Transform variadic template parameters to another types

你说的曾经没有我的故事 提交于 2019-11-30 03:43:13

You can do more than just expand a variadic parameter pack as a plain list: you can expand an expression too. You can therefore have m_sequences be a tuple of vectors rather than a tuple of the elements:

template <typename... T>
struct sequences
{
   std::tuple<std::vector<T>...> m_sequences;
};

You can also do nifty tricks with parameter packs to pick the appropriate element from the vector:

template<size_t ... Indices> struct indices_holder
{};

template<size_t index_to_add,typename Indices=indices_holder<> >
struct make_indices_impl;

template<size_t index_to_add,size_t...existing_indices>
struct make_indices_impl<index_to_add,indices_holder<existing_indices...> >
{
    typedef typename make_indices_impl<
        index_to_add-1,
        indices_holder<index_to_add-1,existing_indices...> >::type type;
};

template<size_t... existing_indices>
struct make_indices_impl<0,indices_holder<existing_indices...> >
{
    typedef indices_holder<existing_indices...>  type;
};

template<size_t max_index>
typename make_indices_impl<max_index>::type make_indices()
{
    return typename make_indices_impl<max_index>::type();
}



template <typename... T>
struct sequences
{
    std::tuple<std::vector<T>...> m_sequences;

    template<size_t... Indices>
    std::tuple<T...> get_impl(size_t pos,indices_holder<Indices...>)
    {
        return std::make_tuple(std::get<Indices>(m_sequences)[pos]...);
    }

    std::tuple<T...> get(size_t pos)
    {
        return get_impl(pos,make_indices<sizeof...(T)>());
    }
};

OK, this might seem a bit like overkill but how about this: As far as I know the only option to "iterate" variadics is using the <head, tail...> notation with a template specialization for the simple <head-only> case.

Therefore you could try something like this:

simple case:

template <typename T>
struct sequences
{
   std::tuple<T> get(size_t pos)
   {
     return values[pos];
   }

   std::vector<T> get_sequence()
   {
      return values;
   }

   std::vector<T> values;
};

recursive case:

template <typename T, typename ...U>
struct sequences
{
   std::tuple<T, std::tuple<U...> > get(size_t pos)
   {
     return std::make_tuple(values[pos], remainder->get(pos));
   }

  template <size_t Idx>
  std::vector<
      typename std::tuple_element<Idx, std::tuple<T...>>::type
    > get_sequence()
  {
    return get_sequence_internal<
         typename std::tuple_element<Idx, std::tuple<T...>>::type, Idx
       >();
   }

   template <typename V, 0>
   std::vector<V> get_sequence_internal()
   {
      return values;
   }

   template <typename V, size_t Idx>
   std::vector<V> get_sequence()
   {
      return remainder->getSequence_internal<V, Idx-1>();
   }



   std::vector<T> values;
   sequences<U...>* remainder;
};

Disclaimer: not tested, not even compiled, but I suppose you get the basic idea. At least two problems remain:

  1. The return value of get() is not your single struct but a tuple chain. Perhaps you can unchain it recursively with std::get<0>...
  2. I don't know if the specialization of get_sequence_internal generates a compile time error because the type of V can differ from T.
易学教程内所有资源均来自网络或用户发布的内容,如有违反法律规定的内容欢迎反馈
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!