Add rfl::AddNamespacedTagsToVariants for rfl::Generic and same-name-different-namespace structs

[bryceschober]

As an alternative to #467, I offer this draft proposal.

Details:

• Add a separate rfl::AddNamespacedTagsToVariants processor to support the usage of rfl::Generic directly in auto-tagged variants.
• Reuse code paths for rfl::AddTagsToVariants by adding a defaulted non-type template parameter to rfl::VariantAlternativeWrapper
• Build a set of robust and parallel test suites for all 4 combinations of variant auto-tagging
• Rename the previously-existing variant-tagging test files to better match the processor name.

Outstanding items:

• Update documentation to cover the new functionality and use-cases.
• Is adding the parallel processor's functionalty to the existing AddTagsToVariants.hpp and is_add_tags_to_variants_v.hpp acceptable? Those files were named specifically matching the old processor, and now I've added to them.

[liuzicheng1987]

@bryceschober , other than the one comment I made, which you can basically just copy + paste, I like this very much. This is basically ready to be merged.

[bryceschober]

@liuzicheng1987 I've implemented your requested static_assert().

I also added CBOR test coverage representative of our use-case, and using a new method of detailed CBOR test expectations that I think will be more practical going forward. It kinda sucks to depend on outside tooling like https://cbor.me, but it's also relatively easy to use and help connect the dots with the test failure formatter I implemented.

[liuzicheng1987]

@bryceschober thank you!

[jmcken8]

Looks good to me, thanks for doing this work @bryceschober