Implement major math functions

[tompng]

Implement these methods

BigMath.cbrt(x, prec)
BigMath.hypot(x, y, prec)
BigMath.asin(x, prec)
BigMath.acos(x, prec)
BigMath.atan2(y, x, prec)
BigMath.sinh(x, prec)
BigMath.cosh(x, prec)
BigMath.tanh(x, prec)
BigMath.asinh(x, prec)
BigMath.acosh(x, prec)
BigMath.atanh(x, prec)
BigMath.log2(x, prec)
BigMath.log10(x, prec)
BigMath.log1p(x, prec)
BigMath.expm1(x, prec)

Remaining methods

Math.singleton_methods - BigMath.singleton_methods
# => [:erf, :erfc, :gamma, :lgamma, :frexp, :ldexp]

[mrzasa]

Just curious: why do we decide what prec2 is needed? It differs between the functions and I'm wondering how it's calculated.

[tompng]

Normally, BigMath functions calculate with prec+BigDecimal.double_fig precision.

I think BigMath.log2(1024, prec) is expected to return exactly 10. To calculate exact value for 2**n numbers, we need to round the result to precision = prec+BigDecimal.double_fig.
log(x)/log(2) should be calculated with prec + BigDecimal.double_fig + extra precision.
I decided to use extra=BigDeciaml.double_fig/2 here.

[mrzasa]

Thanks!