FreeFem++-cs
an integrated environment for FreeFem++
Antoine Le Hyaric, Jacques-Louis Lions Laboratory, Pierre and Marie Curie University.

 

Download & Install

  1. Check the License

    2.  
  2. Download
     
    Platform Latest Version Contains
    Linux Ubuntu 10.04 64 LTS 12.4 FreeFem++ 3.19
    Linux Ubuntu 11.10 64 12.4 FreeFem++ 3.19
    Linux Ubuntu 12.04 64 LTS 12.4 FreeFem++ 3.19
    Linux Ubuntu 12.04 LTS 12.4 FreeFem++ 3.19
    MacOS 10.7 12.4 FreeFem++ 3.19
    Source 12.4 FreeFem++ 3.19
    Windows 12.4 FreeFem++ 3.19
    Windows 64 12.4 FreeFem++ 3.19

    • All available packages are listed in the packages directory. If you are interested in other binary packages, please leave a note below. The most popular requests will be added to the list.
    • Because of some compilation issues, MPI is not available with the MacOS 10.7 version of FreeFem++-cs yet.


     
  3. Install

    4.  
  4. Unstable Source Tree

    Many FreeFem++-cs users get back to this site with very valuable improvement ideas that they have already developed. This is a perfect case for distributed development, and Mercurial is a very useful tool in this situation. So I have published the Mercurial development repository of FreeFem++-cs. Every user who wants to publish his own improvements of the source code is welcome to "clone" this repository and publish his own modifications as another Mercurial copy (provided that it is in agreement with the license, of course).

    To download the main FreeFem++-cs source tree : 

    	hg clone http://www.ljll.math.upmc.fr/~lehyaric/ffcs/dev ffcs-dev

    All published Mercurial repositories will be mentioned on this page with a link and a description. And I will try to keep up with all external developments by pulling back the changes into the main repository ("hg pull") when everyone can profit from them.

    Thank you all for your valuable input!



Add your comments

Name :

Comments :


Please copy the following numbers (sorry, this is for computer-generated spam prevention) :





Previous comments (newest first)
Mark B. 14/05/12 21:58
It works! Thank you so much!!!
antoine 14/05/12 11:09
Hi Mark,

Version 12.4 for 32-bit Ubuntu is now available for download. Can you try it and let me know if it works for you?

Antoine.
Mark B. 12/05/12 03:16
No problem, I really appreciate your help!
antoine 11/05/12 09:37
Hi Mark,

The 32-bit version is not ready yet. I will only have access to a 32-bit Ubuntu system on monday. I will let you know as soon as this version is available for download.

Antoine.
Mark B. 11/05/12 01:08
Hey Antoine,

Where can I download the compiled 32 bit version?

Thanks!!
antoine 10/05/12 21:29
Hi Mark,

32-bit software can often run on 64-bit systems, but 64-bit software imperatively needs a 64-bit system. I will build a 32-bit version of FFCS. It should work fine.

Regards,

Antoine.
Mark B. 10/05/12 20:53
I tried 12.4 with both Ubuntu 11.10 and the upgraded Ubuntu 12.04 LTS with no luck. I no longer get the "encoding error" file, but no window opens up. Does it matter if I'm running a 32 bit system instead of 64 bit?

Thanks!
antoine 10/05/12 15:14
Hi Mark,

Could you try and download version 12.4 for Ubuntu 11.10 and let me know if it works?

Thank you in advance,

A.
antoine 10/05/12 10:33
Hi Mark,

Ubuntu updates sometims produce that sort of problems. Let me try and reproduce that bug locally and I will get back to you.
Mark B. 10/05/12 10:25
I am running Ubuntu 11.10 and recently installed some updates and now my FreeFem cs is no longer working. I get a strange file in the directory that I\'m executing from that says something about an encoding error.

I\'ve tried multiple compilations including older versions without any luck. I have also tried compiling it from the source, but I wasn\'t able to due to errors.

Is there anything I can do to fix the problem?
Ivan 21/04/12 02:36
it looks like i found it in 'other packages' Thanks a lot for developing this, it's very nice!
Ivan 21/04/12 02:28
Where is ubuntu 32 bit install? I can't run the 64 version, it says 'command not found' and some weird file appears in that folder with incorrect encoding..
Fred 03/02/12 23:26
I was able to compile 12.1 on both my server and a netbook with some "adjustments" I am re compiling now so that I can get all of the make errors and post them here if that is OK. I ran an example (Lapalce-Adapt-aniso-3d.edp) and got the following error message in the ffcs console window:


-- FreeFem++ v 3.180000 (date Fri Feb 3 12:46:57 CST 2012)
Load: lg_fem lg_mesh lg_mesh3 eigenvalue
1 :
2 : load "msh3" (load: dlopen /opt/ffcs/FreeFem++-cs 12.1/Contents/Linux/msh3.so 0x8a1b740) load: msh3

3 : load "tetgen" (load: dlopen /opt/ffcs/FreeFem++-cs 12.1/Contents/Linux/tetgen.so 0x8a1c4f0) load: tetgen

4 : load "mshmet" (load: dlopen /opt/ffcs/FreeFem++-cs 12.1/Contents/Linux/mshmet.so 0x8a1cf40) load: mshmet

5 : load "mmg3d-v4.0"
load error : mmg3d-v4.0
fail :
list prefix: '/opt/ffcs/FreeFem++-cs 12.1/Contents/Linux/' '' list suffix : '' , '.so'

Error line number 5, in file /home/fred/.ffcs26491.edp, before token mmg3d-v4.0
Error load
current line = 5
Compile error : Error load
line number :5, mmg3d-v4.0
error Compile error : Error load
line number :5, mmg3d-v4.0
code = 1 mpirank: 0
FreeFem++ returned error 1
fred 03/02/12 14:59
I noticed this with superlu when building freefem++-3.18 as well. I will try to keep track as much as I can. I let this machine "crunch away" and check it periodically.

Thanks Again!

Cheers!

Fred
antoine 03/02/12 13:38
Hi fred,

Many thanks to you for keeping up. Please do not expect too much from the new source. I am running quite a lot behind schedule, so my first move was to include the latest FF, with the hope that some of the current bugs would be solved automatically. But there are new ones as well. For instance I suspect that superlu only builds properly when calling make _twice_. Please let me know if you notice anything else of the same kind.

Cheers,

Antoine.
fred 03/02/12 12:50
Thanks for the source update. I will try it today and let you know how it goes.

Cheers!

Fred
antoine 24/01/12 11:36
Hi fred, Hi Pauline,

Sorry for the delay. Many things are happening at the same time here. I don't know what is causing this bug at the moment, but I will get back to you as soon as I have a solution.

Best regards,

Antoine.
Pauline 22/01/12 17:26
Hello,

I have the same error that Fred in the previous comment. Did you arrive to solve this problem or does anyone have the solution?

Thank you in advance,

Pauline
fred 18/01/12 12:46
When trying to build 11.14 I get the following error:
./ff-c++ -auto freeyams.cpp
ERROR in auto dependance seach. sorry : freeyams.cpp -I/home/fred/src/ffcs-11.14/ff/upstream/download/include/libMesh ERROR-missing-lib:freeyams
make[3]: *** [freeyams.so] Error 1
make[3]: Leaving directory `/home/fred/src/ffcs-11.14/ff/upstream/examples++-load'
make[2]: *** [load_compile] Error 2
make[2]: Leaving directory `/home/fred/src/ffcs-11.14/ff/upstream/examples++-load'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/fred/src/ffcs-11.14/ff/upstream'
make: *** [all] Error 2


Also, I had to comment out the patch for mmg3d in the Makefile. I am using the following configuration options:

./configure --prefix=/opt/ffcs --with-mpi-prefix=/usr/include --with-freefem-options="--enable-download --enable-optim --enable-opengl --with-flib=/usr/lib/libgfortran.so --with-mpi=yes"

Thank you again,

Fred
antoine 07/12/11 10:01
Hi fred,

Here is the source. Sorry for the delay and thank you very much for your interest in FF and FFCS.

Best regards,

Antoine.
fred 05/12/11 14:10
Hello,

Will the source version of 11.14 be available?

Thank You for all of your efforts!

Fred
anonymous 18/11/11 19:37
thank you very much for sharing
anonymous 06/07/11 21:18
The configuration script does not seem to generate a required vtk make file (for the included vtk).

I am building from source on a debian 64 bit sid/testing box with gcc-4.6.

Could I tell it to use the (OS's) vtk in /usr/lib ?

V/R

Scott

make[4]: Leaving directory `/home/shared/usr/src/ffcs-11/ffcs-11.10/vtk/cmake-2.8.4'
make[3]: Leaving directory `/home/shared/usr/src/ffcs-11/ffcs-11.10/vtk/cmake-2.8.4'
+ cd ..
++ pwd
+ cmakeprog=/home/shared/usr/src/ffcs-11/ffcs-11.10/vtk/cmake-2.8.4/bin/cmake
+ vtk_tgz=vtk-5.6.1.tar.gz
+ vtk_dir=VTK
+ test '!' -f vtk-5.6.1.tar.gz
+ test '!' -d VTK
++ pwd
+ alhcwd=/home/shared/usr/src/ffcs-11/ffcs-11.10/vtk
+ cd VTK
+ test '!' -f CMakeCache.txt
+ make -j 4
make[3]: Entering directory `/home/shared/usr/src/ffcs-11/ffcs-11.10/vtk/VTK'
make[3]: warning: -jN forced in submake: disabling jobserver mode.
make[3]: *** No targets specified and no makefile found. Stop.
make[3]: Leaving directory `/home/shared/usr/src/ffcs-11/ffcs-11.10/vtk/VTK'
make[2]: *** [vtkbuild] Error 2
make[2]: Leaving directory `/home/shared/usr/src/ffcs-11/ffcs-11.10/vtk'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/shared/usr/src/ffcs-11/ffcs-11.10'
make: *** [all] Error 2
anonymous 27/06/11 15:22
thanks~
fred 16/04/11 21:10
Antoine,

Thank you for the response. I will continue to test this version on Slackware. Migrating back to g++4.5 on Archppc may be difficult. Slackware-13.x uses gcc/g++ <= 4.5 (4.3.3 on 13.0 and 4.4.4 on 13.1)

Take Care,

Fred
antoine 15/04/11 12:03
Hi Fred,

It's nice to hear from you. FFCS configuration has been modified lately to take debian packaging into account. This is probably the reason why I need to check the other platforms again. I also noticed that you are using g++ 4.6. Quite a number of components of FFCS still break with g++ 4.6, so if you can use CXX=g++-4.5 with configure, that would be best for the moment (I won't be able to change this before a couple of weeks).

Thank you very much for the bug reports.

Best regards,

Antoine.
fred 13/04/11 23:45
Here are the errors from build attempt on Archppc, iBook G4 for ffcs-11.9:

[ 17%] Building CXX object Utilities/vtkmetaio/CMakeFiles/vtkmetaio.dir/metaUtils.cxx.o
cd /usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio && /usr/bin/c++ -DVTK_IN_VTK -Wno-deprecated -O3 -DNDEBUG -I/usr/src/build/ffcs-11.9/vtk/VTK -I/usr/src/build/ffcs-11.9/vtk/VTK/Common -I/usr/src/build/ffcs-11.9/vtk/VTK/Utilities -I/usr/src/build/ffcs-11.9/vtk/VTK/VolumeRendering -I/usr/src/build/ffcs-11.9/vtk/VTK/Rendering -I/usr/src/build/ffcs-11.9/vtk/VTK/Charts -I/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkalglib -I/usr/src/build/ffcs-11.9/vtk/VTK/Infovis -I/usr/src/build/ffcs-11.9/vtk/VTK/Geovis -I/usr/src/build/ffcs-11.9/vtk/VTK/Views -I/usr/src/build/ffcs-11.9/vtk/VTK/Hybrid -I/usr/src/build/ffcs-11.9/vtk/VTK/Widgets -I/usr/src/build/ffcs-11.9/vtk/VTK/Rendering/Testing/Cxx -I/usr/src/build/ffcs-11.9/vtk/VTK/IO -I/usr/src/build/ffcs-11.9/vtk/VTK/Imaging -I/usr/src/build/ffcs-11.9/vtk/VTK/Graphics -I/usr/src/build/ffcs-11.9/vtk/VTK/GenericFiltering -I/usr/src/build/ffcs-11.9/vtk/VTK/Filtering -I/usr/src/build/ffcs-11.9/vtk/VTK/Common/Testing/Cxx -I/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtklibproj4 -I/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/DICOMParser -I/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkfreetype/include -I/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtknetcdf -I/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkexodus2/include -I/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/MaterialLibrary -I/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/verdict -I/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/utf8/source -I/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/ftgl/src -I/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/ftgl -I/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio -o CMakeFiles/vtkmetaio.dir/metaUtils.cxx.o -c /usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx: In function 'bool vtkmetaio::MET_StringToWordArray(const char*, int*, char***)':
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:805:3: error: 'ptrdiff_t' was not declared in this scope
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:805:3: note: suggested alternatives:
/usr/lib/gcc/powerpc-unknown-linux-gnu/4.6.0/../../../../include/c++/4.6.0/powerpc-unknown-linux-gnu/bits/c++config.h:156:28: note: 'std::ptrdiff_t'
/usr/lib/gcc/powerpc-unknown-linux-gnu/4.6.0/../../../../include/c++/4.6.0/powerpc-unknown-linux-gnu/bits/c++config.h:156:28: note: 'std::ptrdiff_t'
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:805:13: error: expected ';' before 'l'
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:807:13: error: expected ';' before 'p'
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:808:9: error: 'p' was not declared in this scope
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:808:11: error: 'l' was not declared in this scope
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:814:13: error: expected ';' before 'pp'
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:816:9: error: 'pp' was not declared in this scope
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:816:12: error: 'l' was not declared in this scope
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:829:3: error: 'pp' was not declared in this scope
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:829:6: error: 'l' was not declared in this scope
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:845:13: error: expected ';' before 'i'
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:846:7: error: 'i' was not declared in this scope
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:848:8: error: 'p' was not declared in this scope
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:854:11: error: 'p' was not declared in this scope
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:858:5: error: 'j' was not declared in this scope
/usr/src/build/ffcs-11.9/vtk/VTK/Utilities/vtkmetaio/metaUtils.cxx:859:11: error: 'p' was not declared in this scope
make[5]: *** [Utilities/vtkmetaio/CMakeFiles/vtkmetaio.dir/metaUtils.cxx.o] Error 1
make[5]: Leaving directory `/usr/src/build/ffcs-11.9/vtk/VTK'
make[4]: *** [Utilities/vtkmetaio/CMakeFiles/vtkmetaio.dir/all] Error 2
make[4]: Leaving directory `/usr/src/build/ffcs-11.9/vtk/VTK'
make[3]: *** [all] Error 2
make[3]: Leaving directory `/usr/src/build/ffcs-11.9/vtk/VTK'
make[2]: *** [vtkbuild] Error 2
make[2]: Leaving directory `/usr/src/build/ffcs-11.9/vtk'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/usr/src/build/ffcs-11.9'
make: *** [all] Error 2

Configure options are as follows:
./configure --prefix=/opt/ffcs --localstatedir=/var --sysconfdir=/etc --with-freefem-options="--enable-download --enable-optim --enable-opengl --with-flib=/usr/lib/libgfortran.so.3.0.0 --with-mpi=yes"
fred 12/04/11 16:03
Hi,

I get the following error when running the configure script for 11.9 on both Slackware and Archppc Linux:

checking if MPI is specified explicitely... no
checking for 32-bit MPICH on 64-bit Windows... no
checking for mpicxx... yes
./configure: line 6082: test: too many arguments
checking for X... libraries , headers
checking for gethostbyname... yes
checking for connect... yes
checking for remove... yes
checking for shmat... yes
checking for IceConnectionNumber in -lICE... yes
checking for XkbGetKeyboard in -lX11... yes
checking for XextFindDisplay in -lXext... yes
checking for XtAppLock in -lXt... yes
checking for XpmCreateXpmImageFromImage in -lXpm... yes
checking for version name attributes... ""
checking for final directory name... "FreeFem++-cs 11.9"
configure: creating ./config.status
config.status: creating Makefile
config.status: creating src/Makefile
config.status: creating src/mpi/Makefile
config.status: creating tests/Makefile
config.status: creating doc/Makefile
config.status: creating pack/Makefile
config.status: creating ff/Makefile
config.status: creating keywords/Makefile
config.status: creating fltk/Makefile
config.status: creating rxspencer/Makefile
config.status: creating vtk/Makefile
config.status: creating mingw/Makefile
config.status: creating vtkFlRWI/Makefile
config.status: creating config.h
config.status: executing depfiles commands

My configure options are:
./configure --prefix=/opt/ffcs --localstatedir=/var --sysconfdir=/etc --with-freefem-options="--enable-download --enable-optim --enable-opengl --with-flib=/usr/lib/libgfortran.so.3.0.0 --with-mpi=yes"

Which works for 11.7 and previous.

The build is successful on Slackware but msh3 is missing somehow when running the examples. Build fails on Archppc when compiling vtk. I don't have the error messages copied from my iBook yet but will get them for you soon.

Thanks again & Take Care,

Fred

Mikhail Artemiev 11/04/11 09:42
Sorry.
It's my fault!
Now I understood!
Of course,
"int asd = 1;
border c1(t=0,1){x=t;y=0;label=asd;}"

Thanks
Mikhail Artemiev 11/04/11 09:33
Good morning, Antoine!
I use FreeFem++-cs 11.4 with Windows XP.
My problem is:
when I write
"border c1(t=0,1){x=t;y=0;}" all is OK;
but when I write
"border c1(t=0,1){x=t;y=0;label=asd;}" compiler sais that
" 1 : border c1(t=0,1){x=t;y=0;label=asd The Identifier asd does not exist "
and highlights the word "asd" by red.
What is the mistake?

Thanks
C. Trophime 01/04/11 22:57
My problem with the different rendering in FF come from me.
I switch from medit to plot which means I use ffglut instead of ffmedit.
So I think there is no trouble :)

Anyway I will run the tests in FF
Should I run them as "root" just like in FFCS?


C.
antoine 01/04/11 18:40:12
> Now ffcs 11.7 compiles fine with the latest mysql
> So we can have both freefem++ and ffcs in Debian Testing.

I was also trying to fix the same issue. Well done.

> I have also tested freefem++ and ffcs (with vtk included). Both codes are ok.
> There is only a slight problem. It seems that freefem++ no longer use the
> same default windows to display the results. The window used to be with black
> as backgroundcolor now it's white... Can ffcs alter the freefem++ medit
> default options?

No, FFCS does not interact with Medit directly. It happens probably somewhere deep inside FF. Please let me know if you want me to investigate.

> To package ffcs for Debian I think we still have to split the package...

I am testing this right now...

> I have also some troubles detecting parms, superlu-dist in freefem++ and a
> problem with freeyams (cannot find automatically libMesh)

If these are the only problems left, it means that many FF tests work fine already.

I'll keep you posted as soon as --disable-client/server works again.

A.
C. Trophime 01/04/11 17:11:06
Now ffcs 11.7 compiles fine with the latest mysql
So we can have both freefem++ and ffcs in Debian Testing.

I have also tested freefem++ and ffcs (with vtk included). Both codes are ok.
There is only a slight problem. It seems that freefem++ no longer use the same default windows to display the results. The window used to be with black as backgroundcolor now it's white... Can ffcs alter the freefem++ medit default options?

To package ffcs for Debian I think we still have to split the package...
I have also some troubles detecting parms, superlu-dist in freefem++ and a problem with freeyams (cannot find automatically libMesh)

Best
C.
C. Trophime 01/04/11 16:32:19
Hi,
I am trying to use vtk libraries from Debian testing but I end with a problem
which is similar to :

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=617240

It seems to be due to mysql.
I am trying to check that using myql from Debian Unstable solve the problem
But it takes ages to rebuild the mysql package
Otherwise I'll have to wait for the bug to be closed :(
antoine 01/04/11 07:59:16
Hi,

I am not sure why this happens. Could you do the following test: remove the file src/ffrgraph.cpp, run make again and copy the final lines of the output back to this page?
ujack 01/04/11 07:22:38
i am getting the following error when compiling. any help

mv -f .deps/ffs-ffs.Tpo .deps/ffs-ffs.Po
g++ -DFFS -I ../ff/upstream/src/Graphics -I ../ff/upstream -I ../ff/upstream/src/fflib -O3 -DNDEBUG -O3 -DDRAWING -DBAMG_LONG_LONG -DNCHECKPTR -rdynamic -fPIC -o ffs ffs-ffs.o ffs-ffrgraph.o ffs-ffcsstream.o ffs-parallelempi-empty.o ffs-mymain.o libffs.a ../ff/upstream/src/lglib/liblg.a ../ff/upstream/src/fflib/libff.a /home/jurombo/Downloads/ffcs-11.7/ff/upstream/download/lib/libumfpack.a /home/jurombo/Downloads/ffcs-11.7/ff/upstream/download/lib/libamd.a -L/home/jurombo/Downloads/ffcs-11.7/ff/upstream/download/lib -larpack -llapack -L/home/jurombo/Downloads/ffcs-11.7/ff/upstream/download/blas -lcblas -lf77blas -lXxf86vm -lXext -lX11 -lXpm -lm -ldl /usr/lib/gcc/i586-manbo-linux-gnu/4.4.3/libgfortran.so -lXpm -lXt -lXext -lX11 -lpthread -lSM -lICE -lm
ffs-ffrgraph.o: In function `main':
ffrgraph.cpp:(.text+0x1050): multiple definition of `main'
ffs-ffs.o:ffs.cpp:(.text+0x9b0): first defined here
collect2: ld returned 1 exit status
make[4]: *** [ffs] Error 1
make[4]: Leaving directory `/home/jurombo/Downloads/ffcs-11.7/src'
make[3]: *** [all-recursive] Error 1
make[3]: Leaving directory `/home/jurombo/Downloads/ffcs-11.7/src'
make[2]: *** [all] Error 2
make[2]: Leaving directory `/home/jurombo/Downloads/ffcs-11.7/src'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/jurombo/Downloads/ffcs-11.7'
make: *** [all] Error 2
[root@juMobile ffcs-11.7]#
antoine 31/03/11 15:30
Hi,

Thank you for letting me know about the mail. I don't know how long the problem will last. Fortunately, the comment forms work fine. I'll keep you posted as soon as I have a solution for the --disable-client/server problem.

Regards,

Antoine.
C. Trophime 31/03/11 13:56
I have some trouble with --disable-client and --disable-server:

* --disable-client : fail when trying to \"wget\" vtk with an unknown version number
* --disable-server : fail to install ffmedit...

Best C.

PS: it seems that your email is \"down\"

ujack 24/03/11 14:09
I am getting the following error when compiling.

configure: C++ preprocessor "/lib/cpp" fails security check

I am doing this on a Mandriva 2010.2 platform

any quick fix??
fred 08/03/11 15:17
Just finished compiling and running on ArchLinux ppc 32 bit (iBook G4). No problems with build or executables.

Thanks Again!

Take Care,

Fred
antoine 07/03/11 16:58
Hi Fred,

Thank you very much for your feedback.

Cheers,

A.
fred 07/03/11 15:28
11.6 compiles and runs on Slackware-13.0 32 bits. I am running examples right now and most everything seems to be OK.

Thanks Again!

Fred
fred 22/02/11 02:58
Antoine,

I removed the installation directory (/opt/ffcs) and tried to re-install and it succeeded using python2.7 instead of python3. I think I had an older version of ffcs installed and didn't uninstall it first. Everything seems to work for now. I will post if I find anything else on Archlinux or Slackware.

Thanks & Take Care,

Fred
fred 21/02/11 16:52
Antoine,

Sorry for the duplicate post! I accidentally hit the reload button on my browser. I am using XFS as a filesystem both on Arch and Slackware.

Take Care,

Fred
fred 21/02/11 16:50
Antoine,

I made the changes that you suggested. I ran across the following issues after a sucessful compile from the tarball distribution.

1) For Arch Linux, /usr/bin/python points to python3 so install fails with this error:

make[1]: Entering directory `/home/fred/src/ffcs/ffcs-11.4/pack'
rsync -avq --delete work/ "FreeFem++-cs 11.4"
cd "FreeFem++-cs 11.4" && ./install_as_root /opt/ffcs /opt/ffcs/bin
++ python -c 'import os,sys;print os.path.realpath(sys.argv[1])' ./install_as_root
File "<string>", line 1
import os,sys;print os.path.realpath(sys.argv[1])
^
SyntaxError: invalid syntax
+ canonlink=
make[1]: *** [install] Error 1
make[1]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/pack'
make: *** [install-recursive] Error 1

For python2 on Arch, /usr/bin/python2.7 must be used. I am not sure how to configure for this though.

2) After correcting python problem temporarialy, and trying to install with corrected install.m4 I still get the dangling symbolic link problem. I don't on Slackware though.

Anyway, hope this helps to stomp out some bugs.

Take Care,

Fred
antoine 21/02/11 16:04
Hi Fred,

Thank you very much for all the bug reports. Python should not even get called on Linux (it's used as a MacOS-only replacement for "readlink -f") so I really need to fix this.

About the symlink problem, does your Arch Linux setup use HFS as a filesystem or ext[234]?

Cheers,

A.
fred 21/02/11 15:29
Antoine,

I made the changes that you suggested. I ran across the following issues after a sucessful compile from the tarball distribution.

1) For Arch Linux, /usr/bin/python points to python3 so install fails with this error:

make[1]: Entering directory `/home/fred/src/ffcs/ffcs-11.4/pack'
rsync -avq --delete work/ "FreeFem++-cs 11.4"
cd "FreeFem++-cs 11.4" && ./install_as_root /opt/ffcs /opt/ffcs/bin
++ python -c 'import os,sys;print os.path.realpath(sys.argv[1])' ./install_as_root
File "<string>", line 1
import os,sys;print os.path.realpath(sys.argv[1])
^
SyntaxError: invalid syntax
+ canonlink=
make[1]: *** [install] Error 1
make[1]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/pack'
make: *** [install-recursive] Error 1

For python2 on Arch, /usr/bin/python2.7 must be used. I am not sure how to configure for this though.

2) After correcting python problem temporarialy, and trying to install with corrected install.m4 I still get the dangling symbolic link problem. I don't on Slackware though.

Anyway, hope this helps to stomp out some bugs.

Take Care,

Fred
antoine 20/02/11 09:32
Hi Fred,

You have found the root cause of the problem. The compiler is not able to generate correct code when specifying "-mcpu=powerpc -maltivec -mcpu=common" in that order. To remove -mcpu=common everywhere, just comment out the following lines in ff/upstream/acoptim.m4:

# Generic code
if test "$enable_debug" != yes \
-a "$enable_optim" != no \
-a "$enable_generic" = yes
then
CHECK_COMPILE_FLAG(C,-mcpu=common,CFLAGS)
CHECK_COMPILE_FLAG(C++,-mcpu=common,CXXFLAGS)
CHECK_COMPILE_FLAG(Fortran 77,-mcpu=common,FFLAGS)
fi

then remove config.status in the same directory and run "autoreconf". "./configure ..." from the main directory should be ok after that. I will do the same in my own code.

Many thanks for your help.

Cheers,

Antoine.
fred 20/02/11 02:05
Antoine,

Scotch compiles if I remove this "-DNDEBUG -O3 -mcpu=common" from the Makefile in the scotch directory. I have this set to "-DNDEBUG -O3 -mcpu=powerpc -maltivec" for my arcitechure. Build then fails with this error:

/home/fred/src/ffcs/ffcs-11.4/ff/upstream/download/lib/libarpack.a(ivout.o): In function `ivout_':
ivout.f:(.text+0x23c): undefined reference to `__mulh'
ivout.f:(.text+0x3e4): undefined reference to `__mulh'
ivout.f:(.text+0x534): undefined reference to `__mulh'
ivout.f:(.text+0x654): undefined reference to `__mulh'
ivout.f:(.text+0x76c): undefined reference to `__mulh'
/home/fred/src/ffcs/ffcs-11.4/ff/upstream/download/lib/libarpack.a(ivout.o):ivout.f:(.text+0x87c): more undefined references to `__mulh' follow
collect2: ld returned 1 exit status
make[6]: *** [FreeFem++-x11] Error 1
make[6]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff/upstream/src/x11'
make[5]: *** [all-recursive] Error 1
make[5]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff/upstream/src'
make[4]: *** [all-recursive] Error 1
make[4]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff/upstream'
make[3]: *** [all] Error 2
make[3]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff/upstream'
make[2]: *** [compile] Error 2
make[2]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4'
make: *** [all] Error 2


Thanks & Take Care,

Fred
fred 19/02/11 21:34
Antoine,

Thanks for the reply. I corrected the install.m4 file and executed autoreconf and configure on Slackware. Then tried to make install and it succeded without errors. I will try the Scotch fix this evening and let you know how it went on my laptop. Perhaps I should not try to have such esoteric hardware but I really like kde4 on this iBook :)

Thanks Again & Take Care,

Fred
antoine 19/02/11 17:55
Hi fred,

Thank you for all your reports.

About the Scotch problem: by default, FF compilation does not stop when one of the downloads does not compile. I have changed this recently in FFCS to make sure that all the downloads work on the regular platforms. But we cannot be sure that all the downloads support all our platforms. I need to add a configuration option for this. Would it be ok for you if you could disable scotch on request?

- If you want to disable Scotch right now, remove it from the line "MPI_SOFT=..." in ff/upstream/download/Makefile. Also remove all references to pastix from ff/upstream/examples++-mpi/Makefile.

- If you want to keep Scotch, it will probably be more work. Please let me know.

About the FreeFem++-ffg problem: this looks like a dangling symbolic link in the install script. Please replace the line "if test -e "$2"" by "if test -e "$2" || test -L "$2"" in pack/install.m4. I have corrected it in my development version as well. I will publish it as soon as I can (probably on monday).

Cheers,

Antoine.
fred 19/02/11 17:05
The following errors occur during a make install on Slackware 13.0 32 bits:

+ test -e /opt/ffcs/bin/FreeFem++-ffg
+ ln -s '/opt/ffcs/FreeFem++-cs 11.4/FreeFem++' /opt/ffcs/bin/FreeFem++-ffg
ln: creating symbolic link `/opt/ffcs/bin/FreeFem++-ffg': File exists
make[1]: *** [install] Error 1
make[1]: Leaving directory `/home/fred/src/ffcs-11.4/pack'
make: *** [install-recursive] Error 1

Thanks,

Fred

fred 19/02/11 12:51
The following errors occur on ARCH Linux PPC (iBook G4 CFLAGS='-mcpu=powerpc -maltivec', CXXFLAGS='-mcpu=powerpc -maltivec'):

make[10]: Entering directory `/home/fred/src/ffcs/ffcs-11.4/ff/upstream/download/scotch/scotch_5.1_esmumps/src/scotch'
gcc -mcpu=powerpc -maltivec -DNDEBUG -O3 -DNDEBUG -O3 -mcpu=common -rdynamic -fPIC -DLONG -Drestrict=__restrict -DCOMMON_PTHREAD -DCOMMON_FILE_COMPRESS_GZ -DCOMMON_RANDOM_FIXED_SEED -DCOMMON_TIMING_OLD -DSCOTCH_RENAME -DSCOTCH_DETERMINISTIC -I../../include -I../libscotch -DSCOTCH_VERSION=\""5.1"\" acpl.c -o acpl -L../../lib -lscotch -lscotcherrexit -lz -lm -lpthread
../../lib/libscotch.a(arch_mesh.o): In function `_SCOTCHarchMesh2DomTerm':
arch_mesh.c:(.text+0x170): undefined reference to `__divss'
arch_mesh.c:(.text+0x180): undefined reference to `__quoss'
../../lib/libscotch.a(arch_mesh.o): In function `_SCOTCHarchMesh2DomBipartU':
arch_mesh.c:(.text+0x5b4): undefined reference to `__mulh'
arch_mesh.c:(.text+0x600): undefined reference to `__mulh'
../../lib/libscotch.a(arch_mesh.o): In function `_SCOTCHarchMesh3DomTerm':
arch_mesh.c:(.text+0x804): undefined reference to `__divss'
arch_mesh.c:(.text+0x814): undefined reference to `__quoss'
arch_mesh.c:(.text+0x81c): undefined reference to `__divss'
arch_mesh.c:(.text+0x82c): undefined reference to `__quoss'
../../lib/libscotch.a(arch_torus.o): In function `_SCOTCHarchTorus2DomTerm':
arch_torus.c:(.text+0x170): undefined reference to `__divss'
arch_torus.c:(.text+0x180): undefined reference to `__quoss'
../../lib/libscotch.a(arch_torus.o): In function `_SCOTCHarchTorus3DomTerm':
arch_torus.c:(.text+0x62c): undefined reference to `__divss'
arch_torus.c:(.text+0x63c): undefined reference to `__quoss'
arch_torus.c:(.text+0x644): undefined reference to `__divss'
arch_torus.c:(.text+0x654): undefined reference to `__quoss'
../../lib/libscotch.a(common_integer.o): In function `_SCOTCHintPerm':
common2_integer.c:(.text+0x31c): undefined reference to `__divus'
../../lib/libscotch.a(common_integer.o): In function `_SCOTCHintSort3asc1':
common2_integer.c:(.text+0xacc): undefined reference to `__mulh'
collect2: ld returned 1 exit status
make[10]: *** [acpl] Error 1
make[10]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff/upstream/download/scotch/scotch_5.1_esmumps/src/scotch'
make[9]: *** [scotch] Error 2
make[9]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff/upstream/download/scotch/scotch_5.1_esmumps/src/scotch'
make[8]: *** [scotch] Error 2
make[8]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff/upstream/download/scotch/scotch_5.1_esmumps/src'
make[7]: *** [scotch_5.1_esmumps/FAIT] Error 2
make[7]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff/upstream/download/scotch'
make[6]: *** [compile-dir] Error 2
make[6]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff/upstream/download'
make[5]: *** [tag-compile-pkg] Error 1
make[5]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff/upstream/download'
make[4]: *** [all-recursive] Error 1
make[4]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff/upstream/download'
make[3]: *** [all-recursive] Error 1
make[3]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff/upstream'
make[2]: *** [all] Error 2
make[2]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff/upstream'
make[1]: *** [compile] Error 2
make[1]: Leaving directory `/home/fred/src/ffcs/ffcs-11.4/ff'
make: *** [install-recursive] Error 1

configure options are:
./configure --prefix=/opt/ffcs --prefix=/opt/freefem++ --localstatedir=/var --sysconfdir=/etc --with-freefem-options="--enable-download --enable-optim --enable-opengl --with-flib=/usr/lib/libgfortran.so --with-mpi=yes"

Thanks

Fred
antoine 24/01/11 11:05
@Joel (11/01/11 23:21):

I have just checked Version 11.2 for 64-bit Ubuntu 10.04 on 64-bit Ubuntu 10.10. It runs fine for me (the title graphics do not pop back up). Could you try it and let me know if it's ok for you as well?
antoine 17/01/11 15:35
@Eliseo Chacón Vera:

Sorry, the MacOS version of FFCS is not up-to-date compared to the other platforms. I will get back to you as soon as I have new version.
Eliseo Chacón Vera 13/01/11 16:09
Dear Antoine,

when I install the soft on MAC OS X 10.6.6 I get the FreeFem++-cs 10.18.app

When I double click on it I get the name FreeFem++-cs 10.18.app on the left side of the upper bar
but no window to work with.

Do you have an idea of what the problem might be?

Cordially yours,

Eliseo
Joel 12/01/11 14:51
@antoine,

I will see if I can recreate the VTK error for you. I will hold off on MPI for now (shouldn't need it anyway). As far as the title graphic, I had that both with the precompiled and self-compiled versions. In fact it was this that had me try to compile it myself. One thing I didn't try was turning off compiz which has been known to do silly things like this. I will see if that helps as well. Thanks again for this great tool!
antoine 12/01/11 11:04
@Joel:

Thank you very much for our feedback. No, you have done nothing wrong. There are issues with MPI that I hope to solve in the coming version (11.1). I would be interested in the error message for VTK, if you still have it. About the title graphic, it's the first time I hear about this problem, I don't know what causes it. Let me try it on the same system as yours and I will get back to you.
Joel 11/01/11 23:21
Also the title graphic that covers the space where the generated results will be shown keeps popping back over the results after the calculation is run. Is there something that would prevent this?
Joel 11/01/11 23:17
Hello, I have successfully built the new version on my Ubuntu 10.10 64bit but I had a couple issues. I had to download vtk*.tar.gz and put it in the directory as it couldn't find my system install and failed to download it. This may have been a one time glitch.

Also it had issues finding mpi.h, so I had to compile without mpi. Is there something to this or have I done something wrong here? (I have openmpi etc.)
antoine 07/12/10 11:14
@Shyam Sunder:

It took me some time to realise this, but this was a simple bug in the install step above. One should type "make package" and not "make pack". I have now corrected the documentation. Sorry for the trouble.
fred 04/11/10 02:31
10.16 compiles and runs for Slackware 13.0 32 bit. Use configure options given in earlier post.

Thanks Again!

Fred
Shyam Sunder 26/10/10 11:23
Hi,
I had downloaded ffcs-10.16 today. I am using a Linux Ubuntu 10.04 LTS machine.
After extracting the files, I gave ./configure, which was completed successfully.
Then I gave the command make. After that I decided to give make check. It showed that 2 out of 4 tests failed. When I give make pack, it says, \\\'Nothing to be done for \\\'pack\\\'\\\'.

I feel that I have not installed it successfully, cause when I run a freefem++-cs program (which runs successfully in my Windows laptop), it fails to complete.
It says
err code 5 , mpirank 0
ffc: error: FreeFem++ returned error 5

I would be grateful if you could help me.

Thanks in advance.
anonymous 26/10/10 11:22
Hi,
I had downloaded ffcs-10.16 today. I am using a Linux Ubuntu 10.04 LTS machine.
After extracting the files, I gave ./configure, which was completed successfully.
Then I gave the command make. After that I decided to give make check. It showed that 2 out of 4 tests failed. When I give make pack, it says, \'Nothing to be done for \'pack\'\'.

I feel that I have not installed it successfully, cause when I run a freefem++-cs program (which runs successfully in my Windows laptop), it fails to complete.
It says
err code 5 , mpirank 0
ffc: error: FreeFem++ returned error 5

I would be grateful if you could help me.

Thanks in advance.
antoine 16/10/10 22:36
> 64-bit Windows 7?

Configuring FreeFem++-cs and its dependencies for 64-bit Windows with Cygwin and MinGW is not straightforward. I am working on it now (this is the reason why FFCS for Windows is not at version 10.15 yet), but I also have other tasks to do concurrently so it may still take some time. As soon as it is ready, I will place it on this page and update the RSS feed. Please keep watching!
anonymous 16/10/10 11:46
64-bit Windows 7?
antoine 01/10/10 14:26
Hi Fred,

Thank you very much for the detailed solution.

Cheers,

Antoine.
fred 01/10/10 14:09
Oops!

For 32 bit systems:

./configure --prefix=/opt/ffcs --localstatedir=/var --sysconfdir=/etc --with-freefem-options="--enable-download --enable-optim --enable-opengl --with-flib=/usr/lib/libgfortran.so.3.0.0 --with-mpi=yes"
fred 01/10/10 13:28
It works on both Slackware 13.1 32 bits and 13.0 64 bits. Many thanks! The configure options were:

./configure --prefix=/opt/ffcs --localstatedir=/var --sysconfdir=/etc --libdir=/usr/lib64 --with-freefem-options="--enable-download --enable-optim --enable-opengl --with-flib=/usr/lib64/libgfortran.so.3.0.0 --with-mpi=yes"

Take Care,

Fred
fred 01/10/10 12:31
Thanks for the reply and the help. I'm re-compiling it right now, actually on two different systems. One is 64 bit and the other 32 bit. I found the same problem on a "vanilla" slackware64 box with version 13.0 on it. I made your suggested changes to the configure options and will let them run to see what happens. It should be done in a short while so I will post again soon. The location of libgfortran.so is in /usr/lib on 32 bit systems with Slackware and /usr/lib64 on 64 bit systems.

I have ffcs-10.15 running on my old P3 laptop with Slackware-12.2 on it. I'ts a little slow but very nice! A great tool! Many Thanks!

Take Care,

Fred
antoine 01/10/10 11:03
Hi Fred,

This is definitely a Fortran library problem, most probably caused by a bug in ff/local/configure.ac. The last item on the problematic link command should not be "-lf2c" (is "-ff2c" a copy/paste artifact?) but something like "/usr/lib/gcc/i486-linux-gnu/4.4.5/libgfortran.so" (this is what I get on a Debian system). Could you find the corresponding location of libgfortran.so on your system and run the modified link command manually? If this manual check works, I will modify ff/local/configure.ac accordingly.

Thank you for the time you take in testing and reporting these bugs on Slackware. This is very valuable to me as I don't have an easy access to a Slackware system.
fred 01/10/10 02:53
On "Plain install" of Slackware 13.1 (32 bits smp kernel) the following errors occurs (looks like previous problem in a different way?):

make[6]: Entering directory `/home/fred/src/ffcs-10.15/ff/local/src/x11'
g++ -DHAVE_CONFIG_H -I. -I../.. -I../fflib -I../Graphics -I../.. -I../femlib -DNDEBUG -O3 -march=pentium4 -DDRAWING -DBAMG_LONG_LONG -DNCHECKPTR -rdynamic -fPIC -MT Xrgraph.o -MD -MP -MF .deps/Xrgraph.Tpo -c -o Xrgraph.o `test -f '../Graphics/Xrgraph.cpp' || echo './'`../Graphics/Xrgraph.cpp
mv -f .deps/Xrgraph.Tpo .deps/Xrgraph.Po
g++ -DHAVE_CONFIG_H -I. -I../.. -I../fflib -I../Graphics -I../.. -I../femlib -DNDEBUG -O3 -march=pentium4 -DDRAWING -DBAMG_LONG_LONG -DNCHECKPTR -rdynamic -fPIC -MT ffcsstream.o -MD -MP -MF .deps/ffcsstream.Tpo -c -o ffcsstream.o `test -f '../../../ffcsstream.cpp' || echo './'`../../../ffcsstream.cpp
mv -f .deps/ffcsstream.Tpo .deps/ffcsstream.Po
g++ -DNDEBUG -O3 -march=pentium4 -DDRAWING -DBAMG_LONG_LONG -DNCHECKPTR -rdynamic -fPIC -o FreeFem++-x11 Xrgraph.o ffcsstream.o ../lglib/liblg.a ../fflib/libff.a /home/fred/src/ffcs-10.15/ff/local/download/lib/libumfpack.a /home/fred/src/ffcs-10.15/ff/local/download/lib/libamd.a -L/home/fred/src/ffcs-10.15/ff/local/download/lib -larpack -llapack -L/home/fred/src/ffcs-10.15/ff/local/download/blas -lcblas -lf77blas -lXxf86vm -lXext -lX11 -lXpm -lm -ldl -ff2c
/home/fred/src/ffcs-10.15/ff/local/download/lib/libarpack.a(dnaupd.o): In function `dnaupd_':
dnaupd.f:(.text+0x3a0): undefined reference to `_gfortran_st_write'
dnaupd.f:(.text+0x3ae): undefined reference to `_gfortran_st_write_done'
dnaupd.f:(.text+0x3f6): undefined reference to `_gfortran_st_write'
dnaupd.f:(.text+0x410): undefined reference to `_gfortran_transfer_integer'
dnaupd.f:(.text+0x424): undefined reference to `_gfortran_transfer_integer'

Configure was as follows:

./configure --prefix=/opt/ffcs --localstatedir=/var --sysconfdir=/etc --with-freefem-options="--enable-download --enable-optim --enable-opengl --with-flib=-ff2c --with-mpi=yes"

Works great and builds fine on Slackware 13.0 (32bits, smp kernel)

Thanks!

Fred
antoine 06/09/10 11:42
Thank you for letting me know. It should be ok now.
anonymous 06/09/10 11:37
I'd like to try latest Freefem++-cs but downloads links seem to be all "broken"