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]: fix PR testsuite/12325, inf unavailable on vax


On Sun, Jan 14, 2007 at 11:06:27AM -0500, Kaveh R. GHAZI wrote:
> On Sat, 13 Jan 2007, Kaveh R. GHAZI wrote:
> 
> > I suspect that SPU is failing this test as well.  If you like, I can add
> > spu-*-* to the target list.  The macro on the line in question expands to
> > test all three floating point types.  The dg-warning doesn't care if it
> > gets 1, 10 or a zillion matching warnings, it eliminates all of them like
> > grep -v.
> >
> > In this case there are two builtin_inf calls for each type making six
> > warnings total on vax.  You're probably getting two less since double
> > works.  You're getting two warnings for float, and I don't know what
> > you're getting for long double.  In any case, adding spu to the target
> > list should "fix it" for spu until you add the target_supports macro.
> 
> So I built a quick cross-compiler to spu-unknown-elf and found that only
> float yields the warning, not double or long double.  Nevertheless as I
> explained they all appear on one line with the same warning text.  So the
> one dg-warning I put in is good either way regardless of how many types
> yield the diagnostic.
> 
> So I added spu-*-* to the target list.
> 
> Okay everywhere?

Yes.

Janis


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