Testing Asynchronous Code Continued

This past week John Sundell wrote a blog post (Unit testing asynchronous Swift code) about unit testing asynchronous code.

As I was trying to apply some of the techniques when I came across a method I couldn’t mock because I was getting an error.

Objective-C method addNotificationRequest:withCompletionHandler: provided by method add(_:withCompletionHandler:) does not match the requirement’s selector (add:withCompletionHandler:)

Consider the following method for UNUserNotificationCenter in Swift.

func add(_ request: UNNotificationRequest, withCompletionHandler completionHandler: ((Error?) -> Swift.Void)? = nil)

If you would want to create a Swift protocol to mimic the interface to UNUserNotificationCenter that would be pretty simple.

@objc
protocol UserNotificationCenter {
	func add(_ request: UNNotificationRequest, withCompletionHandler completionHandler: ((Error?) -> Swift.Void)?)
}

The problem is that the method signature in Swift is different from the method signature in Objective-C. Here is the method definition in Objective-C.

- (void)addNotificationRequest:(UNNotificationRequest *)request withCompletionHandler:(nullable void(^)(NSError *__nullable error))completionHandler;

Since the Objective-C method has a different method signature than the Swift version, we have to specify the Objective-C method signature in our protocol.

@objc(addNotificationRequest:withCompletionHandler:)
func add(_ request: UNNotificationRequest, withCompletionHandler completionHandler: ((Error?) -> Swift.Void)?)

Now we can implement that method in our mock object!