Get 'tox -e lint' woking again

[tseaver]

Closes #303.

[dataflake]

I wonder why this is using pylint at all instead of flake8.

[tseaver]

@dataflake

I wonder why this is using pylint at all instead of flake8.

• For the config comments in the code, I reckon it is what @jimfulton calls "hysterical raisins."
• The GHA config that has installs but then disables a pylint check seems to be down to the c-code config in the meta repositories:

$ grep -l pylint zope.*/.github/workflows/tests.yml 
zope.container/.github/workflows/tests.yml
zope.hookable/.github/workflows/tests.yml
zope.i18nmessageid/.github/workflows/tests.yml
zope.index/.github/workflows/tests.yml
zope.interface/.github/workflows/tests.yml
zope.proxy/.github/workflows/tests.yml
zope.security/.github/workflows/tests.yml

I'd be in favor of ripping out all of the config comments mentioning pylint. I'd really prefer as well that we get the meta config updated to actually run isort and flake8 in the same way that tox.ini runs them.

[tseaver]

Another comment, especiialy after digging through the tests to get flake8 happy last night: the overall quality of the tests varies a lot, and the amount of mindless boilerplate in some of them is a bit shocking: my twenty-year older self doesn't appreciate the lack of factoring out fixtures, etc., at all, and even less the verbosity of the unittest.TestCase-style. If I had my "druthers," they would magically be replaced by well-factored, readable tests using the pytest features.