Swift参数的类型没有正确推断？

Question

String has the following initializers (amongst many):

init(count: Int, repeatedValue: Character)
init(count: Int, repeatedValue: UnicodeScalar)

so shouldn't the following work?

let str = String(count:8, repeatedValue:"*")    // Error: "Extra argument 'count' in call"

Shouldn't the compiler be able to infer the type?! Even if double quotes are strictly for String literals, both Character and UnicodeScalar conform to the UnicodeScalarLiteralConvertible protocol. So what's happening here?

Sure I could always do repeatedValue:("*" as Character) but then what's the point of having types inferred?!

Answer 1

Well the problem here (as @Thilo already mentioned) is the ambiguity of protocol conformance.

Character > ExtendedGraphemeClusterLiteralConvertible > UnicodeScalarLiteralConvertible
UnicodeScalar > UnicodeScalarLiteralConvertible

The compiler checks these protocols, but when it hits UnicodeScalarLiteralConvertible it has no idea which initializer to choose.

You could make this easier by extending String with a unique initializer:

extension String {
    init(count: Int, repeatedCharacter: Character) {
        self.init(count: count, repeatedValue: repeatedCharacter)
    }
}

let str1 = String(count: 8, repeatedCharacter: "*")
println(str1)

or (although a bit over the top I think) a custom operator which makes it explicit that it's a Character :

postfix operator • {}
postfix func •(c: Character) -> Character { return c }

let str2 = String(count: 8, repeatedValue: "*"•)
println(str2)