math: avoid runtime conversions of floating-point constants
[musl] / src / math / __math_invalidf.c
2019-04-17 Szabolcs Nagymath: add single precision error handling functions