On 08/11/11 11:04, Maciej Bliziński wrote:
2011/8/11 Roberto Bagnara<bagnara@cs.unipr.it>:
On 08/11/11 10:21, Maciej Bliziński wrote:
Hi Roberto,
2011/8/11 Roberto Bagnara<bagnara@cs.unipr.it>:
can you please define the environment variable `PPL_NOISY_TESTS' to anything, just before running `bounds1' and `smashproduct1' by hand? Then please send us the output.
Here's the output: http://buildfarm.opencsw.org/~maciej/bounds1.log http://buildfarm.opencsw.org/~maciej/smashproduct1.log
Where are the segfaults?
Interestingly, segfaults go away when stdout goes to a pipe instead of the terminal. For example:
maciej@unstable9s :~/src/opencsw/pkg/ppl/trunk> PPL_NOISY_TESTS=wibble work/solaris9-sparc/build-isa-sparcv8/ppl-0.11.2/tests/Partially_Reduced_Product/smashproduct1
/dev/null Bus Error (core dumped) maciej@unstable9s :~/src/opencsw/pkg/ppl/trunk> PPL_NOISY_TESTS=wibble work/solaris9-sparc/build-isa-sparcv8/ppl-0.11.2/tests/Partially_Reduced_Product/smashproduct1 /dev/null | cat maciej@unstable9s :~/src/opencsw/pkg/ppl/trunk>
Perhaps the segfaults are a bit of a red herring; these two tests also fail, e.g:
maciej@unstable9s :~/src/opencsw/pkg/ppl/trunk> (work/solaris9-sparc/build-isa-sparcv8/ppl-0.11.2/tests/Partially_Reduced_Product/smashproduct1 || echo "Test failed" | cat) Test failed
We could investigate the failures first, and after fixing them, we might find out that the segfaults are gone too.
When a test fails in the PPL camp, the output with PPL_NOISY_TESTS ends with a list of the test functions that failed. I don't see that in the files you linked. I suggest you use gdb on smashproduct1 to find out more. Cheers, Roberto -- Prof. Roberto Bagnara Applied Formal Methods Laboratory Department of Mathematics, University of Parma, Italy http://www.cs.unipr.it/~bagnara/ mailto:bagnara@cs.unipr.it