Bug#848859: reproducible here

Paolo Greppi paolo.greppi at libpf.com
Fri Feb 10 08:27:27 UTC 2017


Hi, on a sid 2-core amd64 VM after locally building the package with "dpkg-buildpackage -A" I have launched 200 times the test suite alone with "time make check &> xxx" (it takes about 3 min each so it's quicker than building the entire package).

I get these 11 failures:

112:FAIL: make_gmm_test.pl
112:# FAIL:  1
131:FAIL: test_continuation.pl
131:# FAIL:  1
153:FAIL: make_gmm_test.pl
153:# FAIL:  1
160:FAIL: test_continuation.pl
160:# FAIL:  1
164:FAIL: make_gmm_test.pl
164:# FAIL:  1
185:FAIL: test_continuation.pl
185:# FAIL:  1
197:FAIL: test_continuation.pl
197:# FAIL:  1
35:FAIL: make_gmm_test.pl
35:# FAIL:  1
48:FAIL: test_continuation.pl
48:# FAIL:  1
85:FAIL: make_gmm_test.pl
85:# FAIL:  1
95:FAIL: make_gmm_test.pl
95:# FAIL:  1

So the test failure rate is indeed 5%.

I'll now proceed to forward upstream, but I have to relauch everything with set -e so that I can send the test artifacts (such as auto_gmm_torture15_sub.cc produced in directory "tests" in the case of your last test log) so that they can reproduce.

Paolo



More information about the debian-science-maintainers mailing list