cot reports wrong value for (0, 0)

[hsmyers]

b = math.complex(0,0) then b.cot() reports (NaN, NaN) as compared to Mathematica's ComplexInfinity. Not that the correct answere is always useful, but it (the NaN answer) did cause a problem in testing extended precision complex functions.

[harrysarson]

I can't remember exact what the status is but mathjs definitely has some issues regarding complex infinite values. You are definitely right that the current behaviour is not ideal.

refs: #804

[josdejong]

Thanks for reporting.

@infusion what do you think about this edge case?