Add a batch_cast equivalent for batch_bool

[marzer]

Hiya,

Recently I've found myself having to do math on floats, but do testing against associated integer masks, and I'd like to reduce the pain associated with working with select() in these scenarios.

The library has batch_cast for regular batch, but no equivalent for their batch_bool friends. There's some undocumented bool_cast() functions but they aren't generalized at all and don't appear to be what I'm looking for. Also the somewhat ominous //FIXME comment next to them suggests that I maybe shouldn't depend on these being present 😅

I have hacked this together in my own codebase, and it seems to serve the purpose:

	template <class Dest, class Source, class A>
	inline batch_bool<Dest, A> batch_bool_cast(const batch_bool<Source, A>& x) noexcept
	{
		static_assert(batch_bool<Source, A>::size == batch_bool<Dest, A>::size);

		if constexpr (std::is_same_v<Source, Dest>)
		{
			return x;
		}
		else if constexpr (std::is_same_v<typename batch_bool<Source, A>::register_type,
										  typename batch_bool<Dest, A>::register_type>)
		{
			return batch_bool<Dest, A>(x.data);
		}
		else
		{
			return batch_bool<Dest, A>(batch_cast<Dest>(batch<Source, A>(x)).data);
		}
	}

I think something like this would be a useful addition to the library itself to complement the existing batch_cast().

[serge-sans-paille]

+1 for that one. Please not thet we're still in C++11 though. Waiting for you PR :-)

[marzer]

Waiting for you PR :-)

Sure, I'd be happy to give this a crack.