Make more deterministic test for shooting point selector

[sroet]

In the doc update #963, I saw from codecov that the following line was not reliably hit in shooting.py:

This adds a test to force that code path in the first commit: 1842ed9

The second commit (31e11bf) drops nose in the test_shooting.py and uses pytest instead as an incremental update towards #756

Edit: (dc6e428) removes the duplicate of the test added in the first commit (I had copied it to a different spot around in the first commit, but forgot to remove the original one 🤦 )

[codecov]

Codecov Report

Merging #964 (7495374) into master (5384e3f) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #964   +/-   ##
=======================================
  Coverage   80.25%   80.25%           
=======================================
  Files         136      136           
  Lines       14449    14449           
=======================================
  Hits        11596    11596           
  Misses       2853     2853           

---

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 5384e3f...7495374. Read the comment docs.

[dwhswenson]

LGTM. I think I mentioned recently that there may be a couple tests that are flaky with regards to coverage -- that is, some small fraction of lines are sometimes covered, sometimes not. This is one are where I remember that being the case. On Coveralls we allowed some small drop; we don't (yet) on CodeCov. Thanks for the fix, and the update to pytest-style!