Can we make a class constructor virtual in C++ to create polymorphic objects?

  No. C++ being static typed (the purpose of RTTI is different) language, it is meaningless to the C++ compiler to create an object polymorphically. The compiler must be aware of the class type to create the object.

  In other words, what type of object to be created is a compile time decision from C++ compiler perspective. If we make constructor virtual, compiler flags an error. In fact except inline, no other keyword is allowed in the declaration of constructor.

  In practical scenarios(情节) we would need to create a derived class object in a class hierarchy based on some input. Putting in other words, object creation and object type are tightly coupled which forces modifications to extended. The objective of virtual constructor is to decouple object creation from it’s type.

  How can we create required type of object at runtime? For example, see the following sample program.

 1 #include <iostream>
2 using namespace std;
3
4 //// LIBRARY START
5 class Base
6 {
7 public:
8 Base()
9 {
10 }
11 virtual ~Base() // Ensures to invoke actual object destructor
12 {
13 }
14
15 // An interface
16 virtual void DisplayAction() = 0;
17 };
18
19 class Derived1 : public Base
20 {
21 public:
22 Derived1()
23 {
24 cout << "Derived1 created" << endl;
25 }
26
27 ~Derived1()
28 {
29 cout << "Derived1 destroyed" << endl;
30 }
31
32 void DisplayAction()
33 {
34 cout << "Action from Derived1" << endl;
35 }
36 };
37
38 class Derived2 : public Base
39 {
40 public:
41 Derived2()
42 {
43 cout << "Derived2 created" << endl;
44 }
45
46 ~Derived2()
47 {
48 cout << "Derived2 destroyed" << endl;
49 }
50
51 void DisplayAction()
52 {
53 cout << "Action from Derived2" << endl;
54 }
55 };
56
57 //// LIBRARY END
58
59 class User
60 {
61 public:
62
63 // Creates Drived1
64 User() : pBase(0)
65 {
66 // What if Derived2 is required? - Add an if-else ladder (see next sample)
67 pBase = new Derived1();
68 }
69
70 ~User()
71 {
72 if( pBase )
73 {
74 delete pBase;
75 pBase = 0;
76 }
77 }
78
79 // Delegates to actual object
80 void Action()
81 {
82 pBase->DisplayAction();
83 }
84
85 private:
86 Base *pBase;
87 };
88
89 int main()
90 {
91 User *user = new User();
92
93 // Need Derived1 functionality only
94 user->Action();
95
96 delete user;
97 }

  Output:

  Derived1 created
  Action from Derived1
  Derived1 destroyed

  In the above sample, assume that the hierarchy Base, Derived1 and Derived2 are part of library code. The class User is utility class trying to make use of the hierarchy. The main function is consuming Base hierarchy functionality via User class.

  The User class constructor is creating Derived1 object, always. If the User‘s consumer (the main in our case) needs Derived2 functionality, User needs to create "new Derived2()" and it forces recompilation. Recompiling is bad way of design, so we can opt for the following approach.

  Before going into details, let us answer, who will dictate to create either of Derived1 or Derived2 object? Clearly, it is the consumer of User class. The User class can make use of if-else ladder to create either Derived1 or Derived2, as shown in the following sample,

  1 #include <iostream>
2 using namespace std;
3
4 //// LIBRARY START
5 class Base
6 {
7 public:
8 Base() { }
9
10 virtual // Ensures to invoke actual object destructor
11 ~Base() { }
12
13 // An interface
14 virtual void DisplayAction() = 0;
15 };
16
17 class Derived1 : public Base
18 {
19 public:
20 Derived1()
21 {
22 cout << "Derived1 created" << endl;
23 }
24
25 ~Derived1()
26 {
27 cout << "Derived1 destroyed" << endl;
28 }
29
30 void DisplayAction()
31 {
32 cout << "Action from Derived1" << endl;
33 }
34 };
35
36 class Derived2 : public Base
37 {
38 public:
39 Derived2()
40 {
41 cout << "Derived2 created" << endl;
42 }
43
44 ~Derived2()
45 {
46 cout << "Derived2 destroyed" << endl;
47 }
48
49 void DisplayAction()
50 {
51 cout << "Action from Derived2" << endl;
52 }
53 };
54
55 //// LIBRARY END
56
57 class User
58 {
59 public:
60
61 // Creates Derived1 or Derived2 based on input
62 User() : pBase(0)
63 {
64 int input; // ID to distinguish between
65 // Derived1 and Derived2
66
67 cout << "Enter ID (1 or 2): ";
68 cin >> input;
69
70 while( (input != 1) && (input != 2) )
71 {
72 cout << "Enter ID (1 or 2 only): ";
73 cin >> input;
74 }
75
76 if( input == 1 )
77 {
78 pBase = new Derived1;
79 }
80 else
81 {
82 pBase = new Derived2;
83 }
84
85 // What if Derived3 being added to the class hierarchy?
86 }
87
88 ~User()
89 {
90 if( pBase )
91 {
92 delete pBase;
93 pBase = 0;
94 }
95 }
96
97 // Delegates to actual object
98 void Action()
99 {
100 pBase->DisplayAction();
101 }
102
103 private:
104 Base *pBase;
105 };
106
107 int main()
108 {
109 User *user = new User();
110
111 // Need either Derived1 or Derived2 functionality
112 user->Action();
113
114 delete user;
115 }

  The above code is *not* open for extension, an inflexible design. In simple words, if the library updates the Base class hierarchy with new class Derived3. How can the User class creates Derived3 object? One way is to update the if-else ladder that creates Derived3 object based on new input ID 3 as shown below.

 1 #include <iostream>
2 using namespace std;
3
4 class User
5 {
6 public:
7 User() : pBase(0)
8 {
9 // Creates Drived1 or Derived2 based on need
10
11 int input; // ID to distinguish between
12 // Derived1 and Derived2
13
14 cout << "Enter ID (1 or 2): ";
15 cin >> input;
16
17 while( (input != 1) && (input != 2) )
18 {
19 cout << "Enter ID (1 or 2 only): ";
20 cin >> input;
21 }
22
23 if( input == 1 )
24 {
25 pBase = new Derived1;
26 }
27 else if( input == 2 )
28 {
29 pBase = new Derived2;
30 }
31 else
32 {
33 pBase = new Derived3;
34 }
35 }
36
37 ~User()
38 {
39 if( pBase )
40 {
41 delete pBase;
42 pBase = 0;
43 }
44 }
45
46 // Delegates to actual object
47 void Action()
48 {
49 pBase->DisplayAction();
50 }
51
52 private:
53 Base *pBase;
54 };

  The above modification forces the users of User class to recompile, bad (inflexible) design! And won’t close User class from further modifications due to Base extension.

  The problem is with the creation of objects. Addition of new class to the hierarchy forcing dependents of User class to recompile. Can’t we delegate(代表) the action of creating objects to class hierarchy itself or to a function that behaves virtually? By delegating the object creation to class hierarchy (or to a static function) we can avoid the tight coupling between User and Base hierarchy.

  Enough theory, see the following code,

  1 #include <iostream>
2 using namespace std;
3
4 //// LIBRARY START
5 class Base
6 {
7 public:
8
9 // The "Virtual Constructor"
10 static Base *Create(int id);
11
12 Base()
13 {
14 }
15
16 virtual ~Base() // Ensures to invoke actual object destructor
17 {
18 }
19
20 // An interface
21 virtual void DisplayAction() = 0;
22 };
23
24 class Derived1 : public Base
25 {
26 public:
27 Derived1()
28 {
29 cout << "Derived1 created" << endl;
30 }
31
32 ~Derived1()
33 {
34 cout << "Derived1 destroyed" << endl;
35 }
36
37 void DisplayAction()
38 {
39 cout << "Action from Derived1" << endl;
40 }
41 };
42
43 class Derived2 : public Base
44 {
45 public:
46 Derived2()
47 {
48 cout << "Derived2 created" << endl;
49 }
50
51 ~Derived2()
52 {
53 cout << "Derived2 destroyed" << endl;
54 }
55
56 void DisplayAction()
57 {
58 cout << "Action from Derived2" << endl;
59 }
60 };
61
62 class Derived3 : public Base
63 {
64 public:
65 Derived3()
66 {
67 cout << "Derived3 created" << endl;
68 }
69
70 ~Derived3()
71 {
72 cout << "Derived3 destroyed" << endl;
73 }
74
75 void DisplayAction()
76 {
77 cout << "Action from Derived3" << endl;
78 }
79 };
80
81 // We can also declare "Create" outside Base
82 // But it is more relevant to limit it's scope to Base
83 Base *Base::Create(int id)
84 {
85 // Just expand the if-else ladder, if new Derived class is created
86 // User code need not be recompiled to create newly added class objects
87
88 if( id == 1 )
89 {
90 return new Derived1;
91 }
92 else if( id == 2 )
93 {
94 return new Derived2;
95 }
96 else
97 {
98 return new Derived3;
99 }
100 }
101 //// LIBRARY END
102
103 //// UTILITY START
104 class User
105 {
106 public:
107 User() : pBase(0)
108 {
109 // Receives an object of Base heirarchy at runtime
110
111 int input;
112
113 cout << "Enter ID (1, 2 or 3): ";
114 cin >> input;
115
116 while( (input != 1) && (input != 2) && (input != 3) )
117 {
118 cout << "Enter ID (1, 2 or 3 only): ";
119 cin >> input;
120 }
121
122 // Get object from the "Virtual Constructor"
123 pBase = Base::Create(input);
124 }
125
126 ~User()
127 {
128 if( pBase )
129 {
130 delete pBase;
131 pBase = 0;
132 }
133 }
134
135 // Delegates to actual object
136 void Action()
137 {
138 pBase->DisplayAction();
139 }
140
141 private:
142 Base *pBase;
143 };
144
145 //// UTILITY END
146
147 //// Consumer of User (UTILITY) class
148 int main()
149 {
150 User *user = new User();
151
152 // Action required on any of Derived objects
153 user->Action();
154
155 delete user;
156 }

  The User class is independent of object creation. It delegates that responsibility to Base, and provides an input in the form of ID. If the library adds new class Derived4, the library modifier will extend the if-else ladder inside Create to return proper object. Consumers of User need not recompile their code due to extension of Base.

  Note that the function Create used to return different types of Base class objects at runtime. It acts like virtual constructor, also referred as Factory Method in pattern terminology.

  Pattern world demonstrate different ways to implement the above concept. Also there are some potential design issues with the above code. Our objective is to provide some insights into virtual construction, creating objects dynamically based on some input. We have excellent books devoted to the subject, interested reader can refer them for more information.

  这篇文章讲述的是有设计模式中的工厂模式Factory Pattern,这个我不太懂哦。

  Please write comments if you find anything incorrect, or you want to share more information about the topic discussed above.

  转载请注明:http://www.cnblogs.com/iloveyouforever/

  2013-11-26  21:00:57

Advanced C++ | Virtual Constructor的更多相关文章

  1. Java设计模式-工厂方法模式(Virtual Constructor/Polymorphic Factory)

    工厂方法模式(Virtual Constructor/Polymorphic Factory) 工厂方法模式是类的创建模式,又叫做虚拟构造子模式(Virtual Constructor)或者多态性工厂 ...

  2. Advanced C++ | Virtual Copy Constructor

    这个不懂,等看会了再写...

  3. Objective-C设计模式——工厂方法模式virtual constructor(对象创建)

    工厂方法模式 工厂方法模式可以控制对象的创建过程,屏蔽对象创建的细节,可以直接创建出我们所需要的已经配置好的对象. 工厂方法模式定义了创建方法的接口,让子类决定实例化哪一个类,工厂方法模式使得一个类的 ...

  4. 【java设计模式】【创建模式Creational Pattern】工厂方法模式Factory Method Pattern(多态性工厂模式Polymorphic Factory Pattern、虚拟构造子模式Virtual Constructor Pattern)

    public class Test { public static void main(String[] args){ Creator ca=new ConcreteCreatorA(); ca.cr ...

  5. 为什么内联函数,构造函数,静态成员函数不能为virtual函数

    http://blog.csdn.net/freeboy1015/article/details/7635012 为什么内联函数,构造函数,静态成员函数不能为virtual函数? 1> 内联函数 ...

  6. Virtual Table

    C++对象模型——吴泰 C/C++杂记 C++中的虚函数(表)实现机制以及用C语言对其进行的模拟实现 C++ 多继承和虚继承的内存布局 [已翻译100%] (虚继承参考,推荐) 图说C++对象模型:对 ...

  7. MoreEffectiveC++Item35 条款25 将constructor和non-member functions虚化

    1.virtual constructor 在语法上是不可将构造函数声明成虚函数,虚函数用于实现"因类型而异的行为",也就是根据指针或引用所绑定对象的动态类型而调用不同实体.现在所 ...

  8. 工厂方法模式——创建型模式02

    1. 简单工厂模式     在介绍工厂方法模式之前,先介绍一下简单工厂模式.虽然简单工厂模式不属于GoF 23种设计模式,但通常将它作为学习其他工厂模式的入门,并且在实际开发中使用的也较为频繁. (1 ...

  9. C++异常处理:try,catch,throw,finally的用法

    写在前面 所谓异常处理,即让一个程序运行时遇到自己无法处理的错误时抛出一个异常,希望调用者可以发现处理问题. 异常处理的基本思想是简化程序的错误代码,为程序键壮性提供一个标准检测机制. 也许我们已经使 ...

随机推荐

  1. vue事件绑定

    事件绑定 vue官方文档:<https://cn.vuejs.org/v2/guide/events.html#%E4%BA%8B%E4%BB%B6%E4%BF%AE%E9%A5%B0%E7%A ...

  2. Solon & Solon Cloud 1.5.62 发布,轻量级 Java 基础开发框架

    Solon 已有120个生态扩展插件,此次更新主要为细节打磨,且对k8s和docker-compose更友好: 1.插件 solon.coud ,事件总线增加支持本地同主题多订阅模式(以支持同服务内, ...

  3. 使用Adobe Acrobat进行Word转PDF遇到的问题及解决方法

    软件版本:Adobe Acrobat 9 Pro 使用场景:Word转PDF 问题1: 我以为先要在Adobe Acrobat 9 Pro中打开Word文件,然后在执行类似转换/导出操作.但是始终无法 ...

  4. 【JAVA】编程(5)---递归

    作业要求: 利用递归来计算出 从1加到100的数和 : public class 递归 { public static void main(String[] args) { System.out.pr ...

  5. c++学习笔记3(内联函数)

    函数调用是有开销的,调用时需将参数放入栈中,返回地址也要放入,返回时还需从栈中取出,跳转返回地址去执行,需几条语句的时间,如果本身程序代码短,则会显得十分浪费,所以引入了内联函数的机制 写法:在函数前 ...

  6. [源码解析] PyTorch分布式(5) ------ DistributedDataParallel 总述&如何使用

    [源码解析] PyTorch 分布式(5) ------ DistributedDataParallel 总述&如何使用 目录 [源码解析] PyTorch 分布式(5) ------ Dis ...

  7. Python 爬取 豆瓣

    ... import urllib.request import time from bs4 import BeautifulSoup def url_open(url): response = ur ...

  8. 100_第一个vue-cli项目

    目录 什么是vue-cli 主要的功能 需要的环境 Node.js : http://nodejs.cn/download/ Git : https://git-scm.com/downloads 安 ...

  9. Github树型插件--Octotree

    octotree 是一款chrome插件,用于将 GitHub 项目代码以树形格式展示,而且在展示的列表中,我们可以下载指定的文件,而不需要下载整个项目. 官网地址:https://www.octot ...

  10. 单&双&单|双||的区别

    System.out.println(true|true);//true System.out.println(true|false);//true System.out.println(false| ...