Bug in _GG_LOS_calc_signal()

[Didou09]

Trying to debug PR #207, I realized the output of the script:
examples/tutorials/plot_custom_emissivity.py

Depends a lot on whether newcalc=True or False

True: (new optimized routines) - wrong result (for example: no negative values, sharp edges...)

False: (old reference routine) - good result

[Didou09]

High priority, we shoul probably issue a new released when fixed, as current users may run into false results

[Didou09]

Once the bug is identified, solved and similar routines are double-checked, I suggested adding a new unit test based on this particular tutorial to check that the results of newcalc=True and newcalc=False are very close / identical (use np.allclose()), what do you think ?

[Didou09]

Fixed in devel by PR #249

[jdhare]

Resolved in 1.4.2, many thanks!