GHCi crash due to unsafe coerce ?

[Mathew Mills]

My first darcs-pull after the big System-FC HEAD change has certain  
problems with impossibilities.  I find that I cannot start 'ghci' for  
example.  AFAIK, I do not have any ghci startup scripts.  I am  
rolling back patches now to see if I can get a working system again.


Mathew-Mills-Computer% ghci
    ___         ___ _
   / _ \ /\  /\/ __(_)
/ /_\// /_/ / /  | |      GHC Interactive, version 6.5, for Haskell 98.
/ /_\\/ __  / /___| |      http://www.haskell.org/ghc/
\____/\/ /_/\____/|_|      Type :? for help.

ghc-6.5: panic! (the 'impossible' happened)
   (GHC version 6.5 for powerpc-apple-darwin):
         ByteCodeGen.schemeE: unhandled case
     (base:GHC.Base.>>{v 01Q} [gid]
        @ <nt>base:GHC.IOBase.IO{tc 32I}
        base:GHC.IOBase.$f16{v reO} [gid]
        @ base:GHC.Base.(){(w) tc 40}
        @ base:GHC.Base.(){(w) tc 40}
        sat_syB{v} [lid]
        sat_syD{v} [lid])
     `cast` (base:GHC.Prim.CoUnsafe{(w) tc 34K}  
(<nt>base:GHC.IOBase.IO{tc 32I} base:GHC.Base.(){(w) tc 40})
                                                base:GHC.Base.(){(w)  
tc 40}
             :: {<nt>base:GHC.IOBase.IO{tc 32I} base:GHC.Base.(){(w)  
tc 40}
                   :=:
                 base:GHC.Base.(){(w) tc 40}})

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

Mathew-Mills-Computer%