This is the mail archive of the
mailing list for the GCC project.
Re: patch: cleanup gcc.dg/vect/ (take 2)
On Mon, Nov 08, 2004 at 05:33:54PM -0500, James A. Morrison wrote:
> Janis Johnson <firstname.lastname@example.org> writes:
> Are the following run time tests or dejagnu tests?
They would be procs in *.exp files similar to
check_named_sections_available in lib/target-supports.exp. They could
be used via related dg-require-* procs or using dg-target-list if that
patch is accepted. Each check_* proc would compare the current target
to a list of targets for which the attribute is known to be true, or
return a result depending on whether a test case could compile, or
some combination of those.
> > Is this a sufficient set of new check_* procs?
> > vect_int_fp: target supports both integer and floating point vectors
> These need to be split. MIPS supports floating point, SPARC with VIS 1
> only supports integers, but I think VIS > 2 supports floating point vectors.
Good point, thanks.
> > vect_align: target handles mis-aligned vectors
> perhaps vect_misalign?
> > vect_bitwise: target provides vector bitwise operations
> > vect_max: target provides a vector max operation
> > New support in http://gcc.gnu.org/ml/gcc-patches/2004-11/msg00417.html
> > would allow combining check_* procs and targets, so a test could use
> > [dg-target-list xfail [expr [check_vect_int_fp]] && [check_vect_align]]].
> What about using unsupported instead of xfail?
That could be done by putting a target list in dg-final instead of an
xfail list. A particular dg-final check would use a target list if it
is only relevant for particular targets, or an xfail list if the check
should be tried for every target that runs the test but is a known
failure (with an open PR) on some targets.