[Numpy-tickets] [NumPy] #306: Location of arrayobject.h

NumPy numpy-tickets at scipy.net
Thu Jan 4 18:32:12 CST 2007


#306: Location of arrayobject.h
--------------------+-------------------------------------------------------
 Reporter:  gpk     |        Owner:  somebody
     Type:  defect  |       Status:  reopened
 Priority:  normal  |    Milestone:          
Component:  Other   |      Version:          
 Severity:  normal  |   Resolution:          
 Keywords:          |  
--------------------+-------------------------------------------------------
Changes (by goddard):

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

Comment:

 Can this issue be revisited?

 The installation location of the numpy C header files

   PREFIX/lib/python2.x/site-packages/numpy/core

 does not follow the established python convention of putting C headers in

   PREFIX/include/python2.x

 This ticket was closed with the explanation that the user may not have
 write permission on the standard include directory or that directory may
 not exist. It is likely that if the user has write permission on
 PREFIX/lib/python2.x/site-packages then they also have it on
 PREFIX/include/python2.x.  If the latter directory does not exist it would
 be reasonable to try to create it.  If the directory cannot be created or
 cannot be written it would be reasonable to complete the installation
 without the header files.  In most installations numpy header files will
 not be used and a developer will have the experience to remedy the problem
 in the rare circumstance that they find the headers missing.  Finally many
 C extensions that depend on numpy will not use distutils so
 numpy.get_include does not remedy the awkward header file location.

 Thanks,

  Tom Goddard
  Programmer for Chimera molecular visualization system
 (www.cgl.ucsf.edu/chimera)

-- 
Ticket URL: <http://projects.scipy.org/scipy/numpy/ticket/306#comment:2>
NumPy <http://projects.scipy.org/scipy/numpy>
The fundamental package needed for scientific computing with Python.


More information about the Numpy-tickets mailing list