This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: ACATS c380004
- From: Duncan Sands <duncan dot sands at math dot u-psud dot fr>
- To: kenner at vlsi1 dot ultra dot nyu dot edu (Richard Kenner)
- Cc: gcc at gcc dot gnu dot org
- Date: Fri, 7 Jan 2005 15:59:51 +0100
- Subject: Re: ACATS c380004
- References: <10501071439.AA14755@vlsi1.ultra.nyu.edu>
Hi Richard, thanks for replying.
On Friday 07 January 2005 15:39, Richard Kenner wrote:
> ACATS c380004 fails with current CVS. I ran it under valgrind, and got:
>
> Normally, it fails only on 64-bit machines.
This is 32 bit intel (linux).
> In any event, this is a known problem and relates to a rather pathalogical
> usage in that ACATS test. A protest of the test failed, so at some point,
> the implementation of entry families will have to change to allow this obscure
> case. But this isn't on anybody's high-priority list since it will never
> occur in real programs.
>
> If we had an XFAIL mechanism in ACATS, it should be applied to this test.
Well, the information from valgrind shows that bad code is being produced (eg:
silly arg (-2147483640) to malloc()). So it looks like the compiler is doing
something worse than just not conforming to the RM. But maybe I misunderstood
what you meant by a "known problem"? By the way, gcc at the pre-ssa tag doesn't
fail this test, and valgrind doesn't show any problems for it.
All the best,
Duncan.