darcs patch: Don't charge for int2Word# and word2Int# -- they don't...

[Samuel Bronson]

On 11/6/06, Simon Marlow <simonmarhaskell at gmail.com> wrote:
> Samuel Bronson wrote:
> > Mon Oct 30 07:41:11 EST 2006  Samuel Bronson <naesten at gmail.com>
> >   * Don't charge for int2Word# and word2Int# -- they don't do anything, after all
>
> Sounds like a reasonable change to me - if it makes things slower, I'd be
> interested to know why.

It actually seems to interact badly with GHC's notorious sensitivity
to order of inlining... like, things are inlined into their wrappers
when that is not such a good idea... at least, it *looks* like that
happens.