On Sun, Sep 06, 2009 at 07:14:03PM +0200, Johan Hake wrote: > On Saturday 05 September 2009 23:40:50 DOLFIN wrote: > > One or more new changesets pushed to the primary dolfin repository. > > A short summary of the last three changesets is included below. > > > > changeset: 6950:e685bced17258c83ec9c55badc4ce3adf991c1d8 > > tag: tip > > user: Anders Logg <l...@simula.no> > > date: Sat Sep 05 23:40:39 2009 +0200 > > files: ChangeLog demo/parameters/cpp/main.cpp > > dolfin/main/SubSystemsManager.cpp dolfin/main/SubSystemsManager.h > > dolfin/parameter/GlobalParameters.cpp dolfin/parameter/GlobalParameters.h > > dolfin/parameter/Parameters.cpp dolfin/parameter/Parameters.h description: > > Filter out PETSc argument when parsing command-line parameters. > > The following should now work: > > > > ./demo <other options> --petsc.foo bar <other options> > > > > provided the program in question parses command-line arguments by > > > > parameters.parse(argc, argv). > > > > All arguments starting with "--petsc." will be sent to PETSc. In the > > above example, > > > > -foo bar > > > > will be sent to PETSc. (Note '-' instead of '--'.) > > > > Other arguments will be parsed by the parameter set in question. > > > > This only happens for the global DOLFIN parameter set (or any other > > parameter set named "dolfin"). > > It is a bit fragile to rely on the naming of a particular parameter. > > Couldn't the petsc part be implemented in GlobalParameters::parse() instead, > for example by letting 'parse' be a virtual function?
Yes, good idea. I'll change it when I fix the () / [] issue. -- Anders
signature.asc
Description: Digital signature
_______________________________________________ DOLFIN-dev mailing list DOLFIN-dev@fenics.org http://www.fenics.org/mailman/listinfo/dolfin-dev