The main disadvantage of this way is that we end up with twice as many classes. What's more, mutable subclasses often exist only as a way to initialize and modify their immutable versions. Many bugs can be caused by using a mutable subclass by accident. For example, a mental burden shows in setting up properties. We have to always check if a mutable subclass exists, and if so use copy modifier instead of strong for the base class.
大致意思是创建一个可变子类，它将所有属性重新定义为readwrite。这种方式的主要缺点是我们最终得到两倍的类。而且，可变子类通常仅作为初始化和修改其不可变版本的方式存在。偶然使用可变子类可能会导致许多错误。例如，在设置属性时会出现心理负担。我们必须始终检查是否存在可变子类。
还有一点这种方式只能在Objective-C中使用。
Builder pattern
Builder pattern 模式需要我们使用一个Builder来创建目标对象，目标对象的property依旧是readonly，但是Builder的对应property却可以选择为readwrite。依旧用UserModel为例，我们需要为其进行适当的改造，改造之后：
typealias UserModelBuilderBlock = (UserModelBuilder) -> UserModelBuilder
public class UserModel: NSObject{
public let userId: NSNumber
public private(set) var name: String?
public private(set) var email: String?
public private(set) var age: Int?
public private(set) var address: String?
init(userId: NSNumber) {
self.userId = userId
super.init()
}
convenience init(userId: NSNumber ,with block: UserModelBuilderBlock){
let userModelBuilder = block(UserModelBuilder.init(userId: userId))
self.init(userId: userModelBuilder.userId)
self.email = userModelBuilder.email
self.name = userModelBuilder.name
self.age = userModelBuilder.age
self.address = userModelBuilder.address
}
}
之后是对应的Builder
class UserModelBuilder: NSObject {
public let userId: NSNumber
public var name: String?
public var email: String?
public var age: Int?
public var address: String?
init(userId: NSNumber) {
self.userId = userId
super.init()
}
}
