[swift-users] A somewhat stupid question about protocol conformances

[Adrian Zubarev]

Sorry yeah you’re right, the example also requires conditional conformances. If you do not need dynamic dispatch in your use case then you can workaround the issue for now by hiding the protocol requirements and abusing the protocol extension:


import UIKit

protocol AreaProtocol { /* keep it empty */ }

extension AreaProtocol where Self : Sequence, Self.Element == CGPoint {
	func area() -> CGFloat {
		return 0.0
	}
		
	func volume(height: CGFloat) -> CGFloat {
		return height * area()
	}
}

// or directly
extension Sequence where Element == CGPoint {
	// stuff from above
}

extension Array : AreaProtocol {}

let p0 = CGPoint(x: 0.0, y: 0.0)
let p1 = CGPoint(x: 2.0, y: 0.0)
let p2 = CGPoint(x: 2.0, y: 2.0)
let p3 = CGPoint(x: 0.0, y: 2.0)

let poligon = [p0, p1, p2, p3]
let a = poligon.area()
let v = poligon.volume(height: 10.0)


Am 25. November 2017 um 13:58:44, Antonino Ficarra (antonino.ficarra at gmail.com) schrieb:

Adrian,
thanks, but don’t work :-(


import Foundation

protocol AreaProtocol {
// required
func area() -> CGFloat

// implemented in protocol extension
func volume( height:CGFloat ) -> CGFloat
// ....
}

extension AreaProtocol {
func volume( height:CGFloat ) -> CGFloat {
return height * area()
}
// ....
// ....
}

protocol CGPointSequence: Sequence, AreaProtocol where Element == CGPoint {}

extension CGPointSequence {
func area() -> CGFloat {
return 0.0 // ... poligon area
}
}


extension Array:CGPointSequence {} // !'CGPointSequence' requires the types 'Element' and 'CGPoint' be equivalent




Il giorno 25 nov 2017, alle ore 09:11, Adrian Zubarev <adrian.zubarev at devandartist.com> ha scritto:

This is correct. None of your types does conform to your protocol because you never conformance explicitly. 

The extenstion you wrote is just a conditional extension which says if the Self type conforms to AreaProtocol and the associated typed Element is a CGPoint then add the area method to it. Nothing more, nothing less.

Now you may think that you have to rewrite that extension to `extension Sequence : AreaProtocol where Element == CGPoint`, but that won't work because retroactively conforming protocols to other protocols is not supported yet. Nor are conditional conformances included yet, but they will in Swift 4.1 or 5.

You have one option left.
protocol MySequence : Sequence, AreaProtocol where Element == CGPoint {}

extension Array : MySequence {}

Then also conform other sequences you need to support that functionality.

Double check if the code compiles, I wrote it on my iPhone, but you get the idea now.

-- 
Adrian Zubarev
Sent with Airmail
Am 25. November 2017 um 07:40:25, Antonino Ficarra via swift-users (swift-users at swift.org) schrieb:

In this code example:

import Foundation

protocol AreaProtocol {
func area() -> CGFloat

// implemented in protocol extension
func volume( height:CGFloat ) -> CGFloat
// ....
}

extension AreaProtocol {
func volume( height:CGFloat ) -> CGFloat {
return height * area()
}
// ....
// ....
}


// conform CGPoint sequences to AreaProtocol
extension Sequence
where Self : AreaProtocol, Element == CGPoint
{
func area() -> CGFloat {
return 0.0 // ... poligon area
}
}

let p0 = CGPoint(x: 0.0, y: 0.0)
let p1 = CGPoint(x: 2.0, y: 0.0)
let p2 = CGPoint(x: 2.0, y: 2.0)
let p3 = CGPoint(x: 0.0, y: 2.0)

let poligon = [p0,p1,p2,p3]
let a  = poligon.area()// ! Type '[CGPoint]' does not conform to protocol 'AreaProtocol'
let v  = poligon.volume( height:10.0 )// ! Value of type '[CGPoint]' has no member 'volume'


An array of CGPoint is a CGPoint sequence? Why the array don't gets automatic conformance to AreaProtocol?
How can conform an array of CGPoint to AreaProtocol?

Sorry for my bad english,
Antonino

_______________________________________________
swift-users mailing list
swift-users at swift.org
https://lists.swift.org/mailman/listinfo/swift-users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.swift.org/pipermail/swift-users/attachments/20171125/cf83e80c/attachment.html>