Splitting out ifstrstream into a header and an implementation.  This
means that much of the code that depends on it can be compiled only
once and the parallel I/O is concentrated into a few files.  To do
this, a number of files that relied on basic_ifstrstream to load the
mpi header had to be modified to manage their own headers.

More changes for parallel

Creating busticated version of OpenMD

Fixed a bug in Torsion (and GhostTorsion) which is triggered by
a configuration with colinear atoms.  The problem was discovered by
Brett Donovan.   Thanks, Brett!

Adding property set to svn entries

Fixed bugs in Restraint, refactored RestReader and RestWriter

Added new restraint infrastructure
Added MolecularRestraints
Added ObjectRestraints
Added RestraintStamp
Updated thermodynamic integration to use ObjectRestraints
Added Quaternion mathematics for twist swing decompositions
Significantly updated RestWriter and RestReader to use dump-like files
Added selections for x, y, and z coordinates of atoms
Removed monolithic Restraints class
Fixed a few bugs in gradients of Euler angles in DirectionalAtom and RigidBody
Added some rotational capabilities to prinicpalAxisCalculator

Returned old LD forces to SMIPDForceManager.

Fixes for Inversions for use with Amber

Many fixes

Changes required for Inversions and Base Atom types.  This will
break OOPSE badly for a few days or so...

A few formatting changes to prettify the code

Massive update to do virials (both atomic and cutoff-group) correctly.
The rigid body constraint contributions had been missing and this was
masked by the use of cutoff groups...

Massive changes preparing for release of OOPSE-4: The main difference
is that the new MD file format (.md, .dump, .eor) now contains meta-data
information along with the configuration information.

Adding spherical boundary conditions to LD integrator

fixed a bug in rigid bodies that contain directional atoms

OptionSectionParser get compiled

Fix the other branch of force calculation in Torision.cpp

There is still something wrong with inertia tensor projection

fix a bug in creating cutoffGroup. When cutoffGroup is turned off, there is a mismatch between group and center of mass array

fix a bug in projecting the inertia tensor of directional atom in rigibody into rigidbody's body frame

putting spaceballs back where they belong

fix some warning messages

adding exclude pairs for rigidbody and cutoff group

reactionfield get fixed

begin bug fix

add getRigidBodyAt  function into Molecule class

adding wildcard class; NameFinder is finished

(1) complete section parser's error message
(2) add GhostTorsion
(3) accumulate inertial tensor from the directional atoms before calculate rigidbody's inertial tensor

fix a bug in global index

merging new_design branch into OOPSE-2.0

Shapes looks like it's working

change #include "mpi.h" to #include <mpi.h> in AbstractClasses.hpp

change Makefile to relative path

Import of OOPSE v. 2.0