Wrong azimuth angle

[dinhnhat0401]

Sorry but i find that the sun azimuth must plus 180 degree so:
function getAzimuth(H, phi, dec) {
return atan(sin(H), cos(H) * sin(phi) - tan(dec) * cos(phi));
}

must be:
function getAzimuth(H, phi, dec) {
return PI + atan(sin(H), cos(H) * sin(phi) - tan(dec) * cos(phi));
}

[mourner]

You're right, I used south-based azimuth whereas north-based azimuth is usually used for sun position...

[florianmski]

Thank you ! I'm currently writing an Android app and I've ported this awesome tool to java and I didn't understand where was my mistake concerning the sun azimuth :)

[xqjibz]

I created PR #35 to address this with tests.

[ghost]

just noticed this bug myself, thanks for the great work.

The pull request seems to fix it for me, will be double checking in the next days.

[spencerthayer]

Will fix this be merged into the main branch soon?

[mourner]

Probably, but I'd suggest just adding a PI instead of waiting for it to be merged.

[warpling]

Is there a reason #35 hasn't been merged?

[mourner]

Yeah, hesitating to break backwards compatibility, leading to even more confusion :) I'll deal with this eventually, just didn't get my hands on it yet.

[warpling]

Makes sense! Crazy how many forks are downstream also…

[ghost]

If breaking backwards compatibility is a concern than it might be good to add a new interface method which would give the correct result and leave the old one for those who were doing their own workarounds.

[xqjibz]

I did not consider backwards functionality when I did the PR, that's my bad, I simply updated my library here and ran with it. @mourner if you have another solution, I'd be happy to work something up, otherwise no harm no foul on the PR.

[ultrafez]

According to semver, the way to avoid breaking users existing apps is to release a new version with a major version number change (i.e. release as 2.0.0). This way, the vast majority of users won't pull it into their project automatically.