nightly-2024-05-07: fix: Compute the correct slice length when coercing from a literal ar…

…ray of complex types (#4986)

# Description

## Problem\*

Resolves #4967

## Summary\*

There is a problem in the [array to slice
coercion](https://github.com/noir-lang/noir/blob/07930d4373a393146210efae69e6ec40171f047b/compiler/noirc_evaluator/src/ssa/ir/instruction/call.rs#L87)
code. When evaluating the coercion, it does not take into account that
for SSA, complex types such as tuples and structs are flattened,
resulting in an array representation with length of elements times the
size of the complex type (as a result of the flattening). This results
in the replacement slice having an incorrect size.

Then, because the slice has an inflated size, the code which decodes the
elements of the slice for printing or tracking (in the debugger) crashes
when it attempts to decode beyond the values that it has available.

## Additional Context

There are two additional minor changes introduced in this PR:

1. Removed the `Optional<>` for `PrintableType::Array`'s length. This
was introduced to be able to print slices (setting the length to
`None`), but those are now represented properly with
`PrintableType::Slice`.
2. Added support for printing slices.

Both are pretty small changes which I made while investigating the
issue, but I'm willing to send separate PRs for those if required.

I also expanded the existing test cases to verify both the coerced slice
length and printing of slices.

## Documentation\*

Check one:
- [X] No documentation needed.
- [ ] Documentation included in this PR.
- [ ] **[For Experimental Features]** Documentation to be submitted in a
separate PR.

# PR Checklist\*

- [X] I have tested the changes locally.
- [X] I have formatted the changes with [Prettier](https://prettier.io/)
and/or `cargo fmt` on default settings.