新视角:通过函数式编程的范畴理论来看待这个问题会更容易理解;

在低层类型无法很好表达的类型,可以将其转化为高阶类型进行表示。

将协议的实现类型转化为monad类型;

解决将具有关联类型的协议当作类型的

swift所谓的类型擦除是指通过协议的泛型化、继承、封装,将协议类型抹除的过程;

泛化的最终目的:

将实现了同一协议,并且关联类型一样的类型作为同一类型来看待。

接口类型最终目的:

1、接口类型用实例类型初始化后的行为和实例类型的行为一致;

2、实现了协议的可类型化,和实例类型的封装;

3、隐藏了底层的实现细节,合并了类型功能和封装功能;

let testGo: [_AnyCupBase<Coffee>] = [_AnyCupBox(CoffeeMug()), _AnyCupBox(CoffeeCup())]

终极限制:关联类型的一致性必须的到满足;

涉及到的元素:

关联类型的可指定类型;

协议的实例类型;

协议的泛化类型:协议的可类型化;用于做类型使用;用于确保关联类型的一致性;用于将容器类型从关联类型的功能中解放出来

容器类型、泛化类型的继承:1、用于包装协议的实例类型;2、用于给协议的泛化类型指定类型;

接口类型:1、接口层面只包含泛型化的关联类型;2、将协议的泛化类型作为类型来引用容器类型;3、初始化容器类型,赋值给协议的泛化类型类型;

协议的泛化是协议可类型化的基础;

消息的传递通过类似与代理机转发机制来实现。

Introduction

If you have ever tried to use Swift protocols with associated types as types, you probably had (just like me) a very rough first (and possibly second) ride.  Things that should be intuitively doable are just not and the compiler gives infuriating error messages.  It comes down to the fact that the language is not yet “finished” (and, as a result, protocols are not really usable as types — yet).  The good news is that problems have workarounds and one such workaround is “type erasure” (not the Java compiler feature, something else that goes by the same name).  Apple uses the type erasure pattern every time you read Any<Something> and cleverer people than myself described it well (more information in the references at the end).  This is my attempt at understanding the pattern.

So what’s the problem?

Imagine that you have a protocol Cup  that describes a generic cup, with an associated type Content  that describes the cup’s content.  Imagine that this protocol has some methods and some properties.  For fun (and for realism) let’s make these methods mutating  and returning and instance of Self  (the specific instance of Cup we are dealing with).

 
1
2
3
4
5
6
7
8
protocol Cup {
    // Have an associated type
    associatedtype Content
    var description: String { get }
    // Use this associated type
    mutating func replaceContent(with newContent: Content) -> Self
    func showContent() -> Content?
}

Let’s define some concrete types for Content .

 
1
2
3
4
5
6
7
8
9
struct Tea {
    let description = "tea"
    let hasMilk: Bool
}
 
struct Coffee {
    let description = "coffee"
    let isDecaf: Bool
}

Let’s also add some concrete types for the protocol Cup . A cup of tea and a cup of coffee perhaps.  Honouring the spirit of Swift, I’ll make them value types ( struct ) but it would work in the same way for a class , without any further adjustment.

 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
struct TeaMug: Cup {
    // Assign a concrete type to the associated type as your first action. This is not necessary but will unleash the power of autocomplete.
    typealias Content = Tea
    let description = "tea mug"
    private var content: Tea?
    // Please note that after specialising Content, these lines will autocomplete nicely
    mutating func replaceContent(with newContent: Tea) -> TeaMug {
        content = newContent
        return self
    }
    func showContent() -> Tea? {
        return content
    }
}
 
struct CoffeeMug: Cup {
    typealias Content = Coffee
    let description = "coffee mug"
    private var content: Coffee?
    mutating func replaceContent(with newContent: Coffee) -> CoffeeMug {
        content = newContent
        return self
    }
    func showContent() -> Coffee? {
        return content
    }
}
 
struct CoffeeCup: Cup {
    typealias Content = Coffee
    let description = "coffee cup"
    private var content: Coffee?
    mutating func replaceContent(with newContent: Coffee) -> CoffeeCup {
        content = newContent
        return self
    }
    func showContent() -> Coffee? {
        return content
    }
}

We can nicely create different instances of cups with different contents.

 
1
2
3
4
5
var myCoffeeCup = CoffeeCup()
myCoffeeCup.replaceContent(with: Coffee(isDecaf: false))
 
var myTeaMug = TeaMug()
myTeaMug.replaceContent(with: Tea(hasMilk: true))

And here are the problems that I invariably face when I start thinking of Cup  as a type (it’s not a type! It’s a protocol!).

(1) Collection of protocol types

 
1
2
3
4
let cupboardOne: [Cup] = [CoffeeMug(), CoffeeCup()]
let cupboardTwo: [Cup] = [TeaMug(), TeaMug()]
// The following works but it's probably not what you want, the type is not generic and is a concrete [TeaMug] rather than a generic [Cup]
let cupboardThree = [TeaMug(), TeaMug()]

(2) Returning a protocol type in a function

 
1
2
3
func giveMeACup() -> Cup? {
    return cupboardOne.first
}

(3) Taking a protocol type as a parameter  to a function

 
1
2
3
func describe(cup: Cup) {
    print(cup.description)
}

(4) Using a protocol type as a type for class or struct properties

 
1
2
3
class mySuperDuperClass {
    var cup: Cup
}

The error message

The error message, in all cases above, is always the same: protocol ‘Cup’ can only be used as a generic constraint because it has Self or associated type requirements.  In other words, if a protocol has associated types you cannot use such protocol as a type. It makes sense, the Swift compiler will not know what exact protocol type you are dealing with unless the associated type is specified.  A nice mnemonic way to put could be:

Swift does not like multidimensional type uncertainty, but one dimensional type uncertainty is OK.

If we can somehow fix the associated type Content , we can make it work.  There are individual workarounds that work for problems (2) to (4) but they have different limitations: we are not going to discuss them here.  We are going to present a more general solution that takes care of all four problems.

The Solution

The architecture for the solution requires a few steps and a lot of patience to follow.  I will try my best to be clear.  Here we go.

(1) Base class mocking the protocol

First let’s define a base class that describes the protocol Cup .  The associated type Content  is now turned into a generic parameter for the class.  The class cannot be used directly because it cannot be initialised.  Furthermore, every property and every method throw a fatal error upon access or call: they must be overridden, all of them.  Why do we need this base class?  It’s a trick.  As we will see later, we need to be able to store a subclass of it but we cannot refer directly to an instance of such subclass.

 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
private class _AnyCupBase<Content>: Cup {
 
    // All the properties in the Cup protocol go here.
    var description: String {
        get { fatalError("Must override") }
    }
    
    // Let's make sure that init() cannot be called to initialise this class.
    init() {
        guard type(of: self) != _AnyCupBase.self else {
            fatalError("Cannot initialise, must subclass")
        }
    }
    
    // All the methods in the Cup protocol here
    func replaceContent(with newContent: Content) -> Self {
        fatalError("Must override")
    }
    func showContent() -> Content? {
        fatalError("Must override")
    }
}

(2) A box containing a concrete instance of the protocol

Let’s define a box, where the concrete type for protocol Cup  is stored: such box will inherit from our previously define abstract class.  By inheriting from _AnyBase , it also follows the Cup  protocol.  The generic parameter ConcreteCup  also follows the protocol Cup .  All methods and properties of ConcreteCup  are wrapped by this class. Note that when referring to this class we need to specify a concrete type for Cup . For example: _AnyCupBox<TeaCup> .  That is not what we want.  When referring to its super class instead, we only need to specify a concrete type for its Content .  For example: _AnyCupBase<Tea> .  This should shed some light on why we need a base abstract class to inherit from and this trick lays at the heart of the technique.

 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
private final class _AnyCupBox<ConcreteCup: Cup>: _AnyCupBase<ConcreteCup.Content> {
    // Store the concrete type
    var concrete: ConcreteCup
    
    // Override all properties
    override var description: String {
        get { return concrete.description }
    }
    
    // Define init()
    init(_ concrete: ConcreteCup) {
        self.concrete = concrete
    }
    
    // Override all methods
    override func replaceContent(with newContent: ConcreteCup.Content) -> Self {
        concrete.replaceContent(with: newContent)
        return self
    }
    override func showContent() -> ConcreteCup.Content? {
        return concrete.showContent()
    }
}

(3) The pseudo-protocol

And finally we can define the public class that we are going to use.

 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
final class AnyCup<Content>: Cup {
    // Store the box specialised by content.  
    // This line is the reason why we need an abstract class _AnyCupBase. We cannot store here an instance of _AnyCupBox directly because the concrete type for Cup is provided by the initialiser, at a later stage.
    private let box: _AnyCupBase<Content>
    
    // All properties for the protocol Cup call the equivalent Box proerty
    var description: String {
        get { return box.description }
    }
    
    // Initialise the class with a concrete type of Cup where the content is restricted to be the same as the genric paramenter
    init<Concrete: Cup>(_ concrete: Concrete) where Concrete.Content == Content {
        box = _AnyCupBox(concrete)
    }
    
    // All methods for the protocol Cup just call the e quivalent box method
    func replaceContent(with newContent: Content) -> Self {
        box.replaceContent(with: newContent)
        return self
    }
    func showContent() -> Content? {
        return box.showContent()
    }
}

See how this works!

 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
// Use the generic type in a collection
 
let cupboardOne: [AnyCup<Coffee>] = [AnyCup(CoffeeMug()), AnyCup(CoffeeCup())]
let cupboardTwo: [AnyCup<Tea>] = [AnyCup(TeaMug()), AnyCup(TeaMug())]
 
// Return the generic type in a function (explicitely or implicitely)
 
func giveMeACupOfCoffee() -> AnyCup<Coffee>? {
    return cupboardOne.first
}
 
// Use the generic type as an input to a function
 
func describe(cup: AnyCup<Coffee>) {
    print(cup.description)
}
 
// Store a generic property
 
struct myShelf<Content> {
    var cup: AnyCup<Content>
}

Conclusion

While I am still not very comfortable with this pattern (a lot of boilerplate, difficult to fully grasp), I will give it a go as it seems very general.   You can also refer to two other excellent blog posts by Realm and by The Big Nerd Ranch basically outlining the same architecture and source of inspiration for this writing.

https://blog.jayway.com/2017/05/12/a-type-erasure-pattern-that-works-in-swift/

A “Type Erasure” Pattern that Works in Swift:类型域的转换的更多相关文章

  1. Breaking Down Type Erasure in Swift

    Type Erasure Pattern We can use the type erasure pattern to combine both generic type parameters and ...

  2. 关于type erasure

    哇,好久没有写blog了,再不写的话,blog的秘密都要忘记了,嘿嘿. 最近在试着参与一个开源项目,名字叫avim(A Vibrate IM),别想多了哟.地址是:https://github.com ...

  3. Java魔法堂:解读基于Type Erasure的泛型

    一.前言 还记得JDK1.4时遍历列表的辛酸吗?我可是记忆犹新啊,那时因项目需求我从C#转身到Java的怀抱,然后因JDK1.4少了泛型这样语法糖(还有自动装箱.拆箱),让我受尽苦头啊,不过也反映自己 ...

  4. java,swift,oc互相转换,html5 web开发跨平台

    java,swift,oc互相转换,html5 web开发跨平台 写一个java->swift的程序,这个程序是做跨平台系统的核心部分swift和oc到java也在考虑之列Swift->J ...

  5. Type Erasure with Pokemon---swift的类型擦除

    我感觉这个是swift的设计缺陷. 类型擦除:解决泛型类型作为公用类型的问题 是抽象的公用机制的一种实现方式. 1)类型擦除并不能解决类型不一致的兼容问题,只能解决类似继承一致性的兼容问题. 2)擦除 ...

  6. [ios][swift]Swift类型之间转换

    http://www.ruanman.net/swift/learn/4741.html

  7. Swift中如何转换不同类型的Mutable指针

    在Swift中我们拥有强大高级逻辑抽象能力的同时,低级底层操作被刻意的限制了.但是有些情况下我们仍然想做一些在C语言中的hack工作,下面本猫就带大家看一看如何做这样的事. hacking is ha ...

  8. [Swift]LeetCode756. 金字塔转换矩阵 | Pyramid Transition Matrix

    We are stacking blocks to form a pyramid. Each block has a color which is a one letter string, like ...

  9. [Swift]字符串大小写转换,同时实现本地化或设置语言环境

    在NSString中提供了3种字符串大小写转换方式:1. 转换字符串大小写2. 转换字符串大小写,并实现本地化3. 转换字符串大小写,并设置语言环境. 一. 转换字符串大小写如果只是想单纯的将字符串进 ...

随机推荐

  1. [转]WordPress 主题教程 #2:模板文件和模板

    本文转自:http://blog.wpjam.com/m/wp-theme-lesson-2-template-files-and-templates/ 模板文件(template files)和模板 ...

  2. Razor 中的@rendersection

    在使用布局页时,可以指定页面中某处的渲染,具体的用@rendersection来做.如在布局页中要渲染一段自定义的脚本, @RenderSection("scripts", req ...

  3. [日常] Go语言圣经-函数多返回值习题

    Go语言圣经-函数多返回值1.在Go中,一个函数可以返回多个值2.许多标准库中的函数返回2个值,一个是期望得到的返回值,另一个是函数出错时的错误信息3.如果一个函数将所有的返回值都显示的变量名,那么该 ...

  4. 【JVM】5、JVM内存管理机制

    转自:http://blog.csdn.net/lengyuhong/article/details/5953544 近期看了看Java内存泄露的一些案例,跟原来的几个哥们讨论了一下,深入研究发现JV ...

  5. Charlie's Change(完全背包+路径记忆)

    Charlie's Change Time Limit: 1000MS   Memory Limit: 30000K Total Submissions: 3176   Accepted: 913 D ...

  6. 存储过程TYPE类型参数调试

    当我们写完一段存储过程后,必然需要调试运行一下写的代码是否能成功运行,当参数是字符,数字或日期时,可以直接在测试窗口输入值.但是类型如果是定义好的TYPE,就无法简单的输入. 一.自定义TYPE TY ...

  7. POJ2478(SummerTrainingDay04-E 欧拉函数)

    Farey Sequence Time Limit: 1000MS   Memory Limit: 65536K Total Submissions: 16927   Accepted: 6764 D ...

  8. 排序算法(6)--exchang Sorting--交换排序[1]--Bubble Sort--冒泡排序

    1.基本思想 冒泡排序的基本思想是,对相邻的元素进行两两比较,顺序相反则进行交换,这样,每一趟会将最小或最大的元素“浮”到顶端,最终达到完全有序 2.实现原理 冒泡排序是一种简单的排序算法,根据顺序两 ...

  9. AutoCAD.net-错误消息大全

    case Acad::eOk:lstrcpy(Glb_AcadErrorInfo,_T("正确"));break;case Acad::eNotImplementedYet:lst ...

  10. Nginx控制客户端请求的速率

    使用ngx_http_limit_req_module模块的两个参数 ngx_http_limit_req_module模块用于限制每个IP访问每个定义key的请求速率 1.limit_req_zon ...