On Sat, Apr 25, 2009 at 08:10:45PM +0100, Max Bolingbroke wrote: > 2009/4/25 Ian Lynagh <ig...@earth.li>: > > Ooops, sorrry about that. > > > >> Thanks Max, I now understand the reasoning behind this change a bit > >> better. However, I'm still not convinced that it is a good idea, > > > > All in my opinion: > > > > * The fact that you are even allowed to let/where-bind values with > > unlifted type is surprising. Making it clearer that something odd is > > going on is a good thing. If it were up to me I think I would disallow > > them completely. > > Where this really breaks down is in a not-entirely-hypothetical > language extension where you are able to write functions that are > polymorphic in the strictness of certain type variables, and unlift > arbitrary lifted types by prefixing them with a bang. Then, should you > be allowed to write the following? > > map f (x:xs) = y : map f xs > where y = f x > > By your rules, we should not be allowed to, because "f x" may > potentially return an unlifted computation that is eagerly evaluated > and hence may diverge. IMHO this is far too restrictive.
I think this is only a problem if you have kind polymorphism. If f is merely strict then it is fine. It's only if the return type is unlifted that you have a problem. Thanks Ian _______________________________________________ Cvs-ghc mailing list Cvs-ghc@haskell.org http://www.haskell.org/mailman/listinfo/cvs-ghc
