populate pshell's env_help with docstring of values

[jessedhillon]

This changes pshell to check the docstring of values assigned to env, and if present, use it to as the value for env_help for that value instead of the repr. For example:

Before

Environment:
  app          The WSGI application.
  pp           <function pprint at 0x7f1b37960050>
  registry     Active Pyramid registry.
  request      Active request object.
  root         Root of the default resource tree.
  root_factory Default root factory used to create `root`.
  transction   <module 'transaction' from '/home/jesse/.virtualenvs/boutique/local/lib/python2.7/site-packages/transaction/__init__.pyc'>

After

Environment:
  app          The WSGI application.
  pp           Pretty-print a Python object to a stream [default is sys.stdout].
  registry     Active Pyramid registry.
  request      Active request object.
  root         Root of the default resource tree.
  root_factory Default root factory used to create `root`.
  transaction  Exported transaction functions.  $Id$

SooooOOO nice!

💇 👯 👍

[digitalresistor]

Please fix coverage in tests, and repush to this PR: https://travis-ci.org/Pylons/pyramid/jobs/84021793

[jessedhillon]

I'm not familiar with this CI system, can you give any suggestions about what's the issue and how to repro?

[stevepiercy]

@jessedhillon It's not really the CI, it's the lack of coverage for this feature. To avoid this, always run tests locally before committing. Instructions are in HACKING.txt.

For this specific error, click Details then click on the failing job. Here's the pertinent line.

Note to self, I think I need to finally get around to creating that contributing guidelines file for Pyramid.

[digitalresistor]

Sure, Travis CI just runs all of our tox environments, you can do so locally too:

1. Download and install tox (easy_install tox or pip install tox)
2. At the top-level directory run (to get the coverage report):
   • tox -e py2-cover,py3-cover,coverage

You'll need to make sure that you have both Python 2 (as a binary named python2.7) and Python 3 (as python3.4) installed.

This will run the tests under python 2, collecting information about what lines are covered/not covered, then as python 3 and find out what lines are covered/not covered by tests, and then the last one will merge the two results so that any python 2 or python 3 only code paths don't count against the total lines covered.

[jessedhillon]

I see what's going on now, thanks. It's complaining that there's no test that touches the case where the object has a falsy docstring. I'll add that test and resubmit. Aside from the test, do you have any comments on the feature itself or the way it's been implemented?

[digitalresistor]

Nope, looks good to me :-).

[digitalresistor]

Sorry, I should have been better about including what was wrong, rather than just giving you a link to Travis. That's my bad!

[jessedhillon]

No worries, I saw the build was failing but didn't really understand the output of the test runner because I was looking for a failing test. I've added a test (did a forced push, didn't want to create a new commit for two lines) that adds None to the env. None.__doc__ is None also, so that hits the falsy case now.

[jessedhillon]

Pyramid is the only sane framework AFAIAC, so I'm happy to be able to make a contribution and look forward to making more.

[digitalresistor]

Last but not least to speed up getting this accepted, add a commit that signs CONTRIBUTORS.txt please.

Looks good 👍

[jessedhillon]

Signed

[digitalresistor]

@mmerickel This is good to go :-)

[mmerickel]

Great, I'll merge this when I have a second to backport it to 1.6-branch and update the changelog.