Bug in F77 : exp() of complex arguments

[axt654 at csc.anu.oz]

This is just a note in case anyone else is haveing trouble with exp() on 
a 3100. Basically the algorithm for exponentiation of a complex
argument is faulty. Taking exp of anything with a large imaginary part
gives a result of zero. It is easily fixed by taking mod 2*pi of the imaginary
part before calling exp. (or rewriting zexp to do this).

If you're getting funny results when using exp(complex) then check on this.

Andrew Tridgell
tridge at aerodec.anu.oz.au