powered by
Taylor expansion of digamma(a+b) for a>>b
dfun(x, y, mag = 1e+08)
numeric
first argument
second argument
cutoff magnitude for switching approximations