C++ 对象的内存布局（下）

C++ 对象的内存布局(下)<?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" />

 

 

 

陈皓

 

http://blog.csdn.net/haoel

 

 

＜＜＜点击这里查看上篇

 

重复继承

 

 

下面我们再来看看，发生重复继承的情况。所谓重复继承，也就是某个基类被间接地重复继承了多次。

 

 

下图是一个继承图，我们重载了父类的f()函数。

 

 

<?xml:namespace prefix = v ns = "urn:schemas-microsoft-com:vml" /><shapetype id="_x0000_t75" stroked="f" filled="f" path="m@4@5l@4@11@9@11@9@5xe" o:preferrelative="t" o:spt="75" coordsize="21600,21600"><stroke joinstyle="miter"></stroke><formulas><f eqn="if lineDrawn pixelLineWidth 0"></f><f eqn="sum @0 1 0"></f><f eqn="sum 0 0 @1"></f><f eqn="prod @2 1 2"></f><f eqn="prod @3 21600 pixelWidth"></f><f eqn="prod @3 21600 pixelHeight"></f><f eqn="sum @0 0 1"></f><f eqn="prod @6 1 2"></f><f eqn="prod @7 21600 pixelWidth"></f><f eqn="sum @8 21600 0"></f><f eqn="prod @7 21600 pixelHeight"></f><f eqn="sum @10 21600 0"></f></formulas><path o:connecttype="rect" gradientshapeok="t" o:extrusionok="f"></path><lock aspectratio="t" v:ext="edit"></lock></shapetype><shape id="_x0000_i1025" style="WIDTH: 223.5pt; HEIGHT: 294pt" o:ole="" type="#_x0000_t75"><imagedata o:title="" src="file:///C:%5CDOCUME~1%5CHAO~1.CHE%5CLOCALS~1%5CTemp%5Cmsohtml1%5C13%5Cclip_image001.emz"></imagedata></shape>

 

其类继承的源代码如下所示。其中，每个类都有两个变量，一个是整形（4字节），一个是字符（1字节），而且还有自己的虚函数，自己overwrite父类的虚函数。如子类D中，f()覆盖了超类的函数， f1() 和f2() 覆盖了其父类的虚函数，Df()为自己的虚函数。

 

 

class B

 

{

 

public:

 

int ib;

 

char cb;

 

public:

 

B():ib(0),cb('B') {}

 

 

virtual void f() { cout << "B::f()" << endl;}

 

virtual void Bf() { cout << "B::Bf()" << endl;}

 

};

 

class B1 : public B

 

{

 

public:

 

int ib1;

 

char cb1;

 

public:

 

B1():ib1(11),cb1('1') {}

 

 

virtual void f() { cout << "B1::f()" << endl;}

 

virtual void f1() { cout << "B1::f1()" << endl;}

 

virtual void Bf1() { cout << "B1::Bf1()" << endl;}

 

 

};

 

class B2: public B

 

{

 

public:

 

int ib2;

 

char cb2;

 

public:

 

B2():ib2(12),cb2('2') {}

 

 

virtual void f() { cout << "B2::f()" << endl;}

 

virtual void f2() { cout << "B2::f2()" << endl;}

 

virtual void Bf2() { cout << "B2::Bf2()" << endl;}

 

 

};

 

 

class D : public B1, public B2

 

{

 

public:

 

int id;

 

char cd;

 

public:

 

D():id(100),cd('D') {}

 

 

virtual void f() { cout << "D::f()" << endl;}

 

virtual void f1() { cout << "D::f1()" << endl;}

 

virtual void f2() { cout << "D::f2()" << endl;}

 

virtual void Df() { cout << "D::Df()" << endl;}

 

 

};

我们用来存取子类内存布局的代码如下所示：（在VC++ 2003和G++ 3.4.4下）

typedef void(*Fun)(void);

 

int** pVtab = NULL;

 

Fun pFun = NULL;

 

 

D d;

 

pVtab = (int**)&d;

 

cout << "[0] D::B1::_vptr->" << endl;

 

pFun = (Fun)pVtab[0][0];

 

cout << " [0] "; pFun();

 

pFun = (Fun)pVtab[0][1];

 

cout << " [1] "; pFun();

 

pFun = (Fun)pVtab[0][2];

 

cout << " [2] "; pFun();

 

pFun = (Fun)pVtab[0][3];

 

cout << " [3] "; pFun();

 

pFun = (Fun)pVtab[0][4];

 

cout << " [4] "; pFun();

 

pFun = (Fun)pVtab[0][5];

 

cout << " [5] 0x" << pFun << endl;

 

 

cout << "[1] B::ib = " << (int)pVtab[1] << endl;

 

cout << "[2] B::cb = " << (char)pVtab[2] << endl;

 

cout << "[3] B1::ib1 = " << (int)pVtab[3] << endl;

 

cout << "[4] B1::cb1 = " << (char)pVtab[4] << endl;

 

 

cout << "[5] D::B2::_vptr->" << endl;

 

pFun = (Fun)pVtab[5][0];

 

cout << " [0] "; pFun();

 

pFun = (Fun)pVtab[5][1];

 

cout << " [1] "; pFun();

 

pFun = (Fun)pVtab[5][2];

 

cout << " [2] "; pFun();

 

pFun = (Fun)pVtab[5][3];

 

cout << " [3] "; pFun();

 

pFun = (Fun)pVtab[5][4];

 

cout << " [4] 0x" << pFun << endl;

 

 

cout << "[6] B::ib = " << (int)pVtab[6] << endl;

 

cout << "[7] B::cb = " << (char)pVtab[7] << endl;

 

cout << "[8] B2::ib2 = " << (int)pVtab[8] << endl;

 

cout << "[9] B2::cb2 = " << (char)pVtab[9] << endl;

 

 

cout << "[10] D::id = " << (int)pVtab[10] << endl;

 

cout << "[11] D::cd = " << (char)pVtab[11] << endl;

 

 

程序运行结果如下：

GCC 3.4.4	VC++ 2003
[0] D::B1::_vptr->   [0] D::f()   [1] B::Bf()   [2] D::f1()   [3] B1::Bf1()   [4] D::f2()   [5] 0x1   [1] B::ib = 0   [2] B::cb = B   [3] B1::ib1 = 11   [4] B1::cb1 = 1   [5] D::B2::_vptr->   [0] D::f()   [1] B::Bf()   [2] D::f2()   [3] B2::Bf2()   [4] 0x0   [6] B::ib = 0   [7] B::cb = B   [8] B2::ib2 = 12   [9] B2::cb2 = 2   [10] D::id = 100   [11] D::cd = D	[0] D::B1::_vptr->   [0] D::f()   [1] B::Bf()   [2] D::f1()   [3] B1::Bf1()   [4] D::Df()   [5] 0x00000000   [1] B::ib = 0   [2] B::cb = B   [3] B1::ib1 = 11   [4] B1::cb1 = 1   [5] D::B2::_vptr->   [0] D::f()   [1] B::Bf()   [2] D::f2()   [3] B2::Bf2()   [4] 0x00000000   [6] B::ib = 0   [7] B::cb = B   [8] B2::ib2 = 12   [9] B2::cb2 = 2   [10] D::id = 100   [11] D::cd = D

下面是对于子类实例中的虚函数表的图：

 

 

<shape id="_x0000_i1027" style="WIDTH: 315pt; HEIGHT: 240pt" type="#_x0000_t75"><imagedata o:title="" src="file:///C:%5CDOCUME~1%5CHAO~1.CHE%5CLOCALS~1%5CTemp%5Cmsohtml1%5C13%5Cclip_image003.png"><font color="#000000" size="3"></font></imagedata></shape>

 

 

 

我们可以看见，最顶端的父类B其成员变量存在于B1和B2中，并被D给继承下去了。而在D中，其有B1和B2的实例，于是B的成员在D的实例中存在两份，一份是B1继承而来的，另一份是B2继承而来的。所以，如果我们使用以下语句，则会产生二义性编译错误：

 

 

D d;

 

d.ib = 0;//二义性错误

 

d.B1::ib = 1; //正确

 

d.B2::ib = 2; //正确

 

注意，上面例程中的最后两条语句存取的是两个变量。虽然我们消除了二义性的编译错误，但B类在D中还是有两个实例，这种继承造成了数据的重复，我们叫这种继承为重复继承。重复的基类数据成员可能并不是我们想要的。所以，C++引入了虚基类的概念。

 

 

 

 

钻石型多重虚拟继承

 

 

虚拟继承的出现就是为了解决重复继承中多个间接父类的问题的。钻石型的结构是其最经典的结构。也是我们在这里要讨论的结构：

 

 

上述的“重复继承”只需要把B1和B2继承B的语法中加上virtual 关键，就成了虚拟继承，其继承图如下所示：

 

 

 

 

<shape id="_x0000_i1026" style="WIDTH: 221.25pt; HEIGHT: 302.25pt" o:ole="" type="#_x0000_t75"><imagedata o:title="" src="file:///C:%5CDOCUME~1%5CHAO~1.CHE%5CLOCALS~1%5CTemp%5Cmsohtml1%5C13%5Cclip_image005.emz"><font color="#000000" size="3"></font></imagedata></shape>

 

上图和前面的“重复继承”中的类的内部数据和接口都是完全一样的，只是我们采用了虚拟继承：其省略后的源码如下所示：

 

 

class B {……};

 

class B1 : virtual public B{……};

 

class B2: virtual public B{……};

 

class D : public B1, public B2{ …… };

 

 

 

在查看D之前，我们先看一看单一虚拟继承的情况。下面是一段在VC++2003下的测试程序：（因为VC++和GCC的内存而局上有一些细节上的不同，所以这里只给出VC++的程序，GCC下的程序大家可以根据我给出的程序自己仿照着写一个去试一试）：

 

 

int** pVtab = NULL;

 

Fun pFun = NULL;

 

 

B1 bb1;

 

 

pVtab = (int**)&bb1;

 

cout << "[0] B1::_vptr->" << endl;

 

pFun = (Fun)pVtab[0][0];

 

cout << " [0] ";

 

pFun(); //B1::f1();

 

cout << " [1] ";

 

pFun = (Fun)pVtab[0][1];

 

pFun(); //B1::bf1();

 

cout << " [2] ";

 

cout << pVtab[0][2] << endl;

 

 

cout << "[1] = 0x";

 

cout << (int*)*((int*)(&bb1)+1) <<endl; //B1::ib1

 

cout << "[2] B1::ib1 = ";

 

cout << (int)*((int*)(&bb1)+2) <<endl; //B1::ib1

 

cout << "[3] B1::cb1 = ";

 

cout << (char)*((int*)(&bb1)+3) << endl; //B1::cb1

 

 

cout << "[4] = 0x";

 

cout << (int*)*((int*)(&bb1)+4) << endl; //NULL

 

 

cout << "[5] B::_vptr->" << endl;

 

pFun = (Fun)pVtab[5][0];

 

cout << " [0] ";

 

pFun(); //B1::f();

 

pFun = (Fun)pVtab[5][1];

 

cout << " [1] ";

 

pFun(); //B::Bf();

 

cout << " [2] ";

 

cout << "0x" << (Fun)pVtab[5][2] << endl;

 

 

cout <<