C++17 折叠表达式计算阶乘时出错

Question

The following program gives errors in both g++ and clang...

#include <iostream>
#include <utility>

using namespace std;

//Since C++17 one can use fold expression to calculate factorial:
template <class T, T N, class I = std::make_integer_sequence<T, N>>
struct factorial;

template <class T, T N, T... Is>
struct factorial<T,N,std::index_sequence<T, Is...>> {
   static constexpr T value = (static_cast<T>(1)* ... *(Is + 1));
};


int main(int argc, char** argv) {

    std::cout << factorial<int, 5>::value << std::endl;

    return 0;
}

Errors listed in gcc are as follows:

 error: type/value mismatch at argument 1 in template parameter list for 'template<long unsigned int ..._Idx> using index_sequence = std::integer_sequence<long unsigned int, _Idx ...>'
   24 | struct factorial<T,N,std::index_sequence<T, Is...>> {
      |                                               ^~~
note:   expected a constant of type 'long unsigned int', got 'T'
error: template argument 3 is invalid
   24 | struct factorial<T,N,std::index_sequence<T, Is...>> {
      |                                                  ^~
In function 'int main(int, char**)':
error: incomplete type 'factorial<int, 5>' used in nested name specifier
   31 |     std::cout << factorial<int, 5>::value << std::endl;
      |                                     ^~~~~

Clang errors are as follows:

error: template argument for non-type template parameter must be an expression
struct factorial<T,N,std::index_sequence<T, Is...>> {
                                         ^
D:\Programs\msys64\mingw64\include\c++\10.1.0\utility:344:22: note: template parameter is declared here
  template<size_t... _Idx>
                     ^
main.cpp:32:18: error: implicit instantiation of undefined template 'factorial<int, 5, std::integer_sequence<int, 0, 1, 2, 3, 4> >'
    std::cout << factorial<int, 5>::value << std::endl;
                 ^
main.cpp:22:8: note: template is declared here
struct factorial;
       ^

This example is shown in many sites.

Request somebody to help me correcting it.

Answer 1

You used the wrong type for your result object.

std::index_sequence is a template that accepts only a pack of std::size_t . It doesn't accept a type parameter, because the type is fixed. It is an alias template for std::integer_sequence<std::size_t, Ints...> .

So you trying to specify a type as its the first argument in std::index_sequence<T,...> is wrong.

What you want is the more general type trait std::integer_sequence , making your specialization into this

template <class T, T N, T... Is>
struct factorial<T,N,std::integer_sequence<T, Is...>> {
   static constexpr T value = (static_cast<T>(1)* ... *(Is + 1));
};

Live code example