On Wed, Aug 18, 2010 at 08:31:55PM +0530, Harish Narayanan wrote: > On 8/18/10 8:23 PM, Garth N. Wells wrote: > > On Wed, 2010-08-18 at 16:16 +0200, Anders Logg wrote: > >> On Wed, Aug 18, 2010 at 03:40:47PM +0200, Anders Logg wrote: > >>> On Wed, Aug 18, 2010 at 06:58:24PM +0530, Harish Narayanan wrote: > >>>> On 8/18/10 6:32 PM, Anders Logg wrote: > > > > > >>>> > >>>> ================================================================================ > >>>> SLEPc Configuration > >>>> ================================================================================ > >>>> > >>>> SLEPc source directory: > >>>> /Users/harish/Work/FEniCS/test/src/slepc-3.1-p1 > >>>> SLEPc install directory: > >>>> /Users/harish/Work/FEniCS/test > >>>> PETSc directory: > >>>> /Users/harish/Work/FEniCS/test > >>>> Architecture "unknown" with double precision real numbers > >>>> > >>>> ** > >>>> ** Before running "make" your PETSC_ARCH must be specified with: > >>>> ** ** setenv PETSC_ARCH unknown (csh/tcsh) > >>>> ** ** PETSC_ARCH=unknown ; export PETSC_ARCH (sh/bash) > >>>> ** > >>>> > >>>> makefile:30: darwin10.0.0-cxx-opt/conf/slepcvariables: No such file or > >>>> directory > >>>> /Users/harish/Work/FEniCS/test/src/slepc-3.1-p1/conf/slepc_common_variables:2: > >>>> /Users/harish/Work/FEniCS/test/src/slepc-3.1-p1/darwin10.0.0-cxx-opt/conf/slepcvariables: > >>>> No such file or directory > >>>> /Users/harish/Work/FEniCS/test/src/slepc-3.1-p1/conf/slepc_common_rules:78: > >>>> /Users/harish/Work/FEniCS/test/src/slepc-3.1-p1/darwin10.0.0-cxx-opt/conf/slepcrules: > >>>> No such file or directory > >>>> make: *** No rule to make target > >>>> `/Users/harish/Work/FEniCS/test/src/slepc-3.1-p1/darwin10.0.0-cxx-opt/conf/slepcrules'. > >>>> Stop. > >>>> Failure with exit status: 2 > >>>> Exit message: There was a problem building slepc-3.1-p1. > >>> > >>> I get exactly the same thing. > >> > >> I have something that appears to work now. Will push when I have > >> tested it. > >> > > > > Check that it works with petsc commented out in the .platform file (but > > already complied, e.g. on a previous run of dorsal). > > Dorsal is not designed to this way. You are supposed to > > skip:petsc > > (instead of commenting it out) in the packages list. Commenting out > means dorsal knows nothing of petsc and slepc should fail. skip:petsc > means dorsal knows about it but will skip compiling it. slepc will pick > this up, and so will DOLFIN. > > Harish
I made an attempt to fix this. Looking at petsconf.py in SLEPc, it seems that one should not set PETSC_ARCH when building against an installed version of PETSc, so I just made sure that PETSC_ARCH is unset when Dorsal configures and builds SLEPc and that PETSC_DIR is set to the installation directory. It also seems that it doesn't really matter what we set PETSC_ARCH to since we just build one version of PETSc. It's just a name and it's not used by SLEPc. -- Anders
signature.asc
Description: Digital signature
_______________________________________________ Mailing list: https://launchpad.net/~dorsal Post to : [email protected] Unsubscribe : https://launchpad.net/~dorsal More help : https://help.launchpad.net/ListHelp
