Multiplying Python float to numpy.array of adouble works but fails with a numpy.float64

[b45ch1]

Problem description:
x = numpy.array([1.,2.])
ay = numpy.array([adouble(3.), adouble(4.)])

then:
ay * x[0] # NOT OK!
This is not a bug of pyadolc but a bug of numpy (http://projects.scipy.org/numpy/ticket/1136).

WORKAROUND:
x[0] * ay # OK!