implement all the math functions for all the floating point types

[andrewrk]

Boxes without check marks do not have test coverage.

	f16	f32	f64	f128	c_longdouble
@sqrt	✓	✓	✓
@sin	✓	✓	✓
@cos	✓	✓	✓
@exp	✓	✓	✓
@exp2	✓	✓	✓
@ln	✓	✓	✓
@log2	✓	✓	✓
@log10	✓	✓	✓
@fabs	✓	✓	✓
@floor	✓	✓	✓
@ceil	✓	✓	✓
@trunc	✓	✓	✓
@nearbyInt
@round

[LewisGaul]

I've just come across a need for exp and ln with f128, so I'm going to take a look at this :)

[LewisGaul]

I haven't got anywhere with implementations yet, but I've learnt a lot about how floats work and implementations of the exp2() function. I'm making notes and will track progress at https://github.com/LewisGaul/zig-f128math (taking inspiration from the initial work by @tiehuis in https://github.com/tiehuis/zig-fmath).

Mainly leaving this comment in case useful to anyone else interested in working on this.

I do also have a question about c_longdouble (since it's included in the table above). Quoting from tiehuis/zig-fmath#7:

The c_longdouble type is not implemented at all. The rationale is if required, the user should use the libc math library, which makes some sense given it is a c abi type anyway. The long double functions as well generally are more complex to implement as well.

Is there a reason this can't just map to f128? The C standard just says that long double must be at least as accurate as double I believe. The Musl implementation modifies its behaviour based on LDBL_MANT_DIG and LDBL_MAX_EXP macros (https://www.gnu.org/software/libc/manual/html_node/Floating-Point-Parameters.html), would we need to do something like respect those parameters when interpreting a c_longdouble?