A notice on utilizing pointer to vector elements

vectorelementpointer.cpp

#include <iostream>
#include <vector>
#include <memory>

struct A { int v; };
struct B { const A* a; };

int main()
{
	// NG. Taking raw pointer of an element of the vector.
	// The pointer to the element is changed in the reallocation in push_back.
#if 0
	std::vector<A> v;
	std::vector<B> v2;

	for (int i = 0; i < 10; i++)
	{
		B b;
		{
			A a;
			a.v = i;
			v.push_back(std::move(a));
			b.a = &v.back();
		}
		v2.push_back(std::move(b));
	}

	for (int i = 0; i < 10; i++)
	{
		std::cout << v2[i].a->v << std::endl;
	}
#endif

	// OK. Utilizing std::unique_ptr.
#if 1
	std::vector<std::unique_ptr<A>> v;
	std::vector<B> v2;
	
	for (int i = 0; i < 10; i++)
	{
		B b;
		{
			std::unique_ptr<A> a(new A);;
			a->v = i;
			b.a = a.get();
			v.push_back(std::move(a));
		}
		v2.push_back(std::move(b));
	}

	for (int i = 0; i < 10; i++)
	{
		std::cout << v2[i].a->v << std::endl;
	}
#endif

	return 0;
}