unique_ptr的成员函数在上一篇博客中几乎全部涵盖,其实还有一个很有踢掉,即std::unique_ptr::get_deleter
字面已经很明显了,就获得deleter:
Returns the stored deleter
The stored deleter is a callable object. A functional call to this object with a single argument of member type pointer is expected to delete the managed object, and is automatically called when the unique_ptr is itself destroyed, assigned a new value, or resetted while non-empty.
直接看例子:
#include <iostream>
#include <memory>
class state_deleter { // a deleter class with state
int count_;
public:
state_deleter() : count_(0) {}
template <class T>
void operator()(T* p) {
std::cout << "[deleted #" << ++count_ << "]\n";
delete p;
}
};
int main() {
state_deleter del;
std::unique_ptr<int> p; // uses default deleter
// alpha and beta use independent copies of the deleter:
std::unique_ptr<int, state_deleter> alpha(new int);
std::unique_ptr<int, state_deleter> beta(new int, alpha.get_deleter());
return 0;
}
运行程序,输出结果:
[deleted #1]
[deleted #1]
说明了,在函数结束的时候,alpha beta的析构函数被调用了,而且析构函数是我们自己定义的类。
在上面的代码上加入对alpha beta的操作:
int main() {
state_deleter del;
std::unique_ptr<int> p; // uses default deleter
// alpha and beta use independent copies of the deleter:
std::unique_ptr<int, state_deleter> alpha(new int);
std::unique_ptr<int, state_deleter> beta(new int, alpha.get_deleter());
std::cout << "resetting alpha..."; alpha.reset(new int);
std::cout << "resetting beta..."; beta.reset(new int);
return 0;
}
输出:
resetting alpha…[deleted #1]
resetting beta…[deleted #1]
[deleted #2]
[deleted #2]
说明在reset方法后,也调用了自定义的析构~
再继续添加代码:
#include <iostream>
#include <memory>
class state_deleter { // a deleter class with state
int count_;
public:
state_deleter() : count_(0) {}
template <class T>
void operator()(T* p) {
std::cout << "[deleted #" << ++count_ << "]\n";
delete p;
}
};
int main() {
state_deleter del;
std::unique_ptr<int> p; // uses default deleter
// alpha and beta use independent copies of the deleter:
std::unique_ptr<int, state_deleter> alpha(new int);
std::unique_ptr<int, state_deleter> beta(new int, alpha.get_deleter());
// gamma and delta share the deleter "del" (deleter type is a reference!):
std::unique_ptr<int, state_deleter&> gamma(new int, del);
std::unique_ptr<int, state_deleter&> delta(new int, gamma.get_deleter());
std::cout << "resetting alpha..."; alpha.reset(new int);
std::cout << "resetting beta..."; beta.reset(new int);
std::cout << "resetting gamma..."; gamma.reset(new int);
std::cout << "resetting delta..."; delta.reset(new int);
return 0;
}
输出:
resetting alpha…[deleted #1]
resetting beta…[deleted #1]
resetting gamma…[deleted #1]
resetting delta…[deleted #2]
[deleted #3]
[deleted #4]
[deleted #2]
[deleted #2]
这里我们重点关注delta,这里使用的是按引用传递。
还有一个地方需要说明:
析构的顺序与构造的顺序相反~~
go on:
#include <iostream>
#include <memory>
class state_deleter { // a deleter class with state
int count_;
public:
state_deleter() : count_(0) {}
template <class T>
void operator()(T* p) {
std::cout << "[deleted #" << ++count_ << "]\n";
delete p;
}
};
int main() {
state_deleter del;
std::unique_ptr<int> p; // uses default deleter
// alpha and beta use independent copies of the deleter:
std::unique_ptr<int, state_deleter> alpha(new int);
std::unique_ptr<int, state_deleter> beta(new int, alpha.get_deleter());
// gamma and delta share the deleter "del" (deleter type is a reference!):
std::unique_ptr<int, state_deleter&> gamma(new int, del);
std::unique_ptr<int, state_deleter&> delta(new int, gamma.get_deleter());
std::cout << "resetting alpha..."; alpha.reset(new int);
std::cout << "resetting beta..."; beta.reset(new int);
std::cout << "resetting gamma..."; gamma.reset(new int);
std::cout << "resetting delta..."; delta.reset(new int);
//std::cout << "calling gamma/delta deleter...";
gamma.get_deleter()(new int);
alpha.get_deleter() = state_deleter(); // a brand new deleter for alpha
// additional deletions when unique_ptr objects reach out of scope
// (in inverse order of declaration)
return 0;
}
此时输出:
//输出:
//resetting alpha…[deleted #1]
//resetting beta…[deleted #1]
//resetting gamma…[deleted #1]
//resetting delta…[deleted #2]
//calling gamma / delta deleter…
//[deleted #3]
//[deleted #4]
//[deleted #5]
//[deleted #2]
//[deleted #1]
最后再来一个:
#include <iostream>
#include <memory>
using namespace std;
void deleter(int* ptr) {
delete ptr;
ptr = nullptr;
std::clog << "shared_ptr delete the pointer." << std::endl;
}
int main(void) {
//定义函数类型
typedef void(*tp) (int*);
typedef decltype (deleter)* dp;
using up = void(*) (int*);
std::shared_ptr<int> spi(new int(10), deleter);
std::shared_ptr<int> spi2(new int, deleter);
spi2 = std::make_shared<int>(15);
std::cout << "*spi = " << *spi << std::endl;
std::cout << "*spi2 = " << *spi2 << std::endl;
//unique_ptr是模板函数需要删除器(deleter)类型, 再传入具体的删除器
std::unique_ptr<int, decltype(deleter)*> upi(new int(20), deleter);
std::unique_ptr<int, tp> upi2(new int(25), deleter);
std::unique_ptr<int, dp> upi3(new int(30), deleter);
std::unique_ptr<int, up> upi4(new int(35), deleter);
std::cout << "*upi = " << *upi << std::endl;
std::cout << "*upi2 = " << *upi2 << std::endl;
std::cout << "*upi3 = " << *upi3 << std::endl;
std::cout << "*upi4 = " << *upi4 << std::endl;
return 0;
}
//输出:
//shared_ptr delete the pointer.
//*spi = 10
//* spi2 = 15
//* upi = 20
//* upi2 = 25
//* upi3 = 30
//* upi4 = 35
//shared_ptr delete the pointer.
//shared_ptr delete the pointer.
//shared_ptr delete the pointer.
//shared_ptr delete the pointer.
//shared_ptr delete the pointer.
只是分析这两句:
std::shared_ptr<int> spi2(new int, deleter);
spi2 = std::make_shared<int>(15);
这时候会调用deleter删除器!