CI/TST provide more informative traceback in case of a deadlock in tests

[fcharras]

A set of changes proposed by @tomMoral in #588

[codecov]

Codecov Report

Base: 94.01% // Head: 93.38% // Decreases project coverage by -0.64% ⚠️

Coverage data is based on head (8f06914) compared to base (6836640).
Patch coverage: 100.00% of modified lines in pull request are covered.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1379      +/-   ##
==========================================
- Coverage   94.01%   93.38%   -0.64%     
==========================================
  Files          52       52              
  Lines        7300     7342      +42     
==========================================
- Hits         6863     6856       -7     
- Misses        437      486      +49     

Impacted Files	Coverage Δ
joblib/test/test_memmapping.py	96.99% <ø> (-2.26%)	⬇️
joblib/test/test_testing.py	100.00% <ø> (ø)
joblib/test/test_numpy_pickle.py	94.24% <100.00%> (+<0.01%)	⬆️
joblib/testing.py	95.23% <100.00%> (ø)
joblib/test/testutils.py	50.00% <0.00%> (-50.00%)	⬇️
joblib/backports.py	54.08% <0.00%> (-15.31%)	⬇️
joblib/_parallel_backends.py	92.27% <0.00%> (-2.21%)	⬇️
joblib/disk.py	90.47% <0.00%> (-1.59%)	⬇️
joblib/_memmapping_reducer.py	94.38% <0.00%> (-1.50%)	⬇️
... and 11 more

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[ogrisel]

There are plenty of unrelated changes in this PR. I will open a new PR with only the required change to fix the flake8 problem.

[ogrisel]

I changed the title to make it easy to understand the purpose of the PR without making any assumption about knowledge of concurrent PRs.

I think the new benchmark and the flake8 fix would have deserved independent PRs as those are unrelated and would have made it easier to review.

[ogrisel]

Actually I haven't reviewed the new benchmark.

Could you please rename
benchmarks/bench_njobs_1.py to benchmarks/bench_sequential_fast_tasks.py?

[ogrisel]

@tomMoral can you please push your updated version of the bench script to this PR to be able to merge it?

[glemaitre]

@ogrisel we will remove the bench from this PR and added in a separate PR.

[ogrisel]

+1

[ogrisel]

The memory.rst doctest failure is unrelated and is being addressed in #1396.

The other failure (non silent resource tracker) has also been observed elsewhere so is probably unrelated as well.