improved printing of Cyclotomic

[SoongNoonien]

I tried to address #47. While comparing my results to AbstractAlgebra I found some weird things though. I'm not sure if this is intended but it could to be related to the Expressions of AbstractAlgebra.

julia> 1//(i*j)
1//i*j

julia> 1//i*j
j//i

julia> e2p(1//q*i)//(3*q)
exp(2π𝑖(1//q*i))//3*q

julia> e2p(1//q*i)//3*q
1//3*q*exp(2π𝑖(1//q*i))

Maybe a simple * without spaces is supposed to represent implied multiplication.

[SoongNoonien]

Ok, I forgot to adjust the doctests...