Hi Chris, I'm not sure why this would be happening; is this new behavior after setting the F77 environmental variable, or a continuation of the existing issue? Is there a new set of config.log/make.out/etc files to look at, or are they the same? If there are new ones, would you mind providing them and I'll take a look to sort out what's going on? If not, what happens if you try setting F77? Looking at the `Makefile.am` file in the fortran/ directory, it appears that the files in `libsrc/` are only used if `BUILD_F03` is not defined. I'm not completely certain, but I believe that this is because the `libsrc` directory provides code for interfacing with the C library via `cfortran.h`. On systems that are F03 compliant, there is a different infrastructure in place that is contained within the `fortran/` directory. My apologies for the vague answer, but I'm uncertain of where we're currently at with the F77 variable, and also because I'm not primarily a fortran developer. I hope this helps, -Ward > Hi Ward, > > It appears that the code in the the subdirectory "libsrc" is not being > complied and/or added to the libnetcdff.a. > Rather, it appears as an include "-I../libsrc" in the compilation of the > "fortran" directory. > > The "libsrc" directory seems to be the only place where the unresolved > symbols (e.g. nf_create_par in fort-nc4.c) > are bound to the netcdf C-functions. > > Any idea why the "libsrc" files are not being compiled? > > Thanks, > Chris > > > > On May 19, 2016, at 9:42 AM, Unidata netCDF Support > <address@hidden<mailto:address@hidden>> > wrote: > > Hello Chris, > > From the file `make.out`, I see the following: > > /bin/sh ../libtool --tag=F77 --mode=link -I../fortran -I../fortran > -Msignextend -L/home/rousculp/Platforms/DarwinPGI/lib -o nf_test > f03lib_f_interfaces.o test_get.o test_put.o nf_error.o nf_test.o test_read.o > test_write.o util.o f03lib.o ../fortran/libnetcdff.la -lnetcdf > > It appears that you need to set your `F77` environmental variable to reflect > your compiler; it is currently empty, and instead of passing `-I../fortran` > to the compiler, it is being passed to `libtool`, which complains that it is > an unrecognized option. > > I hope this helps, > > -Ward > > Full Name: chris rousculp > Email Address: address@hidden<mailto:address@hidden> > Organization: LANL > Package Version: netcdf-4.4.0, netcdf-fortran-4.4.4 > Operating System: osx 10.8.5 > Hardware: quad core intel xenon > Description of problem: Hi, > > I'm getting a netcdf fortran package build error when make check. See > attached files with build script, output, and configure.status for both C and > Fortran. > > Thanks, > Chris > > > > > > Ticket Details > =================== > Ticket ID: EOO-961163 > Department: Support netCDF > Priority: Normal > Status: Closed > > > ----------------------------------------------- > Christopher L. Rousculp > Los Alamos National Laboratory > P.O. Box 1663, MS-F699 > Los Alamos NM 87545 > (505) 665-3678 Phn > (505) 500-2712 Mbl > (505) 665-7725 Fax > ---------------------------------------------- > > > Ticket Details =================== Ticket ID: EOO-961163 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.