[Issue 22740] const float values should be rounded to float

[d-bugmail at puremagic.com]

https://issues.dlang.org/show_bug.cgi?id=22740

--- Comment #1 from Iain Buclaw <ibuclaw at gdcproject.org> ---
The introduction of `core.math.toPrec' was made to specifically handle this
case, as compile-time floats are stored at full precision.

--