feat: Added type hierarchy support for arrays and slices

[sirasistant]

Description

Problem*

Arrays are a subtype of slice and SSA & brillig gen wasn't handling that properly.

Summary*

This adds support for type hierarchy handling in defunctionalization and in brillig gen. Also fixes various bugs found along the way. The changes to support type hierarchy itself are contained so it should be easy to remove when we remove type hierarchy from the SSA. In the meanwhile, this PR should allow us to continue advancing in other tasks.

Documentation

• This PR requires documentation updates when merged.

  • I will submit a noir-lang/docs PR.
  • I will request for and support Dev Rel's help in documenting this PR.

Additional Context

PR Checklist*

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

[sirasistant]

I'm working on an approach that I think can be better, only casting as needed and only in function calls

[jfecher]

If we decide against type hierarchy in SSA, we can just revert this PR.

This wouldn't be needed then if arrays were not subtypes of slices? Perhaps we should just remove that and make them distinct types. Separating their syntax as well.

[sirasistant]

This PR can be used as a stopgap until #1963 is finalized

[jfecher]

This PR can be used as a stopgap until #1963 is finalized

I don't think we should merge PRs if they're only going to be used temporarily while we explore something else. Unless that other thing is quite a long time away and the stopgap would be both non-breaking and easy to remove. I'll work on #1963 today if getting this fixed quickly is important.

[sirasistant]

The stopgap should be non-breaking and removing it should be as simple as removing the is method in SSA::Type, the usage of those casts in defunctionalization::CallSignature::can_call and the three cast methods in brillig_block. This PR includes other changes that are fixes to issues found testing this and other improvements. I will separate this in two PRs!

[sirasistant]

separated into this one and #1973

[kevaundray]

@jfecher and @sirasistant where did we leave off in regards to this PR?

[jfecher]

It was my impression this PR is no longer needed after #1963 was merged. Pinging @sirasistant for confirmation

[sirasistant]

Not needed anymore after 1963, closing!