Protocol Types Cannot Conform to Protocols

[tera]

From Protocol Types Cannot Conform to Protocols document:

Currently, even if a protocol P requires no initializers or static members, the existential type P does not conform to P (with exceptions below). This restriction allows library authors to add such requirements (initializers or static members) to an existing protocol without breaking their users' source code.

The word "Currently" is encouraging. Are there plans to make amendments here? For example have a notion of restricted protocol that can't have initialisers / static members, which would allow the existential of type P conform to P.

bikeshedding:

---

Edited:

Wrong example

restricted protocol Animal {
    init() // 🛑 Error, can't have initializers in restricted protocols
    static func foo() // 🛑 Error, can't static members in restricted protocols
    func makeNoise()
}

func declareAnimalSpecies<T: Animal>(_ animal: T) {
    animal.makeNoise()
}

var animal: Animal = Dog()
declareAnimalSpecies(animal) // ✅

Corrected example:

restricted protocol CanMakeNoise {
    func makeNoise()
}

struct Dog: CanMakeNoise {
    func makeNoise() { print("Woof") }
}

func declareAnimalSpecies<T: CanMakeNoise>(_ animal: T) {
    animal.makeNoise()
}

extension Array: CanMakeNoise where Element: CanMakeNoise {
    func makeNoise() {
        forEach { element in
            element.makeNoise()
        }
    }
}

var animal: CanMakeNoise = Dog()
declareAnimalSpecies([animal]) // ✅
// currently gives error

Or perhaps a version of "frozen" attribute, which if apply to a protocol that doesn't have inits/statics would make it "restricted".

[smkuehnhold (Samuel Kuehnhold)]

@tera

func declareAnimalSpecies<T: Animal>(_ animal: T) {
    animal.makeNoise()
}

var animal: Animal = Dog()
declareAnimalSpecies(animal) // ✅

Doesn't this example already work because of existential unwrapping?

[tera]

Thank you, corrected example above.

[xwu (Xiaodi Wu)]

@tera

The word "Currently" is encouraging. Are there plans to make amendments here?

No, it was meant to signify that this is a design decision, not a law of the universe, which is a distinction (but not a contradiction) from the preceding paragraph.