Implementation of continuous integration workflow. It follows the same steps that we currently request when testing branches. The main differences are:
It uses gfortran, which changes the outputs. The changes are generally smaller than 0.07%. CALIPSO backscatter calculations show bigger relative errors, but for very small absolute differences (<1.0e-37).
I have developed a new python test script that runs in python 3.
The new reference file (a.k.a. KGO: known good output) contains new variables. I believe these variables were introduced in #14, but the reference file was not updated.
Implementation of continuous integration workflow. It follows the same steps that we currently request when testing branches. The main differences are:
It uses gfortran, which changes the outputs. The changes are generally smaller than 0.07%. CALIPSO backscatter calculations show bigger relative errors, but for very small absolute differences (<1.0e-37).
I have developed a new python test script that runs in python 3.
The new reference file (a.k.a. KGO: known good output) contains new variables. I believe these variables were introduced in #14, but the reference file was not updated.