Add and derive Deep class

[paf31]

The idea here is to obviate the need for the old generics implementation and deriving.

This is still a work in progress, but it works for everything except records. I think if we want to go this route, then we should

• combine the two libraries back into purescript-generics
• deprecate this one
• make a major release of generics
• remove old generic deriving from the compiler.

Thoughts?

[matthewleon]

Question: Are there uses for Deep other than compatibility with non-GHC style generics? In other words, if I am writing a new library, is it an option I should look at?

[paf31]

@matthewleon That is the main use case, and I'm not sure if it's worth it to be honest. I don't think there is anything you can do with generics that you can't do with generics-rep (this PR is sort of a proof of that 😄)

[paf31]

I actually do have one use case which would work with old-style generics, but not with new-style generics, which is to reflect a generic signature as an actual type with a Generic instance.

So this PR is a proof that every type with a generics-rep instance also has an old Generic instance, but not the other way round, and when we're creating types at runtime, that's what we care about.

[JordanMartinez]

Should this PR be closed?

[hdgarrood]

I think this hasn’t had enough interest to justify inclusion in core, yeah.