← Home ← Back to /g/

Thread 106378981

149 posts 52 images /g/
Anonymous No.106378981 >>106379447 >>106379469 >>106379832 >>106379916 >>106380249 >>106380594 >>106380736 >>106383072 >>106389474 >>106390979 >>106391251
Rust for Linux
Was it doomed from the start or is it the future and just experiencing growing pains with the old C-programmers fearing for their position?

Was it a mistake for Linus to allow both Rust and C at the same time and not allow only one of them for new parts?
Anonymous No.106379097 >>106379114
It's up to Linus to decide.

>and not allow only one of them for new parts?
This would never work and it makes no sense whatsoever.
Having Rust drivers in Linux makes total sense and properly formalizing its APIs via Rust will be beneficial to whole Linux. It had already exposed some mistakes and underspecifications in the existing APIs.
Anonymous No.106379114 >>106379164
>>106379097
you are essentially introducing mixed code base where C isnt just used for maintaining older parts of the kernel
it must cause pain
Anonymous No.106379164 >>106379191 >>106379229
>>106379114
You can't use Rust for core parts of Linux because it doesn't have wide enough platform support. You can only use it for platform specific stuff like drivers(which is the majority of Linux anyway).
Mixed codebase is not an issue in project like this technically speaking, but many contributors do not know Rust so it would make no sense to enforce some Rust only policy for new stuff.
Also, you would need to write Rust interfaces for every kernel API out there, that's a lot of work that won't happen overnight.

Adding Rust support is generally non intrusive and can benefit the kernel by improving safety and attracting younger contributors(something Linus pointed out himself), but anything more than that would be unnecessarily detrimental to the pace at which Linux is being developed.
Anonymous No.106379191 >>106379532
>>106379164
It maybe attracting younger people but is it attracting the kind of people you want? The whole Hector Martin kerfuffle says otherwise (also the Rust community swatting the guy for the crime of saying he didn't like Rust on stream, not sure if that was kernel dev related though.).
Anonymous No.106379229 >>106379532
>>106379164
> you would need to write Rust interfaces for every kernel API out there,
You don’t. The R4L maintainers do it for you.
Anonymous No.106379447
>>106378981 (OP)
It was not a mistake, it does makes technical sense, newer devs don't give a crap what boomer C devs think, and that's about it.
Anonymous No.106379469
>>106378981 (OP)
Yes it was doomed from the start. Linux code is pajeet tier. It only works at all due to sheer momentum and billions of dollars in salaried wagies to hammer shit til it works. Of course after decades of this trying to shove a language with a real type system like C++ or Rust causes a lot of ire, pain, seethe, butthurt and more.
Anonymous No.106379532 >>106379555 >>106387485
>>106379191
>The whole Hector Martin kerfuffle says otherwise
And what did he do? Brought support for apple GPU.
The drama that followed was purely management failure and started from c programmer trying to sabotage R4L, not from Hector himself. His reaction was retarded too but all of this could have been resolved and neither of them would leave Linux if Linus got his shit together and stepped in.

>>106379229
R4L is part of Linux, retard.
Anonymous No.106379555 >>106379581 >>106379747 >>106387103
>>106379532
He tried to cancel and get a bunch of kernel devs kicked off because he was ideologically opposed to them / didn't like him, which is a common attitude with you Rust faggots. Your kind are a cancer that needs to be removed from all serious projects.
Anonymous No.106379581 >>106379610
>>106379555
Are you talking about Helwig? Helwig lost because of his own brain dead arrogance.
Anonymous No.106379610 >>106379768
>>106379581
You full well know what I am talking about you disingenuous fuck. This is what you Rust fags do, you ignore the majority of a situation then focus on something almost completely irrelevant to create a little strawman based on some sort of half truth in order to twist the situation to look good in your and your closeted community's favor. You are a prime example of why Rust and it's community should be gate-kept from Linux kernel development.
Anonymous No.106379747 >>106379806
>>106379555
>He tried to cancel and get a bunch of kernel devs kicked off because he was ideologically opposed to them / didn't like him,
Lol no. Have you even bothered to read the mailing list or do you get all your news from /g/?
Helwig was intentionally blocking changed made by other people(not Hector), calling it cancer and abusing his privileges. This caused some people to lose interest in R4L, which frustrated Hector and caused him to write social media post where he said he is considering making a "hall of fame" for people like Helwig. To which, Linus finally decided to step in and told him this is not acceptable way of handling such issues.

There was no talk about canceling or trying to kick anyone off. There was no problem with any ideology and Helwig never said he personally doesn't like Hector or any other contributor.

Wtf are you talking about?
Anonymous No.106379768
>>106379610
>disingenuous
The only disingenuous person here is you.
Post proof of Helwig saying he doesn't like Hector or that he disagrees with him on some ideological ground or stfu.
The way you are presenting this drama is completely wrong.
Anonymous No.106379806
>>106379747
And of course, Linus later confronted Helwig for his behavior and told him he is crossing the line.
This caused Helwig to get upset and he left Linux as well.

Total management failure.
Anonymous No.106379832 >>106379854 >>106380195 >>106389648
>>106378981 (OP)
itll be fine.
if you know c then rust isn't that hard to pick up and i honestly do not understand the weird aversion to it considering how so many other companies and projects are shifting to rust
feels a bit batshit listening to people whine and bitch about rust and when you ask them what they hate about it they just kinda incoherently jibber at you like the dude who tried to cause a scene at the conference and mixed up java with rust
Anonymous No.106379854 >>106379926
>>106379832
>if you know c then rust isn't that hard to pick up
what I've seen from the maintainers is that its really fucking hard at least in their opinion
Anonymous No.106379916 >>106380202
>>106378981 (OP)
he was obviously forced to put this infection into his kernel.
that's why he doesn't want it in the arch itself but only as external driver support
rust trannies are deluded enough to believe that complex infrastructures with decades of history can be replaced with their rust vibe coding shit.
benchmarking or "security" aside this is not practically possible
Anonymous No.106379926
>>106379854
Old people cry the second they have to learn a slightly different paradigm, which is why all the boomers at my work write horribly overcomplicated OOP code with multiple inheritance.
Anonymous No.106380195 >>106380233
>>106379832
>i honestly do not understand the weird aversion to it considering how so many other companies and projects are shifting to rust
Why do you think C is so popular here despite having little relevancy on its own outside of some specific environments and academia?
It's all just freshman students who just got taught C and want to feel superior over other freshman students that only were taught Java. They think they are special because they now know a baby first systems programming language. That's why you see so many people claiming C is a low level language or that it's "the closest we got to assembly without writing assembly", etc. Ideas made up entirely of misconceptions, with gaps in their knowledge and experience filled only with memes.
Anonymous No.106380202 >>106380535
>>106379916
>he was obviously forced to put this infection into his kernel.
>that's why he doesn't want it in the arch
[citation needed]
Anonymous No.106380233 >>106380481
>>106380195
people hate rust because its dogshit
and love c bc its easy.
simple as fucking ass

youre not doing yourself or your community any favors by being an annoying retard
Anonymous No.106380247
The Rust ecosystem is full of mentally ill troons and that's reason enough to keep it out. Cope, seethe and dilate.
Anonymous No.106380249
>>106378981 (OP)
Yes, please. Add more tooling requirements and points of failure to such a crucial project. I can only wonder why perl wasn't added prior.
Anonymous No.106380481 >>106380725
>>106380233
Pointless post.
Anonymous No.106380535
>>106380202
his actions indicate his motives.
each time a random tranny try to inject his rust shit as an internal component he cut him off
Anonymous No.106380594 >>106381106
>>106378981 (OP)
>Was it doomed from the start or is it the future and just experiencing growing pains with the old C-programmers fearing for their position?
a dual-language codebase is an abomination, the only path forward is toward a monolanguage codebase, it's impossible to know who will win because it's never the best technical solution that win, socials are more important.
linux is unforkable, we'll have to live with the management failures of neo-linus for the next 100 years at the very least
the only way to replace linux is with a microkernel with support for linux drivers
the core of the kernel is doable by a small team, replacing thousands of drivers for millions of devices isn't and it's the strength of linux
>Was it a mistake for Linus to allow both Rust and C at the same time and not allow only one of them for new parts?
it's a mistake to mix two languages, no matter what they are.
Anonymous No.106380725
>>106380481
no, i wanted to say something so i did
and it got you seething so its all the better :)
Anonymous No.106380736 >>106381237 >>106382216
>>106378981 (OP)
>Was it doomed from the start
yeah i guess
its a dogshit lang that isnt even finished yet
it has no business in the kernel
not to mention grafting a new language onto an established codebase is always a bad idea
Anonymous No.106381106 >>106381237
>>106380594
>a dual-language codebase is an abomination
Linux codebase already includes at least 4 language not counting Rust.
Anonymous No.106381237 >>106381253
>>106381106
>linux being multi language
lmao
nta btw, im this guy>>106380736
Anonymous No.106381253 >>106381298
>>106381237
Looks like 6 languages. Maybe more.
Anonymous No.106381298 >>106382321
>>106381253
yeah
notice how the only two that serve a similar niche are c and rust
its a dogshit idea to have two languages within the kernel
just as much as it is a dogshit idea to use two competing workflows in any department
its organization 101
and from the technological standpoint its a dogshit idea to retrain your workforce
its much more efficient to rotate them away
which didnt happen

so not only its a dogshit idea to begin with
but keeping rust in the kernel is a dogshit idea too, all things considered
Anonymous No.106381391 >>106381507
>Was it a mistake
If it was just for drivers no, but trying to push it anywhere else doesn't really make sense. The problem is that Linux isn't a corporation that can just push these huge changes. If this was a corporate environment they would just end C support and force their developers to learn rust and move forward with that. This being open source you have a large camp of people who have been using C for a long time and have no desire to switch to rust. These veterans control a ton of subsystems and if Linus was to force a transition he might risk a fork and lose a ton of maintainers, so he is stuck with the half hearted you can use both. The kernel might eventually switch fully to rust, but that will take something like 20 years and will depend on if rust came maintain its popularity. Personally I would be surprised to see a rust victory, because C has so much inertia it wouldn't be surprising to see rust just become a niche language and fade away like perl.
Anonymous No.106381507 >>106381517 >>106381590 >>106389767
>>106381391
>mfw C's inertia
you meant the inability to dislodge c on part of rust
rust is a complex thing
and no, its not good that it is

for people who do c since a while its red tape
for newcomers c is easy and straightforward
the industry agrees.
cuda is c. ill say it straight, you write it in a sepples framework but you wont be working with anything more complex than an array
not rust, not lisp, c.
it makes sense on several levels
also that beyond a certain point, you really dont need safe and effective (tm)
then you want something that helps you accomplish your job as opposed to standing in your way
Anonymous No.106381517
>>106381507
aw forgot picrel
it was perfect for the occasion
Anonymous No.106381590 >>106381766
>>106381507
>Rust is a complex thing
You seem to misunderstand me. I like C/C++. I don't really like rust, but I am telling you what is probably going to happen. Rust makes promises C isn't willing to make. Managers and product leaders like these promises so they'll push rust. You'd think this wouldn't effect open source, but considering that all the major corps are supporting Linux they'll push rust because it is becoming part of their culture. The only reason I am not a doomer on Linux is because of the strong and healthy culture behind C.
Anonymous No.106381766 >>106389767
>>106381590
>Managers and product leaders
yeah, thats marketing
im not trying to be insulting here
its just that depending on the field they have more less or even zero say on the technical aspect of things
thats why i challenge the idea of inertia in the context of c
its not a matter of inertia
its that there isnt a better offer for many applications
personally im very good at interfacing so i could deal with an arbitrarily sized project in pure C
but even considering a heterogenous workforce, c has its raison d'etre even in tyotl 2025 x86-64 applications
namely: libs for lack of a better umbrella term
in my opinion the major usecase for C is whenever you need 95% of performance for 5% of the effort
this implies unsafe in its rust equivalent for most cases
and means the code can easily be contained and kept correct

kek, im focused on everything BUT linux
i dont have an opinion on the matter beyond a vague expectation that the project will be forked into several exemplaries once linus steps down
i expect at least one of them being programmed in C
because of the already existing maintainers
because of the cult status of c
but also because c stands on its own merit so new maintainers will follow
bruce3434 No.106382216 >>106382519 >>106382582
>>106380736
Why is it dogshit? And what’s wrong with being unfinished? C isn’t finished yet either.
Anonymous No.106382321 >>106382519
>>106381298
>assembly
>not a language
Anonymous No.106382519 >>106382582 >>106383409
>>106382216
>Why is it dogshit?
red tape
bad visual design
cargo is unauditable + a supply chain attack factory

>>106382321
>assembly serves the same role as c
Anonymous No.106382582 >>106383409
>>106382216
>>106382519
im talking from my perspective mind you
i have something like 300k locs of C under my belt
but the cargo mess and bad visual design are universal
picrel im 100% sure is a supply chain attack caught twice
Anonymous No.106383072
>>106378981 (OP)
If it had been introduced with the intention of eventually replacing all the C code, maybe it could work out long term. I don't think that's the plan though and I don't think it's likely that will end up being the case. So it's left in this weird limbo in between state with a lot of growing pains that aren't easy to solve. It would have been better to go to C++ which would have most of the same benefits but way less friction. More would have been accomplished in a shorter time frame as well.
bruce3434 No.106383409 >>106383601
>>106382519
>>106382582
>supply chain attack
Not a language level problem. It can happen to any languages, no?
I mean even this year we had a few security incidents happening in popular C libraries too.
Anonymous No.106383601 >>106384229
>>106383409
well, it is in the case of rust because it ships with a package manager
of course c code has it fair share of issues
but the mechanisms in place dont make it easier to SCA unlike with rust, cargo and its recursive dependencies
cargo downloads source code, not binaries if im not mistaken

its been more than a year since the spat about bcachefs-tools, it was pretty high profile in the linux community
a simple fix would have been to provide a tool that would extract live code form the source
such a tool hasnt been created yet
naturally, questions arise-
how come the rust community didnt fill in the voids? how come they all responded with dialectical tactics instead?
why the rust devs hadnt thought of that?
or heard about that idea, i was fairly open about this on g at a certain point, one would think the message would worm its way, and that assuming nobody had that obvious idea besides me
and considering that
occams razor tells us that theres ulterior motives

cant blame the c-maintainers for not having to build one for the crabs either
crabs should have been nicer people. calling people retards and cniles isnt very conducive to their cooperation lets say
Anonymous No.106384229 >>106384277
>>106383601
r4l doesn't use external dependencies, so you will have to use another (invalid) talking point.
and a repository of source code, central or ad-hoc (crates.io or source-based distros) is by its very nature infinitely more safe than a central repository of binaries, which is what all mainstream distributions provide, as the latter has every attack vector of the former IN ADDITION TO potential attacks from compiling to distribution.
>but muh distro vets...
yeah, as we saw with the xz debacle lmao.

to make things more laughable, there actually exists a new(ish) retarded (genuinely this time) trend going on that is tremendously dangerous while being sold as a safety improvement. it's called flatpaks. but the nocoder internet gantry here doesn't talk about it because they probably find it useful, since they probably can't compile projects themselves for shit, and unlike the rust ecosystem ;), a lot of projects are not reliably trivial to build. lol.
Anonymous No.106384277 >>106384314 >>106384415
>>106384229
>r4l doesn't use external dependencies
yeah but im talking in the context of the whole rust framework
sorry but your whole post is invalid, your premice is wrong
Anonymous No.106384314 >>106384368
>>106384277
>i will skip the rest of the post which dismantles the talking points e-celebs and /g/eets gave me
i accept your concession
Anonymous No.106384368 >>106384413
>>106384314
no, i skipped the rest of your post because i dont like your tone, crab
centralized repo?
what good of it if anybody is allowed to post?

its like sharing needles you found under a park bench, in an empty beercan
Anonymous No.106384413 >>106384442
>>106384368
>continuous display of complete lack if knowledge about how the open-source world always worked
>random analogies
i accept your second concession
Anonymous No.106384415 >>106384442
>>106384277
>ur hole post and pre-mice is wrong hurr but I give no explanation
Anonymous No.106384442 >>106385145
>>106384413
>>106384415
>tranny pileup
fine technical arguments
but between the sad homonems you lack consistence

all you mustered until now is "nuh uh"
Anonymous No.106384563 >>106384602 >>106385903
Hypothetically, if the linux kernel was in rust wouldn't that make it easier for maintainers to validate code because it's easier to debug?
Anonymous No.106384602
>>106384563
somewhat. not even 100%
and all that for the low effort cost of rewriting the whole codebase and retooling your workforce
not to mention rust itself looks like something soemone stomped with studded boots repeatedly and passionately
Anonymous No.106385145 >>106385514
>>106384442
>oh yeah! you're a tranny. and you're a tranny.
i continue to accept your concessions
Anonymous No.106385514
>>106385145
>t. ranny
Anonymous No.106385903 >>106385941 >>106385965
>>106384563
Rust makes it easier to validate code because it is more formal, it allows you to express more precisely the semantics of your API on the type level as you can see in OP's image.
Rust is not really any easier to debug, it might be harder depending on your tool. But it is much less likely to require debugging in runtime.
Anonymous No.106385941 >>106385955 >>106385965
>>106385903
if you actually coded, you would have known why rust is easier to debug in practice, instead of that generic waffling you just did.
Anonymous No.106385948
Ideally c++ would be in the kernel, but Linus hates c++ for autistic reasons so we have to settle for its younger, developmentally disabled brother, rust.
Anonymous No.106385955 >>106386617
>>106385941
I have been programming in Rust for 7 years.
Anonymous No.106385965 >>106386617
>>106385903
>>106385941
>because algorythmic mistakes arent a thing
lol. lmao even
Anonymous No.106386617 >>106387975 >>106388566
>>106385955
if you have been, you would have been exposed to at least Display/Debug/Error::source()/#[tracing::instrument()]/etc. and this is just the basic stuff. let's not get to "fancier" stuff, that may not be relevant to kernel's development, like tokio's console. or fancy stuff that is very relevant to "unsafe" development like miri.
so my assumption is that you're either a nocoder, or yet another freshman/sophomore tard, or someone who codes at the webshit level and thinks all debugging is done inside gdb/lldb. and even then, rust is nicer there since you won't be ever crashing/pausing at a site where the context is full of fully type-erased void* values, or some other callback fuckery with pointers, unions or whatever.
>>106385965
>let me imagine a strawman
>wow very funny. lol. lmao even
Anonymous No.106386865
This bone headed shit is going to backfire on them so hard lol.
Anonymous No.106387103
>>106379555
>help me I am getting cancelled!
If rather than arguing over what happened or providing context this is the best defense you can muster then it's because you know you fucked up and the only way to get past it is for others to not know about it. So get fucked. Retard thought he's the big man and gets to block everything because he decided no dual language. Not his project, not his call, simple as
Anonymous No.106387485 >>106387980
>>106379532
Nah, Hector Martin situation happened due to unrelated drama with his vtubing career. Some other vtuber was harassing him or something
Anonymous No.106387975 >>106388759 >>106389553
>>106386617
>Display/Debug/Error::source()/#[tracing::instrument()]/etc.
When I say debugging in runtime, I mean using a debugger, not implementing a trait for displaying types wtf
Anonymous No.106387980 >>106388049
>>106387485
Post more info
Anonymous No.106388049 >>106388099
>>106387980
google asahi lina and luna the foxgirl drama
Anonymous No.106388099
>>106388049
How about you provide context to your claims.
I have never seen any of this mentioned in Linux mailing lists.
Anonymous No.106388350 >>106390555
This thread and board is filled to the brim with samefagging shills.
Anonymous No.106388566
>>106386617
>let me imagine a strawman
>implying algorythmic errors are not part of debugging

crabs are mental retards
Anonymous No.106388618 >>106388696
>error are part of debugging
Anonymous No.106388696 >>106389470
>>106388618
i look like that and i say that

and even when im sleeping...
...crabs are still retarded
Anonymous No.106388759 >>106388802 >>106388810
>>106387975
>debugging is only when you're in gdb/lldb
if you continued to read, you would have seen that i expected that's what you had in mind.
>not implementing a trait for displaying types
that confirms my presumption that you don't know much.
Anonymous No.106388802 >>106389470
>>106388759
>displaying/checking types
c's compiler takes care of that

are you telling me that you have to do that manually in rust???
Anonymous No.106388810 >>106388865
>>106388759
Whenever you have expected me to call you out or not doesn't change the fact that when I say debugging in runtime, I mean using a debugger. Not implementing a trait for displaying types.
Anonymous No.106388865 >>106389326 >>106389767 >>106395520
>>106388810
>program shows garbage values
>i know, ill print out the types, something that gets checked at compile time, surely thats gonna help
the absolute state of crabs
Anonymous No.106389326 >>106389341
>>106388865
Cool story.
It has nothing to do with what I said so far.
Anonymous No.106389341 >>106389596
>>106389326
>algorythmic bugs arent bugs
idk who taught you english, but they arent your friend
Anonymous No.106389410 >>106389447 >>106389466
why are /g/'s rules unenforced? even /v/ bans rule-breaking posts, most of the time.
Anonymous No.106389447 >>106389466 >>106389470
>>106389410
wym?
whomst did you report and what rule you think theyre breaking?
Anonymous No.106389466
>>106389410
>>106389447
btw, complaining about 4chan staff or its policies is a potentially bannable offence so idk if you wanna be a stifler about rules
Anonymous No.106389470 >>106389512 >>106389541 >>106389603
>>106389447
really tard? I haven't reported anyone. no point in it. /g/ is probably the worst board at this point.
>>106388696
>>106388802
for starters, this tard is breaking global rule 13
https://www.4chan.org/rules#global13
bunch of low quality banter not far above my post post (namecalling, no technical discussion).
let alone literal flamewar shit ACTIVELY OCCURING in this thread.
Anonymous No.106389474 >>106389745
>>106378981 (OP)
Rust glows extremely bright, expect massive glownigger backdoors to be revealed in the near future. The fact that it's championed mainly by the usual suspects should be a clear enough sign.

Didn't one of the main developers of Rust bail now that USAID money dried up?
Anonymous No.106389512 >>106389541
>>106389470
>low quality banter
you started
internal, unwritten rules of this forum are that if you start shitposting, then youre on your own

i provided technical arguments time and again
you shitposted
so now i can do whatever i want with you, within limits of whats acceptable
if i smother the catty with our bullshit ill get b&
if i post animal abuse, or gore or other distasteful shit ill get b&
but i can call you a fat faggot for having zero riposte and calling on the janies because you lost the argument
and i can sprinkle some insults over the injury by posting tech related humrous memes making fun of the rust language AND its community

i know it sucks for you but on this board your ideas and arguments have to stand on their own
Anonymous No.106389541 >>106389603
>>106389470
>>106389512
also
>flamewar
rust vs c threads get a pass
its part of g's culture now
and discussions can get heated, were all pasionate about what we do
janies know this and remain impartial
turbo-based, elite janies
Anonymous No.106389553 >>106389571 >>106389635
>>106387975
how do you expect your debugger to have rich information about your types to introspect without implementing this information somewhere? what better place than in the language instead of writing C++ specific hacks into gdb for libstdc++ types. have you actually debugged anything like... ever? debuggers can do more than step through code as well, btw.
Anonymous No.106389571 >>106389642
>>106389553
>to introspect without implementing this information somewhere?
wrong use of the term impolementing
and you do that whenever you declare a variable in a statically typed language
Anonymous No.106389596 >>106389664 >>106391229
>>106389341
>algorythmic bugs arent bugs
I never claimed otherwise.
Anonymous No.106389603
>>106389470
>>106389541
going back to the flamewar:
there is some entertainment AND educational value to our shit
our flamewars consist in flexing our respective knowledge and abilities
and we get flushed down the shitter whenever we go overboard

i think our shitflinging creates value and thats why were tolerated
Anonymous No.106389623
>Rust for Linux developers created a new library "pinned-init" to safely and fallibly initialize memory that must not be relocated.[3] It was first included in Linux 6.4,[16] and been improved in later versions.
yikes!
Anonymous No.106389635
>>106389553
I haven't written my own debugger, I don't know how they work.
I just press Debug button in RustRover and it let's me step through my code, set up breakpoints, view variables, look up assembly, memory etc. That's all I ever needed. None of this involved implementing traits he mentioned.
When I was talking about debugging in runtime, I was obviously talking about using a debugger. Of course you can debug your programs by inserting debug prints everywhere but this has nothing to do with the argument at hand, which was about whenever Rust makes it easier to debug. Debug prints can be done in any language.
Anonymous No.106389642
>>106389571
ok, you really never have debugged anything, like ever, I take it. maybe you enjoy writing a bunch of special gdb scripting to pretty print your types, but I don't want to.
Anonymous No.106389648 >>106389657
>>106379832
rust is a cargo cult and rust itself with its cargo implementation is the biggest supply chain attack vector ever and you shouldnt have rust anywhere in your project
Anonymous No.106389657
>>106389648
>the biggest supply chain attack vector
Yet nothing ever happened comparable to xz.
Anonymous No.106389664 >>106389671 >>106391167
>>106389596
yeah then why you try to recenter the discussion on static analysis?
i mean
thats rhetorical
rusts assurances falls flat when you factor in real life use parameters

i dont need static analysis
i have rigorous methods
the nr1 error in my code, by occurence, is infinite loops
because i eliminated everything else
the nr2 error is mistakes in interfacing with syscalls but thats because the documentation is shit, but thats another can of worms
this eliminates 99% of what rust has to offer
and the remaining 1% is not worth it when i have to fight the fucking red fucking tape

picrel is a bucketsort
actually an american star flag something sort
a bucketsort where you sort one byte at a time

post the equivalent in rust
afterwards we will get into the intricacies, because theres nuance to the way memory is structured
it took me 2 lines to set that up
we will compare what it takes in rust to do the same

#include "nouvo.digitize.h"

#define DATA_ORDER_PATHS_ITERATIONS 4

__attribute__((always_inline)) inline void data_order_paths_reset(t_list *list)
{
list->first = 0;
list->last = 0;
list->count = 0;
}

size_t data_order_paths(t_list *src_list)
{
t_list table[256] = {0};
int i = 0;
while (i < DATA_ORDER_PATHS_ITERATIONS)
{
t_cell_path* current = (t_cell_path *)(src_list->first);
t_cell_path* next;
int c = 0;
while (current)
{
next = current->next;
c = (current->timestamp.c)[i];
data_get_path_append(&(table[c]), current);
current = next;
}
data_order_paths_reset(src_list);
c = 0;
while (c <= 255)
{
current = (t_cell_path *)(table[c].first);
while (current)
{
next = current->next;
data_get_path_append(src_list, current);
current = next;
}
data_order_paths_reset(&(table[c]));
c++;
}
i++;
}
}
Anonymous No.106389671 >>106389702
>>106389664
>yeah then why you try to recenter the discussion on static analysis?
I haven't said anything about static analysis.

The rest of your post is built on top of this false premise.
Anonymous No.106389702 >>106389715
>>106389671
you made fun of debuggers and retorted with printing the fucking types
thats static analysis and its redundant bc its dealt with by the compiler.

now post bucketsort
im fairly certain its gonna show why people who are good with c can only LOATHE rust
Anonymous No.106389715 >>106389725
>>106389702
>you made fun of debuggers
I haven't.

I only stated that when I say debugging in runtime, I mean using a debugger. Not implementing a trait for displaying types.
Anonymous No.106389725 >>106389730
>>106389715
ah youre the OTHER guy
then why are you antagonistic?
Anonymous No.106389730 >>106389767
>>106389725
>then why are you antagonistic?
I'm not. I just clarified what I said.
Anonymous No.106389745 >>106389788
>>106389474
They spent all their money on DEI and didn't pay their core compiler devs.
Anonymous No.106389767 >>106389825
>>106389730
ok, now i get it
i was agreeing with you here >>106388865
didnt know youre a rust user

but, uh
now you see why the rest of us, in turn, are antagonistic towards rust users, as a rule (ive tried to remain civil. these are mine too >>106381766
>>106381507

some fukken dweeb who knows shit from fuck starts screaming at you
and its a rust problem
rust and lisp, but i think with lisp its autism thats the culprit (so its part and parcel of dealing with the programming world)
Anonymous No.106389788
>>106389745
Rust core team gets paid. Most of they money goes into administration, infrastructure and community grants.

You know that their fundings are public, right? Why lie?

https://rustfoundation.org/resource/annual-report-2024/
Anonymous No.106389825 >>106389838 >>106389870
>>106389767
>now you see why the rest of us, in turn, are antagonistic
There is no us. It's literally just you trying to bait people into arguing with you over meaningless stuff.

I don't care about you and whatever you think think or who you are.
No one actually cares.
I have already made clear what I meant by my original post. Your attempts to start a flamewar are futile. We have nothing to discuss.

This is my last response to you.
Anonymous No.106389838
>>106389825
>It's literally just you trying to bait people into arguing with you over meaningless stuff.
ah, so you too, are of ill will, after all

i guess the rule isnt a rule without a reason, eh?

youre on 4 keks my man
the whole purpose of your activity here is wasting time
Anonymous No.106389870
>>106389825
>the whole purpose of your activity here is wasting time
unless it isnt
then the only thing thats left on the table is influencing perceptions

funny how this discussion turned out
Anonymous No.106389885
btw
you might be 140 iq
but my 125 live in several cores
there isnt a configuration of the universe you can compete

in other words: i have voices in my head but they all say sound engineering shit
Anonymous No.106389979 >>106390006 >>106390038
>yet another REEEEEEEE RUST thread
Jesus fucking christ, get a grip on your life. Literally rent free.
Anonymous No.106390006
>>106389979
rust btfo
2 years ago we had multiple rust threads, concurrently
today?
>no, plis dont make rust threads it hurts...
NNNNNNNNNNNNNNNNNNNNNNNNNN
Anonymous No.106390038
>>106389979
but more seriously
its ok
rust is at 0.999, not 1.0
reforge your shit

wtf is wrong with that?
u n fucking ironically
i was the spark to btfo rust, apparently
and even im saying: you can do better

you need to understand that rust hate is not hate
its just a symmetrical response of a self regulating system
Anonymous No.106390127
the bodybuilder meme was completely unintentional btw
i just fukken felt that the gymbro attitude and whats around the culture fits ...c
if youre a gymbro, ytou like results
when youre a c bro, you like results
some kungfu master can do his chi or what the fuck ever, but you can come in with thicc dummy strength and just deal with teh problem in bruteforce
like in c
like a gymbro would
the similarities dont end there

and thats the gist of the meme
that and the reaction to it
when at its core its something simple and pure
it makes the hatersn otherwise, corrupted make oust themselves
its things within things within things like with most shit i do
but at its core
its simple

lift shit
youll get better at lifting shit
and dont cerr for fagot sois
bc they need them machines
when your value is inherent to your person
/rant i guess
fkn do better, crabs
Anonymous No.106390306
also therad theeem
i think i deserve the right to post a thread theme
https://www.youtube.com/watch?v=vrW3NlemNQw
Anonymous No.106390555
>>106388350
When is the read lock dropped and released in this Rust snippet? And is there a deadlock in this snippet? You may not assume anything about which edition it is compiled and run in.

fn f(value: &RwLock>) {
if let Some(x) = *value.read().unwrap() {
println!("value is {x}");
} else {
let mut v = value.write().unwrap();
if v.is_none() {
*v = Some(true);
}
}
}
Anonymous No.106390979 >>106391107 >>106391141 >>106391209
>>106378981 (OP)
I don't like Rust and I hope it won't be a requirement for anything worthwhile, because it would make bootstrapping Linux (like LFS) very tedious. I use CRUX and LLVM and Rust and whatever depends on them is fucking cancer to maintain. Literally breaking my ports randomly with minor version bump of a package. A package uses newer feature, because of "bleeding edge nonsense". Now you need to recompile everything upwards. Fuck this shit. Imagine if it was a hard dependency for something more crucial and would prompt upgrading more frequently. C is good language because it is stable. No package using C will force me to upgrade whole fucking toolchain for a single package in a single minor version change.
Anonymous No.106391107 >>106391141
>>106390979
>A package uses newer feature, because of "bleeding edge nonsense".
Wait.

Surely this is not a thing in Rust. Publishing unstable packages, for people to use in production.

Right?
Anonymous No.106391141 >>106391205 >>106391245 >>106391521
>>106390979
>>106391107
Are you too retarded to use version pinning? What am I reading here?
Anonymous No.106391167 >>106391229 >>106391285
>>106389664
cont

still no bucketsort in rust btw
im really fucking cu rious how one would look like

same rules as for mine, you can abstract away the absolute fucking mess linked lists are in rust so actually you have a boonus
Anonymous No.106391205 >>106391305
>>106391141
How do you feel about https://github.com/rust-lang/rust/issues/127343#issuecomment-2295313993 ?
Anonymous No.106391209
>>106390979
>Literally breaking my ports randomly with minor version bump of a package. A package uses newer feature, because of "bleeding edge nonsense".
If you don't need bleeding edge then why do you keep updating the package?
Anonymous No.106391229
>>106389596
>>106391167
and then afterwards
lets compare what it takes to tell the computer that we want complete paths vs a representation thereof in separate memory pools
im curious as fuck
bc im very sure of myself i wanna confront it
Anonymous No.106391245
>>106391141
Unstable as in solved by pinning release versions of the package, or unstable as in using unstable features only present in at-the-time trunk versions of Rust and rustc?
How would version pinning work if a Rust package uses unstable features, rustc is upgraded, and then things break because rustc changed or removed the unstable features used in that crate?
Anonymous No.106391251
>>106378981 (OP)
It's just eh, decent idea, subpar execution.
Anonymous No.106391285 >>106391294 >>106391450 >>106391510 >>106391618 >>106392393 >>106395754
>>106391167
NTA but here's a bucketsort implementation in Rust.
Anonymous No.106391294 >>106391314
>>106391285
yea then you see my point
yea you can do it in rust but its way messier than c
Anonymous No.106391305 >>106391488 >>106391521
>>106391205
The same way I feel about any compiler regression. Wait, do you think compiler regressions never happens in C or something?
Anonymous No.106391314 >>106391330
>>106391294
To each their own I guess, your C code was unreadable to me.
Anonymous No.106391330 >>106391352
>>106391314
yeah but youre biased
only stats can tell
Anonymous No.106391352 >>106391409 >>106391451
>>106391330
>yeah but youre biased
As are you. My code implements bucketsort for 11 different number types and a test within that snippet. It's very expressive for what it does.
Anonymous No.106391409 >>106391443 >>106391451
>>106391352
>As are you
no, im not
programming is just one of my skills
im almoost as good in programming as im in psychology. in the large term, i know my shit through marketing training

>It's very expressive for what it does.
yeah but the idea is how descriptive your symbols are

when you optimize your workflow, not many things you do are multiplicative so to speak
as in: do this and you will code 10% faster

in my experience thats not how it works
not with c at least

no, you create a workflow where you delegeate as much as you can to automatisms, conventions
like hard example, not necessarily algo-related, but an example of the virtue of conventions:
i work in c
and bc i use the same name for variables as for the funcion parameters i can add a parameter to a function im refactoring with nigh zero effort
this kind of shit
Anonymous No.106391443 >>106391458
>>106391409
>incoherent babbling
You are good at neither. You should start by learning English first.
Looks like you have a weird mental urge to have the last word so I will be the bigger man and let you have it out of pity. Enjoy.
Anonymous No.106391450 >>106391491 >>106395744
>>106391285
>that generic templating over to_ne_bytes()
holy fuck I kneel.
Anonymous No.106391451
>>106391352
>>106391409
>no, you create a workflow where you delegeate as much as you can to automatisms, conventions
so your programming time is additive bc theres some stuff you delegate to tools, others to method
and you end up reducing certain prosneedures to 0

my original point: efficiency improvements are not multiplicative
theyre additive
bc theres just shit you dont do anymore bc you have methods and tools

and thats why optimizing for what would be mere 3% of devtime seems ridiculous when youre at the beginning of your journey
but when your nearing mastery, that 3% is some 3% of the remaining 20 of your devtime
which actually amounts to 14% of the devtime remaining

again
bc devtime opti is not multiplicative, its additive
Anonymous No.106391458
>>106391443
>t. seething
ask chat gpt to translate it for you

im the smart cookie and youre the learner
and now you have the tools i have 0 reason to slow down just for you
no
try to catch up, fgt snowflake
Anonymous No.106391488
>>106391305
But then you are completely and utterly retarded, for it was not a compiler regression according to the rustc developers. https://github.com/rust-lang/rust/issues/127343#issuecomment-2218261296
Anonymous No.106391491
>>106391450
Pretty cool isn't it?
I love the Rust APIs
Anonymous No.106391510 >>106391528 >>106391575 >>106391647
>>106391285
how the fuck does that weird trick work? I didn't even know that was a thing.
Anonymous No.106391521
>>106391141
>>106391305
This comment is intriguing.
https://github.com/rust-lang/rust/issues/127343#issuecomment-2295640382

>A lint for easily breakable inference items would indeed help. Another long-term idea would be inferred-types.lock, a lock file that would contain types inferred for the build. (Exactly in what format, seems like a challenging problem, but that doesn't sound like an impossible problem.) That would help keeping unmaintained or slowly maintained binaries from breaking with new compilers with type inference changes.

>inferred-types.lock
Anonymous No.106391528
>>106391510
imagine having to hack your lang to build a bucketsort...
Anonymous No.106391575
>>106391510
btw
heres how i init the thing
the first two lines im creating two memory pools i wont have to deal at any point
later on i just allocate my node object from one, and the full path (bc thats what were dealing with) in another to help with memory locality

2 lines
to set up a whole optimization

i dont need your red tape
i legit just dont
lemme do my job, tf do i care about your bells and whistles?
Anonymous No.106391618
>>106391285
>that font
b-bruce? Is that you again?
Anonymous No.106391647
>>106391510
ok. I didn't know about const field trait members. neat I guess. not sure what I'd use it for.
Anonymous No.106392393 >>106393259 >>106393876 >>106395744 >>106395744
>>106391285
Get real
Anonymous No.106393259
>>106392393
only difference is bytemuck is doing the heavy lifting of turning T into a &[u8]. basically sammu shittu otherwise.
Anonymous No.106393876
>>106392393
Nigga if I was real I would not be implementing any sorting algo
Anonymous No.106393939
where in practice would you need to make up sorting algorithms? do people in kernel do that?
Anonymous No.106395520
>>106388865
"complete nocoder" theory confirmed
Anonymous No.106395744 >>106395754
>>106392393
>>106391450
>>106392393
to_*_bytes() methods are in num's ToBytes trait. that's how you do this generically, while allowing any type in the ecosystem to benefit from such an implementation. it also removes the need for using macros.
but really, you shouldn't bother with freshmen/sophomores who are too retarded to even do a simple search:
https://crates.io/search?q=bucketsort
you get 2 results here. the second one (algos) tells you in its README that it implements a shitload of things including bucketsort (didn't look at code).
the first one (stringsort) tells you in its README why bucketsort can be shit.
Anonymous No.106395754
meant to mention >>106391285 in >>106395744.
apologies