Update test_processor.py

[pvinci]

Workaround for change in behavior in python3.6
See: https://bugs.python.org/issue29130

[coveralls]

Coverage remained the same at 68.114% when pulling a47bb37 on pvinci:patch-1 into ff9b688 on mtreinish:master.

[mtreinish]

I'm not sure about this, based on the bug linked in the commit message it looks like starting in python 3.6 python will exit will exit with 120 if there is an error raised during the closing of the interpreter. The example in that bug was that flush() was called on sys.stdout or sys.stderr. This seems like an actual error that we want to report to the user since it's something wrong with the code. Especially if this is happening during import of a module (which is all that happens during discovery).

Was there a use case or issue you were hitting that required making this change? I'm just trying to understand the reason behind why we would want to do this and maybe we can come up with an alternative solution.

[pvinci]

Running tests with stdout/stderr redirection via tox/stestr. Tests run under py27 but fail with 120 error under py36. It may not be your issue in the end, but if we can't run the tests, we can't fix the underlying issue.

…

On Thu, Dec 5, 2019 at 10:30 AM Matthew Treinish ***@***.***> wrote: ***@***.**** commented on this pull request. I'm not sure about this, based on the bug linked in the commit message it looks like starting in python 3.6 python will exit will exit with 120 if there is an error raised during the closing of the interpreter. The example in that bug was that flush() was called on sys.stdout or sys.stderr. This seems like an actual error that we want to report to the user since it's something wrong with the code. Especially if this is happening during import of a module (which is all that happens during discovery). Was there a use case or issue you were hitting that required making this change? I'm just trying to understand the reason behind why we would want to do this and maybe we can come up with an alternative solution. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#283?email_source=notifications&email_token=AAIQDL53HB3YMTCMDHH4KHTQXENBBA5CNFSM4JOC7V2KYY3PNVWWK3TUL52HS4DFWFIHK3DMKJSXC5LFON2FEZLWNFSXPKTDN5WW2ZLOORPWSZGOCODVWHI#pullrequestreview-327637789>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAIQDL7ALNDDQC3ZNIAZQP3QXENBBANCNFSM4JOC7V2A> .