Hacker News Reader

Comment by miew | Hacker News Reader

Why C and not Rust or even Zig?

Because (a) I felt like it, (b) it would be more difficult to make raw syscalls in Rust, and (c) I use[1] flexible array members to minimize allocations, whereas Rust support for those quite bad[2]. It doesn't even look possible to allocate one with size known at runtime.

[1]: https://codeberg.org/201984/dut/src/branch/master/main.c#L16...

[2]: https://doc.rust-lang.org/nomicon/exotic-sizes.html#dynamica...

> it would be more difficult to make raw syscalls in Rust

Would you like to expand on this? Is it because of type conversions that you'd have to do?

> I use[1] flexible array members to minimize allocations

I was under the impression that FAM was a non-standard extension, but alas it is part of C99.

From what I'm seeing you have intrusive list where each `entry` points to the previous and next element, where the path itself is a bag of bytes as part of the entry itself. I'm assuming that what you'd want from Rust is something akin to the following when it comes to the path?

    struct Entry<const NAME_LEN: usize> {
        ..
 mode: Mode,
 name: InlineName<NAME_LEN>,
    }

    struct InlineName<const NAME_LEN: usize> {
 value: [u8; NAME_LEN],
    }

For syscalls, I would have needed to either pull in dependencies or write FFI bindings, and neither of those options are appealing when I could simply write the program in Linux's native language.

For the FAM, your example looks like it requires a compile-time constant size. That's the same as hardcoding an array size in the struct, defeating the whole point. Short names will waste space, and long ones will get truncated.

> For the FAM, your example looks like it requires a compile-time constant size. That's the same as hardcoding an array size in the struct, defeating the whole point. Short names will waste space, and long ones will get truncated.

You made me realize that the const generics stabilization work hasn't advanced enough to do what I was proposing (at least not in as straightforward way): https://play.rust-lang.org/?version=nightly&mode=debug&editi...

Those are const arguments, not const values, which means that you can operate on values as if they didn't have a size, while the compiler does keep track of the size throughout.

I'd have to take a more detailed look to see if this level of dynamism is enough to work with runtime provided strings, which they might not be, unless you started encoding things like "take every CStr, create a StackString<1> for each char, and then add them together".

This is not an appropriate question to ask, that I see sometimes in these threads. "Because the author wanted" is good enough a reason for them to write a program in C. It being a new project written in C can also be good enough reason for you not to use it: dust already exists and is written in Rust, which you can use instead.

Why Rust or Zig and not C?

Better for the author's resume, if they want to make it hype driven.

Also some nebulous "being more secure". Never mind that this tool does not have elevated privileges. You gotta watch out for those remote root exploits even for a local only app, man.

I mean you could extract an archive you've downloaded to your filesystem and said archive could have funky file names and then you use this tool..

But I suppose it's not a very likely bug to have in this kind of tool.

Or “they” could enter your house at 2 am, drug you and hit you with a $5 wrench until they get access to your files :)

Never understood this line of thought. Not everything needs to be super secure. Not everything is going to be an attack vector. No one is going to deploy this onto a production server where this program specifically is going to be the attack vector.

Memory safety is cool and all, but a program that effectively sums a bunch of numbers together isn't going to cause issues. Worst case the program segfaults

Because C is like a cult Toyota Supra with twin turbo and Rust or Zig is like another cool boring Corvette roadster.

Reply on news.ycombinator.com