[petsc-dev] MatTranspose_MPIAIJ broken

[Mark F. Adams]

On May 25, 2013, at 8:25 PM, Jed Brown <jedbrown at mcs.anl.gov> wrote:

> "Mark F. Adams" <mark.adams at columbia.edu> writes:
> 
>> I just updated petsc and rebuilt and am now getting this error:
> 
> Have you reconfigured since the commit below?  

Yes, I reconfigured after my pull today.

> What do you have for
> PETSC_HAVE_MPI_COMBINER_DUP in your petscconf.h?
> 

madams/petsc> grep PETSC_HAVE_MPI_COMBINER_DUP ./arch-xc30-opt/include/petscconf.h
#ifndef PETSC_HAVE_MPI_COMBINER_DUP
#define PETSC_HAVE_MPI_COMBINER_DUP 1

I can't reproduce this problem on my Mac with ex56.  This error was on Edison at NERSc.  The new Cray.

I'll try rebuilding everything from scratch to double check.

> commit 77e230d68da10c8f2d04dd9c231f8a28ef7937a2
> Author: Jed Brown <jedbrown at mcs.anl.gov>
> Date:   Fri May 10 16:51:35 2013 -0500
> 
>    MPI.py: fix check for MPI_COMBINER_DUP
> 
> diff --git a/config/BuildSystem/config/packages/MPI.py b/config/BuildSystem/config/packages/MPI.py
> index 80585dc..8ac81e4 100644
> --- a/config/BuildSystem/config/packages/MPI.py
> +++ b/config/BuildSystem/config/packages/MPI.py
> @@ -766,8 +766,13 @@ class Configure(config.package.Package):
>     for f in funcs:
>       if self.libraries.check(self.dlib, f):
>         self.addDefine('HAVE_' + f.upper(),1)
> +
> +    oldFlags = self.compilers.CPPFLAGS # Disgusting save and restore
> +    self.compilers.CPPFLAGS += ' '+self.headers.toString(self.include)
>     if self.checkCompile('#include <mpi.h>', 'int combiner = MPI_COMBINER_DUP;'):
>       self.addDefine('HAVE_MPI_COMBINER_DUP',1)
> +    self.compilers.CPPFLAGS = oldFlags
> +
>     if self.libraries.check(self.dlib, "MPIDI_CH3I_sock_set"):
>       self.addDefine('HAVE_MPICH_CH3_SOCK', 1)
>     if self.libraries.check(self.dlib, "MPIDI_CH3I_sock_fixed_nbc_progress"):
>