Getting valid substitutions to work for subsumption

Richard Eisenberg rae at
Fri May 19 00:59:48 UTC 2017

> On May 18, 2017, at 5:27 PM, Simon Peyton Jones <simonpj at> wrote:
> I don't agree.   If you call (tcSubsumes hole_ty ty) with closed types
> hole_ty, ty, it should return True if 

I agree here. But it looks like Matthías's function gets the expected type as pushed down by bidirectional typechecking. This type will always be deeply skolemized before tcSubsumes can get a hold of it, so it won't be closed.


More information about the ghc-devs mailing list