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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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.
> Hmmm, that wasn't a great success > > Where netcdf-4.1.2-beta1 fails only the two tests I mentioned, > netcdf-4.1.2-beta2-snapshot2010111112 fails 23, most of them with "Illegal > Instruction". Here is the output from the first failed test: > > *** Testing netCDF-3 Fortran 77 API. > /bin/sh: 234416 Illegal instruction > FAIL: nf_test > > and here's the end of the output > > ================================================ > 23 of 25 tests failed > Please report to address@hidden > ================================================ > make[2]: *** [check-TESTS] Error 1 > make[2]: Leaving directory > `/gpfs_hpcf/filesets/hpcf/scratch/hadfield/netcdf/netcdf-4.1.2-beta2-snapshot2010111112/nf_test' > make[1]: *** [check-am] Error 2 > make[1]: Leaving directory > `/gpfs_hpcf/filesets/hpcf/scratch/hadfield/netcdf/netcdf-4.1.2-beta2-snapshot2010111112/nf_test' > make: *** [check-recursive] Error 1 Oops, that's the trouble with development snapshots, they're usually not ready for release. If you really want to get rid of the two "make check" errors you saw with 4.1.2-beta1, you could go into the public svn trees for netCDF and try revision 442 from Sept 22, but I can't guarantee that wouldn't have other AIX errors. We generally don't test on lots of platforms until we're very close to release ... http://svn.unidata.ucar.edu/repos/netcdf/trunk/ --Russ > >>> On 2010-11-11 at 18:31, "Unidata netCDF Support" <address@hidden> wrote: > > Mark, > > > As I reported in the message earlier today, HDF5 (1.8.5-patch1) was > > broken, as you suggested. I rebuilt it with --disable-shared and now > > it passes "make check" with no failures. > > > > Then I rebuilt netCDF 4.1.2-beta1 and linked it with the new HDF5. It > > passes all tests except these 2: > > > > *** Testing netCDF-4 groups from Fortran 90. > > Error: NetCDF: Numeric conversion not representable > > STOP 2 > > FAIL: f90tst_grps > > *** Testing definition of netCDF-4 vars from Fortran 90. > > Error: NetCDF: Numeric conversion not representable > > STOP 2 > > FAIL: f90tst_vars3 > > > > These are advanced features that I don't use yet, so I will probably > > install anyway, but do you have any idea what they're about? > > Yes, I believe Ed fixed these in a change to libsrc4/nc4var.c, checked > in with the comment > > fixed bug with nf90_inquire_variable returning range error because > of chunksizes for a contiguous variable > > The fix was after 4.1.2-beta1, so currently it's only available in the > daily snapshot until the next beta release (or 4.1.2 release > candidate). If you want it you can get it from > > ftp://ftp.unidata.ucar.edu/pub/netcdf/snapshot/netcdf-4-daily.tar.gz > > --Russ > > Russ Rew UCAR Unidata Program > address@hidden http://www.unidata.ucar.edu > > > > Ticket Details > =================== > Ticket ID: RTE-131011 > Department: Support netCDF > Priority: Urgent > Status: Closed > > > Please consider the environment before printing this email. > > NIWA is the trading name of the National Institute of Water & Atmospheric > Research Ltd. > > Russ Rew UCAR Unidata Program address@hidden http://www.unidata.ucar.edu Ticket Details =================== Ticket ID: RTE-131011 Department: Support netCDF Priority: Urgent Status: Closed