Programming thread

  • Want to keep track of this thread?
    Accounts can bookmark posts, watch threads for updates, and jump back to where you stopped reading.
    Create account
Apparently Mozilla's attempt rewrite GNU Core Utilities in Rust has hit a snag.

Ironic considering the project lead went on stage a few years ago and claimed a speedup with sort.

photo_2025-09-16 14.50.59.webp photo_2025-09-16 14.51.03.webp
 
Anti Snigger said:
17x slowdown is a small price to pay for SAFE code, chud
Also lmao @ the issue poster for using a non LTS version
Click to expand...
I get why he is using the development branch, Ubuntu is going to replace GNU Core Utils with the Rust version.

I'll hand it to Canonical, they keep giving people more and more reasons to not use Ubuntu.
 
ellroy said:
Apparently Mozilla's attempt rewrite GNU Core Utilities in Rust has hit a snag.

Ironic considering the project lead went on stage a few years ago and claimed a speedup with sort.

View attachment 7924643View attachment 7924645
Click to expand...
another effortless W for the GNU system
every time a cuck-licensed rustlet fails this hard, the hurd gets a patch that lets it run 2% more of the debian archive

also it's funny how unimaginative rust is, they could have included some sort of totality checker that could have helped prevent some infinite loops like this one, but they didn't because rust was designed by AMATEUR LANGUAGE DESIGNERS who decided to make an exact copy of c++, down to the incredibly long compilation times (but it forces you to use smart pointers really hard, and has a worse abi)
Anti Snigger said:
17x slowdown is a small price to pay for SAFE code, chud
Also lmao @ the issue poster for using a non LTS version
Click to expand...
at this point why not write it in an actually good application language like common lisp
i would use a set of coreutils written in lisp if the implementation was properly tuned to start incredibly fast

i guess having a little garbage collection is absolutely horrible when allocating and deallocating 5,000 small objects on the heap malloc-style is definitely so much faster right
 
DumbDude43 said:
i haven't use rust a lot myself, but i imagine having to debug anything serious in it must be a total nightmare
Click to expand...
i've heard horror stories of a dependency of a dependency of a dependency having nasty segfaults (ironic)

to its credit i will believe the rustniggers who say it tends to work well the first time when the compiler doesn't reject your program
some of the (limited) ideas in rust would be applied excellently in a better language
 
Rust the language is very usable. They managed to stitch together a whole lot of shitty code in the compiler, but the tooling is nice from a user perspective.

My issue with it is its ecosystem and the mentality of Rust Trannies. They have made the entire ecosystem full of mediocre, over-engineered packages. Each depending on dozens of other packages. Also to blame are the many novices who use Rust as their first language, because hey, it's the BEST language right? Or at least, that's what they're told.
 
if you dont like rustc you can always use grust from gcc
it doesnt have a borrow checker iirc, but it does implement the latest (at the time of gcc 15 dropping) version of rust
 
Rango Dango said:
if you dont like rustc you can always use grust from gcc
it doesnt have a borrow checker iirc, but it does implement the latest (at the time of gcc 15 dropping) version of rust
Click to expand...
why would you use rust if you aren't gonna use the borrow checker? all the grandiose promises about memory safety go out the window if you drop the borrow checker, and that's the only reason for using rust in the first place. might as well just write C (++) at that point.
 
DumbDude43 said:
why would you use rust if you aren't gonna use the borrow checker? all the grandiose promises about memory safety go out the window if you drop the borrow checker, and that's the only reason for using rust in the first place. might as well just write C (++) at that point.
Click to expand...
i mean i didnt say it was a good idea, but it still is a possibility if you wanna just fiddle with the language
 
Lard Shart said:
My issue with it is its ecosystem and the mentality of Rust Trannies. They have made the entire ecosystem full of mediocre, over-engineered packages. Each depending on dozens of other packages. Also to blame are the many novices who use Rust as their first language, because hey, it's the BEST language right? Or at least, that's what they're told.
Click to expand...
this is actually my biggest issue with rust tbh
it would be an ok language if it didn't pervasively want you to statically link 47 dependencies, with none of them having anything resembling a stable abi
DumbDude43 said:
why would you use rust if you aren't gonna use the borrow checker? all the grandiose promises about memory safety go out the window if you drop the borrow checker, and that's the only reason for using rust in the first place. might as well just write C (++) at that point.
Click to expand...
well c is kind of slightly too small for writing complex things and c++ is a bit of an abomination from hell
so rust-sans-borrow-checker might be sort of good idk
also if grust is written in c it would be good for bootstrapping the llvm-based rust compiler
 
for me personally i would just want c with a saner function as a type syntax, differences of character encodings between linux and windows abstracted away, and maybe foreach cuz its clean
 
Rango Dango said:
c with a saner function as a type syntax
Click to expand...
yeah this would be ideal tbh
prescheme is basically this, i hope it gets properly revived eventually
Rango Dango said:
differences of character encodings between linux and windows abstracted away
Click to expand...
this is really hard to abstract away
if we're talking about ideal c replacements i want to go all the way on the "impossible wishes" train and i wish for the complete deprecation and obsolescence of windows
also unix doesn't really have a "character encoding" per se: file names are any byte other than \0 and '/' and everything else is pretty much just byte streams
it just happens that most things conventionally use utf-8 and lf line endings
Rango Dango said:
maybe foreach cuz its
Click to expand...
fuck foreach, i hate gay nigger syntax sugar, give me a FUNCTION called foreach and not some dumb fucking syntax
if you want you can write a gay macro that implements your idea of foreach but don't expect anybody to use it because higher-order functions are just better than adding redundant shit syntax for no reason
in fact if your language doesn't exclusively do looping through functional tail calls, it's a shitty imperativecoal 1950s brainrot abomination
Rango Dango said:
clean
Click to expand...
clean by indian standards i guess
 
canalCrab said:
CSS is really not that hard, you just gotta sit down and actually bother learning it for a few hours
Click to expand...
> https://www.w3.org/Style/CSS/specs.en.html
> few hours
Right...

But if one were to read all that, surely he could express anything with elegant simplicity, right?
(Solution to width dependent height)

Since height percentages are relative to the height of the parent element, we can't rely on it. We must rely on something else. Luckily padding is relative to the width - whether it's horizontal or vertical padding. In padding-xyz: 100%, 100% equals 100% of the box's width. Unfortunately, padding is just that, padding. The content-box's height is 0. No problem! Stick an absolutely positioned element, give it 100% width, 100% height and use it as your actual content box. The 100% height works because percentage heights on absolutely positioned elements are relative to the padding-box of the box their relatively positioned to.

- Alexis Wilke
Click to expand...

Ok, fine, I'm being overly spergy. While it certainly is gigabloated, one can ignore it.
However, the real problem is that the hundreds of properties (how many hundreds? I don't know, I don't think anyone knows how many there are today, but I digress) each have different inheritance, dependencies and most importantly: interactions.
Every time you fix something with CSS, it carries the risk of breaking two other things.

I am sincerely impressed that there are mortals that wield CSS effectively.
My simple programmer brain is not fit for that kind of strain.
 
Gentlemen of the programming thread, I have a conundrum.

In my C project, I am using the terminal / command prompt shells to render my project. The issue is, when working on it on my main rig, Windows, it's a tad slow, and the text scroll I have implemented does it job nicely. However, on Linux, Fedora specifically as I haven't tested other distros, the text scrolls at what I believe to be the screen refresh rate, making it otherwise unreadable. I was told I should consider ncurses, but I am not entirely sure how I would go about using that, as the ncurses website, this is the one I used. I'm not sure how to go about integrating it with my project in Visual Studio 2022 (I like it, leave me alone) so I can use it on windows, then go about similarly on linux... maybe it's the same solution and I'm just retarded. Many such cases. Any suggestions?
 
agent_donkey said:
> https://www.w3.org/Style/CSS/specs.en.html
> few hours
Right...
Click to expand...
You don't have to know all the intricacies of the specs to be able to use it competently. Just like you don't have to know ALL of the JS spec to be able to use the good parts.
iScream into the void said:
Gentlemen of the programming thread, I have a conundrum.

In my C project, I am using the terminal / command prompt shells to render my project. The issue is, when working on it on my main rig, Windows, it's a tad slow, and the text scroll I have implemented does it job nicely. However, on Linux, Fedora specifically as I haven't tested other distros, the text scrolls at what I believe to be the screen refresh rate, making it otherwise unreadable. I was told I should consider ncurses, but I am not entirely sure how I would go about using that, as the ncurses website, this is the one I used. I'm not sure how to go about integrating it with my project in Visual Studio 2022 (I like it, leave me alone) so I can use it on windows, then go about similarly on linux... maybe it's the same solution and I'm just retarded. Many such cases. Any suggestions?
Click to expand...
make the text scroll work on delta/elapsed time, rather than just outputting it as fast as possible?
 
agent_donkey said:
Ok, fine, I'm being overly spergy. While it certainly is gigabloated, one can ignore it.
However, the real problem is that the hundreds of properties (how many hundreds? I don't know, I don't think anyone knows how many there are today, but I digress) each have different inheritance, dependencies and most importantly: interactions.
Every time you fix something with CSS, it carries the risk of breaking two other things.

I am sincerely impressed that there are mortals that wield CSS effectively.
My simple programmer brain is not fit for that kind of strain.
Click to expand...
how to css: save, refresh, see that the thing looks weird, look up some shit, repeat
iScream into the void said:
Any suggestions?
Click to expand...
you deserve everything that is happening to you :smug:
but for real i think there is one implementation called ncurses for gnu/linux and another that you use if you're trying to build for windows for some reason
if you know how to use the linker it should be pretty simple. i have no idea how your shit ide does it but on my machine i can invoke something like gcc `pkg-config --cflags --libs ncurses` -o program program.c and ncurses will be properly included
if you don't want to use a library, you could try the humble art of using ansi terminal escape sequences for shit
 
You must log in or register to reply here.
Back
Top Bottom