Re: [GHC] #14661: Cannot derive (newtype I a b = I (F a -> F b) deriving newtype Category) for type family F

12 Jan 2018

      #14661: Cannot derive (newtype I a b = I (F a -> F b) deriving newtype Category)
for type family F
-------------------------------------+-------------------------------------
        Reporter:  Iceland_jack      |                Owner:  (none)
            Type:  feature request   |               Status:  closed
        Priority:  normal            |            Milestone:
       Component:  Compiler          |              Version:  8.4.1-alpha1
      Resolution:  wontfix           |             Keywords:
                                     |  DerivingStrategies, deriving,
                                     |  TypeFamilies
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  None/Unknown      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------
Changes (by RyanGlScott):

 * status:  new => closed
 * resolution:   => wontfix

Comment:

 Apologies simonpj, comment:6 wouldn't have made sense except to
 Iceland_jack, myself, and a handful of other people.

 `DerivingVia` is the code name we're using to describe a work-in-progress
 deriving strategy that we're developing, based on this
 [https://skillsmatter.com/skillscasts/10934-lightning-talk-stolen-
 instances-taste-just-fine HaskellX talk] that Iceland_jack gave. This
 would allow you to write something of the form:

 {{{#!hs
 newtype Ixed a b = Ixed (Interp a -> Interp b)
   deriving Category
     via (WRAP InterpSym0 (->) a b)
 }}}

 Where `WRAP` is a separate data type that has exactly the right `Category`
 instance that you'd want for `Ixed`. It would generate the following code:

 {{{#!hs
 instance Category Ixed where
   id  = coerce @(forall a. Ixed a a) @(forall a. WRAP InterpSym0 (->) a a)
 id
   (.) = coerce @(forall a b c. Ixed b c -> Ixed a b -> Ixed a c) @(forall
 a b c. WRAP InterpSym0 (->) b c -> WRAP InterpSym0 (->) a b -> WRAP
 InterpSym0 (->) a c) (.)
 }}}

 Another way of thinking about it is that it's an even more generalized
 form of `GeneralizedNewtypeDeriving` where the user is allowed to specify
 their own type to `coerce` from (instead of being required to use a
 newtype's underlying representation type).

 It's not fully documented at the moment, but we do have a WIP
 specification [https://github.com/Icelandjack/deriving-
 via/blob/06cb4fffbac68a7ca788ce1778f76c971906911e/paper/Specification_sketch.markdown
 here] if you're curious.

 Anyways, I'll opt to close this, since Iceland_jack seems OK with the idea
 of using `DerivingVia` to accomplish this task.

-- 
Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/14661#comment:10
GHC http://www.haskell.org/ghc/
The Glasgow Haskell Compiler