Yes it's deliberate. (Turned out to be easy and convenient.) Yes it goes beyond H98, so GHC types a few too many programs even in H98 mode. I guess I should document it. Simon | -----Original Message----- | From: glasgow-haskell-users-bounces@haskell.org [mailto:glasgow-haskell-users- | bounces@haskell.org] On Behalf Of Ross Paterson | Sent: 17 November 2005 11:31 | To: glasgow-haskell-users@haskell.org | Subject: kind inference | | The Haskell 98 Report says kind inference is to be done over dependency | groups with polymorphic kinds defaulted to *, and hence the following | example (from 4.6) is illegal: | | data Tree a = Leaf | Fork (Tree a) (Tree a) | | type TreeList = Tree [] | | GHC 6.4 seems to perform kind inference across all the data, newtype, | type and class declarations of a module before defaulting, and so accepts | the above. An improvement, I think, but unadvertised. Was it intentional? | | _______________________________________________ | Glasgow-haskell-users mailing list | Glasgow-haskell-users@haskell.org | http://www.haskell.org/mailman/listinfo/glasgow-haskell-users