Always return an actual bool from has_* functions

[Avasam]

Reduce changes from #343

Technically a user-facing change, but not sure how to present that in a changelog for setuptools, considering the changelog entry "Always return an actual bool from has_* functions" wouldn't apply to all setuptools' own methods.

[jaraco]

After reflecting on this change for a bit, I realized I don't want to apply this technique generally.

One of the nice things about Python is the ability to use the boolean inference to make checks on objects that aren't themselves boolean. It's nice not to have to write if bool(list_obj) and just write if list_obj. I'd like to see the type checkers start supporting a SupportsBool protocol or similar... or allow any object to be used in place of a bool.

I'm guessing the typechecker authors have already considered this problem and realized that since Any python object can be treated as a boolean, there's little strictness in allowing a declared 'bool' to resolve from Any.

I do want to try avoid sprinkling bool() all over the place simply because the typecheckers are being nitpicky. The simple elegance of boolean inference would be lost, so I'll be resistant to changes like these in the future.