Frac(inf) results in invalid floating point operation

When using frac on an infinity value (and using softfloat routines), programs receive an invalid floating point operation. This seems to be caused by https://gitlab.com/freepascal.org/fpc/source/-/blob/main/rtl/inc/genmath.inc#L1781 where subtracting a value from inf causes that exception to be raised in float64_sub