I'm looking at the example from the “Unowned References and Implicitly Unwrapped Optional Properties” section of the book “The Swift Programming Language.”
Their example code is
class Country {
let name: String
let capitalCity: City!
init(name: String, capitalName: String) {
self.name = name
self.capitalCity = City(name: capitalName, country: self)
}
}
class City {
let name: String
unowned let country: Country
init(name: String, country: Country) {
self.name = name
self.country = country
}
}
This works if I want to deal exclusively with Countries and the only purpose of the City type is to be a capital of a Country. But what happens if I want to create a City?
This creates a runtime exception because no reference to the City's Country is retained since it is an unowned variable:
var chicago = City(name:"Chicago", country: Country(name: "USA", capitalName: "Washington DC"))
chicago.country.name // Playground execution failed: error: Execution was interrupted, reason: EXC_BAD_ACCESS (code=EXC_I386_GPFLT).
How would I allow something like this without creating a Strong Reference Cycle?
There are two typical solutions:
If you want to primarily deal with cities, invert the relationship so that City has a strong reference to Country, and Country points back to an unowned instance.
If you want to have cities and countries as primary objects that cross reference each other, put all cities and countries into collections (or other form of store that owns them), and make both references weak. That way they don't own each other, and you don't have a cycle.
The best way to avoid retain cycles is to consider who owns every object. Objects can own each other, but that should be a clear hierarchy (i.e. a tree). If you have connections that go sidewards and and upwards in the hierarchy, make them weak or unowned.
Solution one is the upwards case, solution two is sidewards case.
Edit
A third option is, to have Country own a collection of all its cities. I think that makes most sense in this simple case, but it means the Country needs to create all cities in it's initialization, or have a method that adds cities.
Here's an example for the second case. It's quite complex, probably too much so for this simple case, but it illustrates extracting a common owner. I would normally use it if there are a lot of cross references. (Think relational database. The records don't own each other.)
class Country {
let name: String
weak var capitalCity: City?
init(name: String) {
self.name = name
}
}
class City {
let name: String
unowned let country: Country
init(name: String, country: Country, isCapital: Bool) {
self.name = name
self.country = country
if isCapital {
country.capitalCity = self
}
}
}
class Planet {
var countries: [Country] = []
var cities: [City] = []
}
let earth = Planet()
earth.countries = [
Country(name: "USA"),
Country(name: "Canada"),
]
earth.cities = [
City(name: "Washington DC", country: earth.countries[0], isCapital: true),
City(name: "Chicago", country: earth.countries[0], isCapital: false),
City(name: "Ottawa", country: earth.countries[1], isCapital: true),
]
In your example, nobody is owning the Country instance. That means it gets deallocated (freed) immediately.
var country = Country(name: "USA", capitalName: "Washington DC")
var chicago = City(name:"Chicago", country: country)
chicago.country.name
will fix it because our coutry variable will keep USA from deallocating
If you use an unowned reference, you always have to keep a strong reference somewhere else.
Related
as shown in the code below:
struct Person {
var name: String
}
struct Group {
var person: Person
func callAsFunction() -> Person {
// Person is immutable value
person
}
}
var james = Person(name: "James")
var group = Group(person: james)
group().name = "Wong" //ERROR: Cannot assign to property: function call returns immutable value
group() return an immutable value, that can't be changed! So Is there any way to make the callAsFunction() method return a mutable value?
Thanks ;)
Updated:
My idea is to transfer all the calls and visits of the Group to the Person object in the Group, just like using Person directly.
I can't use dynamicMemberLookup because I don't know what method or property there will be in Person. For example, there may be 100 methods and properties in Person (not only one name property as demonstrated), and it is impossible for me to write 100 subscript methods with dynamicMemberLookup.
My needs are a bit like proxy objects in the Ruby language. Accessing an object (Group) actually accesses another object (Person) inside it, as if the Group does not exist.
ruby proxy patterns:
https://refactoring.guru/design-patterns/proxy/ruby/example
CallAsFunction is the closest implementation so far, but requires that Person cannot be a Struct, otherwise it cannot be assigned to its properties.
Maybe it's not possible to implement this feature in Swift yet?
You're using the wrong dynamic method. What you want is dynamicMemberLookup. Watch closely. First, the preparation:
struct Person {
var name: String
}
#dynamicMemberLookup
struct Group {
var person: Person
subscript(dynamicMember kp:WritableKeyPath<Person,String>) -> String {
get { self.person[keyPath:kp] }
set { self.person[keyPath:kp] = newValue }
}
}
Now look at what that allows you to say:
var group = Group(person: Person(name: "James"))
group.name = "Wong"
print(group.person) // Person(name: "Wong")
Do you see? We set the name of the Group even though it has no name property, and the result was that we set the name of the Group's person which does have a name property.
The callAsFunction simply returns (a copy of the) Person, which is a value type. You cannot then mutate the property of it like that. It is equivalent to the following:
struct Person {
var name: String
}
Person(name: "Foo").name = "Bar"
That returns the same error:
If Person was a reference type, it would have worked, but not for a value type. And even if you took your value type, and first assigned it to a variable before mutating it, you would only be mutating your copy, not the original.
If you want the behavior you want, you would use a #dynamicMemberLookup as suggested by matt (+1) and outlined in SE-0195.
You said:
I can't use dynamicMemberLookup because I don't know what method or property there will be in Person. For example, there may be 100 methods and properties in Person (not only one name property as demonstrated), and it is impossible for me to write 100 subscript methods with dynamicMemberLookup.
You do not need “100 subscript methods.” It is the motivating idea behind #dynamicMemberLookup, namely that the properties will be determined dynamically. E.g., here is Person with two properties, but Group only has the one #dynamicMemberLookup.
struct Person {
var name: String
var city: String
}
#dynamicMemberLookup
struct Group {
var person: Person
subscript(dynamicMember keyPath: WritableKeyPath<Person, String>) -> String {
get { person[keyPath: keyPath] }
set { person[keyPath: keyPath] = newValue }
}
}
var group = Group(person: Person(name: "James", city: "New York"))
group.name = "Wong"
group.city = "Los Angeles"
print(group.person) // Person(name: "Wong", city: "Los Angeles")
If you want to handle different types, make it generic:
struct Person {
var name: String
var city: String
var age: Int
}
#dynamicMemberLookup
struct Group {
var person: Person
subscript<T>(dynamicMember keyPath: WritableKeyPath<Person, T>) -> T {
get { person[keyPath: keyPath] }
set { person[keyPath: keyPath] = newValue }
}
}
And
var group = Group(person: Person(name: "James", city: "New York", age: 41))
group.name = "Wong"
group.city = "Los Angeles"
group.age = 42
print(group.person) // Person(name: "Wong", city: "Los Angeles", age: 42)
i am somewhat new to swift and still trying to get a grasp on the knowledge. which is probably why i've run into waht feels like a simple mistake. i know that class instances are reference types but even knowing that im not entirely sure how to get the result im after. i want to change my SHIPTO without modifying my BILLTO. does anyone know a fix for this?
class Home {
var house: String
var street: String
var postalCode: String
init(house: String, street: String, postalCode: String) {
self.house = house
self.street = street
self.postalCode = postalCode
}
}
class DeliveryAdress {
var orderNumber: Int
var Shipto: Home
var Billto: Home
init(orderNumber: Int, Shipto: Home, Billto: Home) {
self.orderNumber = orderNumber
self.Shipto = Shipto
self.Billto = Billto
}
}
var homeAddress = Home(house: "appartment style", street: "Pizza Pie Road", postalCode: "90210")
var pizzaDelivery = DeliveryAdress(orderNumber: 001, Shipto: homeAddress, Billto: homeAddress)
pizzaDelivery.Shipto.street = "123 Street"
print(pizzaDelivery.Billto.street)
// this should NOT print "123 street"
print(pizzaDelivery.Shipto.street)
// this SHOULD print 123 street
You have 2 solutions:
Use value types - change class to struct
Create separate Home instances for shipTo and billTo
I'd suggest sticking with structs, since unless you explicitly need reference-type behaviour, you should always prefer structs over classes in Swift.
struct Home {
var house: String
var street: String
var postalCode: String
}
struct DeliveryAddress {
var orderNumber: Int
var shipTo: Home
var billTo: Home
}
var homeAddress = Home(house: "appartment style", street: "Pizza Pie Road", postalCode: "90210")
var pizzaDelivery = DeliveryAddress(orderNumber: 001, shipTo: homeAddress, billTo: homeAddress)
pizzaDelivery.shipTo.street = "123 Street"
print(pizzaDelivery.billTo.street)
print(pizzaDelivery.shipTo.street)
There are a couple of general issues with your code:
The Swift naming convention is lowerCamelCase for variable+function names, and UpperCamelCase for type names. So shipTo and billTo, not Shipto.
By default, everything should be immutable, only use mutable properties when you really need them to be mutable.
Use struct instead of class unless you explicitly need reference-type behaviour.
As you said classes are reference type so the property of street has an adress lets say "A". pizzaDelivery.Shipto.street and pizzaDelivery.billTo.street refers to same address , its A. If you change pizzaDelivery.billTo.street this value .Its actually changing value which address is A . So In that situations You need to use Struct. If you change Home type yo Struct , issue gonna be fixed
struct Home {
var house: String
var street: String
var postalCode: String
}
I've created the following class
class Person {
var firstName: String
var lastName: String
init(firstName: String, lastName: String) {
self.firstName = firstName
self.lastName = lastName
}
func fullName() -> String {
return "\(firstName) \(lastName)"
}
}
Then I instantiated a constant value from the class
let john = Person(firstName: "Johnny", lastName: "Applessed")
Question: Why I can change the content of the variable john? Isn't it a constant? Can someone explain that for me, thanks a lot.
john.firstName = "John"
print(john.firstName) // -> John
As #Wain has said – it's due to the nature of reference types. The instance being a let constant only means you cannot assign a new reference to it – but says nothing about the actual mutability of the instance itself.
If you change your class to a struct, you'll see how the behaviour differs with value types, as changing a property changes the actual value of your Person – therefore you are unable to do so if it's a let constant. However I somewhat doubt you'll want to make your Person a struct, as two people with the same name shouldn't be considered to be the same person.
If you only wish your properties to be assigned upon initialisation (and then read-only for the lifetime of the instance), then I would recommend making them let constants (instead of making their setters private). This will ensure that you cannot even change their value from within your class, once assigned.
The rule is as long you give a property a value before the super.init() call – you can make it a let constant (in this case, you just have to assign them in the initialiser before using self).
class Person {
let firstName: String
let lastName: String
init(firstName: String, lastName: String) {
self.firstName = firstName
self.lastName = lastName
}
...
The class instance itself is a constant, so you can't change it to reference another instance, but the instance is mutable because it's properties are created as vars.
Change firstName to have a private setter and see what you can do:
private(set) var firstName: String
When you're using a constant instance of a class in swift, doesn't mean you can't change the class attributes. It' means you can't instantiate a new object in this constant
let person = Person(firstName: "Johnny", lastName: "Appleseed")
person = Person(firstName: "John", lastName: "Appleseed") //--->It gets error: Cannor assign to value: 'person' is a 'let' constant
But you can create a constant inside class and set this values in the init
class Person {
let firstName: String
let lastName: String
init(firstName: String, lastName: String) {
self.firstName = firstName
self.lastName = lastName
}
func fullName() -> String {
return "\(firstName) \(lastName)"
}
}
//Tip: Don't init the class constants in declaration time or will get the same above error. Just init this constants at constructor/initialization of class.
And Now you have the expected result you want, even if create a 'var' instance of this object
var person = Person(firstName: "Johnny", lastName: "Appleseed")
person.firstName = "John" //--->It gets error: Cannor assign to value: 'person' is a 'let' constant
person = Person(firstName: "John", lastName: "Snow")
person.firstName = "Johnny" //--->It gets error: Cannor assign to value: 'person' is a 'let' constant
Your thinking was not wrong, but a little confuse cause you would be totally right if it was a struct instead a class.
I'm working through a learn-swift playground and upgrading it to Swift 2.0 as I learn the language. The following code (which likely worked with prior versions of Swift) now generates two errors: "'self' used before all stored properties are initialized" and "Constant 'self.capitalCity' used before initialized"
class Country
{
let name: String
let capitalCity: City!
init(name: String, capitalName: String)
{
self.name = name
self.capitalCity = City(name: capitalName, country: self)
}
}
class City
{
let name: String
unowned let country: Country
init(name: String, country: Country)
{
self.name = name
self.country = country
}
}
reading an answer to a similar question I see that I can change let capitalCity: City! to var capitalCity: City! and the syntax error is resolved.
I realize that in this contrived example a country's capital city can change, so that would be fine, but what if there were a case where the value really was a constant...
Is there any way to resolve the syntax error while keeping capitalCity a constant?
In this case I would suggest you to make the property a variable but hiding it (make it seem like a constant) through a computed property:
class Country {
let name: String
private var _capitalCity: City!
var capitalCity: City {
return _capitalCity
}
init(name: String, capitalName: String) {
self.name = name
self._capitalCity = City(name: capitalName, country: self)
}
}
Is there any way to resolve the syntax error while keeping capitalCity a constant?
Not the way you have things set up. The source of the problem is actually that in order to set capitalCity, you have to create a City whose country is self. That is the use of self to which the compiler is objecting:
self.capitalCity = City(name: capitalName, country: self)
^^^^
Since you have configured City's country as a constant, you must supply this value when you initialize your City. Thus you have no way out; you must make capitalCity an Optional var so that it has some other initial value that is legal, namely nil. Your proposed solution actually works like this:
class Country
{
let name: String
var capitalCity: City! = nil // implicit or explicit
init(name: String, capitalName: String)
{
self.name = name
// end of initialization!
// name is set (to name), and capitalCity is set (to nil)...
// ... and so the compiler is satisfied;
// now, we _change_ capitalCity from nil to an actual City,
// and in doing that, we _are_ allowed to mention `self`
self.capitalCity = City(name: capitalName, country: self)
}
}
Just do:
private(set) var capitalCity: City!
which gives you the read-only public interface you want.
I understand that you find var capitalCity: City! contrived. I'd say that the selected answer is truly contrived; it adds lines of code that have no purpose other than resolving a language-related issue. Lines like that are meaningless and meaning is what matters most in code.
Came across this recently while struggling with a similar problem and as of swift 5.5 (or possibly lower) there is another alternative that is interesting. If you convert Capital City to a lazy var you are actually able to use self in initialization.
class Country
{
let name: String
lazy var capitalCity: City = {
City(name: capitalName, country: self)
}()
private let capitalName: String
init(name: String, capitalName: String)
{
self.name = name
self.capitalName = capitalName
}
}
class City
{
let name: String
unowned let country: Country
init(name: String, country: Country)
{
self.name = name
self.country = country
}
}
Cheers!
I have been chewing on this for a week now, and unless I am missing a very fundamental concept in Swift, this seems like a pretty serious bug.
I have a CoreData model with one-to-many relationships like so:
City -> has many Street
Also each entity has a name property of type string.
The corresponding classes are:
class City: NSManagedObject {
#NSManaged var name: String
#NSManaged var streets: NSOrderedSet
}
class Street: NSManagedObject {
#NSManaged var name: String
}
And here comes the really weird stuff:
var city:City = getCities().first!
var streets = city.streets.array as! [Street]
var streetToDelete:Street? = nil
//This works
streetToDelete = streets.first
//This is not working !!!
for street in streets {
if(street.name == "champs elysees"){
streetToDelete = street
}
}
deleteStreet(city, street:streetToDelete!)
And the deleteStreet function:
func deleteStreet(var city:City, var street:Street){
var streets = city.streets
var streetsSet:NSMutableOrderedSet = streets.mutableCopy() as! NSMutableOrderedSet
streetsSet.removeObject(street)
city.streets = streetsSet.copy() as! NSOrderedSet
saveContext()
}
So if I use streets.first, the deletion works as accepted.
But if I use the foreach loop instead nothing happens - no error is thrown, and no street is deleted. (off course I need the foreach loop to find a specific street and delete it).
How come the .first and the foreach are different?
They suppose to return me a pointer to a street in both cases.
(getCities() is a simple, straight from the book, fetch operation)
Update:
After further investigation, it seems that some how the iteration itself over the streets array causes the issue.