Unit test failure with the COBYLA algorithm for spectral fitting

[AdamGoldstein-USRA]

The COBYLA algorithm in scipy.optimize.minimize is failing its unit test with the following error:

________________________ TestCOBYLA.test_parameters _________________________

self = <tests.core.spectra.test_fitting.TestCOBYLA testMethod=test_parameters>

    def test_parameters(self):
>       self.assertAlmostEqual(self.fitter.parameters[0], 0.05, places=2)
E       AssertionError: 0.0001000001 != 0.05 within 2 places (0.049899999900000006 difference)

tests/core/spectra/test_fitting.py:739: AssertionError

In the unit test, a power law with index = -1.3 and amplitude = 0.05 is simulated and fit. The failure appears to be on fitting the amplitude. Unit tests for other algorithms pass, so maybe this is a problem in scipy?