Apologies forgot to CC the list. ---------- Forwarded message ---------- From: Ata ollah Mesgarnejad <a.mesgarne...@gmail.com> Date: Tue, Jul 5, 2016 at 2:56 PM Subject: Re: [Libmesh-users] libpetsc.so: undefined reference to `nc_def_var_deflate' To: John Peterson <jwpeter...@gmail.com>
configure.log <https://drive.google.com/file/d/0B3Oq4XoGCifvbXFRRWJsZVM5RjA/view?usp=drive_web> Hi John, I compiled PETSc with NetCDF and now I want to compiler libMesh. No I don't particularly want libMesh to use this NetCDF + I'm not sure it's possible (I didn't see any options in libMesh configure to let libMesh grab PETSc's NetCDF). PETSc's configure.log is ~35.3 Mbs so I can't attach it directly. I attached it using google drive this time. Best, Ata On Tue, Jul 5, 2016 at 2:47 PM, John Peterson <jwpeter...@gmail.com> wrote: > > > On Tue, Jul 5, 2016 at 12:25 PM, gmail <a.mesgarne...@gmail.com> wrote: > >> Dear developers, >> >> (I appologize if you are getting multiple copies of this, I sent this >> email previously using my other email address that I don’t think is in the >> list.) >> >> I’m getting a NetCDF symbol error when I try to compile libMesh (commit >> a58bb96d07c20eb541b557b0053269e6d8c92f15) vs PETSc 3.7.2 (commit >> 6d2c40479446d0366a4edb124b2d1250692ca322): >> >> make[1]: Entering directory >> `/home/mesgarnejad/Packages/libmesh/build/Linux-gnu-O' >> CXXLD fparser_parse-opt >> /home/mesgarnejad/Packages/petsc/Linux-gnu-O/lib/libpetsc.so: undefined >> reference to `nc_def_var_deflate' >> collect2: error: ld returned 1 exit status >> make[1]: *** [fparser_parse-opt] Error 1 >> make[1]: Leaving directory >> `/home/mesgarnejad/Packages/libmesh/build/Linux-gnu-O' >> make: *** [all-recursive] Error 1 >> >> It’s confusing since the libpetsc.so knows where to find the correct >> libnetcdf.so: >> >> >> linux-vdso.so.1 => (0x00007ffce6388000) >> libsuperlu_dist.so.5 => >> /home/mesgarnejad/Packages/petsc/Linux-gnu-O/lib/libsuperlu_dist.so.5 >> (0x00007f24966f4000) >> libparmetis.so => >> /home/mesgarnejad/Packages/petsc/Linux-gnu-O/lib/libparmetis.so >> (0x00007f24964b6000) >> libmetis.so => >> /home/mesgarnejad/Packages/petsc/Linux-gnu-O/lib/libmetis.so >> (0x00007f2496240000) >> libsuperlu.so.5 => >> /home/mesgarnejad/Packages/petsc/Linux-gnu-O/lib64/libsuperlu.so.5 >> (0x00007f2495fc0000) >> libmpi_cxx.so.1 => >> /shared/opt/openmpi/gcc-5.4.0/lib/libmpi_cxx.so.1 (0x00007f2495da6000) >> libstdc++.so.6 => /shared/opt/gcc/lib64/libstdc++.so.6 >> (0x00007f2495a18000) >> libhwloc.so.5 => /usr/lib64/libhwloc.so.5 (0x00007f24957d9000) >> libnetcdf.so.7 => >> /home/mesgarnejad/Packages/petsc/Linux-gnu-O/lib/libnetcdf.so.7 >> (0x00007f24954cf000) >> > > > I'm not clear on exactly what you've done here. You built NetCDF as an > external package in PETSc? Did you somehow tell libmesh to use this NetCDF > rather than the NetCDF that is bundled with libmesh? I'm looking at our > netcdf.m4, it does not look like we intended this to be possible (there is > not --with-netcdf= flag) so I'm not surprised that it doesn't quite work... > > > > I attached my config.log from libMesh + you can get my configure.log < >> https://www.dropbox.com/s/82yp61eequjr0p3/configure.log?dl=0>from PETSc. > > > Unfortunately dropbox links are blocked at my office, so I can't see this > link at the moment. > > -- > John > -- Best, Ata -- Best, Ata ------------------------------------------------------------------------------ Attend Shape: An AT&T Tech Expo July 15-16. Meet us at AT&T Park in San Francisco, CA to explore cutting-edge tech and listen to tech luminaries present their vision of the future. This family event has something for everyone, including kids. Get more information and register today. http://sdm.link/attshape _______________________________________________ Libmesh-users mailing list Libmesh-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-users
