NOTICE: This version of the NSF Unidata web site (archive.unidata.ucar.edu) is no longer being updated.
Current content can be found at unidata.ucar.edu.
To learn about what's going on, see About the Archive Site.
NOTE: The netcdf-hdf
mailing list is no longer active. The list archives are made available for historical reasons.
"Robert E. McGrath" <mcgrath@xxxxxxxxxxxxx> writes: > Greetings, > > This note is to kick off a discussion of how netcdf should configure, in > order to discover hdf5. > > The current setup (as of alpha 2) uses h5cc. > > I.e., HDF 5 is installed, then in the netcdf4 directory, you > > 1. set CC to h5cc or h5pc > 2. ./configure --enable-netcdf-4 [--enable-parallel] [...] > > We have encountered problems, apparently due to subtle interactions of > h5cc and automake's generated makefiles. > > I libtoolized netcdf4, which fixed some, but not all, of the problems > wiht > h5cc. > > > > We need to come up with some way to reliably build netCDF4. > > One suggestion has been to abandon h5cc/h5pcc, substituting other > methods > of discovering which HDF5 to use, what compilers and paths should be > used, > etc.. > > I pose now pose the question: > > What is the recommended approach? > > I think Elena made the point that perhaps I should not be using h5cc, but that other users might find it useful. In other words, netCDF-4 is not the typical HDF5 user, so perhaps you don't want to generalize any netCDF-4 solution to the broader HDF5 community. Let's explore the issue of not using h5cc, and see how difficult it would be. To compile I need -I paths (for include files), -L paths (for library files), and the names of the libraries. Is it true that there is more than one library involved? Ed -- Ed Hartnett -- ed@xxxxxxxxxxxxxxxx
netcdf-hdf
archives: