Use flake8 for sanity testing

[willthames]

SUMMARY

flake8 catches more errors than pep8, so will result in fewer bugs making the codebase

Would have caught the error introduced in #20214 addressed by #22818

ERROR: Found 1 pep8 issue(s) which need to be resolved:
ERROR: lib/ansible/modules/cloud/amazon/ec2_vpc_nat_gateway_facts.py:118:13: F821 undefined name 'camel_dict_to_snake_dict'

ISSUE TYPE

• Bugfix Pull Request

COMPONENT NAME

test/sanity

ANSIBLE VERSION

ansible 2.4.0 (flake8_tests e6e557cd76) last updated 2017/03/21 15:06:56 (GMT +1000)
  config file = /etc/ansible/ansible.cfg
  configured module search path = Default w/o overrides
  python version = 2.7.13 (default, Jan 12 2017, 17:59:37) [GCC 6.3.1 20161221 (Red Hat 6.3.1-1)]

ADDITIONAL INFORMATION

[mattclay]

Since we're still working through cleaning up existing issues caught by pep8, this is a little premature. We'll also need to discuss which of the new tests need to be excluded.

[mattclay]

I've added this to the TWG agenda for further discussion.

[willthames]

@mattclay fair enough, it does introduce more checks than I was thinking of

http://flake8.pycqa.org/en/latest/user/error-codes.html

There are some tests in there that simplify existing checks (e.g. module importing '*'), but mostly I wanted to avoid catchable bugs, not make things more complicated.

[sivel]

I feel like it's a bit too early for this. We still have the 5 legacy pep8 issues to resolve, and 30 current.

In some cases we have some overlapping checks. import * is allowed in "legacy" modules but not new, and validate-modules is handling that.

My feeling is that this would be a change targeted to 2.5 at earliest.

I'm all for it, I just don't want it distract us from our other already planned efforts.

[mattclay]

This was discussed at the March 23rd TWG meeting. The general consensus is that flake8 will be an improvement over pep8. As long as the change doesn't interfere with the ongoing work to clean up the remaining issues currently reported by pep8, this should be fine.

[willthames]

I suspect that work on pylint will have similar effect, so I'll close this one.