core lint error with arrow notation and GADTs

[benmos]

The following code panics GHC (with 7.0.3, 7.2 and 7.4.0.20111219):

{-# LANGUAGE Arrows, GADTs #-}
import Control.Arrow

data Value a where BoolVal :: Value Bool

class ArrowInit f where
    arrif :: f b -> ()

instance ArrowInit Value where
    arrif = proc BoolVal -> returnA -< ()
    -- arrif = arr (\BoolVal -> ())

I am attaching the -dcore-lint from 7.4.

[benmos]

Core Lint output

[benmos]

Source file

[ross]

Simplified version, avoiding classes:

{-# LANGUAGE Arrows, GADTs #-}
module ArrowBug5777 where

import Control.Arrow

data Value a where BoolVal :: Value Bool

arrif :: Value Bool -> ()
arrif = proc BoolVal -> returnA -< ()

[benmos]

The workaround is to move the GADT data constructor out of the 'proc' pattern and into a 'let':

data Value a where BoolVal :: Int -> Value Bool

arrif :: Value Bool -> Int
arrif = proc x -> do
          let BoolVal i = x
          returnA -< i

[altaic]

This also happens on GHC 7.4.2 x86_64.

$ ghc --make Test3.hs 
[1 of 1] Compiling Main             ( Test3.hs, Test3.o )
ghc: panic! (the 'impossible' happened)
  (GHC version 7.4.2 for x86_64-apple-darwin):
	cgLookupPanic (probably invalid Core; try -dcore-lint)
    ( $dArrow{v amL} [lid] :: base:Control.Arrow.Arrow{tc r4n}
                                ghc-prim:GHC.Prim.(->){(w) tc 3D} )
    static binds for:
    local binds for:
    ( main:Main.arrif{v rci} [gid[ClassOp]] :: forall (( f{tv acm} [tv] :: ghc-prim:GHC.Prim.*{(w) tc 34d}
                                                                           -> ghc-prim:GHC.Prim.*{(w) tc 34d} ) :: ghc-prim:GHC.Prim.*{(w) tc 34d}
                                                                                                                   -> ghc-prim:GHC.Prim.*{(w) tc 34d}).
                                               main:Main.ArrowInit{tc rch}
                                                 ( f{tv acm} [tv] :: ghc-prim:GHC.Prim.*{(w) tc 34d}
                                                                     -> ghc-prim:GHC.Prim.*{(w) tc 34d} ) =>
                                               forall ( b{tv acn} [tv] :: ghc-prim:GHC.Prim.*{(w) tc 34d} ).
                                               ( f{tv acm} [tv] :: ghc-prim:GHC.Prim.*{(w) tc 34d}
                                                                   -> ghc-prim:GHC.Prim.*{(w) tc 34d} ) ( b{tv acn} [tv] :: ghc-prim:GHC.Prim.*{(w) tc 34d} )
                                               -> ghc-prim:GHC.Tuple.(){(w) tc 40} )
    ( main:Main.$WBoolVal{v rcH} [gid[DataConWrapper]] :: main:Main.Value{tc rcj}
                                                            ghc-prim:GHC.Types.Bool{(w) tc 3c} )

Please report this as a GHC bug:  http://www.haskell.org/ghc/reportabug

[simonpj]

Ross, do you feel able to help with this. I have no clue.

[ross]

I'll put it on my list, but first priority is ArrForm? (#5267/#5609), followed by infixr (#5333), with this and #344 as the lowest.

[simonpj@…]

commit c3ad38d7dc39ef583ddfb586413baa2e57ca3ee8

Author: Simon Peyton Jones <simonpj@microsoft.com>
Date:   Mon Mar 4 09:40:56 2013 +0000

    Rearrange the typechecking of arrows, especially arrow "forms"
    
    The typechecking of arrow forms (in GHC 7.6) is known to be bogus, as
    described in Trac #5609, because it marches down tuple types that may
    not yet be fully worked out, depending on when constraint solving
    happens.  Moreover, coercions are generated and simply discarded.  The
    fact that it works at all is a miracle.
    
    This refactoring is based on a conversation with Ross, where we
    rearranged the typing of the argument stack, so that the arrows
    have the form
       a (env, (arg1, (arg2, ...(argn, ())))) res
    rather than
       a (arg1, (arg2, ...(argn, env))) res
    as it was before.
    
    This is vastly simpler to typecheck; just look at the beautiful,
    simple type checking of arrow forms now!
    
    We need a new HsCmdCast to capture the coercions generated from
    the argument stack.
    
    This leaves us in a better position to tackle the open arrow tickets
     * Trac #5777 still fails.  (I was hoping this patch would cure it.)
     * Trac #5609 is too complicated for me to grok.  Ross?
     * Trac #344
     * Trac #5333

 compiler/deSugar/Coverage.lhs     |    3 +
 compiler/deSugar/DsArrows.lhs     |  486 +++++++++++++++++++++----------------
 compiler/hsSyn/HsExpr.lhs         |   13 +-
 compiler/hsSyn/HsUtils.lhs        |    7 +-
 compiler/parser/Parser.y.pp       |    4 +-
 compiler/parser/RdrHsSyn.lhs      |    4 +-
 compiler/rename/RnExpr.lhs        |    4 +-
 compiler/rename/RnTypes.lhs       |    2 +-
 compiler/typecheck/TcArrows.lhs   |  272 ++++++++++-----------
 compiler/typecheck/TcHsSyn.lhs    |    6 +-
 docs/users_guide/glasgow_exts.xml |   48 +++--
 11 files changed, 461 insertions(+), 388 deletions(-)