Frank Thomas

Ok, let's do this. I hope that instances for `Greater[N] ==> Greater[M]` won't be too hard to write.

I wrote a bit about supporting `FiniteDuration` [here](https://github.com/fthomas/refined/issues/633#issuecomment-488071074). @etspaceman It would help if you could provide one or two examples how you would refine `FiniteDuration`.

Ok, this would be possible with the `LengthInSeconds` predicate mentioned in #633, i.e. `FiniteDuration Refined LengthInSeconds[NonNegative]`, etc.

Oh dear! I haven't seen this one yet. Stack traces I've seen so far are https://github.com/fthomas/refined/issues/3 and https://github.com/scalacenter/scalafix/issues/47.

@tpolecat I've just released refined 0.8.0 which contains a potential fix for these compiler crashes. If you see them again after the upgrade, please reopen this issue.

This isn't fixed, but because of #332 and #334 these crashes will hopefully happen less often since we now try to avoid calling `scala.reflect.macros.contexts.Context.eval` if possible.

@NeQuissimus Do you know the predicate for which the macro is called in this case? If it is provided by the library, we can fix this for this particular predicate...

@NeQuissimus They are not primitive predicates but composed of primitive predicates that are defined by the library so in principle we could add their `Validate` instances to the cache of...

I've just read this comment by @retronym in https://github.com/scala/bug/issues/9218#issuecomment-565285839: ```scala class Predicate[A, B] { self => // self type needed to trigger the bug } ``` I think we can...

The self type is now gone in 0.9.12. If anyone observes this compiler crash with 0.9.12 or a newer version, please leave a comment here.