internal::same_cont_new_t should preserve allocators

[Dobiasd]

Minimal example:

#include <fplus/fplus.hpp>
#include <boost/align/aligned_allocator.hpp>

template <typename T>
using aligned_vector = std::vector<T, boost::alignment::aligned_allocator<T, 32>>;

int main()
{
    typedef aligned_vector<int> AlignedIntVector;
    AlignedIntVector aligned_xs = {1,2,2,3,2};
    AlignedIntVector aligned_ys =
        fplus::transform([](auto x) { return x*x; }, aligned_xs);
}

internal::same_cont_new_t can not handle this:

g++:

error: conversion from ‘std::vector<int, std::allocator<int> >’ to non-scalar type ‘AlignedIntVector {aka std::vector<int, boost::alignment::aligned_allocator<int, 32ul> >}’ requested
         fplus::transform([](auto x) { return x*x; }, aligned_xs);

clang++:

error: no viable conversion from 'vector<[...],
      std::allocator<int>>' to 'vector<[...],
      boost::alignment::aligned_allocator<int, 32>>'
    AlignedIntVector aligned_ys =
                     ^

The current workaround is to provide the output container explicitly. fplus::transform_convert can be used for convenience:

#include <fplus/fplus.hpp>
#include <boost/align/aligned_allocator.hpp>

template <typename T>
using aligned_vector = std::vector<T, boost::alignment::aligned_allocator<T, 32>>;

int main()
{
    typedef aligned_vector<int> AlignedIntVector;
    AlignedIntVector aligned_xs = {1,2,2,3,2};
    AlignedIntVector aligned_ys =
        fplus::transform_convert<AlignedIntVector>(
            [](auto x) { return x*x; }, aligned_xs);
}

But since not only fplus::transform is affected, it would be very cool if internal::same_cont_new_t would support that.