The following code fails to compile with Intel C++ 2013.
#include <type_traits>
#include <iostream>
template <
typename T,
typename std::enable_if<std::is_integral<T>::value>::type
>
void myfunc( T a)
{
std::cout << a << std::endl;
}
template <
typename T,
typename std::enable_if<!std::is_integral<T>::value>::type
>
void myfunc( T a)
{
std::cout << a << std::endl;
}
int main()
{
double a;
int b;
myfunc(a);
myfunc(b);
return 0;
}
Here is the error output:
ConsoleApplication1.cpp(33): error : no instance of overloaded function "myfunc" matches the argument list
1> argument types are: (double)
1> myfunc(a);
1> ^
1>
1>ConsoleApplication1.cpp(34): error : no instance of overloaded function "myfunc" matches the argument list
1> argument types are: (int)
1> myfunc(b);
1> ^
1>
Where am I going wrong?
The usual and correct way to use enable_if
in a function is to stick it in the return type.
template <typename T>
typename std::enable_if<std::is_integral<T>::value>::type myfunc(T a) {
std::cout << a << " (integral)" << std::endl;
}
template <typename T>
typename std::enable_if<!std::is_integral<T>::value>::type myfunc(T a) {
std::cout << a << " (non-integral)" << std::endl;
}
For your variant, the correct way is:
template <typename T,
typename = typename std::enable_if<std::is_integral<T>::value>::type>
void myfunc(T a) {
std::cout << a << " (integral)" << std::endl;
}
... the "enable_if" is a default template argument. And it does not work in your case because that function is not overloaded.