On 29-07-15 22:39, Charlie Zender wrote:
> Any hints here would be appreciated...
> NCO builds fine on many different machines.
> It appears to me that new Debian netCDF package seems
> to cause NCO's configure test for strdup() to fail,
> so NCO uses its own internal prototype for strdup()
> instead, and then this prototype runs into a conflict
> with the strdup() prototype in "string.h".
> Not sure why the new netCDF package would alter previous behavior.
The biggest changes in the new netcdf library is that it's built with
the HDF5 MPI variants (openmpi on most architectures). It may be that an
MPI optimized strdup() is used now?
I'm having some doubts if it's wise to only build netcdf with the HDF5
MPI variant, instead of in addition to the HDF5 serial variant used by
most other applications.
GPG Key ID: 4096R/6750F10AE88D4AF1
Fingerprint: 8182 DE41 7056 408D 6146 50D1 6750 F10A E88D 4AF1
Pkg-grass-devel mailing list