IsARealSquareBesideAConstantOffTheRealLine​(AngouriMath.​Entity,​AngouriMath.​Entity[],​AngouriMath.​Entity.​Variable)

There, with s = x + q1/(2 q2) and c = q0/q2 - q1^2/(4 q2^2), which is not
real, Q = q2 (s^2 + c) keeps the sign of its imaginary part on the whole real line,
so its root and ln(2 sqrt(q2) S + 2 q2 x + q1) are continuous there. And beside a real
linear with its root at s0, 1 - w^2 for the artanh's argument w is
c/((s0^2 + c)(s^2 + c)) times a square: for c = r e^(i theta) each factor
below has its argument between 0 and theta, so that of the quotient lies in
[-theta, theta), and w never reaches the artanh's cuts. The logarithm of
(1 + w)/(1 - w) is then continuous on either side of the linear's root, whatever
the phase of H, and its derivative needs only sqrt(z)^2 = z. Welz's
1/((1 + x)^2 sqrt(x^2 + i)) was declined for the sign of H = 1 + i.
https://github.com/asc-community/AngouriMath/issues/233