Worked around previous bug, then ran into this fellah:
swift.ucc.ie:~/freq: make G62.o
make: Circular G62.hi <- G62.o dependency dropped.
rm -f G61.o
ghc-2.01 -c G61.hs -H30m -K2M -recomp -fglasgow-exts -cpp -hi-diffs
-ansi -syslib ghc
ghc-2.01:compile:Output file G61.o doesn't exist
"G61.hs", line 627: Compiler bug: kind mis-match between
t{-a4LD-} :: *b and GHCbuiltins.Int# :: *u
Expected: GHCbuiltins.Int# -> GHCbuiltins.Int# -> a{-a4LE-} ->
G62.HappyState a{-a4LE-} b{-a4LF-} -> Prelude.Int -> [a{-a4LE-}] ->
[G62.HappyState a{-a4LE-} b{-a4LF-}] -> b{-a4LF-}
Inferred: t{-a4LD-} -> GHCbuiltins.Int# -> Token -> G62.HappyState Token
([G62.HappyAbsSyn a{-a4LG-} b{-a4LH-}] -> a{-a4LI-}) -> Prelude.Int -> [Token]
-> [G62.HappyState Token ([G62.HappyAbsSyn a{-a4LG-} b{-a4LH-}] -> a{-a4LI-})]
-> [G62.HappyAbsSyn a{-a4LG-} b{-a4LH-}] -> a{-a4LI-}
In the first argument of happyGoto namely action_171
In an equation for function action_76:
action_76 82 = happyGoto action_171
I'm not even going to begin to guess at the causes of this one.
Around this time, hand-splitting .hc files is starting to look like a
(relatively) attractive option...
Mutually,
Alex.
