A strange bug that seems compiler/platform specific. I've removed the first data point in GC transform test because of this.
This test simulates GOs but collects GC data (using the option parameter). When I run this on my desktop the test runs properly and all collected orbit points are valid. When it is run on Github action the test fails because the first collected orbit point contains dummy data (e.g. R=0 which leads to NaNs in postprocessing).
This is a minor bug (no hotfix needed) but I just want to record it here.
A strange bug that seems compiler/platform specific. I've removed the first data point in GC transform test because of this.
This test simulates GOs but collects GC data (using the option parameter). When I run this on my desktop the test runs properly and all collected orbit points are valid. When it is run on Github action the test fails because the first collected orbit point contains dummy data (e.g. R=0 which leads to NaNs in postprocessing).
This is a minor bug (no hotfix needed) but I just want to record it here.