[GHC] #15231: UndecidableInstances validity checking is wrong in the presence of QuantifiedConstraints

GHC ghc-devs at haskell.org
Tue Jun 5 20:14:34 UTC 2018


#15231: UndecidableInstances validity checking is wrong in the presence of
QuantifiedConstraints
-------------------------------------+-------------------------------------
           Reporter:  RyanGlScott    |             Owner:  (none)
               Type:  bug            |            Status:  new
           Priority:  normal         |         Milestone:  8.6.1
          Component:  Compiler       |           Version:  8.5
  (Type checker)                     |
           Keywords:                 |  Operating System:  Unknown/Multiple
  QuantifiedConstraints              |
       Architecture:                 |   Type of failure:  GHC rejects
  Unknown/Multiple                   |  valid program
          Test Case:                 |        Blocked By:
           Blocking:                 |   Related Tickets:
Differential Rev(s):                 |         Wiki Page:
-------------------------------------+-------------------------------------
 Consider this program:

 {{{#!hs
 {-# LANGUAGE ConstraintKinds #-}
 {-# LANGUAGE GADTs #-}
 {-# LANGUAGE KindSignatures #-}
 {-# LANGUAGE QuantifiedConstraints #-}
 module Bug where

 import Data.Kind

 data ECC :: Constraint -> Type -> Type

 class Y a
 class Z a

 instance  c         => Y (ECC c a)
 instance (c => Z a) => Z (ECC c a)
 }}}

 I would expect both of these instances to work. But while GHC accepts the
 `Y` instance, it rejects the `Z` instance:

 {{{
 $ ~/Software/ghc5/inplace/bin/ghc-stage2 Bug.hs
 [1 of 1] Compiling Bug              ( Bug.hs, Bug.o )

 Bug.hs:15:10: error:
     • Variable ‘c’ occurs more often
         in the constraint ‘c’ than in the instance head
       (Use UndecidableInstances to permit this)
     • In the instance declaration for ‘Z (ECC c a)’
    |
 15 | instance (c => Z a) => Z (ECC c a)
    |          ^^^^^^^^^^^^^^^^^^^^^^^^^
 }}}

 That error message seems completely bogus to me, since `c` appears once in
 both the context and instance head in both instances.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15231>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler


More information about the ghc-tickets mailing list