[swift-evolution] specialize a generic type in a method

Vladimir.S svabox at gmail.com
Tue Jun 21 10:20:49 CDT 2016


On 20.06.2016 21:38, Charlie Monroe wrote:
> Something like this should work:
>
> protocol P1 {
> 	associatedtype T
> 	func foo(t: T)
> 	func bar<U where U == T, U: Equatable>(t: U)
> }
>

Thank you. Strange, but also can't check this code - compiler crashes on it.

IMO it will be good if we can write just

func bar(t: T & Equatable)

>
>> On Jun 20, 2016, at 8:07 PM, Vladimir.S <svabox at gmail.com> wrote:
>>
>> Thank you for reply. Yes, seems this could be solved by extension. No other solutions?
>>
>> But how should I be if I need such a protocol?
>>
>> protocol P1 {
>>    associatedtype T
>>    func foo(t: T)
>>    func bar(t: T) // where T:Equatable  ??
>> }
>>
>> I.e. I need to specify T & Equatable for bar in requirements
>>
>> On 20.06.2016 20:03, Charlie Monroe wrote:
>>> This is IMHO better solved via an extension - it even makes more sense to put together methods whose use is limited only to certain generics constraint:
>>>
>>> extension Foo where I: Equatable {
>>>
>>> 	func bar(i: I) {
>>> 		///
>>> 	}
>>>
>>> }
>>>
>>>
>>>> On Jun 20, 2016, at 6:57 PM, Vladimir.S via swift-evolution <swift-evolution at swift.org> wrote:
>>>>
>>>> Should we be able to specialize a generic type in a method, if that generic type declared in type's definition like here? :
>>>>
>>>> struct Foo<I> {
>>>>   func bar(i: I) where I: Equatable { } // should this work?
>>>> }
>>>>
>>>> As I understand, for concrete instance of foo: Foo<X>, if X is not Equatable, then compiler should prevent us from calling foo.bar(x)
>>>>
>>>> Or, probably, the better way to express this could be :
>>>>
>>>>   func bar(i: I & Equatable) { }
>>>> or
>>>>   func bar(i: Any<I, Equatable>) { }
>>>>
>>>> Can't check if it possible now because my compiler crashes because of 'where I:Equatable' text (submitted to bugs.swift.org)
>>>> _______________________________________________
>>>> swift-evolution mailing list
>>>> swift-evolution at swift.org
>>>> https://lists.swift.org/mailman/listinfo/swift-evolution
>>>
>>>
>
>


More information about the swift-evolution mailing list