Testsuite failures with Intel OSX 10.5.0

[Jack Howarth]

FX,
   I think you are over interpreting those results. Bradley always builds
with -mcpu=970 which is non-standard for powerpc-apple-darwin builds.
Regress is a better metric for the failures from normal builds.

http://gcc.gnu.org/ml/gcc-testresults/2007-11/msg00862.html

On ppc, the -m64 and -m32 failures for gfortran are identical.
I do wish Apple would get around to setting up a regress server for
Macintel though (considering they want it to be a primary platform
for gcc).
            Jack

On Fri, Nov 16, 2007 at 05:45:31PM +0000, François-Xavier Coudert wrote:
> > I have been able to build gfortran on Intel OSX 10.5.0 without
> > trouble. However I see three failures
> 
> Can you post the testresults to gcc-testresults, and open PRs for
> these failures?
> 
> I saw that Leopard on ppc has quite a few failures
> (http://gcc.gnu.org/ml/gcc-testresults/2007-11/msg00719.html),
> fortunately all of them are excess errors@-O3 -g, so they're
> probably related.
> 
> FX