Consider adding instance MonoFunctor Set

[tysonzero]

Thus, even though a MonoFunctor instance for Set could theoretically be defined, it is omitted since it could violate the functor law of omap f . omap g = omap (f . g).

The current docs in Data.Eq specifies that Eq is expected to follow the substitutivity law (modulo private functions, such as showTree and friends), so this functor law could only be violated via an unlawful instance.

Since it is well established that laws are not expected to hold in the presence of law violating instances (Foldable for Set onto a non-associative Monoid violates laws for example), I think it's worth considering adding this instance.

[snoyberg]

I don't have any strong thoughts on this to be honest.