[NumPy-Tickets] [NumPy] #1671: scalars don't raise overflow/underflow/divbyzero/invalid

NumPy Trac numpy-tickets@scipy....
Thu Nov 11 13:17:10 CST 2010


#1671: scalars don't raise overflow/underflow/divbyzero/invalid
------------------------+---------------------------------------------------
 Reporter:  m-paradox   |       Owner:  somebody
     Type:  defect      |      Status:  new     
 Priority:  normal      |   Milestone:  2.0.0   
Component:  numpy.core  |     Version:  devel   
 Keywords:              |  
------------------------+---------------------------------------------------
 All these cases should be raising, I believe:

 {{{
 >>> import numpy as np
 >>> err = np.seterr(all='raise')

 >>> np.float32(1e-38) / np.float32(1e30)
 0.0
 >>> np.float32(1e-38) / np.array(1e38,np.float32)
 0.0
 >>> np.array(1e-38,np.float32) / np.float32(1e30)
 Traceback (most recent call last):
   File "<stdin>", line 1, in <module>
 FloatingPointError: underflow encountered in divide

 >>> np.float64(1e-300) / np.float64(1e300)
 0.0
 >>> np.float64(1e-300) / np.array(1e300, dtype=np.float64)
 0.0
 >>> np.array(1e-300, dtype=np.float64) / np.float64(1e300)
 Traceback (most recent call last):
   File "<stdin>", line 1, in <module>
 FloatingPointError: underflow encountered in divide

 >>> np.float32(1) / np.float32(0)
 inf
 >>> np.array(1,dtype=np.float32) / np.float32(0)
 Traceback (most recent call last):
   File "<stdin>", line 1, in <module>
 FloatingPointError: divide by zero encountered in divide

 >>> np.inf-np.inf
 nan
 >>> np.array(np.inf)-np.inf
 Traceback (most recent call last):
   File "<stdin>", line 1, in <module>
 FloatingPointError: invalid value encountered in subtract

 }}}

-- 
Ticket URL: <http://projects.scipy.org/numpy/ticket/1671>
NumPy <http://projects.scipy.org/numpy>
My example project


More information about the NumPy-Tickets mailing list