The irony of this absurd statement... A language is not judged like that. Especially since there are multiple Rust communities and not only one obviously. Nor is C made of angels either. I have also different even political views than the Rust foundation. But Rust fixes a lot of issues that neither C nor its perfect community is fixing. You can wait and dream with the bold claim like I was told that somehow C got abandoned because it was easier that way. No, son. Until your fantastic C community did not deliver an actually better language then I am afraid the Rust community is going to be the least of your problems. Even Windows is already getting oxidized... Resistance is futile. Big corporations essentially already pushing Rust whether it actually is a good language or not. Though it is not that hard to do better than C... Lol
More specifically a "memory safe" kernel would only be suitable for virtualization because it would not have multiple fail safes to handle actual hardware. Essentially its Communism trying to reinvent the currency.
And the very things that make Rust better than C for some tasks IMPEDE its use in a kernel. The Linux kernel has been around for at least 33 years. Any major bugs that Rust's trappings can prevent are long-gone...and more importantly, the finer control over things like memory means that the kernel is tuned to an extent that Rust would never allow. Even when it comes to new components, why not write them in the fastest, most efficient language, and simply take the extra care to not write in the kind of bugs that Rust is supposed to prevent? This is a question of development speed versus efficiency and stability. In the case of the kernel, efficiency is THE priority.
This sounds plausible, but not realistic. When you talk about new stuff, the amount of device drivers added to linux kernel is increasing by the day and not every driver developer is Greg Kroah-Hartman or Takashi Iwai level developers that don't make silly mistakes at all. They turn out to be random dudes working at random hardware manufacturers. Rust solves majority of the memory bugs they might introduce to the kernel and speeds up the process by a big margin.
@@freedoompictures6839 They also produce slower, clumsier, and in the long run less stable drivers. Better to do a little actual work and get it right.
It's a growing problem, devs are working on the linux kernel to fix issues that affect them, but they're not pushing their patches upstream because dealing with the politics of the upstream project is a waste of time.
People confuse their own abilities with ownership of a project. Most do not realise their changes can have a profound impact, many others simply refuse to go along with the rules. Most people able to do kernel development are lone sigma wolves, very good at doing their own thing (they are usually brilliant) but terrible at collaborating.
@@sillonbono3196 Bruh, most people in open-source are terrible at collaborating. The fact that they think "this software is free to use, modify, and distribute" also means "this software and its developers are free from criticism" (and make no effort to act as though they don't) proves as much. And saying someone's "good at doing their own thing" is like saying they're "good at breathing."
@@AClockworkHellcat "The fact that they think "this software is free to use, modify, and distribute" also means "this software and its developers are free from criticism" Yes, it's the entire point of OpenSource software. Hackers aren't exactly social butterflies. It's never going to be smooth sailing.
@@sillonbono3196 I was with you until you made the false claim that kernel developers are lone wolves. If they weren't good at collaborating, then the kernel wouldn't be what it is today. If they didn't abide by the rules, then Linus would've smacked them down, and he has done so quite a few times. The thing to remember is that the core of the kernel is basically from fewer than 100 devs. Those are the ones that really count, not the thousands who develop single purpose drivers for a weird piece of hardware they bought at a flea market 10 years ago.
Finally clicked for me: It's another power grab: >Take over a shiny new language (Rust) >Make sure there are mainly comrades in Rust Leadership >Make banning people for political opinions outside the norm >Replace C with Rust If this works out, Linux is dead because nobody with any technical skill is left. But I think the tide is turning.
To add, The entire "safety/security" argument doesn't work when what you have is a software that has been out in production for quite some time, and has most (lim -> inf) of its bugs and security issues ironed out. Slipping a non-bootsrap-able alpha-stage non-standard toolchain into a codebase as big, important and tested as linux kernel smelled like rotting fish from the very beginning.
Please do not politicise programming languages. And, yes, I believe - all other thins being equal - that Rust can produce more memory safe code, but has downsides as well (many high abstractions) and unsafe Rust is in some ways worse than C. C developers not helping Rust developers is natural.
I am with the C guys here... Rust ppl are like a weird woke group of weird cult... Not a "woke cult", but mostly woke furries who are also cultist about rewriting everything in rust - awful combo!
Just look at Pop OS Cosmic. The main dev is a mentally ill man that think is a woman and wears animal ears during conferences and post trans flag with trans right message in his presentation. While Pop OS spokesmen refer to him as "her" "she".
People who have worked on the kernel in C for years obviously won't magically hop onto rust completely destroying their workflow and productivity. Learning a new language can take years to become proficient in not to mention that the person need to want to learn it which they obviously don't have to. It's funny how Rust support was added because people wanted it but where are the Rust developers now? I don't see majority of submitted code being Rust how the loud people pushing for support made it seem like. "Add support for Rust and you'll get more young talent contributing to the kernel" yeah that didn't happen
Eh, I don't think Python _use_ has much drama anymore. It's a great glue language that everyone technical can understand and its libraries are usually in the top 5 for anything so a great starting point before investing the time and money to develop something more efficient.
4 หลายเดือนก่อน +6
It infects, but not with drama, but with abstractions C developers don't like. Better to start over (Redox OS).
Honestly if I hear something is written in Rust, I take it as a red flag and simply move on. It's a fully pozzed ecosystem now, and at this point people seem to use Rust primarily as a political statement with technical suitability taking a backseat to ideological uniformity/gatekeeping.
It was always a political statement, that's why Rust had people filing GitHub issues on random projects saying "rewrite it in Rust or you probably want your software to be insecure you prick", and they wouldn't take no for an answer.
@@sunderkeenin I disagree regarding technical merit. There's technical merit to learning how to properly use C++ and the various *_ptr types along with when to use move and forward and r-value and l-value references and their semantics, especially if you already know C++ and/or C. Moving to a completely new language because it makes false promises and looks like Perl barf is just a massive waste of time.
Relax. Rust is actually quite suitable for writing small CLI tools. Consider it a very modern C++ in this case. Lifetimes are very simple in CLI tools, so the type system aren't that difficult to deal with. Not much ideology involved here.
@@generic694 You just described a half dozen languages, my guy, but I didn't ask for use cases-use cases exist for _every_ language. How many languages have communities that actively target the most engaged devs in a naked power grab to gain control over the project so that they can start revising codes of conduct, banning the most prolific developers & refactoring the codebase to be more "inclusive" in a blatant attempt to gatekeep influential projects from being retaken by dissident contributors? We _both_ know this is ideology-driven behavior; it might as well have a fingerprint at this point. Oh wait-it does! Google around a little & see if you can guess what Saul Alinsky might say about this interesting series of developments. :^)
Not that my two cents mean anything, but, always one of those. If you want to port something, in this case rust, by all means have fun with that. But don't expect others to bend backwards for your "cause".
I don't think he was suggesting anyone needs to bend backwards. It seems there are some Linux developers who have a cult like backing of C and C++ who do not want to support rust for no good technical reason. It's political sabotage. I don't care either way, I'm not a Linux or Rust dev but that's what it appears to be from the outside.
@@ImperiumLibertas Rust isn't a certainty. Having to support it and replace with it what has always worked and every veteran developer is an expert in and agrees is still the best option puts the core project and everything that depends on it at severe risk. Linux is the core of a lot of mission critical infrastructure across the planet. Forcing Rust into mainline and all veteran devs to support it at this point is extremely irresponsible. That is what source branches are for. Leave everything not rock solid and proved outside of the trunk and main release.
@@ImperiumLibertas Its no that simple to "not want to support rust", you like rust and want to port some code? good luck but its not the job of a 20+expertise of C developer to help you.
@@nempk1817 the person from the talk didn't in fact say anything about that c programmers must rewrite their code in rust. Rust has a ways to express some conditions in it's type system so they needed help with understanding that these conditions are to write safer code. I think it's a really big misunderstanding from the side of c developers, because rfl just needed information about what contracts were upheld in file systems code and other implicit behaviors.
@@ImperiumLibertas cult-like backer of C here. people have a cult-like backing of C because it has worked properly, without standing in the way of the developer, since 1972. a fun new toy is just that, a fun new toy. once you're done playing with it, you'll move on to the next. rust is neither cool nor particularly needed in the linux kernel.
Definitely has the tone of a new python dev getting mad that their special-use code isn't being put into a very technical portion of code that runs in hard realtime and has its own pseudolanguage/involved semaphore system to keep everything from bumping into itself.
Rust is just bloated. It takes a whole day to compile its compiler, leaving 40gb of junk files and refreshing the cargo database downloads an absurd amount of data.
Yep. Even a trivial rust application takes too much memory and storage to compile. I quit my previous job because it was making things impossible and I got fed up with the rest of the team not acknowledging the problem.
You must be on a worse computer than my potato. It only takes me around three or four hours to compile each new version. Of course, I delete the whole archive each time so I don't have to put up with the bloat hanging about, and it took me 10 minutes to write this comment while I was waiting on `du` to show me the current usage. For the source folder where I've got the `git` repo, it's "only" 19GB for the version I'm currently on, 1.77.0 compiled on 3-17, and the binaries take up 3.2GB. I don't remember what GCC uses up, but since it handles more than one language it's not so easy to quantify, but it still uses up less space than that and compiles significantly faster. I don't do incremental builds like some developers, so compile time matters more to me because things aren't generally cached.
@@anon_y_mousse I mean, I am not a web developer, but I think the way the cargo database works is stupid. It is downloading the whole database (around 200MB as of now), and searching it locally. And of course updating every time it can. Why don't they make a serverside PHP application to query the DB without downloading these 200MB of garbage??
@@gargamel3478 I assume they learned the wrong lesson from how `git` works. Of course, even if that's the case, the database is too large and there's undoubtedly far too much duplication of effort. They should be more discriminating with what goes into it and prune stuff as needed. Here they could learn something from `git` such as branching, but maybe with the branches not all being necessary locally to have a complete picture of the most recent view of the database.
Doesn't the deep tech state aim to create an AI system to write programs & operating systems that are in C into Rust? That said if were Ms. Beep state, I'd hire human coders to do much of the primary conversion of code.
@@Kirfx rust is ok, but too big of a leap for C devs to get into, and thus limits the changes C devs can make to the kernel, including C code that is used by Rust code
I'm simple > They hate C I love C > They love Rust I hate Rust > They hate the leads of NixOS I will support the leads of NixOS > They hate Linux I love Linux There is a clear pattern there Mediocre and scared people hates talented and passionate people
nixos is overthrown by people like jeremy, wtf are you talking about? they banned all of the so called passionate people because they were not woke enough, dead project by now
The actual maintainers are pretty okay on Nixos. Usually not too political outside of a few bad apples. The Nixos moderation and governance teams on the other hand are out right leftists who slap their friends on the wrist when they break the rules and outright ban conservatives for simply having conversations that other people get upset about.
So, for the patronizing narcissists, anyone who tells them “no, not that way” is a “patronizing pedantic megalomaniac”. It’s no surprise that Rust, a language invented to save mediocre developers from themselves, is their primary weapon of choice. Edit: typo
The funniest thing about Rust, is that it's showcase project, for which it was originally made, (the new web engine for mozilla) have never materialized 10 years later. ;) Yet they try to push it everywhere else. Like its a huge success story.
yeah this is what i wanted to say, popOS and system76 can still get by without being associated with contributing to the linux kernel. system76 previously had a big linux presence where they were solving regressions upstream and sending patches and all that. them saying that they don't want to be involved speaks more than anything. lunduke is extrapolating the comments made by system76 in ways that the author didn't mean and it shows. edit: big fan of lunduke before, got into his work when he was covering the mozilla situation and the nix os stuff, i thought he actually reported what was happening objectively. popOS can get by not agreeing with where the kernel is headed because lets be honest with ourselves, linux does suck. and while alot of people share this sentiment with various other tooling like systemd and linux itself it doesn't exactly mean that they should just not use linux, the alternatives out there like windows are far worse. you can apply this to the system76 folk too. disappointed lunduke.
Nobody is going to recompile their Kernel so that their System76 laptop will work on another distro besides Pop!_OS (worst name ever). People will just stop buying their machines.
@@owlmostdead9492 system76 machines run on other distros out of the box, the hardware is generic. the patches they make are for optimizations. alot of sched optimizations on the linux scene come either from system76 or surprisingly the android custom rom scene.
Rust programmers: I'm asking for help to C programmers to migrate the code to Rust so you end without job and we get your positions C programmers: It's not gonna happen easily. An that type of conflict is as old as the first cutting stone industry.
Should be trivial for them to migrate to Rust. They should be able to read the C code easily enough. If they can't... maybe they're not the programmers we need writing kernel code.
>buhu woe is me people disagree with me Rust is a literal cult by god, if c, c++, python or js devs were as extremely fragile as them the whole computing world would end. Why do these people not just keep out of public conversations and remain within their bubbles like "the rust porting project"
Jordan Peterson is right. When people deny God, the first thing they do is to seek other gods and create cults. The cult of vax science, the cult of Python, the cult of Rust. All fanatical people.
I'm a complete amateur with Rust and haven't even touched it in almost a year so my perspective may not be the best but while I really enjoyed simple rust, at a certain point it gets super complex. The thing is, to get over that hump, you kind of have to really live and breathe it which can lead to a little bit of a myopic view of things. Something like Zig I feel like is a more likely inheritor for the Linux kernel guys.
I like Rust. I'm learning it, myself, though I've known C/C++ for decades, and modded my own Linux kernel 30+ years ago. BUT: Complaining that the Linux kernel devs won't either rewrite the kernel in Rust or spend time supporting others who do sounds like a modern Participation Trophy sort of sense of whining entitlement. If you want to make a port, do it. But you aren't owed the time and effort of others who aren't interested.
There's a reason for it, its called "if aint broke, dont fix it" everyone's fear is that because the whippersnappers don't know what they are doing, we gonna end up with another systemd or pulseaudio, but in the kernel this time.
Has everyone forgotten that one of the 'goals' of rust is to get rid of the idea of 'coding skills, and quality?' They state it plainly in their c.o.c.
Their code of conduct is a woke joke: "We will exclude you from interaction if you insult, demean or harass anyone. That is not welcome behavior. We interpret the term “harassment” as including the definition in the Citizen Code of Conduct; if you have any lack of clarity about what might be included in that concept, please read their definition. In particular, we don’t tolerate behavior that excludes people in socially marginalized groups." Who's more socially marginalized than C devs? /s But seriously, it's like they don't know how the internet or digital technology came into being. It wasn't _that_ long ago that IBM was selling computers to the National Socialist Workers Party of Germany. Should we cancel IBM? It wasn't _that_ long ago that Wernher von Braun helped NASA with their rocket programs. Should we cancel NASA? We're making sausage here - not picking and choosing political ideologies. Do the personal opinions of the charcutier matter when we're purely talking about how to make the best sausage? We as humans have the opportunity to learn from the underlying process, and truth, and reality - regardless the messenger. Here in the United States, it _used_ to be the case we have opinions as individuals, and agree or disagree with the opinions of others, without our access to income or livelihood being revoked. Words are not violence. Violence is violence. if my use of the English lexicon bothers someone, that's a _them_ problem, and 'they' can kindly go f*ck themselves. Is that how 'polite' society treat one another? No - but that's politics. This is sausage.
People need to wake up to the fact that all these horrible people, who think they're a moral authority punishing throught-crime, have been raised years ago and will have children, who will be even worse than them. Your very own children are raised into this as well, in kindergartens, schools, universities.
I was and am still a huge fan of System76, and I thought Jeremy Soller to be some kind of programming god. But then I made the mistake of looking at his twitter. I saw him retweeting stuff that included there is a "literal trans genocide". I had to shudder and gtfo twitter. Jeremy, please keep making awesome software. I hope you don't think there is a "trans genocide", but yeah. Please keep making good software. But thankfully, I can rest easy knowing that Linux will move forward with or without System76 and with or without Rust evangelists who got twisted up with woke ideology and ways of behaving online...
You're so right.... it's really troubling, that these devs are so political now - I never realized how much time they spend yapping about their left-wing politics. At least, you could switch to Ubuntu if Pop OS starts putting their woke ideology into their OS - and it's possible - for e.g. firing ppl who aren't woke enough for them....stuff like that would turn me off, anyway.
I think the core of the issue is that C, for all of it's shortcomings, is still one of the most highly flexible languages around. C programmers like it because if the problem you are trying to solve changes, C lets you squish it and pull it to fit the problem with relatively little resistance. The more often the problem changes and you need to adjust the more that strength of C shines through. Rust on the other hand is noticeably more inflexible. Rust data structures are more resistive to change, it can slow development down or worst case make certain features prohibitively expensive to implement. I think this is intuitively apparent, but people will differ on how much they value "Flexibility" VS "Mathematical Provability". Personally, I think we need a C replacement, but Rust is not it. Getting rid of undefined behavior would be a good start. It would be nice to have well defined default behavior regardless of implementation trade-offs, then have instructions that are hardware specific that make the most optimum trade-off for that hardware. This way we reduce the number of surprises and foot guns.
Rust isn't any less flexible than it needs to be. When you really think through the hidden assumptions in C semantics and attempt to make the implicit stuff explicit, and really make an effort to prevent the most common classes of memory bugs, you end up with something like Rust semantics. Read Aria Beingessner's blog for examples; she, like many Rust devs, actually knows C better than most C devs do. "Squishing and pulling" the code to fit the problem is a lazy, cowboy-coder approach; what you need to do is think deeply about the problem and analyze all of the desired semantics and failure cases and be explicit about those in the code, encoding as much of it in the type system as you can so that errors are caught at compile time rather than throwing exceptions (or being silently ignored, leading to bugs and security issues) at run time. Rust has a type system robust enough to do this; C does not. Rust wins.
@@bitwizeafter seeing a women's name and Rust mentioned, my first instinct was "that's not actually a woman" so I checked and it was correct lmao. The Rust community is infested with these mentally unstable individuals, the same ones that ruin everything else open source they get their hands on.
@@bitwize brother, you haven't programmed a single day in your life. "any less flexible than it needs to be?" c, instead, is not "flexible". c is literally 99.1% pure assembly code in english. it's not flexible, it's literally mapped as close as humanly possible to 1:1 to the way the computer needs its instructions fed into it. and you can do one better: to use a subset of ansi c is to get even closer to assembly. hidden c assumptions? name and explain one yourself, don't give the promise that some random mentally ill individual did so on the internet. "\"Squishing and pulling\" the code to fit the problem is a lazy, cowboy-coder approach" brother, pick one! these are complete opposites! you're either LAZY, i.e. you're bored and maybe lack the technical know-how so you use c alternatives (shame on you), or the balls on you are so LARGE that you're embarking on an adventure to solve the given issue with some sort of optimization-minded competency. "what you need to do is think deeply about the problem and analyze all of the desired semantics and failure cases and be explicit about those in the code, encoding as much of it in the type system as you can so that errors are caught at compile time" people don't like using paper these days to solve problems. and it's a damn shame. people think that rust is gonna code for you as well! if you don't understand the problem, _who the heck does?_ don't shoot first, think later. THINK FIRST, then SHOOT! as for the final statement, rust is a tool for tools -- promising safety to people who simply do not know any better. c has been working properly and as intended since *1972.* c wins.
@@freedomgoddess Actually look up the term "cowboy coder". It's not a compliment. Over 50 years and countless CVEs suggest that even them most experienced, conscientious C programmers are prone to falling into UB and memory-usage traps. It's simply not humanly possible to be as careful as it takes to reliably produce C code that's free of memory issues. It's prohibitively difficult to even determine if such issues exist in the first place. Rust doesn't write your code for you. It just helps you write more correct code by providing static, compile-time verification that it does not contain common classes of memory issues. Again, the Rust programmers know C better than most C programmers do -- which is why they use Rust.
@@bitwize "is a lazy, cowboy-coder approach" This reminds me of a story my uncle told me. (Sorry about the wall of text, but it is a long story.) My uncle worked at a factory that made parts for machines. They did plastic injection, they had mills and laths, ect. And they had contracts to produce thousands of a part per day. Anyway one day there was a change in management. The son of the CEO was put in charge of managing this particular factory. The new head manager was fresh out of college, he knew more about running a factory than my uncle or any of the other machinists. The first thing the new manager did was to change the workflow of the machine operators with the goal of making things "more scientific". You see, mills and laths have lots knobs and dials that control the settings of the machines, and when the machine operators were making parts they were just guessing what the settings should be until the parts came out ok. This meant that each operator was using different settings. One guy might have the knob set at 35, another at 36, and another at 33. And they are all making the same part at the same dimensions. The change that was made was to standardize and require each machine operator to set their knobs to the same exact values. My uncle of coarse disobeyed this order, but many novice operators went along with the new mandates. That's not to say my uncle couldn't read the room, if certain people asked him, of coarse he had his knobs set "correctly". What happened after this change was no surprise. Hundreds of useless parts came out of the machines, off by just a fraction of an inch but not within acceptable tolerances. This happened for days. Operators where not meeting their quotas and they were afraid they would lose their jobs. They would go to my uncle and ask "help me, please, how do you get your parts to come out right?" The answer was simple, do what was done before, put in the initial settings, measure the outputted part and adjust the knob, repeat until it is spot on, then periodically after every few hundred parts measure again and readjust if necessary. Soon the factory was back to making parts. You see, no two machines that manufacture parts are identical, even if they are the same brand. Some machines are more worn than others. Vibration differences because of proximity to other machines. Some are sitting on the ground slightly less level than others. The tools that cut the metal go from sharp when new, to dull and need to be replaced. Temperature of the gears in the machine as it is running. All of these things, many of which you have no control over directly, contribute to making the behavior of the machines different enough that it matters when attempting to produce precision parts. It is still unclear if management has learned from its mistake, or if they even know that a large amount of their workers are ignoring them. It's not a great situation to build a work culture where information cannot not flow back to the people in charge. I think management probably does know "more" about my uncle's job, but how much you know is sometimes less important than what you know. (He does not work there anymore, the factory has since fired nearly everyone and is flying non US citizens in. I mean if you don't want to listen to people it doesn't matter if they don't speak english. LUL)
They're still working on the linux kernel, they're just not pushing their changes upstream due to the problems stated. I hope that clarifies. It doesn't really affect their business model.
@@MechaFenris No. As long as their changes are public, linux can merge the patches themselves without wasting everyone else's time with their pedantic nonsense.
@@MechaFenris it doesn't, their patches are public, and anyone can apply it if it benefits them. and even if they are closed off it doesn't violate the gpl because there's no source of linux associated with a .patch file
i don't know anything about kernel programming, c, and rust. but if some people join on my project then expect me to rewrite my code on another language or help them when it was not needed then I'll be angry too 😅. I don't understand the situation, you want rust? just do it, why expect anyone to do something what 'you' want to do..
In a recent video did the research not show C was more energy and memory efficient than Rust? Seems like a step backwards to me but, I am not a programmer.
I know next to nothing about programming languages, but my understanding is that Rust can prevent certain exploits, specifically it enforces memory safety.
Optimization is not always a complete win-win of "faster and use less memory", sometimes you need to consider different aspects (like memory safety in case of Rust) or more complex factors that could be a bigger priority. Or in short: no tool is perfect, programming languages included.
it is very much accurate to write code in c rather than rust for energy efficiency reason. it's just factual that writing c for such a case is better. it's just that rust is injected by trigger happy people where there is c already so it's hard to make any sort of jump to it with such little interest to it by oldheads.
Imo yes. Sounds like a diva that didn't get her sparkling water just right. The resistance sounds reasonable. C devs wouldn't, shouldn't stop their work to accommodate and assist their supposed replacements, especially when the replacement can't even help themselves and only offer negligible to nonexistent benefits to the kernel. The incumbents joined the linux club to develop, not translate.
What these old C developers in the Linux kernel community fail to understand… Mommy’s little girl has a right to be included in the kernel community, because that’s what she wants Mommy’s little girl doesn’t want to use C, she wants to use Rust - so the Linux kernel community needs to accommodate her wishes Mommy’s little girl has a right to be safe from opposing points of view Mommy’s little girl deserves equal outcomes compared to other kernel contributors, so all her PRs should be merged just like everyone else’s The Linux kernel developers need to share the rewards around in a fair and equal manner Soon it will be time to hand over the reigns, and give Mommy’s little sweetie a go at running the whole project When that happens, be sure that Mommy will be there to help as well
@Bryan Lunduke if you watch whole video, not the small part, it changes perspective. The guy from video was trying to do presentation and claims "I don't want you to write rust code, just give me semantic of the C code" then he received small backlash. Jeremi is not contributing to mainline kernel, pop_os has backports. Anyway I don't care. In IT industry there are a lot of holy wars like vim vs emac etc.
The maintainer of Pop!_OS doesn't have to submit patches upstream. Their focus is Pop!_OS working on System76 machines. This might impede other distributions from working out of the box, but that's not their problem. Their kernel patches are likely available for other distributions to apply, it's not uncommon for distributions to have their own downstream patches. Running down System76 is inappropriate when there is an issue with upstream development.
@mrfuzzy2954 This, right here. ^ Unless the terms of his employment actually require upstream patches, he is under no _obligation_ to work with the kernel devs on enacting them. So, no, it's _not_ like a trucker who refuses to deliver to certain states because they don't like the laws.
Tbh Rust is already problematic because of it's weird foundation and it's own megalomania, anyone dependent on it on a project as critical as the Kernel is asking for trouble, even past technical concerns about dependency management that saw Debian devs abandon it, the foundation is going to be a headache for years to come and i prefer the memory issues that can be resolved with a commit to legal issues or (additional) temper tantrums involving the foundation of Rust making demands to remove Torvalds or something because they don't like him...
Arch and Manjaro don't do this kinda stuff. I only use Arch so I can legally say "I use Arch btw" to randos at walmart or at your grandmother's funeral. Manjaro is still my favorite distro, it's the one I started on. I know this is a hot take, especially among Arch users, but I actually liked that they held back packages. The stuff about "dependency hell with AUR packages" is WAY overblown and I never once have had it happen to me. I still use Manjaro on like 3 laptops and I'm using the stable branch with TONS of AUR software. Particularly on my pentesting/hacking VMs that I use to practice and they are ALL super stable and never fail me at all. Gaming on Arch has been great with Steam, Lutris and WINE/Proton, but it was much easier to get things working when my desktop was running Manjaro.
My experience of the AUR is 95% of packages I want are there and I can just build it myself and it works. Only issues I have are with closer source stuff
@@hardrocklobsterroll395 exactly. I used to use Ubuntu way back in the day, but I constantly found myself not booting into it at all. Then, I found Manjaro and I don’t have windows on my machines at all. The AUR made everything so easy, I can even use .deb packages with that conversion tool, but it’s rare and why I can’t remember the name of the tool 😂 Literally EVERYTHING is in the AUR. Debian based distros are clunky and you have to nuke it every time you want to truly update when a new one comes out. With Arch based, sudo pacman -Syu and everything is done.
@@hardrocklobsterroll395 pamac has been the best GUI package manager, by far, in all of Linux. You just can’t beat it. The one in Ubuntu is just clunky and looks really old.
1. Keep in mind that this comes from a Rust developer. 2. Keep in mind that this comes from a Rust developer. 3. Keep in mind that this comes from a Rust developer. 4. Continue the evaluation of the statement by the Rust developer.
Every 5 years there's a lemming-like rush to learn and port the codebase to some faddish new language, which is considered to be the Greatest Thing Ever until the Next Thing. (Instead of maybe focusing on more elegant code and efficient PM processes?) Not saying progress is bad, but this isn't the path of progress.
This entire video is bizarre. Linux is open source, they can fork and not contribute upstream. They are not required to especially if it's not welcome (guessing here based on the megalomaniac comment). They can ship their product without giving anything back. I really don't know what this video is going on about.
Yes, they can ship their product, but their product is influenced by upstream decisions upstream. So, any OS that relies on the Linux kernel has no choice, unless they fork a previous version of the kernel that was not rusted yet.
And also misrepresents the issue quite a bit. I'm starting to see why people have issues with Lunduke. My understanding of this issue is that some of the C code in the kernel is undocumented and very hard to follow. The Rust developers need to interface with those parts of the code, but when reach out to the C developers/maintainers with questions, they get turned away. And because of the difficulty in working with them to document or atleast explain the interface to a crucial part of maybe the most important piece of software on the planet, they figure they rather not bother working with them and instead look to just re-write those systems in Rust. Naturally, this pisses off the C developers because they see it as Rust encroaching in their spaces. I know both C and Rust, and while Rust definitely has it's pitfalls, it's design reduces the chances of many footguns going off with new contributors. And being such a popular language, it'll invite more and more contributors over time. Linus definitely sees it's potential and understands it's benefits. And until/unless a new language with all of Rust's benefits but none of it's failings show up, it is the way forward in the kernel.
I do like to listen to you, but I think you're over dramatising the situation here. People, I would assume and also meaning I could be completely wrong, purchase their machines because they want a working setup. System76 make it work and ship it, they don't need to push jack upstream. Also for the Rust guys, if you want something done, do it yourself, stop whinging that other people won't help.
Jeremy is also not the only person with the ability to contribute patches to the kernel. Although with open firmware, it's not really that necessary anymore. Lunduke is out of his depth and doesn't comprehend what he's talking about. He assumes rather than investigates. He's taking everything out of context, so he completely omitted the reasons why he had troubles getting the thelio-io driver upstreamed to the hwmon interface. It's the same scenario that Asahi Lina faced with the drm maintainer demanding that she write her driver the same as amdgpu, which is a completely different hardware design.
@@mmstick Since I've got your attention. Excited for Cosmic (on nixos). Thanks for all the awesome work you guys are doing. I'll wait until beta before I start playing with it though.
Maybe system76 will fork the Linux kernel and rewrite it all in Rust. Lol This guy needs to calm down. Linus Torvalds is open to rust in the Linux kernel and this guy is fuming Rust isn't being adopted fast enough. Is this guy a Rust fundamentalist or something?
Well. They replaced Gnome DE with Cosmic DE which is written from ground up in Rust. RedoxOS is fully written in Rust including kernel. Funny thing is that as a non Linux OS that RedoxOS is also going to use Cosmic DE it seems.
Exactly...there's a lot of infighting with tech programmers - you can read about Linus blasting ppl at times - and they infight in forums or their comm. channels - but, a separate distro programmer thinking he can insult ppl and complain out loud in public to various programmers of the kernel is wild. Really illogical and crazy, imho.
Pedantic absolutely. Arguably that's how it should be. Patronizing definitely. Most of the time this is "don't be an idiot and we won't have to patronize you" which I'm 100% okay with. Megalomaniacs? Probably also true. I do get the impression that some of these kernel devs need to check their egos. Rust vs C kernel developers... well that's a complicated mess. The rust people probably shouldn't be making presentations and demands about how the C people need to reimplement things just so Rust can use it better. Maybe there's other good reasons, start with those. Or do the work yourself. On the flipside, the "Rust religion" comment was unnecessary. If you disagree on technical grounds, bring those up. Don't just say "no because I don't like you." Basically there's a serious culture conflict here. System76 and Jeremy Soller have had a pretty public spat with the GNOME project as well. You could interpret these cases as System76 having the balls to call out people who are hard to work with. And that does seem to be the case to me. But you could pretty easily interpret these as System76 being a crybaby who will publicly throw a fit when they don't get their way. Where does all this lead? Who knows? Find out next time on LINUX DRAMA Z!
Whether or not you agree with the statement or not, if someone owns a project it's his good right to be a "patronizing pedantic megalomaniac". And coming from PopOS something makes me think they're saying it for all the wrong reasons. Rust devs tend to be a little insane.
surely if your contribution is "do this in rust for me" or "explain how this works so I can do this in rust, just keep explaining because I don't know what this does" then... no, go back to your own hobby os. you're not helping.
So Rust is basically the compiled equivalent of Lua then? Hot garbage syntax + useless error messages + fanatical cult trying to force it into everything? Or is it more the programming equivalent of Scientology? Programming is one of the few fields in which pedantry (defined as "an obsessive concern with details and rules") is a good thing, honestly. Maybe if more developers were "pedantic," software as a whole would be better. Someone needs to sit Jeremy Soller down and tell him to quit sh!tting where he sleeps before he lands himself in the gutter.
To continue the pepsi analogy. It sounds to me more like the lead beverage guy wants to use nitros instead of co2 and is now butthurt because people dont want to use nitros.
Linus Torvalds is pushing for Rust. See his remarks from KubeCon this month. He is disappointed that Rust adoption isn't going faster, and added that basic memory safety issues are still affecting the kernel today. This whole ordeal is about a kernel maintainer that refuses to fix memory safety issues in their API.
If I was a kernel maintainer, I too would be sparing in my assistance to Rust conversion purely for the sake of being trendy. I can't stand the Rust community, obnoxious overtly political kids with inadequate life experience trying to shove a fad language down everyones throat. Nobody wants to put in the pointless effort to convert the kernel for style points, just to have Rust go out of fashion in 5-10 years and be left holding the bag. Pearl Ruby Ada Cobol etc etc... all came and went and the resulting lack of popularity meant a lack of volunteer support, so all the early adopter projects eventually had to re-adopt a more long-term stable language and reconvert all that code or fade into oblivian.
Yap, It's all come down to this. No one want to maintain other people's legacy code, especially the new toy that has no prove yet that it easier to maintain in the long run and still get the jobs done.
@@chrisnelson414 Ada was designed to be solid, but my point was more about a small talent pool in the long term. Eg Cobol is very stable and some companies still have cobol running on an old mainframe, but they are paying almost double normal industry wages to find the rare cobol maintainer. This is the part that doesn't work well for FOSS projects.
They develop hardware. If a bug is found with the hardware they build the best solution is to submit a bug report along with the solution to the Linux kernel... And the lead engineer seems ideologically opposed to that.
I’ve been on linux mint for a month and i like it, tho there is a ton of compatibility problems with proton for specifically 90’s and early 2000’s games and those games work on windows. I have decided that i will switch back to windows 10 and use the lsc edition that comes with next to no bloat and will be supported until 2032. Once support ends on it, i VOW to use linux only. Once i setup linux by then, I’ll passthrough a gpu to kvm for a virtual windows 10. Theres just way too many bugs and glitches and compatibility issues right now with linux. It is getting there but not fast enough for me. I also don’t like the way linux does things like it not being backwards compatible with it’s own games and programs if not maintained and having to paste in a random line of commands from some random on the internet in the terminal just to install the specific program you want if it’s not in the package manager.
that's because system76 did not write the code for Thinkpad harware. I have a Dart Pro laptop from System76 and their own hardware support is very average.
@@levaniandgiorgi2358 thank God. if only the rust devs were bullied as children instead and they might have grown up to be productive, functional individuals instead of a bunch of degenerates.
If I wrote a piece of software in lets say assembly language, the best thing to do would be to ensure that people do not find it useful. This way the recreational aspect will be preserved.
I don't remember C, or C++, being a religion, or having fanatical people, although I believe fanaticism exists everywhere. But with Python and Rust, it is common to encounter people defending their languages as if there was no tomorrow.
@@samuellourenco1050 cpp is a cult who worships the standard library and their lesser deities RAII and OOP, but they aren't particularly active evangelists.
kernel developers have work to do. porting isnt a priority when the current language is good enough (like in real jobs). devs whine when they dont get resources to do it. Its call open source. you can do it yourself. communists want to make people work on their projects. lol.
It seems to originate on the competition which programming language is worse - C or Rust. The Linux kernel is made by C programmers (and they want to continue to do it in C), while the Rust programmers want a kernel to be in Rust instead - right? The question which is superior and which is inferior, that kind of thing is damaging the teamwork in the development team. And since Linux is a project relying on teamwork, this won't do any good for the entire project. It's also not a good idea to start some 'keyboard wars' blaming other people that they're not doing what you want them to do.
Pop!_OS has to be one of the best operating systems with one of the worst names for software. I really don't think they had a single marketing expert or consultant in the room when they named it. Everyone put names in a hate, shook it up, and picked it at random until they got something that wasn't a joke answer. I think it would have made sense for a hardware company called "System76" to name a software platform they're building to deliver the best possible experience on their hardware something other than "Pop" OS. What the hell? Why not name it "Platform76"?
This same developer Jeremy who is obverly obbssesed with Rust is the reason System76 decided to stop development pop os in favor of Cosmic desktop environment just because he had an argument with Gnome team. This caused Pop Os to have become very outdated and Pop OS 24 LTS is still very buggy.
The sad part is that, as inadvisable as it may be for them to say so, they are right! Have you seen "Filesystem in Rust"? The whole seminar got derailed with relentless bikeshedding the moment a single piece of EXAMPLE code was shown. What does that tell you about the environment. What do you think it is like to contribute in that environment.
I'm not absolutely sure, but from what I could find, it seems Filho was merely asking what the semantics of an existing interface are, not to have an interface be created for his use. Sure, explaining something takes time, but that explanation could be documented and would be useful for other users of that interface, regardless of language. I read some speculation that the precise objection to Rust code was that any change from the C side that breaks Rust code would require fixing the Rust code, which means maintainers would have to learn Rust. I thought C code changes breaking Rust code was considered acceptable and a Rust developer would take care of it.
Linux kernel is not just for desktop, but also for embedded devices and other weird systems (e.g HPC) that don't have Rust support/ecosystem (e.g CUDA); also for low level environment, you need to write unsafe code
They probably have other engineers working on the contributions - he just personally doesn't want to do it. In any case, I'd say he's probably just an entitled jerk.. "Pedantic" probably means he just hates the rules for contributors (which exist for very good reasons)..
He omitted the reason because it wouldn't support his message here. Essentially, the hwmon maintainer refused to accept a driver for the thelio-io board because it is of a unique hardware design that is not similar to existing hwmon devices. The hwmon maintainer told Jeremy that he doesn't care about the reason for why the hwmon driver is written the way it is, and he doesn't want Jeremy to explain it. Given that it is technically impossible to do what the hwmon maintainer demands, it could not be merged. The same happened to Asahi Lina. The DRM maintainer demands that the drm driver be implemented exactly the same as amdgpu, except that it is not possible because it's not an amdgpu chip.
The same dev denied part of 'even try to contribute to the Linux kernel anymore' on kernel, I hope you'll correct it :) And your 'in short' is wrong - its about bindings specifically :)
Linux is meant to be portable. The official Rust compiler, on the other hand, can only produce machine code to a very limited number of targets, and it doesn't even support some very common targets like pre-SSE2 x86 CPUs. Yet for some reason rustaceans just don't want to fix their compiler, but they want everyone to use it. It just doesn't make sense.
No, it's that the kernel maintainer won't accept patches to fix memory safety problems in the C code that the Rust compiler caught while working on the Rust buildings for their interfaces.
I dunno, this seems like the kind of thing that would win you points with Linus lol. Also, I think you might be reading too much into this. The lead might not want to contribute to the Kernel anymore, but I'm sure delegation is happening. It's not great PR, but I really don't think the Kernel team is going to burn bridges with them. There is just going to be a "wow" and then move on.
This is some nonsense. If you want to rewrite storage in rust just do it *then* come back when its done. If its better its easy to adopt, if its not better then only your time is lost.
There is some intense hatred that was shown during that Linux Filesystem in Rust presentation. It wasnt the Q and A portion - basically it was 8 minutes in and the torches came out and derailed the whole presentation, based on a slide about modeling a rust type to help programmers understand what can be done out of an opetation. It was baffling to see as an outsider, I am assuming this was some straw to break the C programmer camels back or something
Good, Linux is and should remain a C project, it has been for over 32 years quite successfully, now these Rust folks come along and try to make the project bend to their will to accommodate their pet language? THAT is real "non-technical nonsense". I don't see Ruby devs trying to get the kernel to support their language. When they pushed for the Rust inclusion in the kernel they met with obvious resistance, but argued it was only optional, it should not be a big deal, now they are trying to make longtime devs accommodate the needs and goals of their language. Give these people a hand and they will take your arm, and then your head.
They are not wrong that a lot of linux developers are pretty conservative and even retrograde. But why do they bother to push theirs ideas down the throat of someone? One can always fork any FOSS project and do with it anything one wants. Same as they did with GNOME creating COSMIC. That's the beauty of FOSS.
I personally love rust.
But there is some weird agenda and weird ideological people connected to it.
The rust discord was pure toxicity since forever.
Nobody cares 😂
Most of the monads are contagious: async/await, exception throwing, nullable, ownerships, RAII and borrow checkers.
Because of the toxic Rust Foundation I prefer no Rust in the Kernel
do you know any Rust-Free kernel for Linux that are actually good?
The irony of this absurd statement... A language is not judged like that. Especially since there are multiple Rust communities and not only one obviously. Nor is C made of angels either. I have also different even political views than the Rust foundation. But Rust fixes a lot of issues that neither C nor its perfect community is fixing. You can wait and dream with the bold claim like I was told that somehow C got abandoned because it was easier that way. No, son. Until your fantastic C community did not deliver an actually better language then I am afraid the Rust community is going to be the least of your problems.
Even Windows is already getting oxidized... Resistance is futile. Big corporations essentially already pushing Rust whether it actually is a good language or not. Though it is not that hard to do better than C... Lol
@@techpriest4787Do not know about you but I got zigidized and not oxidized 😉
More specifically a "memory safe" kernel would only be suitable for virtualization because it would not have multiple fail safes to handle actual hardware. Essentially its Communism trying to reinvent the currency.
This is comically irrational. The language itself is pretty great, regardless of the toxicity of the leaders of the Rust Foundation.
And the very things that make Rust better than C for some tasks IMPEDE its use in a kernel.
The Linux kernel has been around for at least 33 years. Any major bugs that Rust's trappings can prevent are long-gone...and more importantly, the finer control over things like memory means that the kernel is tuned to an extent that Rust would never allow.
Even when it comes to new components, why not write them in the fastest, most efficient language, and simply take the extra care to not write in the kind of bugs that Rust is supposed to prevent?
This is a question of development speed versus efficiency and stability. In the case of the kernel, efficiency is THE priority.
Which means pedantry is what keeps the Linux kernel clean and safe.
@@justanothercomment416 If Rust developers understood kernel development, it wouldn't have to be explained to them.
This sounds plausible, but not realistic. When you talk about new stuff, the amount of device drivers added to linux kernel is increasing by the day and not every driver developer is Greg Kroah-Hartman or Takashi Iwai level developers that don't make silly mistakes at all. They turn out to be random dudes working at random hardware manufacturers. Rust solves majority of the memory bugs they might introduce to the kernel and speeds up the process by a big margin.
@@freedoompictures6839 They also produce slower, clumsier, and in the long run less stable drivers.
Better to do a little actual work and get it right.
It's a growing problem, devs are working on the linux kernel to fix issues that affect them, but they're not pushing their patches upstream because dealing with the politics of the upstream project is a waste of time.
People confuse their own abilities with ownership of a project. Most do not realise their changes can have a profound impact, many others simply refuse to go along with the rules. Most people able to do kernel development are lone sigma wolves, very good at doing their own thing (they are usually brilliant) but terrible at collaborating.
@@sillonbono3196 Bruh, most people in open-source are terrible at collaborating. The fact that they think "this software is free to use, modify, and distribute" also means "this software and its developers are free from criticism" (and make no effort to act as though they don't) proves as much. And saying someone's "good at doing their own thing" is like saying they're "good at breathing."
@@sillonbono3196 Kernel quality code rules are there to ensure.. quality. strange right?
@@AClockworkHellcat "The fact that they think "this software is free to use, modify, and distribute" also means "this software and its developers are free from criticism"
Yes, it's the entire point of OpenSource software.
Hackers aren't exactly social butterflies.
It's never going to be smooth sailing.
@@sillonbono3196 I was with you until you made the false claim that kernel developers are lone wolves. If they weren't good at collaborating, then the kernel wouldn't be what it is today. If they didn't abide by the rules, then Linus would've smacked them down, and he has done so quite a few times. The thing to remember is that the core of the kernel is basically from fewer than 100 devs. Those are the ones that really count, not the thousands who develop single purpose drivers for a weird piece of hardware they bought at a flea market 10 years ago.
Finally clicked for me: It's another power grab:
>Take over a shiny new language (Rust)
>Make sure there are mainly comrades in Rust Leadership
>Make banning people for political opinions outside the norm
>Replace C with Rust
If this works out, Linux is dead because nobody with any technical skill is left. But I think the tide is turning.
To add,
The entire "safety/security" argument doesn't work when what you have is a software that has been out in production for quite some time, and has most (lim -> inf) of its bugs and security issues ironed out.
Slipping a non-bootsrap-able alpha-stage non-standard toolchain into a codebase as big, important and tested as linux kernel smelled like rotting fish from the very beginning.
Yup!
Rust is a mistake.
Please do not politicise programming languages. And, yes, I believe - all other thins being equal - that Rust can produce more memory safe code, but has downsides as well (many high abstractions) and unsafe Rust is in some ways worse than C. C developers not helping Rust developers is natural.
Hey, that's straight outta "The Protocols of Learned Developers of Rust"!
I am with the C guys here... Rust ppl are like a weird woke group of weird cult... Not a "woke cult", but mostly woke furries who are also cultist about rewriting everything in rust - awful combo!
Just look at Pop OS Cosmic. The main dev is a mentally ill man that think is a woman and wears animal ears during conferences and post trans flag with trans right message in his presentation. While Pop OS spokesmen refer to him as "her" "she".
People who have worked on the kernel in C for years obviously won't magically hop onto rust completely destroying their workflow and productivity. Learning a new language can take years to become proficient in not to mention that the person need to want to learn it which they obviously don't have to. It's funny how Rust support was added because people wanted it but where are the Rust developers now? I don't see majority of submitted code being Rust how the loud people pushing for support made it seem like. "Add support for Rust and you'll get more young talent contributing to the kernel" yeah that didn't happen
Rust is starting to sound like Python that will infect your project with nonsense and drama the moment either of them get used.
starting to?
Eh, I don't think Python _use_ has much drama anymore. It's a great glue language that everyone technical can understand and its libraries are usually in the top 5 for anything so a great starting point before investing the time and money to develop something more efficient.
It infects, but not with drama, but with abstractions C developers don't like. Better to start over (Redox OS).
Python is pretty much on its apex. It will decline.
What abstractions would those be?
Honestly if I hear something is written in Rust, I take it as a red flag and simply move on. It's a fully pozzed ecosystem now, and at this point people seem to use Rust primarily as a political statement with technical suitability taking a backseat to ideological uniformity/gatekeeping.
It was always a political statement, that's why Rust had people filing GitHub issues on random projects saying "rewrite it in Rust or you probably want your software to be insecure you prick", and they wouldn't take no for an answer.
which sucks because Rust has reasonable technical merit and the reasons to not deal with the "Rust people" overshadowing that is a shame.
@@sunderkeenin I disagree regarding technical merit. There's technical merit to learning how to properly use C++ and the various *_ptr types along with when to use move and forward and r-value and l-value references and their semantics, especially if you already know C++ and/or C. Moving to a completely new language because it makes false promises and looks like Perl barf is just a massive waste of time.
Relax. Rust is actually quite suitable for writing small CLI tools. Consider it a very modern C++ in this case. Lifetimes are very simple in CLI tools, so the type system aren't that difficult to deal with. Not much ideology involved here.
@@generic694 You just described a half dozen languages, my guy, but I didn't ask for use cases-use cases exist for _every_ language. How many languages have communities that actively target the most engaged devs in a naked power grab to gain control over the project so that they can start revising codes of conduct, banning the most prolific developers & refactoring the codebase to be more "inclusive" in a blatant attempt to gatekeep influential projects from being retaken by dissident contributors?
We _both_ know this is ideology-driven behavior; it might as well have a fingerprint at this point. Oh wait-it does! Google around a little & see if you can guess what Saul Alinsky might say about this interesting series of developments. :^)
Not that my two cents mean anything, but, always one of those. If you want to port something, in this case rust, by all means have fun with that. But don't expect others to bend backwards for your "cause".
I don't think he was suggesting anyone needs to bend backwards. It seems there are some Linux developers who have a cult like backing of C and C++ who do not want to support rust for no good technical reason. It's political sabotage.
I don't care either way, I'm not a Linux or Rust dev but that's what it appears to be from the outside.
@@ImperiumLibertas Rust isn't a certainty. Having to support it and replace with it what has always worked and every veteran developer is an expert in and agrees is still the best option puts the core project and everything that depends on it at severe risk. Linux is the core of a lot of mission critical infrastructure across the planet. Forcing Rust into mainline and all veteran devs to support it at this point is extremely irresponsible. That is what source branches are for. Leave everything not rock solid and proved outside of the trunk and main release.
@@ImperiumLibertas Its no that simple to "not want to support rust", you like rust and want to port some code? good luck but its not the job of a 20+expertise of C developer to help you.
@@nempk1817 the person from the talk didn't in fact say anything about that c programmers must rewrite their code in rust.
Rust has a ways to express some conditions in it's type system so they needed help with understanding that these conditions are to write safer code. I think it's a really big misunderstanding from the side of c developers, because rfl just needed information about what contracts were upheld in file systems code and other implicit behaviors.
@@ImperiumLibertas
cult-like backer of C here. people have a cult-like backing of C because it has worked properly, without standing in the way of the developer, since 1972.
a fun new toy is just that, a fun new toy. once you're done playing with it, you'll move on to the next. rust is neither cool nor particularly needed in the linux kernel.
Definitely has the tone of a new python dev getting mad that their special-use code isn't being put into a very technical portion of code that runs in hard realtime and has its own pseudolanguage/involved semaphore system to keep everything from bumping into itself.
I guess PopOS needs some more "diversity and inclusion" very very soon 😂
Or just hire you
@@neilpatrickhairless That'd be horrendous since I've left programming a long time ago. How about you? Want the job?
Dont even joke. Open source has suffered enough
They need JC Denton
DEI is a crime leading to a multitude of diverse crimes disguised to compensate for highly responsible sane practices painted as crimes.
Rust is just bloated. It takes a whole day to compile its compiler, leaving 40gb of junk files and refreshing the cargo database downloads an absurd amount of data.
Yep. Even a trivial rust application takes too much memory and storage to compile. I quit my previous job because it was making things impossible and I got fed up with the rest of the team not acknowledging the problem.
You must be on a worse computer than my potato. It only takes me around three or four hours to compile each new version. Of course, I delete the whole archive each time so I don't have to put up with the bloat hanging about, and it took me 10 minutes to write this comment while I was waiting on `du` to show me the current usage. For the source folder where I've got the `git` repo, it's "only" 19GB for the version I'm currently on, 1.77.0 compiled on 3-17, and the binaries take up 3.2GB. I don't remember what GCC uses up, but since it handles more than one language it's not so easy to quantify, but it still uses up less space than that and compiles significantly faster. I don't do incremental builds like some developers, so compile time matters more to me because things aren't generally cached.
@@anon_y_mousse I mean, I am not a web developer, but I think the way the cargo database works is stupid. It is downloading the whole database (around 200MB as of now), and searching it locally. And of course updating every time it can. Why don't they make a serverside PHP application to query the DB without downloading these 200MB of garbage??
@@gargamel3478 I assume they learned the wrong lesson from how `git` works. Of course, even if that's the case, the database is too large and there's undoubtedly far too much duplication of effort. They should be more discriminating with what goes into it and prune stuff as needed. Here they could learn something from `git` such as branching, but maybe with the branches not all being necessary locally to have a complete picture of the most recent view of the database.
Doesn't the deep tech state aim to create an AI system to write programs & operating systems that are in C into Rust? That said if were Ms. Beep state, I'd hire human coders to do much of the primary conversion of code.
Linus should never have brought this Rust nonsense in the kernel.
Rust is ok. The people doing stupid things.
@@Kirfx That was Linus's issue with C++: the programmers.
@@Kirfx rust is ok, but too big of a leap for C devs to get into, and thus limits the changes C devs can make to the kernel, including C code that is used by Rust code
Keep crying boomer 😂
I'm simple
> They hate C
I love C
> They love Rust
I hate Rust
> They hate the leads of NixOS
I will support the leads of NixOS
> They hate Linux
I love Linux
There is a clear pattern there
Mediocre and scared people hates talented and passionate people
... right about everything except NixOS ..
I thought the leads of NixOS dislike anyone to the right of Mao.
nixos is overthrown by people like jeremy, wtf are you talking about?
they banned all of the so called passionate people because they were not woke enough, dead project by now
The actual maintainers are pretty okay on Nixos. Usually not too political outside of a few bad apples. The Nixos moderation and governance teams on the other hand are out right leftists who slap their friends on the wrist when they break the rules and outright ban conservatives for simply having conversations that other people get upset about.
it's not too late to re-make up your mind about nixos.
So, for the patronizing narcissists, anyone who tells them “no, not that way” is a “patronizing pedantic megalomaniac”. It’s no surprise that Rust, a language invented to save mediocre developers from themselves, is their primary weapon of choice.
Edit: typo
The funniest thing about Rust, is that it's showcase project, for which it was originally made, (the new web engine for mozilla) have never materialized 10 years later. ;) Yet they try to push it everywhere else. Like its a huge success story.
"don't contribute to" =/= "don't work on"
There is a difference, no?
That's what I was thinking. The proof would be, does Pop!_OS work, on System 76 hardware?
If it does, he must be working inside the kernel!
Lunduke loves shoving politics/culture war where it isn’t for clicks, ironically instigating a lot more political activism than he ‘stops’ 😂
yeah this is what i wanted to say, popOS and system76 can still get by without being associated with contributing to the linux kernel.
system76 previously had a big linux presence where they were solving regressions upstream and sending patches and all that. them saying that they don't want to be involved speaks more than anything. lunduke is extrapolating the comments made by system76 in ways that the author didn't mean and it shows.
edit: big fan of lunduke before, got into his work when he was covering the mozilla situation and the nix os stuff, i thought he actually reported what was happening objectively. popOS can get by not agreeing with where the kernel is headed because lets be honest with ourselves, linux does suck. and while alot of people share this sentiment with various other tooling like systemd and linux itself it doesn't exactly mean that they should just not use linux, the alternatives out there like windows are far worse. you can apply this to the system76 folk too.
disappointed lunduke.
Nobody is going to recompile their Kernel so that their System76 laptop will work on another distro besides Pop!_OS (worst name ever). People will just stop buying their machines.
@@owlmostdead9492 system76 machines run on other distros out of the box, the hardware is generic. the patches they make are for optimizations. alot of sched optimizations on the linux scene come either from system76 or surprisingly the android custom rom scene.
Rust programmers: I'm asking for help to C programmers to migrate the code to Rust so you end without job and we get your positions
C programmers: It's not gonna happen easily.
An that type of conflict is as old as the first cutting stone industry.
Ah yes, Masonry.
Should be trivial for them to migrate to Rust. They should be able to read the C code easily enough. If they can't... maybe they're not the programmers we need writing kernel code.
>buhu woe is me people disagree with me
Rust is a literal cult by god, if c, c++, python or js devs were as extremely fragile as them the whole computing world would end.
Why do these people not just keep out of public conversations and remain within their bubbles like "the rust porting project"
Because they do want it to end, taking everything over and tanking it is the purpose of these operations.
Jordan Peterson is right. When people deny God, the first thing they do is to seek other gods and create cults. The cult of vax science, the cult of Python, the cult of Rust. All fanatical people.
7:00 Doesn't that proof Rust is a religion? No System76 for me or my clients for sure now
I'm a complete amateur with Rust and haven't even touched it in almost a year so my perspective may not be the best but while I really enjoyed simple rust, at a certain point it gets super complex. The thing is, to get over that hump, you kind of have to really live and breathe it which can lead to a little bit of a myopic view of things.
Something like Zig I feel like is a more likely inheritor for the Linux kernel guys.
zig is objectively the superior systems programming language, but the kernel should just remain C.
The guy's got a point; those megalomaniacs should give him all the power he demands.
I like Rust. I'm learning it, myself, though I've known C/C++ for decades, and modded my own Linux kernel 30+ years ago.
BUT:
Complaining that the Linux kernel devs won't either rewrite the kernel in Rust or spend time supporting others who do sounds like a modern Participation Trophy sort of sense of whining entitlement.
If you want to make a port, do it. But you aren't owed the time and effort of others who aren't interested.
That was not what was requested or said.
If a Developer is Not Pedantic the developer is not a developer!!!
It seems to be optional. Literally: _"-Wpedantic"_
@@tinkerwithstuff :)
There's a reason for it, its called "if aint broke, dont fix it" everyone's fear is that because the whippersnappers don't know what they are doing, we gonna end up with another systemd or pulseaudio, but in the kernel this time.
Has everyone forgotten that one of the 'goals' of rust is to get rid of the idea of 'coding skills, and quality?' They state it plainly in their c.o.c.
Their code of conduct is a woke joke:
"We will exclude you from interaction if you insult, demean or harass anyone. That is not welcome behavior. We interpret the term “harassment” as including the definition in the Citizen Code of Conduct; if you have any lack of clarity about what might be included in that concept, please read their definition. In particular, we don’t tolerate behavior that excludes people in socially marginalized groups."
Who's more socially marginalized than C devs? /s
But seriously, it's like they don't know how the internet or digital technology came into being.
It wasn't _that_ long ago that IBM was selling computers to the National Socialist Workers Party of Germany. Should we cancel IBM?
It wasn't _that_ long ago that Wernher von Braun helped NASA with their rocket programs. Should we cancel NASA?
We're making sausage here - not picking and choosing political ideologies.
Do the personal opinions of the charcutier matter when we're purely talking about how to make the best sausage?
We as humans have the opportunity to learn from the underlying process, and truth, and reality - regardless the messenger.
Here in the United States, it _used_ to be the case we have opinions as individuals, and agree or disagree with the opinions of others, without our access to income or livelihood being revoked. Words are not violence. Violence is violence. if my use of the English lexicon bothers someone, that's a _them_ problem, and 'they' can kindly go f*ck themselves. Is that how 'polite' society treat one another? No - but that's politics. This is sausage.
@@lyleesjones Well, National Socialists are currentlty the most socially marginalized group. And I wonder if they will accept one in Rust Community
@ActionGamerAaron Exactly my words. How concise of you to reproduce my thoughts in 87 characters.
People need to wake up to the fact that all these horrible people, who think they're a moral authority punishing throught-crime, have been raised years ago and will have children, who will be even worse than them. Your very own children are raised into this as well, in kindergartens, schools, universities.
I was and am still a huge fan of System76, and I thought Jeremy Soller to be some kind of programming god. But then I made the mistake of looking at his twitter. I saw him retweeting stuff that included there is a "literal trans genocide". I had to shudder and gtfo twitter. Jeremy, please keep making awesome software. I hope you don't think there is a "trans genocide", but yeah. Please keep making good software. But thankfully, I can rest easy knowing that Linux will move forward with or without System76 and with or without Rust evangelists who got twisted up with woke ideology and ways of behaving online...
You're so right.... it's really troubling, that these devs are so political now - I never realized how much time they spend yapping about their left-wing politics. At least, you could switch to Ubuntu if Pop OS starts putting their woke ideology into their OS - and it's possible - for e.g. firing ppl who aren't woke enough for them....stuff like that would turn me off, anyway.
Wish there was one.
I think the core of the issue is that C, for all of it's shortcomings, is still one of the most highly flexible languages around. C programmers like it because if the problem you are trying to solve changes, C lets you squish it and pull it to fit the problem with relatively little resistance. The more often the problem changes and you need to adjust the more that strength of C shines through. Rust on the other hand is noticeably more inflexible. Rust data structures are more resistive to change, it can slow development down or worst case make certain features prohibitively expensive to implement. I think this is intuitively apparent, but people will differ on how much they value "Flexibility" VS "Mathematical Provability".
Personally, I think we need a C replacement, but Rust is not it. Getting rid of undefined behavior would be a good start. It would be nice to have well defined default behavior regardless of implementation trade-offs, then have instructions that are hardware specific that make the most optimum trade-off for that hardware. This way we reduce the number of surprises and foot guns.
Rust isn't any less flexible than it needs to be. When you really think through the hidden assumptions in C semantics and attempt to make the implicit stuff explicit, and really make an effort to prevent the most common classes of memory bugs, you end up with something like Rust semantics. Read Aria Beingessner's blog for examples; she, like many Rust devs, actually knows C better than most C devs do. "Squishing and pulling" the code to fit the problem is a lazy, cowboy-coder approach; what you need to do is think deeply about the problem and analyze all of the desired semantics and failure cases and be explicit about those in the code, encoding as much of it in the type system as you can so that errors are caught at compile time rather than throwing exceptions (or being silently ignored, leading to bugs and security issues) at run time. Rust has a type system robust enough to do this; C does not. Rust wins.
@@bitwizeafter seeing a women's name and Rust mentioned, my first instinct was "that's not actually a woman" so I checked and it was correct lmao. The Rust community is infested with these mentally unstable individuals, the same ones that ruin everything else open source they get their hands on.
@@bitwize
brother, you haven't programmed a single day in your life.
"any less flexible than it needs to be?" c, instead, is not "flexible". c is literally 99.1% pure assembly code in english. it's not flexible, it's literally mapped as close as humanly possible to 1:1 to the way the computer needs its instructions fed into it. and you can do one better: to use a subset of ansi c is to get even closer to assembly.
hidden c assumptions? name and explain one yourself, don't give the promise that some random mentally ill individual did so on the internet.
"\"Squishing and pulling\" the code to fit the problem is a lazy, cowboy-coder approach" brother, pick one! these are complete opposites! you're either LAZY, i.e. you're bored and maybe lack the technical know-how so you use c alternatives (shame on you), or the balls on you are so LARGE that you're embarking on an adventure to solve the given issue with some sort of optimization-minded competency.
"what you need to do is think deeply about the problem and analyze all of the desired semantics and failure cases and be explicit about those in the code, encoding as much of it in the type system as you can so that errors are caught at compile time" people don't like using paper these days to solve problems. and it's a damn shame. people think that rust is gonna code for you as well! if you don't understand the problem, _who the heck does?_ don't shoot first, think later. THINK FIRST, then SHOOT!
as for the final statement, rust is a tool for tools -- promising safety to people who simply do not know any better. c has been working properly and as intended since *1972.* c wins.
@@freedomgoddess Actually look up the term "cowboy coder". It's not a compliment.
Over 50 years and countless CVEs suggest that even them most experienced, conscientious C programmers are prone to falling into UB and memory-usage traps. It's simply not humanly possible to be as careful as it takes to reliably produce C code that's free of memory issues. It's prohibitively difficult to even determine if such issues exist in the first place. Rust doesn't write your code for you. It just helps you write more correct code by providing static, compile-time verification that it does not contain common classes of memory issues.
Again, the Rust programmers know C better than most C programmers do -- which is why they use Rust.
@@bitwize "is a lazy, cowboy-coder approach"
This reminds me of a story my uncle told me. (Sorry about the wall of text, but it is a long story.)
My uncle worked at a factory that made parts for machines. They did plastic injection, they had mills and laths, ect. And they had contracts to produce thousands of a part per day. Anyway one day there was a change in management. The son of the CEO was put in charge of managing this particular factory. The new head manager was fresh out of college, he knew more about running a factory than my uncle or any of the other machinists.
The first thing the new manager did was to change the workflow of the machine operators with the goal of making things "more scientific". You see, mills and laths have lots knobs and dials that control the settings of the machines, and when the machine operators were making parts they were just guessing what the settings should be until the parts came out ok. This meant that each operator was using different settings. One guy might have the knob set at 35, another at 36, and another at 33. And they are all making the same part at the same dimensions. The change that was made was to standardize and require each machine operator to set their knobs to the same exact values.
My uncle of coarse disobeyed this order, but many novice operators went along with the new mandates. That's not to say my uncle couldn't read the room, if certain people asked him, of coarse he had his knobs set "correctly".
What happened after this change was no surprise. Hundreds of useless parts came out of the machines, off by just a fraction of an inch but not within acceptable tolerances. This happened for days. Operators where not meeting their quotas and they were afraid they would lose their jobs. They would go to my uncle and ask "help me, please, how do you get your parts to come out right?" The answer was simple, do what was done before, put in the initial settings, measure the outputted part and adjust the knob, repeat until it is spot on, then periodically after every few hundred parts measure again and readjust if necessary. Soon the factory was back to making parts.
You see, no two machines that manufacture parts are identical, even if they are the same brand. Some machines are more worn than others. Vibration differences because of proximity to other machines. Some are sitting on the ground slightly less level than others. The tools that cut the metal go from sharp when new, to dull and need to be replaced. Temperature of the gears in the machine as it is running. All of these things, many of which you have no control over directly, contribute to making the behavior of the machines different enough that it matters when attempting to produce precision parts.
It is still unclear if management has learned from its mistake, or if they even know that a large amount of their workers are ignoring them. It's not a great situation to build a work culture where information cannot not flow back to the people in charge. I think management probably does know "more" about my uncle's job, but how much you know is sometimes less important than what you know. (He does not work there anymore, the factory has since fired nearly everyone and is flying non US citizens in. I mean if you don't want to listen to people it doesn't matter if they don't speak english. LUL)
They're still working on the linux kernel, they're just not pushing their changes upstream due to the problems stated. I hope that clarifies. It doesn't really affect their business model.
Doesn't that violate the license?
@@MechaFenris No. As long as their changes are public, linux can merge the patches themselves without wasting everyone else's time with their pedantic nonsense.
@@MechaFenris it doesn't, their patches are public, and anyone can apply it if it benefits them. and even if they are closed off it doesn't violate the gpl because there's no source of linux associated with a .patch file
i don't know anything about kernel programming, c, and rust. but if some people join on my project then expect me to rewrite my code on another language or help them when it was not needed then I'll be angry too 😅. I don't understand the situation, you want rust? just do it, why expect anyone to do something what 'you' want to do..
PopBSD when?
Probably never, since the last time I checked the various BSDs had said no to Rust and its weird cult.
Never. BSD's are for real cool old-school gays like Kirk McKusick and Eric Allman, not for these modern hyper-sensitive furry transsomethings.
In a recent video did the research not show C was more energy and memory efficient than Rust? Seems like a step backwards to me but, I am not a programmer.
I know next to nothing about programming languages, but my understanding is that Rust can prevent certain exploits, specifically it enforces memory safety.
Optimization is not always a complete win-win of "faster and use less memory", sometimes you need to consider different aspects (like memory safety in case of Rust) or more complex factors that could be a bigger priority. Or in short: no tool is perfect, programming languages included.
it is very much accurate to write code in c rather than rust for energy efficiency reason. it's just factual that writing c for such a case is better. it's just that rust is injected by trigger happy people where there is c already so it's hard to make any sort of jump to it with such little interest to it by oldheads.
Okay, but the question that I would have is: "Is he wrong?"
Imo yes. Sounds like a diva that didn't get her sparkling water just right.
The resistance sounds reasonable. C devs wouldn't, shouldn't stop their work to accommodate and assist their supposed replacements, especially when the replacement can't even help themselves and only offer negligible to nonexistent benefits to the kernel.
The incumbents joined the linux club to develop, not translate.
What these old C developers in the Linux kernel community fail to understand…
Mommy’s little girl has a right to be included in the kernel community, because that’s what she wants
Mommy’s little girl doesn’t want to use C, she wants to use Rust - so the Linux kernel community needs to accommodate her wishes
Mommy’s little girl has a right to be safe from opposing points of view
Mommy’s little girl deserves equal outcomes compared to other kernel contributors, so all her PRs should be merged just like everyone else’s
The Linux kernel developers need to share the rewards around in a fair and equal manner
Soon it will be time to hand over the reigns, and give Mommy’s little sweetie a go at running the whole project
When that happens, be sure that Mommy will be there to help as well
Does mommy have blue hair?
@@samuellourenco1050 By that time, yeah, probably.
@@samuellourenco1050 the mommy probably has a Y chromosone too
I have to imagine you're some greasy 60-year-old dude writing comments like this. Just pure fucking cringe. (Rust sucks tho)
Classic NPC copypaste 😂
@Bryan Lunduke if you watch whole video, not the small part, it changes perspective. The guy from video was trying to do presentation and claims "I don't want you to write rust code, just give me semantic of the C code" then he received small backlash. Jeremi is not contributing to mainline kernel, pop_os has backports.
Anyway I don't care. In IT industry there are a lot of holy wars like vim vs emac etc.
The maintainer of Pop!_OS doesn't have to submit patches upstream. Their focus is Pop!_OS working on System76 machines. This might impede other distributions from working out of the box, but that's not their problem.
Their kernel patches are likely available for other distributions to apply, it's not uncommon for distributions to have their own downstream patches. Running down System76 is inappropriate when there is an issue with upstream development.
@mrfuzzy2954
This, right here. ^
Unless the terms of his employment actually require upstream patches, he is under no _obligation_ to work with the kernel devs on enacting them.
So, no, it's _not_ like a trucker who refuses to deliver to certain states because they don't like the laws.
Tbh Rust is already problematic because of it's weird foundation and it's own megalomania, anyone dependent on it on a project as critical as the Kernel is asking for trouble, even past technical concerns about dependency management that saw Debian devs abandon it, the foundation is going to be a headache for years to come and i prefer the memory issues that can be resolved with a commit to legal issues or (additional) temper tantrums involving the foundation of Rust making demands to remove Torvalds or something because they don't like him...
Rust is toxic
Demanding to remove the creator???
@@42Cosmic42YourMama is toxic 😂
Arch and Manjaro don't do this kinda stuff. I only use Arch so I can legally say "I use Arch btw" to randos at walmart or at your grandmother's funeral. Manjaro is still my favorite distro, it's the one I started on. I know this is a hot take, especially among Arch users, but I actually liked that they held back packages. The stuff about "dependency hell with AUR packages" is WAY overblown and I never once have had it happen to me. I still use Manjaro on like 3 laptops and I'm using the stable branch with TONS of AUR software. Particularly on my pentesting/hacking VMs that I use to practice and they are ALL super stable and never fail me at all. Gaming on Arch has been great with Steam, Lutris and WINE/Proton, but it was much easier to get things working when my desktop was running Manjaro.
My experience of the AUR is 95% of packages I want are there and I can just build it myself and it works. Only issues I have are with closer source stuff
Arch is trendy with the troons. Avoid, same as Rust.
@@tomthegeek let me guess. Debian?
@@hardrocklobsterroll395 exactly. I used to use Ubuntu way back in the day, but I constantly found myself not booting into it at all. Then, I found Manjaro and I don’t have windows on my machines at all. The AUR made everything so easy, I can even use .deb packages with that conversion tool, but it’s rare and why I can’t remember the name of the tool 😂 Literally EVERYTHING is in the AUR. Debian based distros are clunky and you have to nuke it every time you want to truly update when a new one comes out. With Arch based, sudo pacman -Syu and everything is done.
@@hardrocklobsterroll395 pamac has been the best GUI package manager, by far, in all of Linux. You just can’t beat it. The one in Ubuntu is just clunky and looks really old.
Do System76 actually make new hardware? From a hardware perspective, aren't they just a system builder?
Well, that's pretty much most developers in open source. Have they not met Richard stahlmann?
Who's the german?
1. Keep in mind that this comes from a Rust developer.
2. Keep in mind that this comes from a Rust developer.
3. Keep in mind that this comes from a Rust developer.
4. Continue the evaluation of the statement by the Rust developer.
Every 5 years there's a lemming-like rush to learn and port the codebase to some faddish new language, which is considered to be the Greatest Thing Ever until the Next Thing. (Instead of maybe focusing on more elegant code and efficient PM processes?) Not saying progress is bad, but this isn't the path of progress.
funny how the engineer is behaving like the very thing he accuses others of XD
This entire video is bizarre.
Linux is open source, they can fork and not contribute upstream.
They are not required to especially if it's not welcome (guessing here based on the megalomaniac comment).
They can ship their product without giving anything back.
I really don't know what this video is going on about.
Yes, they can ship their product, but their product is influenced by upstream decisions upstream. So, any OS that relies on the Linux kernel has no choice, unless they fork a previous version of the kernel that was not rusted yet.
@@samuellourenco1050 sure I agree. But they aren't required to contribute back. That's my point.
And also misrepresents the issue quite a bit. I'm starting to see why people have issues with Lunduke. My understanding of this issue is that some of the C code in the kernel is undocumented and very hard to follow. The Rust developers need to interface with those parts of the code, but when reach out to the C developers/maintainers with questions, they get turned away. And because of the difficulty in working with them to document or atleast explain the interface to a crucial part of maybe the most important piece of software on the planet, they figure they rather not bother working with them and instead look to just re-write those systems in Rust. Naturally, this pisses off the C developers because they see it as Rust encroaching in their spaces.
I know both C and Rust, and while Rust definitely has it's pitfalls, it's design reduces the chances of many footguns going off with new contributors. And being such a popular language, it'll invite more and more contributors over time. Linus definitely sees it's potential and understands it's benefits. And until/unless a new language with all of Rust's benefits but none of it's failings show up, it is the way forward in the kernel.
I do like to listen to you, but I think you're over dramatising the situation here. People, I would assume and also meaning I could be completely wrong, purchase their machines because they want a working setup. System76 make it work and ship it, they don't need to push jack upstream.
Also for the Rust guys, if you want something done, do it yourself, stop whinging that other people won't help.
Lunduke's all about drama and sowing division it seems. But I guess it is a way to get income instead of actually contributing to the software.
@@mathiasfriman8927 Always has been
Jeremy is also not the only person with the ability to contribute patches to the kernel. Although with open firmware, it's not really that necessary anymore. Lunduke is out of his depth and doesn't comprehend what he's talking about. He assumes rather than investigates. He's taking everything out of context, so he completely omitted the reasons why he had troubles getting the thelio-io driver upstreamed to the hwmon interface. It's the same scenario that Asahi Lina faced with the drm maintainer demanding that she write her driver the same as amdgpu, which is a completely different hardware design.
@@mmstick Since I've got your attention. Excited for Cosmic (on nixos). Thanks for all the awesome work you guys are doing. I'll wait until beta before I start playing with it though.
Picking a fight with linux kernel people not a very good idea Ubuntu did that guess where it got them...
Maybe system76 will fork the Linux kernel and rewrite it all in Rust. Lol
This guy needs to calm down. Linus Torvalds is open to rust in the Linux kernel and this guy is fuming Rust isn't being adopted fast enough.
Is this guy a Rust fundamentalist or something?
Fundamentalists in open source? Naw, that never happens... 😂
Well. They replaced Gnome DE with Cosmic DE which is written from ground up in Rust. RedoxOS is fully written in Rust including kernel. Funny thing is that as a non Linux OS that RedoxOS is also going to use Cosmic DE it seems.
Fuming? Huh?
Rust in the kernel was a mistake.
Yup. An inch was given, an arm will be taken
Exactly.
thats what happens when you work with the %41 folk
Cry more boomer 😂
@@RustIsWinning What a lame attempt, damn
Jeremy is not acting as leading a company, but as a spoiled little kid. We need to act like grown ups for a change.
Exactly...there's a lot of infighting with tech programmers - you can read about Linus blasting ppl at times - and they infight in forums or their comm. channels - but, a separate distro programmer thinking he can insult ppl and complain out loud in public to various programmers of the kernel is wild. Really illogical and crazy, imho.
this is the old os drama I miss
If you want to code an OS in Rust, do it. Don't expect C coders to accommodate you. Grow up.
Pedantic absolutely. Arguably that's how it should be. Patronizing definitely. Most of the time this is "don't be an idiot and we won't have to patronize you" which I'm 100% okay with. Megalomaniacs? Probably also true. I do get the impression that some of these kernel devs need to check their egos.
Rust vs C kernel developers... well that's a complicated mess. The rust people probably shouldn't be making presentations and demands about how the C people need to reimplement things just so Rust can use it better. Maybe there's other good reasons, start with those. Or do the work yourself. On the flipside, the "Rust religion" comment was unnecessary. If you disagree on technical grounds, bring those up. Don't just say "no because I don't like you." Basically there's a serious culture conflict here.
System76 and Jeremy Soller have had a pretty public spat with the GNOME project as well. You could interpret these cases as System76 having the balls to call out people who are hard to work with. And that does seem to be the case to me. But you could pretty easily interpret these as System76 being a crybaby who will publicly throw a fit when they don't get their way.
Where does all this lead? Who knows? Find out next time on LINUX DRAMA Z!
But if they force developers of other projects to change stuff in to rust, that will stop development of kernel? Also will create loot of new bugs?
Whether or not you agree with the statement or not, if someone owns a project it's his good right to be a "patronizing pedantic megalomaniac". And coming from PopOS something makes me think they're saying it for all the wrong reasons. Rust devs tend to be a little insane.
surely if your contribution is "do this in rust for me" or "explain how this works so I can do this in rust, just keep explaining because I don't know what this does" then... no, go back to your own hobby os. you're not helping.
rust devs really are the vegans of linux
Not really --> 🦀
So Rust is basically the compiled equivalent of Lua then? Hot garbage syntax + useless error messages + fanatical cult trying to force it into everything? Or is it more the programming equivalent of Scientology?
Programming is one of the few fields in which pedantry (defined as "an obsessive concern with details and rules") is a good thing, honestly. Maybe if more developers were "pedantic," software as a whole would be better. Someone needs to sit Jeremy Soller down and tell him to quit sh!tting where he sleeps before he lands himself in the gutter.
To continue the pepsi analogy. It sounds to me more like the lead beverage guy wants to use nitros instead of co2 and is now butthurt because people dont want to use nitros.
why are they pushing rust?
Linus Torvalds is pushing for Rust. See his remarks from KubeCon this month. He is disappointed that Rust adoption isn't going faster, and added that basic memory safety issues are still affecting the kernel today. This whole ordeal is about a kernel maintainer that refuses to fix memory safety issues in their API.
microsoft and amazon pushing
If I was a kernel maintainer, I too would be sparing in my assistance to Rust conversion purely for the sake of being trendy.
I can't stand the Rust community, obnoxious overtly political kids with inadequate life experience trying to shove a fad language down everyones throat.
Nobody wants to put in the pointless effort to convert the kernel for style points, just to have Rust go out of fashion in 5-10 years and be left holding the bag. Pearl Ruby Ada Cobol etc etc... all came and went and the resulting lack of popularity meant a lack of volunteer support, so all the early adopter projects eventually had to re-adopt a more long-term stable language and reconvert all that code or fade into oblivian.
Yap, It's all come down to this. No one want to maintain other people's legacy code, especially the new toy that has no prove yet that it easier to maintain in the long run and still get the jobs done.
Can you even undo it once done? You will have this oddball language in there forever and almost impossible to maintain.
Ada would be a better choice than Rust, at least it has standards, multiple implementations, has been used for decades under sane leadership.
@@chrisnelson414 Ada was designed to be solid, but my point was more about a small talent pool in the long term. Eg Cobol is very stable and some companies still have cobol running on an old mainframe, but they are paying almost double normal industry wages to find the rare cobol maintainer. This is the part that doesn't work well for FOSS projects.
Nobody is reading all of that 😂
I demand you increase your workload for me, pronto!
Do they have to contribute on the Kernal to use it? Can they be forced by the Linux Foundation to not use Linux? PS I'm sticking with Pop!_OS.
They develop hardware. If a bug is found with the hardware they build the best solution is to submit a bug report along with the solution to the Linux kernel... And the lead engineer seems ideologically opposed to that.
They can use the kernel as long as they keep the changes public. Linux foundation can merge their changes themselves
@@markojovanovic9651 yeah but it's a grown up variant of kids playing "I'm not touching you" instead of being adults and doing it themselves.
@@Trahloc True, it's ridiculous that this is even happening
I’ve been on linux mint for a month and i like it, tho there is a ton of compatibility problems with proton for specifically 90’s and early 2000’s games and those games work on windows. I have decided that i will switch back to windows 10 and use the lsc edition that comes with next to no bloat and will be supported until 2032. Once support ends on it, i VOW to use linux only. Once i setup linux by then, I’ll passthrough a gpu to kvm for a virtual windows 10.
Theres just way too many bugs and glitches and compatibility issues right now with linux. It is getting there but not fast enough for me.
I also don’t like the way linux does things like it not being backwards compatible with it’s own games and programs if not maintained and having to paste in a random line of commands from some random on the internet in the terminal just to install the specific program you want if it’s not in the package manager.
To add on, these people acting like communists isn’t really helping with the linux stigma
Pop!OS does a lot of things right, I have it on my ThinkPad and really makes the whole thing much better
cosmic?
that's because system76 did not write the code for Thinkpad harware. I have a Dart Pro laptop from System76 and their own hardware support is very average.
@@barbarella7028 How so? My System76 laptop is 6 years old. Of course they support their own hardware. Sounds like user error.
Stock pop os is mid at best even on their own hardware.
Hope cosmic with 24.04 will be better when it goes out of alpha
@@namename8986 didn't upgrade just yet, but what I have works well for me and Cosmic will just be welcome upgrade
Rust devs being obnoxious again?
No. C Devs bullied the rust dev out
@@levaniandgiorgi2358 thank God. if only the rust devs were bullied as children instead and they might have grown up to be productive, functional individuals instead of a bunch of degenerates.
Nothing to do really with Linux - narcissism is endemic in the profession.
Hot news! Distro developers apply patches on upstream kernel source when building their releases!
The Pop! dude can release his drivers anyways.
If I wrote a piece of software in lets say assembly language, the best thing to do would be to ensure that people do not find it useful. This way the recreational aspect will be preserved.
Really hit the nail on the head with that religion of Rust comment. I've never seen a cult this koolaided since the height of C programming.
I don't remember C, or C++, being a religion, or having fanatical people, although I believe fanaticism exists everywhere. But with Python and Rust, it is common to encounter people defending their languages as if there was no tomorrow.
@@samuellourenco1050 Yea, Python too. It's nuts out there.
Watch your language. We're still around, and we can very easily find out who you are.
YourMoma is a cult 😂
@@samuellourenco1050 cpp is a cult who worships the standard library and their lesser deities RAII and OOP, but they aren't particularly active evangelists.
So, a systems programming language holy war...
kernel developers have work to do. porting isnt a priority when the current language is good enough (like in real jobs). devs whine when they dont get resources to do it. Its call open source. you can do it yourself. communists want to make people work on their projects. lol.
I like Pop_OS!... but just let linux kernel stay in C.... C is fine? why re-write in in Rust lmao? to make it DEI proof or what?
My general thoughts exactly.
yes. literally everything we do in western society is in service of baby-proofing it for muds and gays.
Lets just break everything.
Who needs a stable kernel?
It seems to originate on the competition which programming language is worse - C or Rust.
The Linux kernel is made by C programmers (and they want to continue to do it in C), while the Rust programmers want a kernel to be in Rust instead - right?
The question which is superior and which is inferior, that kind of thing is damaging the teamwork in the development team. And since Linux is a project relying on teamwork, this won't do any good for the entire project.
It's also not a good idea to start some 'keyboard wars' blaming other people that they're not doing what you want them to do.
getting Pop!_corn...
so anyone who doesn't like gnome or ubuntu has to manually patch the kernel to get full support for whatever system76 hardware?
i’m afraid windows will prevail
it already has. decades ago.
I have a feeling the lead engineer's comments have more to do than just Rust vs C. In any case, these public comments don't help anybody.
Pop!_OS has to be one of the best operating systems with one of the worst names for software. I really don't think they had a single marketing expert or consultant in the room when they named it. Everyone put names in a hate, shook it up, and picked it at random until they got something that wasn't a joke answer. I think it would have made sense for a hardware company called "System76" to name a software platform they're building to deliver the best possible experience on their hardware something other than "Pop" OS. What the hell? Why not name it "Platform76"?
This same developer Jeremy who is obverly obbssesed with Rust is the reason System76 decided to stop development pop os in favor of Cosmic desktop environment just because he had an argument with Gnome team. This caused Pop Os to have become very outdated and Pop OS 24 LTS is still very buggy.
I'm surprised nobody named and shamed the developers that want to stay in C or Rust
The sad part is that, as inadvisable as it may be for them to say so, they are right! Have you seen "Filesystem in Rust"? The whole seminar got derailed with relentless bikeshedding the moment a single piece of EXAMPLE code was shown. What does that tell you about the environment. What do you think it is like to contribute in that environment.
I'm not absolutely sure, but from what I could find, it seems Filho was merely asking what the semantics of an existing interface are, not to have an interface be created for his use. Sure, explaining something takes time, but that explanation could be documented and would be useful for other users of that interface, regardless of language.
I read some speculation that the precise objection to Rust code was that any change from the C side that breaks Rust code would require fixing the Rust code, which means maintainers would have to learn Rust. I thought C code changes breaking Rust code was considered acceptable and a Rust developer would take care of it.
I know this sounds like a pipe dream but couldn’t the normal linux devs just fork everything and start new foundations that we could support?
Does Mojira own system76 via a proxy? Is rust a backdoor into linux kernel?
Linux kernel is not just for desktop, but also for embedded devices and other weird systems (e.g HPC) that don't have Rust support/ecosystem (e.g CUDA); also for low level environment, you need to write unsafe code
They probably have other engineers working on the contributions - he just personally doesn't want to do it.
In any case, I'd say he's probably just an entitled jerk.. "Pedantic" probably means he just hates the rules for contributors (which exist for very good reasons)..
He omitted the reason because it wouldn't support his message here. Essentially, the hwmon maintainer refused to accept a driver for the thelio-io board because it is of a unique hardware design that is not similar to existing hwmon devices. The hwmon maintainer told Jeremy that he doesn't care about the reason for why the hwmon driver is written the way it is, and he doesn't want Jeremy to explain it. Given that it is technically impossible to do what the hwmon maintainer demands, it could not be merged.
The same happened to Asahi Lina. The DRM maintainer demands that the drm driver be implemented exactly the same as amdgpu, except that it is not possible because it's not an amdgpu chip.
Kudos to System76 Pop!_OS Lead ^_^ It's time for a hard fork of the kernel before any Rust was able to creep in.
The same dev denied part of 'even try to contribute to the Linux kernel anymore' on kernel, I hope you'll correct it :)
And your 'in short' is wrong - its about bindings specifically :)
It was a bad move to allow rust in the kernel. The rust 'community' is rife with entitled crybabies.
Linux is meant to be portable. The official Rust compiler, on the other hand, can only produce machine code to a very limited number of targets, and it doesn't even support some very common targets like pre-SSE2 x86 CPUs. Yet for some reason rustaceans just don't want to fix their compiler, but they want everyone to use it. It just doesn't make sense.
Is it that they won't contribute because they keep trying to push Rust code, and it gets rejected?
No, it's that the kernel maintainer won't accept patches to fix memory safety problems in the C code that the Rust compiler caught while working on the Rust buildings for their interfaces.
Is that what happened to all the C++ zealots? They gave up and became Rust fanatics?
No, Rust is a downgrade to C++ programmers. Rust lacks ALL features that distinguish C++ from C. Rust is just C with ball and chain.
C++ devs are coping right now because they never made it into the kernel LMAO 😂
I dunno, this seems like the kind of thing that would win you points with Linus lol. Also, I think you might be reading too much into this. The lead might not want to contribute to the Kernel anymore, but I'm sure delegation is happening. It's not great PR, but I really don't think the Kernel team is going to burn bridges with them. There is just going to be a "wow" and then move on.
It seems like a very small thing to blow up to this proportion. I really dont see the point.
Lunduke has made a living on reading too much into the things that feed arguments that bolster his worldview.
This is some nonsense. If you want to rewrite storage in rust just do it *then* come back when its done. If its better its easy to adopt, if its not better then only your time is lost.
It could be they use upstream to ubuntu and ubuntu handles the rest honestly.
There is some intense hatred that was shown during that Linux Filesystem in Rust presentation. It wasnt the Q and A portion - basically it was 8 minutes in and the torches came out and derailed the whole presentation, based on a slide about modeling a rust type to help programmers understand what can be done out of an opetation. It was baffling to see as an outsider, I am assuming this was some straw to break the C programmer camels back or something
Good, Linux is and should remain a C project, it has been for over 32 years quite successfully, now these Rust folks come along and try to make the project bend to their will to accommodate their pet language? THAT is real "non-technical nonsense".
I don't see Ruby devs trying to get the kernel to support their language.
When they pushed for the Rust inclusion in the kernel they met with obvious resistance, but argued it was only optional, it should not be a big deal, now they are trying to make longtime devs accommodate the needs and goals of their language.
Give these people a hand and they will take your arm, and then your head.
I've heard that you can escape Rust's safety feature; sooo, is that an weekness
They are not wrong that a lot of linux developers are pretty conservative and even retrograde. But why do they bother to push theirs ideas down the throat of someone? One can always fork any FOSS project and do with it anything one wants. Same as they did with GNOME creating COSMIC. That's the beauty of FOSS.