Prepare master branch for development of 3.5.0

See merge request !200

Allow spaces in # noqa lists

Closes #356

See merge request !196

To match our new configuration file format and its allowance for spaces
in the list

Closes #356

Fix regression in --select behaviour in v3.4.0

Closes #354

See merge request !195

Previously we were hitting the catch-all return that decided the error
was selected. Instead, let's add specific logic since we now know how we
would hit that and we actually don't want it to be selected.

Closes #354

Use extras_require instead of conditional code.

See merge request !191

Removed unused imported module `os` from flake8.main.git._HOOK_TEMPLATE

See merge request !194

Use of extras_require produces correct package metadata for flake8
regardless of the version of Python the package is built with.
The conditional code produces different metadata for Python 3 and 2.
The latest versions of flake8 were released with Python 3 and metadata
did not contain two dependencies necessary for use with Python 2.7.

See https://gitlab.com/pycqa/flake8/issues/341 for details.

Account for users of setuptools<18 that still need the conditional code.

Refactor decision logic into its own object

See merge request !190

This was mostly to prove to myself that we could possibly reach that
return more than actually covering that return. This just shows how
gnarly this logic actually is. I wish there were a better way to write
it.

Also, this further highlights why naming methods is so hard. I can't
think of a better name for 'more_specific_decision_for' that isn't
wildly long and unnecessarily verbose.

Move all Violation related methods from the StyleGuide to our Violation
class.

Rename `decision_for` to `more_specific_decision_for` and
`should_report_error` to `decision_for`.

Convert it to test the DecisionEngine class directly and put them in
their own file.

In dealing with the decision logic in the StyleGuide recently I
recognized that the logic really doesn't belong strictly on the
StyleGuide. A separate object makes perfect sense especially from the
perspective of testability. This is a minor refactor intended solely to
facilitate further testing and perhaps making the logic easier to
understand for others.

Add documentation for using --select and --ignore seperately and together

See merge request !188

In order to write this documentation, we need to also add documentation
for the --max-complexity option which was heretofore undocumented.

Further improve our handling of --select and --ignore

Closes #318

See merge request !187

Add test to cover branch of decision logic we were not previously
exercising

There was a *very* subtle bug in how we handle blanket select statements
with error codes that are in our DEFAULT_IGNORE. In the specific case,
users were specifying ``--select E`` and E126 was not being properly
reported. This unveiled further logic bugs as we refined this.

Closes #318

Prefer iter(dict) instead of dict.keys()

See merge request !186

They are equivalent for iterating so remove the additional function
call.

Pattern identified as outdated by Lennart Regebro's PyCon 2017 talk
"Prehistoric Patterns in Python"

https://www.youtube.com/watch?v=V5-JH23Vk0I

Avoid bugs when users accidentally forget a comma

Closes #329

See merge request !185

Closes #329