On Apr9, 2014, at 02:55 , David Rowley <dgrowleyml@gmail.com> wrote:
> On Wed, Apr 9, 2014 at 8:48 AM, Florian Pflug <fgp@phlo.org> wrote:
>
> As explain above, invtrans_bool is a bit problematic, since it carries
> a real risk of performance regressions. It's included for completeness'
> sake, and should probably not be committed at this time.
>
> Did you mean to write invtrans_minmax? Otherwise you didn't explain about
> you concerns with bool.
Grmpf. Should have re-read that once more before sending :-(
Yes, I meant invtrans_minmax is problematic! invtrans_bool is fine, the
inverse transition function never fails for BOOL_AND and BOOL_OR. This
is why I factored it out into a separate patch, to make it easy to not
apply the MIN/MAX stuff, while still applying the BOOL stuff. Sorry for
the confision.
best regards,
Florian Pflug