On Fri, 24 Oct 2014, Jakub Jelinek wrote:
> On Fri, Oct 24, 2014 at 03:27:19PM +0200, Richard Biener wrote:
> > As noted by Marc I forgot to actually utilize the iterator variable.
> >
> > Bootstrapped on x86_64-unknown-linux-gnu, testing in progress.
> >
> > Richard.
> >
> > PS: How do we want to refer to patterns in ChangeLogs?
>
> Perhaps the syntax should be (simplify "name" (...) { ... })
> (maybe the name being optional?), where you'd give some name to the
> simplification, say "0 % X" or "0 % X => 0" or "0 % X variant 3"
> or whatever, then you could easily refer to those strings in ChangeLog,
> on gcc-patches, in comments etc.
I ripped out optional name support when I added user-defined predicates
which look like
(match truth_valued_p
(truth_not @0))
or
(match (logical_inverted_value @0)
(bit_not truth_valued_p@0))
(un-)conveniently the parsers for (simplify...) and (match...)
are shared.
I can see to re-add the optional pattern naming. OTOH it will be
fun to invent an unique name for each of them ;) (patternN
anyone? ...)
Richard.
> > 2014-10-24 Richard Biener <rguent...@suse.de>
> >
> > * match.pd (0 % X): Properly use the iterator iterating over
> > all modulo operators.
> > (X % 1): Likewise.
> >
> > Index: gcc/match.pd
> > ===================================================================
> > --- gcc/match.pd (revision 216648)
> > +++ gcc/match.pd (working copy)
> > @@ -64,13 +64,13 @@ (define_predicates
> > (for op (ceil_mod floor_mod round_mod trunc_mod)
> > /* 0 % X is always zero. */
> > (simplify
> > - (trunc_mod integer_zerop@0 @1)
> > + (op integer_zerop@0 @1)
> > /* But not for 0 % 0 so that we can get the proper warnings and errors.
> > */
> > (if (!integer_zerop (@1))
> > @0))
> > /* X % 1 is always zero. */
> > (simplify
> > - (trunc_mod @0 integer_onep)
> > + (op @0 integer_onep)
> > { build_zero_cst (type); }))
> >
> > /* x | ~0 -> ~0 */
