[MOAB-dev] MPI implementations and SEEK_*

[Jason Kraftcheck]

MOAB already defines MPICH_IGNORE_CXX_SEEK for internal use (line 121 of
configure.in .).  Presumably this stuff in MBProcConfig.hpp is for the
benefit of applications including that header (indirectly from MBParallelComm.)

- jason



Jed Brown wrote:
> MBProcConfig.hpp has this kludge so that MPICH2 does not complain about
> the SEEK_* macros (defined in iostream) conflicting with the MPI symbols
> MPI::SEEK_*.
> 
> /* MPICH2 will fail if SEEK_* macros are defined
>  * because they are also C++ enums. Undefine them
>  * when including mpi.h and then redefine them
>  * for sanity.
>  */
> #  ifdef SEEK_SET
> #    define MB_SEEK_SET SEEK_SET
> #    define MB_SEEK_CUR SEEK_CUR
> #    define MB_SEEK_END SEEK_END
> #    undef SEEK_SET
> #    undef SEEK_CUR
> #    undef SEEK_END
> #  endif
> #include "mpi.h"
> #  ifdef MB_SEEK_SET
> #    define SEEK_SET MB_SEEK_SET
> #    define SEEK_CUR MB_SEEK_CUR
> #    define SEEK_END MB_SEEK_END
> #    undef MB_SEEK_SET
> #    undef MB_SEEK_CUR
> #    undef MB_SEEK_END
> #  endif
> 
> 
> As discussed in this FAQ entry
> 
>   http://www.mcs.anl.gov/research/projects/mpich2/support/index.php?s=faqs#cxxseek
> 
> an alternative is to define MPICH_IGNORE_CXX_SEEK which will prevent
> MPICH2 from defining the MPI::SEEK_* symbols.  Note that MPI_SEEK_* are
> always defined.
> 
> I mention this because the development version of Open MPI is
> complaining about the current hack.  I suggest defining
> MPICH_IGNORE_CXX_SEEK and removing the hack.  Note that this also fixes
> the order-dependence alluded to in this comment (in MBCore.cpp)
> 
> /* Leave MBParallelComm.hpp before mpi.h or MPICH2 will fail
>  * because its C++ headers do not like SEEK_* macros.
>  */
> 
> 
> Jed


-- 
"A foolish consistency is the hobgoblin of little minds" - Ralph Waldo Emerson