Hi, I think this may be another case of a problem we have seen with building netCDF on OSX Mavericks. Here are two suggested solutions. The first is easy, if it works, but the second is probably required. Solution 1: Run "make clean" in your netcdf build directory, then try again without changing anything else about how you built netCDF from source. Solution 2, from Ward): The problem appears to be related to the shell scripts used to run some of the tests. They rely on a 'srcdir' environmental variable being set; this variable is not being set properly for OSX, and there is no easy way to set it as it changes related to the particular test being run. A work around for this would be to go into those directories with tests that fail and run a subset of the tests manually. For example: $ netcdf-4.1.3/ncdump> ./run_tests.sh fails, but $ netcdf-4.1.3/ncdump> srcdir=`pwd` ./run_tests.sh passes. I'm sorry I don't have a better fix for you, but at least I believe it's safe to assume the main problem is with the shell scripts, and not the netcdf libraries you've compiled. It is worth noting that netcdf-4.1.3 is a fairly old version and is currently unsupported. NetCDF provides backwards compatibility with previous versions, however, so any application using netcdf-4.1.3 should also work with 4.3.1.1 (the latest release). The 4.3.1.1 release may be found here: http://github.com/Unidata/netcdf-c/releases I'm not sure why the above earlier response to a similar problem didn't get archived in our searchable support responses, but it should have been. --Russ > Following your suggestions with CPPFLAGS and LDFLAGS, I was able to > compile the netcdf-c libraries: > > [birch:local/netcdf-c-4.3.1.1/lib] laura% file libnetcdf.dylib > libnetcdf.dylib: Mach-O 64-bit dynamically linked shared library x86_64 > > However, make check failed to reach the end. So I am not sure if it was > compiled correctly or if the test failed because I used 64-bit. I am > attaching a file check.out which is the output of "make check". > > I am going to try to compile the netcdf-fortran librairies next. > > Thanks for all your help with this. > Laura > > > On 3/17/14 2:51 PM, Laura Fowler wrote: > > Hi Ward: > > Here is an answer to your first question about the hdf5 libraries. The > > content of /birch/laura/local/hdf5-1.8.12/lib looks like this: > > > > drwxr-xr-x 2 laura 15 510 Mar 17 12:04 ./ > > drwxr-xr-x 6 laura 15 204 Mar 17 12:04 ../ > > -rwxr-xr-x 1 laura 15 6151932 Mar 17 12:04 libhdf5.8.dylib* > > -rw-r--r-- 1 laura 15 7919992 Mar 17 12:04 libhdf5.a > > lrwxr-xr-x 1 laura 15 15 Mar 17 12:04 libhdf5.dylib@ -> > > libhdf5.8.dylib > > -rwxr-xr-x 1 laura 15 1054 Mar 17 12:04 libhdf5.la* > > -rw-r--r-- 1 laura 15 3328 Mar 17 12:04 libhdf5.settings > > -rw-r--r-- 1 laura 15 1210296 Mar 17 12:04 libhdf5_fortran.a > > -rwxr-xr-x 1 laura 15 1080 Mar 17 12:04 libhdf5_fortran.la* > > -rwxr-xr-x 1 laura 15 172304 Mar 17 12:04 libhdf5_hl.8.dylib* > > -rw-r--r-- 1 laura 15 209608 Mar 17 12:04 libhdf5_hl.a > > lrwxr-xr-x 1 laura 15 18 Mar 17 12:04 libhdf5_hl.dylib@ -> > > libhdf5_hl.8.dylib > > -rwxr-xr-x 1 laura 15 1118 Mar 17 12:04 libhdf5_hl.la* > > -rw-r--r-- 1 laura 15 485864 Mar 17 12:04 libhdf5hl_fortran.a > > -rwxr-xr-x 1 laura 15 1189 Mar 17 12:04 libhdf5hl_fortran.la* > > > > and doing "file" on the 2 libraries below indicate to me that both > > *.dylib libraries have been compiled with 64-bit. Doing "file" on the > > other libraries (like libhdf5.a) do not provide information on single > > versus double precision (not sure why. I also used ar t to try to get > > that information): > > > > [birch:local/hdf5-1.8.12/lib] laura% file libhdf5.8.dylib > > libhdf5.8.dylib: Mach-O 64-bit dynamically linked shared library x86_64 > > > > [birch:local/hdf5-1.8.12/lib] laura% file libhdf5_hl.dylib > > libhdf5_hl.dylib: Mach-O 64-bit dynamically linked shared library x86_64 > > > > [birch:local/hdf5-1.8.12/lib] laura% file libhdf5.a > > libhdf5.a: current ar archive random library > > > > so it seems to me that I correctly compiled the hdf5 libraries as > > 64-bit libraries. I will work on your other suggestions next. > > > > Thanks, > > Laura > > > > > > > > > > > > > > On 3/17/14 2:17 PM, Unidata netCDF Support wrote: > >> Hi Laura, > >> > >> Thank you for sending your config.log! I've taken a look, and the > >> relevant error appears to be this: > >> > >> configure:4330: checking whether the C compiler works > >> configure:4352: gcc /birch/laura/local/hdf5-1.8.12/include > >> /birch/laura/local/hdf5-1.8.12/lib conftest.c >&5 > >> ld: in /birch/laura/local/hdf5-1.8.12/include, can't map file, > >> errno=22 for architecture x86_64 > >> collect2: ld returned 1 exit status > >> > >> So, your compiler is attempting to build a 64-bit netcdf library, and is > >> failing to link against the hdf5 library. From the terminal, can you cd > >> to /birch/laura/local/hdf5-1.8.12/lib and run 'file [hdf5 library > >> name]' ? The 'file' command should tell us about the hdf5 library it > >> found and what architecture it was compiled for. I would *expect* it to > >> be compiled as a 64-bit library, but this error indicates otherwise. > >> > >> Although having typed that, I wonder if we need to tweak your CPPFLAGS > >> and LDFLAGS environmental variables. For example, under a bash shell, I > >> would expect them to be defined as follows: > >> > >> CPPFLAGS="-I/birch/laura/local/hdf5-1.8.12/include" > >> LDFLAGS="-L/birch/laura/local/hdf5-1.8.12/lib" > >> > >> You have omitted the '-I' and '-L', which also may be causing a > >> problem. What happens if you add them and run configure again? > >> > >> -Ward > >> > >> On 3/17/14, 2:08 PM, Laura Fowler wrote: > >>> New Client Reply: cannot configure > >>> > >>> Hi Ward: > >>> Thanks for your help with this. Yep, I cannot build the netcdf-c > >>> library > >>> either using the following setting: > >>> > >>> setenv CC 'gcc -m64' (I used -m64 to build the hdf5 libraries > >>> first) > >>> setenv FC pgf90 > >>> setenv CPPFLAGS /birch/laura/local/hdf5-1.8.12/include (location of the > >>> hdf5 include directory) > >>> setenv LDFLAGS /birch/laura/local/hdf5-1.8.12/lib (location of the hdf5 > >>> lib directory) > >>> > >>> and using the following configure > >>> > >>> ./configure --prefix=/birch/laura/local/netcdf-c-4.3.1.1 > >>> --enable-netcdf4 --disable-dap > >>> > >>> [birch:local/src/netcdf-c-4.3.1.1] laura% ./configure > >>> --prefix=/birch/laura/local/netcdf-c-4.3.1.1 --enable-netcdf4 > >>> --disable-dap > >>> configure: netCDF 4.3.1.1 > >>> checking build system type... x86_64-apple-darwin11.4.2 > >>> checking host system type... x86_64-apple-darwin11.4.2 > >>> checking target system type... x86_64-apple-darwin11.4.2 > >>> checking for a BSD-compatible install... /usr/bin/install -c > >>> checking whether build environment is sane... yes > >>> checking for a thread-safe mkdir -p... ./install-sh -c -d > >>> checking for gawk... no > >>> checking for mawk... no > >>> checking for nawk... no > >>> checking for awk... awk > >>> checking whether make sets $(MAKE)... yes > >>> checking whether make supports nested variables... yes > >>> checking whether to enable maintainer-specific portions of > >>> Makefiles... no > >>> configure: checking user options > >>> checking whether a win32 DLL is desired... no > >>> checking whether a NCIO_MINBLOCKSIZE was specified... 256 > >>> checking if fsync support is enabled... no > >>> checking whether extra valgrind tests should be run... no > >>> checking whether we should try to build netCDF-4... yes > >>> checking do we require hdf5 dynamic-loading support... xno > >>> checking whether reading of HDF4 SD files is to be enabled... no > >>> checking whether to fetch some sample HDF4 files from Unidata ftp site > >>> to test HDF4 reading (requires wget)... no > >>> checking whether parallel I/O for classic and 64-bit offset files using > >>> parallel-netcdf is to be enabled... no > >>> checking whether extra example tests should be run... no > >>> checking whether parallel IO tests should be run... no > >>> checking whether a default chunk size in bytes was specified... 4194304 > >>> checking whether a maximum per-variable cache size for HDF5 was > >>> specified... 67108864 > >>> checking whether a number of chunks for the default per-variable cache > >>> was specified... 10 > >>> checking whether a default file cache size for HDF5 was specified... > >>> 4194304 > >>> checking whether a default file cache maximum number of elements for > >>> HDF5 was specified... 1009 > >>> checking whether a default cache preemption for HDF5 was > >>> specified... 0.75 > >>> checking whether netCDF-4 logging is enabled... no > >>> checking whether cdmremote client is to be built... no > >>> checking whether DAP client is to be built... no > >>> checking for style of include used by make... GNU > >>> checking for gcc... gcc -m64 > >>> checking whether the C compiler works... no > >>> configure: error: in `/birch/laura/local/src/netcdf-c-4.3.1.1': > >>> configure: error: C compiler cannot create executables > >>> > >>> I am also attaching the log file. > >>> > >>> Thanks, > >>> Laura > >>> > >>> > >>> > >>> > >>> > >>> On 3/17/14 1:34 PM, Unidata netCDF Support wrote: > >>>> Hi Laura, > >>>> > >>>> The netcdf-fortran libraries rely on the netcdf-c libraries. > >>>> netcdf.h is > >>>> part of the netcdf-c libraries. You can download our latest release, > >>>> 4.3.1.1 here: > >>>> > >>>> http://github.com/Unidata/netcdf-c/releases > >>>> > >>>> With instructions for building it found here: > >>>> > >>>> http://www.unidata.ucar.edu/software/netcdf/docs/building.html > >>>> > >>>> If you have any questions about installing netcdf-c, please let us > >>>> know. Once installed, you should be able to build the fortran > >>>> libraries > >>>> without any problem. > >>>> > >>>> Have a great day, > >>>> > >>>> -Ward > >>>> > >>>> On 3/17/14, 1:29 PM, address@hidden wrote: > >>>>> New Ticket: cannot configure > >>>>> > >>>>> Full Name: Laura D Fowler > >>>>> Email Address: address@hidden > >>>>> Organization: NCAR/NESL/MMM > >>>>> Package Version: netcdf-fortran-4.2 > >>>>> Operating System: OSX 10.7.5 > >>>>> Hardware: > >>>>> Description of problem: Hello: > >>>>> > >>>>> I would like to compile netcdf-fortran-4.2 on my local machine, > >>>>> but get the error message: > >>>>> > >>>>> checking for netcdf.h... no > >>>>> configure: error: netcdf.h could not be found. Please set CPPFLAGS. > >>>>> > >>>>> after simply doing the following > >>>>> setenv CC gcc > >>>>> setenv FC pgf90 > >>>>> > >>>>> I cannot see the file netcdf.h anywhere in the sourcecode so I am > >>>>> not sure how to set CPPFLAGS. I read that setting that option > >>>>> would help, but it did not and I do not know what to do next. Any > >>>>> other option that I need to add. > >>>>> > >>>>> Thanks, > >>>>> Laura > >>>>> > >>>>> > >>>>> > >>>>> > >>>>> > >>>>> > >>>>> Ticket Details > >>>>> =================== > >>>>> Ticket ID: ZVE-924213 > >>>>> Department: Support netCDF > >>>>> Priority: Normal > >>>>> Status: Open > >>>>> Link: > >>>>> https://www.unidata.ucar.edu/esupport/staff/index.php?_m=tickets&_a=viewticket&ticketid=23627 > >>>> Ticket Details > >>>> =================== > >>>> Ticket ID: ZVE-924213 > >>>> Department: Support netCDF > >>>> Priority: Normal > >>>> Status: Open > >>>> > >> > >> > >> Ticket Details > >> =================== > >> Ticket ID: ZVE-924213 > >> Department: Support netCDF > >> Priority: Normal > >> Status: Open > >> > > > > > > > -- > !---------------------------------------------------- > Laura D. Fowler > Mesoscale and Microscale Meteorology Division (MMM) > National Center for Atmospheric Research > P.O. Box 3000, Boulder CO 80307-3000 > > e-mail: address@hidden > phone : 303-497-1628 > > !---------------------------------------------------- > > > Russ Rew UCAR Unidata Program address@hidden http://www.unidata.ucar.edu Ticket Details =================== Ticket ID: ZVE-924213 Department: Support netCDF Priority: Normal Status: Closed
NOTE: All email exchanges with Unidata User Support are recorded in the Unidata inquiry tracking system and then made publicly available through the web. If you do not want to have your interactions made available in this way, you must let us know in each email you send to us.