[NumPy-Tickets] [NumPy] #2061: difference between behavior of assignment division and future division not documented

[NumPy Trac]

#2061: difference between behavior of assignment division and future division not
documented
----------------------------+-----------------------------------------------
  Reporter:  AlanFrankel    |       Owner:  pv                           
      Type:  defect         |      Status:  closed                       
  Priority:  normal         |   Milestone:  Unscheduled                  
 Component:  Documentation  |     Version:  1.6.1                        
Resolution:  wontfix        |    Keywords:  assignment division operators
----------------------------+-----------------------------------------------
Changes (by oliphant):

  * status:  reopened => closed
  * resolution:  => wontfix


Comment:

 I can support the "same_kind" casting approach for 2.0 and as a context-
 sensitive parameter in the 1.X series if we choose to spend the time on
 it.     __setitem__ will need the same work.

 The current behavior can be confusing at first, but it is consistent with
 the "truncating" behavior of other things in NumPy (like setitem).   There
 is an important difference between a <op>= value and a = a <op> value
 operation.   The later allows for a complete change in the object that 'a'
 is bound to, while the former does not.    Going to "same kind" can still
 make clear this difference but perhaps remove the most glaring surprises.

-- 
Ticket URL: <http://projects.scipy.org/numpy/ticket/2061#comment:12>
NumPy <http://projects.scipy.org/numpy>
My example project