[openstack-dev] migration to flake8

Andrew Laski andrew.laski at rackspace.com
Thu Apr 4 17:17:07 UTC 2013

On 04/04/13 at 12:00pm, Dolph Mathews wrote:
>i'm certainly in favor, as I already use flake8 in vim.

Me too, for the same reason.

>It's just an annoyance to me, but the only gotcha I've found is that I have
>no idea how to tell it to specifically ignore '_' being undefined (for

alaski at crypt ~ $ cat ~/.config/pep8                                                                                                         
builtins = _

but this only works for later versions of pep8, so I have to comment it 
out for running unit tests, which is another reason I'm in favor of the 

>On Thursday, April 4, 2013, Monty Taylor wrote:
>> Hey all!
>> I was going to chat with folks about this at the summit, but it seems to
>> be gaining momentum anyway, so there may not be a lot of pushback.
>> Proposal:
>> Replace pep8 and pyflakes and pylint and hacking.py in our testing with
>> flake8
>> Rationale:
>> flake8 is the new pluggable hotness for this sort of thing. Out of the
>> box it runs pep8 and pyflakes. The combo of those two winds up being
>> very consistent and easy to deal with, so the extra goodness we get from
>> pylint isn't really needed anymore. pyflakes regularly catches actual
>> bugs for me, so rather than being just more style stuff, I'm pretty
>> pleased with it as a useful bit of testing. It's also totally possible
>> and easy to write pyflakes clean code.
>> flake8 supports reading options from tox.ini, which we already have in
>> our trees - even if you aren't using tox to run it. So we can set
>> project options easily (like ignore=E125) and just running flake8 from
>> the command line - or inside of vim/emacs - will work with the right
>> settings.
>> pyflakes issues show up prefixed by "F" and each one can be ignored now
>> just like with pep8.
>> flake8 also supports configurable additions to builtins, which lets us
>> deal with the normal pyflakes issue with how we do gettext.
>> I've also taken hacking.py from nova and turned it into a flake8 plugin,
>> which means we can use that without copying it around. The hacking
>> plugin is in openstack-dev/hacking in gerrit. If you install it, it will
>> inject into the flake8 workflow and show its errors prefixed with "H".
>> If there is a massive trainwreck on a line and pyflakes/hacking just
>> can't deal with reality, you can add a # noqa tag to the line and it'll
>> be skipped. We should probably be very limited in our usage of those.
>> End result: no need for wrapper scripts around pep8, pyflakes or pylint.
>> And a single, easy to deal with static-analysis check.
>> Several of the infra projects have switched already, Maru just put up a
>> quantum patch and I've got a ceilometer patch up for the switch. Patches
>> for everyone else should be super easy.
>> Thoughts?
>> Monty
>> _______________________________________________
>> OpenStack-dev mailing list
>> OpenStack-dev at lists.openstack.org <javascript:;>
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

>OpenStack-dev mailing list
>OpenStack-dev at lists.openstack.org

More information about the OpenStack-dev mailing list