[PATCH v4 16/21] math: Remove the SVID error handling from sinhf

[Wilco Dijkstra]

Hi Adhemerval,

> We can simplify with:
>
>  float sgn = copysignf (2.0f, x);
>   return __math_oflowf_value (sgn * 0x1.fffffep127f);

That's still literally equivalent to __math_oflowf (x < 0)...

Cheers,
Wilco