sloppy bug report

Hi,

Apologies for the sloppiness of this bug report.

in ghc 5.04.2 I get this message:

ghc-5.04.2: panic! (the `impossible' happened, GHC version 5.04.2):
        coreSyn/Subst.lhs:385: Non-exhaustive patterns in function zip_ty_env

I would love to try and give you a simple test case, but I can't. It happens
when I am building a (buddha-transformed) raytracer which I am also trying
to profile (there is a lot of code involved - and it is mangled because of
the debugging transformation, so finding a simplified expression of the
bug is somewhat tricky).

The bug disappears if I re-compile all my code from scratch.

In any case the code for zip_ty_env is simple enough that someone who
understands it may want to consider its assumptions:

   zip_ty_env []       []       env = env
   zip_ty_env (tv:tvs) (ty:tys) env = .... 

Obviously one of the lists is empty and the other is not - though the code
assumes this will never happen.

Cheers,
Bernie.