> For the complete documentation index, see [llms.txt](https://mqjyl2012.gitbook.io/algorithm/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://mqjyl2012.gitbook.io/algorithm/c-cpp/c++-syntax/auto-increment-attenuation.md). # 自增(++) & 自减(--) ## :pencil2: 1、内置类型 ```cpp ++a a++ a=a+1 a+=1 ``` 对于内置类型，对于现代编译器而言，这四个的效率都是没有区别的，在`VS2019`下查看汇编代码如下： ![](/files/-MEpnPVYE9Csbhkc36u-) 注意：不要在一条语句中使用多个++，因为在不同系统中对这样的情况处理可能不一样，比如y = (4 + x++) + (6 + x++)。这条语句只能保证程序执行到下一条语句之前，x被递增两次，并不能保证4 + x++后立即自增。 ## :pencil2: 2、自定义类型 ### :pen\_fountain: 2.1、a++与++a区别 1. a++是先赋值再自增，++a是先自增再赋值。 2. a++是先用临时对象保存原来的对象，然后对原对象自增，再返回临时对象，不能作为左值；++a是直接对于原对象进行自增，然后返回原对象的**引用**，可以作为左值。 3. 由于要生成临时对象，a++需要调用两次拷贝构造函数与析构函数（将原对象赋给临时对象一次，临时对象以值传递方式返回一次）；++a由于不用生成临时变量，且以引用方式返回，故没有构造与析构的开销，效率更高。 左值一般是可以放在赋值符号左边的值，其在内存中有实体；右值一般只能放在赋值符号右边，不具有内存实体，无法通过取地址获得相应对象。如： ```cpp class Point { int x_; int y_; public: Point(int x = 0, int y = 0); Point(const Point&); ~Point(); Point& operator++(); //前置 const Point operator++(int); //后置，int仅仅为了重载 Point operator+(const Point&); Point& operator+=(const Point&); void DisplayPoint(); }; Point::Point(int x, int y) { x_ = x; y_ = y; cout << "this is constructor" << endl; } Point::Point(const Point& b) { this->x_ = b.x_; this->y_ = b.y_; cout << "this is copy constructor" << endl; } Point::~Point() { cout << "this is destructor" << endl; } Point& Point::operator+=(const Point& _right) { this->x_ += _right.x_; this->y_ += _right.y_; return *this; } Point Point::operator+(const Point& _right) { Point temp; temp.x_ = this->x_ + _right.x_; temp.y_ = this->y_ + _right.y_; return temp; } Point& Point::operator++() { ++x_; ++y_; return *this; } const Point Point::operator++(int) { Point temp(*this); this->x_++; this->y_++; return temp; } void Point::DisplayPoint() { cout << "x: " << this->x_ << endl; cout << "y: " << this->y_ << endl; } ``` #### :gem: 2.1.1、效率检测 ```cpp int main() { Point a(1, 1); cout << endl << "this is a++: " << endl; a++; cout << endl << "this is ++a: " << endl; ++a; return 0; } // 输出 this is constructor this is a++: this is copy constructor this is copy constructor this is destructor this is destructor this is ++a: this is destructor ``` `a++`将会有两次的拷贝构造与析构的调用，效率非常低。 #### :gem: 2.1.2、左右值检测 ```cpp class Point { int x_; int y_; public: Point(int x = 0, int y = 0); Point(const Point&); ~Point(); Point& operator++(); //前置 Point operator++(int); //后置，去掉了const限定 Point operator+(const Point&); Point& operator+=(const Point&); void DisplayPoint(); }; Point::Point(int x, int y) { x_ = x; y_ = y; cout << "this is constructor" << endl; } Point::Point(const Point& b) { this->x_ = b.x_; this->y_ = b.y_; cout << "this is copy constructor" << endl; } Point::~Point() { cout << "this is destructor" << endl; } Point& Point::operator+=(const Point& _right) { this->x_ += _right.x_; this->y_ += _right.y_; return *this; } Point Point::operator+(const Point& _right) { Point temp; temp.x_ = this->x_ + _right.x_; temp.y_ = this->y_ + _right.y_; return temp; } Point& Point::operator++() { ++x_; ++y_; return *this; } Point Point::operator++(int) // int仅仅为了重载 { Point temp(*this); this->x_++; this->y_++; return temp; } void Point::DisplayPoint() { cout << "x: " << this->x_ << endl; cout << "y: " << this->y_ << endl; } int main() { Point b(2, 2); cout << endl << "this is &b: " << &b << endl; Point* c; c = &(++b); cout << endl << "this is c = &(++b): " << c << endl; c->DisplayPoint(); c = &(b++); cout << endl << "this is c = &(b++): " << c << endl; c->DisplayPoint(); ++b = *c; b.DisplayPoint(); b++ = *c; b.DisplayPoint(); return 0; } // 输出 this is constructor this is &b: 008FFD38 this is c = &(++b): 008FFD38 x: 3 y: 3 this is copy constructor this is copy constructor this is destructor this is destructor this is c = &(b++): 008FFC5C x: 3 y: 3 x: 3 y: 3 this is copy constructor this is copy constructor this is destructor this is destructor x: 4 y: 4 this is destructor ``` 可以看到`++b`返回的对象指针跟`b`原来的地址是一样的，而`b++`返回的对象地址跟原来的`b`地址不一样（应该是临时对象的地址），虽然可以取到地址，但当成左值将有可能导致错误。比如`b++ = c`就不会将`c`给`b`，达不到原来的目的。**为此，我们应该将后置的`++`函数返回值定义为const类型：`const Point Point::operator++(int)`，就可以避免这种当成左值情况出现**。另外发现返回temp的引用可以减少一次拷贝和析构，但是不建议返回局部变量的引用！因为函数退出，局部变量将析构，引用就会指向不确定内存。 ### :pen\_fountain: 2.2、a+=b与a=a+b的区别 `a+=b`返回的是a的引用，中间不涉及构造与析构，效率与`++a`一样。而`a=a+b`则会生成临时变量，而且以值传递方式返回，会有两次的构造与析构，与`a++`一样。 ## :pencil2: 3、求值过程中的副作用与顺序点 ```cpp int i = 1; i = (++i) + (++i); // i = ? ``` 对于该代码，gcc和clang给出了不同的结果。clang给出了5，这很好理解；gcc却给出了6。为什么是6？ 如果把`i`换成不同的变量，我们能写出这样的代码：`k = (++i) + (++j)`。尽管`i`和`j`自增先后不确定，但对结果是没有影响的。编译器可以这样拆分它： ```cpp ++i; ++j; k = i + j; ``` 那么，我们把三个`i`代入，就得到了： ```cpp ++i; // i = 2 ++i; // i = 3 i = i + i; // i = 3 + 3 ``` 因此结果是6。 ### 幕后的“顺序点” 之所以`(++i) + (++i)`未定义，这涉及到“顺序点”的概念。 “副作用”，即代码中对数据的改变，`++i`和`i = ...`都会产生副作用。C/C++为了让编译器能更好地优化代码，定义了一些位置，编译器能保证这些位置之后的副作用开始前，之前的副作用完成。这些位置，也就是顺序点。 顺序点包括： * 两个语句之间，`for`循环的括号中包含了三个语句； * 函数调用的参数与函数返回之间（但是参数自身没有顺序保证）； * 短路求值操作符前后； * 问号表达式的条件与结果之间； * 变量初始化之间、之后。 --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://mqjyl2012.gitbook.io/algorithm/c-cpp/c++-syntax/auto-increment-attenuation.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.