I just watched Greg Kroah-Hartman talking about Rust in the kernel and explaining the reasons.I am now convinced and I hope they will deprecate all C code for good. The kernel is the largest software project in the world and gets way too many changes to allow unsafe code.https://www.youtube.com/watch?v=HX0GH-YJbGw
>>107262606You are not a kernel developer and likely not much of a developer at all. Nobody cares about your opinions on which languages belong in a project you know nothing about.
>>107262653yeah well he is the future king of kernelIndia calling back
>>107262606>>107235820
>>107262846counterpoint: rust has literally been invented to tardwrangle interns... and jeets
>>107262846C and C++ kind of ruined the reputation of the ISO by standardizing shit languages. Ada, Fortran, and the rest of the ISO standards were pretty good and battle-tested languages.
>>107262927>c is a shit languagesuuuureand thats why cuda is an extension thereof>muh seppleslol.if you use iteratorshit in cuda code youll get beaten to death with keyboards
>>107262653
>>107262962>and thats why cuda is an extension thereofCUDA looks like C for the same reason Java and JavaScript do, because C programmers would complain.
>>107262606"that's the whole point, is to make things fun to work on :D"YEAAHI'M GONNA HAVE FUN TODAYI'M GONNA EAT BANANAS ALL DAY*40+ years of magnificent infrastructure stops working, wealth gets siphoned upwards even harder*
>>107263022yeah bca) we dont need no fancy shit, bccccccccb) were good enough not to need bells and whistlesandc) you need to be at least this good to play around with gpus. which are a very interesting architecture, but you wont be able to follow if youre a midwitand d)anything else than c-like just doesnt make sense in the context of gpgpuyou're scope is the kernel to begin with.and cross-scope manipulation comes at a hefty performance pricewhich maps to an array of arraysiterators make zero senseespecially iterations themselves work differently
>>107262914you are confusing rust with gorust IS indeed easy, but it's easy for the competent/capable, and targets fixing their actual pain points.go was explicitly made to be "easy" for retards (explicitly and publicly stated goal).
>>107263078if youre competent you dont need bells and whistlesyou can keep your ducks in a rowim the lower end of it>help with memorydoes not rhyme with>superior skillsidk who told you the opposite but theyre not your friend
>>107263067>anything else than c-like just doesnt make sense in the context of gpgpuCUDA isn't C-like at all, it just looks like C because C programmers have a tendency of complaining about anything that doesn't look like C. It's an array language made to look like C, the same way JavaScript is a Lisp-like language made to look like C.
>>107263098thats not why it looks like cit looks like c bc youre working on vector typesin the sense of mini arraysand anything else would be retarded bloatplease proceed to shut the fuck up already, thanks in advanceor at least learn what a work item is, and cross reference it with how memeory is fetched on a gpu, and the independence in the order of kernel execution (per workitem)it makes negative sense to use iteratorshit
>>107262927in case it wasn't clear, >>107235820 is referring to the fact that the linux kernel is not written in standard C lol.hell, they even just enabled something called "Microsoft C Extensions", which you would think would have been spammed here, if only for the meme potential given by the misleading name. but nope, nothing. nada. which should tell you how "organic" the anti-Rust spam here is.
>>107263124>it looks like c bc youre working on vector types>in the sense of mini arraysMost languages have arrays and vector types. C is one of the languages that doesn't have real arrays. The whole thing happens in parallel on the GPU, so the C syntax is lying to you.>and anything else would be retarded bloatHaving real arrays isn't "retarded bloat," it's making the language match how it actually works.
>>107263146>t. crybullyno, rust is hated here bc its users are a bunch of cuntsnotice how theres no hate towards perl for example
>>107263165yeah but nobody gives a shit about your bloatyoure working at the level of chunks of 1's and 0's on gpus>propa araysno, its just you cant be trusted with a counter, thats what>make programming easier for melul
>>107262606The reason is obvious, because rust is easier to use and guarantee quality while having the equivalent performance. The problem is all the politics around it.
>>107263202>best performance saaarhelo, reddit
>>107263095i cut my comment short intentionally in anticipation of this retarded nocoder response, which also references fictional creatures.>automatically and statically checking and formalizing logic for you with zero runtime overhead is "bells and whistles"or my favourable shorter version, since it shows the full retardation of the argument at the very basic level:>automation considered harmful
>>107263214You're comparing a several years code with recent code. Of course the recent code is going to perform worse. If you want to complain pick the right target. The target is Ubuntu and Debian using unfinished software.
>>107263172perl is not putting (((anyone's))) controlling position in danger.
>>107263225>static analyzers catch everythingyoure boring, redit
>>107263240>even uses poltard notationur pathetic. go back to /lgbt/, faggot
>>107262606anything is better than pretending to be a verbose assembler of a pdp11 which is Ccommon guys, c was already bad back then, it was never good. i am not arguing to go back to lisp machines but there is a ton of better options
>>107263243static analysis is as good as the info it actually has about code.it also needs to not break most code it's supposed to be running against.the fact that C/C++ static analyzers are shit is not something inherent to the concept itself, but to the limitations within which they have to operate, which are the exact limitations rust was created to not be bound by.
>>107263279you sound like youre seething that youre filteredwhich is goodthe good stuff should be gatekept from drooling retards
>>107263214And btw another target is the wokistadores that invaded rust that are the problem. If they didn't exist none of this would have been a problem, and yet here we are with these obnoxious zealots telling rust is perfect and must have politics when it's a kitchen sink programming language, somewhat better than C or C++ at maintenance. It's not that impressive but I understand the technical appeal. Also it's only really useful for some scenarios, like system tools or lower level code.
>>107263288the actual instructions you run depend on the data.im not reading the rest of your brainvomit, sorry, im just notyou just pretended that static analysis is the end all be all
>>107263172There was a lot of hate towards Perl but it's dead now, so nobody cares anymore.
>>107263325were way past the point where stopping shilling rust would stop the hatei think the whole internet wants to smother you nowlmao
>>107263291no, the worse part is that i do program in c and i am good at it, and i program microcontrollers and industrial shit in which there is no alternatives, well there is actual assembly or some proprietary bullshit that makes you attached to the vendor toolchain, and unable to change vendors in the future.why the hell C is still used in computers where you can chose alternatives is something that i cannot understand. you guys have no idea how good you have it and yet choose masochism
>>107263291>the good stuff should be gatekept from drooling retardsIt is, for example that's why you're not using Rust.
>>107262786>lubricate the transitionyou just can't help yourself can't you rustranny?
>>107263369>no, the worse part is that i do program in c and i am good at it, and i program microcontrollers and industrial shit youre not as good as me thenand its not a matter of iq, but of work practices-designing an algo also means identifying its limits. thats where you put your error checking mechanism. document limits youre not checking for performance reasons-test very often. even when you add a couple lines, nobody is infallible, especially you-when testing, test for destruction. try to make your algo fail, not succeed. tests your error managing mechanisms too. stresstest your code. fuzz your code-follow the lifetimes of your data. print it out to double check the changes youre doingyou suck at c not bc of iqyou suck at c bc you have add zoomie brains and you cant be arsed to do the boring shit
greg always was a massive faggot and probably some sort of red hat agent
>>107263464nigga do you understand what a fucking microcontroller is? it is pretty obvious that you are not readingthe "fun" part is designing the machine and seeing it work, programing it is just another chore, programing languages are not to be enjoyable
>>107263583a tool is not supposed to get in the wayand, no, when i use a hammer, i dont need its to be a multifunction cordless drill at the same time (even if the opposite can be true)thats my position.yeah i know programming is a chore. someone smart once said:"programming is 99% of chores like mopping the floor of a mcdonalds toilet vs 1% of the purest creative joy one can hope to experience"now imagine using a shitty mop on top of that...
How come Rustsissies can never create something new from scratch?They always take someone else's work and just transfigure it into a new grotesque form.
>>107263629lack of skills
>>107263622A 32kg mop called Samantha?
>>107263673i *did think about that
>>107263629servotypsttiny glade
>>107263751these names look like a cat ran over a keyboardwhat do they do?
>>107263622>now imagine using a shitty mop on top of that...that is what is using C... i dont need to imagine shitassembly would be not even having the mop, needing to usebend down and clean directly on the floor
check out my post guys>>107263044
>>107263803depends on what youre doingif youre doing low level shitfuckery or gpushit,anything else- is>muh migrogondroolersok, try x86-64 now
>>107263095>if youre competent you dont need bells and whistlesIt's the opposite, competent developers are the ones who understand their limitations.Jeets, on the other hand, are overconfident.
>>107263850thats cope for being a webshitteryes it took me 300klocs to write error free cbut it doesnt mean you couldnt do the same if you wanted todidnt do that for the lulz btwi work on a number crunching project since 3 years now, things do add up
>>107263848That's worse and uglier than just writing it in assembly. C is the enshittification of programming.
>>107262606>comments are turned offto prevent any kind of debate or questioning of the government prescribed death
>>107262606mature code > new buggy code in a more secure language. Every item they switch to a new componenet will get exploited.
>>107263917its not even actual low level, bc c doesnt follow properly the low level logic of x86so is just more difficult for no real reason, which was my point from the start, he is proving my point
>>107263917>nuh asemblieyoure retarded.i have no other word for that>juggle everything while staving off a register pressure of 16 vector registers,>the lookaround capability of the processor>logical memory fancing>no, do that instead of using a tool which automates all that, aka, a compileryeah noautism is a daily struggle, isnt it?even in mu c code i have to take all that into account but at least its manageable+ i benefit of the big fukken brains of the compiler people>no, i'd winbe my guest
>>107263877Best C developers can't write error free code.But you, a jeet, can. I believe you.
>>107263993>best c programmers>aka unpaid volounteers, working after hourssuuuure, suuuuure>inb4 you didnt even trybc why would i? its optimization, bishh
>>107263966The compiler's not optimizing anything. You still have to write all the instructions, except the names are longer and the code is uglier. A real language would let you write (x[0] + x[2]) + (x[1] + x[3]) and it would optimize it and generate the vector instructions automatically (for whatever CPU architecture and feature set you choose).
>>107263993>>107264022contalternativelyits either:you expect corpos to write compilers for anything more complex than cor you expect that a bunch of unpaid volounteers are gonna write the translation layers for the embbeded and shitsuuurewhy the fuck not, just dont ask whos gonna get sued if things inevitably break downyour optimizm is refreshing, rustrany
>>107262606(((Greg Kroah-Hartman)))
>40 million lines of C>deprecate all C codeOk bro. Good luck.
>>107264078they'll find a way, don't worry....the power of open sores
>>107264078That's like 100,000 lines of a real language.
meanwhile apparently open source projects can't be fucked to support 5+ year old hardware anymore, which used to be one of its main upsidesi guess it's because of how bureaucratic the process has become, and the fact that every single project is a clusterfuck with 10,000 dependencies, trying to do too much, to the point where a single user probably can't modify it to their own needsand even if they could, rolling-release updates and dependency sprawl means you'll have to keep updating that shit and making sure it doesn't break every 3 dayshell world
>>107264036it rearranges your code to better fit your processor specificationsi did test it.it turned explicit inline directives into jumps.i checked if i could do better if i inlined instructions manually, but i saw a degradation in performancetheres things you dont have a concept of, asmletyou learnt asm and thought youre king of the worldbut turns out even asm gets abstractedi suspect what came into play here is the lookaround capability of the cpu crossed with its out of order computation capability.it looks around code, and estimates if code is gonna be ran again on a different set of dataif it is, then code is ran preemptively on a wider amount of dataif everything is linear, parallelizable instructions are further apart, too far for the cpus lookaround capabilityi think thats what happens, and why counter intuitively using jumps actually accelerates code-bc it gives hints to the processor itselfi bet you didnt know that, asmlet
>>107263993nobody can write perfectly error free code in any languagebut mediocre c programmers can write code that doesn't have memory problems, since that's a retardedly small subset of errors that you can manifest
>>107264036>>107264139actuallyinlining that doesnt change a thingi wanted to post this pigrelthe prv one is just something i did to contain the spaghettiand left it in bc it doesnt have any impact on performancenot an appreciable onenonoits when i started manually inlining *this picler that i saw quite substantial performance dropsthe formatting hints to the compiler the logical unitsthe compiler didnt think much, it chose to put that behind jumpsi though further inlining could helpit didnt
>>107264089>That's like 100,000 lines of a real language.a real language which was probably written with C
>>107264166you can write perfectly error free codeyou guard your entry points and then pay attention to correctness on the insideits not iq even if theres a cutoff at 100its method.>but what if its someone/something else that does the allocationsyou follow the lifetimes of your data.yes, even if it takes looking into others code>but what if icant do thatthen the failure is 100% on them for not following interface specifications
>>107264257nta. but i wouldn't use this argument since C is not implemented in C anymore ;)
>>107264257>a real language which was probably written with CReal languages need the compiler to be written in their own language.
>>107264313>since C is not implemented in C anymore ;):odid it used to be? i actually don't knowi thought C was implemented in assembly or something
>>107264335i seeseems prone to weird types of failure or exploitation thougho.o
>>107264284cont>then the failure is 100% on them for not following interface specificationsand this is where rust intervenes.which is a thing, it is a nicheonly rusts rendition of that looks like half a limp dick hanging out of an assalso whatever lang that has the ambition to fill this nicheby design attempts to step in for the lack of competencetheres no three ways around thisif someone cant follow simple instructions, they shouldnt be programmingbut industry wants to devaluate the trade because of the demandand so were here, with inadequate training, and toddlrelangs bc people cant be arsed to pay attention
>>107264336GCC moved to C++ (C with classes style) a few years ago.CLANG/LLVM was C++ from the start.all other implementations are irrelevant.
>>107264370why do they do this......
you're all boring
>>107264063this.kreg croak should give his last croak and gtfo of linux.kreg is the faggot who wrote linux defive drivers 20 years ago when he wanted to teach people about linux and bring new devs.when asked why he hasn't written an updated ldd book, he said "we have all the people we need now" so there's no need for the book.
>>107262606>Rust in the Linux Kernel, why?Greg KH explained it very well in the video, what's your question?Rust is popular among those who actually write code.C is popular among tinker trannies and "tech enthusiasts" like lunduke.
>>107264467Because C is trash.
>>107264657;-;was it always trash?
>>107264641
>>107264663i guess the edgy twitter shitflinging dopamine response would be "Yes." with no elaborationi'll kill everyone herenormies stop using my computers stop programming stop having """tech""" jobs etc
>>107264663yes, page 211 and followinghttps://web.mit.edu/~simsong/www/ugh.pdfor se what jwz says of c at the start of his java ranthttps://www.jwz.org/doc/java.html
>>107264701oh shitty anon
>>107264663yes. but for a while, the only (mainstream) alternative was 1000x trashier, and that ratio covers both quality and quantity.
>>107264701>>107264705>>107264736holy shit3-way group therapywhere did C touch you?
>>107264747>oppenheimer-gun.jpg (183 KB, 980x653)i do that
>>107264759more people should be doing that
>>107264767no...
>>107264767nic egif though/.....................................
>>107264819yes.>>107264831i think you should reconsider
>>107264705the unix hater handbook is outdated, it has 30 years, but its a good read to understand that many shit was not designed at all and was hacked around, bc it was written by the people that knew how it was when only mainframes existedeven if you like posix and c is good to read the opposition argument to know where to stand and how shit can be improved, bc if not you end with the idea that is the ideal system, and thus it never improvesbut yeah the point is that C was a hack into a pdp11 which was really basic and dumbled down computer compared with the big boys, and C was written with that hardware in mind, and still has all its limitations, bc it barely advanced at all.
>>107264863nigger we built the world trade centeranyone can do anything in this worldnihilism bad>>107264869tyvm anon
>>107264869>C was created with pdp 11 in mindand thats why it runs everywhere.the handbook of unix haters sounds like an asshurt cope post the size of a book
>>107262914>>107263078Rust is just a cock cage for the trannies, they love their fashy daddy borrow checker.Jeets unironically get filtered by Rust, same thing with AI. They will produce spaghetti that won't ever compile, but the existence of unsafe means Jeets will just do that.
>>107264942conformism is the true nihilismwhen you work for someone else youre not working for yourselfand nietzche is not nihilismits anarchism, appliedonly you seem to have gotten filtered by the distillation processsomething something stepping through ones shadows etc
>>107264985hmmm...
>>107264979thats what big corpo wants from rust doetheres a massive trend to outsource everything rnthat its not exactly designed for purpose is another thing
>>107264985>when you work for someone else youre not working for yourselfbut what is the nature of working for yourself?i mean, aren't there cases where the money you make from working is worth more to you (in some way) than the time/effort/blood spent?it's not a zero sum game either, is it?maybe that work will be worth more to the one paying you than the money is worth to that thingright?is it possible/feasible to truly "work for yourself"? what do yu mean by that?
>>107264947>and thats why it runs everywhere.Why do you want something bad to run everywhere?
>>107264985>its anarchism, applied>only you seem to have gotten filtered by the distillation processwhat the fuck......
>>107264671So either nobody uses Rust according to /g/ or Rust runs the entire internet. Which one is it?
>>107265042that you have an objectivethat you chose the objective yourselfand that youre doing activities that bring you towards the completion of the goal you set yourself.most people work because "its just a thing you do" then end up threading water and at age 40 they have midlife crisis because they never worked for themselves.their whole life revolved around facilitating their work for others
>>107265094>most people work because "its just a thing you do" then end up threading water and at age 40 they have midlife crisis because they never worked for themselves.>their whole life revolved around facilitating their work for othersi see, ty...
>>107265014The best way to kill Rust for good would be to teach Pajeet about unsafe and promote Rust in India.When all of crates.io is unsafe Pajeet garbage, the trannies will revolt, try to make a new language without unsafe, and fail.
>>107264947runs everywhere without using any of the modern features of the hardware, which is missing the point of such a "low" level languagec is everywhere bc is so basic and uses so little of the hardware bellow that an a compiler can easily ported, and that is a huge point against ityou only need to implement the compiler once, and nowadays we have only a handful of relevant architectures, thus negating the need for such easy to make new compilers. If you forget the more iot and industrial isas, there is only two relevant ones, x86-64 and arm64
>>107265103cheers>>107265073why immediate extremes?how bout- rust is very rarely used and when it is it crashes and burns
>>107265073nobody uses it and it doesnt run the fucking intenert it doesnt run ANYTHING it CRASHES the internet like it crashes everytihbg else it toughes why do you have to repeat their lies
>>107265123The whole PDP-11 meme is very incorrect and your ignorance shows.The meme originated because people wanted LISP machines, which for obvious reasons are not a good idea.Even Stallman came around to liking C eventually.It's been retrofitted to mean concurrency, but concurrency is still NOT A SOLVED PROBLEM.
>>107265121thats probably gonna happen of itself>>107265123even with full idiomatic c you can write architecture-tailored code, even down to os specifics.you can unroll loops to maximize ilpyou can use prefrred data types and preferred data sizesthe performance of your threading will depend on the number of cores availableprolly other stuff i dont think aboutand then theres extended cwith compiler intrinsicsand c derivatives/extensions like cuda or opencl, tailor-made for a specific architecture typeand then the wealth of microcontrollers in the wildi dont think you have any idea of what youre talking aboutbut yeah, c could see some modernization/standardizationwe dont use c bc its perfect, we use c bc its the best for the jobthe bar's pretty high tho.c undefeated since 50 years now and counting
>>107265123>>107265182Furthermore, x86 has always just been a faster PDP-11.Intel's concurrency approach has been to run multiple PDP-11 in parallel.And when they try to do something fancy like speculative execution? It blows up in their face.
>>107265237this take is so bad it must be a false flag to make lisptranies look bad
>>107262606>Saaar so we like we are retarded RUST IS SAFE AND EFFECTIVE as we seen today it fucked half of the internet saar These people need to fuck off with Rust. C is the only way.
>>107265131>rust is very rarely used>literally runs the internetHmm doesnt add up
>>107265280>literally runs the internettahts firmware of your routers etci think that tends to be written in cbtw>run the internetwe all saw that lmao
>>107265280or it doesn't you fucking microsoft jeet agent or troon you are one of those since you propagate this troon language.
>>107265280can you elaborate?>>107265293>>run the internet>we all saw that lmaodid they "edit" a typo from their post??
>>107265182>which for obvious reasons are not a good idea.if its so obvious you probably can tell them
>>107265306stop being a boring shitflingerjust ask what they meant
>>107265322Rust is objectively terrible shit also cloudflare is retarded for using .unwrap in a fucking prod. retards people don't have enough experience with this language SAAAR so they do mistakes like this...
>>107265331that's a great new thing you just brought up
>>107265322also... your body my choice ... chud.
>>107265280>>literally runs the internetruins*
>>107265237>Furthermore, x86 has always just been a faster PDP-11.x86 has string instructions and segmentation, which make it more advanced than the PDP-11. Even the 8086 had those features. The 286 is actually closer to a mainframe than a PDP-11.
but you should realize, posts like this >>107265280 are probably intentionally vague glowie bait postsprobably trying to disrupt the threadthere's usually no better way to respond to that shit than to ask what they mean
>>107265353It's a jeet agent from microshit im telling you this people want to ruin linux with troon language and they cope hard.
>>107265359that's great, manhttps://x.comsmartphones were a mistake
>>107265351imagine if they had managed to do a mainframe on a chip wikipedia.org/wiki/Intel_iAPX_432
>>107262606>Stay strong sis
>MY LANGUAGE IS BETTER<NO MINE ISPure assembly should be mandated everywhere and anything else should be banned
>>107262606Here is what breaks rust. Rust relies on reliable hardware. If your hardware is unreliable it immediately invalidates the entirety of the rust stack. Literally, the whole premise is that you believe you have the value the compiler asserts. What happens when you just don't? Well, in C and every other language that exists everything is built defensively. You assert your invariants and check them.If your hardware can be subverted then you can dynamically execute whatever you want under rust.Async cancellation is what broke me. Such an absolutely shitty implementation for futures.
All this could be avoided if everything was written in HolyC
>>107265631i today learned(llm) that dividing is between 50 to 80 cycle of cpu while it's 5 to 8 for multiplying and it's pretty easy to check that out on godbolt.
>>107264947>the handbook of unix haters sounds like an asshurt copeit has some good points but much of the book is silly. ranting about things isn't terribly productive at best, and quite annoying at worst (read: skill issue), but when paired with *actually fucking suggesting improvements* or changes it can be very useful. but i do love reading a good rant
>>107265631this would unironically improve software quality across the board as normies would get filtered
>>107265455OS/400 is one of the most reliable and secure operating systems, and it was written in PL/I and later re-written mostly in C++It can run for years with zero downtime and programs run without hardware process and memory protection! The trusted translator was mathematically proven correct.Yes, It’s that good.Linux was always fundamentally garbage, and should have never even taken off with it’s “giant lock” it has for so long… if people had critical thinking skills, that is. It’s a cheap minix clone.>>107262606Who cares what this guy says. Just another fad follower.I hope this puts things into perspective.
lmao @ these easily offended cniles>>107265293Your router didn't help when the cloudflare was gone did it>>107265306Can you try again without sounding mad?>>107265311How come Rust runs the internet if nobody uses it? That was my initial question.
>>107266463>How come Rust runs the internet if nobody uses it? That was my initial question.that's my question tooi forget which anon it was forit's a big claim, and i just thought the technology board would know about this
>>107266672It doesn't. Most backends are in Java and frontends in React nowadays. That's not going to change anytime soon. If that anon is talking about routers most are written in C or C++.
>>107265073You're probably confused with cloudflare that did recently move code to rust but that doesn't mean most of the internet runs rust. Cloudflare is hardly the internet. It's a DNS, reverse proxy, ddos protection, etc., and many companies use it, but not the internet. Also only some components they have are in Rust.
>>107267678Nigga everyone uses cloudflare
>>107267678Cloudflare mitms most of the internet you retard.
>>107262606> why?To deGPL Linux so it can be used by Microsoft to make Windows 12.
>>107267944>>107267997Wrong and wrong. They have around 20% of the market. You two stupids didn't even do any research. Besides what cloudflare provides is not the internet, it's just a part. Just because they got big clients it doesn't mean it's most of the internet.
>>107268049the NT kernel is miles ahead of the stinky linux kernel lol
>>107262606HRT destroying the nerd community like crack destroying the Black communities in the 80s
>>107269338literally the best that has come out of microsoft
>>107269525Dave Cutler got a well-deserved payday, for sure.
>>107269338And yet crashes a lot more and doesn't even have a proper graphics abstraction. The best kernel for sure.
>put rust in tranny kernel>shocked
>>107268057Do you understand what market share means?Cloudflare covers at least 80% of the entire traffic
>>107262606It's a modern language with modern features and improved safety features. C is very dated, everyone upgraded to C++ around the year 2000.