Tests for 5-th order polynomials failed, after some investigation it seems like the implementation of ::Complex ^ ::Float has changed since 0.6 and now has a bit less precision. Fixed by replacing the a ^ b operator with exp(log(a) * b) (this is basically how 0.6 implements exponentiation). I'm going to file a bug report against Julia when I get the time to do so.
I also fixed a deprecation warning regarding loop scope changes by introducing another auxiliary variable.
Tests for 5-th order polynomials failed, after some investigation it seems like the implementation of
::Complex ^ ::Float
has changed since 0.6 and now has a bit less precision. Fixed by replacing thea ^ b
operator withexp(log(a) * b)
(this is basically how 0.6 implements exponentiation). I'm going to file a bug report against Julia when I get the time to do so.I also fixed a deprecation warning regarding loop scope changes by introducing another auxiliary variable.