This is the mail archive of the
mailing list for the GCC project.
Re: [RFC] Nonnegative improvements
On Mon, 2005-06-20 at 08:57 -0400, Diego Novillo wrote:
> On Mon, Jun 20, 2005 at 12:33:51AM -0400, James A. Morrison wrote:
> > Thanks, I think vr_value is a good indicator if range information is
> > available.
> It is.
> The one thing you will have to coordinate with Jeff Law is the
> change to simplify_using_ranges. That function needs to change
> into a subroutine of substitute_and_fold or some such. We are
> currently doing to IL scans during final substitution and that is
> wasteful and unnecessary.
Yup. I was going to combine the two routines after adding support
for simplifying conditionals using VRP data. I'm in a minor rathole
right now dealing with some issues that arose as part of that work.