InstantMock aims at creating mocks easily in Swift, and configuring them with expectations or stubbed implementations.
For examples, see Example.playground
.
Swift versions compatibility:
Swift version | InstantMock version |
---|---|
5.0 | 2.5.X |
4.2 | 2.2.X |
4.0 | 2.0/2.1 |
3.X | 1.1.X |
InstantMock enables to create a single mock that can be used in many tests, for a protocol or a class.
The easiest way to create a mock for a protocol is to inherit from the Mock
class.
// MARK: Protocol to be mocked
protocol Foo {
func bar(arg1: String, arg2: Int) -> Bool
}
// MARK: Mock class inherits from `Mock` and adopts the `Foo` protocol
class FooMock: Mock, Foo {
// implement `bar` of the `Foo` protocol
func bar(arg1: String, arg2: Int) -> Bool {
return super.call(arg1, arg2)! // provide values to parent class
}
}
To create a mock for a class, the mock must adopt the MockDelegate
protocol.
// MARK: Class to be mocked
class Foo {
func bar(arg1: String, arg2: Int) -> Bool
}
// MARK: Mock class inherits from `Foo` and adopts the `MockDelegate` protocol
class FooMock: Foo, MockDelegate {
// create `Mock` delegate instance
private let mock = Mock()
// conform to the `MockDelegate` protocol, by providing the `Mock` instance
var it: Mock {
return mock
}
// implement `bar` of the `Foo` class
override func bar(arg1: String, arg2: Int) -> Bool {
return mock.call(arg1, arg2)! // provide values to the delegate
}
}
To work properly, mocks must comply with a few rules regarding return values, due to Swift strong typing.
The syntax is as follow:
func returnsOptional() -> Bool? {
return mock.call()
}
Here, call()
returns nil
or Void
.
For some methods, mocks must return non-optional values. If a return value type adopts the MockUsable protocol (which is the case for the most common types like Bool
, Int
…), just force unwrapping the result to call()
, like in the following example:
func returnsMockUsable() -> Bool { // `Bool` adopts `MockUsable`
return mock.call()! // force unwrapping
}
For other types, make sure to provide a default value, like in the following example:
func returnsCustom() -> CustomType {
return mock.call() ?? CustomType() // return a `CustomType` default value
}
For catching errors on throwing methods, simply use callThrowing()
instead of call()
.
If a return value type adopts the MockUsable protocol (which is the case for the most common types like Bool
, Int
…), just force unwrapping the result to callThrowing()
, like in the following example:
func bazMockUsable() throws -> Bool {
return try callThrowing()!
}
For other types, make sure to provide a default value, like in the following example:
func bazCustom() throws -> CustomType {
return try callThrowing() ?? CustomType() // return a `CustomType` default value
}
It is possible to mock properties declared in a protocol, like in the following example:
// define protocol with a property `prop` that has a getter and a setter
protocol FooProperty {
var prop: String { get set }
}
// mock of `FooProperty`
class FooPropertyMock: Mock, FooProperty {
var prop: String {
get { return super.call()! }
set { return super.call(newValue) }
}
}
Expectations aim at verifying that a call is done with some arguments. They are set using a syntax like in the following example:
// create mock instance
let mock = FooMock()
// create expectation on `mock`, that is verified when `bar` is called
// with "hello" for `arg1` and any value of the type of `arg2`
mock.expect().call(
mock.bar(arg1: Arg.eq("hello"), arg2: Arg.any())
)
Rejections are the contrary of expectations. They make sure no call is being done with some arguments. Simply use reject()
instead of expect()
.
In addition, expectations and rejections can be set on the number of calls: Use the following syntax:
// create expectation on `mock`, that is verified when 2 calls are done on `bar`
// with "hello" for `arg1` and any value of the type of `arg2`
mock.expect().call(
mock.bar(arg1: Arg.eq("hello"), arg2: Arg.any()),
count: 2
)
Setting expectations on properties can be done using the following syntax:
// create mock instance
let mock = FooPropertyMock()
// create expectation on `mock`, that is verified when the property `prop` is called
mock.expect().call(mock.prop)
// create expectation on `mock`, that is verified when the property `prop` is set
// with the exact value "hello"
mock.expect().call(
mock.property.set(mock.prop, value: Arg.eq("hello"))
)
Verifying expectations and rejections is done this way:
// test fails when any of the expectations or rejections set on `mock` is not verified
mock.verify()
Expecations can be reset this way:
mock.resetExpectations()
Stubs aim at performing actions when a function is called with some arguments. They are set using a syntax like in the following example:
// create mock instance
let mock = FooMock()
// create stubbed implementation of the `bar` method, which returns `true` when called
// with "hello" for `arg1` and any value of the type of `arg2`
mock.stub().call(
mock.bar(arg1: Arg.eq("hello"), arg2: Arg.any())
).andReturn(true)
Set the return value with andReturn(…)
on the stub instance.
This is done with andReturn(closure: { _ in return … })
on the stub instance. This enables to return different values on the same stub, depending on some conditions.
This is done with andDo { _ in … }
on the stub instance.
This is done with andThrow(…)
on the stub instance.
Chaining several actions on the same stub is possible, given they don't confict. For example, it is possible to return a value and call another function, like in andReturn(true).andDo { _ in print("something") }
.
Rules:
- the last closure registered by
andDo
is called first - the last error registered by
andThrow
is thrown - the last return value registered by
andReturn
is returned - otherwise, the last return value computation method, registered by
andReturn(closure:)
, is called
Stubs can be reset this way:
mock.resetStubs()
Example:
// configure mock to return "string" when calling `basic` whatever provided arguments
mock.stub().call(mock.basic(arg1: Arg.any(), arg2: Arg.any())).andReturn("string")
// reset the previously configured stubs
mock.resetStubs()
// calling `basic` does not return "string"
let ret = mock.basic(arg1: "", arg2: 2)
XCTAssertNotEqual(ret, "string")
Expectations are verified only if arguments match what is registered. Same goes for calling stubbed implementations.
Matching an exact value is done with Arg.eq(…)
.
Values can be matched if they:
- conform to the
AnyObject
protocol, which is the case for all classes implicitly, for exampleArg.eq(NSString("hello"))
- or conform to the
MockUsable
protocol, for exampleArg.eq(42)
- or they are types, for example
Arg.eq(String.self)
- or they are tuples, limited to 5 values, for example
Arg.eq(("a string", 42))
Matching any value can be done for types that adopt the MockUsable
protocol, with Arg.any()
.
Matching a value that verifies a certain condition is done with Arg.verify({ _ in return … })
.
Matching a closure is a special case. Use the following syntax: Arg.closure()
.
Limitation: closures can be matched as long as they have less than 5 arguments.
Arguments can also be captured for later use thanks to the ArgumentCaptor
class.
For example:
// create captor for type `String`
let captor = ArgumentCaptor<String>()
// create expectation on `mock`, that is verified when `bar` is called
// with 42 for `arg2`. All values for `arg1` are captured.
mock.expect().call(mock.bar(arg1: captor.capture(), arg2: Arg.eq(42)))
...
// retrieve the last captured value
let value = captor.value
// retrieve all captured values
let values = captor.allValues
Capturing a closure is particularly useful for stubbing the behavior of a method with callbacks, see this conversation.
Capturing a closure is a special case. Use the following syntax:
Limitation: closures can be captured as long as they have less than 5 arguments.
// create captor for type closure `(Int) -> Bool`
let captor = ArgumentClosureCaptor<(Int) -> Bool>()
...
// retrieve the last captured closure, and call it
let ret = captor.value!(42)
MockUsable
is a protocol that makes types easily usable in mocks.
For a given type, it allows to return non-optional values and to match any values.
Adding MockUsable
on an existing type is done by creating an extension that adopts the protocol. For example:
extension SomeClass: MockUsable {
static var any = SomeClass() // any value
// return any value
public static var anyValue: MockUsable {
return SomeClass.any
}
// returns true if an object is equal to another `MockUsable` object
public func equal(to value: MockUsable?) -> Bool {
guard let value = value as? SomeClass else { return false }
return self == value
}
}
Adding MockUsable
on an existing type that uses inheritance, should always be done on the deepest subclass.
Indeed, adding this extension to both a parent and a subclass would create build conflicts.
For now, the following types are MockUsable
:
- Bool
- Int, Int64
- UInt, UInt64
- Float
- Double
- String
- Set
- Array
- Dictionary
- Date
List of changes can be found here.
- Xcode 10
- iOS 9
- osX 10.10
InstantMock is available using CocoaPods, see Podfile
example:
target 'Example' do
# Tests target
target 'ExampleTests' do
inherit! :search_paths
pod 'InstantMock'
end
end
InstantMock is available using the Swift Package Manager, by adding the dependency either with Xcode or by editing the Package.swift
file:
.package(url: "https://github.com/pirishd/InstantMock", from: "2.5.6"),
Patrick Irlande - pirishd@icloud.com
InstantMock is available under the MIT License.