Some lingering issues with dev versions of numpy, which looks a bit like something on the conda side...

Codecov Report

❗ No coverage uploaded for pull request base (master@f8fac84). Click here to learn what that means.
The diff coverage is 100%.

@@            Coverage Diff            @@
##             master     #596   +/-   ##
=========================================
  Coverage          ?   96.58%           
=========================================
  Files             ?       49           
  Lines             ?     4506           
  Branches          ?        0           
=========================================
  Hits              ?     4352           
  Misses            ?      154           
  Partials          ?        0

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f8fac84...7d61674. Read the comment docs.

Hello @StanczakDominik! Thanks for updating your pull request.

Congratulations! There are no PEP8 issues in this pull request. 😸

Comment last updated on January 04, 2019 at 17:21 Hours UTC

EDIT: this is a comment to my previous commit, "Revert "Update Type Hint Annotations for Consistency (#586)". It got push -f'd out of this branch as I found another way to solve the issue without reverting the commit.

Bloody hell. This is exactly what we get for not dealing with the astropy unit bug earlier and pushing to master on broken tests. My bad, apologies >_<

What happened here is that I was so sure #586 was fine since tests were failing but that PR did nothing related to the astropy 3.1.0 unit optimization bug, so it was fine to merge it. It wasn't! That pull request triggered the __all__ glitch. Basically, importing stuff from typing - Callable and the like in file example.py puts them into example's, namespace, right? Well, that's bad news because then sphinx with automodapi, when building these, inspects the __all__ parameter which now contains Callable and tries to print out their definitions. But those can't be found in the file, so it returns warnings such as:

b'<partial node>':: WARNING: toctree contains reference to nonexisting document 'api/plasmapy.utils.pytest_helpers.Dict'
b'<partial node>':: WARNING: toctree contains reference to nonexisting document 'api/plasmapy.utils.pytest_helpers.Callable'
b'<partial node>':: WARNING: toctree contains reference to nonexisting document 'api/plasmapy.utils.pytest_helpers.Dict'
b'<partial node>':: WARNING: toctree contains reference to nonexisting document 'api/plasmapy.utils.pytest_helpers.Callable'
b'<partial node>':: WARNING: toctree contains reference to nonexisting document 'api/plasmapy.utils.pytest_helpers.Dict'

To sum up, when I was in high school I used to play Command and Conquer 3 online a lot, in team games. In fact, that's where I picked up most of my english. There was this guy nicknamed Ulkrond whom me and my group met up with often, whose most memorable quote that we had many a laugh about was "Your hubris shall be your downfall".

And, well, turns out that indeed it was.

The good thing about version control and open development is that it makes it more straightforward to fix things like this when issues arise. Thank you for making these fixes!

And as I like to say, getting things wrong is the first step towards getting things right!