TypeHiearchy.pdf is missing an edge from union to class

[zygoloid]

A union is a kind of class, so there should be a line from union to class in TypeHierarchy.pdf.

(Otherwise this is an excellent and incredibly useful resource! Would you have any interest in contributing it to the standard?)

[HowardHinnant]

I'd be happy to contribute it to the standard. But you may not want it after my response. :-)

I should make this chart more clear that it models the primary and composite type_traits ([meta.unary.cat] and [meta.unary.comp] respectively). Here there is no such relationship between union and class.

This was an intentional deviation from [basic.types] for C++11 justified by the practical needs of the clients of <type_traits>.

That being said, if anyone wants to create a derivative work from this that depicts [basic.types], I don't mind at all.