#7696 closed bug (fixed)

Another kindFunResult panic

Reported by: nwf Owned by:
Priority: normal Milestone:
Component: Compiler Version: 7.6.2
Keywords: Cc: dimitris@…
Operating System: Unknown/Multiple Architecture: Unknown/Multiple
Type of failure: None/Unknown Difficulty: Unknown
Test Case: typecheck/should_fail/T7696 Blocked By:
Blocking: Related Tickets:

Description

While working at the GHCi prompt, I left off a pair of parens, yielding:

> import Control.Applicative
> import Control.Monad.Trans
> :type \x -> (either (fmap Right) id) <$> lift $ x
ghc: panic! (the 'impossible' happened)
  (GHC version 7.7.20130214 for x86_64-unknown-linux):
	kindFunResult
<<details unavailable>>

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

Replacing <$> lift $ x with <$> (lift $ x), which is what I'd meant, makes everybody happier. But as this happens in both 7.6.2 and 7.7.20130214, I thought I should do what it says and report it.

Thanks!

Change History (3)

comment:1 Changed 14 months ago by monoidal

There are two different issues here: one causes panic in HEAD, the other in 7.6.2.

Panic in HEAD (7.6 correctly reports error):

f1 :: (m a, t m)
f1 = undefined

f2 :: ((), m ())
f2 = f1

Panic in 7.6 (HEAD correctly reports error):

lift :: m a -> t m a
lift = undefined

k :: f (Either Int Int)
k = lift

The second one is probably the same cause as #7368, which was left unfixed in 7.6.

comment:2 Changed 14 months ago by simonpj@…

commit c969cc3d2cfcd31fc8e7222c7c9ec116191c136b

Author: Simon Peyton Jones <simonpj@microsoft.com>
Date:   Sun Mar 3 23:04:09 2013 +0000

    Treat equalities with incompatible kinds as "irreducible" constraints
    
    Originally we had the invariant that CTyEqCan and CFunEqCan have LHS
    and RHS with compatible kinds.  This is important because if they have
    different kinds, then a substitution using the CTyEqCan can give rise
    to an ill-kinded type, which in turn makes typeKind crash, and this
    led to Trac #7696.  (The possibility of this happening really only
    occurred when we introduced kind polymorphism.)
    
    I thought at first this was going to be really awkward to solve, but
    happily it turned out to be easy.  We already have CIrredEvCan
    constraints, which are "stuck"; we can't use them and we can't solve
    them.  Yet. After some substitution from solving other constraints we
    may be able to make progress.
    
    So for equality constraints where the LHS and RHS don't have compatible kinds
    (although perhaps not YET compatible, eg k and *, just needing to
    unify k := *), we now generate a CIrredEvCan, plus the necessary kind
    equality constraint.
    
    This entailed some refactoring of course, but only in TcCanonical.  In
    particular, the emitKindConstraint code has gone, in favour of a kind
    check in canEqLeaf.  See Note [Equalities with incompatible kinds] in
    TcCanonical, and Note [CIrredEvCan constraints] in TcRnTypes

 compiler/typecheck/TcCanonical.lhs |  150 +++++++++++++++++++-----------------
 compiler/typecheck/TcRnTypes.lhs   |   33 ++++++--
 2 files changed, 102 insertions(+), 81 deletions(-)

comment:3 Changed 14 months ago by simonpj

  • Cc dimitris@… added
  • Difficulty set to Unknown
  • Resolution set to fixed
  • Status changed from new to closed
  • Test Case set to typecheck/should_fail/T7696

Excellent point! Actually this bug revealed a pretty significant shortcoming in the constraint solver (see the commit message). Fortunately, I found rather a neat solution, so all is good now.

Thanks for identifying it so compactly.

Simon

Note: See TracTickets for help on using tickets.