./bootstrap.sh --prefix=<wherever you want to headers> --libdir<where you want the libs> --with-libraries=python

./bjam address-model=32_64 architecture=x86 variant=release link=shared install

I have been trying to get PyCUDA or PyOpenCL working for the longest time but....they require boost to be working which has been a problem.

Once I did get boost installed, I found out that no matter what I try and specify it links against the system Python libraries. PyOpenCL worked fine

at that point...with the system Python, unfortunately i REALLY want to use the Enthought python distribution and I cant seem to convince the boost

libraries to link those.

this is a good mac os x boost install page

http://wiki.tiker.net/BoostInstallationHowto

On Mar 5, 2010, at 1:10 PM, Ian Wojtowicz wrote:

I’m having a hellish time trying to build the python bindings for boost in Mac
OS X 10.6 (Snow Leopard). I’m using boost_1_42_0, which seems to build okay
(except for hanging on test), and I’ve downloaded bgl-python-0.9 (the boost
python bindings), but the build instructions seem to be old and don’t reflect
the contents of bgl-python-0.9.

I’ve been banging my head against the wall trying to get this library up and
running. I really need a system that can help me calculate betweenness
centrality on a dataset I have.

Has anyone successfully installed boost with its python bindings on Snow Leopard?

Are there any other network graph libraries out there for python (or Java) that
can calculate betweenness well?

Thanks.

Ian

_______________________________________________
Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost-build