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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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.
Andrew, > To: address@hidden > cc: address@hidden > From: address@hidden (Andrew Goodwillie) > Subject: netCDF bug? > Organization: . > Keywords: 199911222258.PAA17893 In the above message, you wrote: > I think that there is a corruption in the netcdf-3.3.1.tar.gz file > that is located in the pub/netcdf/old directory in your FTP site. > I've successfully installed this package many times in the past > and when I use an older ".Z" UNIX-compressed version of this tar > file it install perfectly, as expected. The only difference is that > now the tar file is compressed with gzip and has been moved to the > old sub-directory. Other clues that this gzip version is bad are > that on your FTP site the gzipped tar file of version 3.3.1 is > twice the size of the UNIX-compressed version that I had previously > downloaded, and, when the gzipped version in uncompressed the > resultant tar file is over three times as big as the tar file that > uncompresses from the .Z version. > > This past weekend when trying to install the package (downloaded on > 11/20/99 as file "netcdf-3.3.1.tar.gz" from your FTP site) I repeatedly > ran into the following error when I ran ./configure. I'm working on a > Sun ultra 5, running Solaris 7, 64-bit capability is disabled, and I > have the official Sun C/F77/F90 compilers as well as gcc: > > winter2# ./configure > loading cache ./config.cache > ./configure: -64: not found > checking for m4... (cached) m4 > checking for xlc... no > checking for c89... c89 > checking C compiler... works > checking how to make dependencies... cc -xM > checking for CC... no > checking for cxx... no > checking for c++... c++ > checking C++ compiler "c++"... works > checking how to run the C preprocessor... c89 -E > checking for f77... f77 > checking for FORTRAN .F compiler... > checking if FORTRAN compiler handles *.F files... yes > checking for C-equivalent to FORTRAN routine "SUB"... + cat > doit=$FC -c ${FFLAGS} conftest.f > + eval echo configure:1456: "$FC -c ${FFLAGS} conftest.f" > + echo configure:1456: f77 -c -O conftest.f > + nm conftest.o > + awk > /SUB_/{print "SUB_";exit} > /SUB/ {print "SUB"; exit} > /sub_/{print "sub_";exit} > /sub/ {print "sub"; exit} > FCALLSCSUB=sub_ > + echo sub_ > sub_ > + rm -f conftest.f conftest.o > + exit > winter2# > > > As you can see, the second output line does not look good. Fortunately I > found a version of 3.3.1 that I had downloaded - as a .Z compressed file - > from your site a couple of years ago. I have been able to install this > with no problem whatsoever. Indeed, this weekend I did some tests and have > been able to download and install all other versions of netCDF (both the > older and more recent versions) that are on your FTP site. > > So, there's some problem with the gzipped tar file for version 3.3.1 that's > on your FTP site. > > Can you check it? > > Thanks! > > Dr. Andrew Goodwillie > Geological Data Center > Scripps Institution of Oceanography > La Jolla > CA 92093-0220 Thanks for sending this in. Yes, there appears to be a problem with the "netcdf-3.3.1.tar.gz" distribution. I uncompressed it on my Solaris 7 system and ran into the same problem. Near as I can figure out, the distribution was created by tar-ing up and compressing a working version of the package (with unique modifications to the configure script) rather than a virgin version of the package. Best to stay with version 3.4 of the netCDF package. -------- Steve Emmerson <http://www.unidata.ucar.edu>