Contributed by IBM.  Based on patch 0010 from code discussions.

Contributed by IBM.  Based on patch 0009 from code discussions.

When opening file with inconsistent "amode"s.

Contributed by IBM.  Based on patch 0007 from code discussions.

Subtle: also converts two error checks to use different classes at the
same time.

Contributed by IBM.  Based on patch 0005 from code discussions.

Contributed by IBM.  Based on patch 0004 from code discussions.

Contributed by IBM.  Patch 0001 from code discussions.

Partially reviewed by robl@ & balaji@.

[svn-r10624] Added another test program for debugger interface, in the debugger directory where the other tests live

When --disable-LANG is passed to configure, set the corresponding
language variables to "false" to prevent searching for a working
compiler when the user specifically disabled that language.

If this commit causes any problems, revert it.  We lived without
problems without it for many years, so this isn't a critical feature.

Reviewed by balaji@.

[svn-r10619] Add warnings about the implicit run option for runtests and recognize additional programs to skip

[svn-r10613] Because MPI_WEIGHTS_EMPTY is new in MPI-3, mark Fortran test that uses this value as requiring MPI-3

[svn-r10612] Correct test to mark off part that requires MPI-3 and use Mtest routine for verbose output

Bug report from Dmitry Durnov @ Intel.  This patch also includes a fix
based on code suggested by Dmitry.

No reviewer.

No reviewer.

This commit was actually tested with `--enable-fast`, unlike r10600.

No reviewer.

latest versions.  If we later discover that we have to use a new
version of these tools, we don't have to break ABI by changing them at
that point.

Reviewed by goodell.

Reviewer: none

This is example 11.21 from the MPI 3.0 spec.

Reviewer: buntinas

This patch add a Grequest wait function, is needed to support waitall with RMA
ops.

Reviewer: buntinas

Caught by the special tests and `--enable-fast`.

No reviewer.

We currently expect this test to fail and we do not want sysadmins to be
concerned if they run `make testing` at installation time as a sanity
check.  This is really just a hack covering for our lack of an XFAIL
capability in the `runtests` machinery.

No reviewer.

No reviewer.

Reviewed by goodell.

of the MPI-3 standard).  Fixes ticket #1711.

Reviewed by robl.

This fixes the performance regression that was introduced by concatenation of
per-target lists.

Reviewer: goodell

Moved fence_issued and start_assert MPID_Win members into CH3.  These should
not be a required part of the ADI, since they are not needed above the ADI and
implementations should be free to choose different mechanisms for tracking
the state of synchronization operations.

Reviewer: buntinas

The fence_cnt field in MPID_Win is not a counter, it's a flag that indicates if
fence has been called.

Reviewer: buntinas