coronography unit test failures on Mac only

[stscijgbot-jp]

Issue JP-3614 was created on JIRA by Ned Molter:

The test jwst/coron/tests/test_coron.py::test_align_array started showing 
AssertionError: Not equal to tolerance rtol=1e-07, atol=1e-06 a few days ago when GitHub Actions runs the unit test suite for MacOS py311.
 
The differences are pretty small: Max absolute difference: 9.8773848e-06, Max relative difference: 4.67009943e-05.  However, it would be good to figure out why this happened (e.g., a numpy/scipy release) and ensure that there are no qualitative differences to worry about here before simply increasing the tolerances.

[braingram]

See: #8440

The differences from the leastsq fit are relatively small. However the tolerances for the test had to be raised considerably (see #8439) due to how the test is calculated.

I had some hope that scipy dev fixed this (as there were other leastsq issues). However the last time I checked (sometime last week) the nightly version still caused this test to fail.

[stscijgbot-jp]

Comment by Howard Bushouse on JIRA:

Fixed by #8490