[svn-r8126] Purpose:

Bug fix/optimization

Description:
    Address slowdown in MPI-I/O file metadata operations that was introduced
mid-stream.  We now _require_ a POSIX compliant parallel file system for the
MPI-I/O file driver (as well as for the MPI-POSIX file driver).
    Also optimized file open operation when the file is being created by
reducing the number of collective & syncronizing calls.
    Additionally, refactor the MPI routines into a common place, eliminating
duplicated code.

Platforms tested:
    FreeBSD 4.9 (sleipnir) w/parallel
    h5committest

doc/html/Environment.html

@@ -87,6 +87,10 @@ their defined effects are as follows:
             directory do not remove temporary HDF5 data files.
             The default is for each test to remove the files before exit.
 
+    <dt><tt>HDF5_DISABLE_VERSION_CHECK</tt> 
+        <dd>When set to <code>1</code>, HDF5 will not abort when the version
+            of the HDF5 headers doesn't match the version of the HDF5 library.
+
     <dt><tt>HDF5_MPI_OPT_TYPES</tt> 
             &nbsp;&nbsp;(for parallel beta version only)
         <dd>When set to <code>1</code>, PHDF5 will use the MPI optimized 
@@ -94,13 +98,6 @@ their defined effects are as follows:
             Currently, this optimization fails when accessing extendable
             datasets.  The default is not to use the optimized code.
 
-    <dt><tt>HDF5_MPI_1_METAWRITE</tt> 
-            &nbsp;&nbsp;(for parallel beta version only)
-        <dd>When set to <code>1</code>, PHDF5 will write the metadata 
-            via process 0 of each opened parallel HDF5 file.  This should 
-            improve I/O throughput.  The default is not to use this 
-            optimization.
-
 </dl>
 </dir>