[MOAB-dev] commit/MOAB: iulian07: fix parallel builds configured without mbzoltan

[Iulian Grindeanu]

yes, sorry about it 
We have only a simple build in serial after every commit; 
parallel builds without zoltan all failed last night :( 

I will enable more parallel builds in jenkins-ci, after every commit. (Tomorrow, because I cannot edit jenkins over vpn ) 

the buildbot watcher for git/bitbucket is not working right, I don't know why, so I had to use jenkins . 
Where are the petsc builds that depend on moab? I need to subscribe to that list, at least. 

Iulian 

----- Original Message -----

| Thanks for the fix Iulian. Just a simple preprocessor check for
| zoltan
| configuration got left out crashing all non-zoltan builds.

| Can we please initiate buildbot checks after every commit? Is there a
| way to even do this or do we have to wait till end of the day for
| confirmation ? These could be as simple as running "git pull; make
| check" on the different builds. MOAB master needs to be stable since
| PETSc master relies on it in certain daily configurations.

| Vijay

| On Sun, Oct 20, 2013 at 9:34 AM, <commits-noreply at bitbucket.org>
| wrote:
| > 1 new commit in MOAB:
| >
| > https://bitbucket.org/fathomteam/moab/commits/8dbdd2de6148/
| > Changeset: 8dbdd2de6148
| > Branch: master
| > User: iulian07
| > Date: 2013-10-20 16:32:07
| > Summary: fix parallel builds configured without mbzoltan
| >
| > if build in parallel, without zoltan, the guards will allow only
| > trivial
| > partition for MPAS meshes
| >
| > Affected #: 1 file
| >
| > diff --git a/src/io/NCHelperMPAS.cpp b/src/io/NCHelperMPAS.cpp
| > index 35fdb2b..497c7f1 100644
| > --- a/src/io/NCHelperMPAS.cpp
| > +++ b/src/io/NCHelperMPAS.cpp
| > @@ -302,6 +302,7 @@ ErrorCode NCHelperMPAS::create_mesh(Range&
| > faces)
| >
| >
| > #ifdef USE_MPI
| > +#ifdef HAVE_ZOLTAN
| > int& partMethod = _readNC->partMethod;
| > if (partMethod==ScdParData::RCBZOLTAN && procs >=2) // it does not
| > make sense to partition
| > // if the number of processors is less than 2; trivial partition is
| > good enough
| > @@ -331,12 +332,16 @@ ErrorCode NCHelperMPAS::create_mesh(Range&
| > faces)
| > nLocalCells = localGidCells.size();
| > }
| > else {
| > -#endif /* use mpi */
| > +#endif /* this is end for HAVE_ZOLTAN */
| > +#endif /* if use mpi */
| > + /* without zoltan, only trivial partition is possible */
| > start_cell_idx++; // 0 based -> 1 based
| > localGidCells.insert(start_cell_idx, start_cell_idx + nLocalCells -
| > 1);
| > #ifdef USE_MPI
| > +#ifdef HAVE_ZOLTAN
| > }
| > -#endif
| > +#endif /* end for HAVE_ZOLTAN */
| > +#endif /* end for USE_MPI */
| > // Read number of edges on each local cell, to calculate actual
| > maxEdgesPerCell
| > int nEdgesOnCellVarId;
| > int success = NCFUNC(inq_varid)(_fileId, "nEdgesOnCell",
| > &nEdgesOnCellVarId);
| >
| > Repository URL: https://bitbucket.org/fathomteam/moab/
| >
| > --
| >
| > This is a commit notification from bitbucket.org. You are receiving
| > this because you have the service enabled, addressing the recipient
| > of
| > this email.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/moab-dev/attachments/20131020/6d6d8d3f/attachment.html>