on embedded devices when

I wanna run elle on my esp32

[hoary sluice]

it alreado can

compile to x86, decompile to c, compile to esp32

[stark timber]

look guys I made it so u can search and sort playlists in the browser

[solid gazelle]

racchole

[spark tiger]

kmoe spotted

[placid igloo]

how many minutes

I've got 6 outof 7 minutes

[placid igloo]

thought this was my Spotify activity for a second

[winged mantle]

i just realised i'm kind of not prioritising getting hello world working in my language

hello world is hard...

idk maybe strings can just be another primitive type

[shrewd canopy]

They should be

[winged mantle]

my value.h header is already bloated

but yeah i guess they probably should be lol

why did i think this would be hard

[shrewd canopy]

I mean optimizing strings if they were primitives would be easier

[winged mantle]

if they were objects that'd mean two levels of indirection to the data ig

i had to remind myself that strings are immutable in gc languages so that would be pointless

unless i want to be weird and make strings mutable

[shrewd canopy]

Make another string primitive then

[winged mantle]

i like the way the compile command i made doesn't tell you off for passing no files lol

[shrewd canopy]

Mutable string primitive purely for optimization/performance purposes

[winged mantle]

horror

trying print hello world

that's what i like to see

seems my lexer doesn't like strings

what could i possibly be doing wrong LOL

[shrewd canopy]

[winged mantle]

average c code of somebody who doesn't normally code in c, probably

idk

probably pretty typical c code

&***&*(void **)memcpy(x1, ++++++x2)

[jade stone]

copilot is so smart and so stupid at the same time

[royal nymph]

how do you not have /bin/bash????

[ornate quiver]

i ate it

[valid jetty]

@hoary sluice https://media.discordapp.net/attachments/1331136257597771938/1427928235877859398/RPReplay_Final1760515010.mov?ex=68f0a54d&is=68ef53cd&hm=43229c35da37d6701833c740957aabd5e1597259a1d7b9439d187b8b80dc9de8&

i turned the message action sheets into context menus

it's actually way nicer than the action sheet i love it

tho i have to find a way to bring reacting back

[hoary sluice]

how did u do this

[valid jetty]

client mods

i have this patch ```ts
Patcher.before(ActionSheet, 'openLazy', (self, [component, type, message]) => {
if (type !== "MessageLongPressActionSheet" || !message) return;

component.then(instance => {
const unpatch = Patcher.after(instance, 'default', (_, __, res) => {
unpatch();
const buttons = findInReactTree(res, x => x?.[0]?.type?.name === 'ButtonRow');
if (!buttons) return res;
const state = ContextMenu.useContextMenuState();
const items = React.useMemo(() => buttons.map(({ props: { message, onPressRow, IconComponent } }) => ({
IconComponent,
action: onPressRow,
label: message
})), []);

  React.useEffect(() => {
    ActionSheet.hideActionSheet();
    ContextMenu.showContextMenu({
      key: `uid_${message.id}`,
      x: 33,
      y: 250,
      position: 'below',
      align: 'left',
      height: 126,
      width: 102.33332824707031,
      state,
      items,
      title: "Manage Message",
      requestClose: () => {
        ContextMenu.hideContextMenu();
        const active = state?.activeIndex?.value;
        active >= 0 && items[active].action();
      },
      onClose: () => {}
    })
  }, []);

  return res;
});

});
})

that ive been working on since last night

i finally got it to work this morning

[hoary sluice]

discord uses react native??

[valid jetty]

on mobile yes

[hoary sluice]

I thought they did kotlin

[valid jetty]

on ios, always

on android, since discord 127

[hoary sluice]

when was 127

and were they on kotlin before

[valid jetty]

we're now on discord 300-ish

[hoary sluice]

oh

[valid jetty]

they were on kotlin

that's why aliucord is on such an old version

[hoary sluice]

i heard they were switching to react native like a year ago

[valid jetty]

they're using the last version before discord switched to RN

i'm also running an old version, the last one that has tabsv2

discord 213

[hoary sluice]

is there deobfuscated source code for discord

[valid jetty]

which is from about feb 2024

[hoary sluice]

whats tabsv2

[valid jetty]

no, but you can disassemble the bundle into some sort of assembly-like soup

[valid jetty]

this chat ui https://cdn.discordapp.com/attachments/1205946622995865600/1427021577660272753/RPReplay_Final1760298774.mov?ex=68f0a4a8&is=68ef5328&hm=6ad8f6dba8ef9a42e9a8ccd8588d995965263f2e4522a75ffae58d2aced817f4&

skip to like 30 seconds in

[hoary sluice]

i think i remember that layout

[valid jetty]

it was awesome

that's why i'm bringing it back

[hoary sluice]

rosie are you allowed to work at 17 in the uk

[valid jetty]

uhhh probably

[hoary sluice]

you can open a company at 16

and do b2b contracting

so there are no restrictions on working hours

you should look for part time contracting

to build experience on a resume

(or just lie on ur resume after ur done with school)

[ionic lake]

The children yearn for the mines

[hoary sluice]

speaking of children

i will raise my children as software devs from a young age

then outsource my work to them

and when theyre adults i will apply to jobs for them with my resume and let them do the work and take all the money

i should name my eldest son after myself so this is easier to do

[valid jetty]

make a pr now

@royal nymph you will merge

[hoary sluice]

you should submit this to expressjs itd be funny

[gilded surge]

i square to god i will kms

the quickshell systemd service doesn't correctly load singletons when u reload your config so you have to stop and then start the service

[spark tiger]

It's finally here. Zed is available on Windows.

Built from scratch, and rendering at 1 million pixels/millisecond.

[hoary sluice]

wtf is that unit

[jade stone]

guhhhhhhhhh

me when i want to use regex but vscode says EOL bug

[lucid trail]

fps except it’s dependent on resolution lol

[hoary sluice]

no its fpms

so it runs at 101fps at 4k

thats not that impressive

[winged mantle]

marketting

[valid jetty]

@robust jackal ignore how it's so slow but i did it.. (it turns out making a pop out component here is really hard)

i had to go 4 patches deep lmao

hell.

thanks discord!

yay

[tired vigil]

does the mobile app not have hashed intl keys

[valid jetty]

it does

im running an old version

[tired vigil]

I see

[shrewd canopy]

average react native moment

[valid jetty]

not really

usually its 1-2 max

this is just ultra bad

well i should probably do a findInReactTree for that

but i just took it straight from my repl

its fineeeeeee

@cinder egret can you give me brainrot for a sec i wanna see if it explodes

no obviously not lol

im on 213

oh

212

[cinder egret]

you don’t deserve that ❤️

[valid jetty]

ignore that lmao

its from a really old plugin i made

import { get, set } from '../common/store';

export abstract class Patch {
    static key: string;
    static title: string;
    static subtitle: string;
    static icon: string;

    static get(key: string, def?: any) {
        return get(`${this.key}.${key}`) ?? def
    }

    static set(key: string, value: any) {
        return set(`${this.key}.${key}`, value);
    }

    static get enabled() {
        return this.get('enabled');
    }

    static patch(Patcher) { };
    static render({ disabled }) { };
}

i made the chat menu thing as part of utilitypatches

which i havent touched in at least 2 years

[valid jetty]

@hoary sluice @deep mulch

this one isnt interactive because its a very particular fractal

its (|Re(z)| + i|Im(z)|)^2 + c where c is the pixel coordinate

[deep mulch]

burning ship so cool

show shader code

[deep mulch]

where did you learn how

[valid jetty]

wikipedia

[deep mulch]

is there any guides that isn't just plain math

[deep mulch]

that's plain math

[valid jetty]

it literally just tells you

[deep mulch]

I cannot read that

idk what Im is

tlelll

@valid jetty you will help me

[valid jetty]

a complex number is z = a + bi

Re(z) = a
Im(z) = b

[deep mulch]

I'll try to make it in my compose script

what is Re and Im

[valid jetty]

real and imaginary parts of z

[crude star]

real and imaginary

|x| is abs(x)

[deep mulch]

ik

what is c

[valid jetty]

the pixel coord

basically, its like the mandelbrot set

but instead of just z^2 you take the absolute value of the components

[deep mulch]

o

does x and y map to c somehow

[deep mulch]

oh

so it's a vec2

guh

[valid jetty]

min and max are just the bounds of the graph

[deep mulch]

I hate single variable letter names like that

[valid jetty]

c.x == Re(c)
c.y == Im(c)

[deep mulch]

just call it pos

[valid jetty]

its not pos

[deep mulch]

@valid jetty im ready

teach me

@valid jetty

rosie either died or is typing an essay

i hate silent typing

[valid jetty]

im heree

[deep mulch]

hii

[valid jetty]

what do you want me to do

[deep mulch]

i wanna make burning ship

what if i timeout you @young flicker

[valid jetty]

can you get a red screen shader

[deep mulch]

yes

hold on

gonna add a way to switch shader

[valid jetty]

okiee

[deep mulch]

@valid jetty

[valid jetty]

ok awesome

[deep mulch]

rosinga

[valid jetty]

now define a vec2 for min_x, max_x, and another for min_y, max_y

define a number of max iterations and send it to the gpu

now in the gpu, get these functions so you can perform complex arithmetic

// (a + bi)(c + di) = (ac−bd) + (ad+bc)i
vec2 cplx_mul(vec2 a, vec2 b) {
    return vec2(a.x * b.x - a.y * b.y, a.x * b.y + a.y * b.x);
}

// |a + bi| = sqrt(a^2 + b^2)
float cplx_abs(vec2 a) {
    return sqrt(a.x * a.x + a.y * a.y);
}

you dont really have to know how they work, just that youll be using them

now in your main shader

convert your pixel coordinate into a coordinate from min_x to max_x and same for y

float x = mix(x_min, x_max, fragTexCoord.x);
float y = mix(y_min, y_max, fragTexCoord.y);
vec2 c = vec2(x, y);

that is your value of c

z starts at 0, as defined by the formula

now perform the iterative process the number of times you specified

vec2 z = vec2(0.0, 0.0);

int i = 0;
for (; i < max_iter; ++i) {
    // z = (|Re(z)| + |Im(z)|i)^2 + c
    z = vec2(abs(z.x), abs(z.y));
    z = cplx_mul(z, z) + c;
    
    // if the distance from the center is more than a circle with radius 1, then it diverges
    if (cplx_abs(z) > 2.0) break;
}

now, color this pixel however you like

i chose to shade it smoothly but im sure you can figure out how to color it

if not, i did it like this

float t = float(i) / float(max_iter);
t = sqrt(t);
t = smoothstep(0.0, 1.0, t);
t = 1 - t;

float k = t * 4.0;
int l = int(floor(k));
int m = min(l + 1, 4);
float f = fract(k);

vec3 col = mix(colors[l], colors[m], f);
finalColor = vec4(col, 1.0);

where colors is passed from the cpu

[deep mulch]

LONG

okay

i will try

[valid jetty]

and then after, define in compose some way to change min_x, min_y, max_x, max_y using the mouse

this is how i do it in raylib you can use it as reference

[deep mulch]

funny thing i could theroetically run this in a discord bot

user sends it compose code and it returns rendered image

[valid jetty]

lmao horror

[deep mulch]

jetbrains should hire me

[winged mantle]

horror

how did the typescript language server just disappear

[deep mulch]

sorry

[humble stream]

This is smaller than my brain

[winged mantle]

i guess people use vscode because it just works

[valid jetty]

my whole shader is here if you wanna use it as reference

its very similar to the other fractals i made

[deep mulch]

now define a vec2 for min_x, max_x, and another for min_y, max_y
on the cpu side?

Try catch is not supported around composable function invocations. guh!

[winged mantle]

oh i just deleted the stuff i added to my config and it works

also i forgot to install tsx treesitter parser

[deep mulch]

ill just catch exception inside the script

oh funny i think i found a compiler bug

[valid jetty]

yes

and pass it to the gpu

[deep mulch]

Caused by: java.lang.RuntimeException: error: 4: symbol 'min' was already defined if i dont catch it i get this error

[valid jetty]

because youll be changing it on the cpu side

[deep mulch]

if i do catch it i get java.lang.Exception: ERROR Unresolved reference 'makeShader'. (script.composable.kts:204:38)

[shrewd canopy]

same reason why people use windows

why husk

[deep mulch]

insane statement

[winged mantle]

should i be worried by the fact that an import call is taking 100ms

i'm probably doing too much stuff with global initialisation

or idk i guess node.js automatically runs swc for all ts files

that might be a bit slow

[shrewd canopy]

how it's insane

[winged mantle]

comments still showing white in tsx

[deep mulch]

@valid jetty ar the min and max literally just the bounds of the canvas

[valid jetty]

yes

[deep mulch]

meow

time to see if this works

[valid jetty]

but theyre on the cpu so you can control them

[deep mulch]

negative min?

scary

[valid jetty]

well yes this is the complex plane

values range from -i to i and -1 to 1

[deep mulch]

whats the simplest way to color i

it

[valid jetty]

something like this

float t = float(i) / float(max_iter);
vec3 col = vec3(t, t * t, t * 0.5);
finalColor = vec4(col, 1.0);

[deep mulch]

funny thing

apparently sksl doesnt let me use a non-const in for loops

so i have to hardcode max_iter

idk if theres a way around

[valid jetty]

lmao thats probably fine

you probably wont be changing that during runtime anyway

[deep mulch]

where would i place that? before the break statement?

[valid jetty]

nono afterwards

[deep mulch]

oh

okay

[valid jetty]

when all the iterations are done

whatever the value of i is

[deep mulch]

is there a reason finalColor is camelCase

[valid jetty]

thats just what glsl expects

[deep mulch]

oh i forgot

sksl has you use a return statement

[valid jetty]

typically the

[deep mulch]

i could maybe even return from insie the for loop

[valid jetty]

uhhh yeah probably

[deep mulch]

i think i did something wrong

half4 main(vec2 fragCoord) {
    float x = mix(x_min, x_max, fragCoord.x);
    float y = mix(y_max, y_max, fragCoord.y);
    vec2 c = vec2(x, y);
    
    vec2 z = vec2(0.0, 0.0);
    half4 final_color = vec4(1.0, 1.0, 1.0, 1.0);
    for(int i=0; i < max_iter; i++) {
        z = vec2(abs(z.x), abs(z.y));
        z = cplx_mul(z, z) + c;
        
        if (cplx_abs(z) > 2.0) break;
        
        float t = float(i) / float(max_iter);
        vec3 col = vec3(t, t * t, t * 0.5);
        final_color = half4(col, 1.0);
    }
    
    return final_color;
}

its solid white

[valid jetty]

half4 main(vec2 fragCoord) {
    float x = mix(x_min, x_max, fragCoord.x);
    float y = mix(y_max, y_max, fragCoord.y);
    vec2 c = vec2(x, y);
    
    vec2 z = vec2(0.0, 0.0);
    half4 final_color = vec4(1.0, 1.0, 1.0, 1.0);
    for(int i=0; i < max_iter; i++) {
        z = vec2(abs(z.x), abs(z.y));
        z = cplx_mul(z, z) + c;
        
        if (cplx_abs(z) > 2.0) {
            float t = float(i) / float(max_iter);
            vec3 col = vec3(t, t * t, t * 0.5);
            final_color = half4(col, 1.0);
            break;
        }
    }
    
    return final_color;
}

[deep mulch]

solid black now

[valid jetty]

suspicious

[deep mulch]

is -2f/2f fine for min/max

[valid jetty]

i think its working youre just looking at the wrong region or something

this is what youre supposed to see

[deep mulch]

o

[valid jetty]

can i see the full code

[deep mulch]

BURNING_SHIP(
    shader = """
        uniform float x_min, x_max, y_min, y_max;
        const int max_iter = 256;
        
        // (a + bi)(c + di) = (ac−bd) + (ad+bc)i
        vec2 cplx_mul(vec2 a, vec2 b) {
            return vec2(a.x * b.x - a.y * b.y, a.x * b.y + a.y * b.x);
        }
        
        // |a + bi| = sqrt(a^2 + b^2)
        float cplx_abs(vec2 a) {
            return sqrt(a.x * a.x + a.y * a.y);
        }
        
        vec4 main(vec2 fragCoord) {
            float x = mix(x_min, x_max, fragCoord.x);
            float y = mix(y_max, y_max, fragCoord.y);
            vec2 c = vec2(x, y);
            
            vec2 z = vec2(0.0, 0.0);
            vec4 final_color = vec4(1.0, 1.0, 1.0, 1.0);
            for(int i=0; i < max_iter; i++) {
                z = vec2(abs(z.x), abs(z.y));
                z = cplx_mul(z, z) + c;
                
                if (cplx_abs(z) > 2.0) {
                    float t = float(i) / float(max_iter);
                    vec3 col = vec3(t, t * t, t * 0.5);
                    final_color = vec4(col, 1.0);
                    break;
                }
            }
            
            return final_color;
        }
    """.trimIndent()
) {
    var minX = -2f
    var minY = 2f
    var maxX = -2f
    var maxY = 2f

    override fun buildBuffer(): ByteBuffer {
        return ByteBuffer
            .allocate(16)
            .order(ByteOrder.LITTLE_ENDIAN)
            .putFloat(minX)
            .putFloat(minY)
            .putFloat(maxX)
            .putFloat(maxY)
            .rewind()
    }
};

eventually i wanna store the shader code in a seperate file

[valid jetty]

it should be y_min, y_max

these values are wrong

should be -2, -2, 2, 2

[deep mulch]

GUH!

blame copilot for that

i should disable

[deep mulch]

wdym

oh

yea

i confused my own var order

[valid jetty]

lmao

[deep mulch]

and i pass it in the wrong order too

[valid jetty]

lmao horror

[deep mulch]

hmmstill black

unless that single pixel is something

BURNING_SHIP(
    shader = """
        uniform float x_min, y_min, x_max, y_max;
        const int max_iter = 256;
        
        // (a + bi)(c + di) = (ac−bd) + (ad+bc)i
        vec2 cplx_mul(vec2 a, vec2 b) {
            return vec2(a.x * b.x - a.y * b.y, a.x * b.y + a.y * b.x);
        }
        
        // |a + bi| = sqrt(a^2 + b^2)
        float cplx_abs(vec2 a) {
            return sqrt(a.x * a.x + a.y * a.y);
        }
        
        vec4 main(vec2 fragCoord) {
            float x = mix(x_min, x_max, fragCoord.x);
            float y = mix(y_min, y_max, fragCoord.y);
            vec2 c = vec2(x, y);
            
            vec2 z = vec2(0.0, 0.0);
            vec4 final_color = vec4(1.0, 1.0, 1.0, 1.0);
            for(int i=0; i < max_iter; i++) {
                z = vec2(abs(z.x), abs(z.y));
                z = cplx_mul(z, z) + c;
                
                if (cplx_abs(z) > 2.0) {
                    float t = float(i) / float(max_iter);
                    vec3 col = vec3(t, t * t, t * 0.5);
                    final_color = vec4(col, 1.0);
                    break;
                }
            }
            
            return final_color;
        }
    """.trimIndent()
) {
    var minX = -3.5f
    var minY = -3.5f
    var maxX = 2.5f
    var maxY = 2.5f

    override fun buildBuffer(): ByteBuffer {
        return ByteBuffer
            .allocate(16)
            .order(ByteOrder.LITTLE_ENDIAN)
            .putFloat(minX)
            .putFloat(minY)
            .putFloat(maxX)
            .putFloat(maxY)
            .rewind()
    }
};

tried changing min/maxs

[valid jetty]

i have a feeling this is something to do with the difference between fragCoord in sksl and fragTexCoord in glsl

[deep mulch]

maybe

hm

ill try something

[valid jetty]

i think im right

try this

float x = mix(x_min, x_max, fragCoord.x / iResolution.x);
float y = mix(y_min, y_max, fragCoord.y / iResolution.y);

[deep mulch]

you are i think too

yea

i need to pass in the resolution

yippee

@valid jetty

[valid jetty]

yayyyyyy

[deep mulch]

how make it pretty now

[valid jetty]

now find a way to shade it properly

[deep mulch]

yop

[valid jetty]

i do this

float t = float(i) / float(max_iter);
t = sqrt(t);
t = smoothstep(0.0, 1.0, t);
t = 1 - t;

float k = t * 4.0;
int l = int(floor(k));
int m = min(l + 1, 4);
float f = fract(k);

vec3 col = mix(colors[l], colors[m], f);
finalColor = vec4(col, 1.0);
``` where ```glsl
uniform vec3 colors[5];
``` and
```rs
colors := [
    Vector3::new(0.0, 0.0, 0.0),
    Vector3::new(0.835, 0.176, 0.0),
    Vector3::new(0.937, 0.463, 0.153),
    Vector3::new(1.0, 1.0, 1.0),
    Vector3::new(0.639, 0.0078, 0.384)
];

for i in 0..colors.len() {
    loc := rl::get_shader_location(shader, "colors[{}]".format(i));
    rl::set_shader_value(shader, loc, &colors[i].x, ShaderUniformDataType::SHADER_UNIFORM_VEC3);
}

[deep mulch]

is colors a gradient?

[valid jetty]

nop

well

[deep mulch]

o

[valid jetty]

its not really

[deep mulch]

how doesi t owkr

[valid jetty]

its the lesbian flag lol

you can just mess around with the colors

wait actually

i dont think this one is the lesbian flag

[deep mulch]

grunka grunka

i need to add hot reloading

to my code

[valid jetty]

its these colors

as well as black and white

[deep mulch]

@valid jetty you should run my compose-scripts project on your pc

[valid jetty]

maybe,,

[valid jetty]

i changed it because the lesbian color scheme looks bad here

[deep mulch]

its not very practical but its so cool

burning ship

[valid jetty]

better

if you add controls you should go explore this area

[deep mulch]

love

[valid jetty]

on the left

[deep mulch]

is this just random operations until it looked pretty or is there math behind it

bright

[valid jetty]

normalize, apply smoothing, flip (so dark regions are inside the fractal rather than outside)

i have 5 colors so i map t to [0, 4], aka the indexes of the color
then l is the lower color, m is the upper color
and f is the fractional part used for blending

then i interpolate and thats the color

[deep mulch]

o

neat

[valid jetty]

its colored based on how many iterations it takes to escape

thats why i invert it

[deep mulch]

@valid jetty how many iterations do you use

[valid jetty]

300

[deep mulch]

low iterations look funny

[deep mulch]

@valid jetty got something that isnt painful to look at

[valid jetty]

yayyy

now make interactive

[deep mulch]

soon

how zoom

[valid jetty]

mouse := rl::get_mouse_position();
m := Vector2::new(
    x_min + mouse.x * (x_max - x_min) / SCREEN_WIDTH,
    y_max - mouse.y * (y_max - y_min) / SCREEN_HEIGHT
);

wheel := rl::get_mouse_wheel_move();
if wheel != 0 {
    scale := wheel > 0 ? 0.95 : 1.05;
    x_min = m.x + (x_min - m.x) * scale;
    x_max = m.x + (x_max - m.x) * scale;
    y_min = m.y + (y_min - m.y) * scale;
    y_max = m.y + (y_max - m.y) * scale;
}

[deep mulch]

oh

thats simpler than i tohught

theres a slight problem i need to figure out

[limpid mica]

it's just an affine transform smh

[deep mulch]

its following the window scale, so it gets stretched if the window is wide

[deep mulch]

im dumbb

[valid jetty]

thats to be expected no?

thats why my window is square

[deep mulch]

i dont like that

[limpid mica]

my name rini

[deep mulch]

zoom works

need to add offsets now

[limpid mica]

hi p

[deep mulch]

@limpid mica hiiiii

the lily

@young flicker hats me

[valid jetty]

if rl::is_mouse_button_down(MouseButton::MOUSE_LEFT_BUTTON) {
    dx := rl::get_mouse_delta().x * (x_max - x_min) / SCREEN_WIDTH;
    dy := rl::get_mouse_delta().y * (y_max - y_min) / SCREEN_HEIGHT;

    x_min -= dx;
    x_max -= dx;
    y_min -= dy;
    y_max -= dy;
}

[deep mulch]

i was gonna do that @valid jetty

[valid jetty]

sowee

[deep mulch]

evil roie

[valid jetty]

@deep mulch

@deep mulch

@deep mulch

[deep mulch]

zoom and offset working kinda @valid jetty you love

lumina

[valid jetty]

yayyyyy

[deep mulch]

you are lumina

[valid jetty]

you love

lumina

lumina

lumina

lumina

lumina

..

@young flicker can you react wwith lumina so i dont seem insane

[deep mulch]

as i zoom in it amplifies the existing offset which i need to fix

[valid jetty]

offset

[deep mulch]

offset

[valid jetty]

spam offset

get paige's account locked

[deep mulch]

offset

offset

offset

offset

offset

offset

offset

offset

offset

offset

offset

offset

offset

offset

offset

offset

offset

offset

offset

offset

offset

offset

offset

noo dont do that

[valid jetty]

yeah but thats mean

turn it back on

thank you

@deep mulch now for the fun part

comment out this line and look at the result

leave everything else as-is

[nimble bone]

offset

[deep mulch]

its a mandelbrot!!

how it do that

[valid jetty]

magic

mandelbrot is defined as f(z) = z^2 + c

burning ship is just that but you take the absolute value of the components of z first before you square

[deep mulch]

did you hear something @valid jetty

thought i heard something

[valid jetty]

yeah

[deep mulch]

how

Tuahmisu!

[cinder egret]

@young flicker block zt

true

[valid jetty]

true

[cinder egret]

@valid jetty

[valid jetty]

@cinder egret

[cinder egret]

@valid jetty send that essay in my dms aha

[deep mulch]

i need to fix the zoomm

[valid jetty]

@valid jetty send that essay in my dms aha

oh

[deep mulch]

i zoom in then dragging the mouse moves it so much

[valid jetty]

@cinder egret send that essay in my dms aha

[cinder egret]

dm me

do it no balls

do it or you have balls

[deep mulch]

@young flicker

didnt ask @young flicker

[cinder egret]

I have literally so many Apple products made by Apple

[deep mulch]

didnt ask @cinder egret

[valid jetty]

I have literally so many Apple products made by Apple

[cinder egret]

my dream girl

[deep mulch]

CREEP

[valid jetty]

hiiiii

[deep mulch]

CREEPLORIST

@valid jetty how do you handle the drag properly

[valid jetty]

mouse delta

[cinder egret]

drag deez LOL

@deep mulch

[deep mulch]

idk what im doing worng @valid jetty

the cursor should stay in place

over the ship

[valid jetty]

how are you calculating mouse delta

[deep mulch]

its just given to me

change.x change.y

husk if you love me

@young flicker

[valid jetty]

well then clearly whats given to you is wrong or youre using it wrong

[deep mulch]

well obviously

im doing something wrong

[valid jetty]

dx := rl::get_mouse_delta().x * (x_max - x_min) / SCREEN_WIDTH;
dy := rl::get_mouse_delta().y * (y_max - y_min) / SCREEN_HEIGHT;

x_min -= dx;
x_max -= dx;
y_min -= dy;
y_max -= dy;
``` idk

[valid jetty]

this doesnt seem scaled by (max - min)

[deep mulch]

i only have offsetX and offsetY variables that define the cetner coord

i think

detectDragGestures { change, dragAmount ->
    change.consume()
    val dragDeltaX = dragAmount.x / zoom
    val dragDeltaY = dragAmount.y / zoom

    offsetX += dragDeltaX
    offsetY += dragDeltaY
}

[valid jetty]

you shouldnt be tracking zoom but instead scaling your min and max by the appropriate zoom

and then your drag deltas should be the same no matter what the zoom is

[deep mulch]

wouldnt minX always equal -maxX

oh

yea that works

close up of the ship

[valid jetty]

yayyyyy

awesome

[deep mulch]

i get this insane error I havent figured out with my scripts

Caused by: java.lang.ClassCastException: class Script_composable$Shader$BURNING_SHIP cannot be cast to class Script_composable$Shader (Script_composable$Shader$BURNING_SHIP is in unnamed module of loader org.jetbrains.kotlin.scripting.compiler.plugin.impl.CompiledScriptClassLoader @73f717da; Script_composable$Shader is in unnamed module of loader org.jetbrains.kotlin.scripting.compiler.plugin.impl.CompiledScriptClassLoader @459048ea)
@valid jetty

[valid jetty]

here is my elle render

[deep mulch]

make it 144 fps

@valid jetty what if you animated some property of it

[valid jetty]

chaos

i do wonder what happens if you change the starting position

[deep mulch]

are the min/max complex plane stuff

[valid jetty]

yeah

oh wow

changing the start position z_0

[deep mulch]

roieeee 😭

[valid jetty]

btw i tried controlling the value of c earlier and it resulted in just solid colors and was useless

hmm what else can i control

ooooooo

the exponent

[deep mulch]

its so weird zooming into one of these

since it repeats

[valid jetty]

lol yeah

[deep mulch]

@valid jetty how far can you zoom in

[valid jetty]

as far as the precision of a 32 bit float will take you

lol

[deep mulch]

funny colors

i tried 16384 iterations its very laggy as expected

[valid jetty]

weirdly enough as you increase the number of iterations the eiffel towers start to fade out

[deep mulch]

yop

it gets all stretched if i keep zooming too far

[valid jetty]

yeah thats the 32 bit float exploding

[deep mulch]

is 64 bit possible

[valid jetty]

not in shaders i dont think

[deep mulch]

it would have to render on cpu wouldnt it

[valid jetty]

actually apparently you can use doubles

husk

ok well

[deep mulch]

do

[deep mulch]

what value is changing there

[valid jetty]

starting position

z_0

[deep mulch]

i dont have a z_0

[valid jetty]

you do

this is set to vec2(0) for you

[deep mulch]

oh

just say z

[valid jetty]

husk

im going to sleepm,,

[deep mulch]

niniii thank you for helping

[valid jetty]

@deep mulch https://cdn.discordapp.com/attachments/542264318465671170/1428277628837695498/Screen_Recording_2025-10-16_at_8.04.37.mov?ex=68f1eab2&is=68f09932&hm=f9b9e77037f4933718ce47274759887ebdb3ed34862a2b79e033d2d56e514ca7&

[valid jetty]

@deep mulch https://cdn.discordapp.com/attachments/542264318465671170/1428281549438320700/Screen_Recording_2025-10-16_at_8.19.57.mov?ex=68f1ee59&is=68f09cd9&hm=b192cb29d7d7ff9d5d97046e7fafcd1f396f792055b1761b663808d2446fae96&

[sweet pebble]

@solemn rock @wanton osprey @royal nymph
I love ur loading quotes plugin but can i customize the quotes?

i know i can if i custom download into documents type thing for custom plugins, but cant otherwise?

[royal nymph]

did you really have to ping 3 people for this basic ass question

dumbass

[sweet pebble]

damn bro my fault

[royal nymph]

rules 9, 10 and 12

[runic sundial]

https://cdn.discordapp.com/attachments/1015060231060983891/1427921232875098213/comfytocozy.gif

[winged mantle]

should i just switch from solid because i'm too noob

is that a good reason

reactivity seems to be lost across For each boundary

wait no... across component boundary

oh i'm dunmb

there are so many things you need to wrap in closures

[stark timber]

hi

[stark timber]

u could trial and error but if u want smth easier then I think it’s fair

[winged mantle]

idk idk all frameworks are kind of magic

[stark timber]

what was your first one

[winged mantle]

i just used vanilla js for years

[stark timber]

and then

[winged mantle]

well ok

i think i might have used jquery a bit

then solid when i actually needed a... solid web application

as doing in vanilla js seemed too hard

[stark timber]

alright

well idk u could try react it’s pretty simple and the most popular

or svelte is good too

[winged mantle]

i think i mainly chose solid because of smaller bundle size

also global state seems easier

you can just have a global signal lol

i did manage to figure out my issue

[crude star]

didn't react get a whole lot smaller recently

[stark timber]

why

[winged mantle]

you can just create a global variable

so convenient

[stark timber]

you mean this stuff?

vue is kinda like that in the same way and it’s easy af

vue was my first framework

it’s good but annoying for big projects

[winged mantle]

the main thing i keep running into is instead of this

let name;
if (props.game === "undertale")
    name = "Toby Fox";
else if (props.game === "deltarune")
    name = "Foby Tox";
else
    name = "unknown";
return <div>{name}</div>

you need to do this

const name = () => {
    if (props.game === "undertale")
        return "Toby Fox";
    else if (props.game === "deltarune")
        return "Foby Tox";
    else
        return "unknown";
};
return <div>{name()}</div>;

and i don't really understand... why... it's hard to find resources explaining things

with the former it won't properly react to property changes

i guess it does make sense sort of because fine grained reactivity

but i feel like i don't really understand what's hapenning still

then again i guess most react developers probably don't learn what's going on under the hood and it also has confusing quirks

[stark timber]

idk seems like that means the component isn’t rerendered when the props change

at least not in a way I’m familiar with

[winged mantle]

i think solid probably relies on a lot of compiler magic

doesn't react have a compiler but its optional

[royal nymph]

same way you cant destructure props

[winged mantle]

i think the main issue is unlike react where it tells you when you break the rules solid doesn't

[crude star]

the components function only ever runs once

[royal nymph]

you can do this with react too, just have to write a smell helper or use a contextprovider

[stark timber]

yea ik

[royal nymph]

or smth like zustand

[crude star]

but <>{this}</> creates a rendering scope

so that part in {} basically gets turned into a function

that's what the solidjs magic is

[hoary sluice]

u mean the auto imports?

[stark timber]

with the store concept

[ionic lake]

here name is calculated once on the component, the accessor is what keeps it reactive

[ionic lake]

yes, take a look at dom-expressions

[winged mantle]

ah yeah thats the thing

projects i return to after not touching for months my beloved

[solemn rock]

I think you can

I remember writing this feature.

[deep mulch]

@valid jetty I made a Julia fractal last night

I was trying to get variables passed in but didn't finish

theoretically it's possible to make my script project run in the browser

[spark tiger]

what is that fps counter i keep seeing it

[deep mulch]

it's raylib I think

[valid jetty]

yayyy

[valid jetty]

[deep mulch]

I should add a fps counter

[spark tiger]

oh

[deep mulch]

faints per second

[spark tiger]

idgi ❓

[valid jetty]

i made a flag to turn burning ship into mandelbrot and look at this

everywhere there is a ship, there is a corresponding mini mandelbrot bulb

[lucid trail]

thats really cool its so smooth

[valid jetty]

yeah its awesme

[jade stone]

The react compiler pretty much just helps reduce boilerplate

[young basalt]

im pro skidder

tuff right

[spark tiger]

✅

[young basalt]

ty

[spark tiger]

how do i make game loop run at like consistent rate. do i like do sleep or

while (1) {
  float delta = curTime - lastTime;
  if (delta < UPDATE_RATE)
    continue;
  update();
}

[spark tiger]

fuhh off gng

[valid jetty]

sleeping is imprecise

[spark tiger]

what about nanosleep or same thing

[valid jetty]

look at how raylib does that

https://github.com/raysan5/raylib/blob/master/src/rcore.c#L920-L1034

[spark tiger]

or what’s it called

[spark tiger]

tyy

[valid jetty]

specifically this bit

raylib's wait function uses nanosleep yea

[spark tiger]

oh

[crude star]

that's going to use 100% cpu

nanosleep is fine

[spark tiger]

ic ty

though it seems like there isn’t nanosleep on windows

[lucid trail]

https://blog.bearcats.nl/accurate-sleep-function/#timers-on-windows windows has some interesting stuff. They also have a method for partial busy waiting above

seems good

[spark tiger]

guh that sounds like too much work. i mean i could just blindly copy paste it but id rather figure out how it works. i can just uh not use windows

[nimble bone]

im pro skidder

tuff right

[spark tiger]

oh seems like raylib just uses Sleep()

[lucid trail]

figure out how it works
doesn't the post explain it? or do you mean waitable timers

[spark tiger]

oh i thought the part above was irrelevant so i thought it was only those two paragraphs