On Sun, Aug 30, 2009 at 5:33 AM, Ganesh Sittampalam<[email protected]> wrote:
> Mon Aug  3 06:05:14 BST 2009  Ganesh Sittampalam <[email protected]>
>  * couple of GADT build fixes for recent GHCs

This is dealing with the typical issue of using existential types in a
polymorphic pattern match.  In particular,
-     let Sealed ps = flattenOne t
Causes GHC's head to explode.  The reason is that let and where are
not-monomorphic like "case", lambda, and function definitions.  This
makes existentials very hard to reason about.  GHC tackles it by
giving up (perhaps the only reasonable solution).

The refactorings look fine to me.

Jason
_______________________________________________
darcs-users mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-users

Reply via email to