bug#61460: 30.0.50; Calendar shows eclipse for quarter moon

[Ulrich Mueller]

>>>>> On Tue, 14 Feb 2023, Michael Heerdegen wrote:
> I did not find good discussions of the eclipse limits in the English
> Wikipedia version; I found good explanations in German however:

>   
> https://de.wikipedia.org/wiki/Finsternis-Limit#Finsternis-Limite_bei_Sonnenfinsternissen

Thanks. I followed the legend of the diagram and was pointed to the
following page, which contains a more comprehensive explanation:
https://www.swetzel.ch/astronomie/finster/finster.html#3

> The limits for an eclipse to happen are not really constant, but taking
> this into account is probably out of scope of the currently available
> code in lunar.el.

I wonder why the angles in above explanation are completely different
from the ones given in eclipse-check:

          ((< moon-lat 2.42600766e-1)
           (concat "** " phase-name " Eclipse **"))
          ((< moon-lat 0.37)
           (concat "** " phase-name " Eclipse possible **"))

These are in radians. The first angle is 13.9° (exactly) and the
second one is 21.2°. They are in argument of latitude, while (IIUC)
the angles in "Finsternis-Limit" are given in ecliptic longitude.
Still, with the moon's orbital inclination being only 5°, one wouldn't
expect much of a difference there. Or am I missing something?

I've also found the book mentioned in lunar.el:
https://books.google.de/books?id=vVBPtkABpUoC&lpg=PA186&vq=eclipse&hl=de&pg=PA177#v=onepage&q&f=false

CCing Nicholas Strauss (who had originally submitted the code in
bug #20414).