tox -p raises AttributeError at the end of the run

[Russ Allbery]

Whenever I run tox -p all for a project that uses tox-docker to start Docker containers, I get the following exception backtrace at the end of the run. This doesn't interfere with the run; all of the test environments report success before this happens. However, it does cause tox to exit with a non-zero status.

This seems to only affect tox -p, not running environments individually or sequentially using tox without arguments or tox -e. It happens with both Python 3.9 and Python 3.10.

Traceback (most recent call last):
  File "/home/eagle/dvl/venvs/safir/bin/tox", line 8, in <module>
    sys.exit(cmdline())
  File "/home/eagle/dvl/venvs/safir/lib/python3.9/site-packages/tox/session/__init__.py", line 44, in cmdline
    main(args)
  File "/home/eagle/dvl/venvs/safir/lib/python3.9/site-packages/tox/session/__init__.py", line 69, in main
    exit_code = session.runcommand()
  File "/home/eagle/dvl/venvs/safir/lib/python3.9/site-packages/tox/session/__init__.py", line 197, in runcommand
    return self.subcommand_test()
  File "/usr/lib/python3.9/contextlib.py", line 126, in __exit__
    next(self.gen)
  File "/home/eagle/dvl/venvs/safir/lib/python3.9/site-packages/tox/session/__init__.py", line 205, in cleanup
    self.hook.tox_cleanup(session=self)
  File "/home/eagle/dvl/venvs/safir/lib/python3.9/site-packages/pluggy/hooks.py", line 286, in __call__
    return self._hookexec(self, self.get_hookimpls(), kwargs)
  File "/home/eagle/dvl/venvs/safir/lib/python3.9/site-packages/pluggy/manager.py", line 93, in _hookexec
    return self._inner_hookexec(hook, methods, kwargs)
  File "/home/eagle/dvl/venvs/safir/lib/python3.9/site-packages/pluggy/manager.py", line 84, in <lambda>
    self._inner_hookexec = lambda hook, methods, kwargs: hook.multicall(
  File "/home/eagle/dvl/venvs/safir/lib/python3.9/site-packages/pluggy/callers.py", line 208, in _multicall
    return outcome.get_result()
  File "/home/eagle/dvl/venvs/safir/lib/python3.9/site-packages/pluggy/callers.py", line 80, in get_result
    raise ex[1].with_traceback(ex[2])
  File "/home/eagle/dvl/venvs/safir/lib/python3.9/site-packages/pluggy/callers.py", line 187, in _multicall
    res = hook_impl.function(*args)
  File "/home/eagle/dvl/venvs/safir/lib/python3.9/site-packages/tox_docker/__init__.py", line 370, in tox_cleanup
    stop_containers(venv)
  File "/home/eagle/dvl/venvs/safir/lib/python3.9/site-packages/tox_docker/__init__.py", line 381, in stop_containers
    for container_name, container in envconfig._docker_containers.items():
  File "/home/eagle/dvl/venvs/safir/lib/python3.9/site-packages/tox/config/__init__.py", line 1042, in __getattribute__
    rv = object.__getattribute__(self, name)
AttributeError: 'TestenvConfig' object has no attribute '_docker_containers'

[Dan Crosta]

Looks like this is using tox 3.x ... would you be willing to upgrade to tox4 (and tox-docker 4.0.0a2 -- install with --pre) to see if the issue persists there?

[Russ Allbery]

Ah, thank you, I hadn't noticed tox4. I can confirm that it does fix this issue. (It has a couple of other issues, one of which I see was already reported as tox-dev/tox#2373 and the other of which I'll report in a separate issue.)

[Punleuk Oum]

I am having the same issue with tox-docker 3.1.0 and tox 3.28.0.
PyCharm requires the use of tox < 4.0.0 due to import issues in the tox plugin.

tox-docker==4.0.0a2 works, but I had to add
PATH=/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin to the environment configuration, because docker-credential-desktop is installed in /usr/local/bin