Due to Swift's very static nature, mocking and stubbing is much harder to do than in other languages. There are no dynamic mocking framework like OCMock or Mockito. The usual approach is to just write your mock objects manually, like so:

protocol Foo {
    func doSomething(arg: String) -> Int
}

class FooMock: Foo {
    var doSomethingHasBeenCalled = false
    var doSomethingArgument = String?
    var doSomethingReturnValue: Int!

    func doSomething(arg: String) -> Int {
        doSomethingHasBeenCalled = true
        doSomethingArgument = arg
        return doSomethingReturnValue
    }
}

This is a lot of boilerplate (and this is just a simple example, which doesn't allow for conditional stubbing, for example). This is where Mokka comes in.

Mokka provides a testing helper class called FunctionMock<Args> (and a variant for returning functions called ReturningFunctionMock<Args, ReturnValue>) that takes care of:

• Recording function/method calls for verification (Has the method been called?, How often has the method been called?)
• Capturing the arguments for verification (With which arguments has the method been called?)
• Stubbing return values (also conditionally) (This method should return 42 if called with argument "x")

With these helpers it gets much more convenient to define your mock objects:

class FooMock: Foo {
    let doSomethingFunc = ReturningFunctionMock<String, Int>()
    func doSomething(arg: String) -> Int {
        return doSomethingFunc.recordCallAndReturn(arg)
    }
}

You can now use the function mock object for verification:

func testSomething() {
    // ...
    XCTAssertEqual(myMock.doSomethingFunc.callCount, 2)
    XCTAssertEqual(myMock.doSomethingFunc.argument, "lorem ipsum")
    // ...
}

and for faking the return value:

func testSomething() {
    // static return value
    myMock.doSomethingFunc.returns(100)

    // dynamic return value
    myMock.doSomethingFunc.returns { $0 + 200 }   // $0 is the argument(s) passed to the method

    // conditional return value
    myMock.doSomethingFunc.returns(123, when: { $0 == "foo" })
    myMock.doSomethingFunc.returns(456, when: { $0 == "bar" })
    myMock.doSomethingFunc.returns(789)
}

• Xcode 10.2
• Swift 5.0

To install Mokka via CocoaPods, just add the Mokka pod for your test target to the Podfile:

pod 'Mokka'

You can install Mokka using Swift Package Manager. Just add this repository as a dependency to your Package.swift file (and don't forget to also add "Mokka" as a dependency in your test target):

dependencies: [
    .package(url: "https://github.com/danielr/Mokka", from: "1.0.0")
    // ...
]

To install Mokka with Carthage, add this to your Cartfile:

github "danielr/Mokka"

Then drag the built Mokka.framework to your project and make sure it's added to the unit test target, not the main app target. If you see issues errors like "The bundle xxx couldn’t be loaded because it is damaged or missing necessary resources.", then you might need to tweak your test target's Runtime Search Paths.

There are currently three types of mock helpers available:

• FunctionMock: Allows to record the calls to a function (the call count and the arguments), as well as to optionally stub the function's behavior. Use this for functions that have a Void return value.
• ReturningFunctionMock: Provides the same functionality as FunctionMock, but adds the ability to fake the returned value. Use this for functions that have a non-void return value.
• PropertyMock: Allows to provide fake values for a property and record whether a property has been read or set.

The first step is to implement your mocks using Mokka's helpers. The general approach is the same for all types of mocks: You declare a property for the mock and use that mock object inside your function implementations to record the calls to that function.

For the examples below, let's assume we want to mock the following protocol:

protocol Engine {
    func turnOn()
    func turnOff()
    var isOn: Bool { get }

    func setSpeed(to value: Float)  // kilometers per hour
    func setSpeed(to value: Float, in unit: UnitSpeed)

    func currentSpeed(in unit: UnitSpeed) -> Double
}

For functions that don't return a value, use FunctionMock<Args>. This class has one generic parameter which defines the type(s) of the argument(s).

For functions with no arguments, that should be Void:

class EngineMock: Engine {
    let turnOnFunc = FunctionMock<Void>(name: "turnOn()")
    func turnOn() {
        turnOnFunc.recordCall()
    }
	
    // ...
}

Note: The name parameter in the mock initializers is optional. It is purely informational and might be useful for error messages (e.g. better assertion error messages). It is good practice to provide the names in the standard Swift #selector syntax.

For functions with a single argument, just use that argument's type:

class EngineMock: Engine {
    let setSpeedFunc = FunctionMock<Float>(name: "setSpeed(to:)")
    func setSpeed(to value: Float) {
        setSpeedFunc.recordCall(value)
    }
	
    // ...
}

For functions with more than one argument, you need to use a tuple to represent the arguments (because Swift does not (yet?) support variadic generic parameters). Although you don't have to, it is a good practice to name the tuple elements, which makes it much clearer when referring to them in your testing code.

class EngineMock: Engine {
    let setSpeedInUnitFunc = FunctionMock<(value: Float, unit: UnitSpeed)>(name: "setSpeed(to:unit:)")
    func setSpeed(to value: Float, in unit: UnitSpeed) {
        setSpeedInUnitFunc.recordCall((value: value, unit: unit))
    }
	
    // ...
}

For functions with a return value, use ReturningFunctionMock<Args, ReturnValue>. This works very much the same way as FunctionMock, but adds a second generic parameter for the return value type. It also provides a recordCallAndReturn() method, instead of the recordCall() method.

class EngineMock: Engine {
    let currentSpeedFunc = ReturningFunctionMock<UnitSpeed, Double>(name: "currentSpeed(in:)")
    func currentSpeed(in unit: UnitSpeed) -> Double {
        return currentSpeedFunc.recordCallAndReturn(unit)
    }
	
    // ...
}

For the arguments of returning methods, the same rules apply as for non-returning functions (see above). For example:

• A mock for a function that has no arguments and returns a Bool would be declared as ReturningFunctionMock<Void, Bool>
• A mock for a function that has two arguments of type Int and String? and returns a Double would be declared as ReturningFunctionMock<(arg1: Int, arg2: String?), Double>

In many cases it is enough to just implement property requirements of the mocked protocol by declaring a stored property with a default value in your mock. However, when you want to be able to explicitly track whether a property has been read or written, Mokka's PropertyMock can be helpful. It is generic over the type of the property and its use in the mock implementation is quite self-explanatory: Instead of using a stored property, declare a computed property and delegate the getter and setter (if it's settable property) to the get() and set(_:) methods of the PropertyMock object:

class EngineMock: Engine {
    let isOnProperty = PropertyMock<Bool>(name: "isOn")
    var isOn: Bool {
        get { return isOnProperty.get() }
        set { isOnProperty.set(newValue) }
    }
	
    // ...
}

Note that you don't need to provide a default value for the property (the get() method fails with a preconditionFailure if there is no value).

A common use case for mocks is to verify if a method has been called, and sometimes specifically how often it has been called. For that, both FunctionMock and ReturningFunctionMock provide some properties:

• called: Bool Returns whether the method has been called (once or more).
• calledOnce: Bool Returns whether the method has been called exactly once.
• callCount: Int The number of times the method has been called.

XCTAssertTrue(engineMock.setSpeedFunc.called)
XCTAssertTrue(engineMock.setSpeedFunc.calledOnce)
XCTAssertEqual(engineMock.setSpeedFunc.callCount, 3)

In addition to verifying if a function has been called, you often also want to check the argument(s) with which the function has been called. You can do that via the arguments property:

XCTAssertEqual(engineMock.setSpeedInUnitFunc.arguments.value, 100.0)
XCTAssertEqual(engineMock.setSpeedInUnitFunc.arguments.unit, .kilometersPerHour)

(This requires that you follow the recommended practice of naming the tuple members, see above. If you don't, you have to access the arguments by their index, e.g. arguments.0.)

For single-argument functions (where there's no arguments tuple) you can also use the argument property, which looks a bit nicer:

XCTAssertEqual(engineMock.setSpeedFunc.argument, 100.0)

Sometimes it's necessary to stub the behavior of a function, for example to introduce some important side-effects. One common example for this is calling a delegate method. You can do that by providing a closure that will be executed when the function is called. The closure will be provided with the function arguments:

let delegate = FooDelegateMock()
someMock.myFunction.stub { arg in
    delegate.somethingHappened(with: arg)
}

For returning functions it's crucial to be able to fake the return value. Mokka provides 3 ways of doing that: Static return values, dynamic return values and conditional return values. Let's have a look at each of those.

For most cases it's sufficient to provide a simple static value that should be returned by the mock implementation:

engineMock.currentSpeedFunc.returns(100.0)

Sometimes it's convenient to provide a return value that is dynamically generated, often depending on the function's arguments. You can do that by providing a closure:

engineMock.currentSpeedFunc.returns { unit in
    // always return 100 km/h, converted to the requested target unit
    let kmhValue = Measurement(value: 100, unit: UnitSpeed.kilometersPerHour)
    return kmhValue.converted(to: unit).value
}

Both, static and dynamic return values can also be provided conditionally:

engineMock.currentSpeedFunc.returns(100.00, when: { $0 == .kilometersPerHour })
engineMock.currentSpeedFunc.returns(62.137, when: { $0 == .milesPerHour })
engineMock.currentSpeedFunc.returns(0)	   // otherwise

This is how you use properties that are backed by PropertyMock in your testing code:

someMock.fooProperty.value = 10	// use value to access the underlying property value

// do something

XCTAssertTrue(someMock.fooProperty.hasBeenRead)
XCTAssertFalse(someMock.fooProperty.hasBeenSet

You can find a simple example project in MokkaExample.

It includes

• a subject under test (Car)
• two mocked protocols (Engine and Battery)

It is a minimal example, but it should be enough to get you started with the concepts of Mokka.

Mokka has been created and is maintained by Daniel Rinser, @danielrinser.

Mokka is available under the MIT License.