Yes, it wasn't an exact reversion of just that change (though perhaps that would be a better way to go about it). The logic that you're looking for (if I'm reading this right) is that last bit of the diff you posted ( https://github.com/jcrocholl/pep8/compare/1.5.7...435d1cbf995a659a82d1d4b42d...) from lines (numbered off master) 2041 - 2101. Namely, the former assignment of ``pep8style = StyleGuide(parse_argv=True, config_file=True)`` used to take the config_file=True argument and then ``process_options(...)`` would use that to set the config_file to DEFAULT_CONFIG (now named USER_CONFIG in master). Does that clear up the piece that you were missing? ~ Ian Lee On Tue, Mar 17, 2015 at 10:32 PM, Ian Cordasco <graffatcolmingov@gmail.com> wrote:

On Wed, Mar 18, 2015 at 12:17 AM, Ian Lee <ianlee1521@gmail.com> wrote:

...

Hi All --

I just pushed a commit [1] to the pep8 repo that contains what I hope is a fix for the issues with pep8 / flake8 config files that arose from the way that pep8 #368 was handled, e.g. [2] (I think, I'm getting 500 errors when I try to few issues on the flake8 repo, but I'm hoping that is transient).

This led to flake8 2.4.0 pinning to pep8 < 1.6 in it's latest release [3].

I'm hoping that folks can test out there configurations and ways of running flake8 / pep8 with various configuration files (user and project specific configs) and please report any issues to me.

Thanks in advance!

[1] https://github.com/jcrocholl/pep8/commit/435d1cbf995a659a82d1d4b42d25e345955... [2] https://gitlab.com/pycqa/flake8/issues/35 [3] https://pypi.python.org/pypi/flake8/2.4.0

~ Ian Lee

_______________________________________________ code-quality mailing list code-quality@python.org https://mail.python.org/mailman/listinfo/code-quality

Hey Ian,

Thanks for working on this. Looking at https://github.com/jcrocholl/pep8/compare/1.5.7...435d1cbf995a659a82d1d4b42d... it's not identical to 1.5.7. By default options.config_file was previously None. So the USER_CONFIG wasn't checked/loaded first ( https://github.com/jcrocholl/pep8/compare/1.5.7...435d1cbf995a659a82d1d4b42d...). Am I missing logic where options.config was changed to default to USER_CONFIG? (I know the binding was previously called user_conf, but I'm not entirely convinced this is doing the same thing in effect.)

Cheers, Ian

Not that early here either, so I could also be confusing myself, but I don't think this is an issue. flake8 uses the flake8.engine.get_style_guide() function [1] which takes in ``config_file=DEFAULT_CONFIG`` in just about every place it's called (e.g. [2]) which in pep8 gets passed explicitly to ``process_options()``[3] with the value that is set in ``pep8.StyleGuide.__init__()`` [4] [1] https://gitlab.com/pycqa/flake8/blob/master/flake8/engine.py#L107 [2] https://gitlab.com/pycqa/flake8/blob/master/flake8/main.py#L24 [3] https://github.com/jcrocholl/pep8/blob/master/pep8.py#L1787-L1788 [4] https://github.com/jcrocholl/pep8/blob/master/pep8.py#L1782 ~ Ian Lee On Tue, Mar 17, 2015 at 11:00 PM, Ian Cordasco <graffatcolmingov@gmail.com> wrote:

On Wed, Mar 18, 2015 at 12:50 AM, Ian Lee <ianlee1521@gmail.com> wrote:

...

Yes, it wasn't an exact reversion of just that change (though perhaps that would be a better way to go about it).

The logic that you're looking for (if I'm reading this right) is that last bit of the diff you posted ( https://github.com/jcrocholl/pep8/compare/1.5.7...435d1cbf995a659a82d1d4b42d...) from lines (numbered off master) 2041 - 2101. Namely, the former assignment of ``pep8style = StyleGuide(parse_argv=True, config_file=True)`` used to take the config_file=True argument and then ``process_options(...)`` would use that to set the config_file to DEFAULT_CONFIG (now named USER_CONFIG in master).

Does that clear up the piece that you were missing?

~ Ian Lee

On Tue, Mar 17, 2015 at 10:32 PM, Ian Cordasco < graffatcolmingov@gmail.com> wrote:

...

On Wed, Mar 18, 2015 at 12:17 AM, Ian Lee <ianlee1521@gmail.com> wrote:

...

Hi All --

I just pushed a commit [1] to the pep8 repo that contains what I hope is a fix for the issues with pep8 / flake8 config files that arose from the way that pep8 #368 was handled, e.g. [2] (I think, I'm getting 500 errors when I try to few issues on the flake8 repo, but I'm hoping that is transient).

This led to flake8 2.4.0 pinning to pep8 < 1.6 in it's latest release [3].

I'm hoping that folks can test out there configurations and ways of running flake8 / pep8 with various configuration files (user and project specific configs) and please report any issues to me.

Thanks in advance!

[1] https://github.com/jcrocholl/pep8/commit/435d1cbf995a659a82d1d4b42d25e345955... [2] https://gitlab.com/pycqa/flake8/issues/35 [3] https://pypi.python.org/pypi/flake8/2.4.0

~ Ian Lee

_______________________________________________ code-quality mailing list code-quality@python.org https://mail.python.org/mailman/listinfo/code-quality

Hey Ian,

Thanks for working on this. Looking at https://github.com/jcrocholl/pep8/compare/1.5.7...435d1cbf995a659a82d1d4b42d... it's not identical to 1.5.7. By default options.config_file was previously None. So the USER_CONFIG wasn't checked/loaded first ( https://github.com/jcrocholl/pep8/compare/1.5.7...435d1cbf995a659a82d1d4b42d...). Am I missing logic where options.config was changed to default to USER_CONFIG? (I know the binding was previously called user_conf, but I'm not entirely convinced this is doing the same thing in effect.)

Cheers, Ian

So it looks like the SyleGuide's API has changed in a fairly significant way then, no? Users aren't expected to pass `config_file=True` now? This means the next release of flake8 will either need to specify a higher minimum version or conditionally handle how it creates the StyleGuide based on the version of pep8. It's late here and I'm tired though, so I could be wrong, but it seems like this is somewhat backwards incompatible.