[NumPy-Tickets] [NumPy] #1950: Redundant numeric type classes lead to unreliable behavior of isinstance()

NumPy Trac numpy-tickets@scipy....
Fri Sep 9 11:19:59 CDT 2011


#1950: Redundant numeric type classes lead to unreliable behavior of isinstance()
---------------------------+------------------------------------------------
 Reporter:  stereotype441  |       Owner:  somebody
     Type:  enhancement    |      Status:  new     
 Priority:  normal         |   Milestone:  1.6.2   
Component:  numpy.core     |     Version:  1.6.0   
 Keywords:                 |  
---------------------------+------------------------------------------------
Changes (by charris):

  * type:  defect => enhancement


Comment:

 For historical reasons, the numpy classes represent C types and hence
 different dtypes can have the same precision depending on the platform.
 The upshot is that isinstance is not going to work for determining the
 precision since there is no guarantee that the same dtype will always be
 used to represent a given precision.This needs to be fixed up at some
 point and some progress has been made in numpy casting so as to provide
 more consistent results, but fixing this problem in general won't happen
 soon.

-- 
Ticket URL: <http://projects.scipy.org/numpy/ticket/1950#comment:1>
NumPy <http://projects.scipy.org/numpy>
My example project


More information about the NumPy-Tickets mailing list