Fix #580 by using a fixed-point implementation for unit conversions using integer representations #1780

Job	Run time
Generate build matrix for CMake Test Package CI	4s
fmtlib gsl-lite C++20 Apple Clang 15 Debug	1m 8s
fmtlib none C++20 Apple Clang 15 Release	1m 0s
fmtlib gsl-lite C++20 Clang-16 (x86-64) Debug	1m 11s
fmtlib none C++20 Clang-16 (x86-64) Release	52s
fmtlib none C++20 Clang-17 (x86-64) Debug	1m 52s
std::format gsl-lite C++20 Clang-17 (x86-64) Debug	1m 42s
std::format gsl-lite C++20 Clang-18 (arm64) Debug	1m 13s
fmtlib none C++23 Clang-18 (arm64) Release	1m 14s
std::format gsl-lite C++20 Clang-18 (x86-64) Debug	2m 29s
std::format none C++23 Clang-18 (x86-64) Debug	1m 44s
fmtlib gsl-lite C++20 GCC-12 Debug	43s
fmtlib ms-gsl C++20 GCC-12 Release	41s
fmtlib ms-gsl C++23 GCC-12 Debug	41s
std::format gsl-lite C++20 GCC-13 Debug	43s
std::format none C++20 GCC-13 Release	42s
fmtlib none C++20 GCC-14 Release	38s
std::format gsl-lite C++20 GCC-14 Debug	44s
fmtlib gsl-lite C++20 MSVC 14.4 Debug	2m 20s
std::format gsl-lite C++20 MSVC 14.4 Debug	2m 33s
	24m 14s

Provide feedback