nedbat
commented
7 years ago I'm not sure how to run with unicode-nazi and have it finish the test suite. For me, the whole test run abruptly ends with:
Traceback (most recent call last):
File "/usr/local/pythonz/pythons/CPython-2.7.13/lib/python2.7/runpy.py", line 174, in _run_module_as_main
"__main__", fname, loader, pkg_name)
File "/usr/local/pythonz/pythons/CPython-2.7.13/lib/python2.7/runpy.py", line 72, in _run_code
exec code in run_globals
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/unicodenazi.py", line 128, in <module>
main()
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/unicodenazi.py", line 119, in main
execfile(sys.argv[0], main_mod.__dict__)
File "/usr/local/virtualenvs/coverage/bin/pytest", line 11, in <module>
sys.exit(main())
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/config.py", line 58, in main
return config.hook.pytest_cmdline_main(config=config)
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/vendored_packages/pluggy.py", line 745, in __call__
return self._hookexec(self, self._nonwrappers + self._wrappers, kwargs)
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/vendored_packages/pluggy.py", line 339, in _hookexec
return self._inner_hookexec(hook, methods, kwargs)
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/vendored_packages/pluggy.py", line 334, in <lambda>
_MultiCall(methods, kwargs, hook.spec_opts).execute()
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/vendored_packages/pluggy.py", line 614, in execute
res = hook_impl.function(*args)
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/main.py", line 139, in pytest_cmdline_main
return wrap_session(config, _main)
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/main.py", line 133, in wrap_session
exitstatus=session.exitstatus)
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/vendored_packages/pluggy.py", line 745, in __call__
return self._hookexec(self, self._nonwrappers + self._wrappers, kwargs)
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/vendored_packages/pluggy.py", line 339, in _hookexec
return self._inner_hookexec(hook, methods, kwargs)
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/vendored_packages/pluggy.py", line 334, in <lambda>
_MultiCall(methods, kwargs, hook.spec_opts).execute()
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/vendored_packages/pluggy.py", line 613, in execute
return _wrapped_call(hook_impl.function(*args), self.execute)
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/vendored_packages/pluggy.py", line 250, in _wrapped_call
wrap_controller.send(call_outcome)
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/terminal.py", line 406, in pytest_sessionfinish
outcome.get_result()
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/vendored_packages/pluggy.py", line 280, in get_result
_reraise(*ex) # noqa
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/vendored_packages/pluggy.py", line 265, in __init__
self.result = func()
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/_pytest/vendored_packages/pluggy.py", line 614, in execute
res = hook_impl.function(*args)
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/flaky/flaky_pytest_plugin.py", line 214, in pytest_sessionfinish
if hasattr(self.config, 'slaveoutput'):
File "/Users/ned/coverage/trunk/.tox/py27/lib/python2.7/site-packages/unicodenazi.py", line 27, in warning_encode
warnings.warn(UnicodeWarning('Implicit conversion of unicode to str'),
AttributeError: 'NoneType' object has no attribute 'warn'
Originally reported by David MacIver (Bitbucket: davidmaciver, GitHub: Unknown)
Context: As discussed I am currently doing terrible things to Coverage.
As part of the Hypothesis build we run some of our tests on Python 2 under unicodenazi to check for implicit conversions between unicode and str and turn them into errors. This avoids finding encoding problems the hard way.
When I added coverage code I found a bunch of problems where it was mixing unicode and strings. The source (for me) seems to be that canonical_filename returns unicode, while Python file system operations expect paths to be a str.
In my case the fix was just that I encoded the output of
canonical_filenamewithsys.getfilesystemencoding(). I don't know if that's actually sufficient, but it seems to work.I can't say for sure if this is likely to be causing any problems. It might lead to weird errors when running in directories where there are non-ascii characters in the path? Or it might be a harmless false positive. I don't know, but I thought it worth mentioning.