我对大多数OOP理论都有很好的理解,但最让我困惑的是虚拟析构函数。

我以为析构函数总是被调用,不管是什么,也不管是链中的每个对象。

你打算什么时候让它们虚拟化?为什么?


当前回答

将所有析构函数都设为虚拟,除非你有充分的理由不这样做。

否则会发生这样的邪恶:

假设您有一个包含Apple和Orange对象的Fruit指针数组。

从Fruit对象集合中删除时,除非~Fruit()是虚拟的,否则无法调用~Apple()和~Orange()。

正确完成示例:

#include <iostream>
using namespace std;
struct Fruit { // good
  virtual ~Fruit() { cout << "peel or core should have been tossed" << endl; } 
};
struct Apple:  Fruit { virtual ~Apple()  {cout << "toss core" << endl; } };
struct Orange: Fruit { virtual ~Orange() {cout << "toss peel" << endl; } };

int main() { 
  Fruit *basket[]={ new Apple(), new Orange() };
  for (auto fruit: basket) delete fruit;
};

正品产出量

toss core
peel or core should have been tossed
toss peel
peel or core should have been tossed

错误示例:

#include <iostream>
using namespace std;
struct Fruit { // bad 
  ~Fruit() { cout << "peel or core should have been tossed" << endl; } 
};
struct Apple:  Fruit { virtual ~Apple()  {cout << "toss core" << endl; } };
struct Orange: Fruit { virtual ~Orange() {cout << "toss peel" << endl; } };

int main() { 
  Fruit *basket[]={ new Apple(), new Orange() };
  for (auto fruit: basket) delete fruit;
};

不良输出

peel or core should have been tossed
peel or core should have been tossed

(注意:为了简洁起见,我使用了struct,通常使用class并指定public)

其他回答

我认为这个问题的核心是关于虚拟方法和多态性,而不是具体的析构函数。下面是一个更清晰的例子:

class A
{
public:
    A() {}
    virtual void foo()
    {
        cout << "This is A." << endl;
    }
};

class B : public A
{
public:
    B() {}
    void foo()
    {
        cout << "This is B." << endl;
    }
};

int main(int argc, char* argv[])
{
    A *a = new B();
    a->foo();
    if(a != NULL)
    delete a;
    return 0;
}

将打印出:

This is B.

如果没有虚拟,它将打印出:

This is A.

现在您应该了解何时使用虚拟析构函数。

虚拟构造函数是不可能的,但虚拟析构函数是可能的。让我们做个实验。。。。。。。

#include <iostream>

using namespace std;

class Base
{
public:
    Base(){
        cout << "Base Constructor Called\n";
    }
    ~Base(){
        cout << "Base Destructor called\n";
    }
};

class Derived1: public Base
{
public:
    Derived1(){
        cout << "Derived constructor called\n";
    }
    ~Derived1(){
        cout << "Derived destructor called\n";
    }
};

int main()
{
    Base *b = new Derived1();
    delete b;
}

上述代码输出以下内容:

Base Constructor Called
Derived constructor called
Base Destructor called

派生对象的构造遵循构造规则,但当我们删除“b”指针(基指针)时,我们发现只有基析构函数被调用。但这绝不能发生。为了做适当的事情,我们必须使基析构函数虚拟化。现在让我们看看以下情况:

#include <iostream>

using namespace std;

class Base
{ 
public:
    Base(){
        cout << "Base Constructor Called\n";
    }
    virtual ~Base(){
        cout << "Base Destructor called\n";
    }
};

class Derived1: public Base
{
public:
    Derived1(){
        cout << "Derived constructor called\n";
    }
    ~Derived1(){
        cout << "Derived destructor called\n";
    }
};

int main()
{
    Base *b = new Derived1();
    delete b;
}

输出变化如下:

Base Constructor Called
Derived Constructor called
Derived destructor called
Base destructor called

因此,基指针的销毁(对派生对象进行分配!)遵循销毁规则,即首先是派生指针,然后是基指针。另一方面,没有什么像虚拟构造函数。

我认为讨论“未定义”行为,或者至少讨论在通过没有虚拟析构函数的基类(/struct)删除时可能发生的“崩溃”未定义行为,或者更准确地说,没有vtable,是有益的。下面的代码列出了一些简单的结构(类也是如此)。

#include <iostream>
using namespace std;

struct a
{
    ~a() {}

    unsigned long long i;
};

struct b : a
{
    ~b() {}

    unsigned long long j;
};

struct c : b
{
    ~c() {}

    virtual void m3() {}

    unsigned long long k;
};

struct d : c
{
    ~d() {}

    virtual void m4() {}

    unsigned long long l;
};

int main()
{
    cout << "sizeof(a): " << sizeof(a) << endl;
    cout << "sizeof(b): " << sizeof(b) << endl;
    cout << "sizeof(c): " << sizeof(c) << endl;
    cout << "sizeof(d): " << sizeof(d) << endl;

    // No issue.

    a* a1 = new a();
    cout << "a1: " << a1 << endl;
    delete a1;

    // No issue.

    b* b1 = new b();
    cout << "b1: " << b1 << endl;
    cout << "(a*) b1: " << (a*) b1 << endl;
    delete b1;

    // No issue.

    c* c1 = new c();
    cout << "c1: " << c1 << endl;
    cout << "(b*) c1: " << (b*) c1 << endl;
    cout << "(a*) c1: " << (a*) c1 << endl;
    delete c1;

    // No issue.

    d* d1 = new d();
    cout << "d1: " << d1 << endl;
    cout << "(c*) d1: " << (c*) d1 << endl;
    cout << "(b*) d1: " << (b*) d1 << endl;
    cout << "(a*) d1: " << (a*) d1 << endl;
    delete d1;

    // Doesn't crash, but may not produce the results you want.

    c1 = (c*) new d();
    delete c1;

    // Crashes due to passing an invalid address to the method which
    // frees the memory.

    d1 = new d();
    b1 = (b*) d1;
    cout << "d1: " << d1 << endl;
    cout << "b1: " << b1 << endl;
    delete b1;  

/*

    // This is similar to what's happening above in the "crash" case.

    char* buf = new char[32];
    cout << "buf: " << (void*) buf << endl;
    buf += 8;
    cout << "buf after adding 8: " << (void*) buf << endl;
    delete buf;
*/
}

我并不是建议你是否需要虚拟析构函数,尽管我认为一般来说,拥有它们是一个很好的做法。我只是指出了如果基类(/struct)没有vtable,而派生类(/struck)有vtable,并且通过基类(/ststruct)指针删除对象,那么可能会导致崩溃的原因。在这种情况下,传递给堆的空闲例程的地址是无效的,因此是崩溃的原因。

如果运行上述代码,您将清楚地看到问题发生的时间。当基类(/struct)的this指针与派生类(/struct)的this指示器不同时,您将遇到此问题。在上面的示例中,结构a和b没有vtables。结构c和d确实有vtables。因此,指向c或d对象实例的a或b指针将被修复以说明vtable。如果传递此a或b指针进行删除,则会由于地址对堆的空闲例程无效而崩溃。

如果计划从基类指针中删除具有vtable的派生实例,则需要确保基类具有vtable。一种方法是添加一个虚拟析构函数,您可能无论如何都希望它能正确地清理资源。

将所有析构函数都设为虚拟,除非你有充分的理由不这样做。

否则会发生这样的邪恶:

假设您有一个包含Apple和Orange对象的Fruit指针数组。

从Fruit对象集合中删除时,除非~Fruit()是虚拟的,否则无法调用~Apple()和~Orange()。

正确完成示例:

#include <iostream>
using namespace std;
struct Fruit { // good
  virtual ~Fruit() { cout << "peel or core should have been tossed" << endl; } 
};
struct Apple:  Fruit { virtual ~Apple()  {cout << "toss core" << endl; } };
struct Orange: Fruit { virtual ~Orange() {cout << "toss peel" << endl; } };

int main() { 
  Fruit *basket[]={ new Apple(), new Orange() };
  for (auto fruit: basket) delete fruit;
};

正品产出量

toss core
peel or core should have been tossed
toss peel
peel or core should have been tossed

错误示例:

#include <iostream>
using namespace std;
struct Fruit { // bad 
  ~Fruit() { cout << "peel or core should have been tossed" << endl; } 
};
struct Apple:  Fruit { virtual ~Apple()  {cout << "toss core" << endl; } };
struct Orange: Fruit { virtual ~Orange() {cout << "toss peel" << endl; } };

int main() { 
  Fruit *basket[]={ new Apple(), new Orange() };
  for (auto fruit: basket) delete fruit;
};

不良输出

peel or core should have been tossed
peel or core should have been tossed

(注意:为了简洁起见,我使用了struct,通常使用class并指定public)

通过指向基类的指针调用析构函数

struct Base {
  virtual void f() {}
  virtual ~Base() {}
};

struct Derived : Base {
  void f() override {}
  ~Derived() override {}
};

Base* base = new Derived;
base->f(); // calls Derived::f
base->~Base(); // calls Derived::~Derived

虚拟析构函数调用与任何其他虚拟函数调用都没有区别。

对于base->f(),调用将被分派到Derived::f()中,对于base->~base()也是如此-它的重写函数-将调用Derived::~Derived()。

间接调用析构函数时也会发生同样的情况,例如delete base;。delete语句将调用base->~base(),该函数将被分派到Derived::~Derived()。

具有非虚拟析构函数的抽象类

若您不打算通过指向其基类的指针删除对象,那个么就不需要使用虚拟析构函数。只需保护它,使其不会被意外调用:

// library.hpp

struct Base {
  virtual void f() = 0;

protected:
  ~Base() = default;
};

void CallsF(Base& base);
// CallsF is not going to own "base" (i.e. call "delete &base;").
// It will only call Base::f() so it doesn't need to access Base::~Base.

//-------------------
// application.cpp

struct Derived : Base {
  void f() override { ... }
};

int main() {
  Derived derived;
  CallsF(derived);
  // No need for virtual destructor here as well.
}