This is /dpt/, the best subreddit of /g/.

What are you working on?

Threadly reminder.

the endless ocean of trying to grasp vulkan

>enter
>leaves

Implemented brk syscall in my nostdlib C++ project because sometimes I need dynamic allocations.

In C++ what's a good way to design an interface where there is only 1 active implementation and it can be swapped?

The way I would think about this normally is polymorphism using a smart pointer to the interface and then populating it with my derived class, and if I need to perform a swap I would delete it and construct and assign to the nee derived class.

Is there a better way?

>>105725499
It's called inheritance.

>>105725499
Generally, you have a reference semantics solution which would be smart pointer to base class and virtual methods, and you have a value semantics solution which would be std::variant with std::visit.

>>105724992 (OP)

Bikeshedding over what serialization and protocol I should use for an API (JSON-RPC, XML-RPC, HTTP POST endpoints with JSON or XML, or SOAP)

>>105725880
You should use a binary format.

>>105725893
If I didn't have to make the endpoints usable from a browser with JS, I would

>>105725904
Binary formats are usable from a browser with JS.

>>105725924
They are? Like MessagePack and such?

do people actually use all these esoteric "features" of c++? I'm just not seeing a use case for syntax so convoluted that you need to pull out a reference manual just to understand what it's doing

>>105725933
Why are webshitters so mindbroken? Everything must be le heckin frameworkerino...
https://developer.mozilla.org/en-US/docs/Web/API/XMLHttpRequest_API/Sending_and_Receiving_Binary_Data

>>105725949
Brother I am not manually inventing and deserializing a binary format, I have better things to bikeshed

>>105725961
Ok faggot, then just use jsöy and fuck off.

>>105725969
What a strange thing to get upset about

Trying to implement a Chip-8 interpreter! Learning a lot about bytes and that kind of low level stuff!

>>105726269
BYTES MY ANUS

>>105725949
>Everything must be le heckin frameworkerino...
Today, I saw someone using React just so they could define constants.

>>105724992 (OP)
Trying to decipher how 4chan "algorithm" on bumping and thread deletion works. Here is how I theorize how it works:

https://files.catbox.moe/j9w5t4.README

Haven't checked the leaked code because I can't read PHP (And even if I knew, I doubt I'll would still understand it)

>>105726848
I downloaded this and it killed me irl i'm a ghost

>>105726873
Fine, here is a pastebin:
https://pastebin.com/29LjE40g

>>105726927
I read this and came back to life then died a second time because of how cringe it is
the archived thread is literally just the last recently bumped thread, if you read catalogue it's in bump order it's always the last one
autosage = mod/janny power to make a thread not be bumped by posts even under bump limit
posting bumps to the front except I think in certain cases (i forget which but that's for stuff like spam bumping your own thread). probably happens in the order the server processes it wrt simultaneity
Lurk more also >spoonfeeding

>>105726964
I understand that I misused the word autosage to refer something else, But what I don''t get is the part of archived threads being the last recently bumped thread.

Isn't the reason of threads being achieved the lack of being bumped? How does 4chan determine what threads are inactive enough to delete them and when to do so?

>>105727079
imagine there is a list of threads
at the front of the list are stickied threads
after that they are all in order by how recently bumped they are
when you post in the thread, it may or may not get bumped which moves it to the front (except behind stickied threads)
when a new thread is made, the last thread in the list is archived
the archives are also in bump order, and at the same time the last archived thread is deleted (or maybe that's time based i don't remember)

>>105727114
ooooh, makes total sense.

Thanks for making it clear like to a retard like me.

>>105727167
1 billion hours mspaint
again i might be wrong about when archived stuff is deleted it might have a timer on it as well
if the thread has been set by a janny to get auto saged because it hurt their feelings or if it has reached the bump limit, then the post is added to the thread it just simply doesn't get moved to the front
the catalogue and archived lists are in bump order

>>105727189
archive should say cannot post i guess. the last threads may or may not be threads you can't bump (since auto sage or bump limit doesn't automatically move it to the end, it just no longer gets moved forward so will inevitably die so long as people keep posting)

>>105727195
Archive straight up tells you it has a timer before a post gets deleted.

Also, wanted to test if in fact you were right since in theory, every time a single anon makes a post and bumps the thread (like in the drawing, j gets bumped to the first page), the order of the threads based on bump order would wildly change every time the catalog pages get refreshed.

And wouldn't you know it. It does. So everything you say checks out.

Thanks anon.

>>105727517
>Archive straight up tells you it has a timer before a post gets deleted.
it does, but they're also in bump order i think (which is also the order they archived), i guess you could see if a middle thread ever disappears from the archive when it's older than threads that archived sooner
my guess would be there's some thread limit so that it actually deletes by bump order, plus a timer
here's an extreme example - bump limit for probably over a year (the limit varies by board, normally 300 or 315, idk what it is for /po/ or which post put it over the edge)
>>>/po/622919
because /po/ is a board that is not as frequently posted on
also there's no threads in /po/s archive - which confirms the timer being real

>>105724992 (OP)
I have been too depressed to work on anything since all the maids left.

>>105728205
Only psycho maids survive

>>105726269
I'm kinda giving up

>>105728385
Why are you giving up on learning about it after only 4 hours?

>>105728432
I'm learning about it but it's getting kinda hard and there's too much noise and people are distracting me a lot.
I guess I need a break but it kind of annoys me that I wrote almost no lines of code today.

>>105728456
Try writing down a few functions or even just formula stuff and use it for tomorrow.
Writing something always feels good and helps you to remember + find what you don't understand.
Very few people can just read/listen and just "get" it.

>>105725949
>le heckin frameworkerino...
messagepack IS a binary format you goober

>>105728456
0-5am is the only sane time for reading

>>105726269
I've done this, it's a really good and fun project. Keep on it, take a break if you need to

C# feels like the ideal mix of static vs. dynamic that I've been looking for for a while now.

how do you think about writing code for longer than 5 minutes without falling asleep or watching porn instead

How to I write a coding empire?

>>105729821
Just keep in mind that, no matter how dumb you might think you are, there are plenty of people in /dpt/ who are even dumber than you. In fact there are plenty of people who're even dumber than an LLM, because at least the LLM can admit it's wrong, and people would actually miss the LLM if it was plugged off.

>>105730038
>no matter how dumb you might think you are, there are plenty of people in /dpt/ who are even dumber than you
No, this is a lie. That guy is the dumbest guy in the block.

See what I mean? Confronted by reality the average /dpt/ poster will immediately deny it. The LLM admits to it and strives to be better, at least within the context of the conversation.

Autism is like pedophilia - the only cure is a bullet to the head.

>there are people even dumber than you
thats the nicest thing anyone has ever said about me

>>105728952
A shit one since I never heard about it.

Can someone explain software version to me?

>>105730220
bigger number = more bloat

I've been writing C++ for 5 years and I have never used new/delete because all the modern C++ youtubers have me terrified that so much as typing those keywords will cause my computer to burst into flames.

>>105730265
They're right, but for the wrong reasons.

>>105730220
https://semver.org/

>>105730265
Wat? I've been writing C++ professionally for over 15 years and I use new and delete every day.

>>105730265
They're just cargo cultists trying to neg you into using standard libraries. You should never use standard libraries for your non-standard programs. The only dependencies you should have are the kernel, the hardware architecture, and your choice of compiler. Never let others gaslight you into believing that 50 million lines of VHDL, C, C++, possibly FORTRAN, and multitude of DSLs baked into those codebases is somehow not enough of a dependency (also known as a liability).

>>105730356
Nobody follows this, and neither should you.

>>105729475

Thanks anon. I'm enjoying what I have so far.
So far I learned to load the bytes of the ROM file into my "reserved" memory, load the font, and am successfully fetching instructions. When I got to decoding them is when I really got confused. This is where I realized I do not really have a clear idea of what I'm doing, I understand I have to work with two consecutive but separate bytes at a time but the rest of the terminology has my mind all over the place, nibbles, registers, the PC (program counter) and all that. So I'll just rest for now.

>>105728509
Thanks for the solid tips anon, I'll try to write some stuff on paper.

>>105729070
I'll see if I can get up early!

>>105730626
stop lying, lots of apps use that. both firefox and chrome, for example.

https://www.chromium.org/developers/version-numbers/

>>105730654
Firefox is on 140 something major version yet handles like and changes like 0.x.y software.

>>105730653
All sprites and fonts are part of the programs you load, schizo.
>code is big indian, first 4 bits are the opcode
switch(memory[pc] >> 12)

elementary.

>>105730827
>All sprites and fonts are part of the programs you load, schizo.
But I never had to do something like this
const FONTSET: [u8; 80] = [
0xF0, 0x90, 0x90, 0x90, 0xF0, // 0
0x20, 0x60, 0x20, 0x20, 0x70, // 1
0xF0, 0x10, 0xF0, 0x80, 0xF0, // 2
0xF0, 0x10, 0xF0, 0x10, 0xF0, // 3
0x90, 0x90, 0xF0, 0x10, 0x10, // 4
0xF0, 0x80, 0xF0, 0x10, 0xF0, // 5
0xF0, 0x80, 0xF0, 0x90, 0xF0, // 6
0xF0, 0x10, 0x20, 0x40, 0x40, // 7
0xF0, 0x90, 0xF0, 0x90, 0xF0, // 8
0xF0, 0x90, 0xF0, 0x10, 0xF0, // 9
0xF0, 0x90, 0xF0, 0x90, 0x90, // A
0xE0, 0x90, 0xE0, 0x90, 0xE0, // B
0xF0, 0x80, 0x80, 0x80, 0xF0, // C
0xE0, 0x90, 0x90, 0x90, 0xE0, // D
0xF0, 0x80, 0xF0, 0x80, 0xF0, // E
0xF0, 0x80, 0xF0, 0x80, 0x80, // F
];

>>105730893
>Rust
Found the problem. Enjoy being told that it's unsafe to load 2 bytes because of endianness uncertainty.

>>105730907
Reading 2 bytes was not hard, I think.

I'm doing it as follows:

struct Chip8 {
memory: [u8; 4096],
pc: usize,
// V: [u8; 16],
// I don't really know what I'm doing with the commented fields
// i: [u8; 16],
// stack: [u8; 16],
// delay_timer: u8,
// sound_timer: u8,
}

// Other stuff here

// (Inside a function that "runs the main loop"
pub fn run(&mut self) -> Result<(), Box<dyn Error>> {
let mut opcode: [u8; 2] = [0; 2];
let mut reader = Cursor::new(&self.memory);
reader.set_position(self.pc as u64); // pc = 0x200
loop {
// fetch instruction
if let Err(_) = reader.read_exact(&mut opcode) {
break;
};
self.pc += 2;

// Decode the instruction
match opcode {
//not sure what to do here
//pulled something from chatgippity but idk
}
}
Ok(())
}

the only language i really know is python and im learning rust rn
i find it interesting how a ton of things are just method calls like i'll call eq_ignore_ascii_case to get case independent equality on a string like what? in pythong i would go .lower() and then == like bro i got to remember every rust method there is like that? that's insane fr fr or like map_or or unwrap_or or is_err like where's my control flow structures
it's all method

I hate programming UI, goddamn index offset forms spaghetti that i have to waddle through every time i want to add something. Being fullstack web and having to deal with js+http on top of this has got to be the most miserable shit ever.

>>105724992 (OP)
Still working on my 4chan client.

>>105732294
is the gui framework cross platform?

>>105732485
I haven't tested in on Linux, but theoretically it should be. I don't use any platform specific code.

>>105732697
is it immediate or retained mode? and how do you connect it to platform specific display and input?

>>105732747
Retained mode. I use SDL3 for windowing and input. Sometimes I'll use native Win32 APIs for certain effects that SDL3 doesn't support, but I made sure to contain that.

>>105732766
i see
although i think sdl has too much stuffs not needed for desktop programs so i made my own window and input handling
also are these images and texts for the thread display individual gui elements?
>Sometimes I'll use native Win32 APIs for certain effects that SDL3 doesn't support
interesting
which effects?

>>105732839
>which effects?

Effects like a drop shadow underneath popups.

>although i think sdl has too much stuffs not needed for desktop programs so i made my own window and input handling

I'm considering using Win32 for Windows and SDL3 for other platforms, as it's starting to become more trouble than it's worth. But that might be for later down the road.

>also are these images and texts for the thread display individual gui elements?

Most of them are. Like individual labels and stuff. My framework supports rich text so comments are all one label. In the attached pic, each blue box represents an individual widget.

>>105732877
>Effects like a drop shadow underneath popups.
that doesn't require win32, it can be done internally with alpha composition in rendering overlapping widgets
>I'm considering using Win32 for Windows and SDL3 for other platforms, as it's starting to become more trouble than it's worth. But that might be for later down the road.
for me i target some niche platforms that aren't supported by it so i can't use sdl
>My framework supports rich text so comments are all one label.
if quotelinks and other parts of comments are one element how do you handle clicks on them?
also how do you reply to the thread with your client? do you bypass the cloudflare check for captcha or you just use 4chan pass?

>>105733000
>if quotelinks and other parts of comments are one element how do you handle clicks on them?
Links are just part of the text, just bounds test where your cursor is, easy.
>also how do you reply to the thread with your client? do you bypass the cloudflare check for captcha or you just use 4chan pass?
Your browser stores cookies which 4chan checks, you can use these cookies for posting and passing Cloudflare. So posting does work, but it's a little weird. I need to design a better system of getting those "keys" unobtrusively.

>>105733024
>Links are just part of the text, just bounds test where your cursor is, easy.
just wondering if it's part of the label's internal functionality or it's done externally
>Your browser stores cookies which 4chan checks, you can use these cookies for posting and passing Cloudflare. So posting does work, but it's a little weird. I need to design a better system of getting those "keys" unobtrusively.
not a fan of cookies since they expire and have to open the web browser to get another one is a hassle even when you can grab the cookies directly from browser like how yt-dlp does it

>>105733158
>not a fan of cookies since they expire and have to open the web browser to get another one is a hassle even when you can grab the cookies directly from browser like how yt-dlp does it

There's no real way around it, it's basically a password for 4chan to validate that you're not a bot. But I do want to make this process as painless as possible.

>>105726848
The bump order algorithm is literally just ->

The board has a thread limit of n threads - stickies

When a new thread is created, it's put in position #1(counted after the stickies) and the thread in the last position is removed.

Every post that is not a sage bumps the thread it's posted in to #1 indiscriminately and contributes to a set m post bump limit. Sage posts don't bump the thread or contribute to the bump limit (they do contribute to the image limit if they have images). After the bump limit is reached, threads start autosaging, as in they cannot be bumped anymore.

When a post is bumped to #1, all posts that were above it are shifted down by one position to fill the vaccum.

That's it. No need for speculation or PPH/PPT. It's a simple, but effective sorting method for sorting by activity as long as the thread limit is <~250, allows for both less active threads to get attention by getting bumped up to page 1 while active threads constantly remain somewhere on the top pages by getting bumped up regularly.

Is it possible to use something like Guile or Babashka's bb.edn, or really anything else to replace makefiles on hobby Cpp projects? Any of you have tried?

>>105734009
>replace makefiles
no but GNU make lets you define new commands in guile
https://www.gnu.org/software/make/manual/html_node/Guile-Integration.html

I don't how good support guile has for just executing commands.
People who don't like build systems usually write bash scripts...

>>105728265
>maid deleted
Why are jannies like this?

>>105724992 (OP)
Workin on mah websitter app. Added some axis helpers on my members and also added some structural bay creation tool.

>>105734074
TIL, I'll look into it.
>bash script
Nah i tried that, it's annoying in the long run.

>>105734088
Because you deserve it.

>>105734088
All posts are constructed in nested psyops to localize and misdirect any potential janny action, so we can accept it in good humor. This is how I run /g/, /sci/, and /lit/ with minimal interventions despite a post history only containing faggotry

My cryptography library has a site now
https://reshsix.github.io/maid/
I wanna add some examples here in the future

>>105735589
>poster.webp

>>105735589
Kill the shitty music, holy fuck.
At least put some Nier Automata music on.

>>105735621
Come on, I converted to webp to load faster on everyone's computers...

>>105735629
I grabbed a non royalty one, if you share some other music that fits the atmosphere and I like more I could change it

today I learned they can't TLS fingerprint you if you're using plain HTTP

I just had a genius idea to get pdfs of programming books for free without going to piracy sites and getting buttfucked by book publishers and ISPs.
>Go to library
>Take home books for borrowing
>Use scanner app on phone to scan books to pdf (takes <=5 hours total depending on book)
>Save pdf to computer
>Return books to library
>Repeat

>>105735589
where are the elliptic curve ciphers

Damn, those curves are elliptic

>>105735800
Soon! There's other algorithms and interface changes that I wanna focus first
Elliptic curves might need it's own version, since they have a long way towards optimization like RSA

>>105735707
Dude just google it, what can the ISP even do lmao

the land of the brave and the free

>>105735707
>the programming section at the library be like

Wanna have some fun? Go to sourcegraph and search this
/[01]?\d\/[0-3]?\d\/199[4-9]|[01]?\d\/[0-3]?\d\/9[4-9]\b/ file:\.(cc|cpp|CPP|hpp)$

>>105730893
>>105731134
>>105731639
r*st only exits for crypto and thats dead anyone still pushing it now is grasping at strawmen

Global variables for performance?

Just watched a video on the duke nukem 3d engine and the single developer made every variable global which puts them in .bss instead of .data or something and is more efficient?

Even at a shitzo level is there any merit to this?

>>105736934
>Global variables for performance?
Not a bad idea. Ignore what reddit says

>>105736934
it's a old game where limits are known at compilation time, with no scripts or anything.

>>105736750
kek I remember being interested in programming back in elementary school in the last 90s so I went to the school library and all they had was a book on commodore basic

>>105736934
anon there's no performance penalty to local variables you're just moving the stack pointer more bytes

>>105737670
That's if they even make it to the stack and don't just live entirely in registers.

>>105736934
>is there any merit to this?
Not these days unless you're doing something rather exotic or very very time sensitive. It's rarely worth the trouble, especially in multithreaded code (like what the Duke3D engine definitely wasn't).
The only benefit back in the day would have been avoiding indirection via the stack pointer, but that's a very heavily optimized path on modern silicon so it's not worth avoiding.

why can the compiler not auto vectorize this? seems fairly trivial

>>105737842
my guess is it doesn't know it can access data[i] unless it has checked previous values against UINT8_MAX

>>105737842
because you are using wrong compiler flags.

>>105736873
>>105736873
Any tips for unlearning Rust and moving back to C++?

>>105738091
There's no detransitioning. It doesn't grow back.

>>105738091
You first need to invest into some cutting-edge research if you want a chance.

>>105738091
time travel to stop your past self from learning it and pray it affects your current self

best free ai to give me terrible programming advice?

ouch, chatgippity

public static class ClassExtensions
{
public static T Unwrap<T>(this T? val) where T : class
=> val ?? throw new NullReferenceException();
}


Why isn't this already a thing in C#

>>105738767
>learned C++ from ChatGPT
>only found out afterwards when I went to actually program something that it hallucinated 99% of the shit it told me

>>105738779
isn't this just doing x.Value where x is T?

>>105737842
>zoomer doesn't know how to do screenshots
>uses his phone like a dumb monkey

>>105736934
If you have to ask then that means you don't even understand what you're asking. Learn some assembly.

>>105738914
Can't do that with ref types

i wish C++ had class/object extensions

https://www.youtube.com/watch?v=AAXlmUKPn2c

>>105738891
I just realized a usecase for Rust. Vibe coders can use AI to write Rust code and so long as they don't do any unsafes theoretically they shouldn't create software with horrible memory leaks or worse, critical vulnerabilities.

>>105738996
>learning assembly
/g/: WHY THE FUCK DON'T YOU JUST LEARN ASSEMBLY YOU FUCKING RETARD

>>105739358
Yes.

>>105736934
Oh it's an intentional performance thing? The dos racing game I am decompiling is like that too. I thought it was just 90s cowboy programming.

>>105739374
CPUs weren't several hundred times faster than memory during DOS's lifecycle, and there was almost no address translation fuckery going on (except for segmentation descriptor nonsense which was relatively benign and could even be abused for stuff like unreal mode, allowing for memory accesses beyond 1 MB without needing to constantly switch modes). On top of that x86 didn't have that many architecturally exposed registers in the first place, so there were constant interactions with memory.

>>105739358
yes? nobody's saying you have to program in assembly, but you should definitely understand and be able to read it, especially if you're bizarrely concerned with microscope performance differences

>>105730893
>>105730653
>>105726269
Holy fucking shit, I made it. I can't believe it.
I never thought I could do this kind of programming. I thought I could never do anything more elaborate than some webdev shit.
Granted, I asked chatgippity a lot of stuff but, holy shit.

>>105739710
What is the total size of your project directory?

had to fix my booru client because i broke batch download

>>105739726
du -hs chip-8/
285M chip-8/

uuhhhh.... well what can I do

>>105739811
That's less than I expected to be honest.
Carry on and well done.

is there someway to use the assert macro from cassert like

assert(RegisterClassEx(&windowClass) != 0);

but still have it run outside of debug builds? or do I have to write a custom macro?

>>105739910
Heh, I thought I was going to get bullied to death.
Thanks Anon <3
For now it only handles the IBM logo ROM, I still need to implement all the other instructions to handle more ROMS. But as I far as I understand, the display instruction was the hardest. Glad thats over.

>>105740369
Contracts (C++26)

>>105740374
It's not you it's rust, it's notorious for creating massive projects.

You will get bullied but not by me and not for anything you've done.

>>105737842
Because it can't know that your data is aligned or that you allocated enough for it not to matter anyway, because nocoders use malloc and don't even know themselves how data is allocated.

>>105739710
>args live on stack in memory that cannot be deallocated
>rustards allocate a string for each arg and then shove pointers into those allocations into yet another allocation
and then people wonder why nobody takes rustards seriously.

>>105740738
fucking lol I forgot you need to dilate your arsehole just to get args in rust.

>>105740770

>>105739208
public class Program
{
static void ModifyValue(ref int? number)
{
int n = number.Value;
Console.WriteLine($"Inside method: {n}");
number *= 2;
}

public static void Main()
{
int? x = 10;
ModifyValue(ref x);
}
}

You mean this?

>>105740738
And what else are you supposed to do Mr Cnile? This is how we do it in 2025, unused ram is waster ram!

>>105740738
Tons of software is rewritten and rust and ends up being faster significantly faster than the older stuff. Compare this to the software being rewritten in C that doesn't exist.

>>105741170
I'VE SEEN THIS COPY PASTA BEFORE

>unused ram is waster ram
Spoken like someone who doesn't know what cache associativity is.

>>105741170
That's not difficult, considering that the old stuff's code quality is fucking garbage. Doesn't justify inflicting irreversible brain damage to yourself by learning and using Rust.

>>105740808
https://www.phoronix.com/news/NOVA-Core-Co-Maintainer
https://www.phoronix.com/news/Tyr-Rust-DRM-Graphics-Driver

step-by-step, rust is making his path toward supremacy. you and everyone else here is already running rust code on his machine.

>>105741386
And here I thought it was Ada that was responsible for nVidia's particular expectorations.

jesus christ so much fucking dx12 boilerplate before you can even start writing the rest of the program

>>105742234
Sounds like the same kind of kvetching of incompetent PS3 devs.

>>105742234
Feel free to reimplement graphics drivers yourself.
>>105742265
PS3 doesn't have several different GPUs that can be switched out anytime or even used in pairs for some mysterious reason, PS3 has no boilerplate, it's just you and hardware. Being a console dev must be incredibly comfy.

>>105740834
No I mean https://learn.microsoft.com/en-us/dotnet/csharp/language-reference/keywords/reference-types
Taking a ref of a value type does not make it a reference type.

>>105741170
the rust evangelists that i have seen in the hn and lobsters threads admit that it is because they use newer, more efficient algorithms and also design the api's to be more efficient (like calculating some value/context before making multiple calls to the same function, instead of having the function always recalculate it), not because rust is inherently faster.

Has Linus torvaldsonn learned rust yet or not? If not, how do he accept rust code in linux if he can't audit it? How does he know if its exploit free or not?

>>105743052
Is "rust code in linux" in the room with us right now?

>>105743069
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/rust?h=v6.16-rc3

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/gpu/nova-core?h=v6.16-rc3

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/gpu/drm/nova?h=v6.16-rc3

>>105743102
I don't have novideo GPU, so rust is literally not in my kernel.

I LOVE CODING IN LIGHT MODE

Before Borland’s new spreadsheet for Windows shipped, Philippe Kahn, the colorful founder of Borland, was quoted a lot in the press bragging about how Quattro Pro would be much better than Microsoft Excel, because it was written from scratch. All new source code! As if source code rusted.

The idea that new code is better than old is patently absurd. Old code has been used. It has been tested. Lots of bugs have been found, and they’ve been fixed. There’s nothing wrong with it. It doesn’t acquire bugs just by sitting around on your hard drive. Au contraire, baby! Is software supposed to be like an old Dodge Dart, that rusts just sitting in the garage? Is software like a teddy bear that’s kind of gross if it’s not made out of all new material?

It's copypasta hour today, isn't it? The little faggot is just posting nonsense that's already been debunked ages ago.

>no content, just seething
I bet HE hasn't finished any of his rust rewrites

Do you guys listen to music when coding?

Gotta say, it's pretty comfy living rent-free in your heads.

>>105743454
>guys

>>105743454
Rain sounds to focus

>>105743281
old codebases suck, they dont acquire bugs but they acquire maintenance costs
>>105743454
she looks like she reposts images with "please do not repost" in them
also pokemon and shotas

would you agree with this? i kind of want to do the dinosaur book because meme books make me feel cool

>>105742234
>SDL_Init
Done.

>>105725880
If it's a CRUD API, then stick to HTTP

Otherwise use gRPC, it requires much less boilerplate than everything else you suggested there. It's by far the easiest to work with

>>105724992 (OP)
>What are you working on?
a network filesystem in C (using libFUSE).
I've only discovered Linux `perf` profiler today, but it already taught me a lot about performance.
I assumed the overhead of my program would be miniscule compared to hitting the disks and the network, turns out i was wrong.
Some lessons i've learned:
1. memset is expensive. Don't zero out structs and arrays if not necessary.
2. ifs are expensive. For logging, don't check the loglevel inside the logging function. Instead, use a function pointer for each loglevel, pointing either to the logging function, or an empty function if that loglevel is disabled. (maybe there's an even better way without the function call overhead?)
3. when (de)serializing data, don't do it field-by-field by advancing a pointer. Instead, (de)serialize by using offsets from the pointer, the compiler will then generate fast SIMD code for these paths.
Profile your code anons!

>>105743454
Listening to classical music right now.

>>105744727
>memset is expensive
How big are your structs/arrays, and how many implicit/explicit calls to memset are you making? memset should be highly optimized, but if you have 100s of MBs of data or are zeroing newly allocated pages (unnecessary) you can hit a lot of page faults.
>ifs are expensive
The ifs in logging functions should be very well predicted if the global you're checking is fixed at startup, and should have basically no visible cost. Are your logging calls being inlined or something?

>>105745296
>memset should be highly optimized
lmao

>>105744727
>ifs are expensive
retard

How does a Java interface work internally?
Suppose I have a class C that implements interfaces I,J, and K, and there is source code with a variable x that declared as I
I x;
if interface I has method m, then how does method dispatch work when evaluating
x.m();
I'm assuming that x has type C and the type object for C has a table of methods, so how do the compiler and runtime pick the right method to dispatch?

>>105745367
nta but they are, during testing just reordering an if-else chain to handle the most common case first improved execution speed by a surprising amount (10-15% across 10 million iterations from memory)

>>105742234
DirectX 12 isn't meant to be written by mortals, you fool!

>>105745699
>most common case
strawman

>>105745621
Here's how I look at it: at run time, the JVM knows the identity of the interface I because this can be determined at compile time and baked into the source code, and it knows the address of the type object for class C. Given these two pieces of information, there has to be some table that stores a pointer to a dispatch table that stores the addresses of functions implementing methods in I for class C.
So how does the JVM obtain the pointer to the dispatch table? Is it a hash table lookup?

>>105745707
It's actually not that complicated. If you don't want to micromanage everything you can just use a template and it basically becomes DX11.

>>105739279
low code is slow code

>>105743281
>It doesn’t acquire bugs just by sitting around on your hard drive.
This was written a decade before node.js and npm were released.

>>105745621
not even internally but just consider how it works on the surface:
a Java interface is just a contract where if a class is declared as implementing the interface, it's required to provide an implementation of all methods of that interface.
there's no uniqueness to the methods of each interface, only a requirement that their methods exist: if many interfaces declare the same method, you can call that method from any interface of that implementation, it'll be the same implementation called

so in your example:
if a class C implements interfaces I, J and K, it's required to implement all methods declared by the interface
if multiple of those interfaces declare the same method m(), the class C will fulfill the contract for all interfaces with its one implementation of m()
it doesn't matter how you declare your variable, it'll still call the concrete implementation of m() in C (assuming you assigned an instance of C to the I variable)
calling it from an interface only limits you to the methods that the interface declares, so you won't see methods from other interfaces or the implementing class (at least not unless you cast the variable)

>>105745296
>How big are your structs/arrays, and how many implicit/explicit calls to memset are you making?
depends on filesystem operation, ops that deals with filesystem path need 4k (max path length on Linux) + headers, reads/writes need 128k (max read/write size) + headers, i'd initialize such buffers and structs with `= { 0 };` in each fs operation function. With a big workload, like `ncdu -t 32` and short filepaths, you could imagine quite a bit of amplification in thousands of those getattrs per second. Even when reading a file sequentially, a memset for read buffer means i'm memseting the buffers with the same speed as i'm reading the file (in loopback tests, i get ~450MB/s, that's 450MB/s wasted on memsets).
>The ifs in logging functions should be very well predicted if the global you're checking is fixed at startup, and should have basically no visible cost. Are your logging calls being inlined or something?
No, logging functions don't get inlined, i used a global for loglevel at startup. I have a lot of TRACE logs for every function in my code, but even when they all immediately return early on this if statement, the logging still took 10% or my server binary runtime. I should probably research how branch prediction works, i think perf can also count branch misses too.

>>105746739
he's asking how it's implemented
how it works at runtime

>>105745621
>How does a Java interface work internally?
Likely through a sub-vtable, though it absolutely doesn't have to be done that way. At the spec level, it's done by a JVM instruction called invokeinterface that handles the details.
There's been quite a bit of work on how to make this efficient. See https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=81d1b10935ef9e148a0f75eef57a0d2d4daa2fa7

>microsoft doesn't provide any includes for win32 assembly programming
>have to rely on some sketchy community-made masm64 sdk that's development has practically stopped on because the main guy behind it died

>>105746944
>assembly
>includes
???

>>105746944
>dude I shaved off 3 instructions when performing the one time task of opening a window!
Literally what is even the point of assembly outside of some specific algorithm you're trying to optimize?

>>105746956
nta but if you want to interop with C you still need function prototypes, defines, structures, etc... Take a look at Windows.h, I guarantee you that you don't want to manually translate all of that for your assembler. Probably not hard to automate, though.

>>105747003
Ok but what does that have to do with an include? If you're writing assembly like he's saying then you have to set up all the state and jump, surely?

>>105747017
Could you manually set up the stack according to the Windows x64 calling convention every time you call into win32? Sure. But the linker still needs to know what to link to. You're also going to get really tired of doing that really fast, with the prototype you can write a macro to do all the repetitive work for you.

>>105747070
What kind of macros are we talking about? I don't really have any experience with this. Is there a reason he can't just use windows.h etc?

>>105747083
Like I guess you'd need something that lexes and preprocesses C to work with windows.h/related headers and all its macros, I'm not familiar with assembly macros or anything that manipulates C function prototypes so I'm assuming there's some other sort of prototype for assemblers?

>>105747083
>Is there a reason he can't just use windows.h
Your specific assembler expects things defined in a specific way, and that way isn't the way C defines things.

>>105725961
>am not manually inventing and deserializing a binary format
Why are you saying this as though it's some big task?

I JUST WANT TO BE ABLE TO HOT-RELOAD I HATE RECOMPILING SO MUCH

>>105747187
t. unemployed

>>105746871
What is a sub-vtable?

>>105747017
I'm so glad that I didn't choose to be an assemblycuck. I tried to use Rust, having to convert C headers by hand broke me, and now I noticed that assembly has exact same problem, horseshoe theory.

>>105747354
a vtable within the object vtable (or at an offset) that guides the interface dispatch.
iirc dotnet had basic optimizations that remove the need to walk the entire street when you know the concrete type so instead of ((I)((T)x)).m() (which goes from T vtable to I vtable to get m()) it becomes ((T)x).m() (T vtable m()) and I think similar optimization for when T: B and m() is overriden by T

>>105747544
I hate the antichrist.

>>105747187
All good languages have hot reload, wdym?

>>105747563
Me too

>>105747374
>claims to be a "programmer"
>converts data by hand

>>105747374
Haskell doesn't have this problem

>>105747187
bro your javascript
bro your common lisp

>>105747676
I never needed pure asm, inline asm just works... in C.

>>105747563
did you vote for him

>>105724992 (OP)
The movement of the green cube makes no sense, op

>>105748722
lel. npc detected.

>>105748744
elaborate

>>105748782
the green cube is bouncing inside a cube.

>>105746944
there's not even a use case for assembly outside of writing drivers or optimizing a crypto algorithm

Print all prime numbers below 50, you can only use one statement, a for statement.

[spoiler]
#include <iostream>
int main(){
for (int n=2, in=0, prime, i; n<=50; in||(i=in=prime=2), i*i>n ?
prime&&std::cout<<n<<'\n', ++n, in=0 : n%i++||(prime=0));
}

[/spoiler]

Print all prime numbers below 50, you can only use only one for statement.
#include <iostream>
int main(){
for (/*your*/; /*code*/; /*here*/);
}

>>105749155
ez

#include <iostream>
#include <print>
int main(){
for (; false; ) ;
std::print("2 3 5 7 11 13 17 19 23 29 31 37 41 43 47\n");
}

>>105745296
>memset should be highly optimized
It's not because it cannot make assumptions about alignments.

>>105749466
it can since it takes a size and an address

>>105745296
>memset should be highly optimized

AMD has a collection of highly optimized libraries
https://www.amd.com/en/developer/aocl.html

avx512 memset:
https://github.com/amd/aocl-libmem/tree/main/src/memset

>>105749155
good morning sirs

interface ThingThatRunsInAForLoop {
public boolean condition();
public void step();
}

public static void main(String[] args) {
for (var p = new ThingThatRunsInAForLoop() {
private int number = 2;

@Override
public boolean condition() {
return number < 50;
}

@Override
public void step() {
if (isPrime()) System.out.print(number + " ");
number++;
}

private boolean isPrime() {
int i = 2;
while (i <= number / 2) {
if (number % i == 0) return false;
i++;
}
return true;
}
}; p.condition(); p.step());
}

>>105749658
Nope. It checks for alignment at runtime, which is the opposite of "highly optimized".

>>105749916
You have to check anyway to know how much registers you can leverage.

>>105749938
OK, schizo.

>>105747563
I don't, and I love the anal rape he subjects God and his ridiculous followers to.

>>105750020
>>>/pol/

>>105750184
I didn't bring up sex changes, even if I do believe opposition against them shouldn't receive first amendment protections.

>>105747187
I hate hot reload it feels untrustworthy and suspicious.

>>105740369
On MSVC/MinGW assert is just a wrapper for CRT _assert:
#ifdef NDEBUG
#define assert(_Expression) ((void)0)
#else /* !defined (NDEBUG) */
#if defined(_UNICODE) || defined(UNICODE)
#define assert(_Expression) \
(void) \
((!!(_Expression)) || \
(_wassert(_CRT_WIDE(#_Expression),_CRT_WIDE(__FILE__),__LINE__),0))
#else /* not unicode */
#define assert(_Expression) \
(void) \
((!!(_Expression)) || \
(_assert(#_Expression,__FILE__,__LINE__),0))
#endif /* _UNICODE||UNICODE */
#endif /* !defined (NDEBUG) */


So just use this:
#ifndef my_assert
#define my_assert(_Expression) \
(void) \
((!!(_Expression)) || \
(_assert(#_Expression,__FILE__,__LINE__),0))
#endif

I feel like gnome is being very hostile towards developers.

Should I switch to KDE, qt and c++?

>>105750985
>just implement it yourself bro
This is why I will never take standardslop seriously.

>>105751174
i know macros look scary bro but if you break it down one line at a time even you can understand it probably

>>105751120
Everything FOSS, especially related to Linux, is hostile to developers in current year. It's all been coopted by degenerates with a political agenda. The only things keeping it going are a handful of elder boomers who bury their heads in the sand and write code pretending everything is alright, and they'll be dead soon. At that point all you've have left is a bunch of broken and abandoned partial Rust conversions.

>>105751202
Sorry, I'm just not gonna use libc if I have to keep fixing it every time I want to do something sensible.

>>105751249
Alright then, see ya.

>>105751415
with the added benefit of ARMcucks and RISC-V cucks having to reimplement my code if they want to run it, because my syscalls are only implemented for x86-64.

>>105751633
Bye.

>>105751665
I'm not going anywhere, seethe troon.

>>105751714
Toodles.

>>105751837
I accept your concession. I truly do. Ironic that it's a rare thing that Rust got right, by having debug assert and assert that always runs.

>>105751846
I don't see it.

How do I enjoy programming again?

>>105751880
First, you need a real problem, and not a fizzbuzz.

>>105751894
Alright. You are the problem, and voted doxxing is the solution.

>>105752108
Well, when are you gonna solve it?

>>105752126
Implementation isn't my job.

>>105752163
You'll do it for free anyway, and it won't be considered a job.

>>105752167
Nah.

>>105752270
You will never be a woman.

>>105752274
Noted.

>>105734268
What's running the matrix calculations?

>>105752349
>cutting out the second panel where doc hangs himself

>been learning linux commands, bash stuff, and make commands
>now Im so use to text editing shortcuts, I'm becoming adhd

Wow, no wonder people get addicted to vim/ emacs. Ive only been using the standard stuff like nano, but I guess I'll learn vim at some point.

>submit raw patch for project that improves performance
>no response for an entire month
>ask what's happening
>"can't you supply a pull request"
>fuck no, I'm not gonna install the dependencies for your fuckfest of a code; be glad I bothered to whip up the code in the first place
>"we need to profile it"
>"actually, no, let's just close the issue"

Alright, see you in another fifteen years. That's how long it too them to realize I was right the first time.

>>105737842
managed to do it, but not really sure if the extra code size is worth it
any suggestions for improvements?
also, i only quickly tested it, but it seems to work

>>105755360
lmao the fuck you are doing? -O3 -march=znver5 auto vectorize the original code.

>>105755360
just realized that i have to check if size_head is greater than the limit

>>105755502
i'm stuck on a 5+ year old compiler for reasons

>>105755502
>>105755360
>no compiler generates a simple VPCMPEQB + TZCNT

I've written a server-client pair in C for Winsock backend.
I separated statemachines from backend as much as I could to facilitate changing the backend(dos serial, actual berkley socket, maybe named pipe), but I feel I could have written this better.
How should I improving?

https://gist.github.com/pachuco/bbf8c2dd5a823de1e8a11641ab95a0b0

Also, I asked Claude nicely for initial sketch of functions(commented out), by providing it with step-by-step of the network exchange.
Surprised it worked.

>>105756285
Oh, 4chanX doesn't embed the 2nd file.

>>105755212
Why would you need to make stuff up and claim you contributed to something without actually compiling it? Based dev.

>>105756377
Why do you have to insert your autism in literally every conversation that you have?

>>105756409
It's not a conversation, I'm just glad that the dev told you to go fuck yourself instead of blindly comitting code you didn't even test. Probably used LLM to make it slower and hallucinated that it's faster.

>autism makes him doubling down
And that's why we used to beat the literal shit out of mentally ill people, and I can't wait for society to go back to these golden times.

Nowadays we're sadly limited to ignoring mentally ill retards, closing their incoherent """raw patches""" and not even explaining why, because calling retards retarded is illegal.

Naturally we would also get rid of any and all accomplishments of autistic people. I would happily go back 500 years in technological developments if it means that I can watch autistic people being tortured to death on the local market square. That would be SO MUCH FUN.

>>105756522
It would be especially fun for you, since you'd get to experience it directly from your own POV, something attainable only by VERY important individuals.

Can you imagine how much screaming and screeching the autists would produce as the lawful government of the people cut body parts off of the antisocial fucks who deserve nothing but misery and pain for the rest of their considerably shortened lives? I would happily die off of tuberculosis at the ripe old age of 32 if in exchange this scenario would become my source of daily entertainment. So see autistic people being mutilated and forced to eat their own body parts before death would fill me with a joy I can barely put into words.

>>105754778
cniles don't want anyone to know this, but full-fat IDEs also have keyboard shortcuts

>>105756588
>>>/pol/ptg

>>105756759
they do. Im not leaving my ide, dont worry.

>the autist knows what a PTG is
Luckily I don't. Because I'm not autistic.

>>105756759
How many of them support the same actions without having to install a vim plugin though.

>>105756908
all the useful ones at least, and then more for all the integrated stuff that a plain editor doesn't have
i'd imagine all a vim plugin would add is some esoteric "cool but not very practical" editing stuff

>use mouse or standard IDE shortcuts

There. I invalidated vim shit

>>105757007
>all the useful ones at least
>i'd imagine all a vim plugin would add is some esoteric "cool but not very practical" editing stuff
Ebussy is that you

>>105757461
name the thing vim has that IDEs don't and need a vim plugin for

>>105724992 (OP)
Made this today~
https://ayasequart.org/stats

>>105757620
Non-retarded keybinds that make you actually productive.

I just finished a timer in gtk4 but I'm so fucking lazy instead of implementing a sound handler I just spawn mpv player to play a system sound and make it a dependency.

https://github.com/aussie114/timer

The best part is it's blocking so you have to wait until the sound finishes playing before you can exit the timer.

>>105758793
>non-retarded keybinds
>vim
if you're that kind of warped in the brain, IDEs usually have vim keybind presets
if not, you can find a preset online or reconfigure the keybinds yourself

>>105758934
>tranny projects
>somehow hallucinates that an IDE can just have a preset for a superior editing mode
If your brain wasn't so warped, I'd explain the difference, but you're a nocoder fizzbuzzer and anything actually useful is entirely lost on your kind.

>no argument
>only ad hominem
concession accepted
vim has nothing on IDEs

no reply necessary

>>105740369
Just do #undef NDEBUG before you start #includeing anything.

>>105743454
Sometimes, when I feel like it. Classical, 80s pop, 90s techno, or psytrance all do it for me.

>>105743454
Sometimes but it has to have no vocals or else I can't concentrate.

>>105724992 (OP)
>inb4 /wdg/
/wdg/ is for frontend shitter

Which programming lang should I pair with a cheap 0.5 GB ram VPS?
I am thinking go or rust (have no problem using either or another with docs/google) but I want to know if anons have specific issues with them
It's basically just user auth to gain access to aws compatible object storage

a script to save my body weight

>>105759702
C

>>105759702
dont use rust lol the borrow chekcer allowcatws like 11 gibibityes on startup

>>105759860
>chekcer allowcatws
NTA and I agree Rust stinks but did you just have a stroke

ngl I have literally never read code written by another person beyond some snippets in documentation, I just write however I feel like

These threads are a good example of what I'd end up like if I became unemployed.

>>105759902

>>105749155
I am going to do this tomorrow probably

I finally uploaded my first portfolio/blog personal site. I made it on flutter because I hate html/css/js. Is not the best tool for those kind of sites fyi. But I'm happy with the result. I deployed it on aws s3/cloudfront/route69 and when its done I'll just upload the changes to s3 and put the link on my instagram or linkedin. It was pretty fun Tbh, specially because I'm a backend developer and i was getting tired of programming.
Thanks for reading my blog, I'm also making a blackjack game for the terminal on Go, but i'm just starting

have you tried gemini cli?

I've started exclusively using trailing return types in C++ and it's way easier to read, only thing that fucks me up is forgetting to put const before the arrow when that's warranted

>>105744391
I had a class that used this as a textbook, it was very dry IIRC but I appreciated the depth, particularly with scheduling

>>105758900
It's 2 lines of code, chatgippity bum steered me and said gtk could not play audio.

>>105761893
use gemini not chatgpt

anybody want to give some feedback on my wip textboard rewrite?
https://beta.kuudere.moe/forum/
feel free to spam or try to break things
the neat bit is it should all update in real-time

>>105739696
>Nobody's saying that
I'm saying that.
You have to program in assembly, at least a little, to get an intuitive understanding of what sort of things high level languages are intrinsically bad at. If you can't, I do not respect you as a developer.

>>105761409
>has to prefix functions with auto

>>105762356
High level languages are bad at wasting my time, and I never needed to program in assembly to know that.

>>105756759
>Keyboard shortcuts
Wasn't ever what makes vim better than IDEs. Lack of common lag in most corpo IDEs, highly extensible and integrated with the shell, integrated UX designed to be a force multiplier for competence. Adding the basic vimkey bindings alone are actual cope from IDEs to stop the user hemorrhage. Yes, this implies Emacs is also better than an IDE. It is.

>>105762438
Your post tells me that you're a brown nocoder, IDEs don't lag on my hardware. Vim's modal editing is what's superior, since I actually use it, I'd know. Now go install a shower and a toilet in your crib, should keep you occupied long enough that you never post here again.

>>105762386
>Bad at wasting my time
High level languages waste your time constantly, most notably by giving you more than enough tools to dig an abstraction grave.
Did you just have a stroke or did you intentionally write this vague and self contradictory?

>>105762471
>He doesn't notice the delay
Tell me you're slow without telling me you're slow.

neovim's indentation is so retarded holy FUCK

>>105762499
I don't stare at what I write, I didn't buy a proper keyboard to doubt it and my white hands.

>>105762488
>abstraction is le bad!
Okay tranny.

>>105762512
You're bragging about being unperceptive? IDE user to a T.

>>105762529
I don't believe that not being in total poverty is a brag.

>>105762518
I didn't say that, and you're nigger for pretending I did.

Over abstraction can waste as much time as under abstraction. You'd know this if you had a first or second world IQ.

>>105762549
Luckily I'm not prone to extremism and that's why I wouldn't touch assemblyshit even if I was paid for.

>>105762504

>>105762558
Are you ok? I don't give a fuck about you, why are you pretending I care what you'd be payed for? "I don't like it" isn't convincing to anyone that doesn't already agree. You're fucking retarded.

>>105761918
lol

every tom, dick, and harry is making a textboard/imageboard after the hack

>>105762595
There's nobody who has high IQ and writes in raw assemblyshit. Assembly got obsoleted the moment first compiler was written, now we have "inline assembly", which luckily isn't a fucking waste of time, but more like configuration for specific piece of code. If you weren't a nocoder, you'd know this. Now go back to writing function prologues and epilogues while I'll let the compiler to the unpaid busywork for free, but at consistently high quality with high probability of inlining, meaning my code is both simpler to read and implicitly faster, all because I'm using C.

>>105762356
Oh no. The autist wouldn't respect you.

A bullet to your head would easily cure your mental affliction, just saying.

>>105762621
mine's been around since 2014 or so
i just rewrite it and make it worse every few years

>>105762595
>payed
Alright, retard.

>>105762623
>Assembly got obsoleted the moment first compiler was written
>>105755589

>>105762656
There's hardware that doesn't have tzcnt instruction, so it won't be automatically emitted since compilers are conservative. Knowing your retardation, you probably don't know how to use your compiler, so I'm not surprised that you'd get hung up on technically having to write two instructions by hand (the horror). Tell me what this has to do with being an extremist assemblynigger who hasn't wrote anything besides a fizzbuzz, because it's not physically possible to write anything of value when you have to tinkertranny with something irrelevant like remembering how write dword is.

>>105762692
how wide dword is*

>>105762692
>There's hardware that doesn't have tzcnt instruction
>>105755502
>znver5

>>105762704
Yes, that's why you have to set -march flag to get what you expect, I know, the horror. It's much easier to rewrite everything in assemblyshit than tell compiler what hardware it should generate code for... Where do dumb niggers like you even come from?

>>105762623
Remember when I said "at least a little assembly," not "only in assembly?" Also, it's interesting that you mentioned prologues and epilogues because managing the stack and register allocation around function calls is actually one of the bigger things that you can optimize to be better, only in assembly - compilers do not do a good job here because they can't think in terms of the whole program, and don't automatically design functions with unique ABIs. Also, the inlining shit means nothing. Fairly basic assembly macros can fix the register/memory muxing problem. You'd know all this if you didn't have a very clear brainlet complex, getting filtered by instructions on a screen before you understood how to actually use them.

>>105762724
Just stop, autist. You've lost:
>no compiler generates a simple VPCMPEQB + TZCNT

>>105762739
Ok care to post code to prove it?

>>105762724
Get a godbolt compile to emit vcmpeqb and tzcnt without intrinsics/builtins. I don't care what -m flag you use.

>>105762768
Can you explain what is tzcnt for here?
Don't forget to do it without moving goalposts.

>>105762756
Prove what? The assembly macro thing? It basically just amounts to passing registers to be defined and defining them upon macro invocation to substitute key words in the function code. You use these keywords instead of the register names to write it. A basic sizing lookup deals with different sizes and you make multiple variants of the given keyword like: bkeyword, wkeyword, dkeyword, qkeyword, etc. It's as simple as that. You can write a macro system like this from scratch, in nasm, in an afternoon.

>>105762814
If I wanted to waste an afternoon on something worthless, I'd watch porn.

>>105762786
How do you not know what the connection between vcmpeqb and tzcnt is?
You've never even heard of vpmovmskb lmao

This is why you need to program in assembly, so you don't end up like this retard.

>>105762818
Why "'d"?
https://delamere.com/blog/is-porn-addiction-on-the-rise

>>105762831
Well, you see, you just moved the goalpost, and your concession is accepted.

And that's why we need to torture autistic people until they scream their lungs out. Literally. That would be so much fun to watch.

>>105762840
Because I spend my afternoons outside away from my computer, since I don't have to reimplement something worthless to make assemblyshit palatable. You have to be severely braindamaged to believe that wrapping useless bloat in macros is a solution to anything. Most functions are so small, that prologue and epilogue alone doubles the code size, and it could've been inlined by a compiler.

>>105762864
>moves goalpost from 2 instructions to 3 instructions
>has a melty when called out

>>105762876
>I spend my afternoons outside
Liar.

There is no melty. There is only your pain I take immense pleasure in. Ripping out every single nail out of your body and jamming them into your throat would be worth dying for.

>>105762892
Looks like a severe seething melty, don't forget to breathe deeply, or you'll pass out and hit your head again, brain damage is multiplicative.

>>105762850
I assumed, since you're shittesting me like a girlfriend with trust issues, that you already know what tzcnt already does. So I told you why your next retarded point about one being vectorized, and the other not, is retarded. I didn't move shit, you own goal'd without even realizing it. Unlike you, I'm not a faggot, so I don't play that haughty, passive aggressive snark shit you like to pull. Go suck your boyfriend's dick and shut the fuck up.

>the autist is an expert in brain damage
>to absolutely no ones surprise

>>105762883
>He still doesn't see the connection after it's borderline spelled out
ngmi you're only mad bc you're too dumb for assembly and this just confirms it

>>105762941
You only asked me to make a compiler emit vectorized cmp and tzcnt, which no compiler will do because they're unrelated and worthless in isolation, movmask was never mentioned. You're literally making up additional requirements on the spot. Also known as moving the goalpost.

>>105762956
You have yet to post code.

>the autist is so literally minded he cannot read between the lines
I cannot wait until the government permanently removes your ilk from society.

>>105762957
The fact that you don't understand this fairly useful and somewhat common avx/bmi idiom by heart, even though you "study compiler output" demonstrates your original point to be pants-shittingly retarded.
>>105762966
You have yet to demonstrate you'd understand it even if you managed to specify once what you wanted to see.

>>105762985
Government is likely to remove you first.

>>105762988
Sorry tranny, but going from 2 to 3 instructions is 33% increase in code size and changes semantics significantly, but I'm not surprised at all that assemblyshittery LARPer is only LARPing.

I'm not autistic, but even if I was I would happily get removed if it meant that only 1% of your ilk meets an abysmally gruesome, prolonged, and public death.

>Ignores my point completely
>Makes a non sequitur about code size
never change /g/

>>105763000
Based get

>>105763015
You have no points, you can only seethe that I nitpicked your retardation.
I was also hoping that you'd correct me on the 33% increase since it's simply not true, but you're just a fizzbuzzing tranny LARPer who knows nothing.

It's just a shame I wouldn't get to torture you personally. I would stick two needles into your eye balls and then transmit all of your posts directly into your brain, so that you can repent for your sins.

(Won't do much, but what else will you be able to do before your death)

Didn't correct bc it was irrelevant prima facie, so I don't care enough to even think about it.
You're the one larping, desperately needing to prove your assembly knowledge like the dunning kruger retard you are. You need to actually program in assembly or you will always feel this way when actually challenged.

>>105763047
>I-I-I TORTURETORTUREKIKILLKILLSEETHEMUSTDILATEAAAANOTTHEAUTISMOS
I accept your concession.

>>105763058
It's ironic that you'd call me a Dunning Kruger after I asked you to specify what instructions must be generated exactly, as if it didn't matter to you at all, despite you supposely being obsessed with every instruction being exactly as you want it.
Something tells me that you haven't touched assembly once in your life.

I regularly listen to the death screams in "Ready or not", to remind myself that this is what you will sound like for several hours.

You didn't ask me that.
You're just inventing shit you never said you autist fuck. Learn to communicate.

>>105763091
Sorry, I sometimes forget that Wernicke's Aphasia is terminal.

>autist fixates on the word "Wernicke" to calm himself down
You'll need a lot more than that once society is through with you.

>>105763132
What do you mean by that? How can I possibly "fixate" on the condition you have? It would be better for everyone if you didn't have this specific illness, or better yet, just killed yourself, but apparently mentally ill subhumans are "people" to so we must pretend that there's nothing wrong with your braindamaged subhuman faggotry to shit up this thread daily.

>>105763114
It's ok bro, nobody's really judging you for your condition, just your homosexuality.

>>105763144
White people stopped caring about gays long ago, get with the times, turdworlder.

>>105763149
Speak for yourself, kike.

>>105763154
Israel cares about gays a lot, because it's brown, just like you.

>it is so autistic it doesn't know what "fixating on a word" means
>apparently it's just natural for it

>>105763166
It's not up to me what's the name of the illness that you have.

>>105763162
I'm sure *you* know what this was supposed to mean. I'm so sorry for your caregiver.

>autist is projecting
No wonder most human beings want to see you smeared across the streets.

>>105763179
Yes, I know that mentioning gays out of nowhere is proof of severe swarthiness.

>>105763283
It's a throw away insult, why are you autistically hyper-analyzing it? Stop being a faggot.

>>105763317
>It's a throw away insult,
I don't feel insulted, in fact I accept your concession and your brownness.

>autist is unable to process emotions
Checks out yet again

>>105762438
>Lack of common lag in most corpo IDEs
like the other anon said: get a better computer
the only "slow" part is startup time but once started, IDEs don't lag on a modern computer. both typing and actions are immediate

>Adding the basic vimkey bindings alone are actual cope from IDEs to stop the user hemorrhage
no one's coping except vimfags. IDE people use their IDE's shortcuts, possibly modified with a few of their own. in the end it's still the same actions, just different key combos to trigger them
IDEs don't have "user hemorrhage" - if anything, IDEs are getting more users over time

again, vim has nothing on IDEs

>>105762471
>Vim's modal editing is what's superior
feel free to explain how having to toggle modes to do certain things is better than just being to do any action at any time

>>105763704
Even simple things like undo are superior when you control what exactly the undo will do when you need it.

>>105763742
looks like just another "neat but not very useful in practice" thing
my IDE has more advanced history tracking and control, as well as quick refactorings that go both ways (eg. inline and split/extract) so I can track changes better than when editing history/undoing things (in which case you lose track of things you have done but then undid)

>>105763924
It shines in projects bigger than fizzbuzz, so I'm afraid I cannot explain to you, because it's like explaining what sex feels like to a virgin.

No one cares about your faggot experiences, autist.

>>105763947
that's some cope and irony, given multi-project-wide change tracking is exactly what shines in large projects
>the "virgin" argument (ad hominem)
concession accepted

>>105764133

>>105764110
The irony here is that you'd depend on your IDE for something that should be in a version control system, while the topic is simple undo functionality for you know what. Oh wait, you don't know.