[NumPy-Tickets] [NumPy] #1481: Move non-function API from ndarrayobject.h into arraytypes.h

NumPy Trac numpy-tickets@scipy....
Sat May 8 19:19:30 CDT 2010


#1481: Move non-function API from ndarrayobject.h into arraytypes.h
-----------------------------+----------------------------------------------
  Reporter:  abingham        |       Owner:  somebody   
      Type:  enhancement     |      Status:  closed     
  Priority:  normal          |   Milestone:  Unscheduled
 Component:  numpy.numarray  |     Version:  devel      
Resolution:  fixed           |    Keywords:             
-----------------------------+----------------------------------------------
Changes (by charris):

  * status:  needs_review => closed
  * resolution:  => fixed


Comment:

 OK, I made this change in r8397-r8398 because I couldn't think of a good
 reason not to ;) However, I changed the header name to ndarraytypes.h
 because there is already an arraytypes.h buried down in
 numpy/core/src/multiarray and I thought best to avoid possible confusion.
 This change does separate out two different sorts of things, although I
 think the real culprit here is __multiarray_api.h. I'm not up to dealing
 with that, however, nor it's almost equally ugly cousin __ufunc_api.h.
 There should probably be completely different include files for the
 various varieties of include so that the choice of header corresponds to
 what is currently chosen with the various defines.

 Can you write up a short explanation of the use/need for this change so I
 can include it in the release notes? I could probably put that together
 from the mailing list thread but you can probably do a better job of it.

 Also feel free to improve the file if you have the time. The multiline
 comments need to be put in standard form, i.e.,

 /*
  * blah, blah
  */

 and it would be nice to collect all the includes up at the top so one
 could see the dependencies at a glance.

-- 
Ticket URL: <http://projects.scipy.org/numpy/ticket/1481#comment:2>
NumPy <http://projects.scipy.org/numpy>
My example project


More information about the NumPy-Tickets mailing list