I have below hierarchical inheritance relationship in Swift,
class Base {
func method1() { }
func method2() { }
}
class Child: Base {
override func method1() { }
func method3() { }
}
class GrandChild: Child {
}
Now, I want to enforce GrandChild to override method1() from Base class and method3() from Child class. How can I do so? Any workaround or better approach?
There is no way to do it on the compile time in Swift. The feature of enforcing method override is not present in Swift.
But there is a work around.
You can do that by putting a fatal error fatalError("Must Override").
Consider the following example.
class Base {
func method1() {
fatalError("Must Override")
}
func method2() { }
}
class Child: Base {
override func method1() { }
func method3() {
fatalError("Must Override")
}
}
class GrandChild: Child {
func method1() { }
func method3() { }
}
But the above method will not give any compile time errors. For that there is another workaround.
You can create a protocol.
protocol ViewControllerProtocol {
func method1()
func method3()
}
typealias ViewController = UIViewController & ViewControllerProtocol
So if you implement this protocol and do not implement the methods compiler will generate an error.
As a feature of protocols in Swift you can also provide a default implementation of the methods in a protocol extension.
Hope this helps.
Related
Ideally, I want to create a BaseViewController class that takes in a protocol type (of a delegate) and have a weak variable as the delegate. Something like this:
class BaseViewController<Delegate: AnyObject> {
weak var delegate: Delegate?
init(delegate: Delegate) {
self.delegate = delegate
super.init(...)
}
}
And then inherit from a view controller like so:
protocol MyDelegate: AnyObject {
func funcA()
func funcB()
}
class SomeViewController: BaseViewController<MyDelegate> {
func doSomething() {
delegate?.funcA()
}
}
This doesn't work as the compiler complains:
'BaseViewController' requires that 'MyDelegate' be a class type
How can I work this around to achieve what I need?
Thanks in advance :)
Thats because in swift protocols doesn't confirm to them selves, you can't use "MyProtocol" as concrete type confirming to protocol "MyDelegate"
What you can rather do is
protocol MyDelegate: AnyObject {
func funcA()
func funcB()
}
class BaseViewController<Delegate: MyDelegate> {
weak var delegate: Delegate?
init(delegate: Delegate) {
self.delegate = delegate
super.init(...)
//keeping OPs code as is
}
}
class SomeOtherDelegateClass: MyDelegate {
func funcA() {
//some code here
}
func funcB() {
//some code here
}
}
class SomeViewController: BaseViewController<SomeOtherDelegateClass> {
func doSomething() {
self.delegate?.funcA()
}
}
EDIT 1:
As OP mentioned in comment, he is trying to introduce a generic property in BaseViewController that will simply hold a weak reference to any instance whose class is decided/declared by Child classes of BaseViewController using generics, I am simplifying the above answer a bit
Try this
protocol MyDelegate {
func funcA()
func funcB()
}
class BaseViewController<Delegate> where Delegate: AnyObject {
weak var delegate: Delegate?
init(delegate: Delegate) {
self.delegate = delegate
super.init(...)
//keeping OPs code as is
}
}
class SomeOtherDelegateClass: MyDelegate {
func funcA() {
//some code here
}
func funcB() {
//some code here
}
}
class SomeViewController: BaseViewController<SomeOtherDelegateClass> {
func doSomething() {
self.delegate?.funcA()
}
}
protocol MyDelegate2 {
func funcABCD()
}
class SomeOtherDelegateClass2: MyDelegate2 {
func funcABCD() {
//some code here
}
}
class SomeViewController2: BaseViewController<SomeOtherDelegateClass2> {
func doSomething() {
self.delegate?.funcABCD()
}
}
TBH, I really dont see much of benefit of this design! Probably you need to revisit the code structure and see if you can come up with better code structure :)
You should set your delegate as a constraint for the generic type T in BaseViewController:
protocol MyDelegate: AnyObject {
func funcA()
func funcB()
}
class Delegated1: MyDelegate {
func funcA() { print("A1") }
func funcB() {}
}
class Delegated2: MyDelegate {
func funcA() { print("A2") }
func funcB() {}
}
class BaseViewController<T: MyDelegate>: UIViewController {
var delegate: T?
func doSomething() {
delegate?.funcA()
}
}
class SomeViewController1: BaseViewController<Delegated1> {}
class SomeViewController2: BaseViewController<Delegated2> {}
class TestClass {
let viewController1: SomeViewController1 = {
let viewController = SomeViewController1(nibName: nil, bundle: nil)
viewController.delegate = .init()
return viewController
}()
let viewController2: SomeViewController2 = {
let viewController = SomeViewController2(nibName: nil, bundle: nil)
viewController.delegate = .init()
return viewController
}()
// prints:
// A1
// A2
func myFunc() {
viewController1.doSomething()
viewController2.doSomething()
}
}
I have a family of abstract classes I want to implement to perform a task depending on the instance created. But when I try to call the method using the variable declared as the base class I get “Static member '…' cannot be used on instance of type '…'”
some pieces of the code looks like this
class BaseClass{
class func theTask(){
print("do nothing")
}
}
class SubClassA: BaseClass{
class func theTask(){
print("do class A task")
}
}
class SubClassB: BaseClass{
class func theTask(){
print("do class B task")
}
}
The variable is declared as a property of one of my viewControllers,
class SecondViewController: UIViewController {
var theObject: BaseClass = BaseClass()
override func viewDidLoad() {
super.viewDidLoad()
if(mysettings.selectedPort ==1){
theObject = SubClassA()
}else{
theObject = SubClassB()
}
configureUIElements()
}
#IBAction func scanButtonPressed(_ sender: Any) {
let devices = self.theObject.theTask()
}
}
The line that says
let devices = self.theObject.theTask()
Is the one that gives me the error
Static member 'theTask' cannot be used on instance of type 'BaseClass'
I come from C++ so this kind of coding is very common but Swift doesn´t seems to like it.
As you are creating and using instances of the classes you have to declare theTask as instance method and override it (which you have to do in any case)
class BaseClass{
func theTask(){
print("do nothing")
}
}
class SubClassA: BaseClass{
override func theTask(){
print("do class A task")
}
}
class SubClassB: BaseClass{
override func theTask(){
print("do class B task")
}
}
Or the other way round put the type into the variable and call the method on the type
class BaseClass{
class func theTask(){
print("do nothing")
}
}
class SubClassA: BaseClass{
override class func theTask(){
print("do class A task")
}
}
class SubClassB: BaseClass{
override class func theTask(){
print("do class B task")
}
}
...
var theObject: BaseClass.Type = BaseClass.self
...
if mysettings.selectedPort == 1 {
theObject = SubClassA.self
}else{
theObject = SubClassB.self
}
...
let devices = theObject.theTask()
You can use type(of:) to accomplish that.
Following your example:
type(of: theObject).theTask()
I have a situation where I declare two functions in a protocol, one of them takes an optional generically constrained parameter, the other function takes no parameters, but needs to be implemented in an extension as a default function where it calls the one with a parameter and passes nil. However I get this error:
Argument passed to call that takes no arguments
My code:
public protocol MenuAccessible {
var menuEntryViewController: UIViewController { get }
}
public protocol MenuTabBarControllerProtocol {
func reloadTabs<T>(from uiProvider: T?) where T: MenuAccessible
func reloadTabs()
}
public extension MenuTabBarControllerProtocol {
func reloadTabs() {
reloadTabs(from: nil) // error here, can't infer type
}
}
Obviously the compiler is not able to infer the type. If I for example pass a nil (Optional) of the required type, then the compiler is happy. For example:
struct MenuAccessibleObject: MenuAccessible {
var menuEntryViewController: UIViewController { return UIViewController() }
}
public extension MenuTabBarControllerProtocol {
func reloadTabs() {
let menuAccessible: MenuAccessibleObject? = nil
reloadTabs(from: menuAccessible) // passes nil, but compiler is happpy
}
}
Is there a way to pass nil in my default function implementation and not have to create that dummy object?
I don't understand why you are Using Generic T there if you are defining type is MenuAccessible
Following is simply compiler without any issue
public protocol MenuAccessible {
var menuEntryViewController: UIViewController { get }
}
public protocol MenuTabBarControllerProtocol {
func reloadTabs(from uiProvider: MenuAccessible?)
func reloadTabs()
}
public extension MenuTabBarControllerProtocol {
func reloadTabs() {
reloadTabs(from: nil)
}
}
public extension MenuTabBarControllerProtocol {
func reloadTabs(from uiProvider: MenuAccessible?) {
fatalError() // implement me
}
}
EDIT
I don't know this will work for you or not but try this
public protocol MenuAccessible {
var menuEntryViewController: UIViewController { get }
}
public class UIProvider:NSObject {
}
public protocol MenuTabBarControllerProtocol {
func reloadAllTheItems<T>(from uiProvider: T?) where T: UIProvider, T: MenuAccessible
func reloadTabs()
}
public extension MenuTabBarControllerProtocol {
func reloadTabs() {
self.reloadAllTheItems(from: Temp())
}
func reloadAllTheItems (provider:(UIProvider & MenuAccessible)) {
}
}
class Temp: (UIProvider & MenuAccessible) {
var menuEntryViewController: UIViewController {
return UIViewController()
}
}
I am writing my iOS Application in Swift 3.
I have a UIViewController extension, where I have to check if the controller instance responds to a method. Below is the code that I a trying out.
extension UIViewController {
func myMethod() {
if self.responds(to: #selector(someMethod)) {
}
}}
Here the responds(to:) method throws a compile time error
Use of unresolved identifier "someMethod".
I read in another post, we have to use self inside the selector argument, but even that is throwing some error.
A simple workaround:
#objc protocol SomeMethodType {
func someMethod()
}
extension UIViewController {
func myMethod() {
if self.responds(to: #selector(SomeMethodType.someMethod)) {
//...
self.perform(#selector(SomeMethodType.someMethod))
// or
(self as AnyObject).someMethod?()
//...
}
}
}
A little more Swifty way:
protocol SomeMethodType {
func someMethod()
}
//For all classes implementing `someMethod()`.
extension MyViewController: SomeMethodType {}
//...
extension UIViewController {
func myMethod() {
if let someMethodSelf = self as? SomeMethodType {
//...
someMethodSelf.someMethod()
//...
}
}
}
Create a protocol which requires someMethod()
protocol Respondable {
func someMethod()
}
And a protocol extension which affects only UIViewController instances
extension Respondable where Self : UIViewController {
func myMethod() {
someMethod()
}
}
Adopt the protocol to some of the view controllers
class VC1 : UIViewController, Respondable {
func someMethod() { print("Hello") }
}
class VC2 : UIViewController {}
class VC3 : UIViewController {}
Now call the method in the extension
let vc1 = VC1()
vc1.myMethod() // "Hello"
Otherwise you get a compiler error:
let vc3 = VC3()
vc3.myMethod() // error: value of type 'VC3' has no member 'myMethod'
Swift 4 answer:
If the selector is written as a string you won't get that error.
extension UIViewController {
func myMethod() {
if self.responds(to: "someMethod")) {
}
}
}
And then in the viewcontroller (dont forget the #objc):
#objc func someMethod() -> Void {}
I have a class with a delegate. I create a subclass, which also has a delegate. I wanted to let the protocol used for the second delegate extend the protocol used for first delegate:
protocol MySuperClassProtocol {
func foo()
}
class MySuperClass {
var delegate:MySuperClassProtocol?
}
protocol MySubClassProtocol:MySuperClassProtocol {
func bar()
}
class MySubClass: MySuperClass {
override var delegate:MySubClassProtocol? // compiler error - "cannot override..."
func test() {
delegate?.foo()
delegate?.bar()
}
}
class UserClass:MySubClassProtocol {
func foo() {
println("foo!")
}
func bar() {
println("bar")
}
}
Is there a way to solve this? The only possible solution I see is to make the 2 protocols independent of each other, and use different names. Like this:
protocol MySuperClassProtocol {
func foo()
}
class MySuperClass {
var mySuperClassDelegate:MySuperClassProtocol?
}
protocol MySubClassProtocol {
func bar()
}
class MySubClass: MySuperClass {
var mySubClassDelegate:MySubClassProtocol?
func test() {
mySuperClassDelegate?.foo()
mySubClassDelegate?.bar()
}
}
class UserClass:MySuperClassProtocol, MySubClassProtocol {
func foo() {
println("foo!")
}
func bar() {
println("bar")
}
}
But this looks a bit weird + will not let me use naming convention for delegate- "delegate".
Sorry for necroposting, the only one solution i found is:
protocol SuperClassDelegate {
func first_method()
}
class SuperClass {
var delegate: SuperClassDelegate?
func do_something() {
delegate?.first_method()
}
}
protocol SubClassDelegate: SuperClassDelegate {
func second_method()
}
class SubClass: SuperClass {
private var subDelegate: SubClassDelegate?
override var delegate: SuperClassDelegate? {
get { return self.subDelegate }
set { self.subDelegate = newValue as! SubClassDelegate? }
}
//override func do_something() {
// super.do_something()
// subDelegate?.second_method()
//}
func do_something_other() {
//subDelegate?.first_method()
self.do_something()
subDelegate?.second_method()
}
}
class InheritanceAndDelegation: SubClassDelegate {
let obj = SubClass()
init() {
obj.delegate = self
}
internal func first_method() {
print("Hello from SuperClass")
}
internal func second_method() {
print("Hello from SubClass")
}
func check() {
obj.do_something_other()
}
}
let inheritanceAndDelegation = InheritanceAndDelegation()
inheritanceAndDelegation.check()
//prints:
//Hello from SuperClass
//Hello from SubClass
Commented code works too. Hope it will be useful for someone.
I was trying to find an ideal solution to this for some time, but could not come up with anything better that this:
protocol BaseDelegateProtocol: class { }
class BaseDelegate: BaseDelegateProtocol { }
class BaseActor {
weak var delegate: BaseDelegate? = nil
}
// MARK: -
protocol ConcreteDelegateProtocol: class {
func doSomething()
}
class ConcreteDelegate: BaseDelegate, ConcreteDelegateProtocol {
func doSomething() {
// Do something
}
}
class ConcreteActor: BaseActor {
private weak var concreteDelegate: ConcreteDelegateProtocol? = nil
override var delegate: BaseDelegate? {
didSet {
concreteDelegate = delegate as? ConcreteDelegateProtocol
}
}
}
Above works in XCode 7 / Swift 2.
This pattern allows adopting more and more protocols on the way down inheriting from BaseDelegate.
There is no need to inherit protocols one from the other, which helps keeping things isolated.
didSet observer on delegate property is automatically called for superclasses, therefore no need to call super.<blah> explicitly, and no risk to 'forget' doing so
Concrete delegate properties can be kept private on each level of inheritance, thereby reducing the clutter.
You can do it in another way, you can add the delegate variable in Subclass and use it to access the SuperClassProtocol also using delegate?.foo().
protocol MySuperClassProtocol {
func foo()
}
class MySuperClass {
//var delegate:MySuperClassProtocol?
}
protocol MySubClassProtocol:MySuperClassProtocol {
func bar()
}
class MySubClass: MySuperClass {
var delegate:MySubClassProtocol?
func test() {
delegate?.foo()
delegate?.bar()
}
}
class UserClass:MySubClassProtocol {
func foo() {
println("foo!")
}
func bar() {
println("bar")
}
}
But the issue with this approach is you can never use MySuperClassProtocol independently unless you create a new SubClass of MySuperClass only for declaring delegate variable.