fyi: The AIX bots are in "failure" state for over a week.
My bot has always failed die to issues with xlc versus gcc. The other AIX bot is not something I own. I submitted a PR, with the idea that optional mods (depending on third-party, i.e., not supplied by AIX) software would not be installed by default and clear the last "warnings" that kept it from saying OK to the CI process.
re: my bot - I am switching to gcc for the bot. XLC is not giving any information the project is interested in - and I hope having a bot return, eventually, a clear status - is of some benefit.
In any case, there appears to be a regression that started about 8 days ago - see failure: https://buildbot.python.org/all/#/builders/10/builds/2603 and "warning" https://buildbot.python.org/all/#/builders/10/builds/2602
regards,
Michael
Currently, the main issue on AIX is https://bugs.python.org/issue36084 which broke AIX. I asked to either fix the issue or revert.
Victor
Le lun. 20 mai 2019 à 23:07, Michael <aixtools@felt.demon.nl> a écrit :
My bot has always failed die to issues with xlc versus gcc. The other AIX bot is not something I own. I submitted a PR, with the idea that optional mods (depending on third-party, i.e., not supplied by AIX) software would not be installed by default and clear the last "warnings" that kept it from saying OK to the CI process.
re: my bot - I am switching to gcc for the bot. XLC is not giving any information the project is interested in - and I hope having a bot return, eventually, a clear status - is of some benefit.
In any case, there appears to be a regression that started about 8 days ago - see failure: https://buildbot.python.org/all/#/builders/10/builds/2603 and "warning" https://buildbot.python.org/all/#/builders/10/builds/2602
regards,
Michael
Python-Buildbots mailing list Python-Buildbots@python.org https://mail.python.org/mailman/listinfo/python-buildbots
-- Night gathers, and now my watch begins. It shall not end until my death.
On 20/05/2019 22:50, Michael wrote: Short part - re: one test seems to be a clear regression (https://bugs.python.org/issue36816) - that I hit when building from github manually.
Two tests - are regressions in both bots: test_asyncio test_threading
regards,
Michael
Python-Buildbots mailing list Python-Buildbots@python.org https://mail.python.org/mailman/listinfo/python-buildbots
Le mar. 21 mai 2019 à 12:32, Michael <aixtools@felt.demon.nl> a écrit :
On 20/05/2019 22:50, Michael wrote: Short part - re: one test seems to be a clear regression (https://bugs.python.org/issue36816) - that I hit when building from github manually.
This one is now fixed, no?
Two tests - are regressions in both bots: test_asyncio test_threading
Do you have references to these issues? All bugs must be tracked in bugs.python.org.
I'm aware of 2 issues impacting AIX:
test multiprocessing: test_rapid_restart() crash on AIX : https://bugs.python.org/issue36752
test_thread leaks a core dump on PPC64 AIX 3.x : https://bugs.python.org/issue36273
Victor
On 21/05/2019 12:35, Victor Stinner wrote:
Le mar. 21 mai 2019 à 12:32, Michael <aixtools@felt.demon.nl> a écrit :
On 20/05/2019 22:50, Michael wrote: Short part - re: one test seems to be a clear regression (https://bugs.python.org/issue36816) - that I hit when building from github manually. This one is now fixed, no?
I'll check - I thought I had cloned from yesterday (manually), but the top log entry is from 8 May. Updating my fork - thought I had cloned from cpython... Sigh!
So, yes, fixed. (blush!)
Will do a fresh build manually - but the last bot results showed them both as "broken" - test_asyncio test_threading
Two tests - are regressions in both bots: test_asyncio test_threading Do you have references to these issues? All bugs must be tracked in bugs.python.org.
I'm aware of 2 issues impacting AIX:
test multiprocessing: test_rapid_restart() crash on AIX : https://bugs.python.org/issue36752
test_thread leaks a core dump on PPC64 AIX 3.x : https://bugs.python.org/issue36273
I'll look at those. I was also looking at
https://bugs.python.org/issue36084 - as AIX has native TID values (PID are always even numbers, TID are always odd).
I'll reply to the appropriate issue thread - and "close" here.
Victor
participants (2)
-
Michael -
Victor Stinner