[GHC] #14146: Can GHC propose kind restrictions?

22 Aug 2017


      #14146: Can GHC propose kind restrictions?
-------------------------------------+-------------------------------------
           Reporter:  Iceland_jack   |             Owner:  (none)
               Type:  feature        |            Status:  new
  request                            |
           Priority:  normal         |         Milestone:
          Component:  Compiler       |           Version:  8.2.1
           Keywords:                 |  Operating System:  Unknown/Multiple
       Architecture:                 |   Type of failure:  None/Unknown
  Unknown/Multiple                   |
          Test Case:                 |        Blocked By:
           Blocking:                 |   Related Tickets:
Differential Rev(s):                 |         Wiki Page:
-------------------------------------+-------------------------------------
 This is UX. As my code gets more polykinded I find myself in situations
 such as these all the more often

 {{{#!hs
 instance (Comonad f, Representable f, Comonad g, Representable g) =>
 Comonad (Compose f g) where
   extract   = undefined
   duplicate = undefined

 instance ComonadApply (Compose f g) where
   (<@>) = undefined
 }}}

 {{{
 /tmp/a.hs:20:10-35: error:
        • No instance for (Comonad (Compose f g))
            arising from the superclasses of an instance declaration
        • In the instance declaration for ‘ComonadApply (Compose f g)’
       |
    20 | instance ComonadApply (Compose f g) where
       |          ^^^^^^^^^^^^^^^^^^^^^^^^^^
    Failed, modules loaded: none.
 }}}

 It would help me if GHC looked up a `Comonad (Compose _ _)` instance,
 compared the kinds

 {{{#!hs
 Comonad      (Compose (f :: Type -> Type) (g :: Type -> Type) :: Type ->
 Type)
 ComonadApply (Compose (f :: k    -> Type) (g :: Type -> k)    :: Type ->
 Type)
 }}}

 a simple suggestion like this would be helpful

 {{{
        • No instance for (Comonad (Compose f g))
            arising from the superclasses of an instance declaration
          Try adding a kind signature (ComonadApply (Compose (f :: Type ->
 Type) g)).
 }}}

 Even more amazing would be

 {{{
        • No instance for (Comonad (Compose f g))
            arising from the superclasses of an instance declaration
          Try adding a context
             (Comonad f, Representable f, Comonad g, Representable g)
               => ComonadApply (Compose f g)
 }}}

 Which the compiler could in theory guess, since following GHC's suggestion
 iteratively you end up with

 {{{#!hs
 ComonadApply (Compose (f :: Type -> Type) g)

 -- ==> add (Comonad f) to the context of the instance declaration
 Comonad f => ComonadApply (Compose f g)

 -- ==> add (Comonad g) to the context of the instance declaration
 (Comonad f, Comonad g) => ComonadApply (Compose f g)

 -- ==> add (Representable f) to the context of the instance declaration
 (Comonad f, Comonad g, Representable g) => ComonadApply (Compose f g)

 -- ==> add (Representable f) to the context of the instance declaration
 (Comonad f, Comonad g, Representable g, Representable f) => ComonadApply
 (Compose f g)
 }}}

-- 
Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/14146
GHC http://www.haskell.org/ghc/
The Glasgow Haskell Compiler