David Strawn
David Strawn
@zarthross @hamnis [I've a poc for a non-source breaking fix here.](https://github.com/isomarcte/circe/commit/a711a0eb788e4a6219a8152e0f0d4ca7e71654f4) However, after working through the problem a little bit, I noticed that we are only using `ProductN` for the...
@hamnis I think it will actually, unless I'm misreading that thread. When you use the `("a", 1)` syntax, you are getting an instance of `String *: Int :* EmptyTuple`, which...
cc @armanbilge @satorg
cc @zarthross
Should `Next` imply `Order`? Should it be true that `a < next(a)` for all `A`? Currently `Next` only implies `PartialOrder`.
Should `cycleNext` always include all elements at least once?
@satorg yeah, I admit it is a little strange. We need something which maps to the natural numbers here, and the integer numbers (as in from negative infinity to positive...
@armanbilge @satorg I've updated the first post on this issues with an unreasonably long discussion on where we can go here. Let me know your thoughts (I don't expect any...
@armanbilge ah, that makes sense (and is unfortunate). I think my general proposal still stands, just with the branching class super types being members rather than extending them. Do you...
Actually, let me take another deeper look at this before we proceed. I'm not satisfied with the ergonomics of the solution I proposed, if we have to encode the super...