This is the mail archive of the
gcc-regression@gcc.gnu.org
mailing list for the GCC project.
Re: new FAILs on HEAD
- From: gcc at ds217-115-141-84 dot dedicated dot hosteurope dot de
- To: Jim Wilson <wilson at specifixinc dot com>
- Cc: gcc-regression at gcc dot gnu dot org, edelsohn at gnu dot org, jh at suse dot cz, kazu at cs dot umass dot edu, pcarlini at suse dot de, rsandifo at redhat dot com, charlet at act-europe dot fr
- Date: Mon, 12 Jan 2004 22:53:55 +0100
- Subject: Re: new FAILs on HEAD
- References: <20040111050501.A8572@ds217-115-141-84> <1073941244.1233.35.camel@leaf.tuliptree.org>
- Reply-to: gcc-regression at gcc dot gnu dot org
n Mon, Jan 12, 2004 at 01:00:43PM -0800, Jim Wilson wrote:
> On Sat, 2004-01-10 at 20:05,
> gcc@ds217-115-141-84.dedicated.hosteurope.de wrote:
> > These new FAILs appeared since the last run:
> > FAIL: cd92001
>
> The bug report claims a patch checked in on the 10th broke the compiler,
> but the follow up message claims a patch checked in on the 9th broke
> it. Which is it? I find these messages contradictory.
Well, the messages indeed represent what the regression tester seems. I
manually confirmed these findings. The FAIL shows up in these test runs:
2003-11-17-1.tests
2003-11-18-2.tests
2003-11-21-2.tests
2003-11-21-4.tests
2003-11-22-1.tests
2003-11-22-2.tests
2003-11-22-3.tests
2003-11-23-2.tests
2003-11-24-1.tests
2003-11-24-2.tests
2003-12-30-3.tests
2003-12-31-3.tests
2003-12-31-4.tests
2004-01-02-10.tests
2004-01-03-1.tests
2004-01-03-4.tests
2004-01-04-1.tests
2004-01-04-2.tests
2004-01-06-2.tests
2004-01-07-2.tests
2004-01-07-3.tests
2004-01-07-4.tests
2004-01-08-1.tests
2004-01-10-1.tests
2004-01-10-4.tests
2004-01-11-3.tests
2004-01-12-1.tests
2004-01-12-2.tests
while for all other runs the test works. Note that from 2003-10-28, when the
acats tests were first run on this machine through 2003-11-17 there is no
appearance of that FAIL. The same goes for the period from 2003-11-24 through
2003-12-30. There were no changes of the build machine.
> I am very skeptical of any regression reports I get from this address.
> The bug reports are flawed. They are incomplete, ambiguous, and
> sometimes contradictory. There is no contact address, so there is no
> human to get clarifications from.
Hmm. The people around here confirm my strong belief that indeed I am human ;-)
> My suspicion is that there is something wrong with the Ada testcase or
> the Ada compiler, and that an Ada expert really needs to be looking at
> this instead of asking non-Ada people to look at it.
Yes, this has been brought up before, see
http://gcc.gnu.org/ml/gcc-regression/2004-01/msg00005.html
I'm going to wait whether this still happens after
http://gcc.gnu.org/ml/gcc-patches/2004-01/msg00965.html
has gone in. If it does, I will ignore this FAIL in the future.
I CC'd Arnaud Charlet. Maybe he can provide further insight.
By the way, manual runs on x86_64-linux do not show this particular on/off FAIL,
but they show one for c954026.
Michael