Fix the KCL any type to not panic

[jtran]

I was trying to understand and test the limits of HomArray and MixedArray. I tried to make functions that use any as a type, and this needed fixing.

With this PR, the following KCL works.

fn id(@x: any): any {
  return x
}

fn singleton(@x: any): [any; 1] {
  return [x]
}

fn len(@a: [any]): number(_) {
  return reduce(a, initial = 0, f = fn(_, count) {
    return count + 1
  })
}

[qa-wolf]

QA Wolf here! As you write new code it's important that your test coverage is keeping up.
Click here to request test coverage for this PR!

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
modeling-app	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	May 8, 2025 10:28pm

[codspeed-hq]

CodSpeed Instrumentation Performance Report

Merging #6748 will not alter performance

_{Comparing jtran/test-arrays (383afbf) with main (fc3ce4c)}

Summary

✅ 54 untouched benchmarks

[nrc]

I'd expect all those tests to pass, if they don't then I suspect that the any type hasn't been properly wired up in execution and just exists as syntax for documentation

[nrc]

I think subtype should be altered so that any is a supertype of everything, and then this line is not needed

[jtran]

Updating subtype made my tests fail in a weird way 😖 so this still needs some work.

[jtran]

Because of implicit coercion of single values into arrays, we were coercing an empty MixedArray into HomArray { value: vec![MixedArray { ... }] }. Fix is here: 3a07212.