Hi Francesc, _POSIX_BARRIERS is 200112L in that machine's standard headers. I put in the hack below in blosc.c (the last conditional) and pytables compiled fine. I will run the test set tomorrow.
Thanks, Chris #if defined(_POSIX_BARRIERS) && ( (_POSIX_BARRIERS - 20012L) >= 0 && _POSIX_BARRIERS != 200112L) On Mar 7, 2012, at 8:25 PM, Francesc Alted wrote: > On Mar 7, 2012, at 6:05 PM, Chris Kees wrote: >> On Wed, Mar 7, 2012 at 5:18 PM, Francesc Alted <fal...@pytables.org> wrote: >> On Mar 7, 2012, at 3:14 PM, Francesc Alted wrote: >> >>> On Mar 7, 2012, at 1:57 PM, Chris Kees wrote: >>> >>>> Hi Anthony, >>>> >>>> I'm using 1.8.7 on ranger. Here is the topmost error: >>>> >>>> compile options: '-DNDEBUG=1 -DHAVE_BZ2_LIB=1 -Iblosc >>>> -I/work/01020/cekees/ranger.gnu/build/ranger.gnu/lib/python2.7/site-packages/numpy/core/include >>>> -I/work/01020/cekee >>>> s/ranger.gnu/build/ranger.gnu/include >>>> -I/work/01020/cekees/ranger.gnu/build/ranger.gnu/include/python2.7 -c' >>>> extra options: '-Isrc -DH5_USE_16_API' >>>> gcc: blosc/blosclz.c >>>> gcc: src/H5ARRAY.c >>>> In file included from >>>> /work/01020/cekees/ranger.gnu/build/ranger.gnu/include/python2.7/Python.h:8, >>>> from src/utils.h:1, >>>> from src/H5ARRAY.c:4: >>>> /work/01020/cekees/ranger.gnu/build/ranger.gnu/include/python2.7/pyconfig.h:1155:1: >>>> warning: "_POSIX_C_SOURCE" redefined >>>> In file included from >>>> /work/01020/cekees/ranger.gnu/build/ranger.gnu/include/H5public.h:37, >>>> from >>>> /work/01020/cekees/ranger.gnu/build/ranger.gnu/include/hdf5.h:24, >>>> from src/H5ATTR.h:19, >>>> from src/H5ARRAY.c:1: >>>> /share/apps/gcc_amd/4.4.5/bin/../lib/gcc/x86_64-unknown-linux-gnu/4.4.5/include-fixed/features.h:199:1: >>>> warning: this is the location of the previous definition >>>> In file included from >>>> /work/01020/cekees/ranger.gnu/build/ranger.gnu/lib/python2.7/site-packages/numpy/core/include/numpy/ndarraytypes.h:1972, >>>> from >>>> /work/01020/cekees/ranger.gnu/build/ranger.gnu/lib/python2.7/site-packages/numpy/core/include/numpy/ndarrayobject.h:17, >>>> from >>>> /work/01020/cekees/ranger.gnu/build/ranger.gnu/lib/python2.7/site-packages/numpy/core/include/numpy/arrayobject.h:15, >>>> from src/utils.h:2, >>>> from src/H5ARRAY.c:4: >>>> /work/01020/cekees/ranger.gnu/build/ranger.gnu/lib/python2.7/site-packages/numpy/core/include/numpy/npy_deprecated_api.h:11:2: >>>> warning: #warning "Using deprecated NumPy >>>> API, disable it by #defining NPY_NO_DEPRECATED_API" >>>> /work/01020/cekees/ranger.gnu/build/ranger.gnu/lib/python2.7/site-packages/numpy/core/include/numpy/__multiarray_api.h:1666: >>>> warning: '_import_array' defined but not use >>>> d >>>> gcc: blosc/blosc.c >>>> blosc/blosc.c:79: error: expected '=', ',', ';', 'asm' or '__attribute__' >>>> before 'barr_init' >>> >>> Uf, that looks really ugly. Which compiler and version are you using? Can >>> you look in this machine's thread.h and tell us the value of >>> _POSIX_BARRIERS symbol? >> >> BTW, numexpr also uses the same parallel engine than blosc. As numexpr is a >> requisite for PyTables, I'm curious about if you have been able to compile >> it too. >> >> >> Yes, I was able to compiler numexpr. The versions of pytables dependencies >> I'm using are: >> >> * Found numpy 1.7.0.dev-0b400cb package installed. >> * Found numexpr 1.5.dev package installed. > > Hmm, this is a good sign. Blosc has probably some more hacks for the > threading code, so this is probably why your compiler is behaving > differently. The compiler version should be fine too. Do you have > internally an expert in posix threading that can have a look at the offending > line? I'm a bit out of ideas. > > -- Francesc Alted > > > > > > > > ------------------------------------------------------------------------------ > Virtualization & Cloud Management Using Capacity Planning > Cloud computing makes use of virtualization - but cloud computing > also focuses on allowing computing to be delivered as a service. > http://www.accelacomm.com/jaw/sfnl/114/51521223/ > _______________________________________________ > Pytables-users mailing list > Pytables-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/pytables-users ------------------------------------------------------------------------------ Virtualization & Cloud Management Using Capacity Planning Cloud computing makes use of virtualization - but cloud computing also focuses on allowing computing to be delivered as a service. http://www.accelacomm.com/jaw/sfnl/114/51521223/ _______________________________________________ Pytables-users mailing list Pytables-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/pytables-users
