The array.data() and array.view().data() are inaccessible.

[acdemiralp]

This means there is no way to access the container underneath linearly. Deal breaker when you do file IO.

Example code:

TEST_CASE("experimental/mdarray")
{
  auto array = std::experimental::mdarray<float, std::experimental::dynamic_extent, std::experimental::dynamic_extent, std::experimental::dynamic_extent>(3, 3, 3);

  auto count = 0;
  for (auto x = 0; x < 3; ++x)
    for (auto y = 0; y < 3; ++y)
      for (auto z = 0; z < 3; ++z)
        array(x, y, z) = count++;

  auto recount = 0;
  for (auto i = 0; i < array.size(); ++i)
    REQUIRE(array.data()[i] == recount++); // Compile error.
}

Error: C2039 'data': is not a member of 'std::experimental::__mdarray_version_0::vector_container_policy<T,std::allocator>' array_test C:\development\source\cpp\particle_tracer\build\vcpkg\installed\x64-windows\include\experimental__p1684_bits\basic_mdarray.hpp 387

When changed to array.view().data()[i] error becomes: C2248 'std::experimental::__mdarray_version_0::basic_mdarray<float,std::experimental::extents<-1,-1,-1>,std::experimental::layout_right,std::experimental::_mdarray_version_0::vector_container_policy<T,std::allocator>>::map': cannot access private member declared in class 'std::experimental::__mdarray_version_0::basic_mdarray<float,std::experimental::extents<-1,-1,-1>,std::experimental::layout_right,std::experimental::__mdarray_version_0::vector_container_policy<T,std::allocator>>' array_test C:\development\source\cpp\particle_tracer\build\vcpkg\installed\x64-windows\include\experimental__p1684_bits\basic_mdarray.hpp 75

[dhollman]

Hey there! This is probably just a bug in my implementation; our mdarray implementation is not up to the production standards of our mdspan implementation. As specified in the proposal (wg21.link/p1684r0), the data() member should be accessible and should do what you want here. There's also the view().span() that should give you a std::span for linear access, though note that mdspan (and, subsequently, mdarray) specializations are not required to have contiguous or uniquely mapped underlying memory, which is why linear access doesn't make much sense in a generic context. In non-generic context, the problem becomes much easier because you can just interact with whatever concrete data structure that's being used for storage (which you know about directly in a non-generic context) rather than dealing with accessing it through mdarray or (especially) mdspan. But since generic use cases are the priority here (in particular, because the concrete use cases are much easier to deal with), this isn't a point of emphasis.