This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.
> Hi Glenn- > > Hopefully, this is a fast question... > > Call me a masochist, but I wanted to make sure *I* could build netCDF > on our T90, so I downloaded the tar file and went thru all the > motions. Everything went well during the build process, with the > exception of some warnings like this: > However, the "make test" had some problems. The output is pretty huge, > but here's where it all started: > > -------- > Making `test' in directory /ftmp/jps/tmpdir.093168a/unidata.ucar.edu/pub/netcdf/netcdf-3.3/src/fortran > > + f90 -c -g -i64 -dp -Wp-F -I../libsrc -DNDEBUG ftest.F > FORCE: No action was taken. > f90 -o ftest -g -i64 -dp -Wp-F ftest.o -L../libsrc -lnetcdf > ./ftest > NCCRE: bad flag, did you forget to include netcdf.inc? > ncdimdef: -1 is not a valid cdfid > ncdimdef: -1 is not a valid cdfid > ncdimdef: -1 is not a valid cdfid > ... > --------- > > Have you actually built 3.3 on our T90 fram the "official" tar file? > > I'll paste my command sequence in at the bottom...maybe I'm screwing > something up... > > Thanks- > John To tell you the truth, I'm not sure what the status is on that platform. The offical release of netcdf-3 includes an "new" way of doing the fortran calling C stuff. This has not worked on most exotic platforms, and I've spent the last month beating on the responsible parties to get is as rugged and portable as the C library. The good news is that we are almost there. I'll keep you posted. In the interium, there should be a working version in the 'ncar' account on t90. This uses the old m4 way of doing the fortran to C bridge, and just has the old netcdf-2 Fortran interfaces. -glenn