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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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.
>>To: address@hidden >>From: "Gary Moore" <address@hidden> >>Subject: netCDF Fortran (77 + 90) - Failure of prebuilt libraries >>Organization: UCAR/Unidata >>Keywords: 200503042339.j24NdwV9019166 > > Institution: Earth tech > Package Version: version 3.6.0 > Operating System: Window 2000 and XP > Hardware Information: Intel P4 and AMD XP (no Athlon-64\'s) > Inquiry: Lahey Computer > LF95 Pro v5.6 > Problem: Onerous changes are required to make executables and one must > include > the MSVC linker. Name mangling used by UNIDATA is not standard according to > Lahey support staff. I need a foolproof, simple static library that people > who > have no adminstrative rights to install software like the visual C++ linker > can link to. It seems that writing a kluge atop NCDUMP and NCGEN is my only > hope - this is a far, far cry from NETCDF's original goal for transparency. > Can UNIDATA do something for the corporate IS disabled? We do have to use EPA > data sets which are being supplied (horrors) as NETCDF. If it were not for > the > laborious effort at programming around the unsigned integer oversight on the > typical fortran compiler I would tackle a reprogramming in pure fortran 95 >>From the XDR up. Has anybody done that????? > Gary, I'm aware that there are problems with the fortran libraries in the windows binaries. Have you tried to build from the source? The fortran name mangling is caused by a package that creates a fortran interface from C code, and this is configurable when building from source. Ed -- Ed Hartnett -- address@hidden