`test_perturbations.py` takes a lot of time

While testing using pytest, the test hangs for a sufficiently long time while test_perturbations.py runs.

Running the following shows that it takes ~140s to run test_solar_pressure in test_perturbations.py which when combined with other durations, gives a significant test time as a whole.

pytest --durations=4

Output:

======================================================================== slowest 4 durations =========================================================================
143.81s call     tests/tests_twobody/test_perturbations.py::test_solar_pressure[200-deltas_expected0]
142.41s call     tests/tests_twobody/test_perturbations.py::test_solar_pressure[400-deltas_expected1]
138.72s call     tests/tests_twobody/test_perturbations.py::test_solar_pressure[600-deltas_expected2]
59.31s call     tests/tests_twobody/test_perturbations.py::test_3rd_body_Curtis[test_params2]

Specifically, the problem seems to be in the test_solar_pressure function. This particularly stood out to me since other tests take much less time.

Thanks!

[astrojuanlu]

And yes, unfortunately these tests take a lot of time because we validate against long propagations found in the literature. They are sort of “integration tests” and there’s not much we can do about them, except perhaps give them another pass and see if we can rewrite absolutely everything in numba.

[astrojuanlu]

Partially addressed in #1250 (thanks @jtegedor!), however there is still more work to do. In particular, one of the functions was jitted in object mode because otherwise numba would give a weird error. We need to investigate more.