循环范围内的初始化列表
我有从单个超类型派生的不同类型的对象.我想知道在这样的循环范围内使用 std::initializer
列表是否有任何缺点:
I have objects of different types derived from a single super-type. I wonder if there are any disadvantages in using std::initializer
list in a range for loop like this:
for(auto object: std::initializer_list<Object *>{object1, object2, object3}) {
}
是否完全可以且高效,还是使用数组更好?对我来说,std::array
解决方案似乎对编译器的限制更大,并且显式说明大小有一个缺点:
Is it completely OK and efficient or would it be better to use an array? To me the std::array
solution seems to be more restrictive for the compiler and there is a disadvantage of explicitly stating the size:
for(auto object: std::array<Object*, 3>{object1, object2, object3}) {
}
是否有任何其他或更好的方法来迭代明确给定的对象列表?
Is there any other or nicer way of iterating over an explicitly given list of objects?
推荐答案
循环内部不需要使用冗长的std::initializer_list
There is no need to use the verbose std::initializer_list
inside the loop
#include <iostream>
#include <initializer_list>
struct B { virtual int fun() { return 0; } };
struct D1 : B { int fun() { return 1; } };
struct D2 : B { int fun() { return 2; } };
int main()
{
D1 x;
D2 y;
B* px = &x;
B* py = &y;
for (auto& e : { px, py })
std::cout << e->fun() << "
";
}
实例.
如果您想在不定义 px
和 py
的情况下即时执行此操作,您确实可以使用 std::initializer_list<B*>{ &x, &y }
在循环内.
If you want to do it on-the-fly without defining px
and py
, you can indeed use std::initializer_list<B*>{ &x, &y }
inside the loop.
相关文章