Implement Tuple Name Differences in XML/JSON

[erichkeane]

User rggjan requested that tuples provide a different name for each
element in this issue: #140.

This fix modifies tuple.hpp to generate tuple names in the order which
they are are placed in the tuple. Note that this will appear to be a
countdown (tuple_element2, tuple_element1, tuple_element0) since tuples
are generated in reverse order.

Signed-off-by: Erich Keane erich.keane@verizon.net

[AzothAmmo]

I don't like allocating that string just to get the NVP name. I think we can stringify the number using templates at compile time. It'd be easier/cleaner with constexpr, but we unfortunately don't get to use that (thanks MSVC!).

[AzothAmmo]

Templates: transforming run time solutions into compile time solutions through the magic of many lines of code: https://gist.github.com/AzothAmmo/4e9d145e5cce0a174b36

[erichkeane]

58d1675 now uses Azoth's example to generate compile-time names. Additionally, he mentioned wanting tuples to reverse their order in another comment, so I implemented that as well.

[erichkeane]

Any comments or anything on this one? Should i remove the ordering change?

[AzothAmmo]

Does this actually change the ordering? It looks like it is still counting down instead of counting up. Other than that it looks good - I do want to change the order to be in the natural tuple order (0, 1, 2, ...)

[erichkeane]

It DOES reverse it. Note how I've reversed the call to apply and the ar( call.

This is my test snippet:

std::tuple<int, int, int, int> tup(1,2,3,4);
ostringstream os;
{
    cereal::JSONOutputArchive ar(os);
    ar(tup);
}
std::cout<< os.str()<<std::endl;

With apply happening second, this is the output:

{
    "value0": {
        "tuple_element3": 4,
        "tuple_element2": 3,
        "tuple_element1": 2,
        "tuple_element0": 1
    }
}

However, with the apply second (like in 58d1675) this is my output:

{
    "value0": {
        "tuple_element0": 1,
        "tuple_element1": 2,
        "tuple_element2": 3,
        "tuple_element3": 4
    }
}

[AzothAmmo]

Right you are. My eyes glossed over the unchanged line in the diff. Will fully test and merge this into 1.1.

[AzothAmmo]

Everything seems fine on Linux, I don't expect any problems on Windows. Will compile with MSVC later anyway, but closing now.