Package Details: freefem 4.9.0-1

Git Clone URL: https://aur.archlinux.org/freefem.git (read-only, click to copy)
Package Base: freefem
Description: A PDE oriented language using the finite element method
Upstream URL: https://freefem.org/index.html
Licenses: LGPL
Conflicts: freefem-ext-dl, freefem-git
Provides: freefem
Submitter: haawda
Maintainer: gdolle
Last Packager: gdolle
Votes: 14
Popularity: 0.000270
First Submitted: 2019-06-19 19:31
Last Updated: 2021-04-19 20:23

Latest Comments

1 2 3 4 5 6 ... Next › Last »

lmello commented on 2021-02-08 23:22

Hi @gdolle,

Well let's review those dependencies then? When you say 'all opt depend aren't functional' do you mean those specified by the ./configure arguments in the PKGBUILD?

(...)
./configure --prefix=/usr \
            --enable-summary \
            --with-petsc=${PETSC_DIR}/lib/petsc/conf/petscvariables \
            --with-slepc=${SLEPC_DIR}/lib/slepc/conf/slepcvariables \
            --disable-superlu \
            --disable-mumps_seq
            #--with-ipopt-include 
            #--with-superlu-include=/usr/include/superlu \
./reconfigure
(...)

i.e. petsc, slepc, superlu, mumps and coin-and-ipopt? Because since I previously installed numpy, scipy, matplotlib, julia and octave, freefem will always compile with boost, cblas, fftw, lapack and mepis, all stated as 'optional' by this pkgbuild, and those are automatically detected by the configure script, aren't they?

Anyways, from https://doc.freefem.org/introduction/installation.html#compilation-on-arch-linux we got, at least:

1. Install the following dependencies:

pacman -Syu
pacman -S git openmpi gcc-fortran wget python
    freeglut m4 make patch gmm
    blas lapack hdf5 gsl fftw arpack suitesparse
    gnuplot autoconf automake bison flex gdb
    valgrind cmake texlive-most

Just installed valgrind, and I got this ./configure output using your script:

configure: WARNING: unrecognized options: --with-slepc
configure:   FreeFEM used download: no 
configure:   --  Dynamic load facility: yes 
configure:   --  ARPACK (eigen value): yes 
configure:   --  UMFPACK (sparse solver): yes 
configure:   --  BLAS: yes 
configure:   --  with MPI:     yes
configure:     --  with PETSc: no / PETSc complex: no 
configure:     --  with SLEPc: no / SLEPc complex: no 
configure:     --  with hpddm: no 
configure:     --  with htool: no 
configure:     --  with bemtool: no (need boost: no and htool: no) 
configure:   --  without libs:  tetgen mmg3d mshmet gmm ipopt scalapack 
mumps mumps_seq nlopt scotch superlu yams pipe MMAP NewSolver mkl hpddm bem 
configure:   --  without plugins:  tetgen.so mmg3d-v4.0.so mshmet.so 
aniso.so ilut.so ff-Ipopt.so MUMPS.so MUMPS_mpi.so MUMPS.so MUMPS_mpi.so 
MUMPS_seq.so MUMPS.so ff-NLopt.so scotch.so SuperLu.so freeyams.so pipe.so 
ff-mmap-semaphore.so NewSolver.so PARDISO.so hpddm.so 
hpddm_substructuring.so  bem.so 
configure:     progs: FreeFem++-nw bamg cvmsh2  FreeFem++-mpi ffmedit ffglut 
configure:  Activated libraries 
configure:  - ARPACK: yes
configure:  - BLACS: no
configure:  - BEMTOOL: no
configure:  - BOOST: no
configure:  - FFTW3: yes
configure:  - FORTRAN: no
configure:  - FREEYAMS: no
configure:  - GMM: no
configure:  - GSL: yes
configure:  - HDF5: yes
configure:  - HTOOL: no
configure:  - HPDDM: no
configure:  - IPOPT: no
configure:  - LIBSEQ: no
configure:  - METIS: yes
configure:  - MKL: no
configure:  - MMG3D: no
configure:  - MMG: no
configure:  - PARMMG: no
configure:  - MPI: yes
configure:  - MSHMET: no
configure:  - MUMPS: no
configure:  - MUMPS_ptscotch: no
configure:  - MUMPS_scotch: no
configure:  - MUMPS_seq: no
configure:  - NLOPT: no
configure:  - PARMETIS: yes
configure:  - PETSC: no
configure:  - PETSCCOMPLEX: no
configure:  - PTHREAD: yes
configure:  - PTSCOTCH: no
configure:  - SCALAPACK: no
configure:  - SCOTCH: no
configure:  - SLEPC: no
configure:  - SLEPCCOMPLEX: no
configure:  - SUPERLU: no
configure:  - TETGEN: no
configure:  - UMFPACK: yes
configure:  PETSc directories do not exist, to build do:
configure: cd 3rdparty/ff-petsc && make
==> Starting build()...

So from those opt deps I've got installed, only boost fails to be included.

gdolle commented on 2021-02-07 14:19

Hi @lmello, yes it's not really essential for the current build (no external pkg *), I'll change it next update.

NB: *Be aware this is a minimal dependency ff++ build and several features are missing (all opt depend aren't functional)

lmello commented on 2021-02-07 14:17

Follow-up: I just rebuilt the AUR package with pamac, editing the PKGBUILD accordingly. Opened a new terminal after installation and it worked just fine. So that the extra 'freefem.csh' and 'freefem.sh' source files (profile path/env scripts) are not necessary anymore with '--prefix=/usr'.

I suggest that you update the PKGBUILD with those minor edits, if only for system coherence.

lmello commented on 2021-02-07 13:32

@Sosthene

As of 2021-02-07, this build works flawlessly on a Manjaro box.

I was just wondering if the '--prefix=/opt/freefem' is essential for it in any way; from the looks of it, I don't see why the standard '--prefix=/usr' packaging shouldn't work.

Sosthene commented on 2020-05-28 11:58

@gdolle I became the maintainer of this package because it was orphaned an I needed it for a course I was taking. I don't really use it anymore so I have no idea whether the package works properly.

If you use It and know the build system I'd be grateful if you maintained it, and you'd probably do a better job than I.

gdolle commented on 2020-05-27 13:54

petsc should be a strong dependency (mumps and hpddm will be provided via petsc). Otherwise a lot of solvers won't be available

Sosthene commented on 2019-11-28 21:46

It seems that unzip, m4, flex and bison are missing as dependencies.

configure: to install missing package under debian or ubuntu, try configure: sudo apt-get install unzip configure: error: Sorry missing unzip,m4,bison,flex,patch command ! ==> ERROR: A failure occurred in prepare(). Aborting... Error making: freefem

haawda commented on 2019-06-13 17:07

Patches are welcome.

gdolle commented on 2019-06-13 11:21

@greyltc yes, the build process download dependencies (like mumps, etc). Not sure it's a good idea to do it appart

greyltc commented on 2019-06-13 09:15

Am I right in that the build process downloads something besides the URLs in the source array? Can we avoid that somehow?