New submission from STINNER Victor: It seems like an unit test of test_thread leaks a process. A bisection found the following 2 tests are enough to get the warning:
* test.test_thread.TestForkInThread.test_forkinthread * test.test_thread.LockTests.test_acquire_contended I bet that test_forkinthread() doesn't wait correctly for the child process (fork) completion using os.waitpid() or something like that. Example 1 (from Refleak buildbot): haypo@selma$ ./python -m test -R 3:3 test_thread Run tests sequentially 0:00:00 load avg: 0.55 [1/1] test_thread beginning 6 repetitions 123456 Warning -- reap_children() reaped child process 6984 .Warning -- reap_children() reaped child process 7054 .Warning -- reap_children() reaped child process 7125 .Warning -- reap_children() reaped child process 7195 .Warning -- reap_children() reaped child process 7266 .Warning -- reap_children() reaped child process 7336 . 1 test OK. Total duration: 5 sec Tests result: SUCCESS Example 2 (only run the 2 methods): haypo@selma$ ./python -m test -m test.test_thread.TestForkInThread.test_forkinthread -m test.test_thread.LockTests.test_acquire_contended -v test_thread test_thread 0:00:00 load avg: 0.27 [1/2] test_thread test_acquire_contended (test.test_thread.LockTests) ... ok test_forkinthread (test.test_thread.TestForkInThread) ... ok ---------------------------------------------------------------------- Ran 2 tests in 0.037s OK 0:00:00 load avg: 0.27 [2/2] test_thread test_acquire_contended (test.test_thread.LockTests) ... Warning -- reap_children() reaped child process 10613 ok test_forkinthread (test.test_thread.TestForkInThread) ... ok ---------------------------------------------------------------------- Ran 2 tests in 0.036s OK test_thread failed (env changed) ---------- keywords: easy messages: 299958 nosy: haypo priority: normal severity: normal status: open title: [EASY] test_thread: Warning -- reap_children() reaped child process versions: Python 3.6, Python 3.7 _______________________________________ Python tracker <rep...@bugs.python.org> <http://bugs.python.org/issue31150> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
