I have a type X
, which is copy-constructible but not copy assignable
struct X {
X();
X(const X&);
X& operator=(const X&) = delete; // !!
X(X&&) noexcept;
X& operator=(X&&) noexcept;
int data = 54;
};
I have two vectors of 'X': a
and b
and I want to insert all the contents of b
at the front of a
:
void add_to_front(std::vector<X>& a, const std::vector<X>& b) {
a.insert(a.begin(), b.begin(), b.end());
}
This compiles and works as expected on msvc but fails to compile on clang and gcc. I'm guessing due to poor implementations of libc++ and libstdc++ which need something to compile even though it will never get called (or, worse yet, it will get called?.).
I could write a manual loop to emplace
elements of b
into a
which will produce a correct result, but the complexity of this is a*b instead of a+b, as each call to emplace
will shift all elements of a
over and over again.
So is there an efficient way to do it?
I must admit that I'm not quite sure whether the poor implementations of libc++ and libstdc++ are the issue. Though, I found a quite simple way to circumvent OPs issue:
#include <vector>
struct X {
X() = default;
X(const X&) = default;
X& operator=(const X&) = delete; // !!
X(X&&) noexcept = default;
X& operator=(X&&) noexcept = default;
int data = 54;
};
void add_to_front(std::vector<X>& a, const std::vector<X>& b) {
std::vector<X> b_(b);
a.insert(a.begin(), std::make_move_iterator(b_.begin()), std::make_move_iterator(b_.end()));
}
int main()
{
std::vector<X> a, b;
add_to_front(a, b);
}
This is accepted by
-std=c++11 -O2
-std=c++11 -O2
/std:c++14 /O2
Live Demo on Compiler Explorer
@Evg had a look into the
Type requirements for std::vector::insert()
In OPs case it's this overload:
template< class InputIt > iterator insert( const_iterator pos, InputIt first, InputIt last );
with
- T must meet the requirements of EmplaceConstructible in order to use overload (4,5).
- T must meet the requirements of MoveAssignable and MoveInsertable in order to use overload (4). required only if InputIt satisfies LegacyInputIterator but not LegacyForwardIterator.
This should be granted by OPs struct X
.
So, to me, it looks like OP is right with the complaints.
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.