While futzing with the CI in the release branch, I realized we're currently doing things I no longer would recommend.

This moves us to a build-test-deploy workflow. The tests are broken into full tests across the supported Python range from the repository and an install-package-and-pytest test that will be more typical of downstream users or OS packagers. Finally, we upload if all tests pass.

I created an environment that allows us to manually review releases before pushing the package to PyPI, as this is the nibabel workflow, but given that we haven't been doing that so far I'm happy to revert to the "go on green" approach.

Codecov Report

Base: 65.25% // Head: 65.29% // Increases project coverage by +0.03% 🎉

Coverage data is based on head (c005bdb) compared to base (e5be008).
has no changes to coverable lines.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3513      +/-   ##
==========================================
+ Coverage   65.25%   65.29%   +0.03%     
==========================================
  Files         309      309              
  Lines       40873    40873              
  Branches     5380     5380              
==========================================
+ Hits        26673    26687      +14     
+ Misses      13126    13108      -18     
- Partials     1074     1078       +4     

Flag	Coverage Δ
unittests	65.03% <ø> (ø)

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

Impacted Files	Coverage Δ
nipype/info.py	87.50% <ø> (ø)
nipype/utils/filemanip.py	72.45% <0.00%> (+0.92%)	⬆️
nipype/testing/utils.py	89.09% <0.00%> (+18.18%)	⬆️

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.