The commits you reference are the original commits, and your PR was rebase-merged into master, so those commits didn't make it in. You'll need to look up the commit ids that are actually on master.

Hmm I get numerous errors on macos 10.14 in graspy's tests:

______________________ ERROR collecting tests/test_io.py _______________________
ImportError while importing test module '/private/tmp/nix-build-python3.7-graspy-0.3.drv-0/source/tests/test_io.py'.
Hint: make sure your test modules/packages have valid Python names.
Traceback:
tests/test_io.py:8: in <module>
    import graspy as gs
graspy/__init__.py:9: in <module>
    import graspy.plot
graspy/plot/__init__.py:2: in <module>
    import matplotlib as mpl
/nix/store/xcryb61l6wg6pnfhdh0ywd0cynjlbmhd-python3.7-matplotlib-3.3.1/lib/python3.7/site-packages/matplotlib/__init__.py:174: in <module>
    _check_versions()
/nix/store/xcryb61l6wg6pnfhdh0ywd0cynjlbmhd-python3.7-matplotlib-3.3.1/lib/python3.7/site-packages/matplotlib/__init__.py:159: in _check_versions
    from . import ft2font
E   ImportError: dynamic module does not define module export function (PyInit_ft2font)

This is strange because the master equivalent builds successfully.

Yea, sorry, I did a rebase and merge. so the commits get new hashes as part of the rebase:

1179840
5315ce0

I've cherry-picked the new commits but not yet investigated OS X breakage.

I didn't think of the possibility of this happening, so maybe a caveat should be added to the backporting checklist in CONTRIBUTING.md?

I've cherry-picked the new commits but not yet investigated OS X breakage.

I think this should be done as a separate PR, as this PR is still an improvement

I didn't think of the possibility of this happening, so maybe a caveat should be added to the backporting checklist in CONTRIBUTING.md?

sure