numpy.dot(a, b, a) gives an incorrect result for a (Trac #2154)

[numpy-gitbot]

Original ticket http://projects.scipy.org/numpy/ticket/2154 on 2012-06-08 by trac user ktdrv, assigned to unknown.

Unless this is expected but undocumented behavior, trying to dot product an array in-place, results in a broken result:

>>> a = arange(1, 10, dtype = float64).reshape(3, 3)
>>> a
array([[ 1.,  2.,  3.],
       [ 4.,  5.,  6.],
       [ 7.,  8.,  9.]])
>>> dot(a, identity(3), a)
array([[ 0.,  2.,  3.],
       [ 4.,  5.,  6.],
       [ 7.,  8.,  9.]])