dolfin team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

On Wed, Jul 05, 2006 at 07:18:44PM +0800, LUK ShunTim wrote:
> Hello,
> 
> I tried building dolfin without petsc and mpi, using the hg repository in
> debian/sid. I used the configure.local script and here are my configure options
> 
> "--disable-petsc --disable-debug --enable-optimization --disable-mpi
> --enable-pydolfin --disable-curses --disable-static"
> 
> Linking to build libdolfin.so results in a bunch of name clash errors like
> 
> "../../src/kernel/fem/.libs/FEM.o:(.data+0x0): multiple definition of
> `_ZN5boost7numeric5ublas21scalar_divides_assignIT_T0_E8computedE'
> ../../src/modules/convdiff/.libs/ConvectionDiffusionSolver.o:(.data+0x0): first
> defined here
> ../../src/kernel/fem/.libs/LoadVector.o:(.data+0x0): multiple definition of
> `_ZN5boost7numeric5ublas21scalar_divides_assignIT_T0_E8computedE'
> ../../src/modules/convdiff/.libs/ConvectionDiffusionSolver.o:(.data+0x0): first
> defined here
> ../../src/kernel/fem/.libs/MassMatrix.o:(.data+0x0): multiple definition of
> `_ZN5boost7numeric5ublas21scalar_divides_assignIT_T0_E8computedE'
> ../../src/modules/convdiff/.libs/ConvectionDiffusionSolver.o:(.data+0x0): first
> defined here
> ..."
> 
> I can post the complete make log if it's useful.
> 
> Regards,
> ST

You can't use --enable-pydolfin together with --disable-petsc, see

http://www.fenics.org/cgi-bin/bugzilla/buglist.cgi?query_format=specific&order=relevance+desc&bug_status=__open__&product=&content=

We've done a lot of restructuring of the linear algebra lately and
now provide support for two separate backends: PETSc and uBlas. We
still have a few things to take care of, including updating the Python
interface.

/Anders