[test] Calculate geodesic scale (M12) using high precision arithmetic

Instructions for building GeographicLib are given here:
https://geographiclib.sourceforge.io/html/highprec.html

It requires the use of libquadmath package, which is present
in g++-7 by default. The cmake command used is:
cmake -DCMAKE_CXX_COMPILER=g++-7 -DGEOGRAPHICLIB_PRECISION=4 ..

Internally, GeographicLib makes use of float128 as part of Boost
Multiprecision library. To output full precision values, use:
std::setprecision(std::numeric_limits<float128>::max_digits10)