This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] testsuite (committed) check_effective_target_broken_cplxf_arg


On Mon, Feb 14, 2005 at 09:08:53AM +0000, Richard Sandiford wrote:
> Janis Johnson <janis187@us.ibm.com> writes:
> > Yes, I noticed that when I was in final testing.  Complex float function
> > results didn't change for PowerPC-64 ELF but they did for both SPARC and
> > MIPS; if there are tests affected by that incompatibility then it can
> > use a new check.
> 
> FWIW, as far as MIPS goes, the changes were supposed to bring us in
> line with SGI's tools.  I'm not sure how much there is to be gained
> from running a compat test between (say) gcc 3.3 and gcc 3.4 on the
> affected ABIs.  (Not that I'm saying it's useless.  I'm just saying
> that IMO, it's not something the testsuite really needs to cope with.)

This isn't for compatibility testing, it's for XFAILing a test when it's
run on a system whose libm was built with a version of GCC that didn't
have the ABI fix.  There are 3 gfortran tests and 3 libstdc++ tests
that call cabsf, and those all fail when I'm testing on powerpc64-linux
with an older version of libm and pass when I'm on a system whose libm
was built with a compiler that has the fix.  I'm likely to continue
testing on the older system for quite awhile and wanted to get rid of
the noise.

Janis


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]