Split encoding into floating and integer functions

[drinckes]

Issue #672

This splits the C encoding function into a public function that takes doubles, and one that takes integers.

It explicitly rounds multiplying the degrees by the precision. This will cause a small number of encoding tests to require correcting the codes.

The issue was that the C code:

long long int lat = degrees * precision;
lat += 90 * precision;

and

long long int lat = 90 * precision;
lat += degrees * precision

didn't always produce the same result. In the first, it truncates the value of degrees * precision, in the second it appears to (mostly) round it before adding it.

[drinckes]

I've verified the conversions of the coordinates in the CSV file to the changed generated codes, and the new codes look correct. I'm going to update them - this will break all other implementations until I get around to them one by one.

[drinckes]

@bocops @fulldecent @rsramek would appreciate if you could take a look at the change please, my C isn't that hot.