Open avalentino opened 1 year ago
Looking at the code there is something that I cannot understand:
If Beta
is set to +/- pi/2
then the computation of 1 / cos(Beta)
will result in a nan
.
Probably the expression can be re-written as follows to avoid this kind of issue:
# eqn. 13
- dBeta = ((ell.semiminor_axis * u - ell.semimajor_axis * huE + E**2) * sin(Beta)) / (
- ell.semimajor_axis * huE * 1 / cos(Beta) - E**2 * cos(Beta)
+ dBeta = ((ell.semiminor_axis * u - ell.semimajor_axis * huE + E**2) * sin(Beta)) * cos(Beta) / (
+ ell.semimajor_axis * huE - E**2 * cos(Beta)**2
)
By the way I think that the root cause of this issue is different
@avalentino I attempted a fix in f3a3a53.
Dear @scivision sorry I'm no longer in the condition to be able to quickly check your changes on an ARM platform. I will be able to provide a feedback the next time I upload a new version of pymap3d in debian. Do you already have an idea about when we will have a new release?
What happened?
Recently e have updated the Debian package for pymap3d for v2.9.1 to v3.0 and we have started having some test failues on one of our ARM architectures
See the detailed log at: https://ci.debian.net/data/autopkgtest/testing/armel/p/pymap3d/34582756/log.gz
Relevant log output