2008/3/30, Johan Hake <[EMAIL PROTECTED]>: > On Sunday 30 March 2008 20:45:02 Åsmund Ødegård wrote: > > On Sun, Mar 30, 2008 at 12:51 PM, Johan Hake <[EMAIL PROTECTED]> wrote: > > [snip] > > > > > > > This should be quite straight forward to do with dolfin-scons, as we > > > > > already do this for pycc. One dolfin_module-name.pc file will also > > > > > be created for each module, making linking and compiling against a > > > > > specific module easy. > > > > > > How straight forward is this Åsmund, Johannes, Ola? :) > > > > Just from top of my head: > > > > Very easy, indeed. In principle, just remove the scons.cfg dolfin have > > today, and start making scons.cfg files inside submodules; as it is > > done in PyCC. I belive the building of dolfin_moudle-name.pc is no > > fully automatic yet, so some hooks must be put in SConscript, besides > > that most thing should just work. > > > How is the logic of swig directories and python modules in simula-scons? > > In pycc we have a swig subdirectory in each module directory, which results > in > one python module for each c++ module. In dolfin we have a swig directory > directly in the package directory. I suppose we want to continue with this > after a possible move to modularized c++ libraries. > > > Johan
Having a single swig generated module separately from the C++ modules like it is with PyDOLFIN has many benefits, this is also the solution Trilinos ended up with after having problems with the "one swig module per C++ module" model. So lets not follow the PyCC structure on that part. -- Martin _______________________________________________ DOLFIN-dev mailing list [email protected] http://www.fenics.org/mailman/listinfo/dolfin-dev
