[svn] * Proper SVN structureinit

--HG--
branch : trunk

32 files changed, 9517 insertions, 0 deletions

diff --git a/dep/include/g3dlite/G3D/AABox.h b/dep/include/g3dlite/G3D/AABox.h
new file mode 100644
index 00000000000..b5e862a7721
--- /dev/null
+++ b/dep/include/g3dlite/G3D/AABox.h
@@ -0,0 +1,255 @@
+/**
+  @file AABox.h
+ 
+  Axis-aligned box class
+ 
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+ 
+  @created 2004-01-10
+  @edited  2006-02-10
+
+  Copyright 2000-2006, Morgan McGuire.
+  All rights reserved.
+ */
+
+#ifndef G3D_AABOX_H
+#define G3D_AABOX_H
+
+#include "G3D/platform.h"
+#include "G3D/Vector3.h"
+#include "G3D/debug.h"
+#include "G3D/Array.h"
+
+namespace G3D {
+
+/**
+ An axis-aligned box.
+ */
+class AABox {
+private:
+
+    /** Optional argument placeholder */
+    static int dummy;
+
+    Vector3  lo;
+    Vector3  hi;
+
+public:
+
+    /** Does not initialize the fields */
+    inline AABox() {}
+
+    /**
+     Constructs a zero-area AABox at v.
+     */
+    inline AABox(const Vector3& v) {
+        lo = hi = v;
+    }
+
+    /** Assumes that low is less than or equal to high along each dimension.
+        To have this automatically enforced, use
+        <code>AABox(low.min(high), low.max(high));</code>
+    */
+    inline AABox(const Vector3& low, const Vector3& high) {
+        set(low, high);
+    }
+
+    /** Assumes that low is less than or equal to high along each dimension.
+    */
+    inline void set(const Vector3& low, const Vector3& high) {
+        debugAssert(
+            (low.x <= high.x) &&
+            (low.y <= high.y) &&
+            (low.z <= high.z));
+        lo = low;
+        hi = high;
+    }
+
+
+    inline const Vector3& low() const {
+        return lo;
+    }
+
+    inline const Vector3& high() const {
+        return hi;
+    }
+
+    /**
+     The largest possible finite box.
+     */
+    static inline const AABox& maxFinite() {
+        static const AABox b = AABox(Vector3::minFinite(), Vector3::maxFinite());
+        return b;
+    }
+
+    static inline const AABox& inf() {
+        static const AABox b = AABox(-Vector3::inf(), Vector3::inf());
+        return b;
+    }
+
+    static inline const AABox& zero() {
+        static const AABox b = AABox(Vector3::zero(), Vector3::zero());
+        return b;
+    }
+
+    /**
+      Returns the centroid of the box.
+     */
+    inline Vector3 center() const {
+        return (lo + hi) * 0.5;
+    }
+
+    /**
+     Distance from corner(0) to the next corner along axis a.
+     */
+    inline double extent(int a) const {
+        debugAssert(a < 3);
+        return hi[a] - lo[a];
+    }
+
+    inline Vector3 extent() const {
+        return hi - lo;
+    }
+
+    /**
+	 @deprecated Use culledBy(Array<Plane>&)
+     */
+    bool culledBy(
+        const class Plane*  plane,
+        int                 numPlanes,
+		int32&				cullingPlaneIndex,
+		const uint32  		testMask,
+        uint32&             childMask) const;
+
+    /**
+	 @deprecated Use culledBy(Array<Plane>&)
+     */
+    bool culledBy(
+        const class Plane*  plane,
+        int                 numPlanes,
+		int32&				cullingPlaneIndex = dummy,
+		const uint32  		testMask = 0xFFFFFF) const;
+
+    /**
+     Splits the box into two AABoxes along the specified axis.  low contains
+     the part that was closer to negative infinity along axis, high contains
+     the other part.  Either may have zero volume.
+     */
+    void split(const Vector3::Axis& axis, float location, AABox& low, AABox& high) const;
+
+	/**
+	 Conservative culling test for up to 32 planes.	
+	 Returns true if there exists a <CODE>plane[p]</CODE> for
+     which the entire object is in the negative half space
+     (opposite the plane normal).
+
+	 <CODE>testMask</CODE> and <CODE>childMask</CODE>
+	 are used for optimizing bounding volume hierarchies.
+     The version of this method that produces childMask
+     is slower than the version without; it should only
+     be used for parent nodes.
+
+	 @param cullingPlaneIndex The index of the first plane for which
+	  the entire object is in the negative half-space.  The function
+	  exits early when one plane is found.  -1 when the function
+	  returns false (i.e. when no plane culls the whole object).
+
+	 @param testMask  If bit <I>p</I> is 0, the 
+	   bounding volume automatically passes the culling test for
+	   <CODE>plane[p]</CODE> (i.e. it is known that the volume
+	   is entirely within the positive half space).  The function
+       must return false if testMask is 0 and test all planes
+       when testMask is -1 (0xFFFFFFFF).
+
+     @param childMask Test mask for the children of this volume.
+       
+	 */
+	bool culledBy(
+		const Array<Plane>&		plane,
+		int32&					cullingPlaneIndex,
+		const uint32  			testMask,
+        uint32&                 childMask) const;
+
+    /**
+     Conservative culling test that does not produce a mask for children.
+     */
+	bool culledBy(
+		const Array<Plane>&		plane,
+		int32&					cullingPlaneIndex = dummy,
+		const uint32  			testMask		  = -1) const;
+
+    inline bool contains(
+        const Vector3&      point) const {
+        return
+            (point.x >= lo.x) &&
+            (point.y >= lo.y) &&
+            (point.z >= lo.z) &&
+            (point.x <= hi.x) &&
+            (point.y <= hi.y) &&
+            (point.z <= hi.z);
+    }
+
+    /** @deprecated */
+    inline float surfaceArea() const {
+        Vector3 diag = hi - lo;
+        return 2.0f * (diag.x * diag.y + diag.y * diag.z + diag.x * diag.z);
+    }
+
+    inline float area() const {
+        return surfaceArea();
+    }
+
+    inline float volume() const {
+        Vector3 diag = hi - lo;
+        return diag.x * diag.y * diag.z;
+    }
+
+    Vector3 randomInteriorPoint() const;
+
+    Vector3 randomSurfacePoint() const;
+
+    /** @deprecated use Box constructor */
+    class Box toBox() const;
+
+    /** Returns true if there is any overlap */
+    bool intersects(const AABox& other) const;
+
+    /** Returns true if there is any overlap.
+        @cite Jim Arvo's algorithm from Graphics Gems II*/
+    bool intersects(const class Sphere& other) const;
+
+    /** Return the intersection of the two boxes */
+    AABox intersect(const AABox& other) const {
+        Vector3 H = hi.min(other.hi);
+        Vector3 L = lo.max(other.lo).min(H);
+        return AABox(L, H);
+    }
+
+    inline unsigned int hashCode() const {
+        return lo.hashCode() + hi.hashCode();
+    }
+
+    inline bool operator==(const AABox& b) const {
+        return (lo == b.lo) && (hi == b.hi);
+    }
+
+    inline bool operator!=(const AABox& b) const {
+        return !((lo == b.lo) && (hi == b.hi));
+    }
+
+    void getBounds(AABox& out) const {
+        out = *this;
+    }
+};
+
+}
+
+/**
+ Hashing function for use with Table.
+ */
+inline unsigned int hashCode(const G3D::AABox& b) {
+	return b.hashCode();
+}
+
+
+#endif
diff --git a/dep/include/g3dlite/G3D/Array.h b/dep/include/g3dlite/G3D/Array.h
new file mode 100644
index 00000000000..290563b6719
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Array.h
@@ -0,0 +1,1156 @@
+/** 
+  @file Array.h
+ 
+  @maintainer Morgan McGuire, graphics3d.com
+  @cite Portions written by Aaron Orenstein, a@orenstein.name
+ 
+  @created 2001-03-11
+  @edited  2007-05-12
+
+  Copyright 2000-2007, Morgan McGuire.
+  All rights reserved.
+ */
+
+#ifndef G3D_ARRAY_H
+#define G3D_ARRAY_H
+
+#include "G3D/platform.h"
+#include "G3D/debug.h"
+#include "G3D/System.h"
+#ifdef G3D_DEBUG
+//   For formatting error messages
+#    include "G3D/format.h"
+#endif
+#include <vector>
+#include <algorithm>
+
+#ifdef G3D_WIN32
+#   include <new>
+    
+#   pragma warning (push)
+    // debug information too long
+#   pragma warning( disable : 4312)
+#   pragma warning( disable : 4786)
+#endif
+
+
+namespace G3D {
+
+/**
+ Constant for passing to Array::resize
+ */
+const bool DONT_SHRINK_UNDERLYING_ARRAY = false;
+
+/** Constant for Array::sort */
+const int SORT_INCREASING = 1;
+/** Constant for Array::sort */
+const int SORT_DECREASING = -1;
+
+/**
+ Dynamic 1D array.  
+
+ Objects must have a default constructor (constructor that
+ takes no arguments) in order to be used with this template.
+ You will get the error "no appropriate default constructor found"
+ if they do not.
+
+ Do not use with objects that overload placement <code>operator new</code>,
+ since the speed of Array is partly due to pooled allocation.
+
+ If SSE is defined Arrays allocate the first element aligned to
+ 16 bytes.
+
+
+ Array is highly optimized compared to std::vector.  
+ Array operations are less expensive than on std::vector and for large
+ amounts of data, Array consumes only 1.5x the total size of the 
+ data, while std::vector consumes 2.0x.  The default
+ array takes up zero heap space.  The first resize (or append)
+ operation grows it to a reasonable internal size so it is efficient
+ to append to small arrays.  Memory is allocated using
+ System::alignedMalloc, which produces pointers aligned to 16-byte
+ boundaries for use with SSE instructions and uses pooled storage for
+ fast allocation.  When Array needs to copy
+ data internally on a resize operation it correctly invokes copy
+ constructors of the elements (the MSVC6 implementation of
+ std::vector uses realloc, which can create memory leaks for classes
+ containing references and pointers).  Array provides a guaranteed
+ safe way to access the underlying data as a flat C array --
+ Array::getCArray.  Although (T*)std::vector::begin() can be used for
+ this purpose, it is not guaranteed to succeed on all platforms.
+
+ To serialize an array, see G3D::serialize.
+
+ Do not subclass an Array.
+ */
+template <class T>
+class Array {
+private:
+    /** 0...num-1 are initialized elements, num...numAllocated-1 are not */
+    T*              data;
+
+    int             num;
+    int             numAllocated;
+
+    void init(int n, int a) {
+        debugAssert(n <= a);
+        debugAssert(n >= 0);
+        this->num = 0;
+        this->numAllocated = 0;
+        data = NULL;
+        if (a > 0) {
+            resize(n);
+        } else {
+            data = NULL;
+        }
+    }
+
+    void _copy(const Array &other) {
+        init(other.num, other.num);
+        for (int i = 0; i < num; i++) {
+            data[i] = other.data[i];
+        }
+    }
+
+    /**
+     Returns true iff address points to an element of this array.
+     Used by append.
+     */
+    inline bool inArray(const T* address) {
+        return (address >= data) && (address < data + num);
+    }
+
+
+    /** Only compiled if you use the sort procedure. */
+    static bool __cdecl compareGT(const T& a, const T& b) {
+        return a > b;
+    }
+
+
+    /**
+     Allocates a new array of size numAllocated (not a parameter to the method) 
+     and then copies at most oldNum elements from the old array to it.  Destructors are
+     called for oldNum elements of the old array.
+     */
+    void realloc(int oldNum) {
+         T* oldData = data;
+         
+         // The allocation is separate from the constructor invocation because we don't want 
+         // to pay for the cost of constructors until the newly allocated
+         // elements are actually revealed to the application.  They 
+         // will be constructed in the resize() method.
+
+         data = (T*)System::alignedMalloc(sizeof(T) * numAllocated, 16);
+
+         // Call the copy constructors
+         {const int N = iMin(oldNum, numAllocated);
+          const T* end = data + N;
+          T* oldPtr = oldData;
+          for (T* ptr = data; ptr < end; ++ptr, ++oldPtr) {
+
+             // Use placement new to invoke the constructor at the location
+             // that we determined.  Use the copy constructor to make the assignment.
+             const T* constructed = new (ptr) T(*oldPtr);
+
+             (void)constructed;
+             debugAssertM(constructed == ptr, 
+                 "new returned a different address than the one provided by Array.");
+         }}
+
+         // Call destructors on the old array (if there is no destructor, this will compile away)
+         {const T* end = oldData + oldNum;
+          for (T* ptr = oldData; ptr < end; ++ptr) {
+              ptr->~T();
+         }}
+
+
+         System::alignedFree(oldData);
+    }
+
+public:
+
+    /**
+     C++ STL style iterator variable.  Call begin() to get 
+     the first iterator, pre-increment (++i) the iterator to get to
+     the next value.  Use dereference (*i) to access the element.
+     */
+    typedef T* Iterator;
+    typedef const T* ConstIterator;
+
+    /**
+     C++ STL style iterator method.  Returns the first iterator element.
+     Do not change the size of the array while iterating.
+     */
+    Iterator begin() {
+        return data;
+    }
+
+    ConstIterator begin() const {
+        return data;
+    }
+    /**
+     C++ STL style iterator method.  Returns one after the last iterator
+     element.
+     */
+    ConstIterator end() const {
+        return data + num;
+    }
+
+    Iterator end() {
+        return data + num;
+    }
+
+   /**
+    The array returned is only valid until the next append() or resize call, or 
+	the Array is deallocated.
+    */
+   T* getCArray() {
+       return data;
+   }
+
+   /**
+    The array returned is only valid until the next append() or resize call, or 
+	the Array is deallocated.
+    */
+   const T* getCArray() const {
+       return data;
+   }
+
+   /** Creates a zero length array (no heap allocation occurs until resize). */
+   Array() {
+       init(0, 0);
+   }
+
+   /**
+    Creates an array of size.
+    */
+   Array(int size) {
+       init(size, size);
+   }
+
+   /**
+    Copy constructor
+    */
+   Array(const Array& other) {
+       _copy(other);
+   }
+
+   /**
+    Destructor does not delete() the objects if T is a pointer type
+    (e.g. T = int*) instead, it deletes the <B>pointers themselves</B> and 
+    leaves the objects.  Call deleteAll if you want to dealocate
+    the objects referenced.  Do not call deleteAll if <CODE>T</CODE> is not a pointer
+    type (e.g. do call Array<Foo*>::deleteAll, do <B>not</B> call Array<Foo>::deleteAll).
+    */
+   ~Array() {
+       // Invoke the destructors on the elements
+       for (int i = 0; i < num; i++) {
+           (data + i)->~T();
+       }
+       
+       System::alignedFree(data);
+       // Set to 0 in case this Array is global and gets referenced during app exit
+       data = NULL;
+	   num = 0;
+       numAllocated = 0;
+   }
+
+
+   /**
+    Removes all elements.  Use resize(0, false) or fastClear if you want to 
+    remove all elements without deallocating the underlying array
+    so that future append() calls will be faster.
+    */
+   void clear() {
+       resize(0);
+   }
+
+   /** resize(0, false) */
+   void fastClear() {
+       resize(0, false);
+   }
+
+   /**
+    Assignment operator.
+    */
+   Array& operator=(const Array& other) {
+       resize(other.num);
+       for (int i = 0; i < num; ++i) {
+           data[i] = other[i];
+       }
+       return *this;
+   }
+
+   Array& operator=(const std::vector<T>& other) {
+       resize((int)other.size());
+       for (int i = 0; i < num; ++i) {
+           data[i] = other[i];
+       }
+       return *this;
+   }
+
+   /**
+    Number of elements in the array.
+    */
+   inline int size() const {
+      return num;
+   }
+
+   /**
+    Number of elements in the array.  (Same as size; this is just
+    here for convenience).
+    */
+   inline int length() const {
+      return size();
+   }
+
+   /**
+    Swaps element index with the last element in the array then
+    shrinks the array by one.
+    */
+   void fastRemove(int index) {
+       debugAssert(index >= 0);
+       debugAssert(index < num);
+       data[index] = data[num - 1];
+       resize(size() - 1);
+   }
+
+   /**
+    Resizes, calling the default constructor for 
+    newly created objects and shrinking the underlying
+    array as needed (and calling destructors as needed).
+    */
+   void resize(int n) {
+      resize(n, true);
+   }
+
+   /** Resizes without shrinking the underlying array */
+   void fastResize(int n) {
+      resize(n, false);
+   }
+
+
+   /**
+    Inserts at the specified index and shifts all other elements up by one.
+    */
+   void insert(int n, const T& value) {
+       // Add space for the extra element
+       resize(num + 1, false);
+
+       for (int i = num - 1; i > n; --i) {
+           data[i] = data[i - 1];
+       }
+       data[n] = value;
+   }
+
+    /** @param shrinkIfNecessary if false, memory will never be
+      reallocated when the array shrinks.  This makes resizing much
+      faster but can waste memory. */
+   void resize(int n, bool shrinkIfNecessary) {
+      int oldNum = num;
+      num = n;
+
+      // Call the destructors on newly hidden elements if there are any
+      for (int i = num; i < oldNum; ++i) {
+          (data + i)->~T();
+      }
+
+      // Once allocated, always maintain 10 elements or 32 bytes, whichever is higher.
+      static const int minSize = iMax(10, 32 / sizeof(T));
+
+      if (num > numAllocated) {
+          // Grow the underlying array
+
+          if (numAllocated == 0) {
+              // First allocation; grow to exactly the size requested to avoid wasting space.
+              numAllocated = n;
+              debugAssert(oldNum == 0);
+              realloc(oldNum);
+          } else {
+         
+              if (num < minSize) {
+                  // Grow to at least the minimum size
+                  numAllocated = minSize;
+
+              } else {
+
+                  // Increase the underlying size of the array.  Grow aggressively
+                  // up to 64k, less aggressively up to 400k, and then grow relatively
+                  // slowly (1.5x per resize) to avoid excessive space consumption.
+                  //
+                  // These numbers are tweaked according to performance tests.
+
+                  float growFactor = 3.0;
+
+                  size_t oldSizeBytes = numAllocated * sizeof(T);
+                  if (oldSizeBytes > 400000) {
+                      // Avoid bloat
+                      growFactor = 1.5;
+                  } else if (oldSizeBytes > 64000) {
+                      // This is what std:: uses at all times
+                      growFactor = 2.0;
+                  }
+
+                  numAllocated = (num - numAllocated) + (int)(numAllocated * growFactor);
+
+                  if (numAllocated < minSize) {
+                      numAllocated = minSize;
+                  }
+              }
+
+              realloc(oldNum);
+          }
+
+      } else if ((num <= numAllocated / 3) && shrinkIfNecessary && (num > minSize)) {
+          // Shrink the underlying array
+
+          // Only copy over old elements that still remain after resizing
+          // (destructors were called for others if we're shrinking)
+          realloc(iMin(num, oldNum));
+
+      }
+
+      // Call the constructors on newly revealed elements.
+      // Do not use parens because we don't want the intializer
+      // invoked for POD types.
+      for (int i = oldNum; i < num; ++i) {
+          new (data + i) T;
+      }
+   }
+
+    /**
+     Add an element to the end of the array.  Will not shrink the underlying array
+     under any circumstances.  It is safe to append an element that is already
+     in the array.
+     */
+    inline void append(const T& value) {
+        
+        if (num < numAllocated) {
+            // This is a simple situation; just stick it in the next free slot using
+            // the copy constructor.
+            new (data + num) T(value);
+            ++num;
+        } else if (inArray(&value)) {
+            // The value was in the original array; resizing
+            // is dangerous because it may move the value
+            // we have a reference to.
+            T tmp = value;
+            append(tmp);
+        } else {
+            // Here we run the empty initializer where we don't have to, but
+            // this simplifies the computation.
+            resize(num + 1, DONT_SHRINK_UNDERLYING_ARRAY);
+            data[num - 1] = value;
+        }
+    }
+
+
+    inline void append(const T& v1, const T& v2) {
+        if (inArray(&v1) || inArray(&v2)) {
+            T t1 = v1;
+            T t2 = v2;
+            append(t1, t2);
+        } else if (num + 1 < numAllocated) {
+            // This is a simple situation; just stick it in the next free slot using
+            // the copy constructor.
+            new (data + num) T(v1);
+            new (data + num + 1) T(v2);
+            num += 2;
+        } else {
+            resize(num + 2, DONT_SHRINK_UNDERLYING_ARRAY);
+            data[num - 2] = v1;
+            data[num - 1] = v2;
+        }
+    }
+
+
+    inline void append(const T& v1, const T& v2, const T& v3) {
+        if (inArray(&v1) || inArray(&v2) || inArray(&v3)) {
+            T t1 = v1;
+            T t2 = v2;
+            T t3 = v3;
+            append(t1, t2, t3);
+        } else if (num + 2 < numAllocated) {
+            // This is a simple situation; just stick it in the next free slot using
+            // the copy constructor.
+            new (data + num) T(v1);
+            new (data + num + 1) T(v2);
+            new (data + num + 2) T(v3);
+            num += 3;
+        } else {
+            resize(num + 3, DONT_SHRINK_UNDERLYING_ARRAY);
+            data[num - 3] = v1;
+            data[num - 2] = v2;
+            data[num - 1] = v3;
+        }
+    }
+
+
+    inline void append(const T& v1, const T& v2, const T& v3, const T& v4) {
+        if (inArray(&v1) || inArray(&v2) || inArray(&v3) || inArray(&v4)) {
+            T t1 = v1;
+            T t2 = v2;
+            T t3 = v3;
+            T t4 = v4;
+            append(t1, t2, t3, t4);
+        } else if (num + 3 < numAllocated) {
+            // This is a simple situation; just stick it in the next free slot using
+            // the copy constructor.
+            new (data + num) T(v1);
+            new (data + num + 1) T(v2);
+            new (data + num + 2) T(v3);
+            new (data + num + 3) T(v4);
+            num += 4;
+        } else {
+            resize(num + 4, DONT_SHRINK_UNDERLYING_ARRAY);
+            data[num - 4] = v1;
+            data[num - 3] = v2;
+            data[num - 2] = v3;
+            data[num - 1] = v4;
+        }
+    }
+
+    /**
+     Returns true if the given element is in the array.
+     */
+    bool contains(const T& e) const {
+        for (int i = 0; i < size(); ++i) {
+            if ((*this)[i] == e) {
+                return true;
+            }
+        }
+
+        return false;
+    }
+
+   /**
+    Append the elements of array.  Cannot be called with this array
+    as an argument.
+    */
+   void append(const Array<T>& array) {
+       debugAssert(this != &array);
+       int oldNum = num;
+       int arrayLength = array.length();
+
+       resize(num + arrayLength, false);
+
+       for (int i = 0; i < arrayLength; i++) {
+           data[oldNum + i] = array.data[i];
+       }
+   }
+
+   /**
+    Pushes a new element onto the end and returns its address.
+    This is the same as A.resize(A.size() + 1, false); A.last()
+    */
+   inline T& next() {
+       resize(num + 1, false);
+       return last();
+   }
+
+   /**
+    Pushes an element onto the end (appends)
+    */
+   inline void push(const T& value) {
+       append(value);
+   }
+
+   inline void push(const Array<T>& array) {
+       append(array);
+   }
+
+   /** Alias to provide std::vector compatibility */
+   inline void push_back(const T& v) {
+       push(v);
+   }
+
+   /** "The member function removes the last element of the controlled sequence, which must be non-empty."
+        For compatibility with std::vector. */
+   inline void pop_back() {
+       pop();
+   }
+
+   /** 
+      "The member function returns the storage currently allocated to hold the controlled
+       sequence, a value at least as large as size()" 
+       For compatibility with std::vector.
+   */
+   int capacity() const {
+       return numAllocated;
+   }
+
+   /** 
+      "The member function returns a reference to the first element of the controlled sequence, 
+       which must be non-empty." 
+       For compatibility with std::vector.
+   */
+   T& front() {
+       return (*this)[0];
+   }
+
+   /** 
+      "The member function returns a reference to the first element of the controlled sequence, 
+       which must be non-empty." 
+       For compatibility with std::vector.
+   */
+   const T& front() const {
+       return (*this)[0];
+   }
+
+   /**
+    Removes the last element and returns it.  By default, shrinks the underlying array.
+    */
+   inline T pop(bool shrinkUnderlyingArrayIfNecessary = true) {
+       debugAssert(num > 0);
+       T temp = data[num - 1];
+       resize(num - 1, shrinkUnderlyingArrayIfNecessary);
+       return temp;
+   }
+
+   /** Pops the last element and discards it without returning anything.  Faster than pop.
+      By default, does not shrink the underlying array.*/
+   inline void popDiscard(bool shrinkUnderlyingArrayIfNecessary = false) {
+       debugAssert(num > 0);
+       resize(num - 1, shrinkUnderlyingArrayIfNecessary);
+   }
+
+
+   /**
+    "The member function swaps the controlled sequences between *this and str."
+    Note that this is slower than the optimal std implementation.
+
+    For compatibility with std::vector.
+    */
+   void swap(Array<T>& str) {
+       Array<T> temp = str;
+       str = *this;
+       *this = temp;
+   }
+
+
+   /**
+    Performs bounds checks in debug mode
+    */
+   inline T& operator[](int n) {
+        debugAssertM((n >= 0) && (n < num), format("Array index out of bounds. n = %d, size() = %d", n, num));
+        debugAssert(data!=NULL);
+        return data[n];
+   }
+
+   inline T& operator[](unsigned int n) {
+        debugAssertM(((int)n < num), format("Array index out of bounds. n = %d, size() = %d", n, num));
+        return data[n];
+   }
+
+   /**
+    Performs bounds checks in debug mode
+    */
+    inline const T& operator[](int n) const {
+        debugAssert((n >= 0) && (n < num));
+        debugAssert(data!=NULL);
+        return data[n];
+    }
+
+    inline const T& operator[](unsigned int n) const {
+        debugAssert((n < (unsigned int)num));
+        debugAssert(data!=NULL);
+        return data[n];
+    }
+
+    inline T& randomElement() {
+        debugAssert(num > 0);
+        debugAssert(data!=NULL);
+        return data[iRandom(0, num - 1)];
+    }
+
+    inline const T& randomElement() const {
+        debugAssert(num > 0);
+        debugAssert(data!=NULL);
+        return data[iRandom(0, num - 1)];
+    }
+
+    /**
+    Returns the last element, performing a check in
+    debug mode that there is at least one element.
+    */
+    inline const T& last() const {
+        debugAssert(num > 0);
+        debugAssert(data!=NULL);
+        return data[num - 1];
+    }
+
+    /** Returns element lastIndex() */
+    inline T& last() {
+        debugAssert(num > 0);
+        debugAssert(data!=NULL);
+        return data[num - 1];
+    }
+
+    /** Returns <i>size() - 1</i> */
+    inline int lastIndex() const {
+        debugAssertM(num > 0, "Array is empty");
+        return num - 1;
+    }
+
+    inline int firstIndex() const {
+        debugAssertM(num > 0, "Array is empty");
+        return 0;
+    }
+
+    /** Returns element firstIndex(), performing a check in debug mode to ensure that there is at least one */
+    inline T& first() {
+        debugAssertM(num > 0, "Array is empty");
+        return data[0];
+    }
+
+    inline const T& first() const {
+        debugAssertM(num > 0, "Array is empty");
+        return data[0];
+    }
+
+    /** Returns iFloor(size() / 2), throws an assertion in debug mode if the array is empty */
+    inline int middleIndex() const {
+        debugAssertM(num > 0, "Array is empty");
+        return num >> 1;
+    }
+
+    /** Returns element middleIndex() */
+    inline const T& middle() const {
+        debugAssertM(num > 0, "Array is empty");
+        return data[num >> 1];   
+    }
+
+    /** Returns element middleIndex() */
+    inline T& middle() {
+        debugAssertM(num > 0, "Array is empty");
+        return data[num >> 1];   
+    }
+
+    /**
+    Calls delete on all objects[0...size-1]
+    and sets the size to zero.
+    */
+    void deleteAll() {
+        for (int i = 0; i < num; i++) {
+            delete data[i];
+        }
+        resize(0);
+    }
+
+    /**
+     Returns the index of (the first occurance of) an index or -1 if
+     not found.
+     */
+    int findIndex(const T& value) const {
+        for (int i = 0; i < num; ++i) {
+            if (data[i] == value) {
+                return i;
+            }
+        }
+        return -1;
+    }
+
+    /**
+     Finds an element and returns the iterator to it.  If the element
+     isn't found then returns end().
+     */
+    Iterator find(const T& value) {
+        for (int i = 0; i < num; ++i) {
+            if (data[i] == value) {
+                return data + i;
+            }
+        }
+        return end();
+    }
+
+    ConstIterator find(const T& value) const {
+        for (int i = 0; i < num; ++i) {
+            if (data[i] == value) {
+                return data + i;
+            }
+        }
+        return end();
+    }
+
+    /**
+     Removes count elements from the array
+     referenced either by index or Iterator.
+     */
+    void remove(Iterator element, int count = 1) {
+        debugAssert((element >= begin()) && (element < end()));
+        debugAssert((count > 0) && (element + count) <= end());
+        Iterator last = end() - count;
+
+        while(element < last) {
+            element[0] = element[count];
+            ++element;
+        }
+        
+        resize(num - count);
+    }
+
+    void remove(int index, int count = 1) {
+        debugAssert((index >= 0) && (index < num));
+        debugAssert((count > 0) && (index + count <= num));
+        
+        remove(begin() + index, count);
+    }
+
+    /**
+     Reverse the elements of the array in place.
+     */
+    void reverse() {
+        T temp;
+        
+        int n2 = num / 2;
+        for (int i = 0; i < n2; ++i) {
+            temp = data[num - 1 - i];
+            data[num - 1 - i] = data[i];
+            data[i] = temp;
+        }
+    }
+
+    /**
+     Sort using a specific less-than function, e.g.:
+
+  <PRE>
+    bool __cdecl myLT(const MyClass& elem1, const MyClass& elem2) {
+        return elem1.x < elem2.x;
+    }
+    </PRE>
+
+  Note that for pointer arrays, the <CODE>const</CODE> must come 
+  <I>after</I> the class name, e.g., <CODE>Array<MyClass*></CODE> uses:
+
+  <PRE>
+    bool __cdecl myLT(MyClass*const& elem1, MyClass*const& elem2) {
+        return elem1->x < elem2->x;
+    }
+    </PRE>
+     */
+    void sort(bool (__cdecl *lessThan)(const T& elem1, const T& elem2)) {
+        std::sort(data, data + num, lessThan);
+    }
+
+
+    /**
+     Sorts the array in increasing order using the > or < operator.  To 
+     invoke this method on Array<T>, T must override those operator.
+     You can overide these operators as follows:
+     <code>
+        bool T::operator>(const T& other) const {
+           return ...;
+        }
+        bool T::operator<(const T& other) const {
+           return ...;
+        }
+     </code>
+     */
+    void sort(int direction = SORT_INCREASING) {
+        if (direction == SORT_INCREASING) {
+            std::sort(data, data + num);
+        } else {
+            std::sort(data, data + num, compareGT);
+        }
+    }
+
+    /**
+     Sorts elements beginIndex through and including endIndex.
+     */
+    void sortSubArray(int beginIndex, int endIndex, int direction = SORT_INCREASING) {
+        if (direction == SORT_INCREASING) {
+            std::sort(data + beginIndex, data + endIndex + 1);
+        } else {
+            std::sort(data + beginIndex, data + endIndex + 1, compareGT);
+        }
+    }
+
+    void sortSubArray(int beginIndex, int endIndex, bool (__cdecl *lessThan)(const T& elem1, const T& elem2)) {
+        std::sort(data + beginIndex, data + endIndex + 1, lessThan);
+    }
+
+    /**
+     The StrictWeakOrdering can be either a class that overloads the function call operator() or
+     a function pointer of the form <code>bool (__cdecl *lessThan)(const T& elem1, const T& elem2)</code>
+     */
+    template<typename StrictWeakOrdering>
+    void sortSubArray(int beginIndex, int endIndex, StrictWeakOrdering& lessThan) {
+        std::sort(data + beginIndex, data + endIndex + 1, lessThan);
+    }
+
+    /** Uses < and == to evaluate operator(); this is the default comparator for Array::partition. */
+    class DefaultComparator {
+    public:
+        inline int operator()(const T& A, const T& B) const {
+            if (A < B) {
+                return 1;
+            } else if (A == B) {
+                return 0;
+            } else {
+                return -1;
+            }
+        }
+    };
+
+    /** The output arrays are resized with fastClear() so that if they are already of the same size
+        as this array no memory is allocated during partitioning. 
+        
+        @param comparator A function, or class instance with an overloaded operator() that compares
+        two elements of type <code>T</code> and returns 0 if they are equal, -1 if the second is smaller,
+        and 1 if the first is smaller (i.e., following the conventions of std::string::compare).  For example:
+
+        <pre>
+        int compare(int A, int B) {
+            if (A < B) {
+                return 1;
+            } else if (A == B) {
+                return 0;
+            } else {
+                return -1;
+            }
+        }
+        </pre>
+        */
+    template<typename Comparator>
+    void partition(
+        const T& partitionElement, 
+        Array<T>& ltArray,
+        Array<T>& eqArray,
+        Array<T>& gtArray,
+        const Comparator& comparator) const {
+
+        // Make sure all arrays are independent
+        debugAssert(&ltArray != this);
+        debugAssert(&eqArray != this);
+        debugAssert(&gtArray != this);
+        debugAssert(&ltArray != &eqArray);
+        debugAssert(&ltArray != &gtArray);
+        debugAssert(&eqArray != &gtArray);
+
+        // Clear the arrays
+        ltArray.fastClear();
+        eqArray.fastClear();
+        gtArray.fastClear();
+
+        // Form a table of buckets for lt, eq, and gt
+        Array<T>* bucket[3] = {&ltArray, &eqArray, &gtArray};
+
+        for (int i = 0; i < num; ++i) {
+            int c = comparator(partitionElement, data[i]);
+            debugAssertM(c >= -1 && c <= 1, "Comparator returned an illegal value.");
+
+            // Insert into the correct bucket, 0, 1, or 2
+            bucket[c + 1]->append(data[i]);
+        }
+    }
+
+    /**
+      Uses < and == on elements to perform a partition.  See partition().
+     */
+    void partition(
+        const T& partitionElement, 
+        Array<T>& ltArray,
+        Array<T>& eqArray,
+        Array<T>& gtArray) const {
+
+        partition(partitionElement, ltArray, eqArray, gtArray, typename Array<T>::DefaultComparator());
+    }
+
+    /** 
+     Paritions the array into those below the median, those above the median, and those elements
+     equal to the median in expected O(n) time using quickselect.  If the array has an even
+     number of different elements, the median for partition purposes is the largest value
+     less than the median.
+
+     @param tempArray used for working scratch space
+     @param comparator see parition() for a discussion.*/
+    template<typename Comparator>
+    void medianPartition(
+        Array<T>&           ltMedian, 
+        Array<T>&           eqMedian, 
+        Array<T>&           gtMedian,
+        Array<T>&           tempArray,
+        const Comparator&   comparator) const {
+
+        ltMedian.fastClear();
+        eqMedian.fastClear();
+        gtMedian.fastClear();
+
+        // Handle trivial cases first
+        switch (size()) {
+        case 0:
+            // Array is empty; no parition is possible
+            return;
+
+        case 1:
+            // One element
+            eqMedian.append(first());
+            return;
+
+        case 2:
+            {
+                // Two element array; median is the smaller
+                int c = comparator(first(), last());
+                
+                switch (c) {
+                case -1:
+                    // first was bigger
+                    eqMedian.append(last());
+                    gtMedian.append(first());
+                    break;
+
+                case 0:
+                    // Both equal to the median
+                    eqMedian.append(first(), last());
+                    break;
+
+                case 1:
+                    // Last was bigger
+                    eqMedian.append(first());
+                    gtMedian.append(last());
+                    break;
+                }
+            }
+            return;
+        }
+
+        // All other cases use a recursive randomized median
+
+        // Number of values less than all in the current arrays        
+        int ltBoost = 0;
+
+        // Number of values greater than all in the current arrays        
+        int gtBoost = 0;
+
+        // For even length arrays, force the gt array to be one larger than the
+        // lt array:  
+        //  [1 2 3] size = 3, choose half = (s + 1) /2
+        //
+        int lowerHalfSize, upperHalfSize;
+        if (isEven(size())) {
+            lowerHalfSize = size() / 2;
+            upperHalfSize = lowerHalfSize + 1;
+        } else {
+            lowerHalfSize = upperHalfSize = (size() + 1) / 2;
+        }
+        const T* xPtr = NULL;
+
+        // Maintain pointers to the arrays; we'll switch these around during sorting
+        // to avoid copies.
+        const Array<T>* source = this;
+        Array<T>* lt     = &ltMedian;
+        Array<T>* eq     = &eqMedian;
+        Array<T>* gt     = &gtMedian;
+        Array<T>* extra  = &tempArray;
+
+        while (true) {
+            // Choose a random element -- choose the middle element; this is theoretically
+            // suboptimal, but for loosly sorted array is actually the best strategy
+
+            xPtr = &(source->middle());
+            if (source->size() == 1) {
+                // Done; there's only one element left
+                break;
+            }
+            const T& x = *xPtr;
+
+            // Note: partition (fast) clears the arrays for us
+            source->partition(x, *lt, *eq, *gt, comparator);
+
+            int L = lt->size() + ltBoost + eq->size();
+            int U = gt->size() + gtBoost + eq->size();
+            if ((L >= lowerHalfSize) &&
+                (U >= upperHalfSize)) {
+
+                // x must be the partition median                    
+                break;
+
+            } else if (L < lowerHalfSize) {
+
+                // x must be smaller than the median.  Recurse into the 'gt' array.
+                ltBoost += lt->size() + eq->size();
+
+                // The new gt array will be the old source array, unless
+                // that was the this pointer (i.e., unless we are on the 
+                // first iteration)
+                Array<T>* newGt = (source == this) ? extra : const_cast<Array<T>*>(source);
+                
+                // Now set up the gt array as the new source
+                source = gt;
+                gt = newGt;
+
+            } else {
+
+                // x must be bigger than the median.  Recurse into the 'lt' array.
+                gtBoost += gt->size() + eq->size();
+
+                // The new lt array will be the old source array, unless
+                // that was the this pointer (i.e., unless we are on the 
+                // first iteration)
+                Array<T>* newLt = (source == this) ? extra : const_cast<Array<T>*>(source);
+                
+                // Now set up the lt array as the new source
+                source = lt;
+                lt = newLt;
+            }
+        }
+
+        // Now that we know the median, make a copy of it (since we're about to destroy the array that it
+        // points into).
+        T median = *xPtr;
+        xPtr = NULL;
+
+        // Partition the original array (note that this fast clears for us)
+        partition(median, ltMedian, eqMedian, gtMedian, comparator);
+    }
+
+    /**
+      Computes a median partition using the default comparator and a dynamically allocated temporary 
+      working array.  If the median is not in the array, it is chosen to be the largest value smaller
+      than the true median.
+     */
+    void medianPartition(
+        Array<T>&           ltMedian, 
+        Array<T>&           eqMedian, 
+        Array<T>&           gtMedian) const {
+
+        Array<T> temp;
+        medianPartition(ltMedian, eqMedian, gtMedian, temp, DefaultComparator());
+    }
+
+
+    /** Redistributes the elements so that the new order is statistically independent
+        of the original order. O(n) time.*/
+    void randomize() {
+        T temp;
+
+        for (int i = size() - 1; i >= 0; --i) {
+            int x = iRandom(0, i);
+
+            temp = data[i];
+            data[i] = data[x];
+            data[x] = temp;
+        }
+    }
+
+
+};
+
+
+/** Array::contains for C-arrays */
+template<class T> bool contains(const T* array, int len, const T& e) {
+    for (int i = len - 1; i >= 0; --i) {
+        if (array[i] == e) {
+            return true;
+        }
+    }
+    return false;
+}
+
+} // namespace
+
+#endif
+
+#ifdef G3D_WIN32
+#   pragma warning (push)
+#endif
diff --git a/dep/include/g3dlite/G3D/Box.h b/dep/include/g3dlite/G3D/Box.h
new file mode 100644
index 00000000000..f097c10e18e
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Box.h
@@ -0,0 +1,228 @@
+/**
+  @file Box.h
+ 
+  Box class
+ 
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+ 
+  @cite Portions based on Dave Eberly's Magic Software Library at <A HREF="http://www.magic-software.com">http://www.magic-software.com</A>
+  @created 2001-06-02
+  @edited  2006-01-05
+
+  Copyright 2000-2006, Morgan McGuire.
+  All rights reserved.
+ */
+
+#ifndef G3D_BOX_H
+#define G3D_BOX_H
+
+#include "G3D/platform.h"
+#include "G3D/Vector3.h"
+#include "G3D/Array.h"
+#include "G3D/Plane.h"
+
+namespace G3D {
+
+class CoordinateFrame;
+
+/**
+ An arbitrary 3D box, useful as a bounding box. 
+
+
+  To construct a box from a coordinate frame, center and extent, use the idiom:
+
+ <CODE>Box box = cframe.toObjectSpace(Box(center - extent/2, center + extent/2));</CODE>
+ */
+class Box {
+private:
+
+    static int32 dummy;
+
+    friend class CoordinateFrame;
+
+    /**
+      <PRE>
+       3    2       7    6
+    
+       0    1       4    5
+
+       front    back (seen through front)
+      </PRE>
+     */
+    Vector3 _corner[8];
+
+    /**
+     Unit axes.
+     */
+    Vector3 _axis[3];
+   
+    Vector3 _center;
+
+    /**
+     Extent along each axis.
+     */
+    Vector3 _extent;
+
+    float  _area;
+    float  _volume;
+
+    void init(
+        const Vector3& min,
+        const Vector3& max);
+
+public:
+
+    /**
+     Does not initialize the fields.
+     */
+    Box();
+
+    /**
+      Constructs a box from two opposite corners.
+     */
+    Box(
+        const Vector3&      min,
+        const Vector3&      max);
+
+    Box(const class AABox& b);
+
+
+    /**
+     Returns the object to world transformation for 
+     this box.  localFrame().worldToObject(...) takes
+     objects into the space where the box axes are
+     (1,0,0), (0,1,0), (0,0,1).  Note that there
+     is no scaling in this transformation.
+     */
+    CoordinateFrame localFrame() const;
+
+    void getLocalFrame(CoordinateFrame& frame) const;
+
+    /**
+      Returns the centroid of the box.
+     */
+    inline Vector3 center() const {
+        return _center;
+    }
+
+    inline Vector3 getCenter() const {
+        return center();
+    }
+
+    /**
+     Returns a corner (0 <= i < 8)
+     @deprecated
+     */
+    inline Vector3 getCorner(int i) const {
+        debugAssert(i < 8);
+        return _corner[i];
+    }
+
+    inline Vector3 corner(int i) const {
+        debugAssert(i < 8);
+        return _corner[i];
+    }
+
+    /**
+     Unit length.
+     */
+    inline Vector3 axis(int a) const {
+        debugAssert(a < 3);
+        return _axis[a];
+    }
+
+    /**
+     Distance from corner(0) to the next corner
+     along the box's local axis a.
+     */
+    inline float extent(int a) const {
+        debugAssert(a < 3);
+        return (float)_extent[a];
+    }
+
+    inline Vector3 extent() const {
+        return _extent;
+    }
+
+    /**
+     Returns the four corners of a face (0 <= f < 6).
+     The corners are returned to form a counter clockwise quad facing outwards.
+     */
+    void getFaceCorners(
+        int                 f,
+        Vector3&            v0,
+        Vector3&            v1,
+        Vector3&            v2,
+        Vector3&            v3) const;
+
+/**
+	 @deprecated Use culledBy(Array<Plane>&)
+     */
+    bool culledBy(
+        const class Plane*  plane,
+        int                 numPlanes,
+	int32&              cullingPlaneIndex,
+	const uint32        testMask,
+        uint32&             childMask) const;
+
+    /**
+	 @deprecated Use culledBy(Array<Plane>&)
+     */
+    bool culledBy(
+        const class Plane*  plane,
+        int                 numPlanes,
+	int32&		cullingPlaneIndex = dummy,
+	const uint32  	testMask = -1) const;
+
+	/**
+      See AABox::culledBy
+	 */
+	bool culledBy(
+		const Array<Plane>&		plane,
+		int32&					cullingPlaneIndex,
+		const uint32  			testMask,
+        uint32&                 childMask) const;
+
+    /**
+     Conservative culling test that does not produce a mask for children.
+     */
+	bool culledBy(
+		const Array<Plane>&		plane,
+		int32&					cullingPlaneIndex = dummy,
+		const uint32  			testMask		  = -1) const;
+
+    bool contains(
+        const Vector3&      point) const;
+
+    /** @deprecated */
+    float surfaceArea() const;
+
+    inline float area() const {
+        return surfaceArea();
+    }
+
+    float volume() const;
+
+    void getRandomSurfacePoint(Vector3& P, Vector3& N = Vector3::dummy) const;
+
+    /**
+      @deprecated
+     Uniformly distributed on the surface.
+     */
+    inline Vector3 randomSurfacePoint() const {
+        Vector3 V;
+        getRandomSurfacePoint(V);
+        return V;
+    }
+
+    /**
+     Uniformly distributed on the interior (includes surface)
+     */
+    Vector3 randomInteriorPoint() const;
+
+    void getBounds(class AABox&) const;
+};
+
+}
+
+#endif
diff --git a/dep/include/g3dlite/G3D/CollisionDetection.h b/dep/include/g3dlite/G3D/CollisionDetection.h
new file mode 100644
index 00000000000..4dc8b26538d
--- /dev/null
+++ b/dep/include/g3dlite/G3D/CollisionDetection.h
@@ -0,0 +1,1157 @@
+/**
+  @file CollisionDetection.h
+
+
+  Moving collision detection for simple primitives.
+
+  @author Morgan McGuire, matrix@graphics3d.com
+  @cite Spherical collision based on Paul Nettle's
+  ftp://ftp.3dmaileffects.com/pub/FluidStudios/CollisionDetection/Fluid_Studios_Generic_Collision_Detection_for_Games_Using_Ellipsoids.pdf
+  and comments by Max McGuire.  Ray-sphere intersection by Eric Haines.
+  Box-Box intersection written by Kevin Egan.
+  Thanks to Max McGuire of Iron Lore for various bug fixes.
+
+  @created 2001-11-19
+  @edited  2006-01-10
+
+  Copyright 2000-2006, Morgan McGuire.
+  All rights reserved.
+ */
+
+#ifndef G3D_COLLISIONDETECTION_H
+#define G3D_COLLISIONDETECTION_H
+
+#include "G3D/platform.h"
+#include "G3D/Vector3.h"
+#include "G3D/Plane.h"
+#include "G3D/Box.h"
+#include "G3D/Triangle.h"
+#include "G3D/Array.h"
+#include "G3D/Ray.h"
+#include "G3D/Line.h"
+
+namespace G3D {
+
+
+/**
+  Collision detection primitives and tools for building
+  higher order collision detection schemes.
+
+  These routines provide <I>moving</I> and static collision detection.
+  Moving collision detection allows the calculation of collisions that
+  occur during a period of time -- as opposed to the intersection of 
+  two static bodies.
+  
+  Moving collision detection routines detect collisions between
+  <I>only</I> static primitives and moving spheres or points.  Since the 
+  reference frame can be user defined, these functions can be used to
+  detect the collision between two moving bodies by subtracting
+  the velocity vector of one object from the velocity vector of the 
+  sphere or point the detection is to occur with.  This unified 
+  velocity vector will act as if both objects are moving simultaneously.
+
+  Collisions are detected for single-sided objects only.  That is,
+  no collision is detected when <I>leaving</I> a primitive or passing
+  through a plane or triangle opposite the normal... except for the 
+  point-sphere calculation or when otherwise noted.
+
+  For a sphere, the collision location returned is the point in world
+  space where the surface of the sphere and the fixed object meet.
+  It is <B>not</B> the position of the center of the sphere at
+  the time of the collision.
+
+  The collision normal returned is the surface normal to the fixed
+  object at the collision location.
+
+  <p>
+  <b>Static Collision Detection:</b> (Neither object is moving)
+
+  <table>
+  <tr><td></td><td><b>Vector3</b></td><td><b>LineSegment</b></td><td><b>Ray *</b></td><td><b>Line</b></td><td><b>Plane</b></td><td><b>Triangle</b></td><td><b>Sphere</b></td><td><b>Cylinder</b></td><td><b>Capsule</b></td><td><b>AABox</b></td><td><b>Box</b></td></tr>
+  <tr><td><b>Vector3</b></td><td>Vector3::operator== Vector3::fuzzyEq G3D::distance</td><td bgcolor=#C0C0C0 colspan=10 ></td></tr>
+  <tr><td><b>LineSegment</b></td><td>LineSegment::closestPoint LineSegment::distance CollisionDetection::closestPointOnLineSegment</td><td></td><td bgcolor=#C0C0C0 colspan=9 ></td></tr>
+  <tr><td><b>Ray *</b></td><td>Ray::closestPoint Ray::distance</td><td></td><td></td><td bgcolor=#C0C0C0 colspan=8 ></td></tr>
+  <tr><td><b>Line</b></td><td>Line::closestPoint Line::distance</td><td></td><td>CollisionDetection::closestPointsBetweenLineAndLine</td><td></td><td bgcolor=#C0C0C0 colspan=7 ></td></tr>
+  <tr><td><b>Plane</b></td><td></td><td></td><td></td><td></td><td></td><td bgcolor=#C0C0C0 colspan=6 ></td></tr>
+  <tr><td><b>Triangle</b></td><td></td><td></td><td></td><td></td><td></td><td></td><td bgcolor=#C0C0C0 colspan=5 ></td></tr>
+  <tr><td><b>Sphere</b></td><td>Sphere::contains</td><td></td><td></td><td></td><td></td><td></td><td></td><td bgcolor=#C0C0C0 colspan=4 ></td></tr>
+  <tr><td><b>Cylinder</b></td><td>Cylinder::contains</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td bgcolor=#C0C0C0 colspan=3 ></td></tr>
+  <tr><td><b>Capsule</b></td><td>Capsule::contains</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td bgcolor=#C0C0C0 colspan=2 ></td></tr>
+  <tr><td><b>AABox</b></td><td>AABox::contains</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td bgcolor=#C0C0C0 colspan=1 ></td></tr>
+  <tr><td><b>Box</b></td><td>Box::contains</td><td>(treat as Ray)</td><td>CollisionDetection::collisionTimeForMovingPointFixedBox</td><td>(treat as Ray)</td><td>CollisionDetection::penetrationDepthForFixedBoxFixedPlane</td><td>CollisionDetection::penetrationDepthForFixedBoxFixedPlane</td><td>CollisionDetection::penetrationDepthForFixedSphereFixedBox</td><td>None (use OPCODE)</td><td>CollisionDetection::movingSpherePassesThroughFixedBox</td><td>CollisionDetection::penetrationDepthForFixedBoxFixedBox</td><td>CollisionDetection::penetrationDepthForFixedBoxFixedBox</td></tr>
+  </table>
+
+  <p>
+  <b>Moving Collision Detection:</b>
+
+  <i>* Note: Moving collision detection against certain primitives is equivalent to static collision
+   detection against a bigger primitive.  Ray, Line Segment == ``moving Point''; Capsule ==``moving Sphere''; Plane == ``moving Line''</i>
+ */
+class CollisionDetection {
+private:
+
+	/**
+	 Default parameter if value passed to a function as reference is
+	 not to be calculated.  Must be explicitly supported by function.
+	 */
+	static Vector3 ignore;
+
+	/**
+	 Default parameter if value passed to a function as reference is
+	 not to be calculated.  Must be explicitly supported by function.
+	 */
+    static bool    ignoreBool;
+
+	/**
+	 Default parameter if value passed to a function as reference is
+	 not to be calculated.  Must be explicitly supported by function.
+	 */
+    static Array<Vector3> ignoreArray;
+
+
+    // Static class!
+    CollisionDetection() {}
+    virtual ~CollisionDetection() {}
+
+public:
+
+    /**
+      Converts an index [0, 15] to the corresponding separating axis.
+      Does not return normalized vector in the edge-edge case
+      (indices 6 through 15).
+
+	  @param separatingAxisIndex	Separating axis.
+	  @param box1					Box 1.
+	  @param box2					Box 2.
+
+ 	  @return Axis that separates the two boxes.
+  	 */
+    static Vector3 separatingAxisForSolidBoxSolidBox(
+            const int       separatingAxisIndex,
+            const Box &     box1,
+            const Box &     box2);
+
+    /**
+	  Tests whether two boxes have axes that are parallel to
+	  each other.  If they are, axis1 and axis2 are set to be
+	  the parallel axes for both box1 and box2 respectively.
+
+	  @param ca			Dot products of each of the boxes axes
+	  @param epsilon	Fudge factor (small unit by which the dot
+	  					products may vary and still be considered
+	  					zero).
+	  @param axis1		Parallel Axis 1. [Post Condition]
+	  @param axis2		Parallel Axis 2. [Post Condition]
+
+	  @return true  - If boxes have a parallel axis
+	  @return false - otherwise.
+ 	 */
+    static bool parallelAxisForSolidBoxSolidBox(
+            const double*   ca,
+            const double    epsilon,
+            int &           axis1,
+            int &           axis2);
+
+    /**
+      Calculates the projected distance between the two boxes along
+      the specified separating axis, negative distances correspond
+      to an overlap along that separating axis.  The distance is not
+      divided by denominator dot(L, L), see
+      penetrationDepthForFixedSphereFixedBox() for more details
+
+      @param separatingAxisIndex
+	  @param a Box 1's bounding sphere vector
+	  @param b Box 2's bounding sphere vector
+	  @param D Vector between Box 1 and Box 2's center points
+	  @param c Pointer to array of dot products of the axes of Box 1
+	           and Box 2.
+	  @param ca Pointer to array of unsigned dot products of the axes
+	            of Box 1 and Box 2.
+	  @param ad Pointer to array of dot products of Box 1 axes and D.
+	  @param bd Pointer to array of dot products of Box 2 axes and D.
+
+      @return Projected distance between the two boxes along the
+      specified separating axis.
+     */
+    static float projectedDistanceForSolidBoxSolidBox(
+            const int           separatingAxisIndex,
+            const Vector3 &     a,
+            const Vector3 &     b,
+            const Vector3 &     D,
+            const double*       c,
+            const double*       ca,
+            const double*       ad,
+            const double*       bd);
+
+
+	/**
+	  Creates a set of standard information about two boxes in order to
+	  solve for their collision.  This information includes a vector to
+	  the radius of the bounding sphere for each box, the vector between
+	  each boxes' center and a series of dot products between differing
+	  important vectors.  These dot products include those between the axes
+	  of both boxes (signed and unsigned values), and the dot products
+	  between all the axes of box1 and the boxes' center vector and box2
+	  and the boxes' center vector.
+
+	  @pre The following space requirements must be met:
+      		- c[]  9 elements
+      		- ca[] 9 elements
+      		- ad[] 3 elements
+      		- bd[] 3 elements
+
+	  @cite dobted from David Eberly's papers, variables used in this function
+      correspond to variables used in pages 6 and 7 in the pdf
+      http://www.magic-software.com/Intersection.html
+      http://www.magic-software.com/Documentation/DynamicCollisionDetection.pdf
+
+      @note Links are out-dated. (Kept to preserve origin and authorship)
+
+	  @param box1 Box 1
+	  @param box2 Box 2
+	  @param a Box 1's bounding sphere vector
+	  @param b Box 2's bounding sphere vector
+	  @param D Vector between Box 1 and Box 2's center points
+	  @param c Pointer to array of dot products of the axes of Box 1
+	           and Box 2.
+	  @param ca Pointer to array of unsigned dot products of the axes
+	            of Box 1 and Box 2.
+	  @param ad Pointer to array of dot products of Box 1 axes and D.
+	  @param bd Pointer to array of dot products of Box 2 axes and D.
+	 */
+    static void fillSolidBoxSolidBoxInfo(
+            const Box &     box1,
+            const Box &     box2,
+            Vector3 &       a,
+            Vector3 &       b,
+            Vector3 &       D,
+            double*         c,
+            double*         ca,
+            double*         ad,
+            double*         bd);
+
+	/**
+	  Performs a simple bounding sphere check between two boxes to determine
+	  whether these boxes could <i>possibly</i> intersect.  This is a very
+	  cheap operation (three dot products, two sqrts and a few others).  If
+	  it returns true, an intersection is possible, but not necessarily
+	  guaranteed.
+
+	  @param a Vector from box A's center to an outer vertex
+	  @param b Vector from box B's center to an outer vertex
+	  @param D Distance between the centers of the two boxes
+
+	  @return true - if possible intersection
+	  @return false - otherwise (This does not guarantee an intersection)
+	 */
+    static bool conservativeBoxBoxTest(
+            const Vector3 &     a,
+            const Vector3 &     b,
+            const Vector3 &     D);
+
+	/**
+ 	  Determines whether two fixed solid boxes intersect.
+
+      @note To speed up collision detection, the lastSeparatingAxis from
+      the previous time step can be passed in and that plane can be
+      checked first.  If the separating axis was not saved, or if the
+      two boxes intersected then lastSeparatingAxis should equal -1.
+
+	  @cite Adobted from David Eberly's papers, variables used in this function
+      correspond to variables used in pages 6 and 7 in the pdf
+      http://www.magic-software.com/Intersection.html
+      http://www.magic-software.com/Documentation/DynamicCollisionDetection.pdf
+
+ 	  @param box1				Box 1.
+ 	  @param box2				Box 2.
+ 	  @param lastSeparatingAxis	Last separating axis.
+ 	  							(optimization - see note)
+
+	  @return true  - Intersection.
+	  @return false - otherwise.
+	 */
+    static bool fixedSolidBoxIntersectsFixedSolidBox(
+        const Box&      box1,
+        const Box&      box2,
+        const int       lastSeparatingAxis = -1);
+
+    /**
+	  Calculates the closest points on two lines with each other.   If the
+	  lines are parallel then using the starting point, else calculate the
+	  closest point on each line to the other.
+
+	  @note This is very similiar to calculating the intersection of two lines.
+	  Logically then, the two points calculated would be identical if calculated
+	  with inifinite precision, but with the finite precision of floating point
+	  calculations, these values could (will) differ as the line slope approaches
+	  zero or inifinity.
+
+	  @cite variables and algorithm based on derivation at the following website:
+      	    http://softsurfer.com/Archive/algorithm_0106/algorithm_0106.htm
+
+	  @param line1		Line 1.
+	  @param line2		Line 2.
+	  @param closest1	Closest point on line 1.
+	  @param closest2	Closest point on line 2.
+ 	*/
+    static void closestPointsBetweenLineAndLine(
+            const Line &    line1,
+            const Line &    line2,
+            Vector3 &       closest1,
+            Vector3 &       closest2);
+
+    /**
+	  Calculates the depth of penetration between two fixed boxes.
+      Contact normal faces away from box1 and into box2.  If there is
+      contact, only one contact point is returned.  The minimally
+      violated separating plane is computed
+         - if the separating axis corresponds to a face
+              the contact point is half way between the deepest vertex
+              and the face
+         - if the separating axis corresponds to two edges
+              the contact point is the midpoint of the smallest line
+              segment between the two edge lines
+
+	  @note This is very similiar to calculating the intersection of two lines.
+	  Logically then, the two points calculated would be identical if calculated
+	  with inifinite precision, but with the finite precision of floating point
+	  calculations, these values could (will) differ as the line slope approaches
+	  zero or inifinity.
+
+	  @cite adobted from David Eberly's papers, variables used in this function
+      correspond to variables used in pages 6 and 7 in the pdf
+      http://www.magic-software.com/Intersection.html
+      http://www.magic-software.com/Documentation/DynamicCollisionDetection.pdf
+
+	  @param box1				Box 1
+	  @param box2				Box 2
+	  @param contactPoints		Contact point between boxes. [Post Condition]
+	  @param contactNormals		Surface normal at contact point. [Post Condition]
+	  @param lastSeparatingAxis	Last separating axis. (Used for optimization)
+
+	  @return Depth of penetration between the two boxes.  If there is no
+ 	   intersection between the boxes, then a negative value is returned.
+ 	*/
+    static float penetrationDepthForFixedBoxFixedBox(
+        const Box&          box1,
+        const Box&          box2,
+        Array<Vector3>&     contactPoints,
+        Array<Vector3>&     contactNormals,
+        const int           lastSeparatingAxis = -1);
+
+    /**
+	  Calculates the depth of penetration between two fixed spheres as well
+	  as the deepest point of Sphere A that penetrates Sphere B. The normal
+      returned points <B>away</B> from the object A, although it may
+      represent a perpendicular to either the faces of object B or object A
+      depending on their relative orientations.
+
+	  @param sphereA		Fixed Sphere A.
+	  @param sphereB		Fixed Sphere B.
+	  @param contactPoints	Sphere A's deepest point that penetrates Sphere B.
+	 						[Post Condition]
+	  @param contactNormals	Normal at penetration point. [Post Condition]
+
+	  @return Depth of penetration.  If there is no intersection between the
+	 		 objects then the depth will be a negative value.
+	 */
+    static float penetrationDepthForFixedSphereFixedSphere(
+        const class Sphere& sphereA,
+        const Sphere&       sphereB,
+        Array<Vector3>&     contactPoints,
+        Array<Vector3>&     contactNormals = ignoreArray);
+
+    /**
+	  Calculates the depth of penetration between a fixed sphere and a fixed
+	  box as well as the deepest point of the sphere that penetrates the box
+	  and the normal at that intersection.
+
+	  @note There are three possible intersections between a sphere and box.
+	  - Sphere completely contained in the box
+	  - Sphere intersects one edge
+	  - Sphere intersects one vertex
+
+	  The contact point and contact normal vary for each of these situations.
+	  - Sphere contained in Box:
+	 	- Normal is based on side of least penetration (as is the depth calculation).
+	 	- Point is based on center of sphere
+	  - Sphere intersects one edge
+	 	- Normal is based on vector from the box center to the point of depth.
+	 	- Point is closest point to the sphere on the line
+	  - Sphere intersects one vertex
+	 	- Normal is based on vector from the box center to the vertex of penetration.
+	 	- Point is vertex of penetration.
+
+      @cite Adapted from Jim Arvo's method in Graphics Gems
+      See also http://www.win.tue.nl/~gino/solid/gdc2001depth.pdf
+
+	  @param sphere			Fixed Sphere.
+	  @param box			Fixed Box.
+	  @param contactPoints	Sphere point that penetrates the box. [Post Condition]
+	  @param contactNormals	Normal at the penetration point. [Post Condition]
+
+	  @return Depth of penetration.  If there is no intersection between the
+	 		  objects then the depth will be a negative value.
+ 	 */
+    static float penetrationDepthForFixedSphereFixedBox(
+        const Sphere&       sphere,
+        const Box&          box,
+        Array<Vector3>&     contactPoints,
+        Array<Vector3>&     contactNormals = ignoreArray);
+
+	/**
+	  Calculates the depth of penetration between a Fixed Sphere and a Fixed
+	  Plane as well as the deepest point of the sphere that penetrates the plane
+	  and the plane normal at that intersection.
+
+	  @param sphere			Fixed Sphere.
+	  @param plane			Fixed Plane.
+	  @param contactPoints	Sphere point that penetrates the plane.
+	 						[Post Condition]
+	  @param contactNormals	Normal at penetration point. [Post Condition]
+
+ 	  @return Depth of penetration.  If there is no intersection between the
+	  		 objects then the depth will be a negative value.
+ 	 */
+    static float penetrationDepthForFixedSphereFixedPlane(
+        const Sphere&       sphereA,
+        const class Plane&  planeB,
+        Array<Vector3>&     contactPoints,
+        Array<Vector3>&     contactNormals = ignoreArray);
+
+	/**
+	  Calculates the depth of penetration between a fixed box and a fixed
+	  plane as well as the vertexes of the box that penetrate the plane
+	  and the plane normals at those intersections.
+
+	  @param box			Fixed Box.
+	  @param plane			Fixed Plane.
+	  @param contactPoints	Box points that penetrate the plane.
+	 						[Post Condition]
+	  @param contactNormals	Normals at penetration points [Post Condition]
+
+	  @return Depth of penetration.  If there is no intersection between the
+	 		 objects then the depth will be a negative value.
+ 	*/
+    static float penetrationDepthForFixedBoxFixedPlane(
+        const Box&          box,
+        const Plane&        plane,
+        Array<Vector3>&     contactPoints,
+        Array<Vector3>&     contactNormals = ignoreArray);
+
+	/**
+ 	  Calculates time between the intersection of a moving point and a fixed
+ 	  plane.
+
+ 	  @note This is only a one sided collision test.   The side defined by
+ 	  the plane's surface normal is the only one tested.  For a two sided
+ 	  collision, call the function once for each side's surface normal.
+
+ 	  @param point		Moving point.
+ 	  @param velocity	Point's velocity.
+ 	  @param plane		Fixed plane.
+ 	  @param location	Location of collision. [Post Condition]
+ 	  					(Infinite vector on no collision)
+ 	  @param outNormal	Plane's surface normal. [Post Condition]
+
+ 	  @return Time til collision.  If there is no collision then the return
+ 	          value will be inf().
+ 	*/
+    static float collisionTimeForMovingPointFixedPlane(
+        const Vector3&			point,
+        const Vector3&			velocity,
+        const class Plane&		plane,
+        Vector3&				outLocation,
+        Vector3&                outNormal = ignore);
+
+	/**
+ 	  Calculates time between the intersection of a moving point and a fixed
+ 	  triangle.
+
+ 	  @note This is only a one sided collision test.   The side defined by
+ 	  the triangle's surface normal is the only one tested.  For a two sided
+ 	  collision, call the function once for each side's surface normal.
+
+ 	  @param orig		Moving point.
+ 	  @param dir		Point's velocity.
+	  @param v0 		Triangle vertex 1.
+	  @param v1 		Triangle vertex 2.
+	  @param v2 		Triangle vertex 3
+ 	  @param location	Location of collision. [Post Condition]
+ 	  					(Infinite vector on no collision)
+
+ 	  @return Time til collision.  If there is no collision then the return
+ 	          value will be inf().
+ 	*/
+    inline static float collisionTimeForMovingPointFixedTriangle(
+        const Vector3& orig,
+        const Vector3& dir,
+        const Vector3& v0,
+        const Vector3& v1,
+        const Vector3& v2) {
+        return Ray::fromOriginAndDirection(orig, dir).intersectionTime(v0, v1, v2);
+    }
+
+	/**
+ 	  Calculates time between the intersection of a moving point and a fixed
+ 	  triangle.
+
+ 	  @note This is only a one sided collision test.   The side defined by
+ 	  the triangle's surface normal is the only one tested.  For a two sided
+ 	  collision, call the function once for each side's surface normal.
+
+ 	  @param orig		Moving point.
+ 	  @param dir		Point's velocity.
+	  @param v0 		Triangle vertex 1.
+	  @param v1 		Triangle vertex 2.
+	  @param v2 		Triangle vertex 3
+ 	  @param location	Location of collision. [Post Condition]
+ 	  					(Infinite vector on no collision)
+
+ 	  @return Time til collision.  If there is no collision then the return
+ 	          value will be inf().
+ 	*/
+    inline static float collisionTimeForMovingPointFixedTriangle(
+        const Vector3& orig,
+        const Vector3& dir,
+        const Vector3& v0,
+        const Vector3& v1,
+        const Vector3& v2,
+        Vector3&       location) {
+        float t = collisionTimeForMovingPointFixedTriangle(orig, dir, v0, v1, v2);
+        if (t < inf()) {
+            location = orig + dir * t;
+        }
+        return t;
+    }
+
+	/**
+ 	  Calculates time between the intersection of a moving point and a fixed
+ 	  triangle.
+
+ 	  @note This is only a one sided collision test.   The side defined by
+ 	  the triangle's surface normal is the only one tested.  For a two sided
+ 	  collision, call the function once for each side's surface normal.
+
+ 	  @param orig		Moving point.
+ 	  @param dir		Point's velocity.
+ 	  @param tri		Fixed triangle.
+ 	  @param location	Location of collision. [Post Condition]
+ 	  					(Infinite vector on no collision)
+ 	  @param normal		Triangle's surface normal. [Post Condition]
+
+ 	  @return Time til collision.  If there is no collision then the return
+ 	          value will be inf().
+ 	*/
+    inline static float collisionTimeForMovingPointFixedTriangle(
+        const Vector3&  orig,
+        const Vector3&  dir,
+        const Triangle& tri,
+        Vector3&        location = ignore,
+        Vector3&        normal   = ignore) {
+
+        float t = collisionTimeForMovingPointFixedTriangle(
+            orig, dir, tri.vertex(0), tri.vertex(1), tri.vertex(2));
+        
+        if ((t < inf()) && (&location != &ignore)) {
+            location = orig + dir * t;
+            normal   = tri.normal();
+        }
+        return t;
+    }
+
+	/**
+ 	  Calculates time between the intersection of a moving point and a fixed
+ 	  triangle.
+
+ 	  @note This is only a one sided collision test.   The side defined by
+ 	  the triangle's surface normal is the only one tested.  For a two sided
+ 	  collision, call the function once for each side's surface normal.
+
+ 	  @param orig		Moving point.
+ 	  @param dir		Point's velocity.
+	  @param v0 		Triangle vertex 1.
+	  @param v1 		Triangle vertex 2.
+	  @param v2 		Triangle vertex 3
+ 	  @param location	Location of collision. [Post Condition]
+ 	  					(Infinite vector on no collision)
+ 	  @param normal		Triangle's surface normal. [Post Condition]
+
+ 	  @return Time til collision.  If there is no collision then the return
+ 	          value will be inf().
+ 	*/
+    inline static float collisionTimeForMovingPointFixedTriangle(
+        const Vector3& orig,
+        const Vector3& dir,
+        const Vector3& v0,
+        const Vector3& v1,
+        const Vector3& v2,
+        Vector3&       location,
+        Vector3&       normal) {
+        float t = collisionTimeForMovingPointFixedTriangle(orig, dir, v0, v1, v2);
+        if (t < inf()) {
+            location = orig + dir * t;
+            normal   = (v2 - v0).cross(v1 - v0).direction();
+        }
+        return t;
+    }
+
+    /**
+     Unlike other methods, does not support an output normal.
+     If the ray origin is inside the box, returns inf() but inside
+     is set to true.
+     <B>Beta API</B>
+
+     @cite Andrew Woo, from "Graphics Gems", Academic Press, 1990
+  	 @cite Optimized code by Pierre Terdiman, 2000 (~20-30% faster on my Celeron 500)
+     @cite Epsilon value added by Klaus Hartmann
+     @cite http://www.codercorner.com/RayAABB.cpp
+     */
+    static float collisionTimeForMovingPointFixedAABox(
+        const Vector3&			point,
+        const Vector3&			velocity,
+        const class AABox&      box,
+        Vector3&				outLocation,
+        bool&                   inside = ignoreBool,
+        Vector3&                outNormal = ignore);
+
+    /**
+	 Calculates time between the intersection of a moving point and a fixed
+	 Axis-Aligned Box (AABox).
+
+	 @note Avoids the sqrt from collisionTimeForMovingPointFixedAABox.
+
+	 @param point		Moving point.
+	 @param velocity	Sphere's velocity.
+	 @param box			Fixed AAbox.
+	 @param location	Location of collision. [Post Condition]
+	 @param Inside		Does the ray originate inside the box? [Post Condition]
+	 @param normal		Box's surface normal to collision [Post Condition]
+
+	 @return Time til collision.  If there is no collision then the return
+	         value will be inf().
+ 	*/
+    static bool collisionLocationForMovingPointFixedAABox(
+        const Vector3&			point,
+        const Vector3&			velocity,
+        const class AABox&      box,
+        Vector3&				outLocation,
+        bool&                   inside = ignoreBool,
+        Vector3&                normal = ignore);
+
+    /**
+	 Calculates time between the intersection of a moving point and a fixed
+	 sphere.
+
+	 @note When ray is starts inside the rectangle, the exiting intersection
+	 is detected.
+
+	 @param point		Moving point.
+	 @param velocity	Point's velocity.
+	 @param Sphere		Fixed Sphere.
+	 @param location	Location of collision. [Post Condition]
+	 @param outNormal	Sphere's surface normal to collision [Post Condition]
+
+	 @return Time til collision.  If there is no collision then the return
+	         value will be inf().
+ 	*/
+    static float collisionTimeForMovingPointFixedSphere(
+        const Vector3&			point,
+        const Vector3&			velocity,
+        const class Sphere&		sphere,
+        Vector3&				outLocation,
+        Vector3&                outNormal = ignore);
+
+    /**
+	 Calculates time between the intersection of a moving point and a fixed
+	 box.
+
+	 @note If the point is already inside the box, no collision: inf is returned.
+
+	 @param point		Moving point.
+	 @param velocity	Sphere's velocity.
+	 @param box			Fixed box.
+	 @param location	Position of collision. [Post Condition]
+	 @param outNormal	Box's surface normal to collision [Post Condition]
+
+	 @return Time til collision.  If there is no collision then the return
+	         value will be inf().
+ 	*/
+    static float collisionTimeForMovingPointFixedBox(
+        const Vector3&			point,
+        const Vector3&			velocity,
+        const class  Box&		box,
+        Vector3&				outLocation,
+        Vector3&                outNormal = ignore);
+
+	/**
+	 Calculates time between the intersection of a moving point and a fixed
+	 rectangle defined by the points v0, v1, v2, & v3.
+
+	 @note This is only a one sided collision test.   The side defined by
+	 the rectangle's surface normal is the only one tested.  For a two sided
+	 collision, call the function once for each side's surface normal.
+
+	 @param point		Moving point.
+	 @param velocity	Sphere's velocity.
+	 @param v0 			Rectangle vertex 1.
+	 @param v1 			Rectangle vertex 2.
+	 @param v2 			Rectangle vertex 3
+	 @param v3 			Rectangle vertex 4.
+	 @param location	Location of collision [Post Condition]
+	 @param outNormal	Rectangle's surface normal. [Post Condition]
+
+	 @return Time til collision.  If there is no collision then the return
+	         value will be inf().
+ 	*/
+    static float collisionTimeForMovingPointFixedRectangle(
+        const Vector3&			point,
+        const Vector3&			velocity,
+        const Vector3&			v0,
+        const Vector3&			v1,
+        const Vector3&			v2,
+        const Vector3&			v3,
+        Vector3&				outLocation,
+        Vector3&                outNormal = ignore);
+
+	/**
+	 Calculates time between the intersection of a moving point and a fixed
+	 capsule.
+
+	 @param point		Moving point.
+	 @param velocity	Point's velocity.
+	 @param capsule		Fixed capsule.
+	 @param location	Location of collision. [Post Condition]
+	 @param outNormal	Capsule's surface normal to collision [Post Condition]
+
+	 @return Time til collision.  If there is no collision then the return
+	         value will be inf().
+ 	*/
+	static float collisionTimeForMovingPointFixedCapsule(
+		const Vector3&		    point,
+		const Vector3&		    velocity,
+		const class Capsule&	capsule,
+        Vector3&				outLocation,
+        Vector3&                outNormal = ignore);
+
+	/**
+	 Calculates time between the intersection of a moving sphere and a fixed
+	 triangle.
+
+	 @param sphere		Moving sphere.
+	 @param velocity	Sphere's velocity.
+	 @param plane		Fixed Plane.
+	 @param location	Location of collision -- not center position of sphere
+	 					at the collision time. [Post Condition]
+	 @param outNormal	Box's surface normal to collision [Post Condition]
+
+	 @return Time til collision.  If there is no collision then the return
+	         value will be inf().
+	 */
+    static float collisionTimeForMovingSphereFixedPlane(
+        const class Sphere&		sphere,
+        const Vector3&	    	velocity,
+        const class Plane&		plane,
+        Vector3&				outLocation,
+        Vector3&                outNormal = ignore);
+
+	/**
+	 Calculates time between the intersection of a moving sphere and a fixed
+	 triangle.
+
+	 @param sphere		Moving sphere.
+	 @param velocity	Sphere's velocity.
+	 @param triangle	Fixed Triangle.
+	 @param location	Location of collision -- not center position of sphere
+	 					at the collision time. [Post Condition]
+	 @param outNormal	Box's surface normal to collision [Post Condition]
+
+	 @return Time til collision.  If there is no collision then the return
+	         value will be inf().
+	*/
+    static float collisionTimeForMovingSphereFixedTriangle(
+        const class Sphere&		sphere,
+        const Vector3&		    velocity,
+        const Triangle&       triangle,
+        Vector3&				outLocation,
+        Vector3&                outNormal = ignore);
+
+	/**
+	 Calculates time between the intersection of a moving sphere and a fixed
+	 rectangle defined by the points v0, v1, v2, & v3.
+
+	 @param sphere		Moving sphere.
+	 @param velocity	Sphere's velocity.
+	 @param v0 			Rectangle vertex 1.
+	 @param v1 			Rectangle vertex 2.
+	 @param v2 			Rectangle vertex 3
+	 @param v3 			Rectangle vertex 4.
+	 @param location	Location of collision -- not center position of sphere
+	 					at the collision time. [Post Condition]
+	 @param outNormal	Box's surface normal to collision [Post Condition]
+
+	 @return Time til collision.  If there is no collision then the return
+	         value will be inf().
+ 	*/
+    static float collisionTimeForMovingSphereFixedRectangle(
+        const class Sphere&		sphere,
+        const Vector3&	    	velocity,
+        const Vector3&	    	v0,
+        const Vector3&	    	v1,
+        const Vector3&	    	v2,
+        const Vector3&	    	v3,
+        Vector3&				outLocation,
+        Vector3&                outNormal = ignore);
+
+	/**
+	 Calculates time between the intersection of a moving sphere and a fixed
+	 box.
+
+	 @note This function will not detect an intersection between a moving object
+	 that is already interpenetrating the fixed object.
+
+	 @param sphere		Moving sphere.
+	 @param velocity	Sphere's velocity.
+	 @param box			Fixed box.
+	 @param location	Location of collision -- not center position of sphere
+	 					at the collision time. [Post Condition]
+	 @param outNormal	Box's surface normal to collision [Post Condition]
+
+	 @return Time til collision.  If there is no collision then the return
+	         value will be inf().
+ 	*/
+    static float collisionTimeForMovingSphereFixedBox(
+        const class Sphere&		sphere,
+        const Vector3&		    velocity,
+        const class Box&		box,
+        Vector3&				outLocation,
+        Vector3&                outNormal = ignore);
+
+    /**
+	 Calculates time between the intersection of a moving sphere and a fixed
+	 sphere.
+
+	 @note This won't detect a collision if the sphere is already interpenetrating
+	       the fixed sphere.
+
+	 @param movingSphere	Moving sphere.
+	 @param velocity		Sphere's velocity.
+	 @param fixedSphere		Fixed Sphere.
+	 @param location		Location of collision -- not center position of sphere
+	 						at the collision time. [Post Condition]
+	 @param outNormal		Sphere's surface normal to collision [Post Condition]
+
+	 @return Time til collision.  If there is no collision then the return
+	         value will be inf().
+ 	*/
+	static float collisionTimeForMovingSphereFixedSphere(
+		const class Sphere&		sphere,
+		const Vector3&		    velocity,
+		const class Sphere&	    fixedSphere,
+        Vector3&				outLocation,
+        Vector3&                outNormal = ignore);
+
+    /**
+	 Calculates time between the intersection of a moving sphere and a fixed
+	 capsule.
+
+	 @note This won't detect a collision if the sphere is already
+	       interpenetrating the capsule.
+
+	 @param sphere		Moving sphere.
+	 @param velocity	Sphere's velocity.
+	 @param capsule		Fixed capsule.
+	 @param location	Location of collision -- not center position of sphere
+	 					at the collision time. [Post Condition]
+	 @param outNormal	Capsule's surface normal to the collision [Post Condition]
+
+	 @return Time til collision.  If there is no collision then the return
+	         value will be inf().
+ 	*/
+	static float collisionTimeForMovingSphereFixedCapsule(
+		const class Sphere&		sphere,
+		const Vector3&		    velocity,
+		const class Capsule&	capsule,
+        Vector3&				outLocation,
+        Vector3&                outNormal = ignore);
+
+    /**
+	 Finds the direction of bounce that a sphere would have when it 
+	 intersects an object with the  given time of collision, the 
+	 collision location and the collision normal.
+
+	 @note This function works like a pong style ball bounce.
+
+	 @param sphere				Moving sphere.
+	 @param velocity			Sphere's velocity.
+	 @param collisionTime		Time of collision.
+	 @param collisionLocation	Collision location.
+	 @param collisionNormal		Surface collision normal.
+
+	 @return Direction of bounce.
+ 	*/
+    static Vector3 bounceDirection(
+        const class Sphere&		sphere,
+        const Vector3&			velocity,
+        const float				collisionTime,
+        const Vector3&			collisionLocation,
+        const Vector3&          collisionNormal);
+
+    /**
+	 Finds the direction of slide given a moving sphere, its velocity, the
+	 time of collision and the collision location.  This function works as
+	 if the sphere intersects the surface and continues to hug it.
+
+	 @note The result will work well for calculating the movement of a player
+	 who collides with an object and continues moving along the object instead
+	 of just bouncing off it.
+
+	 @param sphere				Moving sphere.
+	 @param velocity			Sphere's velocity.
+	 @param collisionTime		Time of collision
+	 @param collisionLocation	Collision location.
+
+	 @return Direction of slide.
+ 	*/
+    static Vector3 slideDirection(
+        const class Sphere&		sphere,
+        const Vector3&			velocity,
+        const float				collisionTime,
+        const Vector3&			collisionLocation);
+
+	/**
+	 Finds the closest point on a line segment to a given point.
+
+	 @param v0 line vertex 1.
+	 @param v1 line vertex 2.
+	 @param point External point.
+
+	 @return Closests point to <code>point</code> on the line segment.
+ 	*/
+	static Vector3 closestPointOnLineSegment(
+        const Vector3&			v0,
+        const Vector3&			v1,
+        const Vector3&			point);
+
+    /**
+	 Finds the closest point on a line segment to a given point.
+
+	 @note This is an optimization to closestPointOnLineSegment.  Edge length
+	 and direction can be used in this function if already pre-calculated.  This
+	 prevents doing the same work twice.
+
+	 @param v0 line vertex 1.
+	 @param v1 line vertex 2.
+     @param edgeDirection The direction of the segment (unit length).
+     @param edgeLength The length of the segment.
+     @param point External point.
+
+	 @return Closests point to <code>point</code> on the line segment.
+ 	*/
+    static Vector3 closestPointOnLineSegment(
+        const Vector3&			v0,
+        const Vector3&			v1,
+        const Vector3&          edgeDirection,
+        float                  edgeLength,
+        const Vector3&			point);
+
+    /**
+	 Finds the closest point on the perimeter of the triangle to an external point;
+	 given a triangle defined by three points v0, v1, & v2, and the external point.
+
+	 @param v0 Triangle vertex 1.
+	 @param v1 Triangle vertex 2.
+	 @param v2 Triangle vertex 3.
+	 @param point External point.
+
+	 @return Closests point to <code>point</code> on the perimeter of the
+	 triangle.
+ 	*/
+    static Vector3 closestPointToTrianglePerimeter(
+        const Vector3&			v0,
+        const Vector3&			v1,
+        const Vector3&			v2,
+        const Vector3&			point);
+
+	/**
+	 Finds the closest point on the perimeter of the triangle to an external point;
+	 given a triangle defined by the array of points v, its edge directions and
+	 their lengths, as well as the external point.
+
+	 @note This is an optimization to closestPointToTrianglePerimeter.  Edge length
+	 and direction can be used in this function if already pre-calculated.  This
+	 prevents doing the same work twice.
+
+	 @param v0 Triangle vertex 1.
+	 @param v1 Triangle vertex 2.
+	 @param v2 Triangle vertex 3.
+	 @param point External point.
+
+	 @return Closests point to <code>point</code> on the perimeter of the
+	 triangle.
+ 	*/
+    static Vector3 closestPointToTrianglePerimeter(
+        const Vector3           v[3],
+        const Vector3           edgeDirection[3],
+        const double            edgeLength[3],
+        const Vector3&			point);
+
+    /**
+	 Tests whether a point is contained within the triangle defined by
+	 v0, v1, & v2 and its plane's normal.
+
+	 @param v0 Triangle vertex 1.
+	 @param v1 Triangle vertex 2.
+	 @param v2 Triangle vertex 3.
+	 @param normal Normal to triangle's plane.
+	 @param point The point in question.
+	 @param primaryAxis Primary axis of triangle.  This will be detected
+	        if not given. This parameter is provided as an optimization.
+
+	 @return true  - if point is inside the triangle.
+	 @return false - otherwise
+ 	*/
+    static bool isPointInsideTriangle(
+        const Vector3&			v0,
+        const Vector3&			v1,
+        const Vector3&			v2,
+        const Vector3&			normal,
+        const Vector3&			point,
+        Vector3::Axis  primaryAxis = Vector3::DETECT_AXIS);
+
+     /**
+	  Tests for the intersection of a moving sphere and a fixed box in a
+	  given time limit.
+
+	  @note Returns true if any part of the sphere is inside the box
+     		during the time period (inf means "ever").  Useful for
+		    performing bounding-box collision detection.
+
+	  @param sphere 		Moving sphere.
+	  @param velocity 	Velocity of moving sphere.
+	  @param box 		Fixed box.
+	  @param timeLimit 	Time limit for intersection test.
+
+	  @return true  -  if the two objects will touch.
+	  @return false - if there is no intersection.
+ 	*/
+    static bool movingSpherePassesThroughFixedBox(
+        const Sphere&           sphere,
+        const Vector3&          velocity,
+        const Box&              box,
+        double                  timeLimit = inf());
+
+	/**
+	 Tests for the intersection of a moving sphere and a fixed sphere in a
+	 given time limit.
+
+	 @note This function will not detect an intersection between a moving object
+	 that is already interpenetrating the fixed object.
+
+	 @param sphere 		Moving sphere.
+	 @param velocity 	Velocity of moving sphere.
+	 @param fixedSphere Fixed sphere.
+	 @param timeLimit 	Time limit for intersection test.
+
+	 @return true  -  if the two spheres will touch.
+	 @return false - if there is no intersection.
+ 	*/
+    static bool movingSpherePassesThroughFixedSphere(
+        const Sphere&           sphere,
+        const Vector3&          velocity,
+        const Sphere&           fixedSphere,
+        double                  timeLimit = inf());
+
+	/**
+	 Tests for the intersection of two fixed spheres.
+
+	 @param sphere1 Fixed sphere 1.
+	 @param sphere2 Fixed sphere 2.
+
+	 @return true -  if the two spheres touch.
+	 @return false - if there is no intersection.
+ 	*/
+    static bool fixedSolidSphereIntersectsFixedSolidSphere(
+        const Sphere&           sphere1,
+        const Sphere&           sphere2);
+
+	/**
+	 Tests for the intersection of a fixed sphere and a fixed box.
+
+	 @param sphere Fixed sphere.
+	 @param box    Fixed box.
+
+	 @return true  -  if the two objects touch.
+	 @return false - if there is no intersection.
+ 	*/
+    static bool fixedSolidSphereIntersectsFixedSolidBox(
+        const Sphere&           sphere,
+        const Box&              box);
+
+    /**
+	 Tests whether a point is inside a rectangle defined by the vertexes
+	 v0, v1, v2, & v3, and the rectangle's plane normal.
+
+	 @param v0 Rectangle vertex 1.
+	 @param v1 Rectangle vertex 2.
+	 @param v2 Rectangle vertex 3.
+	 @param v3 Rectangle vertex 4.
+	 @param normal Normal to rectangle's plane.
+	 @param point The point in question.
+
+	 @return true  - if point is inside the rectangle.
+	 @return false - otherwise
+ 	*/
+    static bool isPointInsideRectangle(
+        const Vector3&			v0,
+        const Vector3&			v1,
+        const Vector3&			v2,
+        const Vector3&			v3,
+        const Vector3&			normal,
+        const Vector3&			point);
+
+    /**
+	 Finds the closest point on the perimeter of the rectangle to an
+	 external point; given a rectangle defined by four points v0, v1,
+	 v2, & v3, and the external point.
+
+	 @param v0 Rectangle vertex 1.
+	 @param v1 Rectangle vertex 2.
+	 @param v2 Rectangle vertex 3.
+	 @param v3 Rectangle vertex 4.
+	 @param point External point.
+
+	 @return Closests point to <code>point</code> on the perimeter of the
+ 	 rectangle.
+ 	 */
+    static Vector3 closestPointToRectanglePerimeter(
+        const Vector3&			v0,
+        const Vector3&			v1,
+        const Vector3&			v2,
+        const Vector3&			v3,
+        const Vector3&			point);
+
+     /**
+	  Finds the closest point in the rectangle to an external point; Given
+	  a rectangle defined by four points v0, v1, v2, & v3, and the external
+	  point.
+
+	  @param v0 Rectangle vertex 1.
+	  @param v1 Rectangle vertex 2.
+	  @param v2 Rectangle vertex 3
+	  @param v3 Rectangle vertex 4.
+	  @param point External point.
+
+      @return Closet point in the rectangle to the external point.
+	  */
+	 static Vector3 closestPointToRectangle(
+        const Vector3&			v0,
+        const Vector3&			v1,
+        const Vector3&			v2,
+        const Vector3&			v3,
+        const Vector3&			point);
+};
+
+} // namespace
+
+#endif // G3D_COLLISIONDETECTION_H
diff --git a/dep/include/g3dlite/G3D/CoordinateFrame.h b/dep/include/g3dlite/G3D/CoordinateFrame.h
new file mode 100644
index 00000000000..657323eea6b
--- /dev/null
+++ b/dep/include/g3dlite/G3D/CoordinateFrame.h
@@ -0,0 +1,318 @@
+/**
+ @file CoordinateFrame.h
+
+ @maintainer Morgan McGuire, matrix@graphics3d.com
+ 
+ @created 2001-03-04
+ @edited  2006-04-07
+
+ Copyright 2000-2006, Morgan McGuire.
+ All rights reserved.
+*/
+
+#ifndef G3D_COORDINATEFRAME_H
+#define G3D_COORDINATEFRAME_H
+
+#include "G3D/platform.h"
+#include "G3D/Vector3.h"
+#include "G3D/Vector4.h"
+#include "G3D/Matrix3.h"
+#include "G3D/Array.h"
+#include <math.h>
+#include <string>
+#include <stdio.h>
+#include <cstdarg>
+#include <assert.h>
+
+namespace G3D {
+
+/**
+ A rigid body RT (rotation-translation) transformation.
+    
+CoordinateFrame abstracts a 4x4 matrix that maps object space to world space:
+  
+  v_world = C * v_object
+
+CoordinateFrame::rotation is the upper 3x3 submatrix, CoordinateFrame::translation
+is the right 3x1 column.  The 4th row is always [0 0 0 1], so it isn't stored.  
+So you don't have to remember which way the multiplication and transformation work,
+it provides explicit toWorldSpace and toObjectSpace methods.  Also, points, vectors 
+(directions), and surface normals transform differently, so they have separate methods.
+ 
+Some helper functions transform whole primitives like boxes in and out of object space.
+
+Convert to Matrix4 using CoordinateFrame::toMatrix4.  You <I>can</I> construct a CoordinateFrame
+from a Matrix4 using Matrix4::approxCoordinateFrame, however, because a Matrix4 is more 
+general than a CoordinateFrame, some information may be lost.
+
+See also: G3D::Matrix4, G3D::Quat
+*/
+class CoordinateFrame {
+public:
+
+    /**
+     Takes object space points to world space.
+     */
+    Matrix3							rotation;
+
+    /**
+     Takes object space points to world space.
+     */
+    Vector3							translation;
+
+    /**
+     The direction an object "looks" relative to its own axes.
+     @deprecated This is always -1 and will be fixed at that value in future releases.
+     */
+    static const float				zLookDirection;
+
+    inline bool operator==(const CoordinateFrame& other) const {
+        return (translation == other.translation) && (rotation == other.rotation);
+    }
+
+    inline bool operator!=(const CoordinateFrame& other) const {
+        return !(*this == other);
+    }
+
+    bool fuzzyEq(const CoordinateFrame& other) const;
+
+    bool fuzzyIsIdentity() const;
+
+    bool isIdentity() const;
+
+    /**
+     Initializes to the identity coordinate frame.
+     */
+    inline CoordinateFrame() : 
+        rotation(Matrix3::identity()), translation(Vector3::zero()) {
+    }
+
+	CoordinateFrame(const Vector3& _translation) :
+        rotation(Matrix3::identity()), translation(_translation) {
+	}
+
+    CoordinateFrame(const Matrix3 &rotation, const Vector3 &translation) :
+        rotation(rotation), translation(translation) {
+    }
+
+    CoordinateFrame(const Matrix3 &rotation) :
+        rotation(rotation), translation(Vector3::zero()) {
+    }
+
+    CoordinateFrame(const CoordinateFrame &other) :
+        rotation(other.rotation), translation(other.translation) {}
+
+    /**
+      Computes the inverse of this coordinate frame.
+     */
+    inline CoordinateFrame inverse() const {
+        CoordinateFrame out;
+        out.rotation = rotation.transpose();
+        out.translation = -out.rotation * translation;
+        return out;
+    }
+
+    inline ~CoordinateFrame() {}
+
+    /** See also Matrix4::approxCoordinateFrame */
+    class Matrix4 toMatrix4() const;
+
+    /**
+     Produces an XML serialization of this coordinate frame.
+     */
+    std::string toXML() const;
+
+    /**
+     Returns the heading of the lookVector as an angle in radians relative to
+     the world -z axis.  That is, a counter-clockwise heading where north (-z) 
+     is 0 and west (-x) is PI/2.
+
+     Note that the heading ignores the Y axis, so an inverted
+     object has an inverted heading.
+     */
+    inline float getHeading() const {
+        Vector3 look = rotation.getColumn(2);
+        float angle = -(float) atan2(-look.x, look.z);
+        return angle;
+    }
+
+    /**
+     Takes the coordinate frame into object space.
+     this->inverse() * c
+     */
+    inline CoordinateFrame toObjectSpace(const CoordinateFrame& c) const {
+        return this->inverse() * c;
+    }
+
+    inline Vector4 toObjectSpace(const Vector4& v) const {
+        return this->inverse().toWorldSpace(v);
+    }
+
+    inline Vector4 toWorldSpace(const Vector4& v) const {
+        return Vector4(rotation * Vector3(v.x, v.y, v.z) + translation * v.w, v.w);
+    }
+
+    /**
+     Transforms the point into world space.
+     */
+    inline Vector3 pointToWorldSpace(const Vector3& v) const {
+        return Vector3(
+			rotation[0][0] * v[0] + rotation[0][1] * v[1] + rotation[0][2] * v[2] + translation[0],
+			rotation[1][0] * v[0] + rotation[1][1] * v[1] + rotation[1][2] * v[2] + translation[1],
+			rotation[2][0] * v[0] + rotation[2][1] * v[1] + rotation[2][2] * v[2] + translation[2]);
+    }
+
+    /**
+     Transforms the point into object space.
+     */
+	inline Vector3 pointToObjectSpace(const Vector3& v) const {
+		float p[3];
+		p[0] = v[0] - translation[0];
+		p[1] = v[1] - translation[1];
+		p[2] = v[2] - translation[2];
+		return Vector3(
+			rotation[0][0] * p[0] + rotation[1][0] * p[1] + rotation[2][0] * p[2],
+			rotation[0][1] * p[0] + rotation[1][1] * p[1] + rotation[2][1] * p[2],
+			rotation[0][2] * p[0] + rotation[1][2] * p[1] + rotation[2][2] * p[2]);
+    }
+
+    /**
+     Transforms the vector into world space (no translation).
+     */
+    inline Vector3 vectorToWorldSpace(const Vector3& v) const {
+        return rotation * v;
+    }
+
+    inline Vector3 normalToWorldSpace(const Vector3& v) const {
+        return rotation * v;
+    }
+
+    class Ray toObjectSpace(const Ray& r) const;
+
+    Ray toWorldSpace(const Ray& r) const;
+
+    /**
+     Transforms the vector into object space (no translation).
+     */
+    inline Vector3 vectorToObjectSpace(const Vector3 &v) const {
+        // Multiply on the left (same as rotation.transpose() * v)
+        return v * rotation;
+    }
+
+    inline Vector3 normalToObjectSpace(const Vector3 &v) const {
+        // Multiply on the left (same as rotation.transpose() * v)
+        return v * rotation;
+    }
+
+    void pointToWorldSpace(const Array<Vector3>& v, Array<Vector3>& vout) const;
+
+    void normalToWorldSpace(const Array<Vector3>& v, Array<Vector3>& vout) const;
+
+    void vectorToWorldSpace(const Array<Vector3>& v, Array<Vector3>& vout) const;
+
+    void pointToObjectSpace(const Array<Vector3>& v, Array<Vector3>& vout) const;
+
+    void normalToObjectSpace(const Array<Vector3>& v, Array<Vector3>& vout) const;
+
+    void vectorToObjectSpace(const Array<Vector3>& v, Array<Vector3>& vout) const;
+
+    class Box toWorldSpace(const class AABox& b) const;
+
+    class Box toWorldSpace(const class Box& b) const;
+
+    class Cylinder toWorldSpace(const class Cylinder& b) const;
+
+    class Capsule toWorldSpace(const class Capsule& b) const;
+
+    class Plane toWorldSpace(const class Plane& p) const;
+
+    class Sphere toWorldSpace(const class Sphere& b) const;
+
+    class Triangle toWorldSpace(const class Triangle& t) const;
+
+    class Box toObjectSpace(const AABox& b) const;
+
+    class Box toObjectSpace(const Box& b) const;
+
+    class Plane toObjectSpace(const Plane& p) const;
+ 
+    class Sphere toObjectSpace(const Sphere& b) const;
+
+    Triangle toObjectSpace(const Triangle& t) const;
+
+    /** Compose: create the transformation that is <I>other</I> followed by <I>this</I>.*/
+    CoordinateFrame operator*(const CoordinateFrame &other) const {
+        return CoordinateFrame(rotation * other.rotation,
+                               pointToWorldSpace(other.translation));
+    }
+
+    CoordinateFrame operator+(const Vector3& v) const {
+        return CoordinateFrame(rotation, translation + v);
+    }
+
+    CoordinateFrame operator-(const Vector3& v) const {
+        return CoordinateFrame(rotation, translation - v);
+    }
+
+    void lookAt(const Vector3& target);
+
+    void lookAt(
+        const Vector3&  target,
+        Vector3         up);
+
+    /** @deprecated See lookVector */
+	inline Vector3 getLookVector() const {
+		return rotation.getColumn(2) * zLookDirection;
+	}
+
+    /** The direction this camera is looking (its negative z axis)*/
+	inline Vector3 lookVector() const {
+		return rotation.getColumn(2) * zLookDirection;
+	}
+
+    /** Returns the ray starting at the camera origin travelling in direction CoordinateFrame::lookVector. */
+    class Ray lookRay() const;
+
+    /** Up direction for this camera (its y axis). */
+    inline Vector3 upVector() const {
+        return rotation.getColumn(1);
+    }
+
+    /**
+     If a viewer looks along the look vector, this is the viewer's "left"
+     @deprecated leftVector
+     */
+    inline Vector3 getLeftVector() const {
+		return -rotation.getColumn(0);
+	}
+
+    /** @deprecated See rightVector */
+    inline Vector3 getRightVector() const {
+		return rotation.getColumn(0);
+	}
+
+    /**
+     If a viewer looks along the look vector, this is the viewer's "left".
+     Useful for strafing motions and building alternative coordinate frames.
+     */
+    inline Vector3 leftVector() const {
+		return -rotation.getColumn(0);
+    }
+
+    inline Vector3 rightVector() const {
+		return rotation.getColumn(0);
+    }
+
+    /**
+     Linearly interpolates between two coordinate frames, using
+     Quat::slerp for the rotations.
+     */
+    CoordinateFrame lerp(
+        const CoordinateFrame&  other,
+        float                   alpha) const;
+
+};
+
+} // namespace
+
+#endif
diff --git a/dep/include/g3dlite/G3D/Crypto.h b/dep/include/g3dlite/G3D/Crypto.h
new file mode 100644
index 00000000000..f4e7bfad6cd
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Crypto.h
@@ -0,0 +1,46 @@
+/** 
+  @file Crypto.h
+ 
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+ 
+
+  @created 2006-03-29
+  @edited  2006-04-06
+ */
+
+#ifndef G3D_CRYPTO_H
+#define G3D_CRYPTO_H
+
+#include "G3D/platform.h"
+#include "G3D/g3dmath.h"
+#include <string>
+
+namespace G3D {
+
+/** Cryptography and hashing helper functions */
+class Crypto {
+public:
+
+    /**
+     Computes the CRC32 value of a byte array.  CRC32 is designed to be a hash
+     function that produces different values for similar strings.
+
+     This implementation is compatible with PKZIP and GZIP.
+
+     Based on http://www.gamedev.net/reference/programming/features/crc32/
+    */
+    static uint32 crc32(const void* bytes, size_t numBytes);
+
+    /**
+     Returns the nth prime less than 2000 in constant time.  The first prime has index
+     0 and is the number 2.
+     */
+    static int smallPrime(int n);
+
+    /** Returns 1 + the largest value that can be passed to smallPrime. */
+    static int numSmallPrimes();
+};
+
+}
+
+#endif
diff --git a/dep/include/g3dlite/G3D/GCamera.h b/dep/include/g3dlite/G3D/GCamera.h
new file mode 100644
index 00000000000..3205f3904b1
--- /dev/null
+++ b/dep/include/g3dlite/G3D/GCamera.h
@@ -0,0 +1,251 @@
+/**
+  @file GCamera.h
+
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+
+  @created 2001-06-02
+  @edited  2006-02-11
+*/
+
+#ifndef G3D_GCAMERA_H
+#define G3D_GCAMERA_H
+
+#include "G3D/platform.h"
+#include "G3D/CoordinateFrame.h"
+#include "G3D/Vector3.h"
+#include "G3D/Plane.h"
+
+namespace G3D {
+
+/**
+  There is a viewport of width x height size in world space that corresponds to
+  a screenWidth x screenHeight pixel grid on a
+  renderDevice->getWidth() x renderDevice->getHeight()
+  window.
+
+  All viewport arguments are the pixel bounds of the viewport-- e.g.,
+  RenderDevice::getViewport().
+ */
+class GCamera  {
+private:
+
+    /**
+    Vertical field of view (in radians)
+    */
+    float						fieldOfView;
+
+    /** 
+     The image plane depth corresponding to a vertical field of 
+     view, where the film size is 1x1.  
+     */
+    float						imagePlaneDepth;
+
+    /**
+     Clipping plane, *not* imaging plane.  Positive numbers.
+     */
+    float						nearPlane;
+
+    /**
+     Positive 
+     */
+    float						farPlane;
+
+    CoordinateFrame             cframe;
+
+public:
+
+    class Frustum {
+    public:
+        class Face {
+        public:
+            /** Counter clockwise indices into vertexPos */
+            int             vertexIndex[4];
+
+            /** The plane containing the face. */
+            Plane           plane;
+        };
+        
+        /** The vertices, in homogeneous space.  If w == 0,
+            a vertex is at infinity. */
+        Array<Vector4>      vertexPos;
+
+        /** The faces in the frustum.  When the
+            far plane is at infinity, there are 5 faces,
+            otherwise there are 6.  The faces are in the order
+            N,R,L,B,T,[F].
+            */
+        Array<Face>         faceArray;
+    };
+
+	GCamera();
+
+    virtual ~GCamera();
+
+
+	CoordinateFrame getCoordinateFrame() const;
+	void getCoordinateFrame(CoordinateFrame& c) const;
+	void setCoordinateFrame(const CoordinateFrame& c);
+           
+   /**
+	 Sets the horizontal field of view, in radians.  The 
+     initial angle is toRadians(55).
+	 <UL>
+	  <LI> toRadians(50) - Telephoto
+	  <LI> toRadians(110) - Normal
+	  <LI> toRadians(140) - Wide angle
+	 </UL>
+ 	*/
+	void setFieldOfView(float angle);
+
+	/**
+	Sets the field of view based on a desired image plane depth
+	(<I>s'</I>) and film dimensions in world space.  Depth must be positive.  Width,
+	depth, and height are measured in the same units (meters are
+	recommended).  The field of view will span the diagonal to the
+	image.<P> <I>Note</I>: to simulate a 35mm GCamera, set width =
+	0.36 mm and height = 0.24 mm.  The width and height used are
+	generally not the pixel dimensions of the image.  
+	*/
+	void setImagePlaneDepth(
+        float                                   depth,
+        const class Rect2D&                     viewport);
+
+	inline double getFieldOfView() const {
+		return fieldOfView;
+	}
+
+    /**
+     Projects a world space point onto a width x height screen.  The
+     returned coordinate uses pixmap addressing: x = right and y =
+     down.  The resulting z value is <I>rhw</I>.
+
+     If the point is behind the camera, Vector3::inf() is returned.
+     */
+    G3D::Vector3 project(
+        const G3D::Vector3&                     point,
+        const class Rect2D&                     viewport) const;
+
+    /**
+     Returns the pixel area covered by a shape of the given
+     world space area at the given z value (z must be negative).
+     */
+    float worldToScreenSpaceArea(float area, float z, const class Rect2D& viewport) const;
+
+    /**
+     Returns the world space 3D viewport corners.  These
+     are at the near clipping plane.  The corners are constructed
+     from the nearPlaneZ, getViewportWidth, and getViewportHeight.
+     "left" and "right" are from the GCamera's perspective.
+     */
+    void get3DViewportCorners(
+        const class Rect2D&                     viewport,
+        Vector3&                                outUR,
+        Vector3&                                outUL,
+        Vector3&                                outLL,
+        Vector3&                                outLR) const;
+
+    /**
+     Returns the image plane depth, <I>s'</I>, given the current field
+     of view for film of dimensions width x height.  See
+     setImagePlaneDepth for a discussion of worldspace values width and height. 
+    */
+    float getImagePlaneDepth(
+        const class Rect2D&                     viewport) const;
+
+
+    /**
+      Returns the world space ray passing through the center of pixel
+      (x, y) on the image plane.  The pixel x and y axes are opposite
+      the 3D object space axes: (0,0) is the upper left corner of the screen.
+      They are in viewport coordinates, not screen coordinates.
+
+
+      Integer (x, y) values correspond to
+      the upper left corners of pixels.  If you want to cast rays
+      through pixel centers, add 0.5 to x and y.        
+    */
+    Ray worldRay(
+        float                                  x,
+        float                                  y,
+        const class Rect2D&                     viewport) const;
+
+
+    /**
+      Returns a negative z-value.
+     */
+    inline float getNearPlaneZ() const {
+        return -nearPlane;
+    }
+
+    /**
+     Returns a negative z-value.
+     */
+    inline float getFarPlaneZ() const {
+        return -farPlane;
+    }
+
+	inline void setFarPlaneZ(float z) {
+		debugAssert(z < 0);
+		farPlane = -z;
+	}
+
+	inline void setNearPlaneZ(float z) {
+		debugAssert(z < 0);
+		nearPlane = -z;
+	}
+
+    /**
+     Returns the GCamera space width of the viewport.
+     */
+    float getViewportWidth(
+        const class Rect2D&                     viewport) const;
+
+    /**
+     Returns the GCamera space height of the viewport.
+     */
+    float getViewportHeight(       
+        const class Rect2D&                     viewport) const;
+
+    /**
+     Read back a GCamera space z-value at pixel (x, y) from the depth buffer.
+    double getZValue(
+        double			x,
+        double			y,
+        const class Rect2D&                     viewport,
+        double			polygonOffset = 0) const;
+     */
+
+    void setPosition(const Vector3& t);
+
+    void lookAt(const Vector3& position, const Vector3& up = Vector3::unitY());
+
+   /**
+    Returns the clipping planes of the frustum, in world space.  
+    The planes have normals facing <B>into</B> the view frustum.
+
+    The plane order is guaranteed to be:
+      Near, Right, Left, Top, Bottom, [Far]
+
+    If the far plane is at infinity, the resulting array will have 
+    5 planes, otherwise there will be 6.
+
+    The viewport is used only to determine the aspect ratio of the screen; the
+    absolute dimensions and xy values don't matter.
+    */
+   void getClipPlanes(
+       const Rect2D& viewport,
+       Array<Plane>& outClip) const;
+
+   /**
+    Returns the world space view frustum, which is a truncated pyramid describing
+    the volume of space seen by this camera.
+    */
+   void getFrustum(const Rect2D& viewport, GCamera::Frustum& f) const;
+
+   GCamera::Frustum frustum(const Rect2D& viewport) const;
+
+};
+
+} // namespace G3D
+
+#endif
diff --git a/dep/include/g3dlite/G3D/Line.h b/dep/include/g3dlite/G3D/Line.h
new file mode 100644
index 00000000000..5dd82c19bc7
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Line.h
@@ -0,0 +1,85 @@
+/**
+ @file Line.h
+ 
+ Line class
+ 
+ @maintainer Morgan McGuire, matrix@graphics3d.com
+ 
+ @created 2001-06-02
+ @edited  2006-02-28
+ */
+
+#ifndef G3D_LINE_H
+#define G3D_LINE_H
+
+#include "G3D/platform.h"
+#include "G3D/Vector3.h"
+
+namespace G3D {
+
+class Plane;
+
+/**
+ An infinite 3D line.
+ */
+class Line {
+protected:
+
+    Vector3 _point;
+    Vector3 _direction;
+
+    Line(const Vector3& point, const Vector3& direction) {
+        _point     = point;
+        _direction = direction.direction();
+    }
+
+public:
+
+    /** Undefined (provided for creating Array<Line> only) */
+    inline Line() {}
+
+    virtual ~Line() {}
+
+    /**
+      Constructs a line from two (not equal) points.
+     */
+    static Line fromTwoPoints(const Vector3 &point1, const Vector3 &point2) {
+        return Line(point1, point2 - point1);
+    }
+
+    /**
+      Creates a line from a point and a (nonzero) direction.
+     */
+    static Line fromPointAndDirection(const Vector3& point, const Vector3& direction) {
+        return Line(point, direction);
+    }
+
+    /**
+      Returns the closest point on the line to point.
+     */
+    Vector3 closestPoint(const Vector3& pt) const;
+
+    /**
+      Returns the distance between point and the line
+     */
+    double distance(const Vector3& point) const {
+        return (closestPoint(point) - point).magnitude();
+    }
+
+    /** Returns a point on the line */
+    Vector3 point() const;
+
+    /** Returns the direction (or negative direction) of the line */
+    Vector3 direction() const;
+
+    /**
+     Returns the point where the line and plane intersect.  If there
+     is no intersection, returns a point at infinity.
+     */
+    Vector3 intersection(const Plane &plane) const;
+};
+
+};// namespace
+
+
+#endif
diff --git a/dep/include/g3dlite/G3D/Matrix3.h b/dep/include/g3dlite/G3D/Matrix3.h
new file mode 100644
index 00000000000..19eda497a15
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Matrix3.h
@@ -0,0 +1,311 @@
+/**
+  @file Matrix3.h
+ 
+  3x3 matrix class
+ 
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+ 
+  @cite Portions based on Dave Eberly's Magic Software Library at <A HREF="http://www.magic-software.com">http://www.magic-software.com</A>
+ 
+  @created 2001-06-02
+  @edited  2006-04-05
+ */
+
+#ifndef G3D_MATRIX3_H
+#define G3D_MATRIX3_H
+
+#include "G3D/platform.h"
+#include "G3D/System.h"
+#include "G3D/Vector3.h"
+#include "G3D/Vector4.h"
+
+namespace G3D {
+
+/**
+  3x3 matrix.  Do not subclass.
+ */
+class Matrix3 {
+private:
+
+    float elt[3][3];
+
+    // Hidden operators
+    bool operator<(const Matrix3&) const;
+    bool operator>(const Matrix3&) const;
+    bool operator<=(const Matrix3&) const;
+    bool operator>=(const Matrix3&) const;
+
+public:
+
+    /** Initial values are undefined for performance.  See also 
+        Matrix3::zero(), Matrix3::identity(), Matrix3::fromAxisAngle, etc.*/
+    inline Matrix3() {}
+
+    Matrix3 (const float aafEntry[3][3]);
+    Matrix3 (const Matrix3& rkMatrix);
+    Matrix3 (float fEntry00, float fEntry01, float fEntry02,
+             float fEntry10, float fEntry11, float fEntry12,
+             float fEntry20, float fEntry21, float fEntry22);
+
+    bool fuzzyEq(const Matrix3& b) const;
+
+    /** Constructs a matrix from a quaternion.
+        @cite Graphics Gems II, p. 351--354
+ 	    @cite Implementation from Watt and Watt, pg 362*/
+    Matrix3(const class Quat& q);
+
+
+    /**
+     Sets all elements.
+     */
+    void set(float fEntry00, float fEntry01, float fEntry02,
+             float fEntry10, float fEntry11, float fEntry12,
+             float fEntry20, float fEntry21, float fEntry22);
+
+    /**
+     * member access, allows use of construct mat[r][c]
+     */
+    inline float* operator[] (int iRow) {
+        debugAssert(iRow >= 0);
+        debugAssert(iRow < 3);
+        return (float*)&elt[iRow][0];
+    }
+
+    inline const float* operator[] (int iRow) const {
+        debugAssert(iRow >= 0);
+        debugAssert(iRow < 3);
+        return (const float*)&elt[iRow][0];
+    }
+
+    inline operator float* () {
+        return (float*)&elt[0][0];
+    }
+
+    inline operator const float* () const{
+        return (const float*)&elt[0][0];
+    }
+
+    Vector3 getColumn (int iCol) const;
+    Vector3 getRow (int iRow) const;
+    void setColumn(int iCol, const Vector3 &vector);
+    void setRow(int iRow, const Vector3 &vector);
+
+    // assignment and comparison
+    inline Matrix3& operator= (const Matrix3& rkMatrix) {
+        System::memcpy(elt, rkMatrix.elt, 9 * sizeof(float));
+        return *this;
+    }
+
+    bool operator== (const Matrix3& rkMatrix) const;
+    bool operator!= (const Matrix3& rkMatrix) const;
+
+    // arithmetic operations
+    Matrix3 operator+ (const Matrix3& rkMatrix) const;
+    Matrix3 operator- (const Matrix3& rkMatrix) const;
+    /** Matrix-matrix multiply */
+    Matrix3 operator* (const Matrix3& rkMatrix) const;
+    Matrix3 operator- () const;
+
+    Matrix3& operator+= (const Matrix3& rkMatrix);
+    Matrix3& operator-= (const Matrix3& rkMatrix);
+    Matrix3& operator*= (const Matrix3& rkMatrix);
+
+    /**
+     * matrix * vector [3x3 * 3x1 = 3x1]
+     */
+    inline Vector3 operator* (const Vector3& v) const {
+        Vector3 kProd;
+
+        for (int r = 0; r < 3; ++r) {
+            kProd[r] =
+                elt[r][0] * v[0] +
+                elt[r][1] * v[1] +
+                elt[r][2] * v[2];
+        }
+
+        return kProd;
+    }
+
+
+    /**
+     * vector * matrix [1x3 * 3x3 = 1x3]
+     */
+    friend Vector3 operator* (const Vector3& rkVector,
+                              const Matrix3& rkMatrix);
+
+    /**
+     * matrix * scalar
+     */
+    Matrix3 operator* (float fScalar) const;
+
+    /** scalar * matrix */
+    friend Matrix3 operator* (double fScalar, const Matrix3& rkMatrix);
+    friend Matrix3 operator* (float fScalar, const Matrix3& rkMatrix);
+    friend Matrix3 operator* (int fScalar, const Matrix3& rkMatrix);
+
+private:
+    /** Multiplication where out != A and out != B */
+    static void _mul(const Matrix3& A, const Matrix3& B, Matrix3& out);
+public:
+
+    /** Optimized implementation of out = A * B.  It is safe (but slow) to call 
+        with A, B, and out possibly pointer equal to one another.*/
+    // This is a static method so that it is not ambiguous whether "this"
+    // is an input or output argument.
+    inline static void mul(const Matrix3& A, const Matrix3& B, Matrix3& out) {
+        if ((&out == &A) || (&out == &B)) {
+            // We need a temporary anyway, so revert to the stack method.
+            out = A * B;
+        } else {
+            // Optimized in-place multiplication.
+            _mul(A, B, out);
+        }
+    }
+
+private:
+    static void _transpose(const Matrix3& A, Matrix3& out);
+public:
+
+    /** Optimized implementation of out = A.transpose().  It is safe (but slow) to call 
+        with A and out possibly pointer equal to one another.
+    
+        Note that <CODE>A.transpose() * v</CODE> can be computed 
+        more efficiently as <CODE>v * A</CODE>.
+    */
+    inline static void transpose(const Matrix3& A, Matrix3& out) {
+        if (&A == &out) {
+            out = A.transpose();
+        } else {
+            _transpose(A, out);
+        }
+    }
+
+    /** Returns true if the rows and column L2 norms are 1.0 and the rows are orthogonal. */
+    bool isOrthonormal() const;
+
+    Matrix3 transpose () const;
+    bool inverse (Matrix3& rkInverse, float fTolerance = 1e-06) const;
+    Matrix3 inverse (float fTolerance = 1e-06) const;
+    float determinant () const;
+
+    /** singular value decomposition */
+    void singularValueDecomposition (Matrix3& rkL, Vector3& rkS,
+                                     Matrix3& rkR) const;
+    /** singular value decomposition */
+    void singularValueComposition (const Matrix3& rkL,
+                                   const Vector3& rkS, const Matrix3& rkR);
+
+    /** Gram-Schmidt orthonormalization (applied to columns of rotation matrix) */
+    void orthonormalize();
+
+    /** orthogonal Q, diagonal D, upper triangular U stored as (u01,u02,u12) */
+    void qDUDecomposition (Matrix3& rkQ, Vector3& rkD,
+                           Vector3& rkU) const;
+
+    float spectralNorm () const;
+
+    /** matrix must be orthonormal */
+    void toAxisAngle(Vector3& rkAxis, float& rfRadians) const;
+
+    static Matrix3 fromAxisAngle(const Vector3& rkAxis, float fRadians);
+
+    /**
+     * The matrix must be orthonormal.  The decomposition is yaw*pitch*roll
+     * where yaw is rotation about the Up vector, pitch is rotation about the
+     * right axis, and roll is rotation about the Direction axis.
+     */
+    bool toEulerAnglesXYZ (float& rfYAngle, float& rfPAngle,
+                           float& rfRAngle) const;
+    bool toEulerAnglesXZY (float& rfYAngle, float& rfPAngle,
+                           float& rfRAngle) const;
+    bool toEulerAnglesYXZ (float& rfYAngle, float& rfPAngle,
+                           float& rfRAngle) const;
+    bool toEulerAnglesYZX (float& rfYAngle, float& rfPAngle,
+                           float& rfRAngle) const;
+    bool toEulerAnglesZXY (float& rfYAngle, float& rfPAngle,
+                           float& rfRAngle) const;
+    bool toEulerAnglesZYX (float& rfYAngle, float& rfPAngle,
+                           float& rfRAngle) const;
+    static Matrix3 fromEulerAnglesXYZ (float fYAngle, float fPAngle, float fRAngle);
+    static Matrix3 fromEulerAnglesXZY (float fYAngle, float fPAngle, float fRAngle);
+    static Matrix3 fromEulerAnglesYXZ (float fYAngle, float fPAngle, float fRAngle);
+    static Matrix3 fromEulerAnglesYZX (float fYAngle, float fPAngle, float fRAngle);
+    static Matrix3 fromEulerAnglesZXY (float fYAngle, float fPAngle, float fRAngle);
+    static Matrix3 fromEulerAnglesZYX (float fYAngle, float fPAngle, float fRAngle);
+
+    /** eigensolver, matrix must be symmetric */
+    void eigenSolveSymmetric (float afEigenvalue[3],
+                              Vector3 akEigenvector[3]) const;
+
+    static void tensorProduct (const Vector3& rkU, const Vector3& rkV,
+                               Matrix3& rkProduct);
+	std::string toString() const;
+
+    static const float EPSILON; 
+
+    // Special values.
+    // The unguaranteed order of initialization of static variables across 
+    // translation units can be a source of annoying bugs, so now the static
+    // special values (like Vector3::ZERO, Color3::WHITE, ...) are wrapped
+    // inside static functions that return references to them. 
+    // These functions are intentionally not inlined, because: 
+    // "You might be tempted to write [...] them as inline functions 
+    // inside their respective header files, but this is something you 
+    // must definitely not do. An inline function can be duplicated 
+    // in every file in which it appears – and this duplication 
+    // includes the static object definition. Because inline functions 
+    // automatically default to internal linkage, this would result in 
+    // having multiple static objects across the various translation 
+    // units, which would certainly cause problems. So you must 
+    // ensure that there is only one definition of each wrapping 
+    // function, and this means not making the wrapping functions inline",
+    // according to Chapter 10 of "Thinking in C++, 2nd ed. Volume 1" by Bruce Eckel, 
+    // http://www.mindview.net/
+    static const Matrix3& zero();
+    static const Matrix3& identity(); 
+
+    // Deprecated. 
+    /** @deprecated Use Matrix3::zero() */
+    static const Matrix3 ZERO;
+    /** @deprecated Use Matrix3::identity() */
+    static const Matrix3 IDENTITY;
+
+protected:
+    // support for eigensolver
+    void tridiagonal (float afDiag[3], float afSubDiag[3]);
+    bool qLAlgorithm (float afDiag[3], float afSubDiag[3]);
+
+    // support for singular value decomposition
+    static const float ms_fSvdEpsilon;
+    static const int ms_iSvdMaxIterations;
+    static void bidiagonalize (Matrix3& kA, Matrix3& kL,
+                               Matrix3& kR);
+    static void golubKahanStep (Matrix3& kA, Matrix3& kL,
+                                Matrix3& kR);
+
+    // support for spectral norm
+    static float maxCubicRoot (float afCoeff[3]);
+
+};
+
+
+//----------------------------------------------------------------------------
+/**  <code>v * M == M.transpose() * v</code> */
+inline Vector3 operator* (const Vector3& rkPoint, const Matrix3& rkMatrix) {
+    Vector3 kProd;
+
+    for (int r = 0; r < 3; ++r) {
+        kProd[r] =
+            rkPoint[0] * rkMatrix.elt[0][r] +
+            rkPoint[1] * rkMatrix.elt[1][r] +
+            rkPoint[2] * rkMatrix.elt[2][r];
+    }
+
+    return kProd;
+}
+
+
+} // namespace
+
+#endif
+
diff --git a/dep/include/g3dlite/G3D/Plane.h b/dep/include/g3dlite/G3D/Plane.h
new file mode 100644
index 00000000000..f00c15c0988
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Plane.h
@@ -0,0 +1,156 @@
+/**
+ @file Plane.h
+
+ Plane class
+
+ @maintainer Morgan McGuire, matrix@graphics3d.com
+
+ @created 2001-06-02
+ @edited  2004-07-18
+*/
+
+#ifndef G3D_PLANE_H
+#define G3D_PLANE_H
+
+#include "G3D/platform.h"
+#include "G3D/Vector3.h"
+#include "G3D/Vector4.h"
+
+namespace G3D {
+
+/**
+ An infinite 2D plane in 3D space.
+ */
+class Plane {
+private:
+
+    /** normal.Dot(x,y,z) = distance */
+    Vector3						_normal;
+    float						_distance;
+
+    /**
+     Assumes the normal has unit length.
+     */
+    Plane(const Vector3& n, float d) : _normal(n), _distance(d) {
+    }
+
+public:
+
+    Plane() : _normal(Vector3::unitY()), _distance(0) {
+    }
+
+    /**
+     Constructs a plane from three points.
+     */
+    Plane(
+        const Vector3&      point0,
+        const Vector3&      point1,
+        const Vector3&      point2);
+
+    /**
+     Constructs a plane from three points, where at most two are
+     at infinity (w = 0, not xyz = inf).
+     */
+    Plane(
+        Vector4             point0,
+        Vector4             point1,
+        Vector4             point2);
+
+    /**
+     The normal will be unitized.
+     */
+    Plane(
+        const Vector3&      __normal,
+        const Vector3&      point);
+
+    static Plane fromEquation(float a, float b, float c, float d);
+
+    virtual ~Plane() {}
+
+    /**
+     Returns true if point is on the side the normal points to or 
+     is in the plane.
+     */
+    inline bool halfSpaceContains(Vector3 point) const {
+        // Clamp to a finite range for testing
+        point = point.clamp(Vector3::minFinite(), Vector3::maxFinite());
+
+        // We can get away with putting values *at* the limits of the float32 range into
+        // a dot product, since the dot product is carried out on float64.
+        return _normal.dot(point) >= _distance;
+    }
+
+    /**
+     Returns true if point is on the side the normal points to or 
+     is in the plane.
+     */
+    inline bool halfSpaceContains(const Vector4& point) const {
+        if (point.w == 0) {
+            return _normal.dot(point.xyz()) > 0;
+        } else {
+            return halfSpaceContains(point.xyz() / point.w);
+        }
+    }
+
+    /**
+     Returns true if point is on the side the normal points to or 
+     is in the plane.  Only call on finite points.  Faster than halfSpaceContains.
+     */
+    inline bool halfSpaceContainsFinite(const Vector3& point) const {
+        debugAssert(point.isFinite());
+        return _normal.dot(point) >= _distance;
+    }
+
+    /**
+     Returns true if the point is nearly in the plane.
+     */
+    inline bool fuzzyContains(const Vector3 &point) const {
+        return fuzzyEq(point.dot(_normal), _distance);
+    }
+
+	inline const Vector3& normal() const {
+		return _normal;
+	}
+
+    /**
+      Returns distance from point to plane. Distance is negative if point is behind (not in plane in direction opposite normal) the plane.
+    */
+    inline float distance(const Vector3& x) const {
+        return (_normal.dot(x) - _distance);
+    }
+
+    inline Vector3 closestPoint(const Vector3& x) const {
+        return x + (_normal * (-distance(x)));
+    }
+
+    /** Returns normal * distance from origin */
+    Vector3 center() const {
+        return _normal * _distance;
+    }
+
+    /**
+     Inverts the facing direction of the plane so the new normal
+     is the inverse of the old normal.
+     */
+    void flip();
+
+    /**
+      Returns the equation in the form:
+
+      <CODE>normal.Dot(Vector3(<I>x</I>, <I>y</I>, <I>z</I>)) + d = 0</CODE>
+     */
+    void getEquation(Vector3 &normal, double& d) const;
+    void getEquation(Vector3 &normal, float& d) const;
+
+    /**
+      ax + by + cz + d = 0
+     */
+    void getEquation(double& a, double& b, double& c, double& d) const;
+    void getEquation(float& a, float& b, float& c, float& d) const;
+
+    std::string toString() const;
+};
+
+} // namespace
+
+#endif
diff --git a/dep/include/g3dlite/G3D/Quat.h b/dep/include/g3dlite/G3D/Quat.h
new file mode 100644
index 00000000000..b852fe0fd47
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Quat.h
@@ -0,0 +1,702 @@
+/**
+  @file Quat.h
+ 
+  Quaternion
+  
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+  
+  @created 2002-01-23
+  @edited  2006-05-10
+ */
+
+#ifndef G3D_QUAT_H
+#define G3D_QUAT_H
+
+#include "G3D/platform.h"
+#include "G3D/g3dmath.h"
+#include "G3D/Vector3.h"
+#include "G3D/Matrix3.h"
+#include <string>
+
+namespace G3D {
+
+/**
+  Unit quaternions are used in computer graphics to represent
+  rotation about an axis.  Any 3x3 rotation matrix can
+  be stored as a quaternion.
+
+  A quaternion represents the sum of a real scalar and
+  an imaginary vector: ix + jy + kz + w.  A unit quaternion
+  representing a rotation by A about axis v has the form 
+  [sin(A/2)*v, cos(A/2)].  For a unit quaternion, q.conj() == q.inverse()
+  is a rotation by -A about v.  -q is the same rotation as q
+  (negate both the axis and angle).  
+  
+  A non-unit quaterion q represents the same rotation as
+  q.unitize() (Dam98 pg 28).
+
+  Although quaternion-vector operations (eg. Quat + Vector3) are
+  well defined, they are not supported by this class because
+  they typically are bugs when they appear in code.
+
+  Do not subclass.
+
+  <B>BETA API -- subject to change</B>
+  @cite Erik B. Dam, Martin Koch, Martin Lillholm, Quaternions, Interpolation and Animation.  Technical Report DIKU-TR-98/5, Department of Computer Science, University of Copenhagen, Denmark.  1998.
+ */
+class Quat {
+private:
+    // Hidden operators
+    bool operator<(const Quat&) const;
+    bool operator>(const Quat&) const;
+    bool operator<=(const Quat&) const;
+    bool operator>=(const Quat&) const;
+
+public:
+
+    /**
+     q = [sin(angle / 2) * axis, cos(angle / 2)]
+    
+     In Watt & Watt's notation, s = w, v = (x, y, z)
+     In the Real-Time Rendering notation, u = (x, y, z), w = w 
+     */
+    float x, y, z, w;
+
+    /**
+     Initializes to a zero degree rotation.
+     */
+    inline Quat() : x(0), y(0), z(0), w(1) {}
+
+    Quat(
+        const Matrix3& rot);
+
+    inline Quat(float _x, float _y, float _z, float _w) :
+        x(_x), y(_y), z(_z), w(_w) {}
+
+    /** Defaults to a pure vector quaternion */
+    inline Quat(const Vector3& v, float _w = 0) : x(v.x), y(v.y), z(v.z), w(_w) {
+    }
+
+    /**
+     The real part of the quaternion.
+     */
+    inline const float& real() const {
+        return w;
+    }
+
+    inline float& real() {
+        return w;
+    }
+
+    /** Note: two quats can represent the Quat::sameRotation and not be equal. */
+	bool fuzzyEq(const Quat& q) {
+		return G3D::fuzzyEq(x, q.x) && G3D::fuzzyEq(y, q.y) && G3D::fuzzyEq(z, q.z) && G3D::fuzzyEq(w, q.w);
+	}
+
+    /** True if these quaternions represent the same rotation (note that every rotation is 
+        represented by two values; q and -q).
+      */
+    bool sameRotation(const Quat& q) {
+        return fuzzyEq(q) || fuzzyEq(-q);
+    }
+
+	inline Quat operator-() const {
+		return Quat(-x, -y, -z, -w);
+	}
+
+    /**
+     Returns the imaginary part (x, y, z)
+     */
+    inline const Vector3& imag() const {
+        return *(reinterpret_cast<const Vector3*>(this));
+    }
+
+    inline Vector3& imag() {
+        return *(reinterpret_cast<Vector3*>(this));
+    }
+
+    /** q = [sin(angle/2)*axis, cos(angle/2)] */
+    static Quat fromAxisAngleRotation(
+        const Vector3&      axis,
+        float               angle);
+
+    /** Returns the axis and angle of rotation represented
+        by this quaternion (i.e. q = [sin(angle/2)*axis, cos(angle/2)]) */
+    void toAxisAngleRotation(
+        Vector3&            axis,
+        double&             angle) const;
+
+    void toAxisAngleRotation(
+        Vector3&            axis,
+        float&              angle) const {
+		double d;
+		toAxisAngleRotation(axis, d);
+		angle = (float)d;
+	}
+
+    Matrix3 toRotationMatrix() const;
+
+    void toRotationMatrix(
+        Matrix3&            rot) const;
+    
+    /**
+     Spherical linear interpolation: linear interpolation along the 
+     shortest (3D) great-circle route between two quaternions.
+
+     Note: Correct rotations are expected between 0 and PI in the right order.
+
+     @cite Based on Game Physics -- David Eberly pg 538-540
+     @param threshold Critical angle between between rotations at which
+	        the algorithm switches to normalized lerp, which is more
+			numerically stable in those situations. 0.0 will always slerp. 
+     */
+    Quat slerp(
+        const Quat&         other,
+        float               alpha,
+        float               threshold = 0.05f) const;
+
+	/** Normalized linear interpolation of quaternion components. */
+	Quat nlerp(const Quat& other, float alpha) const;
+
+    /**
+     Negates the imaginary part.
+     */
+    inline Quat conj() const {
+        return Quat(-x, -y, -z, w);
+    }
+
+    inline float sum() const {
+        return x + y + z + w;
+    }
+
+    inline float average() const {
+        return sum() / 4.0f;
+    }
+
+    inline Quat operator*(float s) const {
+        return Quat(x * s, y * s, z * s, w * s);
+    }
+
+	/** @cite Based on Watt & Watt, page 360 */
+    friend Quat operator* (float s, const Quat& q);
+
+    inline Quat operator/(float s) const {
+        return Quat(x / s, y / s, z / s, w / s);
+    }
+
+    inline float dot(const Quat& other) const {
+        return (x * other.x) + (y * other.y) + (z * other.z) + (w * other.w);
+    }
+
+    /** Note that q<SUP>-1</SUP> = q.conj() for a unit quaternion. 
+        @cite Dam99 page 13 */
+    inline Quat inverse() const {
+        return conj() / dot(*this);
+    }
+
+    Quat operator-(const Quat& other) const;
+
+    Quat operator+(const Quat& other) const;
+
+    /**
+     Quaternion multiplication (composition of rotations).
+     Note that this does not commute.
+     */
+    Quat operator*(const Quat& other) const;
+
+    /* (*this) * other.inverse() */
+    Quat operator/(const Quat& other) const {
+        return (*this) * other.inverse();
+    }
+
+
+    /** Is the magnitude nearly 1.0? */
+    inline bool isUnit(float tolerance = 1e-5) const {
+        return abs(dot(*this) - 1.0f) < tolerance;
+    }
+    
+
+    inline float magnitude() const {
+        return sqrtf(dot(*this));
+    }
+
+    inline Quat log() const {
+        if ((x == 0) && (y == 0) && (z == 0)) {
+            if (w > 0) {
+                return Quat(0, 0, 0, ::logf(w));
+            } else if (w < 0) {
+                // Log of a negative number.  Multivalued, any number of the form
+                // (PI * v, ln(-q.w))
+                return Quat((float)G3D_PI, 0, 0, ::logf(-w));
+            } else {
+                 // log of zero!
+                 return Quat((float)nan(), (float)nan(), (float)nan(), (float)nan());
+            }
+        } else {
+            // Partly imaginary.
+            float imagLen = sqrtf(x * x + y * y + z * z);
+            float len = sqrtf(imagLen * imagLen + w * w);
+            float theta = atan2f(imagLen, (float)w);
+            float t = theta / imagLen;
+            return Quat(t * x, t * y, t * z, ::logf(len));
+        }
+    }
+    /** log q = [Av, 0] where q = [sin(A) * v, cos(A)].
+        Only for unit quaternions 
+        debugAssertM(isUnit(), "Log only defined for unit quaternions");
+        // Solve for A in q = [sin(A)*v, cos(A)]
+        Vector3 u(x, y, z);
+        double len = u.magnitude();
+
+        if (len == 0.0) {
+            return 
+        }
+        double A = atan2((double)w, len);
+        Vector3 v = u / len;
+        
+        return Quat(v * A, 0);
+    }
+    */
+
+    /** exp q = [sin(A) * v, cos(A)] where q = [Av, 0].
+        Only defined for pure-vector quaternions */
+    inline Quat exp() const {
+        debugAssertM(w == 0, "exp only defined for vector quaternions");
+        Vector3 u(x, y, z);
+        float A = u.magnitude();
+        Vector3 v = u / A;
+        return Quat(sinf(A) * v, cosf(A));
+    }
+
+
+    /**
+     Raise this quaternion to a power.  For a rotation, this is
+     the effect of rotating x times as much as the original
+     quaterion.
+
+     Note that q.pow(a).pow(b) == q.pow(a + b)
+     @cite Dam98 pg 21
+     */
+    inline Quat pow(float x) const {
+        return (log() * x).exp();
+    }
+
+
+    /**
+     @deprecated
+     Use toUnit()
+     */
+    inline Quat unitize() const {
+        float mag2 = dot(*this);
+        if (G3D::fuzzyEq(mag2, 1.0f)) {
+            return *this;
+        } else {
+            return *this / sqrtf(mag2);
+        }
+    }
+
+    /**
+     Returns a unit quaterion obtained by dividing through by
+     the magnitude.
+     */
+    inline Quat toUnit() const {
+        return unitize();
+    }
+
+    /**
+     The linear algebra 2-norm, sqrt(q dot q).  This matches
+     the value used in Dam's 1998 tech report but differs from the
+     n(q) value used in Eberly's 1999 paper, which is the square of the
+     norm.
+     */
+    inline float norm() const {
+        return magnitude();
+    }
+
+    // access quaternion as q[0] = q.x, q[1] = q.y, q[2] = q.z, q[3] = q.w
+    //
+    // WARNING.  These member functions rely on
+    // (1) Quat not having virtual functions
+    // (2) the data packed in a 4*sizeof(float) memory block
+    const float& operator[] (int i) const;
+    float& operator[] (int i);
+
+    /** Generate uniform random unit quaternion (i.e. random "direction") 
+	@cite From "Uniform Random Rotations", Ken Shoemake, Graphics Gems III.
+   */
+    static Quat unitRandom();
+
+    // 2-char swizzles
+
+    Vector2 xx() const;
+    Vector2 yx() const;
+    Vector2 zx() const;
+    Vector2 wx() const;
+    Vector2 xy() const;
+    Vector2 yy() const;
+    Vector2 zy() const;
+    Vector2 wy() const;
+    Vector2 xz() const;
+    Vector2 yz() const;
+    Vector2 zz() const;
+    Vector2 wz() const;
+    Vector2 xw() const;
+    Vector2 yw() const;
+    Vector2 zw() const;
+    Vector2 ww() const;
+
+    // 3-char swizzles
+
+    Vector3 xxx() const;
+    Vector3 yxx() const;
+    Vector3 zxx() const;
+    Vector3 wxx() const;
+    Vector3 xyx() const;
+    Vector3 yyx() const;
+    Vector3 zyx() const;
+    Vector3 wyx() const;
+    Vector3 xzx() const;
+    Vector3 yzx() const;
+    Vector3 zzx() const;
+    Vector3 wzx() const;
+    Vector3 xwx() const;
+    Vector3 ywx() const;
+    Vector3 zwx() const;
+    Vector3 wwx() const;
+    Vector3 xxy() const;
+    Vector3 yxy() const;
+    Vector3 zxy() const;
+    Vector3 wxy() const;
+    Vector3 xyy() const;
+    Vector3 yyy() const;
+    Vector3 zyy() const;
+    Vector3 wyy() const;
+    Vector3 xzy() const;
+    Vector3 yzy() const;
+    Vector3 zzy() const;
+    Vector3 wzy() const;
+    Vector3 xwy() const;
+    Vector3 ywy() const;
+    Vector3 zwy() const;
+    Vector3 wwy() const;
+    Vector3 xxz() const;
+    Vector3 yxz() const;
+    Vector3 zxz() const;
+    Vector3 wxz() const;
+    Vector3 xyz() const;
+    Vector3 yyz() const;
+    Vector3 zyz() const;
+    Vector3 wyz() const;
+    Vector3 xzz() const;
+    Vector3 yzz() const;
+    Vector3 zzz() const;
+    Vector3 wzz() const;
+    Vector3 xwz() const;
+    Vector3 ywz() const;
+    Vector3 zwz() const;
+    Vector3 wwz() const;
+    Vector3 xxw() const;
+    Vector3 yxw() const;
+    Vector3 zxw() const;
+    Vector3 wxw() const;
+    Vector3 xyw() const;
+    Vector3 yyw() const;
+    Vector3 zyw() const;
+    Vector3 wyw() const;
+    Vector3 xzw() const;
+    Vector3 yzw() const;
+    Vector3 zzw() const;
+    Vector3 wzw() const;
+    Vector3 xww() const;
+    Vector3 yww() const;
+    Vector3 zww() const;
+    Vector3 www() const;
+
+    // 4-char swizzles
+
+    Vector4 xxxx() const;
+    Vector4 yxxx() const;
+    Vector4 zxxx() const;
+    Vector4 wxxx() const;
+    Vector4 xyxx() const;
+    Vector4 yyxx() const;
+    Vector4 zyxx() const;
+    Vector4 wyxx() const;
+    Vector4 xzxx() const;
+    Vector4 yzxx() const;
+    Vector4 zzxx() const;
+    Vector4 wzxx() const;
+    Vector4 xwxx() const;
+    Vector4 ywxx() const;
+    Vector4 zwxx() const;
+    Vector4 wwxx() const;
+    Vector4 xxyx() const;
+    Vector4 yxyx() const;
+    Vector4 zxyx() const;
+    Vector4 wxyx() const;
+    Vector4 xyyx() const;
+    Vector4 yyyx() const;
+    Vector4 zyyx() const;
+    Vector4 wyyx() const;
+    Vector4 xzyx() const;
+    Vector4 yzyx() const;
+    Vector4 zzyx() const;
+    Vector4 wzyx() const;
+    Vector4 xwyx() const;
+    Vector4 ywyx() const;
+    Vector4 zwyx() const;
+    Vector4 wwyx() const;
+    Vector4 xxzx() const;
+    Vector4 yxzx() const;
+    Vector4 zxzx() const;
+    Vector4 wxzx() const;
+    Vector4 xyzx() const;
+    Vector4 yyzx() const;
+    Vector4 zyzx() const;
+    Vector4 wyzx() const;
+    Vector4 xzzx() const;
+    Vector4 yzzx() const;
+    Vector4 zzzx() const;
+    Vector4 wzzx() const;
+    Vector4 xwzx() const;
+    Vector4 ywzx() const;
+    Vector4 zwzx() const;
+    Vector4 wwzx() const;
+    Vector4 xxwx() const;
+    Vector4 yxwx() const;
+    Vector4 zxwx() const;
+    Vector4 wxwx() const;
+    Vector4 xywx() const;
+    Vector4 yywx() const;
+    Vector4 zywx() const;
+    Vector4 wywx() const;
+    Vector4 xzwx() const;
+    Vector4 yzwx() const;
+    Vector4 zzwx() const;
+    Vector4 wzwx() const;
+    Vector4 xwwx() const;
+    Vector4 ywwx() const;
+    Vector4 zwwx() const;
+    Vector4 wwwx() const;
+    Vector4 xxxy() const;
+    Vector4 yxxy() const;
+    Vector4 zxxy() const;
+    Vector4 wxxy() const;
+    Vector4 xyxy() const;
+    Vector4 yyxy() const;
+    Vector4 zyxy() const;
+    Vector4 wyxy() const;
+    Vector4 xzxy() const;
+    Vector4 yzxy() const;
+    Vector4 zzxy() const;
+    Vector4 wzxy() const;
+    Vector4 xwxy() const;
+    Vector4 ywxy() const;
+    Vector4 zwxy() const;
+    Vector4 wwxy() const;
+    Vector4 xxyy() const;
+    Vector4 yxyy() const;
+    Vector4 zxyy() const;
+    Vector4 wxyy() const;
+    Vector4 xyyy() const;
+    Vector4 yyyy() const;
+    Vector4 zyyy() const;
+    Vector4 wyyy() const;
+    Vector4 xzyy() const;
+    Vector4 yzyy() const;
+    Vector4 zzyy() const;
+    Vector4 wzyy() const;
+    Vector4 xwyy() const;
+    Vector4 ywyy() const;
+    Vector4 zwyy() const;
+    Vector4 wwyy() const;
+    Vector4 xxzy() const;
+    Vector4 yxzy() const;
+    Vector4 zxzy() const;
+    Vector4 wxzy() const;
+    Vector4 xyzy() const;
+    Vector4 yyzy() const;
+    Vector4 zyzy() const;
+    Vector4 wyzy() const;
+    Vector4 xzzy() const;
+    Vector4 yzzy() const;
+    Vector4 zzzy() const;
+    Vector4 wzzy() const;
+    Vector4 xwzy() const;
+    Vector4 ywzy() const;
+    Vector4 zwzy() const;
+    Vector4 wwzy() const;
+    Vector4 xxwy() const;
+    Vector4 yxwy() const;
+    Vector4 zxwy() const;
+    Vector4 wxwy() const;
+    Vector4 xywy() const;
+    Vector4 yywy() const;
+    Vector4 zywy() const;
+    Vector4 wywy() const;
+    Vector4 xzwy() const;
+    Vector4 yzwy() const;
+    Vector4 zzwy() const;
+    Vector4 wzwy() const;
+    Vector4 xwwy() const;
+    Vector4 ywwy() const;
+    Vector4 zwwy() const;
+    Vector4 wwwy() const;
+    Vector4 xxxz() const;
+    Vector4 yxxz() const;
+    Vector4 zxxz() const;
+    Vector4 wxxz() const;
+    Vector4 xyxz() const;
+    Vector4 yyxz() const;
+    Vector4 zyxz() const;
+    Vector4 wyxz() const;
+    Vector4 xzxz() const;
+    Vector4 yzxz() const;
+    Vector4 zzxz() const;
+    Vector4 wzxz() const;
+    Vector4 xwxz() const;
+    Vector4 ywxz() const;
+    Vector4 zwxz() const;
+    Vector4 wwxz() const;
+    Vector4 xxyz() const;
+    Vector4 yxyz() const;
+    Vector4 zxyz() const;
+    Vector4 wxyz() const;
+    Vector4 xyyz() const;
+    Vector4 yyyz() const;
+    Vector4 zyyz() const;
+    Vector4 wyyz() const;
+    Vector4 xzyz() const;
+    Vector4 yzyz() const;
+    Vector4 zzyz() const;
+    Vector4 wzyz() const;
+    Vector4 xwyz() const;
+    Vector4 ywyz() const;
+    Vector4 zwyz() const;
+    Vector4 wwyz() const;
+    Vector4 xxzz() const;
+    Vector4 yxzz() const;
+    Vector4 zxzz() const;
+    Vector4 wxzz() const;
+    Vector4 xyzz() const;
+    Vector4 yyzz() const;
+    Vector4 zyzz() const;
+    Vector4 wyzz() const;
+    Vector4 xzzz() const;
+    Vector4 yzzz() const;
+    Vector4 zzzz() const;
+    Vector4 wzzz() const;
+    Vector4 xwzz() const;
+    Vector4 ywzz() const;
+    Vector4 zwzz() const;
+    Vector4 wwzz() const;
+    Vector4 xxwz() const;
+    Vector4 yxwz() const;
+    Vector4 zxwz() const;
+    Vector4 wxwz() const;
+    Vector4 xywz() const;
+    Vector4 yywz() const;
+    Vector4 zywz() const;
+    Vector4 wywz() const;
+    Vector4 xzwz() const;
+    Vector4 yzwz() const;
+    Vector4 zzwz() const;
+    Vector4 wzwz() const;
+    Vector4 xwwz() const;
+    Vector4 ywwz() const;
+    Vector4 zwwz() const;
+    Vector4 wwwz() const;
+    Vector4 xxxw() const;
+    Vector4 yxxw() const;
+    Vector4 zxxw() const;
+    Vector4 wxxw() const;
+    Vector4 xyxw() const;
+    Vector4 yyxw() const;
+    Vector4 zyxw() const;
+    Vector4 wyxw() const;
+    Vector4 xzxw() const;
+    Vector4 yzxw() const;
+    Vector4 zzxw() const;
+    Vector4 wzxw() const;
+    Vector4 xwxw() const;
+    Vector4 ywxw() const;
+    Vector4 zwxw() const;
+    Vector4 wwxw() const;
+    Vector4 xxyw() const;
+    Vector4 yxyw() const;
+    Vector4 zxyw() const;
+    Vector4 wxyw() const;
+    Vector4 xyyw() const;
+    Vector4 yyyw() const;
+    Vector4 zyyw() const;
+    Vector4 wyyw() const;
+    Vector4 xzyw() const;
+    Vector4 yzyw() const;
+    Vector4 zzyw() const;
+    Vector4 wzyw() const;
+    Vector4 xwyw() const;
+    Vector4 ywyw() const;
+    Vector4 zwyw() const;
+    Vector4 wwyw() const;
+    Vector4 xxzw() const;
+    Vector4 yxzw() const;
+    Vector4 zxzw() const;
+    Vector4 wxzw() const;
+    Vector4 xyzw() const;
+    Vector4 yyzw() const;
+    Vector4 zyzw() const;
+    Vector4 wyzw() const;
+    Vector4 xzzw() const;
+    Vector4 yzzw() const;
+    Vector4 zzzw() const;
+    Vector4 wzzw() const;
+    Vector4 xwzw() const;
+    Vector4 ywzw() const;
+    Vector4 zwzw() const;
+    Vector4 wwzw() const;
+    Vector4 xxww() const;
+    Vector4 yxww() const;
+    Vector4 zxww() const;
+    Vector4 wxww() const;
+    Vector4 xyww() const;
+    Vector4 yyww() const;
+    Vector4 zyww() const;
+    Vector4 wyww() const;
+    Vector4 xzww() const;
+    Vector4 yzww() const;
+    Vector4 zzww() const;
+    Vector4 wzww() const;
+    Vector4 xwww() const;
+    Vector4 ywww() const;
+    Vector4 zwww() const;
+    Vector4 wwww() const;
+};
+
+inline Quat exp(const Quat& q) {
+    return q.exp();
+}
+
+inline Quat log(const Quat& q) {
+    return q.log();
+}
+
+inline G3D::Quat operator*(double s, const G3D::Quat& q) {
+    return q * (float)s;
+}
+
+inline G3D::Quat operator*(float s, const G3D::Quat& q) {
+    return q * s;
+}
+
+} // Namespace G3D
+
+// Outside the namespace to avoid overloading confusion for C++
+inline G3D::Quat pow(const G3D::Quat& q, double x) {
+    return q.pow((float)x);
+}
+
+
+
+#include "Quat.inl"
+
+#endif
diff --git a/dep/include/g3dlite/G3D/Quat.inl b/dep/include/g3dlite/G3D/Quat.inl
new file mode 100644
index 00000000000..705cb0455d4
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Quat.inl
@@ -0,0 +1,38 @@
+/**
+  Quat.inl
+ 
+  @cite Quaternion implementation based on Watt & Watt page 363.  
+  Thanks to Max McGuire for slerp optimizations.
+  
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+  
+  @created 2002-01-23
+  @edited  2004-03-04
+ */
+
+namespace G3D {
+
+inline float& Quat::operator[] (int i) {
+    debugAssert(i >= 0);
+    debugAssert(i < 4);
+    return ((float*)this)[i];
+}
+
+inline const float& Quat::operator[] (int i) const {
+    debugAssert(i >= 0);
+    debugAssert(i < 4);
+    return ((float*)this)[i];
+}
+
+
+
+inline Quat Quat::operator-(const Quat& other) const {
+    return Quat(x - other.x, y - other.y, z - other.z, w - other.w);
+}
+
+inline Quat Quat::operator+(const Quat& other) const {
+    return Quat(x + other.x, y + other.y, z + other.z, w + other.w);
+}
+
+}
+
diff --git a/dep/include/g3dlite/G3D/Ray.h b/dep/include/g3dlite/G3D/Ray.h
new file mode 100644
index 00000000000..3b54f145c33
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Ray.h
@@ -0,0 +1,327 @@
+/**
+ @file Ray.h
+ 
+ Ray class
+ 
+ @maintainer Morgan McGuire, matrix@graphics3d.com
+ 
+ @created 2002-07-12
+ @edited  2006-02-21
+ */
+
+#ifndef G3D_RAY_H
+#define G3D_RAY_H
+
+#include "G3D/platform.h"
+#include "G3D/Vector3.h"
+#include "G3D/Triangle.h"
+
+namespace G3D {
+
+/**
+ A 3D Ray.
+ */
+class Ray {
+private:
+    Ray(const Vector3& origin, const Vector3& direction) {
+        this->origin    = origin;
+        this->direction = direction;
+    }
+
+public:
+    Vector3         origin;
+
+    /**
+     Not unit length
+     */
+    Vector3         direction;
+
+    Ray() : origin(Vector3::zero()), direction(Vector3::zero()) {}
+
+    virtual ~Ray() {}
+
+    /**
+     Creates a Ray from a origin and a (nonzero) direction.
+     */
+    static Ray fromOriginAndDirection(const Vector3& point, const Vector3& direction) {
+        return Ray(point, direction);
+    }
+
+    Ray unit() const {
+        return Ray(origin, direction.unit());
+    }
+
+    /**
+     Returns the closest point on the Ray to point.
+     */
+    Vector3 closestPoint(const Vector3& point) const {
+        float t = direction.dot(point - this->origin);
+        if (t < 0) {
+            return this->origin;
+        } else {
+            return this->origin + direction * t;
+        }
+    }
+
+    /**
+     Returns the closest distance between point and the Ray
+     */
+    float distance(const Vector3& point) const {
+        return (closestPoint(point) - point).magnitude();
+    }
+
+    /**
+     Returns the point where the Ray and plane intersect.  If there
+     is no intersection, returns a point at infinity.
+
+      Planes are considered one-sided, so the ray will not intersect
+      a plane where the normal faces in the traveling direction.
+    */
+    Vector3 intersection(const class Plane& plane) const;
+
+    /**
+     Returns the distance until intersection with the (solid) sphere.
+     Will be 0 if inside the sphere, inf if there is no intersection.
+
+     The ray direction is <B>not</B> normalized.  If the ray direction
+     has unit length, the distance from the origin to intersection
+     is equal to the time.  If the direction does not have unit length,
+     the distance = time * direction.length().
+
+     See also the G3D::CollisionDetection "movingPoint" methods,
+     which give more information about the intersection.
+     */
+    float intersectionTime(const class Sphere& sphere) const;
+
+    float intersectionTime(const class Plane& plane) const;
+
+    float intersectionTime(const class Box& box) const;
+
+    float intersectionTime(const class AABox& box) const;
+
+    /**
+     The three extra arguments are the weights of vertices 0, 1, and 2
+     at the intersection point; they are useful for texture mapping
+     and interpolated normals.
+     */
+    float intersectionTime(
+        const Vector3& v0, const Vector3& v1, const Vector3& v2,
+        const Vector3& edge01, const Vector3& edge02,
+        double& w0, double& w1, double& w2) const;
+
+     /**
+     Ray-triangle intersection for a 1-sided triangle.  Fastest version.
+       @cite http://www.acm.org/jgt/papers/MollerTrumbore97/
+       http://www.graphics.cornell.edu/pubs/1997/MT97.html
+     */
+    inline float intersectionTime(
+        const Vector3& vert0,
+        const Vector3& vert1,
+        const Vector3& vert2,
+        const Vector3& edge01,
+        const Vector3& edge02) const;
+
+
+    inline float intersectionTime(
+        const Vector3& vert0,
+        const Vector3& vert1,
+        const Vector3& vert2) const {
+
+        return intersectionTime(vert0, vert1, vert2, vert1 - vert0, vert2 - vert0);
+    }
+
+
+    inline float intersectionTime(
+        const Vector3&  vert0,
+        const Vector3&  vert1,
+        const Vector3&  vert2,
+        double&         w0,
+        double&         w1,
+        double&         w2) const {
+
+        return intersectionTime(vert0, vert1, vert2, vert1 - vert0, vert2 - vert0, w0, w1, w2);
+    }
+
+    /* One-sided triangle 
+       */
+    inline float intersectionTime(const Triangle& triangle) const {
+        return intersectionTime(
+            triangle.vertex(0), triangle.vertex(1), triangle.vertex(2),
+            triangle.edge01, triangle.edge02);
+    }
+
+    inline float intersectionTime(
+        const Triangle& triangle,
+        double&         w0,
+        double&         w1,
+        double&         w2) const {
+        return intersectionTime(triangle.vertex(0), triangle.vertex(1), triangle.vertex(2),
+            triangle.edge01, triangle.edge02, w0, w1, w2);
+    }
+
+    /** Refracts about the normal
+        using G3D::Vector3::refractionDirection
+        and bumps the ray slightly from the newOrigin. */
+    Ray refract(
+        const Vector3&  newOrigin,
+        const Vector3&  normal,
+        float           iInside,
+        float           iOutside) const;
+
+    /** Reflects about the normal
+        using G3D::Vector3::reflectionDirection
+        and bumps the ray slightly from
+        the newOrigin. */
+    Ray reflect(
+        const Vector3&  newOrigin,
+        const Vector3&  normal) const;
+};
+
+
+#define EPSILON 0.000001
+#define CROSS(dest,v1,v2) \
+          dest[0]=v1[1]*v2[2]-v1[2]*v2[1]; \
+          dest[1]=v1[2]*v2[0]-v1[0]*v2[2]; \
+          dest[2]=v1[0]*v2[1]-v1[1]*v2[0];
+
+#define DOT(v1,v2) (v1[0]*v2[0]+v1[1]*v2[1]+v1[2]*v2[2])
+
+#define SUB(dest,v1,v2) \
+          dest[0]=v1[0]-v2[0]; \
+          dest[1]=v1[1]-v2[1]; \
+          dest[2]=v1[2]-v2[2]; 
+
+inline float Ray::intersectionTime(
+    const Vector3& vert0,
+    const Vector3& vert1,
+    const Vector3& vert2,
+    const Vector3& edge1,
+    const Vector3& edge2) const {
+
+    (void)vert1;
+    (void)vert2;
+
+    // Barycenteric coords
+    float u, v;
+
+    float tvec[3], pvec[3], qvec[3];
+    
+    // begin calculating determinant - also used to calculate U parameter
+    CROSS(pvec, direction, edge2);
+    
+    // if determinant is near zero, ray lies in plane of triangle
+    const float det = DOT(edge1, pvec);
+    
+    if (det < EPSILON) {
+        return (float)inf();
+    }
+    
+    // calculate distance from vert0 to ray origin
+    SUB(tvec, origin, vert0);
+    
+    // calculate U parameter and test bounds
+    u = DOT(tvec, pvec);
+    if ((u < 0.0f) || (u > det)) {
+        // Hit the plane outside the triangle
+        return (float)inf();
+    }
+    
+    // prepare to test V parameter
+    CROSS(qvec, tvec, edge1);
+    
+    // calculate V parameter and test bounds
+    v = DOT(direction, qvec);
+    if ((v < 0.0f) || (u + v > det)) {
+        // Hit the plane outside the triangle
+        return (float)inf();
+    }
+    
+
+    // Case where we don't need correct (u, v):
+    const float t = DOT(edge2, qvec);
+    
+    if (t >= 0.0f) {
+        // Note that det must be positive
+        return t / det;
+    } else {
+        // We had to travel backwards in time to intersect
+        return (float)inf();
+    }
+}
+
+
+inline float Ray::intersectionTime(
+    const Vector3&  vert0,
+    const Vector3&  vert1,
+    const Vector3&  vert2,
+    const Vector3&  edge1,
+    const Vector3&  edge2,
+    double&         w0,
+    double&         w1,
+    double&         w2) const {
+
+    (void)vert1;
+    (void)vert2;
+
+    // Barycenteric coords
+    float u, v;
+
+    float tvec[3], pvec[3], qvec[3];
+
+    // begin calculating determinant - also used to calculate U parameter
+    CROSS(pvec, direction, edge2);
+    
+    // if determinant is near zero, ray lies in plane of triangle
+    const float det = DOT(edge1, pvec);
+    
+    if (det < EPSILON) {
+        return (float)inf();
+    }
+    
+    // calculate distance from vert0 to ray origin
+    SUB(tvec, origin, vert0);
+    
+    // calculate U parameter and test bounds
+    u = DOT(tvec, pvec);
+    if ((u < 0.0f) || (u > det)) {
+        // Hit the plane outside the triangle
+        return (float)inf();
+    }
+    
+    // prepare to test V parameter
+    CROSS(qvec, tvec, edge1);
+    
+    // calculate V parameter and test bounds
+    v = DOT(direction, qvec);
+    if ((v < 0.0f) || (u + v > det)) {
+        // Hit the plane outside the triangle
+        return (float)inf();
+    }
+    
+    float t = DOT(edge2, qvec);
+    
+    if (t >= 0) {
+        const float inv_det = 1.0f / det;
+        t *= inv_det;
+        u *= inv_det;
+        v *= inv_det;
+
+        w0 = (1.0f - u - v);
+        w1 = u;
+        w2 = v;
+
+        return t;
+    } else {
+        // We had to travel backwards in time to intersect
+        return (float)inf();
+    }
+}
+
+#undef EPSILON
+#undef CROSS
+#undef DOT
+#undef SUB
+
+}// namespace
+
+#endif
diff --git a/dep/include/g3dlite/G3D/RegistryUtil.h b/dep/include/g3dlite/G3D/RegistryUtil.h
new file mode 100644
index 00000000000..97de8d06725
--- /dev/null
+++ b/dep/include/g3dlite/G3D/RegistryUtil.h
@@ -0,0 +1,86 @@
+/**
+ @file RegistryUtil.h
+
+ @created 2006-04-06
+ @edited  2006-04-06
+
+ Copyright 2000-2006, Morgan McGuire.
+ All rights reserved.
+*/
+
+#ifndef G3D_REGISTRYUTIL_H
+#define G3D_REGISTRYUTIL_H
+
+#include "G3D/platform.h"
+#include "G3D/g3dmath.h"
+
+// This file is only used on Windows
+#ifdef G3D_WIN32
+
+#include <string>
+
+namespace G3D {
+
+/** 
+    Provides generalized Windows registry querying.
+
+    All key names are one string in the format:
+        "[base key]\[sub-keys]\value"
+
+    [base key] can be any of the following:
+        HKEY_CLASSES_ROOT
+        HKEY_CURRENT_CONFIG
+        HKEY_CURRENT_USER
+        HKEY_LOCAL_MACHINE
+        HKEY_PERFORMANCE_DATA
+        HKEY_PERFORMANCE_NLSTEXT
+        HKEY_PERFORMANCE_TEXT
+        HKEY_USERS
+
+    keyExists() should be used to validate a key before reading or writing
+    to ensure that a debug assert or false return is for a different error.
+*/
+class RegistryUtil {
+
+public:
+    /** returns true if the key exists */
+    static bool keyExists(const std::string& key);
+
+    /** returns false if the key could not be read for any reason. */
+    static bool readInt32(const std::string& key, int32& valueData);
+
+    /** 
+      Reads an arbitrary amount of data from a binary registry key.
+      returns false if the key could not be read for any reason.
+    
+      @beta
+      @param valueData pointer to the output buffer of sufficient size. Pass NULL as valueData in order to have available data size returned in dataSize.
+      @param dataSize size of the output buffer.  When NULL is passed for valueData, contains the size of available data on successful return.
+    */
+    static bool readBytes(const std::string& key, uint8* valueData, uint32& dataSize);
+
+    /** returns false if the key could not be read for any reason. */
+    static bool readString(const std::string& key, std::string& valueData);
+
+    /** returns false if the key could not be written for any reason. */
+    static bool writeInt32(const std::string& key, int32 valueData);
+
+    /** 
+      Writes an arbitrary amount of data to a binary registry key.
+      returns false if the key could not be written for any reason.
+    
+      @param valueData pointer to the input buffer
+      @param dataSize size of the input buffer that should be written
+    */
+    static bool writeBytes(const std::string& key, const uint8* valueData, uint32 dataSize);
+
+    /** returns false if the key could not be written for any reason. */
+    static bool writeString(const std::string& key, const std::string& valueData);
+
+};
+
+} // namespace G3D
+
+#endif // G3D_WIN32
+
+#endif // G3D_REGISTRYTUIL_H
\ No newline at end of file
diff --git a/dep/include/g3dlite/G3D/Sphere.h b/dep/include/g3dlite/G3D/Sphere.h
new file mode 100644
index 00000000000..39097e01850
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Sphere.h
@@ -0,0 +1,128 @@
+/**
+ @file Sphere.h
+ 
+ Sphere class
+ 
+ @maintainer Morgan McGuire, matrix@graphics3d.com
+ 
+ @created 2001-06-02
+ @edited  2004-07-05
+ */
+
+#ifndef G3D_SPHERE_H
+#define G3D_SPHERE_H
+
+#include "G3D/platform.h"
+#include "G3D/Vector3.h"
+#include "G3D/Array.h"
+#include "G3D/Sphere.h"
+
+namespace G3D {
+
+/**
+ Sphere.
+ */
+class Sphere {
+private:
+
+    static int32     dummy;
+
+public:
+    Vector3          center;
+    float            radius;
+
+    Sphere() {
+        center = Vector3::zero();
+        radius = 0;
+    }
+
+    Sphere(
+        const Vector3&  center,
+        float           radius) {
+
+        this->center = center;
+        this->radius = radius;
+    }
+
+    virtual ~Sphere() {}
+
+    bool operator==(const Sphere& other) const {
+        return (center == other.center) && (radius == other.radius);
+    }
+
+    bool operator!=(const Sphere& other) const {
+        return !((center == other.center) && (radius == other.radius));
+    }
+
+    /**
+     Returns true if point is less than or equal to radius away from
+     the center.
+     */
+    bool contains(const Vector3& point) const;
+
+/**
+	 @deprecated Use culledBy(Array<Plane>&)
+     */
+    bool culledBy(
+        const class Plane*  plane,
+        int                 numPlanes,
+		int32&				cullingPlaneIndex,
+		const uint32  		testMask,
+        uint32&             childMask) const;
+
+    /**
+	 @deprecated Use culledBy(Array<Plane>&)
+     */
+    bool culledBy(
+        const class Plane*  plane,
+        int                 numPlanes,
+		int32&				cullingPlaneIndex = dummy,
+		const uint32  		testMask = -1) const;
+
+	/**
+      See AABox::culledBy
+	 */
+	bool culledBy(
+		const Array<Plane>&		plane,
+		int32&					cullingPlaneIndex,
+		const uint32  			testMask,
+        uint32&                 childMask) const;
+
+    /**
+     Conservative culling test that does not produce a mask for children.
+     */
+	bool culledBy(
+		const Array<Plane>&		plane,
+		int32&					cullingPlaneIndex = dummy,
+		const uint32  			testMask		  = -1) const;
+    virtual std::string toString() const;
+
+    float volume() const;
+
+    /** @deprecated */
+    float surfaceArea() const;
+
+    inline float area() const {
+        return surfaceArea();
+    }
+
+    /**
+     Uniformly distributed on the surface.
+     */
+    Vector3 randomSurfacePoint() const;
+
+    /**
+     Uniformly distributed on the interior (includes surface)
+     */
+    Vector3 randomInteriorPoint() const;
+
+    void getBounds(class AABox& out) const;
+};
+
+} // namespace
+
+inline unsigned int hashCode(const G3D::Sphere& sphere) {
+    return (unsigned int)(hashCode(sphere.center) + (sphere.radius * 13));
+}
+
+#endif
diff --git a/dep/include/g3dlite/G3D/System.h b/dep/include/g3dlite/G3D/System.h
new file mode 100644
index 00000000000..507b44670da
--- /dev/null
+++ b/dep/include/g3dlite/G3D/System.h
@@ -0,0 +1,122 @@
+/** 
+  @file System.h
+ 
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+ 
+  @cite Rob Wyatt http://www.gamasutra.com/features/wyatts_world/19990709/processor_detection_01.htm
+  @cite Benjamin Jurke http://www.flipcode.com/cgi-bin/msg.cgi?showThread=COTD-ProcessorDetectionClass&forum=cotd&id=-1
+  @cite Michael Herf http://www.stereopsis.com/memcpy.html
+
+  @created 2003-01-25
+  @edited  2006-04-26
+ */
+
+#ifndef G3D_SYSTEM_H
+#define G3D_SYSTEM_H
+
+#include "G3D/platform.h"
+#include "G3D/g3dmath.h"
+#include <string>
+
+#ifdef G3D_OSX
+#	include <CoreServices/CoreServices.h>
+#endif
+
+
+namespace G3D {
+
+typedef double RealTime;
+
+class System {
+public:
+
+    /** Called automatically by the other System routines.*/
+    static void init();
+
+    /**
+     Guarantees that the start of the array is aligned to the 
+     specified number of bytes.
+     */
+    static void* alignedMalloc(size_t bytes, size_t alignment);
+
+    /**
+     Uses pooled storage to optimize small allocations (1 byte to 5 kilobytes).  
+     Can be 10x to 100x faster than calling ::malloc or new.
+
+     The result must be freed with free.
+
+     Threadsafe on Win32.
+
+     @sa calloc realloc OutOfMemoryCallback free
+     */
+    static void* malloc(size_t bytes);
+
+    static void* calloc(size_t n, size_t x);
+
+    /**
+     @param size Size of memory that the system was trying to allocate
+     @param recoverable If true, the system will attempt to allocate again
+            if the callback returns true.  If false, malloc is going to return 
+            NULL and this invocation is just to notify the application.
+     @return Return true to force malloc to attempt allocation again if the
+            error was recoverable.
+     */
+    typedef bool (*OutOfMemoryCallback)(size_t size, bool recoverable);
+
+    /**
+     When System::malloc fails to allocate memory because the system is
+     out of memory, it invokes this handler (if it is not NULL).
+     The argument to the callback is the amount of memory that malloc
+     was trying to allocate when it ran out.  If the callback returns
+     true, System::malloc will attempt to allocate the memory again.
+     If the callback returns false, then System::malloc will return NULL.
+
+     You can use outOfMemoryCallback to free data structures or to 
+     register the failure.
+     */
+    static OutOfMemoryCallback outOfMemoryCallback;
+
+    /**
+     Version of realloc that works with System::malloc.
+     */
+    static void* realloc(void* block, size_t bytes);
+
+    /** Returns a string describing how well System::malloc is using its internal pooled storage.
+        "heap" memory was slow to allocate; the other data sizes are comparatively fast.*/
+    static std::string mallocPerformance();
+    static void resetMallocPerformanceCounters();
+
+    /** 
+       Returns a string describing the current usage of the buffer pools used for
+       optimizing System::malloc.
+     */
+    static std::string mallocStatus();
+
+    /**
+     Free data allocated with System::malloc.
+
+     Threadsafe on Win32.
+     */
+    static void free(void* p);
+
+    /**
+     Frees memory allocated with alignedMalloc.
+     */
+    static void alignedFree(void* ptr);
+
+    /** An implementation of memcpy that may be up to 2x as fast as the C library
+	one on some processors.  Guaranteed to have the same behavior as memcpy
+	in all cases. */
+    static void memcpy(void* dst, const void* src, size_t numBytes);
+
+    /** An implementation of memset that may be up to 2x as fast as the C library
+	one on some processors.  Guaranteed to have the same behavior as memset
+	in all cases. */
+    static void memset(void* dst, uint8 value, size_t numBytes);
+
+};
+
+
+} // namespace
+
+#endif
diff --git a/dep/include/g3dlite/G3D/Table.h b/dep/include/g3dlite/G3D/Table.h
new file mode 100644
index 00000000000..285d774d24a
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Table.h
@@ -0,0 +1,695 @@
+/**
+  @file Table.h
+
+  Templated hash table class.
+
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+  @created 2001-04-22
+  @edited  2006-10-14
+  Copyright 2000-2006, Morgan McGuire.
+  All rights reserved.
+ */
+
+#ifndef G3D_TABLE_H
+#define G3D_TABLE_H
+
+#include "G3D/platform.h"
+#include "G3D/Array.h"
+#include "G3D/debug.h"
+#include "G3D/System.h"
+#include "G3D/g3dmath.h"
+#include "G3D/Crypto.h"
+#include <cstddef>
+#include <string>
+
+#ifdef G3D_WIN32
+#   pragma warning (push)
+    // Debug name too long warning
+#   pragma warning (disable : 4786)
+#endif
+
+
+template<typename Key>
+struct GHashCode{};
+
+template <>
+struct GHashCode<int>
+{
+    size_t operator()(int key) const { return static_cast<size_t>(key); }
+};
+
+template <>
+struct GHashCode<G3D::uint32>
+{
+    size_t operator()(G3D::uint32 key) const { return static_cast<size_t>(key); }
+};
+
+template <>
+struct GHashCode<G3D::uint64>
+{
+    size_t operator()(G3D::uint64 key) const { return static_cast<size_t>(key); }
+};
+
+template <>
+struct GHashCode<void*>
+{
+    size_t operator()(const void* key) const { return reinterpret_cast<size_t>(key); }
+};
+
+template<class T>
+struct GHashCode<T*>
+{
+    size_t operator()(const T* key) const { return reinterpret_cast<size_t>(key); }
+};
+
+template <>
+struct GHashCode<const std::string>
+{
+    size_t operator()(const std::string& key) const { return static_cast<size_t>(G3D::Crypto::crc32(key.c_str(), key.size())); }
+};
+
+template <>
+struct GHashCode<std::string>
+{
+    size_t operator()(const std::string& key) const { return static_cast<size_t>(G3D::Crypto::crc32(key.c_str(), key.size())); }
+};
+
+namespace G3D {
+
+
+/**
+ An unordered data structure mapping keys to values.
+
+ Key must be a pointer, an int, a std::string, a class with a hashCode() method,
+ or provide overloads for: 
+
+  <PRE>
+    template<> struct GHashCode<class Key> {
+        size_t operator()(Key key) const { return reinterpret_cast<size_t>( ... ); }
+    };
+
+   bool operator==(const Key&, const Key&);
+  </PRE>
+
+ G3D pre-defines GHashCode functions for common types (like <CODE>int</CODE> and <CODE>std::string</CODE>).
+ If you use a Table with a different type you must write those functions yourself.  For example,
+ an enum would use:
+
+  <PRE>
+    template<> struct GHashCode<MyEnum> {
+        size_t operator()(MyEnum key) const { return reinterpret_cast<size_t>( key ); }
+    };
+  </PRE>
+
+  And rely on the default enum operator==.
+
+
+  Periodically check that debugGetLoad() is low (> 0.1).  When it gets near
+  1.0 your hash function is badly designed and maps too many inputs to
+  the same output.
+ */
+template<class Key, class Value, class HashFunc = GHashCode<Key> > 
+class Table {
+public:
+
+    /**
+     The pairs returned by iterator.
+     */
+    class Entry {
+    public:
+        Key    key;
+        Value  value;
+    };
+
+private:
+    /**
+     Linked list nodes used internally by HashTable.
+     */
+    class Node {
+    public:
+        size_t      hashCode;
+        Entry       entry;
+        Node*       next;
+
+        
+        /** Provide pooled allocation for speed. */
+        inline void* operator new (size_t size) {
+            return System::malloc(size);
+        }
+        
+        inline void operator delete (void* p) {
+            System::free(p);
+        }
+
+
+        Node(Key key, Value value, size_t hashCode, Node* next) {
+            this->entry.key   = key;
+            this->entry.value = value;
+            this->hashCode    = hashCode;
+            this->next        = next;
+        }
+
+        /**
+        Clones a whole chain;
+        */
+        Node* clone() {
+           return new Node(this->entry.key, this->entry.value, hashCode, (next == NULL) ? NULL : next->clone());
+        }
+    };
+
+    HashFunc m_HashFunc;
+
+    /**
+     Number of elements in the table.
+     */
+    size_t  _size;
+
+    /**
+     Array of Node*. 
+     We don't use Array<Node*> because Table is lower level.
+     Some elements may be NULL.
+     */
+    Node**  bucket;
+    
+    /**
+     Length of the bucket array.
+     */
+    size_t  numBuckets;
+
+    /**
+     Re-hashes for a larger bucket size.
+     */
+    void resize(size_t numBuckets) {
+
+        Node** oldBucket = bucket;
+        bucket = (Node**)System::alignedMalloc(sizeof(Node*) * numBuckets, 16);
+        System::memset(bucket, 0, sizeof(Node*) * numBuckets);
+
+        for (size_t b = 0; b < this->numBuckets; b++) {
+            Node* node = oldBucket[b];
+         
+            while (node != NULL) {
+                Node* nextNode = node->next;
+        
+                // insert at the head of the list for bucket[i]
+                size_t i = node->hashCode % numBuckets;
+                node->next = bucket[i];
+                bucket[i] = node;
+        
+                node = nextNode;
+            }
+        }
+
+        System::alignedFree(oldBucket);
+        this->numBuckets = numBuckets;
+    }
+
+    void copyFrom(const Table<Key, Value>& h) {
+        this->_size = h._size;
+        this->numBuckets = h.numBuckets;
+        this->bucket = (Node**)System::alignedMalloc(sizeof(Node*) * numBuckets, 16);
+        System::memset(this->bucket, 0, sizeof(Node*) * numBuckets);
+        for (size_t b = 0; b < this->numBuckets; b++) {
+            if (h.bucket[b] != NULL) {
+                bucket[b] = h.bucket[b]->clone();
+            }
+        }
+    }
+
+    /**
+     Frees the heap structures for the nodes.
+     */
+    void freeMemory() {
+        for (size_t b = 0; b < numBuckets; b++) {
+           Node* node = bucket[b];
+           while (node != NULL) {
+                Node* next = node->next;
+                delete node;
+                node = next;
+           }
+        }
+        System::alignedFree(bucket);
+        bucket     = NULL;
+        numBuckets = 0;
+        _size     = 0;
+    }
+
+public:
+
+    /**
+     Creates an empty hash table.  This causes some heap allocation to occur.
+     */
+    Table() {
+        numBuckets = 10;
+        _size      = 0;
+        bucket     = (Node**)System::alignedMalloc(sizeof(Node*) * numBuckets, 16);
+        System::memset(bucket, 0, sizeof(Node*) * numBuckets);
+    }
+
+    /**
+    Destroys all of the memory allocated by the table, but does <B>not</B>
+    call delete on keys or values if they are pointers.  If you want to
+    deallocate things that the table points at, use getKeys() and Array::deleteAll()
+    to delete them.
+    */
+    virtual ~Table() {
+        freeMemory();
+    }
+
+    Table(const Table<Key, Value>& h) {
+        this->copyFrom(h);
+    }
+
+    Table& operator=(const Table<Key, Value>& h) {
+        // No need to copy if the argument is this
+        if (this != &h) {
+            // Free the existing nodes
+            freeMemory();
+            this->copyFrom(h);
+        }
+        return *this;
+    }
+
+    /**
+     Returns the length of the deepest bucket.
+     */
+    size_t debugGetDeepestBucketSize() const {
+        size_t deepest = 0;
+
+        for (size_t b = 0; b < numBuckets; b++) {
+            size_t  count = 0;
+            Node*   node = bucket[b];
+            while (node != NULL) {
+                node = node->next;
+                ++count;
+            }
+
+            if (count > deepest) {
+                deepest = count;
+            }
+        }
+
+        return deepest;
+    }
+
+    /**
+     A small load (close to zero) means the hash table is acting very
+     efficiently most of the time.  A large load (close to 1) means 
+     the hash table is acting poorly-- all operations will be very slow.
+     A large load will result from a bad hash function that maps too
+     many keys to the same code.
+     */
+    double debugGetLoad() const {
+        return debugGetDeepestBucketSize() / (double)size();
+    }
+
+    /**
+     Returns the number of buckets.
+     */
+    size_t debugGetNumBuckets() const {
+        return numBuckets;
+    }
+
+    /**
+     C++ STL style iterator variable.  See begin().
+     */
+    class Iterator {
+    private:
+        friend class Table<Key, Value>;
+
+        /**
+         Bucket index.
+         */
+        size_t              index;
+
+        /**
+         Linked list node.
+         */
+        Node*               node;
+        Table<Key, Value>*  table;
+        size_t              numBuckets;
+        Node**              bucket;
+        bool                isDone;
+
+        /**
+         Creates the end iterator.
+         */
+        Iterator(const Table<Key, Value>* table) : table(const_cast<Table<Key, Value>*>(table)) {
+            isDone = true;
+        }
+
+        Iterator(const Table<Key, Value>* table, size_t numBuckets, Node** bucket) :
+            table(const_cast<Table<Key, Value>*>(table)),
+            numBuckets(numBuckets),
+            bucket(bucket) {
+            
+            if (numBuckets == 0) {
+                // Empty table
+                isDone = true;
+                return;
+            }
+
+            index = 0;
+            node = bucket[index];
+            isDone = false;
+            findNext();
+        }
+
+        /**
+         Finds the next element, setting isDone if one can't be found.
+         Looks at the current element first.
+         */
+        void findNext() {
+            while (node == NULL) {
+                index++;
+                if (index >= numBuckets) {
+                    isDone = true;
+                    break;
+                } else {
+                    node = bucket[index];
+                }
+            }
+        }
+
+    public:
+        inline bool operator!=(const Iterator& other) const {
+            return !(*this == other);
+        }
+
+        bool operator==(const Iterator& other) const {
+            if (other.isDone || isDone) {
+                // Common case; check against isDone.
+                return (isDone == other.isDone) && (other.table == table);
+            } else {
+                return
+                    (table == other.table) &&
+                    (node == other.node) && 
+                    (index == other.index);
+            }
+        }
+
+        /**
+         Pre increment.
+         */
+        Iterator& operator++() {
+            node = node->next;
+            findNext();
+            return *this;
+        }
+
+        /**
+         Post increment (slower than preincrement).
+         */
+        Iterator operator++(int) {
+            Iterator old = *this;
+            ++(*this);
+            return old;
+        }
+
+        const Entry& operator*() const {
+            return node->entry;
+        }
+
+        Entry* operator->() const {
+            return &(node->entry);
+        }
+
+        operator Entry*() const {
+            return &(node->entry);
+        }
+    };
+
+
+    /**
+     C++ STL style iterator method.  Returns the first Entry, which 
+     contains a key and value.  Use preincrement (++entry) to get to
+     the next element.  Do not modify the table while iterating.
+     */
+    Iterator begin() const {
+        return Iterator(this, numBuckets, bucket);
+    }
+
+    /**
+     C++ STL style iterator method.  Returns one after the last iterator
+     element.
+     */
+    const Iterator end() const {
+        return Iterator(this);
+    }
+
+    /**
+     Removes all elements
+     */
+    void clear() {
+         freeMemory();
+         numBuckets = 20;
+         _size = 0;
+         bucket = (Node**)System::alignedMalloc(sizeof(Node*) * numBuckets, 16);
+         System::memset(bucket, 0, sizeof(Node*) * numBuckets);
+    }
+
+   
+    /**
+     Returns the number of keys.
+     */
+    size_t size() const {
+        return _size;
+    }
+
+
+    /**
+     If you insert a pointer into the key or value of a table, you are
+     responsible for deallocating the object eventually.  Inserting 
+     key into a table is O(1), but may cause a potentially slow rehashing.
+     */
+    void set(const Key& key, const Value& value) {
+        size_t code = m_HashFunc(key);
+        size_t b = code % numBuckets;
+        
+        // Go to the bucket
+        Node* n = bucket[b];
+
+        // No bucket, so this must be the first
+        if (n == NULL) {
+            bucket[b] = new Node(key, value, code, NULL);
+            ++_size;
+            return;
+        }
+
+        size_t bucketLength = 1;
+
+        // Sometimes a bad hash code will cause all elements
+        // to collide.  Detect this case and don't rehash when 
+        // it occurs; nothing good will come from the rehashing.
+        bool allSameCode = true;
+
+        // Try to find the node
+        do {
+            allSameCode = allSameCode && (code == n->hashCode);
+
+            if ((code == n->hashCode) && (n->entry.key == key)) {
+               // Replace the existing node.
+               n->entry.value = value;
+               return;
+            }
+
+            n = n->next;
+            ++bucketLength;
+        } while (n != NULL);
+
+        const size_t maxBucketLength = 5;
+        if ((bucketLength > maxBucketLength) & ! allSameCode && (numBuckets < _size * 20)) {
+            // This bucket was really large; rehash if all elements
+            // don't have the same hashcode the number of buckets is reasonable.
+            resize(numBuckets * 2 + 1);
+        }
+
+        // Not found; insert at the head.
+        b = code % numBuckets;
+        bucket[b] = new Node(key, value, code, bucket[b]);
+        ++_size;
+   }
+
+   /**
+    Removes an element from the table if it is present.  It is an error
+    to remove an element that isn't present.
+    */
+   void remove(const Key& key) {
+      
+      size_t code = m_HashFunc(key);
+      size_t b = code % numBuckets;
+
+      // Go to the bucket
+      Node* n = bucket[b];
+
+      // Make sure it was found
+      alwaysAssertM(n != NULL, "Tried to remove a key that was not in the table.");
+
+      Node* previous = NULL;
+
+      // Try to find the node
+      do {
+          if ((code == n->hashCode) && (n->entry.key == key)) {
+              // This is the node; remove it
+              if (previous == NULL) {
+                  bucket[b] = n->next;
+              } else {
+                  previous->next = n->next;
+              }
+              // Delete the node
+              delete n;
+              --_size;
+              return;
+          }
+
+          previous = n;
+          n = n->next;
+       } while (n != NULL);
+
+
+      alwaysAssertM(false, "Tried to remove a key that was not in the table.");
+   }
+
+   /**
+    Returns the value associated with key.
+    @deprecated Use get(key, val) or 
+    */
+   Value& get(const Key& key) const {
+
+        size_t  code = m_HashFunc(key);
+        size_t b = code % numBuckets;
+
+        Node* node = bucket[b];
+
+        while (node != NULL) {
+            if ((node->hashCode == code) && (node->entry.key == key)) {
+                return node->entry.value;
+            }
+            node = node->next;
+        }
+
+        debugAssertM(false, "Key not found");
+        // The next line is here just to make
+        // a compiler warning go away.
+        return node->entry.value;
+   }
+
+   /**
+    If the key is present in the table, val is set to the associated value and returns true.
+    If the key is not present, returns false.
+    */
+   bool get(const Key& key, Value& val) const {
+      size_t code = m_HashFunc(key);
+      size_t b = code % numBuckets;
+
+      Node* node = bucket[b];
+
+      while (node != NULL) {
+          if ((node->hashCode == code) && (node->entry.key == key)) {
+             // found key
+             val = node->entry.value;
+             return true;
+          }
+          node = node->next;
+      }
+
+      // Failed to find key
+      return false;
+   }
+
+   /**
+    Returns true if key is in the table.
+    */
+   bool containsKey(const Key& key) const {
+       size_t code = m_HashFunc(key);
+       size_t b = code % numBuckets;
+
+       Node* node = bucket[b];
+
+       while (node != NULL) {
+           if ((node->hashCode == code) && (node->entry.key == key)) {
+              return true;
+           }
+           node = node->next;
+       } while (node != NULL);
+
+       return false;
+   }
+
+
+   /**
+    Short syntax for get.
+    */
+   inline Value& operator[](const Key &key) const {
+      return get(key);
+   }
+
+
+   /**
+    Returns an array of all of the keys in the table.
+    You can iterate over the keys to get the values.
+    */
+   Array<Key> getKeys() const {
+       Array<Key> keyArray;
+       getKeys(keyArray);
+       return keyArray;
+   }
+
+   void getKeys(Array<Key>& keyArray) const {
+       keyArray.resize(0, DONT_SHRINK_UNDERLYING_ARRAY);
+       for (size_t i = 0; i < numBuckets; i++) {
+           Node* node = bucket[i];
+           while (node != NULL) {
+               keyArray.append(node->entry.key);
+               node = node->next;
+           }
+       }
+   }
+
+   /**
+    Calls delete on all of the keys.  Does not clear the table, 
+    however, so you are left with a table of dangling pointers.
+
+    Same as <CODE>getKeys().deleteAll();</CODE>
+
+    To delete all of the values, you may want something like
+    <PRE>
+        Array<Key> keys = table.getKeys();
+        Set<Value> value;
+        for (int k = 0; k < keys.length(); k++) {
+           value.insert(keys[k]);
+        }
+        value.getMembers().deleteAll();
+        keys.deleteAll();
+    </PRE>
+    */
+   void deleteKeys() {
+       getKeys().deleteAll();
+   }
+
+   /**
+    Calls delete on all of the values.  This is unsafe--
+    do not call unless you know that each value appears
+    at most once.
+
+    Does not clear the table, so you are left with a table
+    of dangling pointers.
+    */
+   void deleteValues() {
+       for (int i = 0; i < numBuckets; i++) {
+           Node* node = bucket[i];
+           while (node != NULL) {
+               delete node->entry.value;
+               node = node->next;
+           }
+       }
+   }
+};
+
+} // namespace
+
+#ifdef G3D_WIN32
+#   pragma warning (pop)
+#endif
+
+#endif
diff --git a/dep/include/g3dlite/G3D/Triangle.h b/dep/include/g3dlite/G3D/Triangle.h
new file mode 100644
index 00000000000..cbe46a8cf50
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Triangle.h
@@ -0,0 +1,116 @@
+/**
+ @file Triangle.h
+  
+ @maintainer Morgan McGuire, matrix@graphics3d.com
+ 
+ @created 2003-04-05
+ @edited  2004-03-14
+
+ @cite Random point method by  Greg Turk, Generating random points in triangles.  In A. S. Glassner, ed., Graphics Gems, pp. 24-28. Academic Press, 1990
+
+ Copyright 2000-2006, Morgan McGuire.
+ All rights reserved.
+ */
+
+#ifndef G3D_TRIANGLE_H
+#define G3D_TRIANGLE_H
+
+#include "G3D/platform.h"
+#include "G3D/g3dmath.h"
+#include "G3D/Vector3.h"
+#include "G3D/Plane.h"
+#include <string>
+
+namespace G3D {
+
+/**
+  A generic triangle representation.  This should not be used
+  as the underlying triangle for creating models; it is intended
+  for providing fast property queries but requires a lot of
+  storage and is mostly immutable.
+  */
+class Triangle {
+private:
+    friend class CollisionDetection;
+    friend class Ray;
+
+    Vector3                     _vertex[3];
+
+    /** edgeDirection[i] is the normalized vector v[i+1] - v[i] */
+    Vector3                     edgeDirection[3];
+    double                      edgeMagnitude[3];
+    Plane                       _plane;
+    Vector3::Axis               _primaryAxis;
+
+    /** vertex[1] - vertex[0] */
+    Vector3                     edge01;
+    /** vertex[2] - vertex[0] */
+    Vector3                     edge02;
+
+    float                       _area;
+
+    void init(const Vector3& v0, const Vector3& v1, const Vector3& v2);
+
+public:
+	
+    Triangle();
+    
+    Triangle(const Vector3& v0, const Vector3& v1, const Vector3& v2);
+    
+    ~Triangle();
+
+    /** 0, 1, or 2 */
+    inline const Vector3& vertex(int n) const {
+        debugAssert((n >= 0) && (n < 3));
+        return _vertex[n];
+    }
+
+    double area() const;
+
+    Vector3::Axis primaryAxis() const {
+        return _primaryAxis;
+    }
+
+    const Vector3& normal() const;
+
+    /** Barycenter */
+    Vector3 center() const;
+
+    const Plane& plane() const;
+
+    /** Returns a random point in the triangle. */
+    Vector3 randomPoint() const;
+
+    /**
+     For two triangles to be equal they must have
+     the same vertices <I>in the same order</I>.
+     That is, vertex[0] == vertex[0], etc.
+     */
+    inline bool operator==(const Triangle& other) const {
+        for (int i = 0; i < 3; ++i) {
+            if (_vertex[i] != other._vertex[i]) {
+                return false;
+            }
+        }
+
+        return true;
+    }
+
+    inline unsigned int hashCode() const {
+        return
+            _vertex[0].hashCode() +
+            (_vertex[1].hashCode() >> 2) +
+            _vertex[2].hashCode();
+    }
+
+    void getBounds(class AABox&) const;
+
+};
+
+} // namespace
+
+inline unsigned int hashCode(const G3D::Triangle& t) {
+	return t.hashCode();
+}
+
+#endif
diff --git a/dep/include/g3dlite/G3D/Vector2.h b/dep/include/g3dlite/G3D/Vector2.h
new file mode 100644
index 00000000000..2d9d2662646
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Vector2.h
@@ -0,0 +1,438 @@
+/**
+  @file Vector2.h
+ 
+  2D vector class
+ 
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+  
+  @created 2001-06-02
+  @edited  2006-01-14
+  Copyright 2000-2006, Morgan McGuire.
+  All rights reserved.
+*/
+
+#ifndef G3D_VECTOR2_H
+#define G3D_VECTOR2_H
+
+#include "G3D/platform.h"
+#include "G3D/g3dmath.h"
+#include "Vector2int16.h"
+#include <string>
+
+namespace G3D {
+
+class Vector2;    
+class Vector3;
+class Vector4;
+
+/**
+ Do not subclass-- this implementation makes assumptions about the
+ memory layout.
+ */
+class Vector2 {
+private:
+    // Hidden operators
+    bool operator<(const Vector2&) const;
+    bool operator>(const Vector2&) const;
+    bool operator<=(const Vector2&) const;
+    bool operator>=(const Vector2&) const;
+
+public:
+    // coordinates
+    float x, y;
+
+    // construction
+    Vector2();
+    Vector2(float x, float y);
+    Vector2(float coordinate[2]);
+    Vector2(double coordinate[2]);
+    Vector2(const Vector2& rkVector);
+    Vector2(const class Vector2int16& v); 
+
+    float& operator[] (int i);
+    const float& operator[] (int i) const;
+    operator float* ();
+    operator const float* () const;
+
+    // assignment and comparison
+    Vector2& operator= (const Vector2& rkVector);
+    bool operator== (const Vector2& rkVector) const;
+    bool operator!= (const Vector2& rkVector) const;
+    unsigned int hashCode() const;
+    bool fuzzyEq(const Vector2& other) const;
+    bool fuzzyNe(const Vector2& other) const;
+    /** Returns true if this vector has finite length */
+    bool isFinite() const;
+
+    /** Returns true if this vector has length == 0 */
+    bool isZero() const;
+
+    /** Returns true if this vector has length == 1 */
+    bool isUnit() const;
+
+    // arithmetic operations
+    Vector2 operator+ (const Vector2& rkVector) const;
+    Vector2 operator- (const Vector2& rkVector) const;
+    Vector2 operator* (float fScalar) const;
+    Vector2 operator* (const Vector2& rkVector) const;
+    Vector2 operator/ (const Vector2& rkVector) const;
+    Vector2 operator/ (float fScalar) const;
+    Vector2 operator- () const;
+
+    inline float sum() const {
+        return x + y;
+    }
+
+    /**
+     Linear interpolation
+     */
+    inline Vector2 lerp(const Vector2& v, float alpha) const {
+        return (*this) + (v - *this) * alpha; 
+    }
+
+    inline Vector2 clamp(const Vector2& low, const Vector2& high) const {
+        return Vector2(
+            G3D::clamp(x, low.x, high.x),
+            G3D::clamp(y, low.y, high.y));
+    }
+
+    inline Vector2 clamp(float low, float high) const {
+        return Vector2(
+            (float)G3D::clamp(x, low, high),
+            (float)G3D::clamp(y, low, high));
+    }
+
+    // arithmetic updates
+    Vector2& operator+= (const Vector2& rkVector);
+    Vector2& operator-= (const Vector2& rkVector);
+    Vector2& operator*= (float fScalar);
+    Vector2& operator/= (float fScalar);
+    Vector2& operator*= (const Vector2& rkVector);
+    Vector2& operator/= (const Vector2& rkVector);
+
+    // vector operations
+    float length() const;
+    Vector2 direction() const;
+    /**
+     Potentially less accurate but faster than direction().
+     Only works if System::hasSSE is true.
+     */
+    Vector2 fastDirection() const {
+        return direction();
+    }
+
+    float squaredLength () const;
+    float dot (const Vector2& rkVector) const;
+    float unitize (float fTolerance = 1e-06);
+
+    Vector2 min(const Vector2 &v) const;
+    Vector2 max(const Vector2 &v) const;
+
+    // Random unit vector
+    static Vector2 random();
+
+    // Special values.
+    // Intentionally not inlined: see Matrix3::identity() for details.
+    static const Vector2& zero();
+    inline static const Vector2& one() { static const Vector2 v(1, 1); return v; }
+    static const Vector2& unitX();
+    static const Vector2& unitY();
+    static const Vector2& inf(); 
+    static const Vector2& nan();
+    /** smallest (most negative) representable vector */
+    static const Vector2& minFinite(); 
+    /** Largest representable vector */
+    static const Vector2& maxFinite();
+
+
+
+    // Deprecated. See Matrix3::identity() for details.
+    /** @deprecated Use Vector2::zero() */
+    static const Vector2 ZERO;
+    /** @deprecated Use Vector2::unitX() */
+    static const Vector2 UNIT_S;
+    /** @deprecated Use Vector2::unitY() */
+    static const Vector2 UNIT_T;
+
+    std::string toString() const;
+
+    // 2-char swizzles
+
+    Vector2 xx() const;
+    Vector2 yx() const;
+    Vector2 xy() const;
+    Vector2 yy() const;
+
+    // 3-char swizzles
+
+    Vector3 xxx() const;
+    Vector3 yxx() const;
+    Vector3 xyx() const;
+    Vector3 yyx() const;
+    Vector3 xxy() const;
+    Vector3 yxy() const;
+    Vector3 xyy() const;
+    Vector3 yyy() const;
+
+    // 4-char swizzles
+
+    Vector4 xxxx() const;
+    Vector4 yxxx() const;
+    Vector4 xyxx() const;
+    Vector4 yyxx() const;
+    Vector4 xxyx() const;
+    Vector4 yxyx() const;
+    Vector4 xyyx() const;
+    Vector4 yyyx() const;
+    Vector4 xxxy() const;
+    Vector4 yxxy() const;
+    Vector4 xyxy() const;
+    Vector4 yyxy() const;
+    Vector4 xxyy() const;
+    Vector4 yxyy() const;
+    Vector4 xyyy() const;
+    Vector4 yyyy() const;
+
+};
+
+inline Vector2 operator*(double s, const Vector2& v) {
+    return v * (float)s;
+}
+
+inline Vector2 operator*(float s, const Vector2& v) {
+    return v * s;
+}
+
+inline Vector2 operator*(int s, const Vector2& v) {
+    return v * (float)s;
+}
+
+
+inline unsigned int hashCode(const G3D::Vector2& v) {
+     return v.hashCode();
+}
+
+
+inline Vector2::Vector2 () : x(0.0f), y(0.0f) {
+}
+
+
+inline Vector2::Vector2(float _x, float _y) : x(_x), y(_y) {
+}
+
+
+inline Vector2::Vector2 (float afCoordinate[2]) {
+    x = afCoordinate[0];
+    y = afCoordinate[1];
+}
+
+
+
+inline Vector2::Vector2 (double afCoordinate[2]) {
+    x = (float)afCoordinate[0];
+    y = (float)afCoordinate[1];
+}
+
+
+inline Vector2::Vector2 (const Vector2& rkVector) {
+    x = rkVector.x;
+    y = rkVector.y;
+}
+
+
+inline Vector2::Vector2 (const Vector2int16& v) : x(v.x), y(v.y) {
+}
+
+
+inline float& Vector2::operator[] (int i) {
+    return ((float*)this)[i];
+}
+
+
+inline const float& Vector2::operator[] (int i) const {
+    return ((float*)this)[i];
+}
+
+
+inline Vector2::operator float* () {
+    return (float*)this;
+}
+
+inline Vector2::operator const float* () const {
+    return (float*)this;
+}
+
+
+inline Vector2& Vector2::operator= (const Vector2& rkVector) {
+    x = rkVector.x;
+    y = rkVector.y;
+    return *this;
+}
+
+
+inline bool Vector2::operator== (const Vector2& rkVector) const {
+    return ( x == rkVector.x && y == rkVector.y);
+}
+
+
+inline bool Vector2::operator!= (const Vector2& rkVector) const {
+    return ( x != rkVector.x || y != rkVector.y);
+}
+
+
+inline Vector2 Vector2::operator+ (const Vector2& rkVector) const {
+    return Vector2(x + rkVector.x, y + rkVector.y);
+}
+
+
+inline Vector2 Vector2::operator- (const Vector2& rkVector) const {
+    return Vector2(x - rkVector.x, y - rkVector.y);
+}
+
+
+inline Vector2 Vector2::operator* (float fScalar) const {
+    return Vector2(fScalar*x, fScalar*y);
+}
+
+
+
+inline Vector2 Vector2::operator- () const {
+    return Vector2( -x, -y);
+}
+
+
+
+inline Vector2& Vector2::operator+= (const Vector2& rkVector) {
+    x += rkVector.x;
+    y += rkVector.y;
+    return *this;
+}
+
+
+
+inline Vector2& Vector2::operator-= (const Vector2& rkVector) {
+    x -= rkVector.x;
+    y -= rkVector.y;
+    return *this;
+}
+
+
+
+inline Vector2& Vector2::operator*= (float fScalar) {
+    x *= fScalar;
+    y *= fScalar;
+    return *this;
+}
+
+
+
+
+inline Vector2& Vector2::operator*= (const Vector2& rkVector) {
+    x *= rkVector.x;
+    y *= rkVector.y;
+    return *this;
+}
+
+
+
+inline Vector2& Vector2::operator/= (const Vector2& rkVector) {
+    x /= rkVector.x;
+    y /= rkVector.y;
+    return *this;
+}
+
+
+inline Vector2 Vector2::operator* (const Vector2& rkVector) const {
+    return Vector2(x * rkVector.x, y * rkVector.y);
+}
+
+
+
+inline Vector2 Vector2::operator/ (const Vector2& rkVector) const {
+    return Vector2(x / rkVector.x, y / rkVector.y);
+}
+
+
+inline float Vector2::squaredLength () const {
+    return x*x + y*y;
+}
+
+
+inline float Vector2::length () const {
+    return sqrtf(x*x + y*y);
+}
+
+
+inline Vector2 Vector2::direction () const {
+    float lenSquared = x * x + y * y;
+
+    if (lenSquared != 1.0f) {
+        return *this / sqrtf(lenSquared);
+    } else {
+        return *this;
+    }
+}
+
+
+
+inline float Vector2::dot (const Vector2& rkVector) const {
+    return x*rkVector.x + y*rkVector.y;
+}
+
+
+
+inline Vector2 Vector2::min(const Vector2 &v) const {
+    return Vector2(G3D::min(v.x, x), G3D::min(v.y, y));
+}
+
+
+
+inline Vector2 Vector2::max(const Vector2 &v) const {
+    return Vector2(G3D::max(v.x, x), G3D::max(v.y, y));
+}
+
+
+
+inline bool Vector2::fuzzyEq(const Vector2& other) const {
+    return G3D::fuzzyEq((*this - other).squaredLength(), 0);
+}
+
+
+
+inline bool Vector2::fuzzyNe(const Vector2& other) const {
+    return G3D::fuzzyNe((*this - other).squaredLength(), 0);
+}
+
+
+
+inline bool Vector2::isFinite() const {
+    return G3D::isFinite(x) && G3D::isFinite(y);
+}
+
+
+
+inline bool Vector2::isZero() const {
+    return (x == 0.0f) && (y == 0.0f);
+}
+
+
+
+inline bool Vector2::isUnit() const {
+    return squaredLength() == 1.0f;
+}
+
+}
+
+// Intentionally outside namespace to avoid operator overloading confusion
+inline G3D::Vector2 operator*(double s, const G3D::Vector2& v) {
+    return v * (float)s;
+}
+inline G3D::Vector2 operator*(int s, const G3D::Vector2& v) {
+    return v * (float)s;
+}
+
+
+inline unsigned int hashCode(const G3D::Vector2& v);
+
+
+#endif
diff --git a/dep/include/g3dlite/G3D/Vector2.inl b/dep/include/g3dlite/G3D/Vector2.inl
new file mode 100644
index 00000000000..27abecf02ce
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Vector2.inl
@@ -0,0 +1,20 @@
+/**
+ @file Vector2.inl
+ 
+ @maintainer Morgan McGuire, matrix@graphics3d.com
+ @cite Portions by Laura Wollstadt, graphics3d.com
+ 
+ @cite Portions based on Dave Eberly'x Magic Software Library
+ at http://www.magic-software.com
+ 
+ 
+ @created 2001-06-02
+ @edited  2006-01-14
+
+  Copyright 2000-2006, Morgan McGuire.
+  All rights reserved.
+ */
+
+
+}
+
diff --git a/dep/include/g3dlite/G3D/Vector2int16.h b/dep/include/g3dlite/G3D/Vector2int16.h
new file mode 100644
index 00000000000..cadb4a80c5f
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Vector2int16.h
@@ -0,0 +1,75 @@
+/**
+  @file Vector2int16.h
+  
+  @maintainer Morgan McGuire, matrix@brown.edu
+
+  @created 2003-08-09
+  @edited  2004-01-03
+
+  Copyright 2000-2006, Morgan McGuire.
+  All rights reserved.
+ */
+
+#ifndef VECTOR2INT16_H
+#define VECTOR2INT16_H
+
+#include "G3D/platform.h"
+#include "G3D/g3dmath.h"
+
+namespace G3D {
+
+/**
+ A Vector2 that packs its fields into uint16s.
+ */
+#ifdef G3D_WIN32
+    // Switch to tight alignment
+    #pragma pack(push, 2)
+#endif
+
+class Vector2int16 {
+private:
+    // Hidden operators
+    bool operator<(const Vector2int16&) const;
+    bool operator>(const Vector2int16&) const;
+    bool operator<=(const Vector2int16&) const;
+    bool operator>=(const Vector2int16&) const;
+
+public:
+    G3D::int16              x;
+    G3D::int16              y;
+
+    Vector2int16() : x(0), y(0) {}
+    Vector2int16(G3D::int16 _x, G3D::int16 _y) : x(_x), y(_y){}
+    Vector2int16(const class Vector2& v);
+
+    inline G3D::int16& operator[] (int i) {
+        debugAssert(((unsigned int)i) <= 1);
+        return ((G3D::int16*)this)[i];
+    }
+
+    inline const G3D::int16& operator[] (int i) const {
+        debugAssert(((unsigned int)i) <= 1);
+        return ((G3D::int16*)this)[i];
+    }
+
+
+    inline bool operator== (const Vector2int16& rkVector) const {
+        return ((int32*)this)[0] == ((int32*)&rkVector)[0];
+    }
+
+    inline bool operator!= (const Vector2int16& rkVector) const {
+        return ((int32*)this)[0] != ((int32*)&rkVector)[0];
+    }
+
+}
+#if defined(G3D_LINUX) || defined(G3D_OSX)
+    __attribute((aligned(1)))
+#endif
+;
+
+#ifdef G3D_WIN32
+    #pragma pack(pop)
+#endif
+
+}
+#endif
diff --git a/dep/include/g3dlite/G3D/Vector3.h b/dep/include/g3dlite/G3D/Vector3.h
new file mode 100644
index 00000000000..79c28fb5e7d
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Vector3.h
@@ -0,0 +1,504 @@
+/**
+  @file Vector3.h
+ 
+  3D vector class
+ 
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+
+  @created 2001-06-02
+  @edited  2005-08-23
+  Copyright 2000-2006, Morgan McGuire.
+  All rights reserved.
+ */
+
+#ifndef G3D_VECTOR3_H
+#define G3D_VECTOR3_H
+
+#include "G3D/platform.h"
+#include "G3D/g3dmath.h"
+#include "G3D/Vector2.h"
+#include <iostream>
+#include <string>
+
+namespace G3D {
+
+class Vector2;    
+class Vector3;
+class Vector4;
+
+/**
+  <B>Swizzles</B>
+ Vector classes have swizzle operators, e.g. <CODE>v.xy()</CODE>, that
+ allow selection of arbitrary sub-fields.  These cannot be used as write 
+ masks.  Examples
+
+  <PRE>
+Vector3 v(1, 2, 3);
+Vector3 j;
+Vector2 b;
+
+b = v.xz();
+j = b.xx();
+</PRE>
+
+
+  <B>Warning</B>
+
+ Do not subclass-- this implementation makes assumptions about the
+ memory layout.
+ */
+class Vector3 {
+private:
+    /**
+     Reflect this vector about the (not necessarily unit) normal.
+     Note that if used for a collision or ray reflection you
+     must negate the resulting vector to get a direction pointing
+     <I>away</I> from the collision.
+
+     <PRE>
+       V'    N      V
+                 
+         r   ^   -,
+          \  |  /
+            \|/
+     </PRE>
+
+     See also Vector3::reflectionDirection
+     */
+    Vector3 reflectAbout(const Vector3& normal) const;
+
+    // Hidden operators
+    bool operator<(const Vector3&) const;
+    bool operator>(const Vector3&) const;
+    bool operator<=(const Vector3&) const;
+    bool operator>=(const Vector3&) const;
+
+public:
+    // construction
+    Vector3();
+    Vector3(float _x, float _y, float _z);
+    Vector3(const class Vector2& v, float _z);
+    Vector3(float coordinate[3]);
+    Vector3(double coordinate[3]);
+    Vector3(const Vector3& rkVector);
+    Vector3(const class Vector3int16& v);
+
+    // coordinates
+    float x, y, z;
+
+    // access vector V as V[0] = V.x, V[1] = V.y, V[2] = V.z
+    //
+    // WARNING.  These member functions rely on
+    // (1) Vector3 not having virtual functions
+    // (2) the data packed in a 3*sizeof(float) memory block
+    const float& operator[] (int i) const;
+    float& operator[] (int i);
+
+    inline operator float* () {
+        return (float*)this;
+    }
+
+    operator const float* () const {
+        return (float*)this;
+    }
+
+    enum Axis {X_AXIS=0, Y_AXIS=1, Z_AXIS=2, DETECT_AXIS=-1};
+
+    /**
+     Returns the largest dimension.  Particularly convenient for determining
+     which plane to project a triangle onto for point-in-polygon tests.
+     */
+    Axis primaryAxis() const;
+
+    // assignment and comparison
+    Vector3& operator= (const Vector3& rkVector);
+    bool operator== (const Vector3& rkVector) const;
+    bool operator!= (const Vector3& rkVector) const;
+    unsigned int hashCode() const;
+    bool fuzzyEq(const Vector3& other) const;
+    bool fuzzyNe(const Vector3& other) const;
+
+    /** Returns true if this vector has finite length. */
+    bool isFinite() const;
+
+    /** Returns true if this vector has length ~= 0 */
+    bool isZero() const;
+
+    /** Returns true if this vector has length ~= 1 */
+    bool isUnit() const;
+    
+    // arithmetic operations
+    Vector3 operator+ (const Vector3& v) const;
+    Vector3 operator- (const Vector3& v) const;
+    Vector3 operator* (float s) const;
+    Vector3 operator/ (float s) const;
+    Vector3 operator* (const Vector3& v) const;
+    Vector3 operator/ (const Vector3& v) const;
+    Vector3 operator- () const;
+
+    // arithmetic updates
+    Vector3& operator+= (const Vector3& v);
+    Vector3& operator-= (const Vector3& v);
+    Vector3& operator*= (float s);
+    Vector3& operator/= (float s);
+    Vector3& operator*= (const Vector3& v);
+    Vector3& operator/= (const Vector3& v);
+
+    /** @deprecated Use magnitude */
+	float G3D_DEPRECATED length() const;
+
+    float magnitude() const;
+    
+    /**
+     The result is a nan vector if the length is almost zero.
+     */
+    Vector3 direction() const;
+
+    /**
+     Potentially less accurate but faster than direction().
+     Only works if System::hasSSE is true.
+     */
+    Vector3 fastDirection() const;
+
+
+    /**
+      See also G3D::Ray::reflect.
+      The length is 1. 
+     <PRE>
+       V'    N       V
+                 
+         r   ^    /
+          \  |  /
+            \|'-
+     </PRE>
+     */
+    Vector3 reflectionDirection(const Vector3& normal) const;
+    
+
+    /**
+     Returns Vector3::zero() if the length is nearly zero, otherwise
+     returns a unit vector.
+     */
+    inline Vector3 directionOrZero() const {
+        float mag = magnitude();
+        if (G3D::fuzzyEq(mag, 0.0f)) {
+            return Vector3::zero();
+        } else if (G3D::fuzzyEq(mag, 1.0f)) {
+            return *this;
+        } else {
+            return *this * (1.0f / mag);
+        }
+    }
+
+    /**
+     Returns the direction of a refracted ray,
+     where iExit is the index of refraction for the
+     previous material and iEnter is the index of refraction
+     for the new material.  Like Vector3::reflectionDirection,
+     the result has length 1 and is 
+     pointed <I>away</I> from the intersection.
+
+     Returns Vector3::zero() in the case of total internal refraction.
+
+     @param iOutside The index of refraction (eta) outside
+     (on the <I>positive</I> normal side) of the surface.
+
+     @param iInside The index of refraction (eta) inside
+     (on the <I>negative</I> normal side) of the surface.
+
+     See also G3D::Ray::refract.
+     <PRE>
+              N      V
+                  
+              ^    /
+              |  /
+              |'-
+          __--
+     V'<--
+     </PRE>
+     */
+    Vector3 refractionDirection(
+        const Vector3&  normal,
+        float           iInside,
+        float           iOutside) const;
+
+    /** Synonym for direction */
+    inline Vector3 unit() const {
+        return direction();
+    }
+
+    /** Returns a normalized vector.  May be computed with lower precision than unit */
+    inline Vector3 fastUnit() const {
+        return fastDirection();
+    }
+
+    /** @deprecated Use squaredMagnitude */
+    float G3D_DEPRECATED squaredLength() const;
+
+    float squaredMagnitude () const;
+	
+    /** @deprecated Use squaredMagnitude */
+    inline float G3D_DEPRECATED norm() const {
+        return squaredMagnitude();
+    }
+
+    float dot(const Vector3& rkVector) const;
+    
+    float G3D_DEPRECATED unitize(float fTolerance = 1e-06);
+
+    /** Cross product.  Note that two cross products in a row
+        can be computed more cheaply: v1 x (v2 x v3) = (v1 dot v3) v2  - (v1 dot v2) v3.
+      */
+    Vector3 cross(const Vector3& rkVector) const;
+    Vector3 unitCross (const Vector3& rkVector) const;
+
+    /**
+     Returns a matrix such that v.cross() * w = v.cross(w).
+     <PRE>
+     [ 0  -v.z  v.y ]
+     [ v.z  0  -v.x ]
+     [ -v.y v.x  0  ]
+     </PRE>
+     */
+    class Matrix3 cross() const;
+
+    Vector3 min(const Vector3 &v) const;
+    Vector3 max(const Vector3 &v) const;
+
+    std::string toString() const;
+
+    inline Vector3 clamp(const Vector3& low, const Vector3& high) const {
+        return Vector3(
+            G3D::clamp(x, low.x, high.x),
+            G3D::clamp(y, low.y, high.y),
+            G3D::clamp(z, low.z, high.z));
+    }
+
+    inline Vector3 clamp(float low, float high) const {
+        return Vector3(
+            G3D::clamp(x, low, high),
+            G3D::clamp(y, low, high),
+            G3D::clamp(z, low, high));
+    }
+
+    /**
+     Linear interpolation
+     */
+    inline Vector3 lerp(const Vector3& v, float alpha) const {
+        return (*this) + (v - *this) * alpha; 
+    }
+
+    /** Gram-Schmidt orthonormalization. */
+    static void orthonormalize (Vector3 akVector[3]);
+
+    /** Random unit vector, uniformly distributed */
+    static Vector3 random();
+
+    /** Random unit vector, distributed
+        so that the probability of V is proportional 
+        to max(V dot Normal, 0).
+
+        @cite Henrik Wann Jensen, Realistic Image Synthesis using Photon Mapping eqn 2.24
+    */
+    static Vector3 cosRandom(const Vector3& normal);
+
+
+    /**
+     Random vector distributed over the hemisphere about normal.
+     */
+    static Vector3 hemiRandom(const Vector3& normal);
+
+    // Input W must be initialize to a nonzero vector, output is {U,V,W}
+    // an orthonormal basis.  A hint is provided about whether or not W
+    // is already unit length.
+    static void generateOrthonormalBasis (Vector3& rkU, Vector3& rkV,
+                                          Vector3& rkW, bool bUnitLengthW = true);
+
+    inline float sum() const {
+        return x + y + z;
+    }
+
+    inline float average() const {
+        return sum() / 3.0f;
+    }
+
+    // Special values.
+    inline static const Vector3& zero()     { static Vector3 v(0, 0, 0); return v; }
+    inline static const Vector3& one()      { static Vector3 v(1, 1, 1); return v; }
+    inline static const Vector3& unitX()    { static Vector3 v(1, 0, 0); return v; }
+    inline static const Vector3& unitY()    { static Vector3 v(0, 1, 0); return v; }
+    inline static const Vector3& unitZ()    { static Vector3 v(0, 0, 1); return v; }
+    inline static const Vector3& inf()      { static Vector3 v((float)G3D::inf(), (float)G3D::inf(), (float)G3D::inf()); return v; }
+    inline static const Vector3& nan()      { static Vector3 v((float)G3D::nan(), (float)G3D::nan(), (float)G3D::nan()); return v; }
+    /** Smallest (most negative) representable vector */
+    inline static const Vector3& minFinite(){ static Vector3 v(-FLT_MAX, -FLT_MAX, -FLT_MAX); return v; }
+    /** Largest representable vector */
+    inline static const Vector3& maxFinite(){ static Vector3 v(FLT_MAX, FLT_MAX, FLT_MAX); return v; }
+
+    // Deprecated. See Matrix3::identity() for details.
+    /** @deprecated Use Vector3::zero() */
+    static const Vector3 ZERO;
+    /** @deprecated Use Vector3::zero() */
+    static const Vector3 ZERO3;
+    /** @deprecated Use Vector3::unitX() */
+    static const Vector3 UNIT_X;
+    /** @deprecated Use Vector3::unitY() */
+    static const Vector3 UNIT_Y;
+    /** @deprecated Use Vector3::unitZ() */
+    static const Vector3 UNIT_Z;
+    /** @deprecated Use Vector3::inf() */
+    static const Vector3 INF3;
+    /** @deprecated Use Vector3::nan() */
+    static const Vector3 NAN3;
+
+    // 2-char swizzles
+
+    Vector2 xx() const;
+    Vector2 yx() const;
+    Vector2 zx() const;
+    Vector2 xy() const;
+    Vector2 yy() const;
+    Vector2 zy() const;
+    Vector2 xz() const;
+    Vector2 yz() const;
+    Vector2 zz() const;
+
+    // 3-char swizzles
+
+    Vector3 xxx() const;
+    Vector3 yxx() const;
+    Vector3 zxx() const;
+    Vector3 xyx() const;
+    Vector3 yyx() const;
+    Vector3 zyx() const;
+    Vector3 xzx() const;
+    Vector3 yzx() const;
+    Vector3 zzx() const;
+    Vector3 xxy() const;
+    Vector3 yxy() const;
+    Vector3 zxy() const;
+    Vector3 xyy() const;
+    Vector3 yyy() const;
+    Vector3 zyy() const;
+    Vector3 xzy() const;
+    Vector3 yzy() const;
+    Vector3 zzy() const;
+    Vector3 xxz() const;
+    Vector3 yxz() const;
+    Vector3 zxz() const;
+    Vector3 xyz() const;
+    Vector3 yyz() const;
+    Vector3 zyz() const;
+    Vector3 xzz() const;
+    Vector3 yzz() const;
+    Vector3 zzz() const;
+
+    // 4-char swizzles
+
+    Vector4 xxxx() const;
+    Vector4 yxxx() const;
+    Vector4 zxxx() const;
+    Vector4 xyxx() const;
+    Vector4 yyxx() const;
+    Vector4 zyxx() const;
+    Vector4 xzxx() const;
+    Vector4 yzxx() const;
+    Vector4 zzxx() const;
+    Vector4 xxyx() const;
+    Vector4 yxyx() const;
+    Vector4 zxyx() const;
+    Vector4 xyyx() const;
+    Vector4 yyyx() const;
+    Vector4 zyyx() const;
+    Vector4 xzyx() const;
+    Vector4 yzyx() const;
+    Vector4 zzyx() const;
+    Vector4 xxzx() const;
+    Vector4 yxzx() const;
+    Vector4 zxzx() const;
+    Vector4 xyzx() const;
+    Vector4 yyzx() const;
+    Vector4 zyzx() const;
+    Vector4 xzzx() const;
+    Vector4 yzzx() const;
+    Vector4 zzzx() const;
+    Vector4 xxxy() const;
+    Vector4 yxxy() const;
+    Vector4 zxxy() const;
+    Vector4 xyxy() const;
+    Vector4 yyxy() const;
+    Vector4 zyxy() const;
+    Vector4 xzxy() const;
+    Vector4 yzxy() const;
+    Vector4 zzxy() const;
+    Vector4 xxyy() const;
+    Vector4 yxyy() const;
+    Vector4 zxyy() const;
+    Vector4 xyyy() const;
+    Vector4 yyyy() const;
+    Vector4 zyyy() const;
+    Vector4 xzyy() const;
+    Vector4 yzyy() const;
+    Vector4 zzyy() const;
+    Vector4 xxzy() const;
+    Vector4 yxzy() const;
+    Vector4 zxzy() const;
+    Vector4 xyzy() const;
+    Vector4 yyzy() const;
+    Vector4 zyzy() const;
+    Vector4 xzzy() const;
+    Vector4 yzzy() const;
+    Vector4 zzzy() const;
+    Vector4 xxxz() const;
+    Vector4 yxxz() const;
+    Vector4 zxxz() const;
+    Vector4 xyxz() const;
+    Vector4 yyxz() const;
+    Vector4 zyxz() const;
+    Vector4 xzxz() const;
+    Vector4 yzxz() const;
+    Vector4 zzxz() const;
+    Vector4 xxyz() const;
+    Vector4 yxyz() const;
+    Vector4 zxyz() const;
+    Vector4 xyyz() const;
+    Vector4 yyyz() const;
+    Vector4 zyyz() const;
+    Vector4 xzyz() const;
+    Vector4 yzyz() const;
+    Vector4 zzyz() const;
+    Vector4 xxzz() const;
+    Vector4 yxzz() const;
+    Vector4 zxzz() const;
+    Vector4 xyzz() const;
+    Vector4 yyzz() const;
+    Vector4 zyzz() const;
+    Vector4 xzzz() const;
+    Vector4 yzzz() const;
+    Vector4 zzzz() const;
+
+    /** A value that can be passed to ignore a parameter.  Never look at the result of dummy. */
+    static Vector3 dummy;
+};
+
+inline G3D::Vector3 operator*(float s, const G3D::Vector3& v) {
+    return v * s;
+}
+
+inline G3D::Vector3 operator*(double s, const G3D::Vector3& v) {
+    return v * (float)s;
+}
+
+inline G3D::Vector3 operator*(int s, const G3D::Vector3& v) {
+    return v * (float)s;
+}
+
+std::ostream& operator<<(std::ostream& os, const Vector3&);
+
+}
+
+unsigned int hashCode(const G3D::Vector3& v);
+
+#include "Vector3.inl"
+
+#endif
diff --git a/dep/include/g3dlite/G3D/Vector3.inl b/dep/include/g3dlite/G3D/Vector3.inl
new file mode 100644
index 00000000000..99110ffc8eb
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Vector3.inl
@@ -0,0 +1,250 @@
+/** 
+  @file Vector3.inl
+
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+
+  @cite Portions based on Dave Eberly's Magic Software Library at http://www.magic-software.com
+
+  @created 2001-06-02
+  @edited  2004-05-21
+  Copyright 2000-2004, Morgan McGuire.
+  All rights reserved.
+ */
+
+//----------------------------------------------------------------------------
+#ifdef SSE
+	// If you receive an error on this line, it is because you do not have the file
+	// xmmintrin.h needed for MMX & SSE extensions.  Download and install
+    //
+    // http://download.microsoft.com/download/vstudio60ent/SP5/Wideband-Full/WIN98Me/EN-US/vs6sp5.exe
+	// and
+    // http://download.microsoft.com/download/vb60ent/Update/6/W9X2KXP/EN-US/vcpp5.exe
+    //
+    // to get this file.
+#   include <xmmintrin.h>
+#endif
+
+inline unsigned int hashCode(const G3D::Vector3& v) {
+     return v.hashCode();
+}
+
+namespace G3D {
+
+//----------------------------------------------------------------------------
+inline Vector3::Vector3() : x(0.0f), y(0.0f), z(0.0f) {
+}
+
+//----------------------------------------------------------------------------
+
+inline Vector3::Vector3 (float fX, float fY, float fZ) : x(fX), y(fY), z(fZ) {
+}
+
+//----------------------------------------------------------------------------
+inline Vector3::Vector3 (float V[3]) : x(V[0]), y(V[1]), z(V[2]){
+}
+//----------------------------------------------------------------------------
+inline Vector3::Vector3 (double V[3]) : x((float)V[0]), y((float)V[1]), z((float)V[2]){
+}
+
+//----------------------------------------------------------------------------
+inline Vector3::Vector3 (const Vector3& V) : x(V.x), y(V.y), z(V.z) {
+}
+
+//----------------------------------------------------------------------------
+
+//inline Vector3::Vector3 (const __m128& m) {
+    // Cast from SSE packed floats
+//    *this = *(Vector3*)&m;
+//}
+
+//----------------------------------------------------------------------------
+inline const float& Vector3::operator[] (int i) const {
+    return ((float*)this)[i];
+}
+
+inline float& Vector3::operator[] (int i) {
+    return ((float*)this)[i];
+}
+
+
+//----------------------------------------------------------------------------
+inline Vector3& Vector3::operator= (const Vector3& rkVector) {
+    x = rkVector.x;
+    y = rkVector.y;
+    z = rkVector.z;
+    return *this;
+}
+
+//----------------------------------------------------------------------------
+
+inline bool Vector3::fuzzyEq(const Vector3& other) const {
+    return G3D::fuzzyEq((*this - other).squaredMagnitude(), 0);
+}
+
+//----------------------------------------------------------------------------
+
+inline bool Vector3::fuzzyNe(const Vector3& other) const {
+    return G3D::fuzzyNe((*this - other).squaredMagnitude(), 0);
+}
+
+//----------------------------------------------------------------------------
+
+inline bool Vector3::isFinite() const {
+    return G3D::isFinite(x) && G3D::isFinite(y) && G3D::isFinite(z);
+}
+
+//----------------------------------------------------------------------------
+inline bool Vector3::operator== (const Vector3& rkVector) const {
+    return ( x == rkVector.x && y == rkVector.y && z == rkVector.z );
+}
+
+//----------------------------------------------------------------------------
+inline bool Vector3::operator!= (const Vector3& rkVector) const {
+    return ( x != rkVector.x || y != rkVector.y || z != rkVector.z );
+}
+
+//----------------------------------------------------------------------------
+inline Vector3 Vector3::operator+ (const Vector3& rkVector) const {
+    return Vector3(x + rkVector.x, y + rkVector.y, z + rkVector.z);
+}
+
+//----------------------------------------------------------------------------
+inline Vector3 Vector3::operator- (const Vector3& rkVector) const {
+    return Vector3(x - rkVector.x, y - rkVector.y, z - rkVector.z);
+}
+
+//----------------------------------------------------------------------------
+inline Vector3 Vector3::operator* (const Vector3& rkVector) const {
+    return Vector3(x * rkVector.x, y * rkVector.y, z * rkVector.z);
+}
+
+inline Vector3 Vector3::operator*(float f) const {
+    return Vector3(x * f, y * f, z * f);
+}
+
+//----------------------------------------------------------------------------
+inline Vector3 Vector3::operator/ (const Vector3& rkVector) const {
+    return Vector3(x / rkVector.x, y / rkVector.y, z / rkVector.z);
+}
+
+//----------------------------------------------------------------------------
+inline Vector3 Vector3::operator- () const {
+    return Vector3(-x, -y, -z);
+}
+
+//----------------------------------------------------------------------------
+inline Vector3& Vector3::operator+= (const Vector3& rkVector) {
+    x += rkVector.x;
+    y += rkVector.y;
+    z += rkVector.z;
+    return *this;
+}
+
+//----------------------------------------------------------------------------
+inline Vector3& Vector3::operator-= (const Vector3& rkVector) {
+    x -= rkVector.x;
+    y -= rkVector.y;
+    z -= rkVector.z;
+    return *this;
+}
+
+//----------------------------------------------------------------------------
+inline Vector3& Vector3::operator*= (float fScalar) {
+    x *= fScalar;
+    y *= fScalar;
+    z *= fScalar;
+    return *this;
+}
+
+//----------------------------------------------------------------------------
+inline Vector3& Vector3::operator*= (const Vector3& rkVector) {
+    x *= rkVector.x;
+    y *= rkVector.y;
+    z *= rkVector.z;
+    return *this;
+}
+
+//----------------------------------------------------------------------------
+inline Vector3& Vector3::operator/= (const Vector3& rkVector) {
+    x /= rkVector.x;
+    y /= rkVector.y;
+    z /= rkVector.z;
+    return *this;
+}
+
+//----------------------------------------------------------------------------
+inline float Vector3::squaredMagnitude () const {
+    return x*x + y*y + z*z;
+}
+
+//----------------------------------------------------------------------------
+inline float Vector3::squaredLength () const {
+    return squaredMagnitude();
+}
+
+//----------------------------------------------------------------------------
+inline float Vector3::magnitude() const {
+    return sqrtf(x*x + y*y + z*z);
+}
+
+//----------------------------------------------------------------------------
+inline float Vector3::length() const {
+    return magnitude();
+}
+
+//----------------------------------------------------------------------------
+inline Vector3 Vector3::direction () const {
+    float lenSquared = squaredMagnitude();
+    float invSqrt = 1.0f / sqrtf(lenSquared);
+    return Vector3(x * invSqrt, y * invSqrt, z * invSqrt);
+}
+
+//----------------------------------------------------------------------------
+
+inline Vector3 Vector3::fastDirection () const {
+    float lenSquared = x * x + y * y + z * z;
+    float invSqrt = rsq(lenSquared);
+    return Vector3(x * invSqrt, y * invSqrt, z * invSqrt);
+}
+
+//----------------------------------------------------------------------------
+inline float Vector3::dot (const Vector3& rkVector) const {
+    return x*rkVector.x + y*rkVector.y + z*rkVector.z;
+}
+
+//----------------------------------------------------------------------------
+inline Vector3 Vector3::cross (const Vector3& rkVector) const {
+    return Vector3(y*rkVector.z - z*rkVector.y, z*rkVector.x - x*rkVector.z,
+                   x*rkVector.y - y*rkVector.x);
+}
+
+//----------------------------------------------------------------------------
+inline Vector3 Vector3::unitCross (const Vector3& rkVector) const {
+    Vector3 kCross(y*rkVector.z - z*rkVector.y, z*rkVector.x - x*rkVector.z,
+                   x*rkVector.y - y*rkVector.x);
+    kCross.unitize();
+    return kCross;
+}
+
+//----------------------------------------------------------------------------
+inline Vector3 Vector3::min(const Vector3 &v) const {
+    return Vector3(G3D::min(v.x, x), G3D::min(v.y, y), G3D::min(v.z, z));
+}
+
+//----------------------------------------------------------------------------
+inline Vector3 Vector3::max(const Vector3 &v) const {
+    return Vector3(G3D::max(v.x, x), G3D::max(v.y, y), G3D::max(v.z, z));
+}
+
+//----------------------------------------------------------------------------
+inline bool Vector3::isZero() const {
+    return G3D::fuzzyEq(squaredMagnitude(), 0.0f);
+}
+
+//----------------------------------------------------------------------------
+
+inline bool Vector3::isUnit() const {
+    return G3D::fuzzyEq(squaredMagnitude(), 1.0f);
+}
+
+} // namespace
diff --git a/dep/include/g3dlite/G3D/Vector3int16.h b/dep/include/g3dlite/G3D/Vector3int16.h
new file mode 100644
index 00000000000..9fd344b2da7
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Vector3int16.h
@@ -0,0 +1,55 @@
+/**
+  @file Vector3int16.h
+  
+  @maintainer Morgan McGuire, matrix@brown.edu
+
+  @created 2003-04-07
+  @edited  2003-06-24
+  Copyright 2000-2004, Morgan McGuire.
+  All rights reserved.
+ */
+
+#ifndef VECTOR3INT16_H
+#define VECTOR3INT16_H
+
+#include "G3D/platform.h"
+#include "G3D/g3dmath.h"
+
+namespace G3D {
+
+/**
+ A Vector3 that packs its fields into uint16s.
+ */
+#ifdef G3D_WIN32
+    // Switch to tight alignment
+    #pragma pack(push, 2)
+#endif
+
+class Vector3int16 {
+private:
+    // Hidden operators
+    bool operator<(const Vector3int16&) const;
+    bool operator>(const Vector3int16&) const;
+    bool operator<=(const Vector3int16&) const;
+    bool operator>=(const Vector3int16&) const;
+
+public:
+    G3D::int16              x;
+    G3D::int16              y;
+    G3D::int16              z;
+
+    Vector3int16() : x(0), y(0), z(0) {}
+    Vector3int16(G3D::int16 _x, G3D::int16 _y, G3D::int16 _z) : x(_x), y(_y), z(_z) {}
+    Vector3int16(const class Vector3& v);
+}
+#if defined(G3D_LINUX) || defined(G3D_OSX)
+    __attribute((aligned(1)))
+#endif
+;
+
+#ifdef G3D_WIN32
+    #pragma pack(pop)
+#endif
+
+}
+#endif
diff --git a/dep/include/g3dlite/G3D/Vector4.h b/dep/include/g3dlite/G3D/Vector4.h
new file mode 100644
index 00000000000..b86bf95b003
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Vector4.h
@@ -0,0 +1,524 @@
+/**
+  @file Vector4.h
+ 
+  Homogeneous vector class.
+ 
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+ 
+  @created 2002-07-09
+  @edited  2005-03-28
+
+  Copyright 2000-2006, Morgan McGuire.
+  All rights reserved.
+ */
+
+#ifndef G3D_VECTOR4_H
+#define G3D_VECTOR4_H
+
+#include "G3D/platform.h"
+#include "G3D/g3dmath.h"
+#include "G3D/Vector3.h"
+#include "G3D/Vector2.h"
+#include <string>
+
+namespace G3D {
+
+class Vector2;    
+class Vector3;
+class Vector4;
+
+/**
+ Do not subclass-- this implementation makes assumptions about the
+ memory layout.
+ */
+class Vector4 {
+private:
+    // Hidden operators
+    bool operator<(const Vector4&) const;
+    bool operator>(const Vector4&) const;
+    bool operator<=(const Vector4&) const;
+    bool operator>=(const Vector4&) const;
+
+public:
+    // construction
+    Vector4();
+    Vector4(float fX, float fY, float fZ, float fW);
+    Vector4(float afCoordinate[4]);
+    Vector4(const Vector4& rkVector);
+    Vector4(const class Color4& c);
+    Vector4(const Vector3& rkVector, float fW);
+    Vector4(const Vector2& v1, const Vector2& v2);
+    Vector4(const Vector2& v1, float fz, float fw);
+
+    // coordinates
+    float x, y, z, w;
+
+    // access vector V as V[0] = V.x, V[1] = V.y, V[2] = V.z, etc.
+    //
+    // WARNING.  These member functions rely on
+    // (1) Vector4 not having virtual functions
+    // (2) the data packed in a 4*sizeof(float) memory block
+    float& operator[] (int i);
+    const float& operator[] (int i) const;
+    operator float* ();
+    operator const float* () const;
+
+    // assignment and comparison
+    Vector4& operator= (const Vector4& rkVector);
+    bool operator== (const Vector4& rkVector) const;
+    bool operator!= (const Vector4& rkVector) const;
+
+    inline void set(float _x, float _y, float _z, float _w) {
+        x = _x;
+        y = _y;
+        z = _z;
+        w = _w;
+    }
+
+    inline void set(const Vector3& v, float _w) {
+        x = v.x;
+        y = v.y;
+        z = v.z;
+        w = _w;
+    }
+
+    inline void set(const Vector2& v, float _z, float _w) {
+        x = v.x;
+        y = v.y;
+        z = _z;
+        w = _w;
+    }
+
+    unsigned int hashCode() const;
+    bool fuzzyEq(const Vector4& other) const;
+    bool fuzzyNe(const Vector4& other) const;
+
+    inline static const Vector4& inf()      { static Vector4 v((float)G3D::inf(), (float)G3D::inf(), (float)G3D::inf(), (float)G3D::inf()); return v; }
+    inline static const Vector4& nan()      { static Vector4 v((float)G3D::nan(), (float)G3D::nan(), (float)G3D::nan(), (float)G3D::nan()); return v; }
+
+    /** sqrt(this->dot(*this)) */
+    float length() const;
+    float squaredLength() const;
+
+    inline float sum() const {
+        return x + y + z + w;
+    }
+
+    /** Returns true if this vector has finite length */
+    bool isFinite() const;
+
+    /** Returns true if this vector has length == 0 */
+    bool isZero() const;
+
+    /** Returns true if this vector has length == 1 */
+    bool isUnit() const;
+
+    // arithmetic operations
+    Vector4 operator+ (const Vector4& rkVector) const;
+    Vector4 operator- (const Vector4& rkVector) const;
+    
+    inline Vector4 operator*(const Vector4& rkVector) const {
+        return Vector4(x * rkVector.x, y * rkVector.y, z * rkVector.z, w * rkVector.w);
+    }
+
+    inline Vector4 operator/(const Vector4& rkVector) const {
+        return Vector4(x / rkVector.x, y / rkVector.y, z / rkVector.z, w / rkVector.w);
+    }
+
+    Vector4 operator* (float fScalar) const;
+    Vector4 operator/ (float fScalar) const;
+    Vector4 operator- () const;
+    friend Vector4 operator* (float, const Vector4& rkVector);
+
+    // arithmetic updates
+    Vector4& operator+= (const Vector4& rkVector);
+    Vector4& operator-= (const Vector4& rkVector);
+    Vector4& operator*= (float fScalar);
+    Vector4& operator/= (float fScalar);
+
+    inline Vector4 clamp(const Vector4& low, const Vector4& high) const {
+        return Vector4(
+            G3D::clamp(x, low.x, high.x),
+            G3D::clamp(y, low.y, high.y),
+            G3D::clamp(z, low.z, high.z),
+            G3D::clamp(w, low.w, high.w));
+    }
+
+    inline Vector4 clamp(float low, float high) const {
+        return Vector4(
+            G3D::clamp(x, low, high),
+            G3D::clamp(y, low, high),
+            G3D::clamp(z, low, high),
+            G3D::clamp(w, low, high));
+    }
+
+    float dot (const Vector4& rkVector) const;
+
+    Vector4 min(const Vector4& v) const;
+    Vector4 max(const Vector4& v) const;
+
+    std::string toString() const;
+
+    /**
+     Linear interpolation
+     */
+    Vector4 lerp(const Vector4& v, float alpha) const;
+
+    // 2-char swizzles
+
+    Vector2 xx() const;
+    Vector2 yx() const;
+    Vector2 zx() const;
+    Vector2 wx() const;
+    Vector2 xy() const;
+    Vector2 yy() const;
+    Vector2 zy() const;
+    Vector2 wy() const;
+    Vector2 xz() const;
+    Vector2 yz() const;
+    Vector2 zz() const;
+    Vector2 wz() const;
+    Vector2 xw() const;
+    Vector2 yw() const;
+    Vector2 zw() const;
+    Vector2 ww() const;
+
+    // 3-char swizzles
+
+    Vector3 xxx() const;
+    Vector3 yxx() const;
+    Vector3 zxx() const;
+    Vector3 wxx() const;
+    Vector3 xyx() const;
+    Vector3 yyx() const;
+    Vector3 zyx() const;
+    Vector3 wyx() const;
+    Vector3 xzx() const;
+    Vector3 yzx() const;
+    Vector3 zzx() const;
+    Vector3 wzx() const;
+    Vector3 xwx() const;
+    Vector3 ywx() const;
+    Vector3 zwx() const;
+    Vector3 wwx() const;
+    Vector3 xxy() const;
+    Vector3 yxy() const;
+    Vector3 zxy() const;
+    Vector3 wxy() const;
+    Vector3 xyy() const;
+    Vector3 yyy() const;
+    Vector3 zyy() const;
+    Vector3 wyy() const;
+    Vector3 xzy() const;
+    Vector3 yzy() const;
+    Vector3 zzy() const;
+    Vector3 wzy() const;
+    Vector3 xwy() const;
+    Vector3 ywy() const;
+    Vector3 zwy() const;
+    Vector3 wwy() const;
+    Vector3 xxz() const;
+    Vector3 yxz() const;
+    Vector3 zxz() const;
+    Vector3 wxz() const;
+    Vector3 xyz() const;
+    Vector3 yyz() const;
+    Vector3 zyz() const;
+    Vector3 wyz() const;
+    Vector3 xzz() const;
+    Vector3 yzz() const;
+    Vector3 zzz() const;
+    Vector3 wzz() const;
+    Vector3 xwz() const;
+    Vector3 ywz() const;
+    Vector3 zwz() const;
+    Vector3 wwz() const;
+    Vector3 xxw() const;
+    Vector3 yxw() const;
+    Vector3 zxw() const;
+    Vector3 wxw() const;
+    Vector3 xyw() const;
+    Vector3 yyw() const;
+    Vector3 zyw() const;
+    Vector3 wyw() const;
+    Vector3 xzw() const;
+    Vector3 yzw() const;
+    Vector3 zzw() const;
+    Vector3 wzw() const;
+    Vector3 xww() const;
+    Vector3 yww() const;
+    Vector3 zww() const;
+    Vector3 www() const;
+
+    // 4-char swizzles
+
+    Vector4 xxxx() const;
+    Vector4 yxxx() const;
+    Vector4 zxxx() const;
+    Vector4 wxxx() const;
+    Vector4 xyxx() const;
+    Vector4 yyxx() const;
+    Vector4 zyxx() const;
+    Vector4 wyxx() const;
+    Vector4 xzxx() const;
+    Vector4 yzxx() const;
+    Vector4 zzxx() const;
+    Vector4 wzxx() const;
+    Vector4 xwxx() const;
+    Vector4 ywxx() const;
+    Vector4 zwxx() const;
+    Vector4 wwxx() const;
+    Vector4 xxyx() const;
+    Vector4 yxyx() const;
+    Vector4 zxyx() const;
+    Vector4 wxyx() const;
+    Vector4 xyyx() const;
+    Vector4 yyyx() const;
+    Vector4 zyyx() const;
+    Vector4 wyyx() const;
+    Vector4 xzyx() const;
+    Vector4 yzyx() const;
+    Vector4 zzyx() const;
+    Vector4 wzyx() const;
+    Vector4 xwyx() const;
+    Vector4 ywyx() const;
+    Vector4 zwyx() const;
+    Vector4 wwyx() const;
+    Vector4 xxzx() const;
+    Vector4 yxzx() const;
+    Vector4 zxzx() const;
+    Vector4 wxzx() const;
+    Vector4 xyzx() const;
+    Vector4 yyzx() const;
+    Vector4 zyzx() const;
+    Vector4 wyzx() const;
+    Vector4 xzzx() const;
+    Vector4 yzzx() const;
+    Vector4 zzzx() const;
+    Vector4 wzzx() const;
+    Vector4 xwzx() const;
+    Vector4 ywzx() const;
+    Vector4 zwzx() const;
+    Vector4 wwzx() const;
+    Vector4 xxwx() const;
+    Vector4 yxwx() const;
+    Vector4 zxwx() const;
+    Vector4 wxwx() const;
+    Vector4 xywx() const;
+    Vector4 yywx() const;
+    Vector4 zywx() const;
+    Vector4 wywx() const;
+    Vector4 xzwx() const;
+    Vector4 yzwx() const;
+    Vector4 zzwx() const;
+    Vector4 wzwx() const;
+    Vector4 xwwx() const;
+    Vector4 ywwx() const;
+    Vector4 zwwx() const;
+    Vector4 wwwx() const;
+    Vector4 xxxy() const;
+    Vector4 yxxy() const;
+    Vector4 zxxy() const;
+    Vector4 wxxy() const;
+    Vector4 xyxy() const;
+    Vector4 yyxy() const;
+    Vector4 zyxy() const;
+    Vector4 wyxy() const;
+    Vector4 xzxy() const;
+    Vector4 yzxy() const;
+    Vector4 zzxy() const;
+    Vector4 wzxy() const;
+    Vector4 xwxy() const;
+    Vector4 ywxy() const;
+    Vector4 zwxy() const;
+    Vector4 wwxy() const;
+    Vector4 xxyy() const;
+    Vector4 yxyy() const;
+    Vector4 zxyy() const;
+    Vector4 wxyy() const;
+    Vector4 xyyy() const;
+    Vector4 yyyy() const;
+    Vector4 zyyy() const;
+    Vector4 wyyy() const;
+    Vector4 xzyy() const;
+    Vector4 yzyy() const;
+    Vector4 zzyy() const;
+    Vector4 wzyy() const;
+    Vector4 xwyy() const;
+    Vector4 ywyy() const;
+    Vector4 zwyy() const;
+    Vector4 wwyy() const;
+    Vector4 xxzy() const;
+    Vector4 yxzy() const;
+    Vector4 zxzy() const;
+    Vector4 wxzy() const;
+    Vector4 xyzy() const;
+    Vector4 yyzy() const;
+    Vector4 zyzy() const;
+    Vector4 wyzy() const;
+    Vector4 xzzy() const;
+    Vector4 yzzy() const;
+    Vector4 zzzy() const;
+    Vector4 wzzy() const;
+    Vector4 xwzy() const;
+    Vector4 ywzy() const;
+    Vector4 zwzy() const;
+    Vector4 wwzy() const;
+    Vector4 xxwy() const;
+    Vector4 yxwy() const;
+    Vector4 zxwy() const;
+    Vector4 wxwy() const;
+    Vector4 xywy() const;
+    Vector4 yywy() const;
+    Vector4 zywy() const;
+    Vector4 wywy() const;
+    Vector4 xzwy() const;
+    Vector4 yzwy() const;
+    Vector4 zzwy() const;
+    Vector4 wzwy() const;
+    Vector4 xwwy() const;
+    Vector4 ywwy() const;
+    Vector4 zwwy() const;
+    Vector4 wwwy() const;
+    Vector4 xxxz() const;
+    Vector4 yxxz() const;
+    Vector4 zxxz() const;
+    Vector4 wxxz() const;
+    Vector4 xyxz() const;
+    Vector4 yyxz() const;
+    Vector4 zyxz() const;
+    Vector4 wyxz() const;
+    Vector4 xzxz() const;
+    Vector4 yzxz() const;
+    Vector4 zzxz() const;
+    Vector4 wzxz() const;
+    Vector4 xwxz() const;
+    Vector4 ywxz() const;
+    Vector4 zwxz() const;
+    Vector4 wwxz() const;
+    Vector4 xxyz() const;
+    Vector4 yxyz() const;
+    Vector4 zxyz() const;
+    Vector4 wxyz() const;
+    Vector4 xyyz() const;
+    Vector4 yyyz() const;
+    Vector4 zyyz() const;
+    Vector4 wyyz() const;
+    Vector4 xzyz() const;
+    Vector4 yzyz() const;
+    Vector4 zzyz() const;
+    Vector4 wzyz() const;
+    Vector4 xwyz() const;
+    Vector4 ywyz() const;
+    Vector4 zwyz() const;
+    Vector4 wwyz() const;
+    Vector4 xxzz() const;
+    Vector4 yxzz() const;
+    Vector4 zxzz() const;
+    Vector4 wxzz() const;
+    Vector4 xyzz() const;
+    Vector4 yyzz() const;
+    Vector4 zyzz() const;
+    Vector4 wyzz() const;
+    Vector4 xzzz() const;
+    Vector4 yzzz() const;
+    Vector4 zzzz() const;
+    Vector4 wzzz() const;
+    Vector4 xwzz() const;
+    Vector4 ywzz() const;
+    Vector4 zwzz() const;
+    Vector4 wwzz() const;
+    Vector4 xxwz() const;
+    Vector4 yxwz() const;
+    Vector4 zxwz() const;
+    Vector4 wxwz() const;
+    Vector4 xywz() const;
+    Vector4 yywz() const;
+    Vector4 zywz() const;
+    Vector4 wywz() const;
+    Vector4 xzwz() const;
+    Vector4 yzwz() const;
+    Vector4 zzwz() const;
+    Vector4 wzwz() const;
+    Vector4 xwwz() const;
+    Vector4 ywwz() const;
+    Vector4 zwwz() const;
+    Vector4 wwwz() const;
+    Vector4 xxxw() const;
+    Vector4 yxxw() const;
+    Vector4 zxxw() const;
+    Vector4 wxxw() const;
+    Vector4 xyxw() const;
+    Vector4 yyxw() const;
+    Vector4 zyxw() const;
+    Vector4 wyxw() const;
+    Vector4 xzxw() const;
+    Vector4 yzxw() const;
+    Vector4 zzxw() const;
+    Vector4 wzxw() const;
+    Vector4 xwxw() const;
+    Vector4 ywxw() const;
+    Vector4 zwxw() const;
+    Vector4 wwxw() const;
+    Vector4 xxyw() const;
+    Vector4 yxyw() const;
+    Vector4 zxyw() const;
+    Vector4 wxyw() const;
+    Vector4 xyyw() const;
+    Vector4 yyyw() const;
+    Vector4 zyyw() const;
+    Vector4 wyyw() const;
+    Vector4 xzyw() const;
+    Vector4 yzyw() const;
+    Vector4 zzyw() const;
+    Vector4 wzyw() const;
+    Vector4 xwyw() const;
+    Vector4 ywyw() const;
+    Vector4 zwyw() const;
+    Vector4 wwyw() const;
+    Vector4 xxzw() const;
+    Vector4 yxzw() const;
+    Vector4 zxzw() const;
+    Vector4 wxzw() const;
+    Vector4 xyzw() const;
+    Vector4 yyzw() const;
+    Vector4 zyzw() const;
+    Vector4 wyzw() const;
+    Vector4 xzzw() const;
+    Vector4 yzzw() const;
+    Vector4 zzzw() const;
+    Vector4 wzzw() const;
+    Vector4 xwzw() const;
+    Vector4 ywzw() const;
+    Vector4 zwzw() const;
+    Vector4 wwzw() const;
+    Vector4 xxww() const;
+    Vector4 yxww() const;
+    Vector4 zxww() const;
+    Vector4 wxww() const;
+    Vector4 xyww() const;
+    Vector4 yyww() const;
+    Vector4 zyww() const;
+    Vector4 wyww() const;
+    Vector4 xzww() const;
+    Vector4 yzww() const;
+    Vector4 zzww() const;
+    Vector4 wzww() const;
+    Vector4 xwww() const;
+    Vector4 ywww() const;
+    Vector4 zwww() const;
+    Vector4 wwww() const;
+
+};
+
+}
+
+inline G3D::Vector4 operator* (float s, const G3D::Vector4& v) {
+    return v * s;
+}
+
+unsigned int hashCode(const G3D::Vector4& v);
+
+#include "Vector4.inl"
+
+#endif
diff --git a/dep/include/g3dlite/G3D/Vector4.inl b/dep/include/g3dlite/G3D/Vector4.inl
new file mode 100644
index 00000000000..48ebf542723
--- /dev/null
+++ b/dep/include/g3dlite/G3D/Vector4.inl
@@ -0,0 +1,193 @@
+/** 
+  @file Vector4.inl
+
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+
+  @created 2002-07-09
+  @edited  2003-02-10
+ */
+
+//----------------------------------------------------------------------------
+
+inline unsigned int hashCode(const G3D::Vector4& v) {
+     return v.hashCode();
+}
+
+namespace G3D {
+
+//----------------------------------------------------------------------------
+inline Vector4::Vector4() {
+    x = y = z = w = 0;
+}
+
+//----------------------------------------------------------------------------
+
+inline Vector4::Vector4 (float fX, float fY, float fZ, float fW) {
+    x = fX;
+    y = fY;
+    z = fZ;
+    w = fW;
+}
+
+//----------------------------------------------------------------------------
+inline Vector4::Vector4 (float afCoordinate[4]) {
+    x = afCoordinate[0];
+    y = afCoordinate[1];
+    z = afCoordinate[2];
+    w = afCoordinate[3];
+}
+
+//----------------------------------------------------------------------------
+inline Vector4::Vector4(const Vector4& rkVector) {
+    x = rkVector.x;
+    y = rkVector.y;
+    z = rkVector.z;
+    w = rkVector.w;
+}
+//----------------------------------------------------------------------------
+inline Vector4::Vector4(const Vector3& rkVector, float fW) {
+    x = rkVector.x;
+    y = rkVector.y;
+    z = rkVector.z;
+    w = fW;
+}
+
+//----------------------------------------------------------------------------
+inline float& Vector4::operator[] (int i) {
+    return ((float*)this)[i];
+}
+
+//----------------------------------------------------------------------------
+inline const float& Vector4::operator[] (int i) const {
+    return ((float*)this)[i];
+}
+
+//----------------------------------------------------------------------------
+inline Vector4::operator float* () {
+    return (float*)this;
+}
+
+inline Vector4::operator const float* () const {
+    return (float*)this;
+}
+
+//----------------------------------------------------------------------------
+inline Vector4& Vector4::operator= (const Vector4& rkVector) {
+    x = rkVector.x;
+    y = rkVector.y;
+    z = rkVector.z;
+    w = rkVector.w;
+    return *this;
+}
+
+//----------------------------------------------------------------------------
+inline bool Vector4::operator== (const Vector4& rkVector) const {
+    return ( (x == rkVector.x) && (y == rkVector.y) && (z == rkVector.z) && (w == rkVector.w));
+}
+
+//----------------------------------------------------------------------------
+inline bool Vector4::operator!= (const Vector4& rkVector) const {
+    return ( x != rkVector.x || y != rkVector.y || z != rkVector.z || w != rkVector.w);
+}
+
+//----------------------------------------------------------------------------
+inline Vector4 Vector4::operator+ (const Vector4& rkVector) const {
+    return Vector4(x + rkVector.x, y + rkVector.y, z + rkVector.z, w + rkVector.w);
+}
+
+//----------------------------------------------------------------------------
+inline Vector4 Vector4::operator- (const Vector4& rkVector) const {
+    return Vector4(x - rkVector.x, y - rkVector.y, z - rkVector.z, w - rkVector.w);
+}
+
+//----------------------------------------------------------------------------
+inline Vector4 Vector4::operator* (float fScalar) const {
+    return Vector4(fScalar*x, fScalar*y, fScalar*z, fScalar*w);
+}
+
+//----------------------------------------------------------------------------
+inline Vector4 Vector4::operator- () const {
+    return Vector4( -x, -y, -z, -w);
+}
+
+//----------------------------------------------------------------------------
+inline Vector4& Vector4::operator+= (const Vector4& rkVector) {
+    x += rkVector.x;
+    y += rkVector.y;
+    z += rkVector.z;
+    w += rkVector.w;
+    return *this;
+}
+
+//----------------------------------------------------------------------------
+inline Vector4& Vector4::operator-= (const Vector4& rkVector) {
+    x -= rkVector.x;
+    y -= rkVector.y;
+    z -= rkVector.z;
+    w -= rkVector.w;
+    return *this;
+}
+
+//----------------------------------------------------------------------------
+
+inline Vector4 Vector4::lerp(const Vector4& v, float alpha) const {
+    return (*this) + (v - *this) * alpha; 
+}
+
+
+//----------------------------------------------------------------------------
+inline Vector4& Vector4::operator*= (float fScalar) {
+    x *= fScalar;
+    y *= fScalar;
+    z *= fScalar;
+    w *= fScalar;
+    return *this;
+}
+
+
+//----------------------------------------------------------------------------
+inline float Vector4::dot(const Vector4& rkVector) const {
+    return x*rkVector.x + y*rkVector.y + z*rkVector.z + w*rkVector.w;
+}
+
+//----------------------------------------------------------------------------
+inline Vector4 Vector4::min(const Vector4 &v) const {
+    return Vector4(G3D::min(v.x, x), G3D::min(v.y, y), G3D::min(v.z, z), G3D::min(v.w, w));
+}
+
+//----------------------------------------------------------------------------
+inline Vector4 Vector4::max(const Vector4 &v) const {
+    return Vector4(G3D::max(v.x, x), G3D::max(v.y, y), G3D::max(v.z, z), G3D::max(v.w, w));
+}
+
+//----------------------------------------------------------------------------
+inline bool Vector4::isZero() const {
+    return (x == 0.0f) && (y == 0.0f) && (z == 0.0f) && (w == 0.0f);
+}
+
+//----------------------------------------------------------------------------
+
+inline bool Vector4::isFinite() const {
+    return G3D::isFinite(x) && G3D::isFinite(y) && G3D::isFinite(z) && G3D::isFinite(w);
+}
+
+//----------------------------------------------------------------------------
+
+inline bool Vector4::isUnit() const {
+    return squaredLength() == 1.0;
+}
+
+//----------------------------------------------------------------------------
+
+inline float Vector4::length() const {
+    return sqrtf(squaredLength());
+}
+
+//----------------------------------------------------------------------------
+
+inline float Vector4::squaredLength() const {
+    return x * x + y * y + z * z + w * w;
+}
+
+}
+
diff --git a/dep/include/g3dlite/G3D/debug.h b/dep/include/g3dlite/G3D/debug.h
new file mode 100644
index 00000000000..0216d1daf11
--- /dev/null
+++ b/dep/include/g3dlite/G3D/debug.h
@@ -0,0 +1,11 @@
+
+#ifndef G3D_LITE_DEBUG_H
+#define G3D_LITE_DEBUG_H
+
+#define debugStatement(x)
+#define debugAssert(x)
+#define debugAssertM(x, y)
+#define alwaysAssertM(x, y)
+
+#endif
+
diff --git a/dep/include/g3dlite/G3D/format.h b/dep/include/g3dlite/G3D/format.h
new file mode 100644
index 00000000000..b031300ab1a
--- /dev/null
+++ b/dep/include/g3dlite/G3D/format.h
@@ -0,0 +1,57 @@
+/**
+ @file format.h
+ 
+ @maintainer Morgan McGuire, matrix@graphics3d.com
+ 
+ @author  2000-09-09
+ @edited  2005-11-03
+
+ Copyright 2000-2005, Morgan McGuire.
+ All rights reserved.
+ */
+
+#ifndef G3D_FORMAT_H
+#define G3D_FORMAT_H
+
+#include "G3D/platform.h"
+#include <string>
+#include <stdio.h>
+#include <cstdarg>
+#include <assert.h>
+#ifndef G3D_WIN32
+    // Don't include varargs.h for some random
+    // gcc reason
+    //#include <varargs.h>
+    #include <stdarg.h>
+#endif
+
+#ifndef _MSC_VER
+    #ifndef __cdecl
+        #define __cdecl __attribute__((cdecl))
+    #endif
+#endif
+
+namespace G3D {
+
+/**
+  Produces a string from arguments of the style of printf.  This avoids
+  problems with buffer overflows when using sprintf and makes it easy
+  to use the result functionally.  This function is fast when the resulting
+  string is under 160 characters (not including terminator) and slower
+  when the string is longer.
+ */
+std::string format(
+    const char*                 fmt
+    ...) G3D_CHECK_PRINTF_ARGS;
+
+/**
+  Like format, but can be called with the argument list from a ... function.
+ */
+std::string vformat(
+    const char*                 fmt,
+    va_list                     argPtr) G3D_CHECK_VPRINTF_ARGS;
+
+
+}; // namespace
+
+#endif
diff --git a/dep/include/g3dlite/G3D/g3dmath.h b/dep/include/g3dlite/G3D/g3dmath.h
new file mode 100644
index 00000000000..1f3eaaae57c
--- /dev/null
+++ b/dep/include/g3dlite/G3D/g3dmath.h
@@ -0,0 +1,525 @@
+/**
+ @file g3dmath.h
+ 
+ Math util class.
+ 
+ @maintainer Morgan McGuire, matrix@graphics3d.com
+ @cite highestBit by Jukka Liimatta
+ 
+ @created 2001-06-02
+ @edited  2006-01-16
+
+ Copyright 2000-2006, Morgan McGuire.
+ All rights reserved.
+ */
+
+#ifndef G3DMATH_H
+#define G3DMATH_H
+
+#ifdef _MSC_VER
+// Disable conditional expression is constant, which occurs incorrectly on inlined functions
+#   pragma warning (push)
+#   pragma warning (disable : 4127)
+// disable: "C++ exception handler used"
+#   pragma warning (disable : 4530)
+#endif
+
+#include "G3D/platform.h"
+#include <ctype.h>
+#include <string>
+#include <float.h>
+#include <limits>
+
+/*These defines enable functionality introduced with the 1999 ISO C
+**standard. They must be defined before the inclusion of math.h to
+**engage them. If optimisation is enabled, these functions will be 
+**inlined. With optimisation switched off, you have to link in the
+**maths library using -lm.
+*/
+
+#define _ISOC9X_SOURCE1
+#define _ISOC99_SOURCE1
+#define __USE_ISOC9X1
+#define __USE_ISOC991
+
+#include <math.h>
+
+#include "G3D/debug.h"
+
+#undef min
+#undef max
+
+namespace G3D {
+
+#if defined(_MSC_VER)
+    
+#if !defined(_WIN64)
+
+/**
+   Win32 implementation of the C99 fast rounding routines.
+   
+   @cite routines are
+   Copyright (C) 2001 Erik de Castro Lopo <erikd AT mega-nerd DOT com>
+   
+   Permission to use, copy, modify, distribute, and sell this file for any 
+   purpose is hereby granted without fee, provided that the above copyright 
+   and this permission notice appear in all copies.  No representations are
+   made about the suitability of this software for any purpose.  It is 
+   provided "as is" without express or implied warranty.
+*/
+
+__inline long int lrint (double flt) {
+    int intgr;
+
+    _asm {
+        fld flt
+        fistp intgr
+    };
+
+    return intgr;
+}
+
+__inline long int lrintf(float flt) {
+    int intgr;
+
+    _asm {
+        fld flt
+        fistp intgr
+    };
+
+    return intgr;
+}
+
+#else
+
+    __inline long int lrint (double flt) {
+        return (long int)floor(flt+0.5f);
+    }
+
+    __inline long int lrintf(float flt) {
+        return (long int)floorf(flt+0.5f);
+    }
+
+
+#endif
+
+#endif
+
+
+const double fuzzyEpsilon = 0.00001;
+
+/** Returns a reference to a static double.
+    This value should not be tested against directly, instead
+    G3D::isNan() and G3D::isFinite() will return reliable results. */
+inline const double& inf() {
+
+// We already have <limits> included but
+// not using it in older gcc for safe compilations
+#if (__GNUC__ == 2)    
+    static const double i = 1.0/sin(0.0);
+#else
+    // double is a standard type and should have infinity
+    static const double i = std::numeric_limits<double>::infinity();
+#endif
+	return i;
+}
+
+/** Returns a reference to a static double.
+    This value should not be tested against directly, instead
+    G3D::isNan() and G3D::isFinite() will return reliable results. */
+inline const double& nan() {
+
+// We already have <limits> included but
+// not using it in older gcc for safe compilations
+#if (__GNUC__ == 2)
+    static const double n = 0.0/sin(0.0);
+#else
+    // double is a standard type and should have quiet NaN
+    static const double n = std::numeric_limits<double>::quiet_NaN();
+#endif
+	return n;
+}
+
+/** Returns a reference to a static double. Use instead of G3D_PI. */
+inline const double& pi() {
+    static const double p = 3.1415926535898;
+    return p;
+}
+
+/** Returns a reference to a static double. Use instead of G3D_HALF_PI. */
+inline const double& halfPi() {
+    static const double p = 1.5707963267949;
+    return p;
+}
+
+/** Returns a reference to a static double. Use instead of G3D_TWO_PI. */
+inline const double& twoPi() {
+    static const double p = 6.283185;
+    return p;
+}
+
+/** @def G3D_PI
+    @deprecated Use G3D::pi() instead. */
+#define G3D_PI      (3.1415926535898)
+/** @def G3D_HALF_PI
+    @deprecated Use G3D::halfPi() instead. */
+#define G3D_HALF_PI (1.5707963267949)
+/** @def G3D_TWO_PI
+    @deprecated Use G3D::twoPi() instead. */
+#define G3D_TWO_PI  (6.283185)
+
+typedef signed char		int8;
+typedef unsigned char	uint8;
+typedef short			int16;
+typedef unsigned short	uint16;
+typedef int				int32;
+typedef unsigned int	uint32;
+
+#ifdef _MSC_EXTENSIONS
+    typedef __int64			   int64;
+    typedef unsigned __int64   uint64;
+#else
+    typedef long long		   int64;
+    typedef unsigned long long uint64;
+#endif
+typedef unsigned int uint;
+
+typedef float			float32;
+typedef double			float64;
+
+int iAbs(int iValue);
+int iCeil(double fValue);
+
+/**
+ Clamps the value to the range [low, hi] (inclusive)
+ */
+int iClamp(int val, int low, int hi);
+double clamp(double val, double low, double hi);
+float clamp(float val, float low, float hi);
+
+/**
+ Returns a + (b - a) * f;
+ */
+inline double lerp(double a, double b, double f) {
+    return a + (b - a) * f;
+}
+
+inline float lerp(float a, float b, float f) {
+    return a + (b - a) * f;
+}
+
+/**
+ Wraps the value to the range [0, hi) (exclusive
+ on the high end).  This is like the clock arithmetic
+ produced by % (modulo) except the result is guaranteed
+ to be positive.
+ */
+int iWrap(int val, int hi);
+
+int iFloor(double fValue);
+
+int iSign(int iValue);
+int iSign(double fValue);
+
+inline int iSign(float f) {
+    return iSign((double)f);
+}
+
+
+/** 
+    Fast round to integer using the lrint routine.
+    Typically 6x faster than casting to integer.
+ */
+inline int iRound(double fValue) {
+    return lrint(fValue);
+}
+
+/** 
+    Fast round to integer using the lrint routine.
+    Typically 6x faster than casting to integer.
+ */
+inline int iRound(float f) {
+    return lrintf(f);
+}
+
+/**
+ Returns a random number uniformly at random between low and hi
+ (inclusive).
+ */
+int iRandom(int low, int hi);
+
+double abs (double fValue);
+double aCos (double fValue);
+double aSin (double fValue);
+double aTan (double fValue);
+double aTan2 (double fY, double fX);
+double sign (double fValue);
+double square (double fValue);
+
+/**
+ Returns true if the argument is a finite real number.
+ */
+bool isFinite(double x);
+
+/**
+ Returns true if the argument is NaN (not a number).
+ You can't use x == nan to test this because all
+ comparisons against nan return false.
+ */
+bool isNaN(double x);
+
+/**
+ Computes x % 3.
+ */
+int iMod3(int x);
+
+/** 
+ [0, 1]
+ @deprecated use uniformRandom()
+ */
+double unitRandom ();
+
+/**
+ Uniform random number between low and hi, inclusive.
+ @deprecated use uniformRandom()
+ */
+double random(double low, double hi);
+
+/**
+ [-1, 1]
+ @deprecated use uniformRandom()
+ */
+double symmetricRandom ();
+
+/**
+ Uniform random number between low and hi, inclusive. [low, hi]
+ */
+float uniformRandom(float low = 0.0f, float hi = 1.0f);
+
+/**
+ Normally distributed random number. 
+ */
+float gaussRandom(float mean = 0.0f, float stdev = 1.0f);
+
+#if defined(_MSC_VER) && (_MSC_VER <= 1200)
+
+    /** VC6 lacks std::min and std::max */
+    inline double min(double x, double y) {
+        return std::_cpp_min(x, y);
+    }
+
+    /** VC6 lacks std::min and std::max */
+    inline float min(float x, float y) {
+        return std::_cpp_min(x, y);
+    }
+
+    /** VC6 lacks std::min and std::max */
+    inline int min(int x, int y) {
+        return std::_cpp_min(x, y);
+    }
+
+    /** VC6 lacks std::min and std::max */
+    inline double max(double x, double y) {
+        return std::_cpp_max(x, y);
+    }
+
+    /** VC6 lacks std::min and std::max */
+    inline float max(float x, float y) {
+        return std::_cpp_max(x, y);
+    }
+
+    /** VC6 lacks std::min and std::max */
+    inline int max(int x, int y) {
+        return std::_cpp_max(x, y);
+    }
+
+#else
+    template <class T>
+    inline T min(const T& x, const T& y) {
+        return std::min<T>(x, y);
+    }
+
+    template <class T>
+    inline T max(const T& x, const T& y) {
+        return std::max<T>(x, y);
+    }
+
+#endif
+
+int iMin(int x, int y);
+int iMax(int x, int y);
+
+double square(double x);
+double sumSquares(double x, double y);
+double sumSquares(double x, double y, double z);
+double distance(double x, double y);
+double distance(double x, double y, double z);
+
+/**
+  Returnes the 0-based index of the highest 1 bit from
+  the left.  -1 means the number was 0.
+
+  @cite Based on code by jukka@liimatta.org
+ */ 
+int highestBit(uint32 x);
+
+/**
+ Note that fuzzyEq(a, b) && fuzzyEq(b, c) does not imply
+ fuzzyEq(a, c), although that will be the case on some
+ occasions.
+ */
+bool fuzzyEq(double a, double b);
+
+/** True if a is definitely not equal to b.  
+    Guaranteed false if a == b. 
+    Possibly false when a != b.*/
+bool fuzzyNe(double a, double b);
+
+/** Is a strictly greater than b? (Guaranteed false if a <= b).
+    (Possibly false if a > b) */
+bool fuzzyGt(double a, double b);
+
+/** Is a near or greater than b? */
+bool fuzzyGe(double a, double b);
+
+/** Is a strictly less than b? (Guaranteed false if a >= b)*/
+bool fuzzyLt(double a, double b);
+
+/** Is a near or less than b? */
+bool fuzzyLe(double a, double b);
+
+/**
+ Computes 1 / sqrt(x).
+ */
+inline float rsq(float x) {
+    return 1.0f / sqrtf(x);
+}
+
+/**
+ Uses SSE to implement rsq.
+ @cite Nick nicolas@capens.net
+ */
+inline float SSErsq(float x) {
+
+    #if defined(SSE) && defined(G3D_WIN32) && !defined(_WIN64)
+        __asm {
+           movss xmm0, x
+           rsqrtss xmm0, xmm0
+           movss x, xmm0
+        }
+        return x;
+    #else
+        return 1.0f / sqrt(x);
+    #endif
+}
+
+/**
+ Return the next power of 2 higher than the input
+ If the input is already a power of 2, the output will be the same 
+ as the input.
+ */
+int ceilPow2(unsigned int in);
+
+/**
+ * True if num is a power of two.
+ */
+bool isPow2(int num);
+
+bool isOdd(int num);
+bool isEven(int num);
+
+double toRadians(double deg);
+double toDegrees(double rad);
+
+/**
+ Returns true if x is not exactly equal to 0.0f.
+ */
+inline bool any(float x) {
+    return x != 0;
+}
+
+/**
+ Returns true if x is not exactly equal to 0.0f.
+ */
+inline bool all(float x) {
+    return x != 0;
+}
+
+/**
+ v / v (for DirectX/Cg support)
+ */
+inline float normalize(float v) {
+    return v / v;
+}
+
+/**
+ a * b (for DirectX/Cg support)
+ */
+inline float dot(float a, float b) {
+    return a * b;
+}
+
+
+/**
+ a * b (for DirectX/Cg support)
+ */
+inline float mul(float a, float b) {
+    return a * b;
+}
+
+/**
+ 2^x
+ */
+inline double exp2(double x) {
+    return pow(2.0, x);
+}
+
+inline double rsqrt(double x) {
+    return 1.0 / sqrt(x);
+}
+
+
+/**
+ sin(x)/x
+ */
+inline double sinc(double x) {
+    double r = sin(x) / x;
+
+    if (isNaN(r)) {
+        return 1.0;
+    } else {
+        return r;
+    }
+}
+
+/**
+ Computes a floating point modulo; the result is t wrapped to the range [lo, hi).
+ */
+inline double wrap(double t, double lo, double hi) {
+    if ((t >= lo) && (t < hi)) {
+        return t;
+    }
+
+    debugAssert(hi > lo);
+
+    double interval = hi - lo;
+
+    return t - interval * iFloor((t - lo) / interval);
+
+}
+
+inline double wrap(double t, double hi) {
+    return wrap(t, 0, hi);
+}
+
+
+} // namespace
+
+#ifdef _MSC_VER
+#   pragma warning (pop)
+#endif
+
+#include "g3dmath.inl"
+
+#endif
+
diff --git a/dep/include/g3dlite/G3D/g3dmath.inl b/dep/include/g3dlite/G3D/g3dmath.inl
new file mode 100644
index 00000000000..1068ed5bf81
--- /dev/null
+++ b/dep/include/g3dlite/G3D/g3dmath.inl
@@ -0,0 +1,289 @@
+/**
+ @file g3dmath.inl
+ 
+  @maintainer Morgan McGuire, matrix@graphics3d.com
+ 
+ @created 2001-06-02
+ @edited  2006-01-14
+ */
+
+#include <stdlib.h>
+
+#ifdef _MSC_VER
+// Disable conditional expression is constant, which occurs incorrectly on inlined functions
+#   pragma warning (push)
+#   pragma warning( disable : 4127 )
+#endif
+
+namespace G3D {
+
+inline bool isNaN(double x) {
+    bool b1  = (x < 0.0);
+    bool b2  = (x >= 0.0);
+    bool b3  = !(b1 || b2);
+    return b3;
+}
+
+inline bool isFinite(double x) {
+    return ! isNaN(x) && (x < G3D::inf()) && (x > -G3D::inf());
+}
+
+//----------------------------------------------------------------------------
+inline int iAbs (int iValue) {
+    return ( iValue >= 0 ? iValue : -iValue );
+}
+
+//----------------------------------------------------------------------------
+inline int iCeil (double fValue) {
+    return int(::ceil(fValue));
+}
+
+//----------------------------------------------------------------------------
+
+inline int iClamp(int val, int low, int hi) {
+    debugAssert(low <= hi);
+	if (val <= low) {
+		return low;
+	} else if (val >= hi) {
+		return hi;
+	} else {
+		return val;
+	}
+}
+
+//----------------------------------------------------------------------------
+
+inline double clamp(double val, double low, double hi) {
+    debugAssert(low <= hi);
+	if (val <= low) {
+		return low;
+	} else if (val >= hi) {
+		return hi;
+	} else {
+		return val;
+	}
+}
+
+inline float clamp(float val, float low, float hi) {
+    debugAssert(low <= hi);
+	if (val <= low) {
+		return low;
+	} else if (val >= hi) {
+		return hi;
+	} else {
+		return val;
+	}
+}
+//----------------------------------------------------------------------------
+
+inline int iWrap(int val, int hi) {
+	if (val < 0) {
+		return ((val % hi) + hi) % hi;
+	} else {
+		return val % hi;
+	}
+}
+
+//----------------------------------------------------------------------------
+inline int iFloor (double fValue) {
+    return int(::floor(fValue));
+}
+
+//----------------------------------------------------------------------------
+inline int iSign (int iValue) {
+    return ( iValue > 0 ? + 1 : ( iValue < 0 ? -1 : 0 ) );
+}
+
+inline int iSign (double fValue) {
+    return ( fValue > 0.0 ? + 1 : ( fValue < 0.0 ? -1 : 0 ) );
+}
+
+//----------------------------------------------------------------------------
+inline double abs (double fValue) {
+    return double(::fabs(fValue));
+}
+
+//----------------------------------------------------------------------------
+inline double aCos (double fValue) {
+    if ( -1.0 < fValue ) {
+        if ( fValue < 1.0 )
+            return double(::acos(fValue));
+        else
+            return 0.0;
+    } else {
+        return G3D_PI;
+    }
+}
+
+//----------------------------------------------------------------------------
+inline double aSin (double fValue) {
+    if ( -1.0 < fValue ) {
+        if ( fValue < 1.0 ) {
+            return double(::asin(fValue));
+        } else {
+            return -G3D_HALF_PI;
+        }
+    } else {
+        return G3D_HALF_PI;
+    }
+}
+
+//----------------------------------------------------------------------------
+inline double aTan (double fValue) {
+    return double(::atan(fValue));
+}
+
+//----------------------------------------------------------------------------
+inline double aTan2 (double fY, double fX) {
+    return double(::atan2(fY, fX));
+}
+
+//----------------------------------------------------------------------------
+inline double sign (double fValue) {
+    if (fValue > 0.0) {
+        return 1.0;
+	}
+
+    if (fValue < 0.0) {
+        return -1.0;
+	}
+
+    return 0.0;
+}
+
+inline double G3D_DEPRECATED unitRandom () {
+    return double(::rand()) / double(RAND_MAX);
+}
+
+inline float uniformRandom(float low, float hi) {
+    return (hi - low) * float(::rand()) / float(RAND_MAX) + low;
+}
+
+
+//----------------------------------------------------------------------------
+inline double G3D_DEPRECATED symmetricRandom () {
+    return 2.0 * double(::rand()) / double(RAND_MAX) - 1.0;
+}
+
+//----------------------------------------------------------------------------
+inline double square(double x) {
+    return x * x;
+}
+
+//----------------------------------------------------------------------------
+inline double sumSquares(double x, double y) {
+    return x*x + y*y;
+}
+
+//----------------------------------------------------------------------------
+inline double sumSquares(double x, double y, double z) {
+    return x*x + y*y + z*z;
+}
+
+//----------------------------------------------------------------------------
+inline double distance(double x, double y) {
+    return sqrt(sumSquares(x, y));
+}
+
+//----------------------------------------------------------------------------
+inline double distance(double x, double y, double z) {
+    return sqrt(sumSquares(x, y, z));
+}
+
+//----------------------------------------------------------------------------
+
+/** @deprecated use G3D::min */
+inline int iMin(int x, int y) {
+    return (x >= y) ? y : x;
+}
+
+//----------------------------------------------------------------------------
+/** @deprecated use G3D::min */
+inline int iMax(int x, int y) {
+    return (x >= y) ? x : y;
+}
+
+//----------------------------------------------------------------------------
+inline int ceilPow2(unsigned int in) {
+    in -= 1;
+
+    in |= in >> 16;
+    in |= in >> 8;
+    in |= in >> 4;
+    in |= in >> 2;
+    in |= in >> 1;
+
+    return in + 1;
+}
+
+inline bool isPow2(int num) {
+    return ((num & -num) == num);
+}
+
+inline bool isOdd(int num) {
+    return (num & 1) == 1;
+}
+
+inline bool isEven(int num) {
+    return (num & 1) == 0;
+}
+
+inline double toRadians(double deg) {
+    return deg * G3D_PI / 180.0;
+}
+
+inline double toDegrees(double rad) {
+    return rad * 180.0 / G3D_PI;
+}
+
+/**
+ Computes an appropriate epsilon for comparing a and b.
+ */
+inline double eps(double a, double b) {
+    // For a and b to be nearly equal, they must have nearly
+    // the same magnitude.  This means that we can ignore b
+    // since it either has the same magnitude or the comparison
+    // will fail anyway.
+    (void)b;
+    const double aa = abs(a) + 1;
+    if (aa == inf()) {
+        return fuzzyEpsilon;
+    } else {
+        return fuzzyEpsilon * aa;
+    }
+}
+
+inline bool fuzzyEq(double a, double b) {
+    return (a == b) || (abs(a - b) <= eps(a, b));
+}
+
+inline bool fuzzyNe(double a, double b) {
+    return ! fuzzyEq(a, b);
+}
+
+inline bool fuzzyGt(double a, double b) {
+    return a > b + eps(a, b);
+}
+
+inline bool fuzzyGe(double a, double b) {
+    return a > b - eps(a, b);
+}
+
+inline bool fuzzyLt(double a, double b) {
+    return a < b - eps(a, b);
+}
+
+inline bool fuzzyLe(double a, double b) {
+    return a < b + eps(a, b);
+}
+
+inline int iMod3(int x) {
+    return x % 3;
+}
+
+} // namespace G3D
+
+#ifdef _MSC_VER
+// Disable conditional expression is constant, which occurs incorrectly on inlined functions
+#   pragma warning (pop)
+#endif
diff --git a/dep/include/g3dlite/G3D/platform.h b/dep/include/g3dlite/G3D/platform.h
new file mode 100644
index 00000000000..fed9c7b5324
--- /dev/null
+++ b/dep/include/g3dlite/G3D/platform.h
@@ -0,0 +1,269 @@
+/**
+ @file platform.h
+
+ #defines for platform specific issues.
+
+ @maintainer Morgan McGuire, matrix@graphics3d.com
+
+ @created 2003-06-09
+ @edited  2006-01-16
+ */
+
+#ifndef G3D_PLATFORM_H
+#define G3D_PLATFORM_H
+
+/**
+ The version number of G3D in the form: MmmBB -> 
+ version M.mm [beta BB]
+ */
+#define G3D_VER 61000
+
+#if defined(G3D_RELEASEDEBUG)
+#   define G3D_DEBUGRELEASE
+#endif
+
+#if defined(G3D_DEBUGRELEASE) && defined(_DEBUG)
+#   undef _DEBUG
+#endif
+
+#if !defined(G3D_DEBUG) && (defined(_DEBUG) || defined(G3D_DEBUGRELEASE))
+#   define G3D_DEBUG
+#endif
+
+#ifdef _MSC_VER 
+    #define G3D_WIN32
+#elif defined(__MINGW32__)
+    #define G3D_WIN32 
+    #define G3D_MINGW32 
+#elif defined(__linux__)
+    #define G3D_LINUX
+#elif defined(__OpenBSD__)
+    #define G3D_LINUX
+#elif defined(__FreeBSD__)
+    #define G3D_LINUX
+#elif defined(__NetBSD__)
+    #define G3D_LINUX
+#elif defined(__APPLE__)
+    #define G3D_OSX
+#else
+    #error Unknown platform 
+#endif
+
+
+// Default to compiling with SSE, but if you want to compile 
+// without installing SP5.0 and the Processor Pack on Windows, compile with NO_SSE
+// defined (can be passed to the compiler command line with /D "NO_SSE")
+#if !defined(NO_SSE)
+   #define SSE
+#endif
+
+#ifdef G3D_WIN32
+// Turn off warnings about deprecated C routines (TODO: revisit)
+#	pragma warning (disable : 4996)
+#endif
+
+// On g++, recognize cases where the -msse2 flag was not specified
+#if defined(SSE) && defined(__GNUC__) && ! defined (__SSE__)
+#   undef SSE
+#endif
+
+#if defined(__GNUC__)
+#    if __STDC_VERSION__ < 199901
+#        define restrict __restrict__
+#    endif
+#endif
+
+
+// Verify that the supported compilers are being used and that this is a known
+// processor.
+
+#ifdef G3D_LINUX
+#   ifndef __GNUC__
+#       error G3D only supports the gcc compiler on Linux.
+#   endif
+
+//#   ifndef __i386__
+//#       error G3D only supports x86 machines on Linux.
+//#   endif
+
+#   define G3D_DEPRECATED __attribute__((__deprecated__))
+
+#   ifndef __cdecl
+#       define __cdecl __attribute__((cdecl))
+#   endif
+
+#   ifndef __stdcall
+#       define __stdcall __attribute__((stdcall))
+#   endif
+
+#   define G3D_CHECK_PRINTF_METHOD_ARGS   __attribute__((__format__(__printf__, 2, 3)))
+#   define G3D_CHECK_VPRINTF_METHOD_ARGS  __attribute__((__format__(__printf__, 2, 0)))
+#   define G3D_CHECK_PRINTF_ARGS          __attribute__((__format__(__printf__, 1, 2)))
+#   define G3D_CHECK_VPRINTF_ARGS         __attribute__((__format__(__printf__, 1, 0)))
+#endif
+
+
+#ifdef G3D_OSX
+    #ifndef __GNUC__
+        #error G3D only supports the gcc compiler on OS X.
+    #endif
+	
+	#if defined(__i386__)
+		#define G3D_OSX_INTEL
+	#elif defined(__PPC__)
+		#define G3D_OSX_PPC
+	#else
+		#define G3D_OSX_UNKNOWN
+	#endif
+
+#   ifndef __cdecl
+#       define __cdecl __attribute__((cdecl))
+#   endif
+
+#   ifndef __stdcall
+#       define __stdcall __attribute__((stdcall))
+#   endif
+
+#   define G3D_DEPRECATED __attribute__((__deprecated__))
+
+#   define G3D_CHECK_PRINTF_METHOD_ARGS   __attribute__((__format__(__printf__, 2, 3)))
+#   define G3D_CHECK_VPRINTF_METHOD_ARGS  __attribute__((__format__(__printf__, 2, 0)))
+#   define G3D_CHECK_PRINTF_ARGS          __attribute__((__format__(__printf__, 1, 2)))
+#   define G3D_CHECK_VPRINTF_ARGS         __attribute__((__format__(__printf__, 1, 0)))
+#endif
+
+
+#ifdef G3D_WIN32
+// Microsoft Visual C++ 7.1	_MSC_VER = 1310
+// Microsoft Visual C++ 7.0	_MSC_VER = 1300
+// Microsoft Visual C++ 6.0	_MSC_VER = 1200
+// Microsoft Visual C++ 5.0	_MSC_VER = 1100
+
+    // Old versions of MSVC (6.0 and previous) don't
+    // support C99 for loop scoping rules.  This fixes them.
+#   if (_MSC_VER <= 1200)
+        // This trick will generate a warning; disable the warning
+#       pragma warning (disable : 4127)
+#       define for if (false) {} else for
+#    endif
+
+#   if (_MSC_VER <= 1200)
+//      Nothing we can do on VC6 for deprecated functions
+#      define G3D_DEPRECATED
+#   else
+#      define G3D_DEPRECATED __declspec(deprecated)
+#   endif
+
+// Prevent Winsock conflicts by hiding the winsock API
+#ifndef _WINSOCKAPI_
+#   define _G3D_INTERNAL_HIDE_WINSOCK_
+#   define _WINSOCKAPI_
+#   endif
+
+// Disable 'name too long for browse information' warning
+#   pragma warning (disable : 4786)
+// TODO: remove
+#   pragma warning (disable : 4244)
+
+#	if defined(_MSC_VER) && (_MSC_VER <= 1200)
+		//	VC6 std:: has signed problems in it	
+#		pragma warning (disable : 4018)
+#	endif
+
+#   define ZLIB_WINAPI
+
+// Mingw32 defines restrict
+#   ifndef G3D_MINGW32
+#          define restrict
+#   endif
+
+#   define G3D_CHECK_PRINTF_ARGS 
+#   define G3D_CHECK_VPRINTF_ARGS
+#   define G3D_CHECK_PRINTF_METHOD_ARGS 
+#   define G3D_CHECK_VPRINTF_METHOD_ARGS
+
+    // On MSVC, we need to link against the multithreaded DLL version of
+    // the C++ runtime because that is what SDL and ZLIB are compiled
+    // against.  This is not the default for MSVC, so we set the following
+    // defines to force correct linking.  
+    //
+    // For documentation on compiler options, see:
+    //  http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vccore/html/_core_.2f.md.2c_2f.ml.2c_2f.mt.2c_2f.ld.asp
+    //  http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vccore98/HTML/_core_Compiler_Reference.asp
+    //
+
+#if 0 //ignore that for mangos
+    // DLL runtime
+    #ifndef _DLL
+	    #define _DLL
+    #endif
+
+    // Multithreaded runtime
+    #ifndef _MT
+	    #define _MT 1
+    #endif
+    // Ensure that we aren't forced into the static lib
+    #ifdef _STATIC_CPPLIB
+	    #undef _STATIC_CPPLIB
+    #endif
+#endif
+
+    #ifdef _DEBUG
+        #pragma comment (linker, "/NODEFAULTLIB:libc.lib")
+        #pragma comment (linker, "/NODEFAULTLIB:libcmt.lib")
+        #pragma comment (linker, "/NODEFAULTLIB:msvcrt.lib")
+        #pragma comment (linker, "/NODEFAULTLIB:libcd.lib")
+        #pragma comment (linker, "/NODEFAULTLIB:msvcrtd.lib")
+    #else
+        #pragma comment(linker, "/NODEFAULTLIB:LIBC.LIB")
+        #pragma comment(linker, "/NODEFAULTLIB:msvcrt.lib")
+        #pragma comment(linker, "/NODEFAULTLIB:libcd.lib")
+        #pragma comment(linker, "/NODEFAULTLIB:libcmtd.lib")
+        #pragma comment(linker, "/NODEFAULTLIB:msvcrtd.lib")
+    #endif
+
+    // Now set up external linking
+
+    #ifdef _DEBUG
+        // zlib and SDL were linked against the release MSVCRT; force
+        // the debug version.
+        #pragma comment(linker, "/NODEFAULTLIB:MSVCRT.LIB")
+#	endif
+
+#   ifndef WIN32_LEAN_AND_MEAN
+#       define WIN32_LEAN_AND_MEAN 1
+#   endif
+
+
+#   define NOMINMAX 1
+#   include <windows.h>
+#   undef WIN32_LEAN_AND_MEAN
+#   undef NOMINMAX
+
+#ifdef _G3D_INTERNAL_HIDE_WINSOCK_
+#   undef _G3D_INTERNAL_HIDE_WINSOCK_
+#   undef _WINSOCKAPI_
+#endif
+
+#endif
+
+#   if defined(_MSC_VER) && (_MSC_VER <= 1200)
+        // VC6 std:: has signed/unsigned problems
+#       pragma warning (disable : 4018)
+#   endif
+
+/** 
+  @def STR(expression)
+
+  Creates a string from the expression.  Frequently used with G3D::Shader
+  to express shading programs inline.  
+
+  <CODE>STR(this becomes a string)<PRE> evaluates the same as <CODE>"this becomes a string"</CODE>
+ */
+#define STR(x) #x
+
+#undef G3D_DEPRECATED
+#define G3D_DEPRECATED
+
+// Header guard
+#endif
diff --git a/dep/include/g3dlite/G3D/stringutils.h b/dep/include/g3dlite/G3D/stringutils.h
new file mode 100644
index 00000000000..9080d00b6a4
--- /dev/null
+++ b/dep/include/g3dlite/G3D/stringutils.h
@@ -0,0 +1,130 @@
+/**
+ @file stringutils.h
+ 
+ @maintainer Morgan McGuire, matrix@graphics3d.com
+ 
+ @author  2000-09-09
+ @edited  2002-11-30
+ */
+
+#ifndef G3D_STRINGUTILS_H
+#define G3D_STRINGUTILS_H
+
+#include "G3D/platform.h"
+#include "G3D/Array.h"
+#include <string>
+
+namespace G3D {
+
+extern const char* NEWLINE;
+
+/**
+ Returns true if the test string begins with the pattern string.
+ */
+bool beginsWith(
+    const std::string&          test,
+    const std::string&          pattern);
+
+/**
+ Returns true if the test string ends with the pattern string.
+ */
+bool endsWith(
+    const std::string&          test,
+    const std::string&          pattern);
+
+/**
+ Produces a new string that is the input string
+ wrapped at a certain number of columns (where
+ the line is broken at the latest space before the
+ column limit.)  Platform specific NEWLINEs
+ are inserted to wrap.
+ */
+std::string wordWrap(
+    const std::string&          input,
+    int                         numCols);
+
+/**
+ A comparison function for passing to Array::sort.
+ */
+int stringCompare(
+    const std::string&          s1,
+    const std::string&          s2);
+
+int stringPtrCompare(
+    const std::string*          s1,
+    const std::string*          s2);
+
+/**
+ Returns a new string that is an uppercase version of x.
+ */
+std::string toUpper(
+    const std::string&          x);
+
+std::string toLower(
+    const std::string&          x);
+
+/**
+ Splits x at each occurance of splitChar.
+ */
+G3D::Array<std::string> stringSplit(
+    const std::string&          x,
+    char                        splitChar);
+
+/**
+ joinChar is not inserted at the beginning or end, just in between
+ elements.
+ */
+std::string stringJoin(
+    const G3D::Array<std::string>&   a,
+    char                        joinChar);
+
+std::string stringJoin(
+    const G3D::Array<std::string>&   a,
+    const std::string&               joinStr);
+
+/**
+ Strips whitespace from both ends of the string.
+ */
+std::string trimWhitespace(
+    const std::string&              s);
+
+/** These standard C functions are renamed for clarity/naming
+   conventions and to return bool, not int.
+   */
+inline bool isWhiteSpace(const char c) {
+    return isspace(c) != 0;
+}
+
+/** These standard C functions are renamed for clarity/naming
+   conventions and to return bool, not int.
+   */
+inline bool isNewline(const char c) {
+    return (c == '\n') || (c == '\r');
+}
+
+/** These standard C functions are renamed for clarity/naming
+   conventions and to return bool, not int.
+   */
+inline bool isDigit(const char c) {
+    return isdigit(c) != 0;
+}
+
+/** These standard C functions are renamed for clarity/naming
+   conventions and to return bool, not int.
+   */
+inline bool isLetter(const char c) {
+    return isalpha(c) != 0;
+}
+
+inline bool isSlash(const char c) {
+    return (c == '\\') || (c == '/');
+}
+
+inline bool isQuote(const char c) {
+    return (c == '\'') || (c == '\"');
+}
+
+}; // namespace
+
+#endif
+