You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Warning -- threading._dangling was modified by test_asyncio on FreeBSD 9 and 10 #74466
Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.
Show more details
GitHub fields:
assignee=Noneclosed_at=<Date2017-06-30.15:22:29.733>created_at=<Date2017-05-05.09:08:35.083>labels= ['3.7', 'tests', 'expert-asyncio']
title='Warning -- threading._dangling was modified by test_asyncio on FreeBSD 9 and 10'updated_at=<Date2017-06-30.15:22:29.732>user='https://github.com/vstinner'
Warning -- threading._dangling was modified by test_asyncio
Before: <_weakrefset.WeakSet object at 0x805f5d810>
After: <_weakrefset.WeakSet object at 0x807929948>
changed the title [-]Warning -- threading._dangling was modified by test_asyncio on FreeBSD 9[/-][+]Warning -- threading._dangling was modified by test_asyncio on FreeBSD 9 and 10[/+]on May 9, 2017
The commit b903067 fixes one leaking thread. I don't know if it's enough.
Koobs gave me a SSH access to his FreeBSD buildbots (9, 10 and CURRENT), but I failed to reproduce the warning issue :-( I tried to run my system_load.py to run the tests when the system load is 5 or higher, but it doesn't help. I tried to run test_asyncio 4 times in parallel, but I was still unable to reproduce the bug.
The commit b903067 not only fixes a bug (cleanup correctly one event loop), but it now also checks after *each* test that we don't leak threads. At least, if the bug reoccurs and the test is run in verbose mode, we will get the name of the test which leaked threads!
Activity
vstinner commentedon May 5, 2017
On the FreeBSD 9 buildbot, I saw this warning:
Warning -- threading._dangling was modified by test_asyncio
http://buildbot.python.org/all/builders/AMD64%20FreeBSD%209.x%203.6/builds/124/steps/test/logs/warnings%20%281%29
http://buildbot.python.org/all/builders/AMD64%20FreeBSD%209.x%203.5/builds/78/steps/test/logs/warnings%20%281%29
vstinner commentedon May 9, 2017
Warning also seen on FreeBSD 10:
http://buildbot.python.org/all/builders/AMD64%20FreeBSD%2010.x%20Shared%203.x/builds/240/steps/test/logs/stdio
Warning -- threading._dangling was modified by test_asyncio
Before: <_weakrefset.WeakSet object at 0x805f5d810>
After: <_weakrefset.WeakSet object at 0x807929948>
[-]Warning -- threading._dangling was modified by test_asyncio on FreeBSD 9[/-][+]Warning -- threading._dangling was modified by test_asyncio on FreeBSD 9 and 10[/+]vstinner commentedon May 15, 2017
Xiang Zhang just fixed a bug to release resources in test_asyncio: bpo-30110, commit 981096f. No idea if it's enough to fix the warning.
vstinner commentedon Jun 30, 2017
New changeset b903067 by Victor Stinner in branch 'master':
bpo-30280: Cleanup threads in ayncio tests (bpo-2501)
b903067
vstinner commentedon Jun 30, 2017
The commit b903067 fixes one leaking thread. I don't know if it's enough.
Koobs gave me a SSH access to his FreeBSD buildbots (9, 10 and CURRENT), but I failed to reproduce the warning issue :-( I tried to run my system_load.py to run the tests when the system load is 5 or higher, but it doesn't help. I tried to run test_asyncio 4 times in parallel, but I was still unable to reproduce the bug.
The commit b903067 not only fixes a bug (cleanup correctly one event loop), but it now also checks after *each* test that we don't leak threads. At least, if the bug reoccurs and the test is run in verbose mode, we will get the name of the test which leaked threads!
vstinner commentedon Jun 30, 2017
New changeset 0e0bc87 by Victor Stinner in branch '3.5':
bpo-30280: Cleanup threads in ayncio tests (bpo-2501) (bpo-2512)
0e0bc87
vstinner commentedon Jun 30, 2017
New changeset 23caf8c by Victor Stinner in branch '3.6':
bpo-30280: Cleanup threads in ayncio tests (bpo-2501) (bpo-2511)
23caf8c