The issue is clear with the following code:
#include <functional>
#include <iostream>
#include <vector>
int main() {
//std::vector<int> a, b;
int a = 0, b = 0;
auto refa = std::ref(a);
auto refb = std::ref(b);
std::cout << (refa < refb) << '\n';
return 0;
}
If I use the commented std::vector<int> a, b;
instead of int a = 0, b = 0;
, then the code does not compile on any of GCC 5.1, clang 3.6, or MSVC'13. In my opinion, std::reference_wrapper<std::vector<int>>
is implicitly convertible to std::vector<int>&
which is LessThanComparable, and thus it should be LessThanComparable itself. Could someone explain this to me?
The issue is that the non-member operator<
for std::vector
is a function template:
template< class T, class Alloc >
bool operator<( const vector<T,Alloc>& lhs,
const vector<T,Alloc>& rhs );
Implicit conversions are not considered when doing template type deduction here, [temp.arg.explicit] emphasis on if:
Implicit conversions (Clause 4) will be performed on a function argument to convert it to the type of the
corresponding function parameter if the parameter type contains no template-parameters that participate
in template argument deduction.
But in this case, the parameter type does participate in deduction. That's why it can't be found. Had we written our own non-template operator<
:
bool operator<(const std::vector<int>& lhs, const std::vector<int>& rhs)
{
return true;
}
Your code would work as expected. To use the generic one though, you will have to explicitly pull out the reference:
std::cout << (refa.get() < refb.get()) << '\n';
Are you certain that
std::vector<int> a, b;
Is doing what it is supposed to? Take this for example
#include <functional>
#include <iostream>
#include <vector>
int main() {
std::vector<int> a, b;
//int a = 0, b = 0;
a.push_back(42);
a.push_back(6);
a.push_back(15);
for (int ii=0; ii<43; ii++) {
b.push_back(ii);
}
auto refa = std::ref(a);
auto refb = std::ref(b);
std::cout<<&refa<<std::endl;
std::cout<<&refb<<std::endl;
std::cout<<"Contents of vector A"<<std::endl;
for(auto n : a)
{
std::cout<<' '<<n;
}
std::cout<<std::endl<<"Contents of vector b: ";
for (auto n : b){
std::cout<<' '<<n;
}
//std::cout << (refa < refb) << '\n';
return 0;
}
Which results in
0x7fff5fbff0c0
0x7fff5fbff0b8
Contents of vector A
42 6 15
Contents of vector b: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42
Ultimately
std::vector<int> a, b;
Creates two separate vectors of integers called a and b, both of which have no contents; this is not how one would declare a single vector with members a and b.
int a=0, b=0;
Declares two separate integers called a and b, which each have a value of 0. Those two code snippets declare completely different variables and should not be used interchangeably.