March 2011 - cgma-dev - mailman3.cels.anl.gov

r4634 - cgm/trunk
by tautges＠mcs.anl.gov 23 Mar '11

23 Mar '11

Author: tautges Date: 2011-03-23 14:47:28 -0500 (Wed, 23 Mar 2011) New Revision: 4634 Modified: cgm/trunk/configure.ac Log: Change ordering of a few OCC libs to match perceived dependencies between them. Modified: cgm/trunk/configure.ac =================================================================== --- cgm/trunk/configure.ac 2011-03-23 19:17:40 UTC (rev 4633) +++ cgm/trunk/configure.ac 2011-03-23 19:47:28 UTC (rev 4634) @@ -354,7 +354,7 @@ if test "x$occ_DIR" != "xno"; then OCC_CORE_LIBS="-lTKMesh -lTKTopAlgo -lTKGeomAlgo -lTKBRep -lTKGeomBase -lTKG3d -lTKG2d -lTKMath -lTKernel" OCC_GEOM_LIBS="-lTKHLR -lTKOffset -lTKShHealing -lTKFillet -lTKFeat -lTKBool -lTKBO -lTKPrim" - OCC_LCAF_LIBS="-lTKCDF -lTKLCAF -lTKBinL" + OCC_LCAF_LIBS="-lTKBinL -lTKLCAF -lTKCDF" OCC_LIBS="$OCC_LCAF_LIBS $OCC_GEOM_LIBS $OCC_CORE_LIBS" CUBIT_OCC_LIB="-lcubit_OCC" OCC_STEP_LIBS="-lTKSTEP -lTKSTEP209 -lTKSTEPAttr -lTKSTEPBase -lTKXSBase"

1 0

r4633 - cgm/trunk
by tautges＠mcs.anl.gov 23 Mar '11

23 Mar '11

Author: tautges Date: 2011-03-23 14:17:40 -0500 (Wed, 23 Mar 2011) New Revision: 4633 Modified: cgm/trunk/configure.ac Log: Added TKCDF to list of OCC libs to pull in (static builds fail w/o this one). Modified: cgm/trunk/configure.ac =================================================================== --- cgm/trunk/configure.ac 2011-03-23 18:58:17 UTC (rev 4632) +++ cgm/trunk/configure.ac 2011-03-23 19:17:40 UTC (rev 4633) @@ -354,7 +354,7 @@ if test "x$occ_DIR" != "xno"; then OCC_CORE_LIBS="-lTKMesh -lTKTopAlgo -lTKGeomAlgo -lTKBRep -lTKGeomBase -lTKG3d -lTKG2d -lTKMath -lTKernel" OCC_GEOM_LIBS="-lTKHLR -lTKOffset -lTKShHealing -lTKFillet -lTKFeat -lTKBool -lTKBO -lTKPrim" - OCC_LCAF_LIBS="-lTKLCAF -lTKBinL" + OCC_LCAF_LIBS="-lTKCDF -lTKLCAF -lTKBinL" OCC_LIBS="$OCC_LCAF_LIBS $OCC_GEOM_LIBS $OCC_CORE_LIBS" CUBIT_OCC_LIB="-lcubit_OCC" OCC_STEP_LIBS="-lTKSTEP -lTKSTEP209 -lTKSTEPAttr -lTKSTEPBase -lTKXSBase"

1 0

Re: [cgma-dev] r4627 - cgm/trunk/itaps
by Iulian Grindeanu 21 Mar '11

21 Mar '11

----- Original Message ----- > On 03/21/2011 04:37 PM, Iulian Grindeanu wrote: > > Thanks Jason, > > > > So those extra include folders no longer appear, and no spurious > > hdf5 link problems appear for me. > > I have only one other problem, but I do not think it is related to > > this, it must be something else > > (test_copymesh fails now, and test_extrudemesh) > > MeshKit/test/algs> test_copymesh > > Running test_load_and_copymove ... > > Running test_copy_quad ... > > terminate called after throwing an instance of 'MeshKit::Error' > > what(): CESets.cpp, line 117: Invalid tag data type > > test_copy_quad: TERMINATED (signal 6) > > test_copy_quad: TERMINATED (signal 11) > > > > This is most likely due to changes in r4628 > > - jason Hi Jason, Yes, you are right; the "local_tag" type is by default iBase_ENTITY_HANDLE, while later it needs to be iBase_ENTITY_SET_HANDLE void CopyMesh::do_copy(iMesh::EntitySetHandle set_handle) { assert(transform); LocalTag local_tag(this->mk_core()); ... link_expand_sets(expandSets, local_tag); process_ce_sets(mesh, copySets.sets(), local_tag);

1 0

Re: [cgma-dev] r4627 - cgm/trunk/itaps
by Jason Kraftcheck 21 Mar '11

21 Mar '11

On 03/21/2011 04:37 PM, Iulian Grindeanu wrote: > Thanks Jason, > > So those extra include folders no longer appear, and no spurious hdf5 link problems appear for me. > I have only one other problem, but I do not think it is related to this, it must be something else > (test_copymesh fails now, and test_extrudemesh) > MeshKit/test/algs> test_copymesh > Running test_load_and_copymove ... > Running test_copy_quad ... > terminate called after throwing an instance of 'MeshKit::Error' > what(): CESets.cpp, line 117: Invalid tag data type > test_copy_quad: TERMINATED (signal 6) > test_copy_quad: TERMINATED (signal 11) > This is most likely due to changes in r4628 - jason

1 0

Re: [cgma-dev] r4627 - cgm/trunk/itaps
by Iulian Grindeanu 21 Mar '11

21 Mar '11

Thanks Jason, So those extra include folders no longer appear, and no spurious hdf5 link problems appear for me. I have only one other problem, but I do not think it is related to this, it must be something else (test_copymesh fails now, and test_extrudemesh) MeshKit/test/algs> test_copymesh Running test_load_and_copymove ... Running test_copy_quad ... terminate called after throwing an instance of 'MeshKit::Error' what(): CESets.cpp, line 117: Invalid tag data type test_copy_quad: TERMINATED (signal 6) test_copy_quad: TERMINATED (signal 11) Thanks again, Iulian ----- Original Message ----- > Author: kraftche > Date: 2011-03-21 15:55:26 -0500 (Mon, 21 Mar 2011) > New Revision: 4627 > > Modified: > cgm/trunk/itaps/iGeom-Defs.inc.in > Log: > No longer try to provide non-C++ ITAPS clients w/ necessary link flags > for linking C++ implementation. Clients must now either a) link with > C++ compiler, b) build MOAB as shared lib, or c) figure out for > themselves how to link C++ itaps impl with whatever language their app > is implemented in > > Modified: cgm/trunk/itaps/iGeom-Defs.inc.in > =================================================================== > --- cgm/trunk/itaps/iGeom-Defs.inc.in 2011-03-21 20:55:06 UTC (rev > 4626) > +++ cgm/trunk/itaps/iGeom-Defs.inc.in 2011-03-21 20:55:26 UTC (rev > 4627) > @@ -8,9 +8,9 @@ > IGEOM_CPPFLAGS = $(IGEOM_DEFINES) $(IGEOM_INCLUDES) > > # The useful flags we are providing > -IGEOM_LDFLAGS = $(IGEOM_CXX_LDFLAGS) @CGM_EXT_LDFLAGS@ > @CGM_CXX_LINKFLAGS@ > +IGEOM_LDFLAGS = $(IGEOM_CXX_LDFLAGS) @CGM_EXT_LDFLAGS@ > IGEOM_LTFLAGS = $(IGEOM_CXX_LTFLAGS) @CGM_EXT_LTFLAGS@ > -IGEOM_LIBS = $(IGEOM_LDFLAGS) -liGeom -lcgm @CGM_EXT_LIBS@ > @CGM_CXX_LIBS@ > +IGEOM_LIBS = $(IGEOM_LDFLAGS) -liGeom -lcgm @CGM_EXT_LIBS@ > > # These are overridden after this file in installed > IGEOM_CXX_LDFLAGS = -L@abs_top_builddir@/.libs -L@abs_builddir@/.libs

1 0

r4627 - cgm/trunk/itaps
by kraftche＠cae.wisc.edu 21 Mar '11

21 Mar '11

Author: kraftche Date: 2011-03-21 15:55:26 -0500 (Mon, 21 Mar 2011) New Revision: 4627 Modified: cgm/trunk/itaps/iGeom-Defs.inc.in Log: No longer try to provide non-C++ ITAPS clients w/ necessary link flags for linking C++ implementation. Clients must now either a) link with C++ compiler, b) build MOAB as shared lib, or c) figure out for themselves how to link C++ itaps impl with whatever language their app is implemented in Modified: cgm/trunk/itaps/iGeom-Defs.inc.in =================================================================== --- cgm/trunk/itaps/iGeom-Defs.inc.in 2011-03-21 20:55:06 UTC (rev 4626) +++ cgm/trunk/itaps/iGeom-Defs.inc.in 2011-03-21 20:55:26 UTC (rev 4627) @@ -8,9 +8,9 @@ IGEOM_CPPFLAGS = $(IGEOM_DEFINES) $(IGEOM_INCLUDES) # The useful flags we are providing -IGEOM_LDFLAGS = $(IGEOM_CXX_LDFLAGS) @CGM_EXT_LDFLAGS@ @CGM_CXX_LINKFLAGS@ +IGEOM_LDFLAGS = $(IGEOM_CXX_LDFLAGS) @CGM_EXT_LDFLAGS@ IGEOM_LTFLAGS = $(IGEOM_CXX_LTFLAGS) @CGM_EXT_LTFLAGS@ -IGEOM_LIBS = $(IGEOM_LDFLAGS) -liGeom -lcgm @CGM_EXT_LIBS@ @CGM_CXX_LIBS@ +IGEOM_LIBS = $(IGEOM_LDFLAGS) -liGeom -lcgm @CGM_EXT_LIBS@ # These are overridden after this file in installed IGEOM_CXX_LDFLAGS = -L@abs_top_builddir@/.libs -L@abs_builddir@/.libs

1 0

r4623 - cgm/trunk/m4
by kraftche＠cae.wisc.edu 21 Mar '11

21 Mar '11

Author: kraftche Date: 2011-03-21 14:23:05 -0500 (Mon, 21 Mar 2011) New Revision: 4623 Modified: cgm/trunk/m4/compiler.m4 Log: add simple checks that C++ compiler works because autoconf doesn't seem to check that if the compiler is specified by the user via CXX= Modified: cgm/trunk/m4/compiler.m4 =================================================================== --- cgm/trunk/m4/compiler.m4 2011-03-21 19:22:59 UTC (rev 4622) +++ cgm/trunk/m4/compiler.m4 2011-03-21 19:23:05 UTC (rev 4623) @@ -10,6 +10,27 @@ ]) +####################################################################################### +# Check if the C++ compiler works. +# Arguments: action on success and action on failure +####################################################################################### +AC_DEFUN([FATHOM_CHECK_CXX_WORKS], [ + AC_LANG_PUSH([C++]) + AC_MSG_CHECKING([if $CXX works]) + AC_COMPILE_IFELSE( + [AC_LANG_PROGRAM( [class Cl { int i; public: Cl(int x) : i(x) {} };] [Cl x(5);])], + [AC_COMPILE_IFELSE( + [AC_LANG_PROGRAM( [], +[#ifdef __cplusplus + choke me + #endif])], + [AC_MSG_RESULT([no (accepted invalid input)]); $2], [AC_MSG_RESULT([yes]); $1])], + [AC_MSG_RESULT([no (failed on trival valid C++ source)]); $2]) + + AC_LANG_POP([C++]) +]) + + ########## Helper function for FATHOM_CHECK_COMPILERS ############# # args: compiler variable, compiler list, path AC_DEFUN([FATHOM_SET_MPI_COMPILER], [

1 0

r4620 - cgm/trunk
by kraftche＠cae.wisc.edu 21 Mar '11

21 Mar '11

Author: kraftche Date: 2011-03-21 14:22:18 -0500 (Mon, 21 Mar 2011) New Revision: 4620 Modified: cgm/trunk/configure.ac Log: add simple checks that C++ compiler works because autoconf doesn't seem to check that if the compiler is specified by the user via CXX= Modified: cgm/trunk/configure.ac =================================================================== --- cgm/trunk/configure.ac 2011-03-21 19:16:35 UTC (rev 4619) +++ cgm/trunk/configure.ac 2011-03-21 19:22:18 UTC (rev 4620) @@ -22,6 +22,8 @@ AC_C_BIGENDIAN( [], [AM_CPPFLAGS="$AM_CPPFLAGS -DLITTLE_ENDIAN=BYTE_ORDER"] ) FATHOM_COMPILER_FLAGS +FATHOM_CHECK_CXX_WORKS([], [AC_MSG_ERROR([Cannot build without C++ compiler])]) + ################################################################################ # Fortran Stuff for ITAPS C bindings ################################################################################

1 0

Re: [cgma-dev] r4616 - in cgm/trunk: . m4
by Iulian Grindeanu 21 Mar '11

21 Mar '11

So, I did a make install in /homes/fathom/3rdparty/mpich2_gcc4.4.3/mpich2-1.2.1/gcc, which seems to have solved the problem Still, I am rather uneasy having to do that. Does anybody worked on that compiler? It seems that everybody has writing permissions there ... Thanks, Iulian ----- Original Message ----- > Thanks, Jason, > Indeed: > /homes/fathom/3rdparty/mpich2/mpich2-1.2.1/gcc/bin> ls -l mpic* > lrwxrwxrwx 1 smithrm fathom 6 2010-07-20 15:58 mpic++ -> mpicxx* > -rwxrwxr-x 1 smithrm fathom 8.4K 2010-07-20 15:58 mpicc* > -rwxrwxr-x 1 smithrm fathom 13K 2010-07-20 15:58 mpich2version* > -rwxrwxr-x 1 smithrm fathom 0 2011-03-21 10:48 mpicxx* > > Somebody/something must have zeroed it out > Iulian > > > ----- Original Message ----- > > On 03/21/2011 12:23 PM, Iulian Grindeanu wrote: > > > Hi Jason, > > > Indeed, something is really wrong with my builds. I just cannot > > > understand what. > > > > > > > From the configure log it looks like you C++ compiler you specified > > (CXX=/homes/fathom/3rdparty/mpich2/mpich2-1.2.1/gcc/bin/mpicxx) > > doesn't work. > > > > - jason > > > > > > > > >> > > >> > > >> I do not see these errors. And the inability to find standard > > >> headers > > >> like > > >> 'fstream' and 'limits.'h is a) indicative of something being > > >> quite > > >> wrong and > > >> b) most likely the reason that Standard_Version.hxx appears > > >> unusable. > > >> > > >> Have you tried regenerating the autotools stuff (autoreconf > > >> -fi)?. > > >> Is > > >> the > > >> detected C++ compiler what you expect it to be? > > > > > > I did try autoreconf -fi > > > > > > I am using mpich2 from /homes/fathom/3rdparty folder. > > > > > > Where should I look? Something changed over the weekend :( it was > > > fine last week. > > > > > > Thanks, > > > Iulian > > > > > > -- > > "A foolish consistency is the hobgoblin of little minds" - Ralph > > Waldo > > Emerson

1 0

Re: [cgma-dev] r4616 - in cgm/trunk: . m4
by Iulian Grindeanu 21 Mar '11

21 Mar '11

Thanks, Jason, Indeed: /homes/fathom/3rdparty/mpich2/mpich2-1.2.1/gcc/bin> ls -l mpic* lrwxrwxrwx 1 smithrm fathom 6 2010-07-20 15:58 mpic++ -> mpicxx* -rwxrwxr-x 1 smithrm fathom 8.4K 2010-07-20 15:58 mpicc* -rwxrwxr-x 1 smithrm fathom 13K 2010-07-20 15:58 mpich2version* -rwxrwxr-x 1 smithrm fathom 0 2011-03-21 10:48 mpicxx* Somebody/something must have zeroed it out Iulian ----- Original Message ----- > On 03/21/2011 12:23 PM, Iulian Grindeanu wrote: > > Hi Jason, > > Indeed, something is really wrong with my builds. I just cannot > > understand what. > > > > From the configure log it looks like you C++ compiler you specified > (CXX=/homes/fathom/3rdparty/mpich2/mpich2-1.2.1/gcc/bin/mpicxx) > doesn't work. > > - jason > > > > >> > >> > >> I do not see these errors. And the inability to find standard > >> headers > >> like > >> 'fstream' and 'limits.'h is a) indicative of something being quite > >> wrong and > >> b) most likely the reason that Standard_Version.hxx appears > >> unusable. > >> > >> Have you tried regenerating the autotools stuff (autoreconf -fi)?. > >> Is > >> the > >> detected C++ compiler what you expect it to be? > > > > I did try autoreconf -fi > > > > I am using mpich2 from /homes/fathom/3rdparty folder. > > > > Where should I look? Something changed over the weekend :( it was > > fine last week. > > > > Thanks, > > Iulian > > > -- > "A foolish consistency is the hobgoblin of little minds" - Ralph Waldo > Emerson

1 0