[PATCH 13/23] math: Fix the expected atan2f (inf) results

[Paul Zimmermann]

       Hi Adhemerval,

> It turns out that although the new atan2f does improve clog10 imaginary
> precision on some archs (on x86_64 max ulps now show 2 instead of 4 for
> the tested inputs); it is still not correctly rounded (I guess because it
> does not calculate in the expected range, since it requires a 
> M_LOG10E multiplication after the atan2 call).

yes, if x is a random floating-point number, then RN(x*M_LOG10E) is not the
correct rounding of x/log(10) with probability about 16%.

Paul