Re: [Boost-testing] Re: [RSI-Sol] Exit status 132?

From: Jeff Garland (jeff_at_[hidden])
Date: 2005-05-08 15:35:47

On Sun, 08 May 2005 22:22:27 +0200, Martin Wille wrote
> >>JOAQUIN LOPEZ MU?Z wrote:
> >>
> >>>One of the tests for Boost.MultiIndex is failing
> >>>with the mentioned exit status, as shown in
> >>>http://tinyurl.com/cplxm. Dopes anyone know
> >>>what this exit code means?
> >>
> >>That's signal 4, Illegal Instruction.
> >>
>....
>....
> > * All other tests work OK for RSI-Sol.
> > * This test works for every other platform.
> > * It used to work OK as of last week.
> > * No significant change (that I'm aware of)
> > has recently been done on the direct or
> > indirect codebase affecting this test.
> 
> This *could* be a buildbot related issue. Of course, there could be
> other reasons, e.g. the compiler produces binaries for a CPU other than
> the one the tests run on (I recently got caught by that problem when 
> I tried to create binaries for a VIA C3 based system. Symptom: everything
> worked except for one program).

I suspect that this might be a 'release mode' issue.  AFAIK Rene is running
his tests in release mode instead of the default debug mode.  So it the best
test might be to shut off optimization on Rene's test to see if that fixes the
problem.  

Jeff