https://issues.apache.org/SpamAssassin/show_bug.cgi?id=6949
--- Comment #3 from Mark Martinec <[email protected]> --- > Solved the t/bayesdbm_flock.t failures mystery: > Bug 6949: fix t/bayesdbm_flock.t: learner leaves timer running, > which can abort sa-learn test runs longer than 10 seconds, > resulting in a test failure This opens up a question if untie_db() is supposed to leave timer running (with a very short timeout of 1 second), and if $sa->{bayes_scanner}->learn() is supposed to leave it running, set at 10 second. So far this hasn't been affecting other operations, so I guess we can ignore it for the time being. Regarding the t/spamd_prefork_stress_4.t, it seems to run fine on that test machine as long as the number of concurrent tasks does not reach the number of spamd child processes, so this looks like hitting some resource limit (other tests on the same host have been observed to terminate with "Out of Memory" every now and them. So I'd not worry about this. So I guess this ticket can now be closed. -- You are receiving this mail because: You are the assignee for the bug.
