This is a bug, or at least infelicity, in 6.6. I'm not inclined to fix it, unless something worse happens, because it's ok in the HEAD.
I will add a regression test though! Simon | -----Original Message----- | From: [EMAIL PROTECTED] [mailto:glasgow-haskell- | [EMAIL PROTECTED] On Behalf Of Donald Bruce Stewart | Sent: 30 November 2006 00:38 | To: David Roundy | Cc: glasgow-haskell-bugs@haskell.org | Subject: Re: bad error message with GADTs | | droundy: | > The following code gives a rather unpleasant error message with ghc | 6.6. | > Adding the type signature fixes the error. | > | > David | > | > {-# OPTIONS -fglasgow-exts #-} | > | > data PatchSeq p a b where | > Nil :: PatchSeq p a a | > U :: p a b -> PatchSeq p a b | > (:-) :: PatchSeq p a b -> PatchSeq p b c -> PatchSeq p a c | > | > -- is_normal :: PatchSeq p a b -> Bool | > is_normal Nil = True | > is_normal (U _) = True | > is_normal (U _ :- _) = True | > is_normal _ = False | > _______________________________________________ | | I.e. this one? | | A.hs:9:10: | Couldn't match kind `?' against `* -> * -> *' | When matching the kinds of `t :: ?' and `t1 :: * -> * -> *' | Expected type: t1 | Inferred type: t | In the pattern: Nil | | _______________________________________________ | Glasgow-haskell-bugs mailing list | Glasgow-haskell-bugs@haskell.org | http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs _______________________________________________ Glasgow-haskell-bugs mailing list Glasgow-haskell-bugs@haskell.org http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs