Handle 32bit and 128bit floats, not just 64bit floats

[pryrt]

as of v0.014, still only handles 64bit double-floats. It would be nice to also handle 32bit single-floats and 128bit floats ("quad"?).

[pryrt]

If I did this, I'd probably then introduce a naming convention of some sort.

IEEE Std 754™-2008 uses <format>-camelCase, which of course wouldn't work in Perl.

My thought: use <format>_camelCase instead: name the per-size functions single_camelCase, double_camelCase, quad_camelCase, and use the %Config{nvsize} to pick the correct alias for the raw camelCase.

Though, perhaps I should instead use the actual 754™ format names: binary32, binary64, binary128.