[GHC] #9123: Need for higher kinded roles
GHC
ghc-devs at haskell.org
Mon Jun 2 18:36:46 UTC 2014
#9123: Need for higher kinded roles
-------------------------------------+------------------------------------
Reporter: simonpj | Owner:
Type: bug | Status: new
Priority: normal | Milestone:
Component: Compiler | Version: 7.8.2
Resolution: | Keywords:
Operating System: Unknown/Multiple | Architecture: Unknown/Multiple
Type of failure: None/Unknown | Difficulty: Unknown
Test Case: | Blocked By:
Blocking: | Related Tickets:
-------------------------------------+------------------------------------
Comment (by goldfire):
`instance Rep f => Rep (Compose f)` where `newtype Compose f g a = Compose
(f (g a))`:
We wish to show `forall a. Coercible (m (f a)) (m (g a))` from `Rep m` and
`Coercible f g`. Fix `a`.
1. Use rule (*) to get `Coercible (f a) (g a)`.
2. Use #9117's eta to get `Coercible f g`. We're done.
`instance Rep p => Rep (WrappedArrow p)` where `newtype WrappedArrow a b c
= WrappedArrow (a b c)`:
We wish to show `forall a. Coercible (p x a) (p y a)` from `Rep p` and
`Coercible x y`. Fix `a`.
1. Use #9117's eta to get `Coercible (p x) (p y)`.
2. Use rule (*), and we're done.
`instance Rep (p a) => Rep (WrappedArrow p a)`:
We wish to show `Coercible (p a x) (p a y)` from `Rep (p a)` and
`Coercible x y`.
1. Use rule (*), and we're done.
--------------------------
Unfortunately, I can't counter the point about `StateT` -- we still can't
express that `s` should be representational if `m`'s argument is. But,
what about this:
{{{
newtype Flip0 f a b = Flip0 (f b a) -- not used; just to better
understand Flip1
newtype Flip1 f a b c = Flip1 (f b a c)
instance Rep m => Rep (Flip1 StateT m)
}}}
Would that work for you?
--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9123#comment:21>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
More information about the ghc-tickets
mailing list