On 1 March 2010 02:46, Massimo Di Stefano <massimodisa...@gmail.com> wrote: > Hi Tont, > > Yes sorry, > > my typo-error writing the mail, > > in my bash profile i have : > > export DYLD_LIBRARY_PATH=/Library/Python/2.6/site-packages/tables/ > >> libhdf5.dylib > > is in : /usr/local/lib > > i guess this libs are already in my path. > > tring to export it : > > export DYLD_LIBRARY_PATH=/usr/local/lib > > do not produce any changes to the log. > > is there anyone using pytables on mac osx, > what's the right provedure to build and runs it ? > > maybe i had to prefer a binary version for hdf ? :-/ > can ( a binary hdf version ) be used by pytaables to build it self ?
So I just installed szip and zlib with a plain ./configure, then hdf5-1.8.4-patch1 with either ./configure --prefix=/Library/Frameworks/Python.framework/Versions/Current ./configure --prefix=/Library/Frameworks/Python.framework/Versions/Current --enable-cxx and pytables-2.2b3 python setup.py build_ext --inplace --hdf5=/Library/Frameworks/Python.framework/Versions/Current and the tests PYTHONPATH=.:$PYTHONPATH python tables/tests/test_all.py all pass, without any *LIBRARY_PATH variables set. The installation docs [1] do suggest using the HDF5 C flavour, so I'd try without the --enable-cxx first. The reason I'm using the */Frameworks/* prefix is that there was a previous version of hdf installed there (1.8.3), and it seemed to conflict with a /usr/local install. Note that my setup will be different to yours; I'm using gcc-4.0 (which defaults to 32-bit mode) and EPD python. You seem to be using Apple's python (in /System/Library/* and /Library/Python/*) and should probably use gcc-4.2. This is the default on OSX 10.6 (both being 64-bit). There's no reason they shouldn't work, maybe try with the beta or a release before the svn. Tony [1] http://www.pytables.org/docs/manual/ch02.html#sourceInstallationDescr ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ Pytables-users mailing list Pytables-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/pytables-users
