Better diagnostic when type has constrained subscript setter override

[stephencelis]

Motivation

Calling a method on a type that is unavailable due to an unmatched conformance does provide a diagnostic that leads the user in the right direction.

For example:

struct A<B> {
  var c: Int {
    fatalError()
  }
  subscript(position: Int) -> B {
    fatalError()
  }
}

extension A where B: Equatable {
  var c: Int {
    get { fatalError() }
    set { fatalError() }
  }
  subscript(position: Int) -> B {
    get { fatalError() }
    set { fatalError() }
  }
  func f() -> Int { fatalError() }
}

func f() {
  let a = A<()>()
  a.c = 42  // 👎 Cannot assign to property: 'c' is a get-only property
  a[0] = () // 👎 Cannot assign through subscript: subscript is get-only
  a.f()     // 👍 Type '()' cannot conform to 'Equatable'
}

Proposed solution

I think it would be nice to provide a diagnostic that includes this kind of hint. For example:

a.c = 42  // 👍 Cannot assign to property: '()' does not conform to 'Equatable'
a[0] = () // 👍 Cannot assign through subscript: '()' does not conform to 'Equatable'
a.f()     // 👍 Type '()' cannot conform to 'Equatable'

Alternatives considered

No response

Additional information

No response