The inverse method of the Ellipsoidal Azimuthal Equidistant projection returns incorrect results about proj4js

[ahocevar]

@EkkoScholtens Thanks for the bug report. Your suggested code change makes sense. Would you be able to create a pull request?

[EkkoScholtens]

I can make the pull request for that issue. But i don't know what the issue is with the longitude. I believe this code is implemented from the book "Map Projections - A Working Manual" By JohnP. Snyder right?

[ahocevar]

I don't know, the code was initially ported from C proj4. But looking again, your suggested code change should not produce a different result, as Math.atan(y / x) does the same as Math.atan(y, x). If you want to dig deeper, you can compare the implementation to the one in mproj's aeqd.js.

[EkkoScholtens]

You would expect that, but they do give different results in this use case. I've checked the PROJ source they do use atan which gives the correct result. However in new versions they made the switch to geographiclib which uses a different way to calculate the result.