Re: [PATCH]: PR middle-end/30250, use MPFR for lgamma_r/gamma_r

["Kaveh R. GHAZI" <ghazi@caip.rutgers.edu>]

On Wed, 23 May 2007, Richard Guenther wrote:

> On 5/19/07, Kaveh R. GHAZI <ghazi@caip.rutgers.edu> wrote:
> > On Thu, 17 May 2007, Richard Guenther wrote:
> >
> > > > +             /* Dereference the arg_sg pointer argument.  */
> > > > +             arg_sg = build_fold_indirect_ref (arg_sg);
> > > > +             /* Proceed iff a valid pointer type was passed in.  */
> > > > +             if (TYPE_MAIN_VARIANT (TREE_TYPE (arg_sg)) == integer_type_node)
> > >
> > > This check should be done earlier based on the pointed to type of the
> > > argument. Ok with this change.
> >
> > Done.  I installed the revised patch below.  Thanks again for your
> > reviews.
>
> This:
>
> /* Expect failures at least until mpfr-2.3.0 is released. */
> /* { dg-xfail-if "mpfr-2.3.0" { *-*-* } } */
>
> doesn't work.  Or did you forget a testsuite/lib patch?
> Richard.

There was no testsuite patch involved, dg-xfail-if should work just fine.
Can you provide some more info as to what exactly doesn't work?  Here are
some 4.3.0 testresults posted for the trunk dated 20070523 on various
platforms.  None of them show builtin-math-4.c failing:

http://gcc.gnu.org/ml/gcc-testresults/2007-05/msg01156.html
http://gcc.gnu.org/ml/gcc-testresults/2007-05/msg01155.html
http://gcc.gnu.org/ml/gcc-testresults/2007-05/msg01153.html
http://gcc.gnu.org/ml/gcc-testresults/2007-05/msg01151.html
http://gcc.gnu.org/ml/gcc-testresults/2007-05/msg01143.html
http://gcc.gnu.org/ml/gcc-testresults/2007-05/msg01132.html

Ah!  I found one from Joe Buck that doesn't pickup the xfail:
http://gcc.gnu.org/ml/gcc-testresults/2007-05/msg01149.html

But I can't explain why.  Maybe he's using an outdated version of dejagnu?

Joe?

		Thanks,
		--Kaveh
--
Kaveh R. Ghazi			ghazi@caip.rutgers.edu