Thread 105682565

>>105685575
What's unreadable about it? It's only a compound type.

>>105682860
>oh my god a hello world that weighs a whooping 4 megabytes?
>and compilation artifacts take a lot of space too!
>the LSP is freezing my IDE!

>my code is shit? Heh, who cares, my PC is good enough for it anyway

>>105685575
>my valuables, in a box, on a seat, of a car, on the road, on earth? What the fuck does that mean?
>oh you mean ConstrainedValuableSecureAutomotiveRoadTransportationPlanet? Then why didn't you just say so from the start?

>>105682602
>Jai
jai sri ram saaaaaaar

>>105682910
Meanwhile C++ has no std::net or std::process library

>>105682565 (OP)
yes, and?

>>105687028
Which is a good thing.

#![no_std]
#![no_main]

use core::panic::PanicInfo;
use core::arch::asm;

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


#[unsafe(no_mangle)] // unmangled corpses very unsafe
pub extern "C" fn _start() -> ! {
let msg = "Hello, World!\n";
unsafe {
asm!(
"syscall",
in("rax") 1,
in("rdi") 1,
in("rsi") msg.as_ptr(),
in("rdx") msg.len(),
options(nostack)
);
asm!(
"syscall",
in("rax") 60,
in("rdi") 0,
options(nostack, noreturn, nomem)
);
}
}

I found a way to make troons and chuds seethe. Reminder: resulting binary is smaller than equivalent C.

>>105687948
> uses inline assembler
> wow! it's small!
so it fucking should be. faggot.

>>105688158
Not an argument. C binary is around 400 bytes bigger.

>>105688177
> 400 bytes bigger
how? did you strip the C object of its symbols and other trash?

>>105688244
I programmed in C for longer than you will in your fizzbuzzing career, you don't have to question me.
>4416 target/x86_64-unknown-linux-gnu/release/nostd-project
llvm's ld further halved the binary size without using mutilating flags such as --nmagic

>>105688276
>I programmed in C for longer than you
you don't know who anyone is. do you have down syndrome?
> you don't have to question me.
i will do whatever i want, especially when you share no c code, or compiler args for c. also, wouldn't it be smaller still without the panic handler? i don't know why the fuck you'd need that if you're just loading up registers and printing text. is that a rust tranny thing?

>>105688330
I know that you're a nobody with no useful opinion of any kind just from the way you write your posts.
>I don't know why the fuck you'd need that
try reading the manual

>>105688344
> can't explain it
so the façade of being this expert programmer sure slipped the fuck off your obese tranny face very quickly. i can't imagine knowing so little about a language you're trying to shill. did you copy paste this code from stackoverflow or was it chatgpt?

>>105688365
I don't get paid to educate illiterate special needs.

>>105685297
>without unsafe code.
Code that does nothing is indeed safe.

>>105685297
wrong
pub trait P: Fn() -> T {}
impl T> P for F {}

fn a P<&'a B>, B>(_: A, a: A::Output) -> &'static B {
a
}

pub fn main() {
println!("{:?}", { let v = vec![vec![1]]; a(|| panic!(), &v) });
}

Exited with signal 11 (SIGSEGV): segmentation violation

https://play.rust-lang.org/?version=stable&mode=release&edition=2024&gist=31dea368e3d7ab63a9ff7830a9126afd

>>105688950
Nobody would write this.

>>105688978
I did

>>105688983
Ok, I'm moving back to C.
t. >>105687948

>>105682565 (OP)
The first four are especially problematic for Rust. Some of this can be solved by using Crust (tsoding's invention)

>>105689029
kill yourself

>>105682565 (OP)
>target folder is larger than node_modules
This is verifiably false. How can anyone even think this, have you used JavaScript recently?

>>105687948
>aarch64 compiler
>code goes ACK

>>105691198
PR welcome.

>>105687906
List of people who will never be women:
> (You)

>>105691364
Women want things done for them, men do things themselves from scratch.

>>105691376
Then go program in assembly, midwit.

>>105691376
Or better yet go program using vacuum tubes.

>>105691483
>>105691493
I accept your concession, tranny.

>ur a troon
>no u
Massive homolust aura on both sides. Rust is tainted and so are C++ and C, you can't escape it. That's why you're all acting like women, why cniles spent decades harassing anyone who didn't bend the knee to their lord of savior.
Let go of these emotions and come back to pascal.

>>105691780
>pascal
Which one

>>105682901
>If you can't have multiple owners to a single mutable data source
Reference counting smart pointer. Same thing you would do in literally any language that supports RAII

>>105683824
C, Rust, brainfuck, redstone and virtually every programming language is Turing complete.

>>105682565 (OP)
The only argument i need is:
It isn't fun.

I don't want to hate my own existence.
And as if the language wouldn't be bad enough, most of the Rust projects are rewrites!
There are people sitting on their PC, looking at C code in one window and writing equivalent code in Rust in the other.
WHO THE FUCK WANTS TO LIVE LIKE THIS?

Did anyone ever hear someone saying
>i enjoy rewriting existing code into Rust
i surely didnt

>>105682690
look at this troon calling others nocoders, while not even knowing what a memory leak is

>>105682565 (OP)
Rust is like Ada. It sounds good to managers.

>>105692095
>And as if the language wouldn't be bad enough, most of the Rust projects are rewrites!
>you wrote a text editor
>EDLIN is a text editor
>YOU'RE CODE IS HECKIN REWRITERINOS OF EDLIN!

>There are people sitting on their PC, looking at C code in one window and writing equivalent code in Rust in the other.
Two different programs that have the same function are not "rewrites." It's like saying all video games are "rewrites" of Pong.

>>105682565 (OP)
my software works
you can't even get a working build system or dependency management.

cope and seethe shitter.

that's my response.

>>105692095
>It isn't fun.
Debugging and writing cmake files isn't fun.
Debugging runtime crashes isn't fun.
Writing header files isn't fun.
Reimplementing hashmaps for every type isn't fun.
Memorizing every library's API semantics isn't fun.
Writing rudimentary code generation isn't fun.
Rust is fun though.

>And as if the language wouldn't be bad enough, most of the Rust projects are rewrites!
So like literally every programming language?
Do you think that majority of projects written in C++, JS, Java, etc are some sort of totally original ideas? Everyone rewrites stuff.

>There are people sitting on their PC, looking at C code in one window and writing equivalent code in Rust in the other.
>WHO THE FUCK WANTS TO LIVE LIKE THIS?
I never met anyone who does this. Are you talking from experience or what?

>Did anyone ever hear someone saying
>>i enjoy rewriting existing code into Rust
>i surely didnt
Me neither. Maybe because barely anyone actually does this and instead go like "oh I like this C/Python/Java/Bash utility, I'm sure I can write something like that but better in C/Python/Java/Bash that I like more!"

>>105692173
People who wrote first pong had much more fun than any of us ever will.

>>105692194
>Writing rudimentary code generation isn't fun.
Rust requires this for every little thing since I can't just include native OS header into Rust and have it just work.

>>105692136
>Rust is like Ada. It sounds good to managers.
Ada is corporate because if its shitty early licensing and DoD roots.
Rust is much more popular in hobby space than it is in corporate.
Neither of these languages "sound good to managers". They both require expensive engineers and are not friendly to agile development. Go is what sounds good to managers. It was designed to appeal to inexperienced developers.

>>105692209
>Rust requires this for every little thing since I can't just include native OS header into Rust and have it just work.
https://github.com/rust-lang/rust-bindgen
You just need to generate bindings. No need to write your own code generation to include C headers.

>>105692230
No thanks, I'll stick to C.

>>105692236
Who asked?

>>105692241
I did, check your emails.

>>105685297
It hasn't been formally proven. That's what the whole RustBelt thing was for.
>>105688919
Vast majority of programs don't need to use unsafe directly. It is usually confined to some library, and even then, the library might not need to use it, and if it does, will confine it to a tiny corner of the implementation.

>>105692342
>directly
nice backpedal

>>105682565 (OP)
i don't like rust but that person is a retard self-reporting that he sucks at writing software.

>>105682860
everything is security sensitive. welcome to the internet age. you don't want your software to be part of somebody's exploit.

>>105692376
>everything is security sensitive
not games

>>105692355
I'm not who you were replying to. Yeah, every program that uses Vec is indirectly using unsafe indirectly, but I'm pretty sure that's not you guys mean when you say that you can't write useful Rust programs without using unsafe.

>>105692402
not him but yes thats what is meant when that statement is said in posts

>>105692402
Yes I do.

>>105692416
>>105692417
Well that's just silly then.

>>105684566
rust by itself, is a stepping stone language wrongly championed as the future.
There's much to learn from it, but we definitely shouldn't be rewriting important things with it

>>105682901
>Does the borrow checker prevent Rust from being turning complete?
no. stop acting retarded.
>If you can't have multiple owners to a single mutable data source then doesn't this make certain algorithms impossible to express in Rust?
turing completeness is about the possible domains for input-output mappings, but even still this is a "no".
>Or is there a way to turn off the borrow checker when needed so you can express these problems?
yes, but it's completely irrelevant because substructural type systems are irrelevant to turing completeness.

>>105692383
if somebody uses your game to cause damage, you're not going to like what happens to your company. if the damage is severe enough, you might just get assfucked by the government.

why isn't there a native language like C#? they all suck

>>105692434
or nboody will care at all not that i work on games i have real shit to do

>>105692429
>rust by itself, is a stepping stone language wrongly championed as the future.
>There's much to learn from it, but we definitely shouldn't be rewriting important things with it
It's much better than C and C++. C and C++ were worse than older languages but wrongly championed as the future.

>>105692464
maybe nobody cares at all that your insecure kernel cost somebody millions of dollars. maybe you'll make it big in the casinos so you should get out there and put your net worth on the tables.

or maybe you could listen to me and realize that this is this is risk management 101, and if you are programming in a professional capacity, yes this is absolutely your fucking problem no matter the industry or how trivial you think your software is.
you have to worry not just about public image, but that being taken to court is a catastrophe and the litigious nature of society makes that a guarantee. jfc.

>>105682910
portable simd is a meme. it's basically LLVM hints at this point. doubtful it will ever actually be anything more.

>>105687906
>w-which is a good thing
not everyone wants to use libc to resolve names and crawl some shitty linked list. there is no excuse why the C++ standard library can't have these basic niceties instead of coping by deferring to libc.

>>105688244
>how?
rust can actually be freestanding without cope flags that don't work nor work portably across C compilers.

>>105692456
because the semantics of the language assume a garbage collector. the closestst C# like language you'll find that's competitive is vala, which compiles to C + glib

>>105692456
you're just a midwit who's afraid of learning new things

>>105692754
C# has pointers, ref and unsafe keywords, and most of the semantics that assume a garbage collector would work just as well with RAII.

>>105692865
ok and? it doesn't refute what I say. you still need a GC and Vala exists.

>>105692949
Yes, it does. RAII is not a GC. There is no reason why a C# like language without a GC can't exist.

>>105692974
which is why Vala exists ffs.

>>105692456
probably because nobody likes c# except for the stockholm syndrome retards who were forced to use it. "retards" because the non-retarded ones eventually learned how to think about problems in ways other than the grain exposed by this one piece of shit microsoft abomination.

>>105685239
This is why letting web devs into embedded is dangerous, rust advertises itself as safe, but it's just empowering retards like you to write unsafe code, because you niggers don't know what you're doing

>>105683824
>C is not a turing complete language either.

>>105693648
That's correct.

>>105693656
I admire your ability to be confident in things you know nothing about

>>105693668
Actually, I know more about it than you. Next time research the topic before you open your mouth.

>>105693656
>

>>105693718
Can't for a counter argument?
Maybe you should try reading books instead of being a meme warrior.

>>105692598
>that being taken to court is a catastrophe and the litigious nature of society makes that a guarantee
And now you understand why every license ever has a line like
>you agree that if shit gets fucked up it’s not our fault, remember you chose to use this software
somewhere in it

>>105693731
you can simulate a turing machine in C
it is turing complete
its not even an argument, none are needed
youre just plainly wrong

>>105693743
He’s trying a gotcha on the technicality that C does not have access to infinite memory. Because of that nobody has ever built a “real” Turing complete language but it can be proven that if the memory limitations were removed C could indeed simulate any Turing machine

>>105693778
>jazz music stops
is he fucking retarded?
that technicality is void because the language doesnt define the properties of the hardware
the only thing that matters is whether said language can be used to simulate a turing machine given the appropriate hardware

>>105693796
Yeah he’s that stupid. And to actually my actually, the real requirement for true Turing completeness is access to an “arbitrarily large” amount of storage. It’s a technically finite number but you can always find a valid program where it’s not enough

>>105693818
>And to actually my actually, the real requirement for true Turing completeness is access to an “arbitrarily large” amount of storage.
thanks for the clarification
have a you
im a self taught so i do have holes in my knowledge

>>105682565 (OP)
>90% fighting the compiler, 10% writing code
dead giveaway of major skill issue

>>105693869
unironically yeah
but this does mean the barrier of entry is high
which makes the language less attractive for corpo given that the mission is to saturate the workforce

>>105682901
this fucking answers..
rust supports pointers. like real pointers a la c. pointers dont use borrow chcker

seriously this board talks about shit they have no clue

>>105692456
Because C# is a managed language.

>>105694029
yeah
we have had an afflux of new C programmers because the poltards who came here *did* learn stuff
and rust is an lgbt cause so plenty of people associated with that group come here to talk out of their ass for virtue points
(poltards shilling c just for the political statement stop rather quick bc they get btfo by both sides. we cultists tend to have a notion of decorum so our community tends to police itself.)

>>105693778
>Because of that nobody has ever built a “real” Turing complete language but it can be proven that if the memory limitations were removed C could indeed simulate any Turing machine
The memory limitations are built into C. There are no such limitations in Lisp, Haskell, Python, and many other languages.

>>105694114
>The memory limitations are built into C
explain in detail or horseshit and lisptrany cope

>>105693624
NTA, but due to forced RAII, memory leaks are not really a common problem in Rust. Unless you are a retard who spams counted shared references in cycles without weaks.
Also webdev influence(async) has been a blessing for embedded(embassy)

why are they like this, bros?
>>105694114
im all for the back and forth of problems vs solutions but they just talk out of their ass

>>105691515
Post code then wise guy

>>105692456
C# is a M$jeet goyslop version of Java.

Rustfmt is so damn good, but I notice tiny inconsistencies:
// There is a trailing comma.
some_func(
some_arg1: i32,
some_arg2: String,
some_arg3: &mut HashMap,
)
// There is **no** trailing comma.
assert_eq!(
some_value,
SomeEnum::SomeMember(SomeOtherEnum::SomeOtherMember)
);

>>105692456
https://www.beeflang.org/

>>105682565 (OP)
All correct excpet the last bit. It was made by trannies, for trannies.

>>105687948
How small?
It's trivial to make a Hello World exe that smaller than 1kb in C.

>>105688978
I guess math isn't real.

>>105694813
It was made by chads you wish were trannies, for chads you wish were trannies.

>slow compiler
This is why I don't get these troons' obsession with trying to use Rust for UI dev. Modern UI frameworks like Flutter and even the new Android stuff support hot reload/hot restart which only works with a fast compiler and a VM in the case of hot reload. Like do they realize that not everything has to be in Rust? I mean if you have a hot path somewhere and want to write it in Rust, sure go for it but in 99% of cases it makes no sense for UI.

>>105694126
A pointer in c has a sizeof
A byte in c has a finite number of bits.

>>105694878
yes, a static language will not have the introspective/reloading capabilities of a live language unless you write a VM for it.
I do not do my UI dev at the same time as my backend dev, however. I write basic bitch HTML and CSS and just have a bunch of sample documents containing made up data. if you are parsing JSON or whatever, there is nothing stopping you from writing mock server responses in your experiments for seeing what it looks like.

>>105694916
> if you are parsing JSON or whatever, there is nothing stopping you from writing mock server responses in your experiments for seeing what it looks like.
Not trying to be rude but is this just a non sequitur? Or what point are you making here.

>>105694889
yeah
just give it more bits
and nothing says you couldnt have a an architecture which accepts arbitrary length adresses
in the you then keep the number in your code in an extensible object and feed it to the cpu piecemeal, whats the problem with that?
>inb4 but what if its an infinite ram, ahah!
infinte ram has room for an infinitely long number, no?
AND an infinite amount of data, thats kinda the deal with infinity, no?

>>105692456
because managed > native

>>105694948
>and nothing says you couldnt have a an architecture which accepts arbitrary length adresses
No, but the C standard doesn't allow it.

>infinte ram has room for an infinitely long number, no?
>AND an infinite amount of data, thats kinda the deal with infinity, no?
That's great if you're using most languages, but it's not allowed by C.

>>105694966
>No, but the C standard doesn't allow it.
+
show me where it says that
+
you can make it work with file descriptors, ackshyually

>>105693743
You can simulate Turing machine in Rust as well. Learn CS.
>>105693778
How is this a gotcha? I am not the one who thinks C is Turning complete.
>>105693796
???
What argument are you trying to make here? There are turning machines written in Rust as well.

>>105694999
>strawmen
>reee
>convulsions
why are you like this, anon?
yes, c is turing complete
i didnt say anything about rust yet
but since you insist...

>>105695015
You are replying to a conversation that mentioned Rust.
>Strawman
No. C is not Turning complete by the very definition. This is why nocoders get filtered. You never make it because you don't know what you are talking about.

>>105694948
You can give it arbitrarily many bits, but you can't give it infinite bits
The fact that c is designed for machines that actually exists is a serious comment, but I don't see it as a critical one.

>>105695113
>No. C is not Turning complete by the very definition
>still didnt post where it says so in the standard
>>105695172
>but you can't give it infinite bits
you dont need them with fseek and fread
you use fwrite to write to the "tape"

>>105695172
also
>but you can't give it infinite bits
why?
you got an infinite tape
it is gonna take an infinite amount of time to traverse with any turing physical-ish machine
that it takes an infinite amount of time to pass an address is irrelevant because inf + inf = inf
its one of its properties

>>105695307
Turing machines can take indefinite amounts of time to halt, but not infinite time. You need indefinite memory to run them, but not infinite memory. It's an important distinction.

>>105694992
>you can make it work with file descriptors, ackshyually
File descriptors are not part of C, not memory, and int in C is not allowed to be infinite, so you can't have an infinite number of file descriptors anyway (real Unix systems limited them to about 10).

>>105695286
>you dont need them with fseek and fread
>you use fwrite to write to the "tape"
They're still limited because fseek, fread, and fwrite can't be infinite.

>>105695473
you alr told me :)
i was indulging our crustacean friend here in an "but if we made the tape infinite anyways" scenario because i thought that you can still do that with C and i dont mind someone hellbent on failing my theory to run the numbers, to the contrary
>>105695483
>File descriptors are not part of C,
defined in iso
https://www.dii.uchile.cl/~daespino/files/Iso_C_1999_definition.pdf
read and open are defined in ansi i think
fd are system-defined but you need only one

>>105695490
>They're still limited because fseek, fread, and fwrite can't be infinite
heres the trick: fread/read advances an internal, abstract pointer so to speak by how many bytes you have read
you dont need an infinite number, you fread repeatedly an infinite amount of times
and you dont even need a number to rewind back to the beginning
>but what about reading PAST the beginning
then you write your program at the infinite place in your tape.
then when you rewind to the beginning, you still only need to read forward to read all your program, even if written backwards, and of infinite size
because infinities

This guy is retarded and should be ignored

>>105692342
>Vast majority of programs don't need to use unsafe directly. It is usually confined to some library, and even then, the library might not need to use it, and if it does, will confine it to a tiny corner of the implementation.
Tiny corner, like all over the place in Rust stdlib?
https://materialize.com/blog/rust-concurrency-bug-unbounded-channels/
Oh, and if some safe Rust code has a bug, and unsafe Rust code relies on that safe Rust code being correct, you may end up in UB territory regardless.

>>105694029
Does Rust require unsafe to dereference a Rust pointer? If yes, then you went into unsafe Rust, which isn't the nicest place to be.

>>105694878
>obsession with trying to use Rust for UI dev.
Name some GUI programs written in Rust.

>>105694889
What is stopping you from having arbitrarily large pointer size and size_t?

>>105695843
It's probably some angry cnile who reposts /g/ shitposts to whore for attention lmao

>>105695843
tests like this are retarded anyway
a compiler could just do the calculation at compile time (like constexpr in c++) and then execution time would be almost 0
doesn't mean the language is fast or slow

>>105684044
>I was banned from using GNU/MIT software
damn bitch, u THAT ugly?

>>105692865
yes and it still assumes GC

>>105692754
shame vala is useless for anything besides writing gnome shitware

>>105682602
jonathan blow spends all his time losing arguments on twitter now, not coming

>>105692456
BeefLang is exactly what you're looking for.

>>105692456
There is and it's called C# with AOT compilation

>>105696990
It still has a GC even with AoT, tardo.

>>105696495
not the OP, and don't know of much about that, but there sure are a lot of prototype Rust UI libraries
https://blog.logrocket.com/state-rust-gui-libraries/

>>105696280
yes it does, just like in c you should always check for nullable before dereferencing it. thats why in c and c++ using handles is considered better than raw pointers. rust just forces you to put and unsafe around it as a reminder.
just like swift and kotlin also force you.

rust doesnt even "force" you, it just requires unsafe. this is a good thing, even linus agrees and demands. nobody uses pointers without checking for nullable thats BAD CODE.

and my reply was to a post about using aliasing without turning off borrow checker, not aboud not using unsafe.

>but using unsafe is le bad!
literally ehat you think youre doing BY DEFAULT when using c and c++?
fyck me you guys are retarded. ive written thousands of c and c++ loc in the last 30 years and ypu retards make me feel ashamed of it. youre all so fucking stupid.

>>105696495
https://github.com/pop-os/cosmic

>>105697982
>>105698404
>few early libraries
>one shell extension
And that's what you call an obsession with UI? What?

>>105697988
>literally ehat you think youre doing BY DEFAULT when using c and c++?
How often do you use the restrict keyword when developing in C? What are your thoughts about tree borrows and pinning?

>>105697988
>yes it does, just like in c you should always check for nullable before dereferencing it.
Checking for null is insufficient for Rust pointers.

>>105698580
This, so much this, you must also check if the pointer is in range of program segment OR in one range of memory maps AND that it was written to already so it is fully initialized.

>>105698609
No, more other things that have to be considered.

>>105682591
Whoever made this gif HATES trannies with unhealthy passion

>>105698609
https://google.github.io/comprehensive-rust/unsafe-rust/dereferencing.html

>>105698650
If trannies were an issue with Rust, and the only issue with Rust, Rust could just be forked.

>>105698609
>>105698657
Most of these things can be differed from the type system invariants. That's probably what that anon meant.

>>105698678
You can just fork it. No one does this because it's not an issue.

>>105682602
Jon is going to release it in sync with his game, he's hinted at it as a marketing strategy for both things

>>105698741
One question is, are there (possibly fundamental) issues with Rust that forking will not fix?

>>105698773
What kind of retarded question is that? C and C++ both have unfixable fundamental issues just like everything else.

>>105682565 (OP)
just use ChatGPT directly as a llvm frontend. programming is a dated concept

>>105698824
>chatgpt as a llvm frontend
The state of /g/

>>105698824
>ChatGPT directly as a llvm frontend
this is the most hilarious words combo i've ever read in this current year

>>105698824
>using chatgpt to generate llvm bitcode
not sure if retarded or giga based.

>>105684566
>corona/mRNA vaccines for example
supercancer in 14 days

It's insane how we have to move to this slower, less readable language full of vulnerabilities just because zoomers can't into C and their discord has the 2SLGBTMAP+ logo. Embarrassing.

>>105687028
asio exists for sockets

>>105701039
>slower
[benchmark needed]

I don't like that Rust takes away control from me. I don't get why Rust programmers want to give up control over their program to a static analysis tool written by someone else. They have the same mindset as vibe coders where they both think the machine they are using knows better than them.

>>105701676
As soon as you want to do anything with multiple mutable owners via a Ref>, which seems to be the recommended way to do said thing, that part of the program is using garbage collection which slows down run time speed. Also Rust default allocator is mid at best.

>>105698799
And what if those unfixable fundamental issues include deal breakers for some cases? Or are sufficiently bad that an alternative or successor language could stay or become very popular?

You ought to have been able to infer that from the context. And speak about yourself better.

>>105698754
If it takes 10 years to make a shitty puzzle game, isnt that bad publicity for the lang

>>105702291
wtf I love AI now

>>105702612
>Sokoblow
lmao

>>105702694
I will pay AAA price for an indie game if he actually names it Sokoblow.

>>105701838
[benchmark needed]

Alternatively, write any reasonably sized program in any language of your choice that you think cannot be expressed in Rust without some significant performance penalty. Also include compiler flags.
I will write my own version in Rust and we will do a benchmark.

This is the only sensible way of arguing over performance difference.

>>105702612
>>105702291
That faggot is making a sokoban game?
I was here thinking he was making some Baldur's Gate type shit; what's taking him so long then

>>105702789
nta. I rather keep my sanity and be more productive in C++ even if the end result is <3% slower (which I doubt)
>>105702806
arguing on twitter.com, mostly.
plus he still uses that external gui grep program instead of basic emacs features like rgrep or etags. No wonder he gets nothing done.

>>105702882
>I rather keep my sanity and be more productive in C++ even if the end result is <3% slower (which I doubt)
C, C++ and Rust all are roughly same speed in practice. You can literally manipulate rustc to generate nearly identical LLVM IR to clang. Matching performance is just matter of putting enough effort into it.

>>105694814
Size: 616 Blocks: 8 IO Block: 4096 regular file

rustc --emit=obj -C panic=abort -C opt-level=z main.rs -o main.o
ld -o main_final -e _start --nmagic --strip-all main.o

>>105682565 (OP)
I tried rust and when it comes to pure Rust then you just have to write their idiomatic way to prevent all those compiler errors but the moment you introduce a C library good luck with that... i don't even want to express the pain i had to endure.
The "rewrite it in rust" meme in reality exists because interfering with libraries not written in rust is not fun nor productive that's why i say that rust in the linux kernel is a dead end, nothing will happen and nothing will be replaced.

>>105682602
I am genuinely interested in Jai from what I've seen, but the 2025 release date seems increasingly unlikely.

>>105703056
>the moment you introduce a C library good luck with that...
Just write a safe abstraction.

>>105702240
a lot of things are deal breakers, you just move the fuck on because you can't get everything you want with any language or even computers. you are just low IQ and can't figure out how to mitigate like a real programmer / engineer.

>>105703074
this. C ABI isn't that hard. C++ stuff, well... that's a different problem.

>>105703109
ABI was never a problem, Rust supports C ABI out of the box, just like C++ does(it literally has same syntax for that too).
API semantics is what might be hard to express in Rust. It's not an issue most of the time, but some large projects with weird patterns might be very hard to do right.

I'm so glad im a webshitter so i don't have to deal with any tranny drama.

>>105703541
i just write my shit and idc either
these threads exist as a challenge doe
crabs have grievances towards c
and c-ultists respond
its not a matter of war, its a matter of stimulation
its a latter of engineering
its a battle of wits even at a psychological level
it is truly the contest of two communities
the meta-relations of it all

rust vs c is not a statement
its a context

>>105703541
>>105703599
(cont)
it drives both communities to higher highs so i dont see why darwinism, which has been proven since, errr idk, life exists?- is a bad thing in ideaspace

>>105682602
>This is why I'm waiting for Jai
Meme language ...

>>105703792
(cont of cont)
rtust benchmarks are never compared with -march=native doe
but they enable simd every time...
funny and fucking cute innit?

fukken
>rust benchmarks against C
>but c when it has a hand tied behind its back
lamao

and even then
you can accelerate atoi merely by padding with 8 bytes
i played around with intrinsics and i stopped caring after 200x speedup
but the crabs benchmark against c with -O3
nigger
you dont have all the horsepwers activated
-O3 means "loazy mode"
which is very good
but its the first fucking step among 10 or 20

ask fukken registryfag
hes the one to ask in how to cut down an operation down to 10%
but even a macro approach destroys the fukken libc
libc is supposed toi be universal
so it can only be shit
generalized vs specialized solutions
the exact effing reason C is still held as golden standard everyone tries to improve upon, 60 FUCKING years after its been invented

(btw i like you registryfag. but i respond in kind.)
(DONT you fukken make smoke with me. esp were on the same side. kindof. bc you implied C is not enough recently and if you troon out i disavow you)

>>105702789
Not the OP.
Graph algorithms? Unless you maybe do integer indexing instead of trying to implement it with references.
It would probably perform well, just be really annoying to implement unless you do stuff like integer indexing.
Also desired architecture can play into decision making.
https://github.com/microsoft/typescript-go/discussions/411

>>105696495
https://gyroflow.xyz/ uses QT
https://zed.dev/ uses a bespoke UI framework
https://graphite.rs/ also uses a bespoke UI framework
https://typst.app/ technically too, on the web
https://en.wikipedia.org/wiki/Tiny_Glade would count too, using the Bevy engine
https://www.kraken.com/de/desktop this doesn't even advertise it

It's getting there

>>105702882
>>105702906
If anything, I would expect the Rust version of the same program to be 3% slower instead, just because of the array bounds checks that can't always be elided, whereas C++ always elides them, just assuming your code is correct.

>>105703076
>you are just low IQ and can't figure out how to mitigate like a real programmer / engineer.
Why not mitigate or deal with it by picking a better option? For instance, a different language if that language is a better option than Rust for a given case?
Also be more kind to yourself, you claiming to have low IQ is not nice.

>>105704173
NTA but just use an assert. Asserts remove a lot of bounds checking if you are repeatedly accessing random indices.

>>105704223
Whoa hold on that sounds unsafe bro

>>105704238
asserting is safe.

>>105704247
ugh.. thank god

>>105704142
>Graph algorithms? Unless you maybe do integer indexing instead of trying to implement it with references.
Graphs have myriad different ways to implement depending on what algorithm you want to do. If you want to implement some sort of dynamic graph based on references/pointers, you will mostly likely want some reference counting in every non-GC language.

>Also desired architecture can play into decision making.
True, but that's different topic. Their situation is specific, they are bunch of web developers who are rewriting JS code into another language and do not want to change it too much. Choosing Rust over Go would probably require a unproportionally more effort for not that many gains and with greater risk of regressions. Go being a managed language makes it often trivial to migrate the code.

>>105704173
In practice, you rarely even index arrays in Rust and rather rely on iterators instead.

>>105704238
Only unchecked_asserts are unsafe. C++ equivalent of this is [[assume(...)]] and C doesn't have a standard way of doing this lol.

>>105682565 (OP)
literally the only true thing is the last one
not reading any of the replies in the thread by the way, kys

>>105688276
> Talks trash about fizzbuzz.
> Can't write fizzbuzz as a generator, one character at a time, to land a job.
So many such cases.

>>105698824
> falls for the vulnerability as a service scam
> gets raped in prison for fraud

>>105704383
There are a bunch of fun answers to a question about graphs here
https://stackoverflow.com/questions/5493474/graph-implementation-c/5493656
How easy would each of them be to implement in Rust? If some of them would not be easy, could one be certain about performance in Rust, given limited options? I am guessing that one could do it as efficiently as in C++, but it may be way more work and more difficult than in C++.
How much reference counting are there in those answers?

Custom allocators and arenas might be easier in for instance Zig relative to Rust https://old.reddit.com/r/rust/comments/1jlopns/turns_out_using_custom_allocators_makes_using/ https://old.reddit.com/r/rust/comments/1km2vej/bump_allocators_in_rust/

https://old.reddit.com/r/rust/comments/1k83suv/do_people_who_use_rust_as_their_main_language/

Architecture can in practice be in some ways related to performance, though one may also be factoring in for instance how much effort it is to implement and maintain.
Discusses game engines and game development in Rust https://old.reddit.com/r/rust/comments/1k83suv/do_people_who_use_rust_as_their_main_language/
This comment complains about Rust, architecture and performance.
https://old.reddit.com/r/rust/comments/1k83suv/do_people_who_use_rust_as_their_main_language/mp3co6x/

>>105682602
>This is why I'm waiting for Jai

>>105706123
>How easy would each of them be to implement in Rust?
Since the first one doesn't even have any deleting, you can just change pointers to references and new to Box::new().leak() and the code will be literally the same.
Next few don't even have any code.
Next one is the same as first except indexes instead of pointers.
etc

>given limited options
What limited options? You can write nearly identical code.

>How much reference counting are there in those answers?
None of these do any cleanup. There is no need to count references.

>Custom allocators and arenas might be easier in for instance Zig relative to Rust https://old.reddit.com/r/rust/comments/1jlopns/turns_out_using_custom_allocators_makes_using/ https://old.reddit.com/r/rust/comments/1km2vej/bump_allocators_in_rust/
How so? You just implement an allocator and pass it to the collection.

>Architecture can in practice be in some ways related to performance, though one may also be factoring in for instance how much effort it is to implement and maintain.
Go is slower than Rust.

>Discusses game engines and game development in Rust https://old.reddit.com/r/rust/comments/1k83suv/do_people_who_use_rust_as_their_main_language/
Rust is unsuitable for gamedev but that has nothing to do with performance. Games do not need top performance either, most of them is written in managed languages. Rust is unsuitable for gamedev because it heavily prefers waterfall approach and doesn't let you rapidly prototype ideas. That's why so many games use scripting languages for game logic, it's just so much faster to write bunch of lua scripts and prototype various ideas than to wrestle with somewhat faster, but much more rigid language.

>This comment complains about Rust, architecture and performance.
>https://old.reddit.com/r/rust/comments/1k83suv/do_people_who_use_rust_as_their_main_language/mp3co6x/
>Games are big amorphous blobs of self-referential mutable data, and working that way performantly in Rust requires a paradigm shift relative to other languages like C++
Having a lot of self referencing data is just as much pain in the ass to write safety and without memory leaks in C++ as it is in Rust.
># already offers a good portion of memory safety tools relative to Rust, and a C++ engine like Unreal also provides memory safety tools through its own garbage collector.
Unreal C++ is not real C++. It's a C++-like managed language that is more like C#. Managed languages like these are more suitable for game development because they are easier to use and faster to iterate over ideas. Something that is crucial in gamedev.
This post says nothing about performance.

>>105682565 (OP)
>87 dependencies to do trivial tasks
No that would be C. Rust needs 87 C virtual machines running OSes to do trivial tasks.

>>105706772
>Since the first one doesn't even have any deleting, you can just change pointers to references and new to Box::new().leak() and the code will be literally the same.
I am beginning to be suspicious about you.
Would the first example, directly transferred to Rust by using references instead of pointers, even fly by the borrow checker? Aliasing? I'd love to see the equivalent Rust code as you describe it. Are the cycles in references really that easy in Rust without using unsafe and without using RC or ARC?


>How so? You just implement an allocator and pass it to the collection.
I see, you're just shitposting and not even reading the comments in that thread.

Please stop shitposting and being incompetent.

>>105706777
Again, please stop shitposting and being incompetent.

>>105682565 (OP)
skill issue.

first time compiles are slow. After that you're alright.

Type aliasing exists. Don't do that.

The borrow checker isn't hard, you're just being retarded. Time to unlearn years of shit programming. Rust has finally got niggerd thinking about how shit Linked Lists are and I'm glad it stops you from making one or any of their variants by default.

The compiler is the dad you never had. It beats your ass so you don't turn into a faggot ass programmer.

Rust community has been the opposite. Very nice, very helpful. Share your code in the c# or go community and they'll tell you to create the simpliest factory abstract proxy hexagon vertical slice before they even give you actual help.

>>105692173
Anon, their firstmost priority is to automaticall trsnspile c into rust and DARPA throws in money for an AI to do so.
This IS most of their existence. Looking at C code and rewriting it into Rust is 99% of their business.
Its ridiculous. Who the fuck wants to do that?

>>105692194
>I never met anyone who does this. Are you talking from experience or what?
well, the only Rust projects that kinda gained traction do that.
In fact, they are even worse.
Coreutils rewrite? Not just years of looking at C and rewriting, but they also have to run and look at the test cases of the original.
That AV1 decoder that ended up slower than the original? It was automatically transpiled and the "Rust developers" look at the auto-generated slop and clean it up and fix it while having the original as reference.

Those are the people who you promote. Nobody wants to do that. Nobody thinks this is fun.
You are a masochist.

>>105692173
>Two different programs that have the same function are not "rewrites."
anon, I...

>>105703599
llm response

>>105684566
>Well known utility or program made in C
>Someone is remaking it in Rust
>goes from a free license to a cuck license
Every time. This is my biggest issue with the Rust community.

>>105708101
>It beats your ass so you don't turn into a faggot ass programmer.
So the compiler doesn't work considering the community

>>105706772
> Go is slower than Rust
That holds if you're not using Rc<>/Arc<> anywhere or if Rust global allocator is good for X program you're writing. I've ran into spots where garbage collection is faster than what Rust offers as default. For example evaluating an ast which is describing a recursive function.

>>105709130
MIT is truest form of freedom. GPL is communism.

>>105707876
>Would the first example, directly transferred to Rust by using references instead of pointers, even fly by the borrow checker?
You just need to annotate adj with internal mutability.

>Aliasing?
There is no aliasing here.

>I'd love to see the equivalent Rust code as you describe it.
use std::collections::HashMap;
use std::cell::RefCell;

struct Vertex {
adj: RefCell>,
name: String,
}

impl Vertex {
fn new(name: String) -> Self {
Vertex {
adj: RefCell::new(vec![]),
name,
}
}
}

struct Graph {
work: HashMap,
}

impl Graph {
fn add_vertex(&mut self, name: String) {
if self.work.get(&name).is_none() {
let v = Vertex::new(name.clone());
let v = Box::leak(Box::new(v));
self.work.insert(name, v);
}
print!("\nVertex already exists!"); // ???
}

fn add_edge(&mut self, from: &str, to: &str, cost: f32) {
let f = self.work.get(from).unwrap();
let t = self.work.get(to).unwrap();
f.adj.borrow_mut().push((cost, t));
}
}


>Are the cycles in references really that easy in Rust without using unsafe and without using RC or ARC?
Rc and Arc is RAII. This code doesn't free anything so you do not need RAII for anything.

>I see, you're just shitposting and not even reading the comments in that thread.
If I was shitposting I would tell you to fuck off back to pleddit you retard. If you want to make an argument then state it yourself. If you rather spam reddit links hoping something sticks then get the fuck out of there.

>>105707882
Not an argument.

>>105708175
>well, the only Rust projects that kinda gained traction do that.
Can you name any popular project that did this other than rav1d?

>Coreutils
>looking at C and rewriting
[citation needed]

>Those are the people who you promote.
Where did I promote these people. I don't even know their names or who they are.

>Nobody wants to do that. Nobody thinks this is fun.
What? Promoting some FOSS developers? I guess?

>You are a masochist.
What?

>>105708180
Do you consider wayland to be rewrite of x?
Is Linux a rewrite of unix?
Is C# a rewrite of Java?

>>105709297
>That holds if you're not using Rc<>/Arc<> anywhere
Adding RAII to your program will not make it magically slower than GC.

>or if Rust global allocator is good for X program you're writing
Global allocator is pretty good. You are free to build a bump allocator on top of it or use any other allocator.

>I've ran into spots where garbage collection is faster than what Rust offers as default. For example evaluating an ast which is describing a recursive function.
See >>105702789

>>105709707
>RefCell
>A mutable memory location with dynamically checked borrow rules
>dynamically checked
How much overhead, Mr. Incompetent Shitposter? More than 0?
Try again.
And please describe yourself in more kind words than "retard" and telling yourself to fuck off.

>>105710115
>How much overhead
This requires a benchmark to measure. See >>105702789

>Try again.
I don't see any reason to try anything again. I responded to all your questions already.

>And please describe yourself in more kind words than "retard" and telling yourself to fuck off.
I didn't describe myself as retard. I only called you a retard for accusing me of shitposting whenever you couldn't respond to my points and for reposting reddit comments on 4chan.

>>105710166
Thank you for proving and admitting that you are incompetent and a shitposter. Please stay far away from any programming job that requires any kind of responsibility.
And do refer to yourself more kindly than as a "retard".

>>105710214
Not an argument.

>>105682591
Whoever made this gif HATES trannies with healthy passion

>>105710219
You already proved it yourself multiple times over, Mr. Dishonest and Incompetent Shitposter.

>>105710244
Yes, I have already stated my arguments.

>>105682565 (OP)
Considering how much traction such kinf of discourse gets someone should start a /programming languages general/ for discussion on the topic's various facets

>>105710363
We had Rust general for a week or two. It was constantly spammed by angry cniles and /pol/ tourists.
You can't have quality discussion on /g/. This board seems to be just for AI slop and twitter screen cap threads(like this one)

>>105710301
That proved what you are, Mr. Dishonest and Incompetent Shitposter.

>>105710439
Who I am is irrelevant to this conversation. Performance is what is relevant. Something that can be measured by a benchmark which you are unwilling to undergo.


This is my last response to you until you write and post some code that can be benchmarked.

>>105682602
It staggers me there are a few people who truly believe a guy who cried when soulja boy didn't """understand""" his hipster platform slop and who argues on twitter all day could truly create something of worth.

>>105710693
But he's so cool and smart and manly and he made two whole games and he speaks authoratively about everything and belittles and rages at anyone who disagrees and I was raised by a single mother with no father figure

>>105710621
Lel, Mr. Dishonest and Incompetent Shitposter.

>>105706772
>the code will be literally the same.
>>105709707
>[uses] RefCell [which has runtime overhead]

Try again, and stay away from any programming job that requires responsibility and competence.