News for hackers. Stuff that matters.
Re:Reflections on Rust, and the Sand Castle Metaphor by csomar on Friday April 27, 2018 @11:42AM (#16941614)
This is indeed very light in content. Also, while Rust fixes a whole set of problems, it doesn't mean that Rust fixes all possible problems.

It is relatively easy to create a situation where you overflow your program. See here: https://github.com/rust-lang/rust/issues/50049

Googling around, there are many situations where bad code can lead to an overflow and your code breaking.

Re:Reflections on Rust, and the Sand Castle Metaphor by urschrei on Friday April 27, 2018 @11:55AM (#16941738)
Who, ever, anywhere, has claimed that "Rust fixes all possible problems"? This is a straw man. As for bad code leading to overflows and crashes: yes, bad code causes overflows and crashes, definitionally. Rust does not, and has never claimed to protect against logic errors in this way. How could it?
Re:Reflections on Rust, and the Sand Castle Metaphor by edflsafoiewq on Friday April 27, 2018 @12:00PM (#16941785)
The language in the article is very strong.

> Rust is another step into the beyond. When finishing a feature and its test suite I’ll run my program to see it in action, but just as a formality – I already know it works. I also know that it’s going to keep working because meticulousness of the compiler is so good at catching regressions.

Re:Reflections on Rust, and the Sand Castle Metaphor by sargas on Friday April 27, 2018 @12:41PM (#16942167)
This is a bold statement, specially because not many other languages have such high quality guarantees at compile time.

> I already know it works

This is the part I don't agree all of the time, but depending on the scope and size of the project, I've seeing this be true. And the compiler is really good at catching regressions, submitting PRs for Rust projects is not easy in the beginning, but it is very hard to insert a regression.

Re:Reflections on Rust, and the Sand Castle Metaphor by yarrel on Friday April 27, 2018 @12:15PM (#16941927)
I've had conversations with enthusiasts who are incorrectly convinced that ownership/borrowing fixes problem X.

Trying to explain otherwise to them is an exercise in futility. Because ownership/borrowing.

So serious Rust project members may never claim this, but there is a large less widely experienced Rust community who have turned the claims Rust does make into something much stronger and more resistant to logic.

Re:Reflections on Rust, and the Sand Castle Metaphor by hathawsh on Friday April 27, 2018 @12:51PM (#16942250)
The same author wrote a Rust article filled with a lot of great content: https://brandur.org/rust-web
Re:Reflections on Rust, and the Sand Castle Metaphor by a_t48 on Friday April 27, 2018 @12:34PM (#16942103)
This is the equivelent of `~Foo() { Foo f; }` isn't it?
Re:Reflections on Rust, and the Sand Castle Metaphor by AndrewGaspar on Friday April 27, 2018 @01:24PM (#16942525)
Yeah, not sure I understand the criticism here. This program also has a bug.

  fn main() {
      let a = 3;
      assert(a == 2, "a should definitely be 2");
  }
Rust doesn't mean no bugs ever, and it doesn't mean no crashes. It protects you from a very specific class of severe bugs, and it has a pretty good set of built-in primitives to help you write bug-free code in the other classes. No more, no less.
Re:Reflections on Rust, and the Sand Castle Metaphor by bvinc on Friday April 27, 2018 @11:47AM (#16941666)
It's even easier than that:

let buf = [0u8; 810241024];