Originally posted by kylew77
View Post
Announcement
Collapse
No announcement yet.
Rust-Written Replacement To GNU Coreutils Progressing, Some Binaries Now Faster
Collapse
X
-
Last edited by mdedetrich; 30 January 2022, 06:29 AM.
- Likes 3
-
Originally posted by sdack View PostWe had many standards after K&R, not to mention C++, and none of them being "heat of the moment". Any programmer, who does not know the risk of using gets(), will be just as bad a programmer regardless of the language they use. The language that stops people from making mistakes has not been invented yet despite all claims of Rust being a magic bullet.
Well I think I'll call it a day for this discussion
- Likes 1
Comment
-
119 comments, 90% are utter garbage.
I see that there is a strong anti-Rust party, made up of people of have never programmed. Otherwise things like "C has no undefined behaviour", "Rust is just hype", "Rust is made to inflate the egos of incompetent programmers", "you must write large swats of your code in unsafe Rust" and other "pearls" shouldn't be here.
I note that very few people understand that the tectonic shift promoted by Rust regards forcing upon programmers a restricted programming model, with strong guarantees and limitations that make possible to write compilers that can reason about the code being compiled.
This is why Rust compiler is able to catch a lot of bugs at compile time. No magic, just science.
And yes, you start from unsafe code and encapsulate it in higher level abstractions ( for those things that aren't enforced by the programming model, like forcing strings to contain valid Unicode code points only ), with appropriate controls on inputs and a careful coding. Nonetheless you reason about small portions of code, of which you can "see" the state by static analysis.
It is appalling to read comments of people that don't understand that software complexity is exponential and the human brain is incapable of following the state of thousands ( or even millions ) of code fragments interacting with each other. More so when the software is updated and other people make the changes. People who obviously could never know every quirk in the code and/or the architecture. Or do you think that every Linux contributor knows every single line of code in the kernel? There are some contributors who have specialized on specific subsystems. And often they don't know how their changes can affect distant parts of the kernel ( ever heard of the butterfly effect? ).
Seriously, reading such bull**** on a technology forum is depressing. And only because there are luddists who don't like innovation. Start programming and we will see how far you will go writing super duper secure software in C.
- Likes 12
Comment
-
Originally posted by catpig View Post
You're saying C++ is the same as C? And you're not joking? Wow
Well I think I'll call it a day for this discussion
He would rather argue that 2+2=5 than admit that he doesn't know what he is talking about and the only reason he is selling his children to argue that C is the bees knees because he is a GNU sellout and can't stand something technically superior is beating gnu coreutils in god forbid, a different language.
- Likes 2
Comment
-
Originally posted by mdedetrich View Post
Why is it hard to believe, fundamentally Rust remains more type information which means the compiler has more information to make better performance optimizations. Its also the case that doing multithreaded programming in Rust is a lot easier/safer which probably gave some performance wins.
- Likes 1
Comment
-
Originally posted by pmorph View Post
Whatever is the case with that, it's also quite normal for the re-implementation to be faster, since the programmer actually aims to be as good or better as the original, whereas the original just aimed to get the job done with good enough performance (from practical use cases point of view).
look at this way, if something in coreutils is slow then a **lot** of software is slow, its a central part of the ecosystem.
Comment
-
Originally posted by sdack View PostNo. A protocol is in itself a set of instructions.
Originally posted by sdack View PostYou are tacking on code onto code to fix other code, because you have not learned why you keep making these mistakes.
Originally posted by sdack View PostYou will make as many mistakes as before, no matter the language you use.
Originally posted by sdack View PostIn the end will you get caught up in so many restrictions that the only reason why you make fewer mistakes is because you get less done.
Originally posted by sdack View PostAnd each of these new wheels is made for a specific purpose that these cannot really be used anywhere else than their intended place. These all do the same and trade simplicity for complexity and lose their usefulness. If you cannot see this then you can also not see why C is still successful.
- Likes 5
Comment
-
Originally posted by catpig View Post
So by that logic we should remove e.g. railings to prevent people from falling down stairs.... Car vendors used to have that absurd attitude, too, claiming "our cars are designed for driving, not crashing" - which was exactly the problem, they should've been designed (in part) for crashing. "Oh we'll just fix the bugs once we encounter them" is a great attitude for hobby projects, but not so much if you're dealing with stuff that matters.
- Likes 1
Comment
-
Originally posted by sdack View PostFire, wheel, blade, hammer, ...
Those are ancient and are still going strong.
You better be ready to add C to this list.
- Likes 1
Comment
-
Originally posted by MadCatX View PostWe know what mistakes we make and why and yet we still keep making them.
A modern car wheel is objectively better and more universal than a wooden chariot wheel. Stainless steel kitchen knife with ergonomic handle is objectively better than a rusting piece of metal. Glock 45 is a better gun than a blunderbuss.
That you mention a Glock 45 is rather curious. Do you believe people should have the right to a gun or that these should be taken away from them?Last edited by sdack; 30 January 2022, 08:02 AM.
Comment
Comment