seemingly inconsistent behavior for "kind-indexed" type constraints in GHC 7.6
eir at cis.upenn.edu
Sat Oct 6 04:25:48 CEST 2012
I think GHC is right in rejecting your instance, but perhaps not for the reasons you expect.
It would be possible to add another instance to NUnits:
instance NUnits Z '[()]
This would require OverlappingInstances, but because of the possibility of further instances, GHC rightly does not assume that a matching instance is the only possible matching instance. So, without that assumption, there is no reason GHC should unify ts with '.
For similar reasons, GHC does not bring the constraints on an instance into the context when an instance matches. So, even if GHC did select the instance you want, it would not bring (' ~ ps) into the context.
To me, both of these problems have solutions: introduce a functional dependency on NUnits, n -> ts; and declare the wanted instance to be "instance NUnits Z '", without the equality constraint. What's interesting is that doing both of these changes didn't get the instance accepted. This *might* be a bug in GHC, but I'd love another opinion before filing, because I'm really not sure.
Another issue at work here is that GHCi, by default, does not enable PolyKinds, even when it has loaded a file with PolyKinds. The effect of this is that printing out polymorphic kinds will see all type variables default to * without :set -XPolyKinds. That may have had a role to play in the commentary in the file, but I'm not sure.
Thanks for posting!
On Oct 5, 2012, at 5:49 PM, Nicolas Frisby wrote:
> GHC 7.6 is rejecting some programs that I think ought to be well-typed.
> Details here https://gist.github.com/3842579
> I find this behavior particularly odd because I can get GHC to deduce
> the type equalities in some contexts but not in others. In particular,
> it does not deduce them inside of type class instances.
> Is this a known issue? I'll create a Trac ticket if that's appropriate.
> Thanks for your time.
> Glasgow-haskell-users mailing list
> Glasgow-haskell-users at haskell.org
More information about the Glasgow-haskell-users