Re: [Numpy-discussion] crash in np.poly

On Thu, May 27, 2010 at 10:00 PM, Charles R Harris wrote:

On Thu, May 27, 2010 at 7:57 PM, Charles R Harris wrote:

...

On Thu, May 27, 2010 at 5:19 PM, wrote:

...

I was tracking a test failure/crash in scipy.signal.ltisys

numpy 1.4.0:

...

...

...

np.poly(np.zeros((0,0)))  ** On entry to DGEEV  parameter number  5 had an illegal value

In current:

In [1]: np.poly(np.zeros((0,0)))

--------------------------------------------------------------------------- ValueError                                Traceback (most recent call last)

/home/charris/<ipython console> in <module>()

/usr/local/lib/python2.6/dist-packages/numpy/lib/polynomial.pyc in poly(seq_of_zeros)     126         pass     127     else: --> 128         raise ValueError, "input must be 1d or square 2d array."     129     130     if len(seq_of_zeros) == 0:

ValueError: input must be 1d or square 2d array.

Looks like it could use a better error message in this case, though. Could you open a low priority ticket for this?

The new Polynomial class does somewhat better:

In [6]: Poly(zeros((0,0))) --------------------------------------------------------------------------- ValueError                                Traceback (most recent call last)

/home/charris/<ipython console> in <module>()

/usr/local/lib/python2.6/dist-packages/numpy/polynomial/polynomial.pyc in __init__(self, coef, domain)

/usr/local/lib/python2.6/dist-packages/numpy/polynomial/polyutils.pyc in as_series(alist, trim)     156     arrays = [np.array(a, ndmin=1, copy=0) for a in alist]     157     if min([a.size for a in arrays]) == 0 : --> 158         raise ValueError("Coefficient array is empty")     159     if any([a.ndim != 1 for a in arrays]) :     160         raise ValueError("Coefficient array is not 1-d")

ValueError: Coefficient array is empty

Chuck

http://projects.scipy.org/numpy/ticket/1495 Thanks, Josef

_______________________________________________ NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion