Call variadic templated function with arguments from a std::vector

Question

I need to convert elements of a std::vector to types based on a template parameter and call a function with these parameters. In pseudocode:

template <typename T...>
void foo(std::vector<std::string> v) {
    if (v.size() != sizeof...(T))
        throw std::runtime_error("Bad");

    bar(convert<T0>(v[0]), convert<T1>(v[1]), ..., convert<Tn>(v[n]));
}

My problem is how to obtain the element indices from the parameter pack, I think there will be some kind of a trick using fold expressions, but I can't figure it out.

Answer 1

If you know that the number of elements in a vector is equal to the parameter pack size, you can solve this problem by adding one level of indirection:

template<typename... T, std::size_t... is>
void foo_impl(const std::vector<std::string>& v, std::index_sequence<is...>) {
    bar(convert<T>(v[is])...);
}


template<typename... T>
void foo(const std::vector<std::string>& v) {
    assert(v.size() == sizeof...(T));
    foo_impl<T...>(v, std::index_sequence_for<T...>{});
}

The idea here is to expand two packs, Ts... and is..., which have equal sizes, simultaneously.

---

C++20 solution:

template<typename... T>
void foo(const std::vector<std::string>& v) {
    assert(v.size() == sizeof...(T));

    [&v]<std::size_t... is>(std::index_sequence<is...>) {
        bar(convert<T>(v[is])...);
    }(std::index_sequence_for<T...>{});
}