Hi Luis, > there is a very subtle bug in netcdf 4.0.1. In netcdf.h the following > mpi entities are defined > > /* These defs added by netCDF configure because parallel HDF5 is not > present. */ > #define MPI_Comm int > #define MPI_Info int > #define MPI_COMM_WORLD 0 > #define MPI_INFO_NULL 0 > > in case mpi.h is included before netcdf.h this defines (may) break the > MPI implementation. > > I think that is a serious design problem and things like this should > definitely be avoided. > > It would be nice to include this problem in the "Know problems" section. I have entered your report in the "known problems" document at http://www.unidata.ucar.edu/software/netcdf/docs/known_problems.html Thanks for reporting this! It would be a difficult problem to diagnose. I will try to make sure it's fixed before the 4.1 release. --Russ Russ Rew UCAR Unidata Program address@hidden http://www.unidata.ucar.edu Ticket Details =================== Ticket ID: YDV-418475 Department: Support netCDF Priority: Normal Status: Closed