math: avoid runtime conversions of floating-point constants
[musl] / VERSION
1 1.2.2