news 2026/4/17 19:36:14

c++智能指针转化:static_pointer_cast、dynamic_pointer_cast、const_pointer_cast、reinterpret_pointer_cast 相关函数梳

作者头像

张小明

前端开发工程师

1.2k 24
文章封面图
c++智能指针转化:static_pointer_cast、dynamic_pointer_cast、const_pointer_cast、reinterpret_pointer_cast 相关函数梳


当我们用“裸”指针进行类层次上的上下行转换时,可以使用dynamic_cast。当然我们也可以使用static_cast,只是dynamic_cast在进行下行转换的时候(即基类到派生类)具有类型检查功能,而static_cast没有。因此存在安全问题。

当我们使用智能指针时,如果需要进行类层次上的上下行转换时,可以使用std::static_pointer_cast()、std::dynamic_pointer_cast、std::const_pointer_cast()和std::reinterpret_pointer_cast()。它们的功能和std::static_cast()、std::dynamic_cast、std::const_cast()和std::reinterpret_cast()类似,只不过转换的是智能指针std::shared_ptr,返回的也是std::shared_ptr类型。

1、std::static_pointer_cast():当指针是智能指针时候,向上转换,用static_cast 则转换不了,此时需要使用static_pointer_cast。

2、std::dynamic_pointer_cast():当指针是智能指针时候,向下转换,用dynamic_cast 则转换不了,此时需要使用dynamic_pointer_cast(此处注意:base基类需要至少有一个virtual成员函数(即多态类型)才能允许动态强制转换,否则编译报错)。

3、std::const_pointer_cast():功能与std::const_cast()类似

4、std::reinterpret_pointer_cast():功能与std::reinterpret_cast()类似

Defined in header <memory>


template< class T, class U >
std::shared_ptr<T> static_pointer_cast( const std::shared_ptr<U>& r ) noexcept;

(1) (since C++11)
template< class T, class U >
std::shared_ptr<T> static_pointer_cast( std::shared_ptr<U>&& r ) noexcept;

(2) (since C++20)
template< class T, class U >
std::shared_ptr<T> dynamic_pointer_cast( const std::shared_ptr<U>& r ) noexcept;

(3) (since C++11)
template< class T, class U >
std::shared_ptr<T> dynamic_pointer_cast( std::shared_ptr<U>&& r ) noexcept;

(4) (since C++20)
template< class T, class U >
std::shared_ptr<T> const_pointer_cast( const std::shared_ptr<U>& r ) noexcept;

(5) (since C++11)
template< class T, class U >
std::shared_ptr<T> const_pointer_cast( std::shared_ptr<U>&& r ) noexcept;

(6) (since C++20)
template< class T, class U >
std::shared_ptr<T> reinterpret_pointer_cast( const std::shared_ptr<U>& r ) noexcept;

(7) (since C++17)
template< class T, class U >
std::shared_ptr<T> reinterpret_pointer_cast( std::shared_ptr<U>&& r ) noexcept;

(8) (since C++20)
基类和派生类的智能指针转换要使用std::dynamic_pointer_cast和std::static_pointer_cast。由于std::dynamic_pointer_cast和dynamic_cast原理一样,std::static_pointer_cast和static_cast原理一样

Creates a new instance of std::shared_ptr whose stored pointer is obtained from r's stored pointer using a cast expression.

If r is empty, so is the new shared_ptr (but its stored pointer is not necessarily null). Otherwise, the new shared_ptr will share ownership with the initial value of r, except that it is empty if the dynamic_cast performed by dynamic_pointer_cast returns a null pointer.

Let Y be typename std::shared_ptr<T>::element_type, then the resulting std::shared_ptr's stored pointer will be obtained by evaluating, respectively:

1-2) static_cast<Y*>(r.get()).

3-4) dynamic_cast<Y*>(r.get()) (If the result of the dynamic_cast is a null pointer value, the returned shared_ptr will be empty.)

5-6) const_cast<Y*>(r.get()).

7-8) reinterpret_cast<Y*>(r.get())

The behavior of these functions is undefined unless the corresponding cast from U* to T* is well formed:

1-2) The behavior is undefined unless static_cast<T*>((U*)nullptr) is well formed.

3-4) The behavior is undefined unless dynamic_cast<T*>((U*)nullptr) is well formed.

5-6) The behavior is undefined unless const_cast<T*>((U*)nullptr) is well formed.

7-8) The behavior is undefined unless reinterpret_cast<T*>((U*)nullptr) is well formed.

After calling the rvalue overloads (2,4,6,8), r is empty and r.get() == nullptr, except that r is not modified for dynamic_pointer_cast (4) if the dynamic_cast fails.

(since C++20)
Parameters
r - The pointer to convert
Notes
The expressions std::shared_ptr<T>(static_cast<T*>(r.get())), std::shared_ptr<T>(dynamic_cast<T*>(r.get())) and std::shared_ptr<T>(const_cast<T*>(r.get())) might seem to have the same effect, but they all will likely result in undefined behavior, attempting to delete the same object twice!

Possible implementation
1、std::static_pointer_cast():

template< class T, class U > std::shared_ptr<T> static_pointer_cast( const std::shared_ptr<U>& r ) noexcept { auto p = static_cast<typename std::shared_ptr<T>::element_type*>(r.get()); return std::shared_ptr<T>(r, p); }


2、std::dynamic_pointer_cast()

template< class T, class U > std::shared_ptr<T> dynamic_pointer_cast( const std::shared_ptr<U>& r ) noexcept { if (auto p = dynamic_cast<typename std::shared_ptr<T>::element_type*>(r.get())) { return std::shared_ptr<T>(r, p); } else { return std::shared_ptr<T>(); } }


3、std::const_pointer_cast()

template< class T, class U > std::shared_ptr<T> const_pointer_cast( const std::shared_ptr<U>& r ) noexcept { auto p = const_cast<typename std::shared_ptr<T>::element_type*>(r.get()); return std::shared_ptr<T>(r, p); }


4、std::reinterpret_pointer_cast()

template< class T, class U > std::shared_ptr<T> reinterpret_pointer_cast( const std::shared_ptr<U>& r ) noexcept { auto p = reinterpret_cast<typename std::shared_ptr<T>::element_type*>(r.get()); return std::shared_ptr<T>(r, p); }


使用示例:

#include <iostream> #include <memory> struct Base { int a; virtual void f() const { std::cout << "I am base!\n";} virtual ~Base(){} }; struct Derived : Base { void f() const override { std::cout << "I am derived!\n"; } ~Derived(){} }; int main(){ auto basePtr = std::make_shared<Base>(); std::cout << "Base pointer says: "; basePtr->f(); auto derivedPtr = std::make_shared<Derived>(); std::cout << "Derived pointer says: "; derivedPtr->f(); // static_pointer_cast to go up class hierarchy basePtr = std::static_pointer_cast<Base>(derivedPtr); std::cout << "Base pointer to derived says: "; basePtr->f(); // dynamic_pointer_cast to go down/across class hierarchy auto downcastedPtr = std::dynamic_pointer_cast<Derived>(basePtr); if(downcastedPtr) { std::cout << "Downcasted pointer says: "; downcastedPtr->f(); } // All pointers to derived share ownership std::cout << "Pointers to underlying derived: " << derivedPtr.use_count() << "\n"; } Output: Base pointer says: I am base! Derived pointer says: I am derived! Base pointer to derived says: I am derived! Downcasted pointer says: I am derived! Pointers to underlying derived: 3 示例2 #include <iostream> // std::cout std::endl #include <memory> // std::shared_ptr std::dynamic_pointer_cast std::static_pointer_cast class base { public: virtual ~base(void) = default; }; class derived : public base { }; class test : public base { }; int main(void) { std::cout << std::boolalpha; // 两个不同的派生类对象 auto derivedobj = std::make_shared<derived>(); auto testobj = std::make_shared<test>(); // 隐式转换 derived->base std::shared_ptr<base> pointer1 = derivedobj; // static_pointer_cast derived->base auto pointer2 = std::static_pointer_cast<base>(derivedobj); // dynamic_pointer_cast base->derived auto pointer3 = std::dynamic_pointer_cast<derived>(pointer1); std::cout << (pointer3 == nullptr) << std::endl; // dynamic_pointer_cast base->derived auto pointer4 = std::dynamic_pointer_cast<test>(pointer1); std::cout << (pointer4 == nullptr) << std::endl; return 0; }


输出结果:

false
true


std::reinterpret_pointer_cast()和std::const_pointer_cast()示例代码:

#include <memory> #include <cassert> #include <cstdint> int main() { std::shared_ptr<int> foo; std::shared_ptr<const int> bar; foo = std::make_shared<int>(10); bar = std::const_pointer_cast<const int>(foo); std::cout << "*bar: " << *bar << std::endl; *foo = 20; std::cout << "*bar: " << *bar << std::endl; std::shared_ptr<std::int8_t[]> p(new std::int8_t[4]{1, 1, 1, 1}); std::shared_ptr<std::int32_t[]> q = std::reinterpret_pointer_cast<std::int32_t[]>(p); std::int32_t r = q[0]; std::int32_t x = (1 << 8) | (1 << 16) | (1 << 24) | 1; assert(r == x); return 0; }


输出:

*bar: 10
*bar: 20

版权声明: 本文来自互联网用户投稿,该文观点仅代表作者本人,不代表本站立场。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。如若内容造成侵权/违法违规/事实不符,请联系邮箱:809451989@qq.com进行投诉反馈,一经查实,立即删除!
网站建设 2026/4/17 23:59:24

Linux密码管理深度解析:passwd与chpasswd的底层机制对比

摘要 本文从Linux安全专家视角&#xff0c;深入剖析passwd与chpasswd命令在密码设置中的底层原理差异&#xff0c;涵盖PAM模块调用机制、密码存储格式、安全考量及实际操作实例&#xff0c;为系统管理员提供全面的技术参考。 1. 核心概览&#xff1a;两种密码管理工具的本质区…

作者头像 李华
网站建设 2026/4/18 8:17:01

Vulkan教程(二十一):顶点输入描述:Vulkan 顶点缓冲的数据格式定义

目录 一、修改顶点着色器:接收外部顶点数据 二、定义 CPU 端顶点数据结构 2.1 引入依赖与定义结构体 2.2 定义顶点数据数组 三、顶点绑定描述(Binding Description) 核心参数解析 四、顶点属性描述(Attribute Description) 4.1 核心参数解析 五、配置图形管线的顶…

作者头像 李华
网站建设 2026/4/18 8:04:51

跨领域Agent集成困局破解(基于IEEE与ISO最新标准解读)

第一章&#xff1a;跨领域 Agent 接口标准的演进与挑战随着人工智能与分布式系统的发展&#xff0c;跨领域 Agent 之间的互操作性成为关键技术瓶颈。为实现不同架构、行业和协议下的智能体协同&#xff0c;接口标准化进程不断推进&#xff0c;但同时也面临语义异构、安全边界与…

作者头像 李华
网站建设 2026/4/18 8:18:09

k8s-pod驱逐

一、Pod驱逐1、为什么要有驱逐pod.spec.containers[].resources中会存在cpu或memory的request和limit。即该pod请求的最小资源和Node结点可以给的最大资源。当一个容器的cpu使用率超过limit时会被进行流控&#xff0c;而当内存超过limit时则会被oom_kill。完全依赖于oom_kill并…

作者头像 李华