add math.signbit

[ringabout]

C99 function

Ref timotheecour#505

[Araq]

We target C89 though. Also, why do we need it? One can always cast to uint64 and check the sign bit.

[timotheecour]

We target C89 though

plenty of code in stdlib use functions that require >= C99, see https://en.wikibooks.org/wiki/C_Programming/math.h, so there's no reason to single this one out; future work can improve general support for C89 (via when not defined(nimc99) maybe) but this shouldn't be at the expense of performance for the vast majority of use cases that can assume >= C99.

One can always cast to uint64 and check the sign bit.

that wouldn't be portable, it'd depend on the FP type, endian-ness, architecture; note that FP endian-ness is not simple, eg see https://en.wikipedia.org/wiki/Endianness

Although the ubiquitous x86 processors of today use little-endian storage for all types of data (integer, floating point), there are a number of hardware architectures where floating-point numbers are represented in big-endian form while integers are represented in little-endian form
here are ARM processors that have half little-endian, half big-endian floating-point representation for double-precision numbers: both 32-bit words are stored in little-endian like integer registers, but the most significant one first

etc...

[ringabout]

It is useful for making copysign work robust in JS backend.
#16609

It is also useful for making nan consistent in CT and RT

timotheecour#499