getfem-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Getfem-users] Problem building getfem++ 3.0.3


From: Madhusudan Singh
Subject: Re: [Getfem-users] Problem building getfem++ 3.0.3
Date: Mon, 10 Dec 2007 16:46:09 -0700

Thanks for your suggestions.

Feeling a little emboldened by your plausible suggested solution, I made a small change to the configure directive after setting the shell variable :

./configure --enable-dd --enable-qd --enable-qhull --enable-matlab --enable-python --with-matlab-toolbox-directory=/usr/local/matlab/toolbox/getfem --with-qd-lib-dir=/usr/lib --enable-matlab-rpc

(enabling matlab-rpc)

and tried again. Now the compilation quits with :

creating libgetfemint.la
(cd .libs && rm -f libgetfemint.la && ln -s ../libgetfemint.la libgetfemint.la)
if gcc -DHAVE_CONFIG_H -I. -I. -I../.. -I../../src -I../../src  -DGMM_USES_BLAS -I/usr/local/include  -O3 -MT gfi_rpc_server.o -MD -MP -MF ".deps/gfi_rpc_server.Tpo" -c -o gfi_rpc_server.o gfi_rpc_server.c; \
        then mv -f ".deps/gfi_rpc_server.Tpo" ".deps/gfi_rpc_server.Po"; else rm -f ".deps/gfi_rpc_server.Tpo"; exit 1; fi
if gcc -DHAVE_CONFIG_H -I. -I. -I../.. -I../../src -I../../src  -DGMM_USES_BLAS -I/usr/local/include  -O3 -MT gfi_rpc_svc.o -MD -MP -MF ".deps/gfi_rpc_svc.Tpo" -c -o gfi_rpc_svc.o gfi_rpc_svc.c; \
        then mv -f ".deps/gfi_rpc_svc.Tpo" ".deps/gfi_rpc_svc.Po"; else rm -f ".deps/gfi_rpc_svc.Tpo"; exit 1; fi
if gcc -DHAVE_CONFIG_H -I. -I. -I../.. -I../../src -I../../src  -DGMM_USES_BLAS -I/usr/local/include  -O3 -MT gfi_rpc_xdr.o -MD -MP -MF ".deps/gfi_rpc_xdr.Tpo" -c -o gfi_rpc_xdr.o gfi_rpc_xdr.c; \
        then mv -f ".deps/gfi_rpc_xdr.Tpo" ".deps/gfi_rpc_xdr.Po"; else rm -f ".deps/gfi_rpc_xdr.Tpo"; exit 1; fi
make[3]: *** No rule to make target `../src/getfem/libgetfem.la', needed by `getfem_server'.  Stop.
make[3]: Leaving directory `/usr/local/src/science/getfem/getfem++-3.0.3/interface/src'
make[2]: *** [install-recursive] Error 1
make[2]: Leaving directory `/usr/local/src/science/getfem/getfem++-3.0.3/interface/src'
make[1]: *** [install-recursive] Error 1
make[1]: Leaving directory `/usr/local/src/science/getfem/getfem++-3.0.3/interface'
make: *** [install-recursive] Error 1

****  Installation failed. Aborting package creation.

Cleaning up...OK

Bye.

I then removed the rpc option and ran make. It compiled without any problems.

Any ideas ?

Thanks,

MS


On Dec 10, 2007 11:49 AM, Richard George < address@hidden> wrote:
I think this could be a missing shell variable.
if you type on the command line:

echo $MATLAB

it should report:
/usr/local/matlab

but I think the lines

>
>            Here
>
>                MATLAB = /
>
>            Please contact:
>

mean this variable is unset and you will see just a blank line.
try typing:

export MATLAB=/usr/local/matlab

(or the equivalent) and then 'make' again

You should also put that command in your shell configuration file so you
don't have to type it each time.

Best wishes

Richard



PS. You might have to run 'make clean' and then './configure ----'
I'm not sure if the $MATLAB variable is expanded when ./configure is
run, or when make is run





On Mon, 2007-12-10 at 10:46 -0700, Madhusudan Singh wrote:
> Hello,
>
> I wish to use Getfem++'s matlab interface to solve Poisson's equation
> in 3 dimensions. I have gmsh installed for providing the initial mesh
> to getfem++. I am new to finite elements (but am well versed in
> electromagnetics), but hope to learn enough to start using this tool
> quickly.
>
> I would like to report a problem and seek its solution.
>
> Build system :
>
> Linux / Ubuntu 7.10 (Gutsy).
> hostname = nanak
> uname -m = i686
> uname -r = 2.6.22-14-generic
> uname -s = Linux
> uname -v = #1 SMP Sun Oct 14 23:05:12 GMT 2007
> QD and qhull libraries installed.
> Matlab 7.4.0.336 (R2007a)
> g++ (GCC) 4.1.3 20070929 (prerelease) (Ubuntu 4.1.2-16ubuntu2)
> Copyright (C) 2006 Free Software Foundation, Inc.
> This is free software; see the source for copying conditions.  There
> is NO
> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR
> PURPOSE.
>
> Next, I created a directory - /usr/local/matlab/toolbox/getfem (my
> matlab installation is under /usr/local/matlab) and issued the
> following configure directive :
>
> ./configure --enable-dd --enable-qd --enable-qhull --enable-matlab
> --enable-python
> --with-matlab-toolbox-directory=/usr/local/matlab/toolbox/getfem
> --with-qd-lib-dir=/usr/lib
>
> Some messages from the configure process :
>
> ...
>
> checking for mex... mex
> checking for matlab path...  /usr/local/matlab
> checking for mex extension...  .mexglx
> Matlab release is : R14
>
> ...
>
> configure: creating ./config.status
> config.status : creating Makefile
> config.status: creating m4/Makefile
> config.status: creating cubature/Makefile
> config.status: creating superlu/Makefile
> config.status: creating src/Makefile
> config.status: creating tests/Makefile
> config.status: creating tests/meshes/Makefile
> config.status: creating tests-2.0/Makefile
> config.status: creating contrib/Makefile
> config.status: creating contrib/icare/Makefile
> config.status: creating contrib/delaminated_crack/Makefile
> config.status: creating contrib/static_friction/Makefile
> config.status: creating contrib/bimaterial_crack_test/Makefile
> config.status: creating contrib/bimat_contact_crack_test/Makefile
> config.status: creating contrib/mixed_elastostatic/Makefile
> config.status: creating contrib/xfem_contact/Makefile
> config.status: creating contrib/crack_plate/Makefile
> config.status: creating contrib/inter_element_test/Makefile
> config.status: creating contrib/aposteriori/Makefile
> config.status: creating bin/Makefile
> config.status: creating interface/Makefile
> config.status: creating interface/bin/Makefile
> config.status: creating interface/src/Makefile
> config.status : creating interface/src/matlab/Makefile
> config.status: creating interface/src/matlab/@gfCvStruct/Makefile
> config.status: creating interface/src/matlab/@gfEltm/Makefile
> config.status: creating interface/src/matlab/@gfFem /Makefile
> config.status: creating interface/src/matlab/@gfGeoTrans/Makefile
> config.status: creating interface/src/matlab/@gfInteg/Makefile
> config.status: creating interface/src/matlab/@gfMesh/Makefile
> config.status: creating interface/src/matlab/@gfMeshFem /Makefile
> config.status: creating interface/src/matlab/@gfSlice/Makefile
> config.status: creating interface/src/matlab/@gfMeshIm/Makefile
> config.status: creating interface/src/matlab/@gfSpmat/Makefile
> config.status : creating interface/src/matlab/@gfPrecond/Makefile
> config.status: creating interface/src/matlab/@gfMdBrick/Makefile
> config.status: creating interface/src/matlab/@gfMdState/Makefile
> config.status: creating interface/src/matlab/@gfLevelSet /Makefile
> config.status: creating interface/src/matlab/@gfMeshLevelSet/Makefile
> config.status: creating interface/src/matlab/private/Makefile
> config.status: creating interface/src/python/Makefile
> config.status : creating interface/src/python/setup.py
> config.status: creating interface/tests/Makefile
> config.status: creating interface/tests/meshes/Makefile
> config.status: creating interface/tests/matlab/Makefile
> config.status : creating interface/tests/matlab/private/Makefile
> config.status : creating interface/tests/python/Makefile
> config.status: creating getfem-config
> config.status: WARNING:  getfem-config.in seems to ignore the
> --datarootdir setting
> config.status: creating getfem-config-notinstalled
> config.status: creating gmm-config
> config.status: WARNING:  gmm-config.in seems to ignore the
> --datarootdir setting
> config.status: creating config.h
> config.status: executing src/getfem/getfem_arch_config.h commands
> config.status: creating src/getfem/getfem_arch_config.h - prefix
> GETFEM for config.h defines
> config.status: executing depfiles commands
>
> ------------------------------------------------------------------------------
> If you want to build the shared library of getfem++, use
> --enable-shared
> (by default, only the static one will be built)
> Using the QD library:
>    high precision (quad-double precision) polynomials and integration
> methods are enabled
> Using the qhull library for delaunay triangulations.
>
> Link options for the BLAS library: -lblas -lg2c
> You can give the location of your prefered blas library with either
> the --with-blas=<lib> option, or the BLAS_LIBS environment variable
> for example: ./configure BLAS_LIBS="-L/usr/lib/atlas/sse2 -lblas"
>
> ---------------------------------------
> Ready to build getfem++
>   building MATLAB interface: YES
>   building PYTHON interface: YES (requires numarray)
> ---------------------------------------
>
>
> I next use checkinstall to create a debian/ubuntu package (this just
> calls make and creates a convenient deb package).
>
>
> Compilation runs for a long time and then stops at :
>
> mex -output gf_matlab -g CDEBUGFLAGS="-O3" LD="g++" \
>         -I. -I./.. -DMATLAB_RELEASE=14 \
>         ./gfm_mex.c ./gfm_common.c ./../gfi_array.c \
>         ../.libs/libgetfemint.a ../../../src/.libs/libgetfem.a -lqhull
> -lblas -lg2c /usr/lib/libqd.a -lm  || (rm gf_matlab.mexglx; false)
> /bin/pwd: couldn't find directory entry in `..' with matching i-node
> /bin/pwd: couldn't find directory entry in `../../..' with matching
> i-node
>
>     Sorry! No options file was found for mex.
>            The mex script must be able to source
>            an options file to define compiler flags
>            and other settings.  This options file
>            is normally found in MATLAB/bin/mexopts.sh.
>            Please check to make sure that your installation
>            is complete and includes this file.
>
>            Here
>
>                MATLAB = /
>
>            Please contact:
>
>                MathWorks Technical Support
>
>            for further assistance.
>
> rm: cannot remove `gf_matlab.mexglx': No such file or directory
> make[4]: *** [gf_matlab.mexglx] Error 1
> make[4]: Leaving directory `/usr/local/src/science/getfem/getfem
> ++-3.0.3/interface/src/matlab'
> make[3]: *** [install-recursive] Error 1
> make[3]: Leaving directory `/usr/local/src/science/getfem/getfem++-
> 3.0.3/interface/src/matlab'
> make[2]: *** [install-recursive] Error 1
> make[2]: Leaving directory `/usr/local/src/science/getfem/getfem
> ++-3.0.3/interface/src'
> make[1]: *** [install-recursive] Error 1
> make[1]: Leaving directory `/usr/local/src/science/getfem/getfem++-
> 3.0.3/interface'
> make: *** [install-recursive] Error 1
>
> ****  Installation failed. Aborting package creation.
>
> Cleaning up...OK
>
> Bye.
>
> Its strange that the process above is using the matlab root directory
> as '/' above after discovering it (correctly) in the configure process
> to be /usr/local/matlab.
> The detected matlab release is also wrong.
>
> Could you please advise as to how these bugs are to be handled ?
>
> I am attaching the config.log file to this email for your convenience.
>
> With regards,
>
> MS
> _______________________________________________
> Getfem-users mailing list
> address@hidden
> https://mail.gna.org/listinfo/getfem-users



reply via email to

[Prev in Thread] Current Thread [Next in Thread]