[zozbot234]

The more compelling scenario is one where the unsafe subset of Rust itself becomes roughly as easy to use as Zig is today, though still with potential challenges wrt. properly interacting with safe code.

That requires literally rethinking every language and standard library facility and asking "is this putting up artificial roadblocks or even invoking straight UB when one tries to use it idiomatically in unsafe contexts?", then coming up with more flexible, more "C like" facilities in that case. It's hard work but quite doable.

[kartoffelsaft]

The roadblock there is a cultural one. Among Rust devs if you ever find the need for an unsafe block then you need an explanation to back it up. If anything, the Rust language would benefit from adding as much friction to unsafe code as possible, so that you're only going to use it when you actually need it.

In other words, the Rust approach to safety is to make as few unsafe LoC as possible, and the Zig approach is to make every unsafe line as safe as possible. As long as their philosophical approach to safety is such that Zig makes writing unsafe code easy and Rust avoids it as much as possible, then writing unsafe code in Zig will always be easier.

[phicoh]

The goal of Rust is not so much to avoid unsafe, the goal is to maximize safe code. The reason is that you don't have to check safe code for all the guarantees to get from the Rust language.

This has a big effect on unsafe code. When unsafe code gets called indirectly from safe code, the unsafe code has to make sure that whatever the safe code does, the result is still safe. This requires very careful design of the interface provided by the unsafe code.

I think it is a research question whether Zig would make writing Rust unsafe code a lot easier. In particular the boundary between safe Rust and unsafe code written in Zig could become very tricky. Possibly tricky enough that it would be as complex to write as unsafe Rust today.

[zozbot234]

The problem is not so much "when unsafe code gets called indirectly by safe code", which is fine if the unsafe happens to be sound. The problem is that "when C-like or Zig-like unsafe code wants to call safe code", the safe code is running in a context where its implied invariants may be violated, leading to insta-UB. Hence code that's intended to provide "library-like" facilities to unsafe blocks cannot be idiomatic Safe Rust, and it needs to be written even more carefully.

For instance, any &mut reference in Rust is assumed not to be aliased, and any &reference not involving UnsafeCell<...> is assumed not to be written to. These implied contracts can be loosened, e.g. by using &Cell<> if applicable (may alias, can be read or written safely, but only "as a whole object": access to the internals does not escape beyond any single operation) which is arguably closer to idiomatic C.

MaybeUninit<> is another common example: C and Zig code often works with possibly-uninitialized data, but this possibility has to be accounted for explicitly in a safe Rust interface. It's always insta-UB if a safe Rust function is passed uninitialized data, even when the equivalent would work idiomatically in C.

[phicoh]

This is what makes Rust Rust and I'd say what makes Rust popular. The way safe Rust is safe. Of course this comes at the expense of making writing unsafe a lot harder.

Though I can imagine that unsafe Rust still has to many of the safe Rust's rules. So there could be a better unsafe language that has fewer restrictions.

[zozbot234]

The real sticking point is not whether there could be a better unsafe language, but a better unsafe library (or unsafe-friendly library that's still conditionally safe in some rigorous sense). That's a much closer goal that could even be achieved within Rust itself as it currently exists today.

[zozbot234]

No, it's not just a cultural thing. There are very real problems with standard Rust facilities not being accessible/usable from an unsafe context even though the C/C++/Zig idiomatic equivalent would be (i.e. would not invoke instant UB), and these problems are solvable. Sophisticated Rust developers are aware of the issue.

> If anything, the Rust language would benefit from adding as much friction to unsafe code as possible

The friction is there already. I'm not advocating for getting rid of explicit 'unsafe' markings, 'SAFETY' comments or even clunky boilerplate, just for closing a very real gap in capability.