This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: results for egcs-2.91.53 586-pc-linux-gnulibc1
- To: jbuck at synopsys dot com, schmid at ltoi dot iap dot physik dot tu-darmstadt dot de
- Subject: Re: results for egcs-2.91.53 586-pc-linux-gnulibc1
- From: mrs at wrs dot com (Mike Stump)
- Date: Thu, 6 Aug 1998 10:40:47 -0700
- Cc: egcs at cygnus dot com
> From: Joe Buck <jbuck@Synopsys.COM>
> To: schmid@ltoi.iap.physik.tu-darmstadt.de (Peter Schmid)
> Date: Wed, 5 Aug 98 17:49:07 PDT
> Cc: egcs@cygnus.com
> Peter Schmid writes:
> It appears that our testing methodology has a severe problem: we can
> report PASS for a test, even though the compiler crashes on the test!
I will agree with you, that in the light of this type of testing, it
is better to assume that the testcase in fact did fail.
The problem is knowing it. One will have to additionally look for
``confused by earlier errors'' and ``Internal compiler error'' in
testsuite/lib/old-deja.exp and if that is present, set a variable.
Then in the loop over messages, in the default case in stead of:
} else {
set ok fail
set uhoh pass
}
do:
} else {
set ok fail
set uhoh pass
if (any_ice)
set uhoh fail
}
I think. Someone want to give it a try? I'd be happy to review the
work, though nothing beats testing it. :-)
> So what should be done? It would be nice to have a separate status,
No! Please, I bed of you, don't ask for that.