Test stats when not all samples are included.

[petrelharp]

Over in #1023 I noticed that some tests were passing that shouldn't, because all of our tests in test_tree_stats.py had the property that the union of all sample sets was equal to the set of all samples. This fixes this oversight (and hopefully everything still works, eek!).

I've also set (temporarily!!) it to run all the tests in test_tree_stats.py, but will need to change that back so that testing doesn't take forever.

[petrelharp]

Hm, this failed but for unrelated reasons!?!? Running it on my laptop atm.

[jeromekelleher]

Yeah, some concurrency bug has shown up in macos latest python 3.6 recently. We should fix ASAP I guess, or just disable the test for now. @benjeffery, have you spotted this one?

[benjeffery]

Not seen that one before #1041

[benjeffery]

@Mergifyio rebase

[mergify]

Command rebase: success

Branch has been successfully rebased

[petrelharp]

passed!

[jeromekelleher]

I think the failures here are all due to the concurrency bug we've figured out, so I think we can reset p back to it's old value and merge here @petrelharp?

[petrelharp]

Reset p; I think we can merge this.

[AdminBot-tskit]

📖 Docs for this PR can be previewed here

[codecov]

Codecov Report

Merging #1039 (a7c9e66) into main (accfac6) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main    #1039   +/-   ##
=======================================
  Coverage   93.68%   93.68%           
=======================================
  Files          26       26           
  Lines       20985    20985           
  Branches      875      875           
=======================================
  Hits        19659    19659           
  Misses       1289     1289           
  Partials       37       37           

Flag	Coverage Δ
c-tests	92.49% <ø> (ø)
lwt-tests	92.80% <ø> (ø)
python-c-tests	94.82% <ø> (ø)
python-tests	98.61% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

---

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 accfac6...a7c9e66. Read the comment docs.