Python Wrapper on Catalina: Segmentation Fault

[ShazAlvi]

After I had issue #462 , I was able import classy in Python27 in a conda environment with gcc-11 installed with home-brew At the moment. I am having issue getting additional quantities like (tCl. mPk etc.) defined in output, however I can get the background quantities. So the following code works perfectly,

import matplotlib.pyplot as plt
from classy import Class

MyClass = Class()
params = {'output': '','omega_b': 0.022383, 'omega_cdm': 0.1201, 'ln10^{10}A_s': 3.044, 'n_s': 0.96605, '100*theta_s': 1.04090, 'tau_reio': 0.0543}
MyClass.set(params)
MyClass.compute()
Bck_Quantity = MyClass.get_background()
plt.loglog(Bck_Quantity['z'], Bck_Quantity['(.)rho_cdm'])
plt.show()

But, adding 'tCl' to the output list returns a segmentation fault, so the following does not work.

import matplotlib.pyplot as plt
from classy import Class

MyClass = Class()
params = {'output': 'tCl','omega_b': 0.022383, 'omega_cdm': 0.1201, 'ln10^{10}A_s': 3.044, 'n_s': 0.96605, '100*theta_s': 1.04090, 'tau_reio': 0.0543}
MyClass.set(params)
MyClass.compute()
Bck_Quantity = MyClass.get_background()
plt.loglog(Bck_Quantity['z'], Bck_Quantity['(.)rho_cdm'])
plt.show()

But if I only ask for background module in MyClass.compute (MyClass.compute(['background'])), everything works again as when tCl is not asked.

I am unable to triangulate why this is happening. I have only recently started working on Mac. On Ubuntu this code works like a charm.

Edit: Just want to add to the previous message that in Ipython the same code works perfectly and I am able to get the tt spectra as expect.

[ThomasTram]

Hi Try to disable --fast-math in the Makefile.

Cheers, Thomas

[ShazAlvi]

Hi Thomas,

Thanks for your suggestion. The option did not help in solving the problem. I still get the same error.

[ShazAlvi]

Hi Thomas,

Just want to add that the same code works fine in IPython. However when the python file is run on the editor or on the terminal it gives a segmentation error.