chore: Allow count of elements in repeated-element array syntax to be any static Field

[ghost]

Description

WIP

Problem

Resolves #445

Summary

WIP

PR Checklist

• I have tested the changes locally.
• I have formatted the changes with Prettier and/or cargo fmt on default settings.

[ghost]

@jfecher, сan you check this and give me a hint?

[jfecher]

@f01dab1e, #445 is only for array expressions like let foo = [1; length];, so we shouldn't need any type changes to this PR. I was thinking this may be quite difficult to implement originally but now that we have slices I think you can leverage those to desugar the above syntax into something like:

let foo = {
    let mut result = [];
    for _ in 0 .. length {
        result = result.push_back(1);
    }
    result.as_array()
};

We could even break this up and add this function to the std library:

// Free function in std::array
fn repeated<T>(element: T, length: Field) -> [T] {
    let mut result = [];
    for _ in 0 .. length {
        result = result.push_back(1);
    }
    result
}

// New method in std::slice
impl<T> [T] {
    fn as_array<N>(self) -> [T; N] {
        assert(self.len() == N);
        ... now somehow return an array without using [elem; len] syntax ...
    }
}

Hmm we may need a builtin for converting a slice to an array since I don't see how you can write as_array otherwise.

[ghost]

@jfecher, how to handle a case like this?

fn main(x : Field) {
    let x = [x];
    let n = x.as_array();
}

The application panicked (crashed).
Message:  Non-numeric type variable used in expression expecting a value
Location: crates/noirc_frontend/src/monomorphization/mod.rs:626

use dep::std;

fn main() {
    let slice = std::array::repeated(0, 5);
    let xs: [Field; 5] = slice.as_array();
    let x = xs[0];
    std::println(x);
}

error: Could not determine loop bound at compile-time
   ┌─ std/array.nr:89:17
   │
89 │     for _ in 0..length {
   │                 ------
   │
   = Call stack:
     1. std/array.nr:89:17

Error: Aborting due to 1 previous error

[jfecher]

@f01dab1e I'm not sure. I'd think N should default to 0 if it cannot be inferred but of the top of my head I'm not sure why your errors are happening. The for loop bounds error are usually either because the length is actually not known at compile-time, or SSA optimizations are not optimizing it out by that point.

[ghost]

@jfecher, closed by mistake, can you open it?

[kevaundray]

What is the status of this PR and can the issue and PR title be updated given we don't have comptime?

[ghost]

This was blocked by the fact that we had no way to indicate the type (size cannot be inferred from the context), now it's blocked by the same thing, but there's an open pull request after which we can resume work, but we encounter another problem, that the method with desugaring will break the formatter.