This should reduce the amount of work for the compiler, and reduce
the chances of an unintelligent compiler/linker dumping a lot of
these symbols into an executable/library.

Suggested by Jed Brown:
http://lists.mpich.org/pipermail/discuss/2013-January/000230.html

No reviewer.

This says that layout compatibility matching should be used instead of
checking for the exact type (e.g., "mpich_struct_mpi_short_int").

Contributed by Jed Brown:
http://lists.mpich.org/pipermail/discuss/2013-January/000226.html

No reviewer.

Fixed misuse of MPI_ namespace in the mutex library.

Reviewer: none

This test case is a variation on the linked list test case that builds a linked
list with a deterministic structure; node k in the list is owned by rank (k %
nproc).  Three variations are included to exercise lock-exclusive, lock-shared,
and lock-all synchronization modes.

Reviewer: none

Also update a couple of other "maint/" files.

Partly reviewed by balaji@.

Now that we are using git, it doesn't make any sense and can only cause
problems.

No reviewer.

git-subtree-dir: src/openpa
git-subtree-split: b48b33c3b0f64456ed86ff0900462593dc159fd8

Also add autogen.sh logic to copy from authoritative sources to the
correct "externals".  So "src/pm/hydra/confdb" is now copied (in a
special fashion) from "confdb" at autogen.sh-time.  Prior to this commit
these externals were simply tracked in the git repository as a
consequence of their (non-trivial) export from SVN.

Non-relative svn:externals, such as "src/openpa" are not dealt with in
this commit.

No reviewer.

Satish pointed out that we were causing problems when the user specified
directories with a `+` in them.  The updated version uses a `|`, which
is a universally poor user choice for directory names (much like `:`
would be).  Single quotes are now also used instead of double quotes.
This sed code is still pretty fragile in the case of exotic path names,
but it's better now.  If this is still a problem for users, the next
step is either to give up on permitting make-time overrides of special
vars like `prefix` or to at least disable that functionality with a
configure switch of some sort.

Reviewed by balaji@.

This is a follow-on to r10622 and tt#1570.  Eric Borisch reported that
shared libraries were not generated on Darwin platforms when
`--disable-f77` and `--disable-fortran` are passed.  The fix in r10622
was buggy, but hopefully this one works a bit better.

If you're debugging this in the future, search for `_lt_disable_F77` in
the generated configure file to help understand why we expect `F77=no`
to prevent libtool problems (and possibly help debug the future
problem).

Reviewed by balaji@.

No reviewer.

No reviewer.

No reviewer.

sense.  Also improve the code to make sure the datatype creation time
does not depend on the data size.

Reviewed by goodell.

Removed duplicate PMPI_Comm_create_group declaration.  Patch provided by Jed
Brown.

The "#include 'mpix.h'" is no longer needed in the mutexes test case; the
needed declarations have been moved to mpi.h.

No reviewer.

1). There was error checking on the comm object in
    MPI_Comm_test_inter(comm, &flag); So if the return value of
    MPI_Comm_test_inter is not MPI_SUCCESS, then the comm is either an
    invalid MPI_Comm handle or intercommunicator handle.
2). A new macro MPIO_CHECK_INFO is added into adioi_error.h. It will
    call MPI_Info_dup, unless there is no more memory space left , as
    long as the info object is valid, this function will return
    MPI_SUCCESS; or it will return an error code. So by checking the
    return value of MPI_Info_dup, we could achieve the purpose of
    checking MPI_Info handles.

Based on patch 0006 from the second round of IBM's error checking
patches.  Replaces 0009 from the first round and augments r10637.

By using MPI_Allreduce to get the maximum value and minimum value of all
sizes, when the two values are identical, all processes have same values
of size.  The problem of checking the sizes with MPI_Bcast is that the
root will pass the check while the others not.

Based on patch 0004 from the second round of IBM's error checking
patches.

This only works when ROMIO is built with MPICH.

Based on patch 0001 from the second round of IBM's error checking
patches.  Replaces 0008 from the first round.

No ABI changes, just hydra fixes, so 10:0:0 becomes 10:1:0.

setting it to nonblocking.  We released 3.0 without this fix.  God
help us with the bug reports!

No reviewer.

so we don't have to keep correcting this.  Also removed section
numbers from the MPI-2 standard that were referenced in the ROMIO
documents.  If we feel that they are necessary, we can add them back
after the release.

No reviewer.

No reviewer.

No reviewer.

Added -h as another argument to access hydra's help message.

Reviewer: balaji

setting it to nonblocking.  We released 3.0 without this fix.  God
help us with the bug reports!

No reviewer.

so we don't have to keep correcting this.  Also removed section
numbers from the MPI-2 standard that were referenced in the ROMIO
documents.  If we feel that they are necessary, we can add them back
after the release.

No reviewer.

No reviewer.

No reviewer.

Added -h as another argument to access hydra's help message.

Reviewer: balaji

No reviewer.