That is good to know. The boost documentation makes it appear as if bjam is the only way to build boost.python, but good to see examples to the contrary!
On Thu, Feb 27, 2014 at 2:19 PM, Toby St Clere Smithe < pyviennacl@tsmithe.net> wrote:
Eelco Hoogendoorn hoogendoorn.eelco@gmail.com writes:
I have a file numpy_boost_python.hpp in one of my projects by Michael Droettboom (can seem to find an online source anymore!), which adds mappings between numpy.ndarray and boost.ndarray, which is very neat and seemless. But like boost.python, it tightly couples with the clusterfuck that is bjam. However, something conceptually like that but integrated with XDress would be great. Indeed it does not sound too complicated; though I don't think I will get around to it anytime soon, unfortunately...
You don't have to use bjam! I have built my projects with distutils and CMake, and never once touched bjam; CMake provides find_package scripts for Boost, Python and NumPy, and for distutils, I just include the relevant files and flags in my project.
See [1] for a CMake example, and [2] for a distutils example.
[1] https://github.com/tsmithe/viennacl-dev/blob/pyviennacl/CMakeLists.txt [2] https://github.com/viennacl/pyviennacl-dev/blob/master/setup.py
Cheers,
Toby
NumPy-Discussion mailing list NumPy-Discussion@scipy.org http://mail.scipy.org/mailman/listinfo/numpy-discussion