lambertw, all branches

[Gil]

Always regarding the Albert Chan's
result = "tiny real part" ± i × n(PI/2), n≠0

here is a possible "patch" for HP50G to fix up such an issue,
where next, final result will be then: 0 ± i × n(PI/2).

Code:

\<< DUP \-> r
  \<< TYPE 1 ==
    IF
    THEN r C\->R DUP ABS DUP .0001 + \pi 2 / \>= SWAP \pi 2 / / \->NUM DUP 0 RND - ABS .0001 \<= AND ROT ABS .00000000001 < AND
      IF
      THEN 0 SWAP R\->C
      ELSE DROP r
      END
    ELSE r
    END
  \>>
\>>