Merge lp:~ewanmellor/swift/lp820185 into lp:~hudson-openstack/swift/trunk
Proposed by
Ewan Mellor
Status: | Merged | ||||
---|---|---|---|---|---|
Approved by: | John Dickinson | ||||
Approved revision: | 336 | ||||
Merged at revision: | 336 | ||||
Proposed branch: | lp:~ewanmellor/swift/lp820185 | ||||
Merge into: | lp:~hudson-openstack/swift/trunk | ||||
Diff against target: |
16 lines (+6/-0) 1 file modified
test/__init__.py (+6/-0) |
||||
To merge this branch: | bzr merge lp:~ewanmellor/swift/lp820185 | ||||
Related bugs: |
|
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
John Dickinson | Approve | ||
gholt (community) | Approve | ||
Review via email: mp+70253@code.launchpad.net |
Commit message
Bug #820185: Intermittent TypeError NoneType in atexit handler of unit tests
Work around what appears to be a Python standard library bug by silencing
exceptions in the logging library.
Description of the change
Bug #820185: Intermittent TypeError NoneType in atexit handler of unit tests
Work around what appears to be a Python standard library bug by silencing
exceptions in the logging library.
To post a comment you must log in.
I've seen this for quite some time in the builds as well, though not on my local machine as of yet. That particular stack trace doesn't seem to cause any failures outside of what it displays. Is this required to be fixed?
I'm worried the side-effect of not raising any exceptions with logging could cause us more and real problems (handlers not set correctly, bugs in formatters, etc.) As a quick example, this logging problem is completely silent and no one would realize a .critical() was even attempted.
import logging
logging. raiseExceptions = False addHandler( logging. StreamHandler( )) critical( 'test %s %s', ('one',))
logger = logging.getLogger()
logger.
logger.