Reputation: 4992
I have a simple templated class which I instantiate for each type that uses it like so:
template<class T> struct meta {
static const std::string& get_name() {return T::class_name;}
};
template<> struct meta<double> {
static constexpr const char* class_name = "double";
static const char* get_name() { return class_name; }
};
template<> struct meta<std::vector<double> {
static constexpr const char* class_name = "std::vector<double>";
static const char* get_name() { return class_name; }
};
template<> struct meta<std::array<double,2>> {
static constexpr const char* class_name = "double[]";
static const char* get_name() { return class_name; }
};
Then I can use these like so:
int main() {
std::cout << meta<double>::get_name() << '\n';
std::cout << meta<std::vector<double> >::get_name() << '\n';
std::cout << meta<std::array<double,2> >::get_name() << '\n';
}
I use macros to generate variants of the template-specialization for each base-type, but I'm stuck on std::array
because it contains a base-type and an integer. I'd like to find out how to add a new templated parameter to an already-templated specialization so that I don't need to specialize for each possible value of integer.
I've tried:
template<int I> struct meta<std::array<double,I>> {
static constexpr const char* class_name = "double[]";
static const char* get_name() { return class_name; }
};
error: ‘class_name’ is not a member of ‘std::array’
Upvotes: 1
Views: 68
Reputation: 3945
From the std::array
C++ documentation:
template<
class T,
std::size_t N
> struct array;
So, you need
template<std::size_t I> struct meta<std::array<double, I>> {
static constexpr const char* class_name = "double[]";
static const char* get_name() { return class_name; }
};
Because int
and std::size_t
are two different types (having different signed-ness), the partial template specialization won't match. Changing int
to size_t
works.
In summary: Read the documentation.
Upvotes: 2