Initiallizer list、auto type 、foreach 、nullptr 代替了c++ 03的NULL、enum class代替了c++03的enum,更安全、override关键标识 for virtual function (更加安全,直观 )、final关键标识 ,主要是class 及virtual function、关键字default标识, compiler generated default constructor、关键标识 delete ,放在函数后面,表示函数不能被调用、lambda function, 让编程更优雅、tuple元祖
1、Initiallizer list
//c++03
std::vector<int> v;
v.push_back(2);
v.push_back(3);
//c++ 11
std::vector<int> v{ 1,2,3 };
2、auto type
std::vector<int> vec = { 1,2,3,4 };
//c++ 03
for (std::vector<int>::iterator it = vec.begin(); it != vec.end(); it++) {
std::cout << *it << std::endl;
}
//c++ 11
for (auto it = vec.begin(); it != vec.end(); it++) {
std::cout << *it << std::endl;
}
auto a = 3 + 2;
auto s = "123";
3. foreach 写c++像是再写c#
std::vector<int> vec = { 1,2,3,4 };
//c++ 03
for (std::vector<int>::iterator it = vec.begin(); it != vec.end(); it++) {
std::cout << *it << std::endl;
}
//c++ 11
//read only(只读)
for (auto i : vec) {
std::cout << i << std::endl;
}
//只读
for (const auto& i : vec) {
std::cout << i << std::endl;
}
//changes the values in v(修改)
for (auto& i : vec) {
i = 3;
}
4. nullptr 代替了c++ 03的NULL
5. enum class代替了c++03的enum,更安全
//c++ 03
enum ship {
bulkship,
tankership
};
//c++ 11
enum class ship1 {
bulkship,
tankership
};
6. override关键标识 for virtual function (更加安全,直观 )
class base {
public:
virtual void fun1(int);
virtual void fun2() const;
void fun3(int);
};
class son :public base {
//c++ 03 存在隐患
/*
void fun1(float); //不小心写错了参数,ok 编译通过,create a new func
void fun2(); //不小心少写了const,ok 编译通过,create a new func
void fun3();
*/
// but in c++ 11 更安全清晰
void fun1(float) override; //编译Error: no func to override
void fun2() override; //编译Error: no func to override
void fun3() override; //编译Error: no func to override
};
7. final关键标识 ,主要是class 及virtual function
//this is means no class can be derived from CPoint2D
class CPoint2D final {
//No class can override Draw
virtual void Draw() final;
};
8. 关键字default标识, compiler generated default constructor
强制编译器生成默认构造函数
- 编译器不再生成默认构造函数
class CPoint2D {
public:
CPoint2D(double x_,double y_) {
x = x_;
y = y_;
}
double x;
double y;
};
int main(){
CPoint2D pt;//编译报错,不存在默认构造函数,因为编译器不再生成
}
- 通过default force compiler to generate the default constructor
class CPoint2D {
public:
CPoint2D(double x_,double y_) {
x = x_;
y = y_;
}
CPoint2D() = default;//告诉编译器强制生成
double x;
double y;
};
int main(){
CPoint2D pt;//ok
}
9. 关键标识 delete ,放在函数后面,表示函数不能被调用
- 看以前的情形,构造函数只接受int,但是输入double也是可以的,因为会自动转换
class dog {
public:
dog(int age_) {
age = age_;
}
int age;
};
int main(){
dog(2); //ok
dog(4.5); //also ok,converted form double to int
}
- c++ 11 解决方案
class dog {
public:
dog(int age_) {
age = age_;
}
dog(double) = delete;
int age;
};
int main(){
dog(2); //ok
dog(4.5); //not ok,已经删除的函数
}
10. lambda function, 让编程更优雅
像c#一样优雅起来
int main(){
std::cout << [](int x, int y) {return x + y; }(3, 5) << std::endl;
}
我经常使用std::sort对结构体配合Lammda匿名函数使用,比如点的数组按照x升序排列
class CPoint2D {
public:
CPoint2D(double x1,double y1) {
x = x1; y = y1;
}
double x;
double y;
};
int main(){
std::vector<CPoint2D> Pts{ {1,1},{3,3},{2,2} };
std::sort(begin(Pts), end(Pts), [](const CPoint2D& pt1, const CPoint2D& pt2) {
return pt1.x < pt2.x;
});
}
11.smart pointer 智能指针
12. tuple元祖
- std::pair<std::string,int>的扩展版,可以当做一个通用的结构体来使用
- tuple某些情况下会让代码的易读性变差
- 什么情况下使用tuple
std::pair<>的扩展版,可多个参数
std::pair<int, std::string> p = std::make_pair(2, "hello");
//std::pair的扩展版
std::tuple<int, std::string, char> t(2,"foo",'a');
auto t1 = std::make_tuple(0, "dog", 'b');
std::cout << std::get<0>(t) << std::endl;
std::cout << std::get<1>(t) << std::endl;
std::cout << std::get<2>(t) << std::endl;
易读性会变差(一个使用结构体,一个使用tuple)
结构体比较清晰 p1.name p1.age
tuple易读性差,通过 std::get<0>(t2) std::get<1>(t2)取值
你会想what stored in position 0&what stored in position 1?
struct person {
std::string name;
int age;
};
person p1;
std::tuple<std::string, int> t2;
//把上面的代码遮挡起来 tuple会导致代码的易读性降低
std::cout << p1.name << p1.age<<std::endl;
std::cout << std::get<0>(t2) << std::get<1>(t2) << std::endl;
//what stored in position 0? what stored in position 1?
when to use tuple
(1) As a one-time only structure to transfer g group of data 比如作为函数的返回值使用,这样就不用声明结构体了
std::tuple<std::string,int> GetNameAge(){
return std::make_tuple("lcl", 20);
}
(2)comparison of tuples
//hours, minutes, seconds
//comparison of tuples
std::tuple<int, int, int> time1, time2;
if (time1 > time2) {
//...
}
(3)Muti index map
std::map<std::tuple<int, char, double>, float> m;