Cleanup PR

[bsipocz]

This PR does a few low prio cleanup tasks:

• Opts in the travis container for the linux builds.
• Removes the locally patched version of the sphinx extension viewcode
• Fixes PEP8, except long lines and except the few third party files hosted in sphinx/ext

I think the only controversial point is the viewcode removal. The patch got included 1.5 years ago into 1.3b1. I think it's a long enough time, and we can remove our local version.

Note that this fails, but when built on top of #225 the tests pass (https://travis-ci.org/bsipocz/astropy-helpers/builds/123386933).

[embray]

This looks like a bug. My guess is pylint or whatever warned that exc is not used, but actually it is on this line but the variable was accidentally renamed err here. Should change back to as exc and change err to exc here.

[bsipocz]

Thanks, indeed.

[eteq]

Hmm, it's a fair question what the oldest version of sphinx we want to support is. I think older sphinxes are less of a priority because it's sort of a "dev-oriented" thing to need to build the sphinx docs. @astrofrog or anyone else have any opinions?

Note, though, @bsipocz, that I think it makes sense to update the need_sphinx in conf.py to 1.3 if you're removing the viewcode.

[bsipocz]

@eteq - changed need_sphinx to 1.3, and rebased. I guess this will need a changelog entry, but not sure which one; 1.2?

[eteq]

Hmm, yeah, now that you ask, I suppose upgrading the required sphinx is something we wouldn't want in a bugfix. So I'd say 1.2

[eteq]

@bsipocz - actually, @astrofrog raised a good point that this affects the affiliated packages, so you should ask on astropy-affiliated-maintainers if anyone minds having to upgrade to 1.3 . If everyone there is OK with it I'd say it's ok to include this for helpers v1.2

[bsipocz]

That's a good point, I'll send out an e-mail.

[bsipocz]

@eteq - There was no reaction to the e-mail about this, so I assume everyone is OK with upgrading to sphinx >=1.3.
There was quite many changes in the helpers recently, so I rebased the PR.

[astrofrog]

@bsipocz - can you rebase?

[bsipocz]

@astrofrog - Done

[astrofrog]

I think AppVeyor is broken for unrelated reasons, so I'll go ahead and merge and debug that in another issue.

[bsipocz]

I was just about to open an issue about the appveyor failures, as said on one of the other PR it's around for a long time.