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 Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated.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.
------- Forwarded Message Date: Wed, 24 Nov 2004 13:29:53 -0700 From: "Paquita Zuidema" <address@hidden> To: Russ Rew <address@hidden> Subject: Re: 20041118: netCDF Fortran (77 + 90) - array size limitation Hi Russ - thanks for the feedback, this is good to know. I did get netcdf to compile successfully last night. Two changes I made: 1. set -YCRFL=1, not -YCRFL=l , i.e. set it to one, not "el". previous mistake on my part. 2. explicitly set CC=/usr/bin/cc. I don't know if this is really necessary, but it didn't hurt. Otherwise, I followed Dierk Polzin's recommended build. I believe that fortran code calling netcdf also needs to be compiled with the -YCRFL=1 flag (not mentioned on his build). This process somehow removed my IDL license file and IDL path ! That was odd, probably a poor previous installation on my part, but still. I wasn't completely done with this project until I relearned how to install the IDL license, but, now things seem to be fine. Thanks again - seeing that this process isn't completely transparent even to a trained professional as yourself helped me to persevere. Paquita Russ Rew wrote: >>To: address@hidden >>From: "Paquita Zuidema" <address@hidden> >>Subject: 20041118: netCDF Fortran (77 + 90) - array size limitation >>Organization: CU/CIRES >>Keywords: 200411182221.iAIMLur5029116 netCDF MacOS-X >> >> > >Hi, > > > >>well, disappointing news. I decided to start over - Ihad >>installed netcdf using Eric Salathe's recommendation, and >>noticed on the email archives that you recommend >>Dierk Polzin's build. got rid of all my netcdf directories, files. >>downloaded netcdf.tar.Z, applied the absoft90_macosx.patch, >>set the environment variables as recommended by Dierk. >>output from ./configure, and from make test (test.log) >>attached. test.log does appear improved over what I >>sent you earlier. >> >> > >The test.log you sent seemed to be truncated after > > Making `test' in directory /Users/pz/netcdf-3.5.1/src/nc_test > > acc -o nc_test nc_test.o error.o test_get.o test_put.o test_read.o > test_write.o util.o -L../libsrc -lnetcdf > ./nc_test -c > ./nc_test > >so it doesn't show whether the nf_test or f90 directories succeeded. >I'll assume those succeeded for you. > > > >>I go to /src/f90 and do a make to create NETCDF.mod and >>TYPESIZES.mod, and copy these over as netcdf.mod and >>typesizes.mod into the directory where "read_sonia.f95" is >>located. I recompile read_sonia using the -YEXT_NAME=LCS >>and -YEXT_SFX=_ flags. >> >>It doesn't compile, and the error messages,attached as make.log, >>don't make sense to me - for one thing they are in upper-case. >> >> > >Also, the symbols all seem to have "_in_NETCDF" appended to them, >which is mystifying. > > ld: Undefined symbols: > ___NF90_CLOSE_in_NETCDF > ___NF90_GET_ATT_TEXT_in_NETCDF > ___NF90_GET_VAR_3D_FOURBYTEREAL_in_NETCDF > ___NF90_INQUIRE_DIMENSION_in_NETCDF > > > >>One thing I noted in config.log is that the c/c++ compiler >>has trouble w/ conftest.c, a clue ? >> >> > >Not much of one, unfortunately. The configure script is expected to >fail in compiling some tests, and all the C tests are named >conftest.c. > >I downloaded a trial version of the AbsoftProFortran for Mac OS X and >installed it, but could not get "make test" to work on the Fortran and >Fortran90 tests with various combinations of flags, although I was using >the cfortran.h version 4.3 from netCDF-3.6.0-beta6 instead of the >patched cfortran.h version 4.1 from netCDF-3.5.1, in hopes that would >work. I kept getting undefined externals, such as _nf_close_ or >_NF_CLOSE_, depending on which flags I used. > >I noticed a 329 page f90 manual came with the Absoft compiler, >but trying the recommendations in "Chapter 8: Interfacing with Other >Languages" didn't work for me either. I notice I was using "cc", the >gcc compiler that comes with Mac OS X and you seemed to be using >"acc", the Absoft C compiler, so there's another difference. > >Sorry I can't be of more help, but this looks like it might take an >expert in using the Absoft compilers on Mac OS X ... > >--Russ > > -- +++++++++++++++++++++++++++++++++++ Paquita Zuidema, Ph.D. CIRES - NOAA Environmental Technology Lab phone: 303-497-4284 fax: 303-497-6181 web: http://www.etl.noaa.gov/~pzuidema mail: R/ET6, 325 Broadway, Boulder, CO 80305 +++++++++++++++++++++++++++++++++++ ------- End of Forwarded Message