[GHC] #14869: Documentation for isLiftedTypeKind is incorrect

[GHC]

#14869: Documentation for isLiftedTypeKind is incorrect
-------------------------------------+-------------------------------------
        Reporter:  RyanGlScott       |                Owner:  (none)
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler (Type    |              Version:  8.2.2
  checker)                           |
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Documentation     |  Unknown/Multiple
  bug                                |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------
Description changed by RyanGlScott:

Old description:

> I [https://github.com/ghc-proposals/ghc-
> proposals/pull/32#issuecomment-369252383 noticed recently] that Template
> Haskell reifies `Constraint` as `Type`:
>
> {{{#!hs
> $ ~/Software/ghc2/inplace/bin/ghc-stage2 --interactive
> GHCi, version 8.5.20180221: http://www.haskell.org/ghc/  :? for help
> Loaded GHCi configuration from /home/rgscott/.ghci
> λ> :set -XTypeFamilies -XTemplateHaskell
> λ> :m + Data.Kind Language.Haskell.TH
> λ> type family Foo :: Constraint
> λ> putStrLn $(reify ''Foo >>= stringE . pprint)
> type family Ghci1.Foo :: *
> }}}
>
> The root of the issue can be traced back to the `isLiftedTypeKind`
> function, which `TcSplice` uses to distinguish `Type` from `Constraint`.
> At least, that's what its
> [http://git.haskell.org/ghc.git/blob/df2c3b3364834d2fd038192c89348fc50a2e0475:/compiler/types/TyCoRep.hs#l721
> documentation] claims:
>
> {{{
>  -- | This version considers Constraint to be distinct from *. Returns
> True
>  -- if the argument is equivalent to Type and False otherwise.
>  isLiftedTypeKind :: Kind -> Bool
>  isLiftedTypeKind = is_TYPE is_lifted
>    where
>      is_lifted (TyConApp lifted_rep []) = lifted_rep `hasKey`
> liftedRepDataConKey
>      is_lifted _                        = False
> }}}
>
> However, in practice this claim about treating `Constraint` and `Type` as
> distinct is false:
>
> {{{
> $ ~/Software/ghc2/inplace/bin/ghc-stage2 --interactive -package ghc
> GHCi, version 8.5.20180221: http://www.haskell.org/ghc/  :? for help
> Loaded GHCi configuration from /home/rgscott/.ghci
> λ> :m + TyCoRep TysWiredIn
> λ> isLiftedTypeKind liftedTypeKind
> True
> λ> isLiftedTypeKind constraintKind
> True
> }}}
>
> Either we should change the implementation of `isLiftedTypeKind` to match
> the documentation's claim, or change the documentation.

New description:

 I [https://github.com/ghc-proposals/ghc-
 proposals/pull/32#issuecomment-369252383 noticed recently] that Template
 Haskell reifies `Constraint` as `Type`:

 {{{
 $ ~/Software/ghc2/inplace/bin/ghc-stage2 --interactive
 GHCi, version 8.5.20180221: http://www.haskell.org/ghc/  :? for help
 Loaded GHCi configuration from /home/rgscott/.ghci
 λ> :set -XTypeFamilies -XTemplateHaskell
 λ> :m + Data.Kind Language.Haskell.TH
 λ> type family Foo :: Constraint
 λ> putStrLn $(reify ''Foo >>= stringE . pprint)
 type family Ghci1.Foo :: *
 }}}

 The root of the issue can be traced back to the `isLiftedTypeKind`
 function, which `TcSplice` uses to distinguish `Type` from `Constraint`.
 At least, that's what its
 [http://git.haskell.org/ghc.git/blob/df2c3b3364834d2fd038192c89348fc50a2e0475:/compiler/types/TyCoRep.hs#l721
 documentation] claims:

 {{{
  -- | This version considers Constraint to be distinct from *. Returns
 True
  -- if the argument is equivalent to Type and False otherwise.
  isLiftedTypeKind :: Kind -> Bool
  isLiftedTypeKind = is_TYPE is_lifted
    where
      is_lifted (TyConApp lifted_rep []) = lifted_rep `hasKey`
 liftedRepDataConKey
      is_lifted _                        = False
 }}}

 However, in practice this claim about treating `Constraint` and `Type` as
 distinct is false:

 {{{
 $ ~/Software/ghc2/inplace/bin/ghc-stage2 --interactive -package ghc
 GHCi, version 8.5.20180221: http://www.haskell.org/ghc/  :? for help
 Loaded GHCi configuration from /home/rgscott/.ghci
 λ> :m + TyCoRep TysWiredIn
 λ> isLiftedTypeKind liftedTypeKind
 True
 λ> isLiftedTypeKind constraintKind
 True
 }}}

 Either we should change the implementation of `isLiftedTypeKind` to match
 the documentation's claim, or change the documentation.

--

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14869#comment:1>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler