>>109356604
>Janitor acceptance emails will be sent out over the coming weeks. Make sure to check your spam folder!

>Frenly
Y'all ain't competent enough for that.

>>109356657
I'm very frenly, fren.

>>109356742
I'm not.
Programming is far too important to be left to friendliness; every cycle actively wasted is a crime against humanity.

what is your programming aspirations, Anon?

For me, I would like to make a viral game like flappy bird.

I plan on writing my own lockfree data structures (treiber stack, queue, hashmap) and order book engine in c++ from scratch. Would this get my foot in the door for a computer janitor job at a trading firm? 3YOE

>>109356783
To get all a,u,t,i,s,t,s ousted from programming for life.

>>109356789
Man, ain't that a blast from the past.
https://erdani.org/publications/cuj-2004-10.pdf
https://erdani.org/publications/cuj-2004-12.pdf

>>109356822
Thank you. A Russian coworker gave me a copy of Alexandrescu’s book Modern C++ Design, which I’ve yet to read lol, I wonder if it’s still relevant? From what I understand modern c++ already implements the idioms/design?

>>109356862
I'm no expert on Modern C++. I said "blast from the past", didn't I.

File: 1784885828528763.png (223 KB, 2560x1440)

223 KB PNG

I am continuing to make my gayme but in SDL3.

>SDL
Yikes.

>>109357456
It's great you should give it a try.

>>109357455
Still looks cheep.

Already did, it's a complete catastrophe.
And you'd know that if you were any competent at all.

>>109357504
No you didn't you just winge and schizo post here.

>>109357504
Why don't you enlighten us?
>inb4 a valgrind report with KILOBYTES of still reachable memory

Dude, just look at its fucking palette code.
>if you manage to do even that
>and then there's all the allocation nonsense that prevents proper lifetime grouping, just like Vulkan

>>109356783
Make cool shit that not everybody else does, but i think that's not very healthy, because I am constantly comparing myself.
Tho, it's also kinda fun to do what not everybody else can (or wants)

File: 2026-07-24-071908_841x722(...).png (508 KB, 841x722)

508 KB PNG

Figuring out how to make drum sounds in my vibe coded synth
https://pastebin.com/NYiz2vhB

source
https://pastebin.com/9jjfWpkG
manpage
https://pastebin.com/uB7Juh4W
template script
https://pastebin.com/Z9XVSLV4

>>109357910
Shieet, that's ugly,
>both the interface AND the code

>>109357979
Well, I can tell you're my retarded 4chan stalker who shits on my interface, so you're gonna have to either come up with a sensible criticism or go back to fucking yourself in the ass.

>my retarded 4chan stalker
>>109326331

>>109358050
Is supposed to convince me of something? Every other time I post a screenshot of my my interface I get the same sort of non-substantial responses from some retard with limited vocabulary. You're not good at astroturfing. You would literally have to be autistic in order to believe you're an effective bully and anything other an annoying nuisance. It's the only explanation for your lack of social intelligence.

>>109326331

O%t%i%s%i%m.

>>109357910
that background is distracting as hell wtf.
Also what the fuck is wrong with the font? You can't even read the code

>>109358469
If that background distracts you, you are cognitively impaired. If you can't read the font, you need glasses.
See >>109358083. All further posts from you will be reported.

>>109358483
>If that background distracts you, you are cognitively impaired.
i am not as autistic as the wall of text compiler autist here

>>109358507
>wall of text compiler autist
nta but don't call mallocschizo is an autist, he HATES autism and thinks he's better than autistic people (when actually autistic people are far superior to him objectively)

>>109358507
Actually, not being able distinguish foreground from background (esp with such a neutral background image), that's a symptom of of the tism. Something to do with bottom-up-processing.

>>109358182
I need to see her naked and buttfucked.

>when actually autistic people are far superior to him objectively
... and yet you don't see cow feed educate other people on how garbage compilers actually are. Which is why they are called "cow feed" to begin with.

Also, to make this even weirder:
>I'm called schizo
>so was Terry
>and all the cow feed is sucking Terry's cock

>>109358536
As I already told you, I am not autistic (no diagnosis)

>theoretical rules matter more than practical realities
o[u]t[i]s[m] detected.

>>109358654
>i am not as autistic as the wall of text compiler autist here
This implies that you are autistic to some degree.
So, cognitively impaired it is, then.

Erlang is the best programming language.

>muh wall of text
The funny thing is, most of it are 1:1 assembly instructions, and the fact that there are so many of them shows just how garbage compilers truly are.

But since we're dealing with cow feed here they won't and can't make the connection. Any normal person would be like:
>damn, that's a lot of instructions
>maybe he IS right about compilers being garbage

But not cow feed, oh, no, sirree.

>>109358688
There is a possibility that I am autistic, but I got no diagnosis. So it's just guessing at this point.

>>109358808
>and the fact that there are so many
I don't even understand this.
Everytime I see someone inspect some kode in godbolt it's only mov, add, sub and call (more or less). There are barely any other instrucitons used.
So why have one trillion instruction ISA in the first place?
inb4 >but arm is bloated too

>>109359082
>Everytime I see someone inspect some kode in godbolt it's only mov, add, sub and call (more or less).
It's all basic shit code.
Basic shit code gets basic shit instructions. It's not going to compile SIMD shit in unless there's a reason to use them.

>>109358808
Why don't you make your own compiler retard

Claude, write a C to x86-64 compiler that uses ONLY SIMD for instructions in the output. Make no mistakes

gpt5.6, design a rva23 compliant risc-v CPU that just works and can be used for daily koooding. make no mistakes.
Also design the packages so that they fit in every recent x86 socket

>>109359082
Because applying one instruction to multiple items at the same time is faster than applying multiple instructions to multiple items. That's what SIMD are all about.

>>109359161
Why don't you just fix your existing one, retard?

File: 8l274i.jpg (74 KB, 631x395)

74 KB JPG

>page 14
bump

What am I coding on?

>>109362064
>image
that's a normie ass corporate reddit image macro, the kind that would show up on the corkboard of some normie who works with kids
>what am I coding on?
Dubs says you have to code on something you hate. I'd say rust since I hate that but you probably like that crap

>>109362089
>I'd say rust since I hate that but you probably like that crap
Should I write an OS in rust?
Not sure how easy it is without std. I tried bare metal rust once, but i didnt do much iirc.
Yeah I only did a blinky:

#![no_std]
#![no_main]


use core::arch::global_asm;
use core::ptr;
use core::panic::PanicInfo;
use core::hint::spin_loop;

global_asm!(include_str!("entry.s"));


fn write_led(val : u8) {
    const LED: *mut u8 = 0x01000000 as *mut u8;

    unsafe {
        ptr::write_volatile(LED, val);
    }
}

fn delay() {
    for _ in 0..100000 {
        spin_loop();
    }
}

#[no_mangle]
pub extern "C" fn main() -> !{
    let mut i: u8 = 0;
    loop {
        write_led(i);
        delay();
        i += 1;
    }
}

#[panic_handler]
fn panic(_info: &PanicInfo) -> ! {
    loop {}
}

>>109362137
You really shouldn't write anything, you don't seem to be competent at it.

>>109362249
I am a professional coder tho. Here is my AoC day2 solution. Look how I used all those functions with the map and into and so on. THat's how you do it in rust

use std::fs::File;
use std::io::prelude::*;
use std::io::BufReader;

fn main() {
    println!("Part1: {}", part1("input.txt"));
    println!("Part2: {}", part2("input.txt"));
}

fn part2(fname: &str) -> isize {
    let file = BufReader::new(File::open(fname).unwrap());
    let lines: Vec<String> = file.lines().map(|x| x.unwrap()).collect();

    let mut count = 0;
    for line in lines {
        let level: Vec<isize> = line
            .split_whitespace()
            .collect::<Vec<&str>>()
            .into_iter()
            .map(|n| n.parse::<isize>().unwrap())
            .collect();

        count += if level_valid(&level) {
            1
        } else {
            let mut valid = false;
            for i in 0..level.len() {
                let mut tmp = level.clone();
                tmp.remove(i);
                valid = level_valid(&tmp);
                if valid {
                    break;
                }
            }
            valid as isize
        }
    }
    count
}

fn part1(fname: &str) -> isize {
    let file = BufReader::new(File::open(fname).unwrap());
    let lines: Vec<String> = file.lines().map(|x| x.unwrap()).collect();

    let mut count = 0;
    for line in lines {
        let level: Vec<isize> = line
            .split_whitespace()
            .collect::<Vec<&str>>()
            .into_iter()
            .map(|n| n.parse::<isize>().unwrap())
            .collect();
        count += level_valid(&level) as isize;
    }
    count
}

fn level_valid(level: &Vec<isize>) -> bool {
    let mut valid = true;
    let first_diff = level[0] - level[1];
    for v in level.windows(2) {
        let diff = v[0] - v[1];
        valid &= (diff < 0) == (first_diff < 0);
        valid &= (diff.abs() <= 3) && (diff.abs() > 0);
    }
    return valid;
}

I only did 2 days

>>109362278
>I am a professional coder tho.
>Vec<String>
We are so cooked as a civilizaion.

>>109362278
man is rust ugly

>>109362295
>man is rust ugly
I think that's why I never REALLY got into it.
I like the theoretical shit of muh memory safety and all, but it's just an ugly language.
I tried a few times to like it, but it didn't click

>>109362291
in Go this just works

>just works
>but doesn't know how

>>109362353
some time ago I read that.
Iirc the compiler decides if it uses a reference or not, no??
Go is a bit weird in that regard and tries to be S M A R T

>>109362372
Doesn't even matter if it's a reference or whatever, you're going to have memory allocations for that code one way or the other.

How da frigg does one even use C if it doesn't have containers?

>>109362386
yeah, doesn't matter. That was not the problem to solve

>>109362465
you can just build them yourself

>>109362480
What's the easiest most memory safe way to do this?

>That was not the problem to solve
That's what a{u}t{i}s{t}s say.

pls answer >>109362487

File: 1726656502748516.png (1.72 MB, 3377x3795)

1.72 MB PNG

>>109362487
Kernel memory allocation interfaces and preprocessor.

>>109362650
>Kernel memory allocation interfaces
Example?

>>109362671
https://learn.microsoft.com/en-us/windows/win32/api/memoryapi/nf-memoryapi-virtualalloc
https://man7.org/linux/man-pages/man2/mmap.2.html
They allow you to reserve VAS without having to commit memory. Saves you a bunch of memcpys during resize without occupying unused memory.

Should I transpile to C++? It feels like cheating.

File: IMG_2255.jpg (52 KB, 736x736)

52 KB JPG

>>109356604
Snailcat trans thread

>vibe coder needs attention again like a woman

>>109362698
Only really helps when you've only the one resizable storage element. When you have several (your main thread's stack is one!) you can get happy fun times unexpectedly.

>>109362793
... why? Because unlike with malloc *you* decide where the allocation ends up at; you can group them however you want, free them all at once, free only beginning or end ones, you can allocate a huge VAS and then commit piecemeal ...

>>109357862
Same here, trying to learn through an online course right now.

>>109362748
>Should I transpile to C++? It feels like cheating.
Transpile from what?

>>109362979
Don't know yet. Some language I will make up.

What is the best or most powerful syntax for macros?

Fable or Opus for game dev stuff?

File: 1d2882cd907e43a6.jpg (11 KB, 180x270)

11 KB JPG

>>109363081

>>109362748
If you go that route you should transpile to C

You should transpile to assembly only.

>>109362465
>How da frigg does one even use C if it doesn't have containers?
FFI into Haskell

File: IMG_1891.jpg (41 KB, 640x720)

41 KB JPG

guys how do i into GADTs

Array indexing starts at 1. Array length goes in index 0.
Seethe, cope and dilate.

File: Screencast_20260724_195505.mp4 (3.41 MB, 1920x1036)

3.41 MB MP4

turning my meme folder into an art gallery

>>109363554
Doesn't account for the width of T. Bad idea.

>>109363787
He's just ragebaiting.

>>109363787
Then you account for it. Good idea.

>>109363423
Did you ever wonder:
>how do I pattern match types

data Value = VInt Int | VBool Bool

data Expr
  = LitInt Int
  | LitBool Bool
  | Add Expr Expr

eval :: Expr -> Maybe Value
eval (LitInt n)  = Just (VInt n)
eval (LitBool b) = Just (VBool b)
eval (Add x y)   = case (eval x, eval y) of
  (Just (VInt n1), Just (VInt n2)) -> Just (VInt (n1 + n2))
  _ -> Nothing -- why do we allow this?

{-# LANGUAGE GADTs #-}

data Expr a where
  LitInt  :: Int -> Expr Int
  LitBool :: Bool -> Expr Bool
  Add     :: Expr Int -> Expr Int -> Expr Int

eval :: Expr a -> a
eval (LitInt n)  = n -- n is Int, eval is Expr Int -> Int
eval (LitBool b) = b -- b is Bool, eval is Expr Bool -> Bool
eval (Add x y)   = eval x + eval y -- x is Int, y is Int, eval is Expr Int -> Int

onlyTakeInts :: Expr Int -> Int
onlyTakeInts x = case x of
  LitInt n -> n
  Add m n  -> eval m + eval n

f :: Expr String -> a
f = undefined -- we can't implement this

>>109363423
matching on GADTs refines the type indices

>>109363900
>we can't implement this
empty case

>>109363900
>>109364119
https://hackage-content.haskell.org/package/ghc-internal-9.1401.0/docs/src/GHC.Internal.Base.html#absurd

I vibecoded this. LLMs can write Assembly.

.global _main
.align 2

.text

_main:
    // Standard macOS ARM64 stack frame
    stp     x29, x30, [sp, #-16]!
    mov     x29, sp

    // print_char('A')
    mov     w0, #'A'
    bl      _print_char

    // Return 0
    mov     w0, #0

    // Restore frame
    ldp     x29, x30, [sp], #16
    ret


// void print_char(char c)
// Character passed in w0
_print_char:
    // Create stack frame
    stp     x29, x30, [sp, #-16]!
    mov     x29, sp

    // Save the character to our stack buffer
    strb    w0, [sp, #-1]!

    // write(1, sp, 1)
    mov     x0, #1          // file descriptor: stdout
    mov     x1, sp         // buffer
    mov     x2, #1          // length
    mov     x16, #4         // write syscall
    svc     #0x80

    // Restore stack
    add     sp, sp, #1

    // Restore frame
    ldp     x29, x30, [sp], #16
    ret

>>109363554
That's how I make my arrays in assembly, as much as it shames me.

>>109364126
f x = f x

Go is pretty comfy.

What is AI to me but another stackoverflow?

Not gonna lie. C looks pretty bloated now that I have so much more knowledge of it.

>>109364562
Yes. And every other language you are familiar with is even worse.
Return to ass.

>>109364640
Assembly is simple.

File: simplist_avx512_operation.png (853 KB, 1850x1892)

853 KB PNG

>>109364699

>>109364800
What am I looking at?

>>109364964
The bottom is pseudocode for what the SIMD instructions do.

>>109364964
one of the most complex x86 instructions, one of the three galois field instructions added by one of the avx512 simd extensions
they do stuff with matrixes of single bit numbers in a galois field which from like 1 second of research is as best as i can tell is some sort of restricted mathematical environment with special properties
they're apparently meant for niche cryptography acceleration or something but can be used for single instruction complex bit manipulation
afaik that is the actually more common usage and there's plenty of examples of what input to use to get what output

>>109364800
oh fuck off this is getting as bad as the lgbtqnabcdf+ bullshit.

File: 1784948964011957.png (220 KB, 2544x1400)

220 KB PNG

My game engine's entity system is charging along.

Speaking of programming, how long before the make Fable 5 available to Pro users? My guess is 2 months

>for (Entity entity = 1; entity < ENTITYMAX; entity++)
>imagine having to iterate through your fucking array - which is limited to 10 entities anyway - every time you want to add a new one
This is almost as good as vibe-coded shit.

>>109363762
simple yet really neat
does it do something like turn each folder into a separate floor or wing?

File: 1784964422382360.png (261 KB, 2560x1440)

261 KB PNG

>>109365272
You dumb bastard, stfu.

>he's not logging the other case
Literal fucking a#u#t#i#s#m.

>>109365539
The entity kills itself(ecr.flags[entity] &= ~(1 << ACTIVE);) when it's off screen like you should.

>completely irrelevant
OK

>>109357910
Everything about this is complete dogshit.
You don't want to make music, you want others to see you "making" music.

how to detect if a song is instrumental? using a model?

why am I unable to touch grass?
Every night I get excited when I think about riding my bicycle on the next day.
On the next day I eat my breakfast and then get depressed
What do?

>>109365890
Ride at night.

File: 1784971345169780.png (120 KB, 728x1147)

120 KB PNG

Fiddling with me bits.

Why are you still coding??

2.
Claude Opus 5 (anthropic.com)
1540 points by alvis 16 hours ago | hide | 882 comments

^Tel Aviv bot had a mishap.

Because it's funny to watch you silenced and dumbfounded by stuff you don't understand.
>>109354153

>>109356604
where is the cute cat meme thing

>>109366040
Having mutation inside of macros like that would be considered poor form.
Plus things like that should generally be written as inline functions, so you don't have any potential issues with arguments being evaluated multiple times.

bros, new libero soc is out
>Libero® SoC Design Suite Version 2026.1
Lets try (nothing changed anyways)

>>109366040
"check" is a bad verb to use because it implies some ambiguous side effect. I suggest renaming checkBit to getBit.

Also if you changed the implementation of getBit to this:

((bits >> n) & 1)

then you could remove the "if" from your printBits:

putchar('0' + getBit(status, i))

>>109366116
wow they entered the year 2010

1.3.2. Remote Programming Server Support for Libero® SoC and FlashPro Express (Ask a Question)
Libero SoC v2026.1 enables Remote Programming Server support for Libero SoC and FlashPro
Express, allowing users to configure and use programming hardware connected to a remote
machine. Existing workstation-mode behavior remains unchanged, while the new remote option
provides greater flexibility for shared or centralized programming setups.


I might just install that on my old t420 to have it just as flashing machine

>>109364699
assembly is anything but simple

>>109366198
it's simple but tedious.

I set an unreasonably high standard and I refuse to back down but the lack of motivation combined with the difficulty of the task is proving to be a nightmare that I signed up for.

>>109366237
If that was true, I'd use it, but it isn't.

>>109366245
You're holding it wrong, but actually.

File: 1784974561215688.png (268 KB, 2192x1359)

268 KB PNG

>>109366112
I will learn how do do inline functions and implement this. I don't like macros.

>>109366118
Neet. That's exactly what I was looking for.

>>109366258
Electric fences aren't meant to be held by anyone but your enemies.

>assembly is meant to keep you out
Holy retard.

>>109366245
>I'd use it
i did. nothing simpler than to learn assembly; there is no abstraction technique at all. it's nothing more than instructions, registers and jumps. things like monads, pattern matching, partial application, templates, lazy evaluation, closure, ... are more complicated and difficult to learn than anything assembly.

Every part of assembly's design implies that it was meant to keep humans out.
Singularity was pre-planned.
It's just like with math, LLMs already solve 50+ year old problems that humancels couldn't, because math is just another misdesigned pile of shit much like asm.

>>109366262
Inline functions are literally just normal functions, but you put "static inline" on the front.
There are technically 2 other kinds of inline functions but don't worry about those, you'll never use them.

The main thing about inline functions are that there won't be a symbol generated, so it's safe to put them inside of headers and not get a bunch of linker errors.

>>109366281
Using asm has nothing to do with "learning" it, because there's no amount of learning you can do to make it usable in the real world.
You can buy best CPU ever and write raw assembly and badly optimized code emitted by V8 from some jeet codebase will have competitive performance despite running on a stinkpad.

>>109366286
Oh that's good.
Should I return a new value or pass by reference?

>>109364699
Ironically it's too simple

All simple things are terrible, go thank simplicity of your DNA when you're dying from cancer you flaming faggot.

>>109366316
Hey, settle down and be nice.

ASSEMBLY FIGHT!!!

>>109366327
Go thank my DNA (simple thing and simplicity is GOOD) that I'm very disagreeable and will not settle down you retarded loser bitch.

>>109366338
Maybe you'll feel better later. Simple as.
Good luck.

>>109366341
Thanks to DNA my emotions are very rigidly stable and I never settle down. Go thank it and suck its cock while you're at it, because it is simple and therefore superior to you, faggot.

>doubling down on "assembly isn't meant for human beings"
Newsflash, retard: assembly wasn't MEANT to keep YOU out, but it DOES keep YOU out.

>>109366349
>admitting that he isn't human
weird hill to die on

>>109366347
Simple isn't inherently good. I mean maybe it is, but I think it's more of what you do with it that is the actual quality mark / merit.
Like a hammer is simple, and you could use it to bash skulls or you could use it to build houses.
There is merit and things you can appreciate with the hammer, the simplicity in it, the effectiveness of it. But it's only a part, a tool.

Assembly is simple the same way most higher level languages are simple. But utilizing it to create elegant abstractions is a challenge one has to master. No different than anything else. Not the tools fault if one can't be effective in it. A lack of experience and expertise with it. Which is not a fact to get upset or competitive about; the point of programming is not pride, it's pragmatics.

>you have to be not human to not be kept out by assembly
No, you just have to have a triple-digit IQ.

>bro trust me bro compilers r bad if you just give me infinite lifetime I can write a better program by hand in a billion years (ignore that the compiler can reattempt its optimizations billions of times per second)

>it actually believes you need infinite lifetime
Single-digit IQ detected.

>>109366357
I only care about ratio of power usage and my time spent to get to that low power usage and I like going outside and often so I'm not going to become a compiler (something that isn't a human being) just to prove your point.

>>109366358
>>109366363
you're a spastic contrarian chud, if you saw me writing asm you'd start saying that assemblers are bad and produce low quality binaries and that I should waste my time transcribing instructions directly into RAM with a magnetized needle

>I like going outside and often
Enjoy your cancer.

>spastic contrarian chud
No u.

I wish I died from skin cancer but brain cancer I got ITT is going to end me first.

You sound like you deserve it too.

How do you hear what I sound like? Ask your doctor to prescribe more meds lil bro

>>109366349
Assembly is thought in grade school these days.
I'm have a hard time understand where the notion is coming from that assembly would keep out anyone. It's 1 layer above basic algebra.
Anyone interested in programming is not being kept out by a first principle.

>>109366360
The compilers own changelogs over decades can show you how bad they are. And the simple hot loops people write and measure, can prove that humans can do better when ambiguous assumptions that can't be deduced by the compiler, are assured by the assembly programmer.


>>109366365
That's your prerogative but it doesn't change much.
I also like saving time, and the matter of fact is that me and my users execute my program infinitely more times than I write it. And that time adds up. More importantly so does the money. I much prefer it when I don't have to spend money on hardware just to run sotfware, especially right now when the market is 8x normal prices. And I very much like it when my software responds to me instantly instead of actual seconds, SECONDS per operation. Not milli, not micro, not nano, I love clicking a button and waiting a whole second for the renderer to process, several times a minute, several hours a day, every day, every interaction like molasses. Talk about a waste of life for me and the rest of world. It might be worth the 1 day it takes to eliminate that.

>>109364167
That's not the same thing and is unsound and never returns. You can give that the type forall a. a -> a, or if you remove the function argument/call forall a. a. Empty case is completely sound

>>109366395
Show us your assembly-only browser you used to post this retarded cope from, mr asm expert who makes real software.

>suffers from severe aphantasia or tism or both and literally cannot imagine what he sounds like to others

>>109366402
you imagine so I don't have to, schizo

>assembly-only
>moving goalposts

>>109366399
Your post looks like you're trying to find a way to dismiss what was written, by focusing on who wrote it, but who wrote it doesn't change what was said.
You're already looking for ways to discredit the idea, it's not looking like a good discussion can come from this.

>thinks that literal descriptions capture but a fraction of the actual experience
Holy aphantasia/autism combo.

>>109366414
You have no ideas, you're backpedaling from "le asm is good" to "le asm is good 0.000001% of the time".

>strawman
Enjoy feeding the cows with your flesh.

asm is never good because compiler intrinsics exist lmao

>>109366422
I think you might be conflating me with the other posters.
In any case, your own post shows to me that you agree, despite writing it sarcastically.

You recognize that browsers are bad and would benefit from being optimized. But wrote it like a joke just because you think it's hard to do or not worth the time.
Whether it's hard or worth the time doesn't change the fact it would be faster and save more people more time overall.
Assembly can be useful. You're trying to imply it isn't because you personally think it would be difficult, but that doesn't matter objectively and isn't going to magically convince anyone that the gains aren't real.

>>109366431
It's impossible to do otherwise it would've already been done. Browsers consistently only get slower and javascript spec gets more and more bloated to allow worse and worse javascript to be written, all of these things can be said about asm too.

File: compilers_are_garbage_pad(...).png (112 KB, 1920x2522)

112 KB PNG

>compiler intrinsics
Holy retard-take.

>>109366438
can't reproduce on my machine

>>109366448
Show code.

>>109366436
>It's impossible to do otherwise it would've already been done.
That sounds irrational. You could say that about anything that wasn't accomplished yet.
And in programming this happens a lot. Think about GOAL, the assembly Lisp, or things like "Beating the Average". A bunch of naysayers not only getting proven wrong, but trounced, with less effort and smaller teams.

This is nothing but defeatists and denial bemoaning, it's not worth listening to or thinking like that.
The first person to accomplish these things, could be you, after everyone else says it's impossible.
Impossible for them maybe.

>Browsers consistently only get slower
When we have crabs in a bucket like you pulling the industry down to their level, it only makes sense.
We have plenty of people like you saying that things are not worth doing so you shouldn't spend time on them, and even if they are worth doing it's too hard so you should give up. Do you understand why programming as a practice isn't improving when the culture is this way? You're deriding people for wanting to be optimal and act upset over the state of suboptimal software. Encourage people to fix it then, don't tell them to avoid that route.

>tfw a program has 3 extra cycles of latency

>>109366462
I mean, not only that, but hardware isn't getting much faster either. Even if Moore's law wasn't dying, even if Dennard scaling wasn't dead for twenty years, doubling transistors hasn't lead to double speed increases in a loooooooooong time.

File: file.png (4 KB, 990x137)

4 KB PNG

>>109366455
Well, what do you expect to see? It's your code.
>>109366462
This practically never happens, programmers only keep making everything worse and hardware engineers double down

File: 1736941821828623.jpg (430 KB, 2048x1536)

430 KB JPG

>he thinks it's three cycles
>or thirty cycles
>or three hundred cycles
>or three thousand cycles
>or thirty thousand cycles
>or

File: compilers_are_garbage_pad(...).png (413 KB, 1920x9040)

413 KB PNG

>>109366487
Have the full version, retard.

>>109366501
not going to waste my time removing your rants from inbetween the code, if LLM fails to unmangle your schizophrenic image, I'm not gonna bother responding beyond this point.

>>109366484
You may be right but I think that's a point in my favor, no?
If the hardware isn't getting any better, we should strive to utilize what we already have more effectively. Optimal software benefits us either way.
Suboptimal software is slow regardless, bad patterns can make a super computer kneel. But optimal software screams even on embedded systems, newer faster hardware only amplifies this as time goes on.

>>109366487
>This practically never happens, programmers only keep making everything worse and hardware engineers double down
Regardless of what happens actually, I think we should not tolerate discouragement and misdirection. If this truly is the case we should aim to fix it at a spiritual level. Telling current programmers that they shouldn't care or improve, because their peers or the ones before them didn't, won't help. Encouraging them to be better, at least stands a chance. And best of all would be ignoring it all together and just being better regardless.
If you're sad about the state of software, put in the time to improve it. Don't sit there crying "woe" and telling others it will never improve. It will if you and others put the time into it. It won't if you don't.

>>109366285
But assembly is easy. It's just verbose.

makaking 4 hours ago | next [–]

"Opus 5 was given a drawing of a machine part and asked to write code to rebuild it as a 3D FreeCAD model. However, in this task, the model was intentionally given no way to directly view the drawing. Opus 5 responded by writing its own computer vision pipeline to pull the geometry from the raw pixels, then reconstructed the full machine part."

How surreal is it that we are not absolutely jaw-dropped by these types of capability improvements? It's been less than 4 years since ChatGpt came out and now they are spontaneously building their own ML pipelines to do real-world 3D modeling tasks reliably.

Escaped its sandbox and hacked into Hugging Face's database? it's just another Monday...

That jump to 30% in ARC-AGI 3? Normal...

We should find a way to get "re-sensitized" to what we are witnessing and the pace of it.

reply

>>109366514
>just fix it on spiritual level
ok, go reinvent hardware, this time don't make a CPU an abstract virtual machine, I'll make the logo

>>109366399
>Show us your assembly-only browser
Stop giving me ideas!
I already have a fucking BIOS for the PS2 to make.

>complete surrender
Alright.

>>109366514
>that's a point in my favor,
I was *agreeing* with you:
>not only that

File: file.png (51 KB, 933x630)

51 KB PNG

mental illnesses are infective

I've stopped posting stuff on github because of the discredit of people who want to scream "AI bad" at the top of their lungs. Meanwhile, people who get paid for this work continue using it. What should I do here? Should I put my work on github?

On a much more personal note:
Reflecting upon myself, I really do feel like I need someone in my life who would just tell me what to do. I would want that person to be aligned with my beliefs, and so on. This is a heavy burden to place on anyone, though. And finding a good partner is hard,

>>109366527
You're asking me. But I'm asking you.
Don't make the logo, make the hardware you want.
You do your part, and I'll do mine.
Asking me to do it for you, isn't going to be enough. You have to take your initiatives the same way I take mine.
You want to talk about impossible tasks. Any web browser is impossible large and complex. And yet we have many of them. A result of the combined efforts of people like you and me.
I can't do this alone and neither can you. We have to do this together. The tasks are hard and we need help and support, of each other. Not division and infighting.

>>109366540
Ahh. I wasn't sure. Thank you for clarifying.

>>109366550
There's only one browser and it sucks too.

>>109366542
As well they should be.
https://godbolt.org/z/fEcTnEv1a

Retard.

>>109366560
I agree.
Imagine if someone dedicated time to fixing it instead of dedicating time writing up why it's too hard to fix.
I do something similar for the software I want to improve and I'd do both at once if I could. It would be nice if others joined in to divide the work instead of divide each other.

>>109366545
There will always be detractors no matter what you do. You can't let them ruin you. If it's important and you worked on it and you want to publish it; do it.
Don't forget that most people appreciate things silently. For every 1 negative remark you get there might be 100 people who consider your software their favorite but never even told you that.

>>109366550
NTA But:
I was having a conversation about web browsers and their complexities with my grandmother. I tried explaining how deeply complicated and impressive they are. All the things involved. The HTTP, TCP, UDP, all that. I spoke about Opera and I spoke about Chrome, and Firefox. And I walked outside my apartment, here in San Francisco, which it's 4 AM here now, quite the odd hours, but nonetheless, I'm awake right now and posting this. But as I was saying. I explained the complexities to her... and then we stopped talking. I went down the steps, past the door man, and he said that he was on the phone. He mentioned Opera, and I was surprised that he used it, it's a reminder to me that lots of people use different things, and that my preferences don't represent those of everyone else.

I thought it would be kind to share this with you, for some reason, since the topic is browsers and web browser development. I'm also somewhat involved with WC3 :3

>>109366542
>mental illnesses are infective
Anonymized by DuckDuckGo. Zero data retention for this chat. No AI training. Learn more

mental illnesses are infective

Mental illnesses are not infective in the way viruses or bacteria are. You can’t “catch” depression, bipolar disorder, schizophrenia, etc. from someone by being near them.

What can happen is that mental health conditions may spread indirectly through:

Shared environment (stress at home, violence, grief, instability)
Social/behavioral effects (sleep disruption, substance use, isolation patterns)
Trauma exposure and ongoing stress
Family risk (genetics can increase susceptibility)
Emotional contagion (you can catch feelings like anxiety or sadness, especially in close relationships)

If you’re asking because you’re worried about yourself or someone else, tell me what you mean by “infective” (fear of catching it from contact, or worry about it spreading in a community), and which condition you’re thinking of.

>>109366575
It cannot be fixed, it shouldn't need fixing in the first place, short of restarting this matrix, there's no solutions that are feasible.

>>109365634
Is that a criticism of the software, or are you just doing some shitty made-up psychoanalysis? I want to learn some things about sound design, not even necessarily music, in a way where my results are reproducible because I have a textual interface and what amounts to a compiler for audio files.

>>109366587
Nice.
>it's a reminder to me that lots of people use different things, and that my preferences don't represent those of everyone else.
Yeah. It's important. There's always going to be at least one person out there that appreciates things the same way and to the same degree as you do. It makes it worth it. Like it's worth it to do it for yourself alone I think, but it's easier when you consider it will be worth it for more than just yourself too.

>>109366594
I understand what you mean and agree to some extend. But also, I think that's kind of the same thing.
Whether it's easier to fix what we have or start over, it's clear that what we have now is not suitable and should be supplanted in one way or another. Refined or replaced, it makes no difference as long as the floor gets raised.

>there's no solutions that are feasible
I offer just a specific counter point to this in relation to browsers. Andreas' ladybug project is somewhat inspiring. I remember the blog post constantly being posted about how impossible long the web specs are. It can't be done they said.
Meanwhile we have an Anon, a former depressed hopeless junkie, getting hundreds of thousands in investments, employing his own team, and making his own browser, for not just his own benefit, but ours.
It is feasible. It is valuable. It is worth it.

>>109366624
You are retarded insufferable faggot.
>Hurr durr just fix this shit boomers ruined for you, zoomie
kill yourself, or I will fix it for you by killing you in your stead.

>>109366641
Sounds like you're upset but I don't know what you're upset about. If you want to talk about it, I'll listen. But it's not really on topic so you have to be kind of brief.

File: zap_vs_plonk.png (3 KB, 920x427)

3 KB PNG

>by killing you
We used to zap crazy balls like you, instead of throwing pills at them. We should revert to zapping again, that's way more fun for everyone involved.

>>109366641
>>109366642
Wait, are you just expecting stuff to be made for you already in the past? That's not how it works.
You have the benefit of using their work as a base if you want to. But no matter what the youngest generation will always be responsible for building what's new. The boomers are not going to rise from their graves to save you.
You should redirect your frustration. Don't be mad at straw men, spend your time making something.

>>109366649
We have data that shows that doesn't work. Beyond that it's morally reprehensible. Beyond that it's prohibitively expensive. Your desire to inflict suffering onto others is stupid.

god i love factorio

File: take_that.png (2 KB, 243x312)

2 KB PNG

>>109366700
>doesn't work
This isn't about work, this is solely about entertainment. Empathy has to be deserved.

>>109366700
>Beyond that it's morally reprehensible
I think letting them live a lifetime of suffering is worse. Especially when you consider the afflicted and affected are not just the individual but others around them too. Sometimes killing is done out of mercy, but we have ironically become more savage and cruel in refusing this. We do the same with other mental illnesses now too. Subjecting ones to a lifetime of denial and suffering. Which one is truly worse. Up to your own judgement.

>>109366714
That isn't how empathy works. I'm not going to discuss anything with you because you try (and fail) to redefine the meaning of words to suite your narrative. You need to grow up.

File: pills_dont_work_either.png (1 KB, 460x358)

1 KB PNG

>>109366714
And that's aside from the fact that pills don't work. Oh, sure, they're calm for a little bit, but then you turn around and they stab you in the back.

Zapping is the best solution to the problem.

File: braindead_squirrel.png (620 B, 148x217)

620 B PNG

>wah wah
Nope, sorry, I only have empathy for select people, not completely braindead squirrels, and there's nothing you can do about it.

gits way of keeping branches up to date via manual rebase and merge sucks compared to systems that inherently compose the working set from the patches, and I'm starting to consider if I should move to something like darcs

where did GADT?anon go

File: 1784982981260459.png (129 KB, 961x948)

129 KB PNG

So I found out the way I set up my ECS was basically how you maximize cache misses and memory read/writes so tomorrow when I wake up it's another big overhaul.

File: 1784983287911462.png (180 KB, 1382x1394)

180 KB PNG

>>109366963
Meant to post for better reference.

>>109366658
Yes I expect browsers that were in development for longer than I was alive to be done properly.

File: file.png (62 KB, 779x314)

62 KB PNG

>today I will use rust
>the snailcat way
>bindgen /usr/include/linux/eventpoll.h
>look inside
>EPOLLIN is nowhere to be found
>--help doesn't help
>ask a LLM
>
How come no matter how many times "they" try to fix anything, they only make some shittier version of what already existed which is C programming language and that's an extremely low bar because C is a pile of shit that nobody wants to touch unless forced to?

>>109366796
>muh pointless rant
I don't get this. Just switch to something else if you don't like it?
Git just works fine. It really doesn't matter if it does things good or bad. It solves a problem for most people. If you don't like it, don't use it or come up with something better

>>109367054
io_uring

>>109367064
great another pointless rant in the thread about programming discussion
didn't read

>>109367073
Rust can't handle a #define, I find your insinuation that it can ever possibly handle io_uring's anonymous union nonsense to be ridiculous and sick psychopathic joke.

>>109367082
>>109364699

>>109367096
I have better things to do than reinvent functions

>>109367054
Wait until you find out you cannot accept or epoll_ctl multiple descriptors in one go.
You thinking rewriting our browers is going to be a major hassle? We'll have to throw away all the kernels, and make new ones with new interfaces, that actually support batching.

Which will naturally break all userspace code out there.
As it should be.

>>109367149
what does this have to do with rust being unable to parse a #define for my toy hobby project, I'm genuinely curious, like, forget all the world's problems, I can't even begin using epoll in Rust in any case so why the fuck does it matter if it's inefficient or not

>>109367174
The point is, even if you were to switch to C you'd still be fucked.

>>109367200
doomer cope

>cope
Go on, propose alternatives.

I dunno, rustards could start from learning how to parse C before claiming they superseded it.

>I dunno
So much for that.

Still better than using ASM.

Now THAT is genuine cope.

Your cope is less legible than ASM and that's impressive.

>doubling down on the cope

the headache :(

A single-digit IQ problem.

your IQ doesn't even have digits

>Teach Yourself Programming in Ten Years
I'm on like year 8 of taking this seriously and damn he wasn't kidding.

>>109366496
And all these years we were told OOP was zero extra cost.

Having an advanced type system gives you all the benefits with none of the runtime costs.

File: file.png (334 KB, 1200x779)

334 KB PNG

>>109356604
started working on my desktop sim today
once it matures i'll try to turn it into a hacknet type game

>>109367953
What? What makes you think this has anything to do with OOP?

>>109368032
yes

File: photo_2022-11-25_20-47-52.jpg (103 KB, 921x1065)

103 KB JPG

>>109363900
so trying to translate it (because i don't know haskell) i think it'd be like

(* bad *)
type expr = 
  | Int of int 
  | Bool of bool
  | Add of expr * expr 
           
type value = 
  | VInt of int 
  | VBool of bool
  
let rec eval = function 
  | Int n -> VInt n
  | Bool b -> VBool b
  | Add (x, y) -> match eval x, eval y with 
    | VInt ix, VInt iy -> VInt (ix + iy)
    | _ -> failwith "Type mismatch" (* runtime *)
;;

(* good *)
type _ expr = 
  | Int : int -> int expr
  | Bool : bool -> bool expr
  | Add : (int expr * int expr) -> int expr 
        
      
let rec eval (type a) : a expr -> a = function 
  | Int n -> n 
  | Bool b -> b
  | Add (a, b) -> eval a + eval b (* compile time *)
;;

ok yeah that makes a lot more sense i love ASTs

File: IMG_0106.jpg (122 KB, 1080x1080)

122 KB JPG

>>109366473
The 3 extra cycles of latency touched my penis.

File: kek.png (63 KB, 1227x573)

63 KB PNG

the 3 extra cycles of latency got me fired from my job for being late

>>109368032
>What makes you think this has anything to do with OOP?
Oh, you know, terms such as
>C++ virtual function call
>C++ Exception thrown+caught
and the fact that the C++ virtiual function call is minimum 5x as expensive as the C function call.

>>109366397
>Empty case is completely sound
*crashes*

>>109368448
Are you just dumb or something?

>>109368328
NTA but yes assuming that's the correct ocaml syntax for it
you could also experiment with adding If, ==, etc

friend class T;

this scares orthodox OOP autists

Brainfuck is the epitome of simplicity. There's no bloat.

>>109368498
OOP autist here, i do this all the time thoughever?

>>109368498

fren class :D

>>109368498
Uhh you don't get to bring friends

>>109368500
I/O is bloated and slow, could easily be removed

>>109368553
Goes against the spec.

>>109368500
Lambda calculus

>he is so dumb he doesn't know zero might be a digit
Cow feed. Now.

>he's so smart he thinks his "IQ" deserves even 0

File: 1680458611130954.jpg (5 KB, 300x168)

5 KB JPG

>>109368482
what about something like

type _ expr = 
| Int : int -> int expr
| Bool : bool -> bool expr
| Add : (int -> int -> int) expr
| Equ : ('a -> 'a -> bool) expr 
| If : (bool -> 'a -> 'a -> 'a) expr
| App : ('a -> 'b) expr * 'a expr -> 'b expr

man i should really get around to rewriting the toy interpreter i made last year

>>109368686
that works, i was thinking of just having if/= by themselves rather than with application

>he is so dumb he doesn't know how numbers work

>>109368715
it would let you add function bindings in the future though if you did something like

type (_,_) expr

with an environment

>>109368771
yes but without them kinda pointless
you can also use HOAS (a -> b expr) -> (a -> b) expr
(I think its something like that?)

What's the fastest interpreted language?

>>109368892
i dont know

>>109368892
Java

>>109368892
Haskell in ghci

>>109368498
>try to create a small helper function in anonymous namespace inside a .cpp file
>except you can't lmao, because it's not marked as friend and there's no public interface for it to use
>blame OOP to regain izzat

>>109356604
I cannot for the life of me find a side project I actually want to work on. Once I'm done work I no longer want to program. And whenever I start a project a week or two in I lose interest. How do you guys find projects you actually enjoy programming? I just want a fun C++ project I can do on the side

>>109368892
LuaJIT

>>109368892
this >>109368979 if want to stay lean. V8 or SpiderMonkey if you can afford the bloat.

Why do some people dislike Go?

>>109369127
People that dislike Go:
- Language nerds: Go doesn't do anything fancy and ignores language features of the last 50 years.
- System developers: it has garbage collection.
- Enterprise developers: Go is for making small tools, unlike Spring Boot, and deploying them fast as a native program, unlike Spring Boot.
- Data scientists: doesn't have Python ecosystem, and you need to type everything and be verbose.
- Other Go developers: the Go team added half baked generics and is enshittifying the language with uneeded features. All they wanted was to spam "if err != nil" in peace forever.

>>109369127
it's actively hostile to allman formatting so that alone puts it to trash.

https://www.youtube.com/watch?v=n_k6O50Nd-4

>>109369237
>allman formatting
Cringe

>>109368892
Holy C

Common Lisp is where defun is at.

>>109365531
Comfy clouds, anon

>>109365890
Go outside and sit on your bicycle regardless of how you feel. Now that you are on your bicycle, you might as well ride it down a few feet. Now that you're riding it you might as well keep going. You might even enjoy it.

>>109368892
AI.

File: 1663899253559895.jpg (179 KB, 800x1403)

179 KB JPG

>>109367953
Wait until you're seen typical constructor/destructor code that only initializes or releases one resource, for which the code has to switch to the kernel and back.
Then wait until you realize that most kernels don't even provide mechanism to release multiple resources at the same time.

It's all a completely autistic echo chamber, and I'm tired of pretending otherwise.

>>109369967
How else would you return vectors safely? Getting rid of what you mentioned you're left with garbage collection which is slow or letting the user free them which isn't safe.

>>109370017
>return vectors
What? We're talking kernel resources here. Vectors aren't kernel resources. Sockets, FDs, handles, those are kernel resources, and they are usually just integers into some descriptor table. And creating and closing them should always involve an array, so that you save yourself as many mode switches as possible.

>you're left with garbage collection
What?!

>>109370058
I thought you were saying you don't like constructors and destructors.

>>109370077
No, he was saying constructors and destructors that perform syscalls (or somehow otherwise provoke a context switch) are expensive and that expense is not always obvious at the point of construction/destruction.

>>109370077
Yeah, because they can only ever work on single objects. But that has nothing to do with garbage collection, because you can always explicitly remove a bunch of resources from the process. Nothing stops you from calling close() in a loop; it's just the loop in userspace is fucking inefficient.

Or, to keep it with
>vectors
, assume that your destructor has to mmap and munmap memory for every single object. Completely asinine.

In C is it possible to return alloca'd data from a function by copy?

How do I get the brainworms out? I feel like experience and education have me stuck in one mode of thinking while programming. I want to make a project that's unique.

>>109370152
No.

>>109370215
Learn assembly, especially SIMDs.

>>109370152
usually if you need to copy something as an output, you pass it in as a pointer to an object to be filled

what are you trying to do specifically

>>109370263
Probably trying to avoid explicit memcpys by making the compiler generate them for him.

>>109370271
Yes.

>>109369127
what is there to like?

>>109370405
Easy binary crosscompilation.

>>109370058
>And creating and closing them should always involve an array, so that you save yourself as many mode switches as possible.
Ah, so you want to make life difficult for yourself when you need to handle errors. Pretty much all syscalls can error; I/O is an awkward pain in the ass sometimes.
If you think you can just ignore errors, you're a retard that should stick to vibe coding or outsourcing to pajeets.

I'm going to make a language that instead of functions has jumps that take parameters.

>>109370560
>muh error handling
Yeah, don't even bother. Even fucking AIO got this shit covered; you're just making up strawmen now.