Add "cajun" for JSON reading and writing, Save searching in client and some more stuff for searcg

7 files changed, 1743 insertions, 0 deletions

diff --git a/src/cajun/elements.h b/src/cajun/elements.h
new file mode 100644
index 0000000..df1a0ca
--- /dev/null
+++ b/src/cajun/elements.h
@@ -0,0 +1,299 @@
+/******************************************************************************
+
+Copyright (c) 2009-2010, Terry Caton
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+    * Redistributions of source code must retain the above copyright 
+      notice, this list of conditions and the following disclaimer.
+    * Redistributions in binary form must reproduce the above copyright
+      notice, this list of conditions and the following disclaimer in the
+      documentation and/or other materials provided with the distribution.
+    * Neither the name of the projecct nor the names of its contributors 
+      may be used to endorse or promote products derived from this software 
+      without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE 
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
+ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+#pragma once
+
+#include <deque>
+#include <list>
+#include <string>
+#include <stdexcept>
+
+/*  
+
+TODO:
+* better documentation (doxygen?)
+* Unicode support
+* parent element accessors
+
+*/
+
+namespace json
+{
+
+namespace Version
+{
+   enum { MAJOR = 2 };
+   enum { MINOR = 0 };
+   enum {ENGINEERING = 2 };
+}
+
+/////////////////////////////////////////////////
+// forward declarations (more info further below)
+
+
+class Visitor;
+class ConstVisitor;
+
+template <typename ValueTypeT>
+class TrivialType_T;
+
+typedef TrivialType_T<double> Number;
+typedef TrivialType_T<bool> Boolean;
+typedef TrivialType_T<std::string> String;
+
+class Object;
+class Array;
+class Null;
+
+
+
+/////////////////////////////////////////////////////////////////////////
+// Exception - base class for all JSON-related runtime errors
+
+class Exception : public std::runtime_error
+{
+public:
+   Exception(const std::string& sMessage);
+};
+
+
+
+
+/////////////////////////////////////////////////////////////////////////
+// UnknownElement - provides a typesafe surrogate for any of the JSON-
+//  sanctioned element types. This class allows the Array and Object
+//  class to effectively contain a heterogeneous set of child elements.
+// The cast operators provide convenient implicit downcasting, while
+//  preserving dynamic type safety by throwing an exception during a
+//  a bad cast. 
+// The object & array element index operators (operators [std::string]
+//  and [size_t]) provide convenient, quick access to child elements.
+//  They are a logical extension of the cast operators. These child
+//  element accesses can be chained together, allowing the following
+//  (when document structure is well-known):
+//  String str = objInvoices[1]["Customer"]["Company"];
+
+
+class UnknownElement
+{
+public:
+   UnknownElement();
+   UnknownElement(const UnknownElement& unknown);
+   UnknownElement(const Object& object);
+   UnknownElement(const Array& array);
+   UnknownElement(const Number& number);
+   UnknownElement(const Boolean& boolean);
+   UnknownElement(const String& string);
+   UnknownElement(const Null& null);
+
+   ~UnknownElement();
+
+   UnknownElement& operator = (const UnknownElement& unknown);
+
+   // implicit cast to actual element type. throws on failure
+   operator const Object& () const;
+   operator const Array& () const;
+   operator const Number& () const;
+   operator const Boolean& () const;
+   operator const String& () const;
+   operator const Null& () const;
+
+   // implicit cast to actual element type. *converts* on failure, and always returns success
+   operator Object& ();
+   operator Array& ();
+   operator Number& ();
+   operator Boolean& ();
+   operator String& ();
+   operator Null& ();
+
+   // provides quick access to children when real element type is object
+   UnknownElement& operator[] (const std::string& key);
+   const UnknownElement& operator[] (const std::string& key) const;
+
+   // provides quick access to children when real element type is array
+   UnknownElement& operator[] (size_t index);
+   const UnknownElement& operator[] (size_t index) const;
+
+   // implements visitor pattern
+   void Accept(ConstVisitor& visitor) const;
+   void Accept(Visitor& visitor);
+
+   // tests equality. first checks type, then value if possible
+   bool operator == (const UnknownElement& element) const;
+
+private:
+   class Imp;
+
+   template <typename ElementTypeT>
+   class Imp_T;
+
+   class CastVisitor;
+   class ConstCastVisitor;
+   
+   template <typename ElementTypeT>
+   class CastVisitor_T;
+
+   template <typename ElementTypeT>
+   class ConstCastVisitor_T;
+
+   template <typename ElementTypeT>
+   const ElementTypeT& CastTo() const;
+
+   template <typename ElementTypeT>
+   ElementTypeT& ConvertTo();
+
+   Imp* m_pImp;
+};
+
+
+/////////////////////////////////////////////////////////////////////////////////
+// Array - mimics std::deque<UnknownElement>. The array contents are effectively 
+//  heterogeneous thanks to the ElementUnknown class. push_back has been replaced 
+//  by more generic insert functions.
+
+class Array
+{
+public:
+   typedef std::deque<UnknownElement> Elements;
+   typedef Elements::iterator iterator;
+   typedef Elements::const_iterator const_iterator;
+
+   iterator Begin();
+   iterator End();
+   const_iterator Begin() const;
+   const_iterator End() const;
+   
+   iterator Insert(const UnknownElement& element, iterator itWhere);
+   iterator Insert(const UnknownElement& element);
+   iterator Erase(iterator itWhere);
+   void Resize(size_t newSize);
+   void Clear();
+
+   size_t Size() const;
+   bool Empty() const;
+
+   UnknownElement& operator[] (size_t index);
+   const UnknownElement& operator[] (size_t index) const;
+
+   bool operator == (const Array& array) const;
+
+private:
+   Elements m_Elements;
+};
+
+
+/////////////////////////////////////////////////////////////////////////////////
+// Object - mimics std::map<std::string, UnknownElement>. The member value 
+//  contents are effectively heterogeneous thanks to the UnknownElement class
+
+class Object
+{
+public:
+   struct Member {
+      Member(const std::string& nameIn = std::string(), const UnknownElement& elementIn = UnknownElement());
+
+      bool operator == (const Member& member) const;
+
+      std::string name;
+      UnknownElement element;
+   };
+
+   typedef std::list<Member> Members; // map faster, but does not preserve order
+   typedef Members::iterator iterator;
+   typedef Members::const_iterator const_iterator;
+
+   bool operator == (const Object& object) const;
+
+   iterator Begin();
+   iterator End();
+   const_iterator Begin() const;
+   const_iterator End() const;
+
+   size_t Size() const;
+   bool Empty() const;
+
+   iterator Find(const std::string& name);
+   const_iterator Find(const std::string& name) const;
+
+   iterator Insert(const Member& member);
+   iterator Insert(const Member& member, iterator itWhere);
+   iterator Erase(iterator itWhere);
+   void Clear();
+
+   UnknownElement& operator [](const std::string& name);
+   const UnknownElement& operator [](const std::string& name) const;
+
+private:
+   class Finder;
+
+   Members m_Members;
+};
+
+
+/////////////////////////////////////////////////////////////////////////////////
+// TrivialType_T - class template for encapsulates a simple data type, such as
+//  a string, number, or boolean. Provides implicit const & noncost cast operators
+//  for that type, allowing "DataTypeT type = trivialType;"
+
+
+template <typename DataTypeT>
+class TrivialType_T
+{
+public:
+   TrivialType_T(const DataTypeT& t = DataTypeT());
+
+   operator DataTypeT&();
+   operator const DataTypeT&() const;
+
+   DataTypeT& Value();
+   const DataTypeT& Value() const;
+
+   bool operator == (const TrivialType_T<DataTypeT>& trivial) const;
+
+private:
+   DataTypeT m_tValue;
+};
+
+
+
+/////////////////////////////////////////////////////////////////////////////////
+// Null - doesn't do much of anything but satisfy the JSON spec. It is the default
+//  element type of UnknownElement
+
+class Null
+{
+public:
+   bool operator == (const Null& trivial) const;
+};
+
+
+} // End namespace
+
+
+#include "elements.inl"
diff --git a/src/cajun/elements.inl b/src/cajun/elements.inl
new file mode 100644
index 0000000..dbfbf2a
--- /dev/null
+++ b/src/cajun/elements.inl
@@ -0,0 +1,442 @@
+/******************************************************************************
+
+Copyright (c) 2009-2010, Terry Caton
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+    * Redistributions of source code must retain the above copyright 
+      notice, this list of conditions and the following disclaimer.
+    * Redistributions in binary form must reproduce the above copyright
+      notice, this list of conditions and the following disclaimer in the
+      documentation and/or other materials provided with the distribution.
+    * Neither the name of the projecct nor the names of its contributors 
+      may be used to endorse or promote products derived from this software 
+      without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE 
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
+ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+#include "visitor.h"
+#include "reader.h"
+#include <cassert>
+#include <algorithm>
+#include <map>
+
+/*  
+
+TODO:
+* better documentation
+
+*/
+
+namespace json
+{
+
+
+inline Exception::Exception(const std::string& sMessage) :
+   std::runtime_error(sMessage) {}
+
+
+/////////////////////////
+// UnknownElement members
+
+class UnknownElement::Imp
+{
+public:
+   virtual ~Imp() {}
+   virtual Imp* Clone() const = 0;
+
+   virtual bool Compare(const Imp& imp) const = 0;
+
+   virtual void Accept(ConstVisitor& visitor) const = 0;
+   virtual void Accept(Visitor& visitor) = 0;
+};
+
+
+template <typename ElementTypeT>
+class UnknownElement::Imp_T : public UnknownElement::Imp
+{
+public:
+   Imp_T(const ElementTypeT& element) : m_Element(element) {}
+   virtual Imp* Clone() const { return new Imp_T<ElementTypeT>(*this); }
+
+   virtual void Accept(ConstVisitor& visitor) const { visitor.Visit(m_Element); }
+   virtual void Accept(Visitor& visitor) { visitor.Visit(m_Element); }
+
+   virtual bool Compare(const Imp& imp) const
+   {
+      ConstCastVisitor_T<ElementTypeT> castVisitor;
+      imp.Accept(castVisitor);
+      return castVisitor.m_pElement &&
+             m_Element == *castVisitor.m_pElement;
+   }
+
+private:
+   ElementTypeT m_Element;
+};
+
+
+class UnknownElement::ConstCastVisitor : public ConstVisitor
+{
+   virtual void Visit(const Array& array) {}
+   virtual void Visit(const Object& object) {}
+   virtual void Visit(const Number& number) {}
+   virtual void Visit(const String& string) {}
+   virtual void Visit(const Boolean& boolean) {}
+   virtual void Visit(const Null& null) {}
+};
+
+template <typename ElementTypeT>
+class UnknownElement::ConstCastVisitor_T : public ConstCastVisitor
+{
+public:
+   ConstCastVisitor_T() : m_pElement(0) {}
+   virtual void Visit(const ElementTypeT& element) { m_pElement = &element; } // we don't know what this is, but it overrides one of the base's no-op functions
+   const ElementTypeT* m_pElement;
+};
+
+
+class UnknownElement::CastVisitor : public Visitor
+{
+   virtual void Visit(Array& array) {}
+   virtual void Visit(Object& object) {}
+   virtual void Visit(Number& number) {}
+   virtual void Visit(String& string) {}
+   virtual void Visit(Boolean& boolean) {}
+   virtual void Visit(Null& null) {}
+};
+
+template <typename ElementTypeT>
+class UnknownElement::CastVisitor_T : public CastVisitor
+{
+public:
+   CastVisitor_T() : m_pElement(0) {}
+   virtual void Visit(ElementTypeT& element) { m_pElement = &element; } // we don't know what this is, but it overrides one of the base's no-op functions
+   ElementTypeT* m_pElement;
+};
+
+
+
+
+inline UnknownElement::UnknownElement() :                               m_pImp( new Imp_T<Null>( Null() ) ) {}
+inline UnknownElement::UnknownElement(const UnknownElement& unknown) :  m_pImp( unknown.m_pImp->Clone()) {}
+inline UnknownElement::UnknownElement(const Object& object) :           m_pImp( new Imp_T<Object>(object) ) {}
+inline UnknownElement::UnknownElement(const Array& array) :             m_pImp( new Imp_T<Array>(array) ) {}
+inline UnknownElement::UnknownElement(const Number& number) :           m_pImp( new Imp_T<Number>(number) ) {}
+inline UnknownElement::UnknownElement(const Boolean& boolean) :         m_pImp( new Imp_T<Boolean>(boolean) ) {}
+inline UnknownElement::UnknownElement(const String& string) :           m_pImp( new Imp_T<String>(string) ) {}
+inline UnknownElement::UnknownElement(const Null& null) :               m_pImp( new Imp_T<Null>(null) ) {}
+
+inline UnknownElement::~UnknownElement()   { delete m_pImp; }
+
+inline UnknownElement::operator const Object& () const    { return CastTo<Object>(); }
+inline UnknownElement::operator const Array& () const     { return CastTo<Array>(); }
+inline UnknownElement::operator const Number& () const    { return CastTo<Number>(); }
+inline UnknownElement::operator const Boolean& () const   { return CastTo<Boolean>(); }
+inline UnknownElement::operator const String& () const    { return CastTo<String>(); }
+inline UnknownElement::operator const Null& () const      { return CastTo<Null>(); }
+
+inline UnknownElement::operator Object& ()    { return ConvertTo<Object>(); }
+inline UnknownElement::operator Array& ()     { return ConvertTo<Array>(); }
+inline UnknownElement::operator Number& ()    { return ConvertTo<Number>(); }
+inline UnknownElement::operator Boolean& ()   { return ConvertTo<Boolean>(); }
+inline UnknownElement::operator String& ()    { return ConvertTo<String>(); }
+inline UnknownElement::operator Null& ()      { return ConvertTo<Null>(); }
+
+inline UnknownElement& UnknownElement::operator = (const UnknownElement& unknown) 
+{
+   // always check for this
+   if (&unknown != this)
+   {
+      // we might be copying from a subtree of ourselves. delete the old imp
+      //  only after the clone operation is complete. yes, this could be made 
+      //  more efficient, but isn't worth the complexity
+      Imp* pOldImp = m_pImp;
+      m_pImp = unknown.m_pImp->Clone();
+      delete pOldImp;
+   }
+
+   return *this;
+}
+
+inline UnknownElement& UnknownElement::operator[] (const std::string& key)
+{
+   // the people want an object. make us one if we aren't already
+   Object& object = ConvertTo<Object>();
+   return object[key];
+}
+
+inline const UnknownElement& UnknownElement::operator[] (const std::string& key) const
+{
+   // throws if we aren't an object
+   const Object& object = CastTo<Object>();
+   return object[key];
+}
+
+inline UnknownElement& UnknownElement::operator[] (size_t index)
+{
+   // the people want an array. make us one if we aren't already
+   Array& array = ConvertTo<Array>();
+   return array[index];
+}
+
+inline const UnknownElement& UnknownElement::operator[] (size_t index) const
+{
+   // throws if we aren't an array
+   const Array& array = CastTo<Array>();
+   return array[index];
+}
+
+
+template <typename ElementTypeT>
+const ElementTypeT& UnknownElement::CastTo() const
+{
+   ConstCastVisitor_T<ElementTypeT> castVisitor;
+   m_pImp->Accept(castVisitor);
+   if (castVisitor.m_pElement == 0)
+      throw Exception("Bad cast");
+   return *castVisitor.m_pElement;
+}
+
+
+
+template <typename ElementTypeT>
+ElementTypeT& UnknownElement::ConvertTo() 
+{
+   CastVisitor_T<ElementTypeT> castVisitor;
+   m_pImp->Accept(castVisitor);
+   if (castVisitor.m_pElement == 0)
+   {
+      // we're not the right type. fix it & try again
+      *this = ElementTypeT();
+      m_pImp->Accept(castVisitor);
+   }
+
+   return *castVisitor.m_pElement;
+}
+
+
+inline void UnknownElement::Accept(ConstVisitor& visitor) const { m_pImp->Accept(visitor); }
+inline void UnknownElement::Accept(Visitor& visitor)            { m_pImp->Accept(visitor); }
+
+
+inline bool UnknownElement::operator == (const UnknownElement& element) const
+{
+   return m_pImp->Compare(*element.m_pImp);
+}
+
+
+
+//////////////////
+// Object members
+
+
+inline Object::Member::Member(const std::string& nameIn, const UnknownElement& elementIn) :
+   name(nameIn), element(elementIn) {}
+
+inline bool Object::Member::operator == (const Member& member) const 
+{
+   return name == member.name &&
+          element == member.element;
+}
+
+class Object::Finder : public std::unary_function<Object::Member, bool>
+{
+public:
+   Finder(const std::string& name) : m_name(name) {}
+   bool operator () (const Object::Member& member) {
+      return member.name == m_name;
+   }
+
+private:
+   std::string m_name;
+};
+
+
+
+inline Object::iterator Object::Begin() { return m_Members.begin(); }
+inline Object::iterator Object::End() { return m_Members.end(); }
+inline Object::const_iterator Object::Begin() const { return m_Members.begin(); }
+inline Object::const_iterator Object::End() const { return m_Members.end(); }
+
+inline size_t Object::Size() const { return m_Members.size(); }
+inline bool Object::Empty() const { return m_Members.empty(); }
+
+inline Object::iterator Object::Find(const std::string& name) 
+{
+   return std::find_if(m_Members.begin(), m_Members.end(), Finder(name));
+}
+
+inline Object::const_iterator Object::Find(const std::string& name) const 
+{
+   return std::find_if(m_Members.begin(), m_Members.end(), Finder(name));
+}
+
+inline Object::iterator Object::Insert(const Member& member)
+{
+   return Insert(member, End());
+}
+
+inline Object::iterator Object::Insert(const Member& member, iterator itWhere)
+{
+   iterator it = Find(member.name);
+   if (it != m_Members.end())
+      throw Exception(std::string("Object member already exists: ") + member.name);
+
+   it = m_Members.insert(itWhere, member);
+   return it;
+}
+
+inline Object::iterator Object::Erase(iterator itWhere) 
+{
+   return m_Members.erase(itWhere);
+}
+
+inline UnknownElement& Object::operator [](const std::string& name)
+{
+
+   iterator it = Find(name);
+   if (it == m_Members.end())
+   {
+      Member member(name);
+      it = Insert(member, End());
+   }
+   return it->element;      
+}
+
+inline const UnknownElement& Object::operator [](const std::string& name) const 
+{
+   const_iterator it = Find(name);
+   if (it == End())
+      throw Exception(std::string("Object member not found: ") + name);
+   return it->element;
+}
+
+inline void Object::Clear() 
+{
+   m_Members.clear(); 
+}
+
+inline bool Object::operator == (const Object& object) const 
+{
+   return m_Members == object.m_Members;
+}
+
+
+/////////////////
+// Array members
+
+inline Array::iterator Array::Begin()  { return m_Elements.begin(); }
+inline Array::iterator Array::End()    { return m_Elements.end(); }
+inline Array::const_iterator Array::Begin() const  { return m_Elements.begin(); }
+inline Array::const_iterator Array::End() const    { return m_Elements.end(); }
+
+inline Array::iterator Array::Insert(const UnknownElement& element, iterator itWhere)
+{ 
+   return m_Elements.insert(itWhere, element);
+}
+
+inline Array::iterator Array::Insert(const UnknownElement& element)
+{
+   return Insert(element, End());
+}
+
+inline Array::iterator Array::Erase(iterator itWhere)
+{ 
+   return m_Elements.erase(itWhere);
+}
+
+inline void Array::Resize(size_t newSize)
+{
+   m_Elements.resize(newSize);
+}
+
+inline size_t Array::Size() const  { return m_Elements.size(); }
+inline bool Array::Empty() const   { return m_Elements.empty(); }
+
+inline UnknownElement& Array::operator[] (size_t index)
+{
+   size_t nMinSize = index + 1; // zero indexed
+   if (m_Elements.size() < nMinSize)
+      m_Elements.resize(nMinSize);
+   return m_Elements[index]; 
+}
+
+inline const UnknownElement& Array::operator[] (size_t index) const 
+{
+   if (index >= m_Elements.size())
+      throw Exception("Array out of bounds");
+   return m_Elements[index]; 
+}
+
+inline void Array::Clear() {
+   m_Elements.clear();
+}
+
+inline bool Array::operator == (const Array& array) const
+{
+   return m_Elements == array.m_Elements;
+}
+
+
+////////////////////////
+// TrivialType_T members
+
+template <typename DataTypeT>
+TrivialType_T<DataTypeT>::TrivialType_T(const DataTypeT& t) :
+   m_tValue(t) {}
+
+template <typename DataTypeT>
+TrivialType_T<DataTypeT>::operator DataTypeT&()
+{
+   return Value(); 
+}
+
+template <typename DataTypeT>
+TrivialType_T<DataTypeT>::operator const DataTypeT&() const
+{
+   return Value(); 
+}
+
+template <typename DataTypeT>
+DataTypeT& TrivialType_T<DataTypeT>::Value()
+{
+   return m_tValue; 
+}
+
+template <typename DataTypeT>
+const DataTypeT& TrivialType_T<DataTypeT>::Value() const
+{
+   return m_tValue; 
+}
+
+template <typename DataTypeT>
+bool TrivialType_T<DataTypeT>::operator == (const TrivialType_T<DataTypeT>& trivial) const
+{
+   return m_tValue == trivial.m_tValue;
+}
+
+
+
+//////////////////
+// Null members
+
+inline bool Null::operator == (const Null& trivial) const
+{
+   return true;
+}
+
+
+
+} // End namespace
diff --git a/src/cajun/reader.h b/src/cajun/reader.h
new file mode 100644
index 0000000..f431670
--- /dev/null
+++ b/src/cajun/reader.h
@@ -0,0 +1,148 @@
+/******************************************************************************
+
+Copyright (c) 2009-2010, Terry Caton
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+    * Redistributions of source code must retain the above copyright 
+      notice, this list of conditions and the following disclaimer.
+    * Redistributions in binary form must reproduce the above copyright
+      notice, this list of conditions and the following disclaimer in the
+      documentation and/or other materials provided with the distribution.
+    * Neither the name of the projecct nor the names of its contributors 
+      may be used to endorse or promote products derived from this software 
+      without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE 
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
+ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+
+
+#pragma once
+
+#include "elements.h"
+#include <iostream>
+#include <vector>
+
+namespace json
+{
+
+class Reader
+{
+public:
+   // this structure will be reported in one of the exceptions defined below
+   struct Location
+   {
+      Location();
+
+      unsigned int m_nLine;       // document line, zero-indexed
+      unsigned int m_nLineOffset; // character offset from beginning of line, zero indexed
+      unsigned int m_nDocOffset;  // character offset from entire document, zero indexed
+   };
+
+   // thrown during the first phase of reading. generally catches low-level problems such
+   //  as errant characters or corrupt/incomplete documents
+   class ScanException : public Exception
+   {
+   public:
+      ScanException(const std::string& sMessage, const Reader::Location& locError) :
+         Exception(sMessage),
+         m_locError(locError) {}
+
+      Reader::Location m_locError;
+   };
+
+   // thrown during the second phase of reading. generally catches higher-level problems such
+   //  as missing commas or brackets
+   class ParseException : public Exception
+   {
+   public:
+      ParseException(const std::string& sMessage, const Reader::Location& locTokenBegin, const Reader::Location& locTokenEnd) :
+         Exception(sMessage),
+         m_locTokenBegin(locTokenBegin),
+         m_locTokenEnd(locTokenEnd) {}
+
+      Reader::Location m_locTokenBegin;
+      Reader::Location m_locTokenEnd;
+   };
+
+
+   // if you know what the document looks like, call one of these...
+   static void Read(Object& object, std::istream& istr);
+   static void Read(Array& array, std::istream& istr);
+   static void Read(String& string, std::istream& istr);
+   static void Read(Number& number, std::istream& istr);
+   static void Read(Boolean& boolean, std::istream& istr);
+   static void Read(Null& null, std::istream& istr);
+
+   // ...otherwise, if you don't know, call this & visit it
+   static void Read(UnknownElement& elementRoot, std::istream& istr);
+
+private:
+   struct Token
+   {
+      enum Type
+      {
+         TOKEN_OBJECT_BEGIN,  //    {
+         TOKEN_OBJECT_END,    //    }
+         TOKEN_ARRAY_BEGIN,   //    [
+         TOKEN_ARRAY_END,     //    ]
+         TOKEN_NEXT_ELEMENT,  //    ,
+         TOKEN_MEMBER_ASSIGN, //    :
+         TOKEN_STRING,        //    "xxx"
+         TOKEN_NUMBER,        //    [+/-]000.000[e[+/-]000]
+         TOKEN_BOOLEAN,       //    true -or- false
+         TOKEN_NULL,          //    null
+      };
+
+      Type nType;
+      std::string sValue;
+
+      // for malformed file debugging
+      Reader::Location locBegin;
+      Reader::Location locEnd;
+   };
+
+   class InputStream;
+   class TokenStream;
+   typedef std::vector<Token> Tokens;
+
+   template <typename ElementTypeT>   
+   static void Read_i(ElementTypeT& element, std::istream& istr);
+
+   // scanning istream into token sequence
+   void Scan(Tokens& tokens, InputStream& inputStream);
+
+   void EatWhiteSpace(InputStream& inputStream);
+   std::string MatchString(InputStream& inputStream);
+   std::string MatchNumber(InputStream& inputStream);
+   std::string MatchExpectedString(InputStream& inputStream, const std::string& sExpected);
+
+   // parsing token sequence into element structure
+   void Parse(UnknownElement& element, TokenStream& tokenStream);
+   void Parse(Object& object, TokenStream& tokenStream);
+   void Parse(Array& array, TokenStream& tokenStream);
+   void Parse(String& string, TokenStream& tokenStream);
+   void Parse(Number& number, TokenStream& tokenStream);
+   void Parse(Boolean& boolean, TokenStream& tokenStream);
+   void Parse(Null& null, TokenStream& tokenStream);
+
+   const std::string& MatchExpectedToken(Token::Type nExpected, TokenStream& tokenStream);
+};
+
+
+} // End namespace
+
+
+#include "reader.inl"
\ No newline at end of file
diff --git a/src/cajun/reader.inl b/src/cajun/reader.inl
new file mode 100644
index 0000000..47311ad
--- /dev/null
+++ b/src/cajun/reader.inl
@@ -0,0 +1,533 @@
+/******************************************************************************
+
+Copyright (c) 2009-2010, Terry Caton
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+    * Redistributions of source code must retain the above copyright 
+      notice, this list of conditions and the following disclaimer.
+    * Redistributions in binary form must reproduce the above copyright
+      notice, this list of conditions and the following disclaimer in the
+      documentation and/or other materials provided with the distribution.
+    * Neither the name of the projecct nor the names of its contributors 
+      may be used to endorse or promote products derived from this software 
+      without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE 
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
+ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+#include <cassert>
+#include <set>
+#include <sstream>
+
+/*  
+
+TODO:
+* better documentation
+* unicode character decoding
+
+*/
+
+namespace json
+{
+
+inline std::istream& operator >> (std::istream& istr, UnknownElement& elementRoot) {
+   Reader::Read(elementRoot, istr);
+   return istr;
+}
+
+inline Reader::Location::Location() :
+   m_nLine(0),
+   m_nLineOffset(0),
+   m_nDocOffset(0)
+{}
+
+
+//////////////////////
+// Reader::InputStream
+
+class Reader::InputStream // would be cool if we could inherit from std::istream & override "get"
+{
+public:
+   InputStream(std::istream& iStr) :
+      m_iStr(iStr) {}
+
+   // protect access to the input stream, so we can keeep track of document/line offsets
+   char Get(); // big, define outside
+   char Peek() {
+      assert(m_iStr.eof() == false); // enforce reading of only valid stream data 
+      return m_iStr.peek();
+   }
+
+   bool EOS() {
+      m_iStr.peek(); // apparently eof flag isn't set until a character read is attempted. whatever.
+      return m_iStr.eof();
+   }
+
+   const Location& GetLocation() const { return m_Location; }
+
+private:
+   std::istream& m_iStr;
+   Location m_Location;
+};
+
+
+inline char Reader::InputStream::Get()
+{
+   assert(m_iStr.eof() == false); // enforce reading of only valid stream data 
+   char c = m_iStr.get();
+   
+   ++m_Location.m_nDocOffset;
+   if (c == '\n') {
+      ++m_Location.m_nLine;
+      m_Location.m_nLineOffset = 0;
+   }
+   else {
+      ++m_Location.m_nLineOffset;
+   }
+
+   return c;
+}
+
+
+
+//////////////////////
+// Reader::TokenStream
+
+class Reader::TokenStream
+{
+public:
+   TokenStream(const Tokens& tokens);
+
+   const Token& Peek();
+   const Token& Get();
+
+   bool EOS() const;
+
+private:
+   const Tokens& m_Tokens;
+   Tokens::const_iterator m_itCurrent;
+};
+
+
+inline Reader::TokenStream::TokenStream(const Tokens& tokens) :
+   m_Tokens(tokens),
+   m_itCurrent(tokens.begin())
+{}
+
+inline const Reader::Token& Reader::TokenStream::Peek() {
+   if (EOS())
+   {
+      const Token& lastToken = *m_Tokens.rbegin();
+      std::string sMessage = "Unexpected end of token stream";
+      throw ParseException(sMessage, lastToken.locBegin, lastToken.locEnd); // nowhere to point to
+   }
+   return *(m_itCurrent); 
+}
+
+inline const Reader::Token& Reader::TokenStream::Get() {
+   const Token& token = Peek();
+   ++m_itCurrent;
+   return token;
+}
+
+inline bool Reader::TokenStream::EOS() const {
+   return m_itCurrent == m_Tokens.end(); 
+}
+
+///////////////////
+// Reader (finally)
+
+
+inline void Reader::Read(Object& object, std::istream& istr)                { Read_i(object, istr); }
+inline void Reader::Read(Array& array, std::istream& istr)                  { Read_i(array, istr); }
+inline void Reader::Read(String& string, std::istream& istr)                { Read_i(string, istr); }
+inline void Reader::Read(Number& number, std::istream& istr)                { Read_i(number, istr); }
+inline void Reader::Read(Boolean& boolean, std::istream& istr)              { Read_i(boolean, istr); }
+inline void Reader::Read(Null& null, std::istream& istr)                    { Read_i(null, istr); }
+inline void Reader::Read(UnknownElement& unknown, std::istream& istr)       { Read_i(unknown, istr); }
+
+
+template <typename ElementTypeT>   
+void Reader::Read_i(ElementTypeT& element, std::istream& istr)
+{
+   Reader reader;
+
+   Tokens tokens;
+   InputStream inputStream(istr);
+   reader.Scan(tokens, inputStream);
+
+   TokenStream tokenStream(tokens);
+   reader.Parse(element, tokenStream);
+
+   if (tokenStream.EOS() == false)
+   {
+      const Token& token = tokenStream.Peek();
+      std::string sMessage = std::string("Expected End of token stream; found ") + token.sValue;
+      throw ParseException(sMessage, token.locBegin, token.locEnd);
+   }
+}
+
+
+inline void Reader::Scan(Tokens& tokens, InputStream& inputStream)
+{
+   while (EatWhiteSpace(inputStream),              // ignore any leading white space...
+          inputStream.EOS() == false) // ...before checking for EOS
+   {
+      // if all goes well, we'll create a token each pass
+      Token token;
+      token.locBegin = inputStream.GetLocation();
+
+      // gives us null-terminated string
+      char sChar = inputStream.Peek();
+      switch (sChar)
+      {
+         case '{':
+            token.sValue = MatchExpectedString(inputStream, "{");
+            token.nType = Token::TOKEN_OBJECT_BEGIN;
+            break;
+
+         case '}':
+            token.sValue = MatchExpectedString(inputStream, "}");
+            token.nType = Token::TOKEN_OBJECT_END;
+            break;
+
+         case '[':
+            token.sValue = MatchExpectedString(inputStream, "[");
+            token.nType = Token::TOKEN_ARRAY_BEGIN;
+            break;
+
+         case ']':
+            token.sValue = MatchExpectedString(inputStream, "]");
+            token.nType = Token::TOKEN_ARRAY_END;
+            break;
+
+         case ',':
+            token.sValue = MatchExpectedString(inputStream, ",");
+            token.nType = Token::TOKEN_NEXT_ELEMENT;
+            break;
+
+         case ':':
+            token.sValue = MatchExpectedString(inputStream, ":");
+            token.nType = Token::TOKEN_MEMBER_ASSIGN;
+            break;
+
+         case '"':
+            token.sValue = MatchString(inputStream);
+            token.nType = Token::TOKEN_STRING;
+            break;
+
+         case '-':
+         case '0':
+         case '1':
+         case '2':
+         case '3':
+         case '4':
+         case '5':
+         case '6':
+         case '7':
+         case '8':
+         case '9':
+            token.sValue = MatchNumber(inputStream);
+            token.nType = Token::TOKEN_NUMBER;
+            break;
+
+         case 't':
+            token.sValue = MatchExpectedString(inputStream, "true");
+            token.nType = Token::TOKEN_BOOLEAN;
+            break;
+
+         case 'f':
+            token.sValue = MatchExpectedString(inputStream, "false");
+            token.nType = Token::TOKEN_BOOLEAN;
+            break;
+
+         case 'n':
+            token.sValue = MatchExpectedString(inputStream, "null");
+            token.nType = Token::TOKEN_NULL;
+            break;
+
+         default:
+         {
+            std::string sErrorMessage = std::string("Unexpected character in stream: ") + sChar;
+            throw ScanException(sErrorMessage, inputStream.GetLocation());
+         }
+      }
+
+      token.locEnd = inputStream.GetLocation();
+      tokens.push_back(token);
+   }
+}
+
+
+inline void Reader::EatWhiteSpace(InputStream& inputStream)
+{
+   while (inputStream.EOS() == false && 
+          ::isspace(inputStream.Peek()))
+      inputStream.Get();
+}
+
+inline std::string Reader::MatchExpectedString(InputStream& inputStream, const std::string& sExpected)
+{
+   std::string::const_iterator it(sExpected.begin()),
+                               itEnd(sExpected.end());
+   for ( ; it != itEnd; ++it) {
+      if (inputStream.EOS() ||      // did we reach the end before finding what we're looking for...
+          inputStream.Get() != *it) // ...or did we find something different?
+      {
+         std::string sMessage = std::string("Expected string: ") + sExpected;
+         throw ScanException(sMessage, inputStream.GetLocation());
+      }
+   }
+
+   // all's well if we made it here
+   return sExpected;
+}
+
+
+inline std::string Reader::MatchString(InputStream& inputStream)
+{
+   MatchExpectedString(inputStream, "\"");
+
+   std::string string;
+   while (inputStream.EOS() == false &&
+          inputStream.Peek() != '"')
+   {
+      char c = inputStream.Get();
+
+      // escape?
+      if (c == '\\' &&
+          inputStream.EOS() == false) // shouldn't have reached the end yet
+      {
+         c = inputStream.Get();
+         switch (c) {
+            case '/':      string.push_back('/');     break;
+            case '"':      string.push_back('"');     break;
+            case '\\':     string.push_back('\\');    break;
+            case 'b':      string.push_back('\b');    break;
+            case 'f':      string.push_back('\f');    break;
+            case 'n':      string.push_back('\n');    break;
+            case 'r':      string.push_back('\r');    break;
+            case 't':      string.push_back('\t');    break;
+			//case 'u':      string.push_back('\u');    break; // TODO: what do we do with this?
+            default: {
+               std::string sMessage = std::string("Unrecognized escape sequence found in string: \\") + c;
+               throw ScanException(sMessage, inputStream.GetLocation());
+            }
+         }
+      }
+      else {
+         string.push_back(c);
+      }
+   }
+
+   // eat the last '"' that we just peeked
+   MatchExpectedString(inputStream, "\"");
+
+   // all's well if we made it here
+   return string;
+}
+
+
+inline std::string Reader::MatchNumber(InputStream& inputStream)
+{
+   const char sNumericChars[] = "0123456789.eE-+";
+   std::set<char> numericChars;
+   numericChars.insert(sNumericChars, sNumericChars + sizeof(sNumericChars));
+
+   std::string sNumber;
+   while (inputStream.EOS() == false &&
+          numericChars.find(inputStream.Peek()) != numericChars.end())
+   {
+      sNumber.push_back(inputStream.Get());   
+   }
+
+   return sNumber;
+}
+
+
+inline void Reader::Parse(UnknownElement& element, Reader::TokenStream& tokenStream) 
+{
+   const Token& token = tokenStream.Peek();
+   switch (token.nType) {
+      case Token::TOKEN_OBJECT_BEGIN:
+      {
+         // implicit non-const cast will perform conversion for us (if necessary)
+         Object& object = element;
+         Parse(object, tokenStream);
+         break;
+      }
+
+      case Token::TOKEN_ARRAY_BEGIN:
+      {
+         Array& array = element;
+         Parse(array, tokenStream);
+         break;
+      }
+
+      case Token::TOKEN_STRING:
+      {
+         String& string = element;
+         Parse(string, tokenStream);
+         break;
+      }
+
+      case Token::TOKEN_NUMBER:
+      {
+         Number& number = element;
+         Parse(number, tokenStream);
+         break;
+      }
+
+      case Token::TOKEN_BOOLEAN:
+      {
+         Boolean& boolean = element;
+         Parse(boolean, tokenStream);
+         break;
+      }
+
+      case Token::TOKEN_NULL:
+      {
+         Null& null = element;
+         Parse(null, tokenStream);
+         break;
+      }
+
+      default:
+      {
+         std::string sMessage = std::string("Unexpected token: ") + token.sValue;
+         throw ParseException(sMessage, token.locBegin, token.locEnd);
+      }
+   }
+}
+
+
+inline void Reader::Parse(Object& object, Reader::TokenStream& tokenStream)
+{
+   MatchExpectedToken(Token::TOKEN_OBJECT_BEGIN, tokenStream);
+
+   bool bContinue = (tokenStream.EOS() == false &&
+                     tokenStream.Peek().nType != Token::TOKEN_OBJECT_END);
+   while (bContinue)
+   {
+      Object::Member member;
+
+      // first the member name. save the token in case we have to throw an exception
+      const Token& tokenName = tokenStream.Peek();
+      member.name = MatchExpectedToken(Token::TOKEN_STRING, tokenStream);
+
+      // ...then the key/value separator...
+      MatchExpectedToken(Token::TOKEN_MEMBER_ASSIGN, tokenStream);
+
+      // ...then the value itself (can be anything).
+      Parse(member.element, tokenStream);
+
+      // try adding it to the object (this could throw)
+      try
+      {
+         object.Insert(member);
+      }
+      catch (Exception&)
+      {
+         // must be a duplicate name
+         std::string sMessage = std::string("Duplicate object member token: ") + member.name; 
+         throw ParseException(sMessage, tokenName.locBegin, tokenName.locEnd);
+      }
+
+      bContinue = (tokenStream.EOS() == false &&
+                   tokenStream.Peek().nType == Token::TOKEN_NEXT_ELEMENT);
+      if (bContinue)
+         MatchExpectedToken(Token::TOKEN_NEXT_ELEMENT, tokenStream);
+   }
+
+   MatchExpectedToken(Token::TOKEN_OBJECT_END, tokenStream);
+}
+
+
+inline void Reader::Parse(Array& array, Reader::TokenStream& tokenStream)
+{
+   MatchExpectedToken(Token::TOKEN_ARRAY_BEGIN, tokenStream);
+
+   bool bContinue = (tokenStream.EOS() == false &&
+                     tokenStream.Peek().nType != Token::TOKEN_ARRAY_END);
+   while (bContinue)
+   {
+      // ...what's next? could be anything
+      Array::iterator itElement = array.Insert(UnknownElement());
+      UnknownElement& element = *itElement;
+      Parse(element, tokenStream);
+
+      bContinue = (tokenStream.EOS() == false &&
+                   tokenStream.Peek().nType == Token::TOKEN_NEXT_ELEMENT);
+      if (bContinue)
+         MatchExpectedToken(Token::TOKEN_NEXT_ELEMENT, tokenStream);
+   }
+
+   MatchExpectedToken(Token::TOKEN_ARRAY_END, tokenStream);
+}
+
+
+inline void Reader::Parse(String& string, Reader::TokenStream& tokenStream)
+{
+   string = MatchExpectedToken(Token::TOKEN_STRING, tokenStream);
+}
+
+
+inline void Reader::Parse(Number& number, Reader::TokenStream& tokenStream)
+{
+   const Token& currentToken = tokenStream.Peek(); // might need this later for throwing exception
+   const std::string& sValue = MatchExpectedToken(Token::TOKEN_NUMBER, tokenStream);
+
+   std::istringstream iStr(sValue);
+   double dValue;
+   iStr >> dValue;
+
+   // did we consume all characters in the token?
+   if (iStr.eof() == false)
+   {
+      char c = iStr.peek();
+      std::string sMessage = std::string("Unexpected character in NUMBER token: ") + c;
+      throw ParseException(sMessage, currentToken.locBegin, currentToken.locEnd);
+   }
+
+   number = dValue;
+}
+
+
+inline void Reader::Parse(Boolean& boolean, Reader::TokenStream& tokenStream)
+{
+   const std::string& sValue = MatchExpectedToken(Token::TOKEN_BOOLEAN, tokenStream);
+   boolean = (sValue == "true" ? true : false);
+}
+
+
+inline void Reader::Parse(Null&, Reader::TokenStream& tokenStream)
+{
+   MatchExpectedToken(Token::TOKEN_NULL, tokenStream);
+}
+
+
+inline const std::string& Reader::MatchExpectedToken(Token::Type nExpected, Reader::TokenStream& tokenStream)
+{
+   const Token& token = tokenStream.Get();
+   if (token.nType != nExpected)
+   {
+      std::string sMessage = std::string("Unexpected token: ") + token.sValue;
+      throw ParseException(sMessage, token.locBegin, token.locEnd);
+   }
+
+   return token.sValue;
+}
+
+} // End namespace
diff --git a/src/cajun/visitor.h b/src/cajun/visitor.h
new file mode 100644
index 0000000..a06299d
--- /dev/null
+++ b/src/cajun/visitor.h
@@ -0,0 +1,65 @@
+/******************************************************************************
+
+Copyright (c) 2009-2010, Terry Caton
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+    * Redistributions of source code must retain the above copyright 
+      notice, this list of conditions and the following disclaimer.
+    * Redistributions in binary form must reproduce the above copyright
+      notice, this list of conditions and the following disclaimer in the
+      documentation and/or other materials provided with the distribution.
+    * Neither the name of the projecct nor the names of its contributors 
+      may be used to endorse or promote products derived from this software 
+      without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE 
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
+ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+#pragma once
+
+#include "elements.h"
+
+namespace json
+{
+
+
+class Visitor
+{
+public:
+   virtual ~Visitor() {}
+
+   virtual void Visit(Array& array) = 0;
+   virtual void Visit(Object& object) = 0;
+   virtual void Visit(Number& number) = 0;
+   virtual void Visit(String& string) = 0;
+   virtual void Visit(Boolean& boolean) = 0;
+   virtual void Visit(Null& null) = 0;
+};
+
+class ConstVisitor
+{
+public:
+   virtual ~ConstVisitor() {}
+
+   virtual void Visit(const Array& array) = 0;
+   virtual void Visit(const Object& object) = 0;
+   virtual void Visit(const Number& number) = 0;
+   virtual void Visit(const String& string) = 0;
+   virtual void Visit(const Boolean& boolean) = 0;
+   virtual void Visit(const Null& null) = 0;
+};
+
+
+} // End namespace
diff --git a/src/cajun/writer.h b/src/cajun/writer.h
new file mode 100644
index 0000000..594d541
--- /dev/null
+++ b/src/cajun/writer.h
@@ -0,0 +1,78 @@
+/******************************************************************************
+
+Copyright (c) 2009-2010, Terry Caton
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+    * Redistributions of source code must retain the above copyright 
+      notice, this list of conditions and the following disclaimer.
+    * Redistributions in binary form must reproduce the above copyright
+      notice, this list of conditions and the following disclaimer in the
+      documentation and/or other materials provided with the distribution.
+    * Neither the name of the projecct nor the names of its contributors 
+      may be used to endorse or promote products derived from this software 
+      without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE 
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
+ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+#pragma once
+
+#include "elements.h"
+#include "visitor.h"
+
+namespace json
+{
+
+class Writer : private ConstVisitor
+{
+public:
+   static void Write(const Object& object, std::ostream& ostr);
+   static void Write(const Array& array, std::ostream& ostr);
+   static void Write(const String& string, std::ostream& ostr);
+   static void Write(const Number& number, std::ostream& ostr);
+   static void Write(const Boolean& boolean, std::ostream& ostr);
+   static void Write(const Null& null, std::ostream& ostr);
+   static void Write(const UnknownElement& elementRoot, std::ostream& ostr);
+
+private:
+   Writer(std::ostream& ostr);
+
+   template <typename ElementTypeT>
+   static void Write_i(const ElementTypeT& element, std::ostream& ostr);
+
+   void Write_i(const Object& object);
+   void Write_i(const Array& array);
+   void Write_i(const String& string);
+   void Write_i(const Number& number);
+   void Write_i(const Boolean& boolean);
+   void Write_i(const Null& null);
+   void Write_i(const UnknownElement& unknown);
+
+   virtual void Visit(const Array& array);
+   virtual void Visit(const Object& object);
+   virtual void Visit(const Number& number);
+   virtual void Visit(const String& string);
+   virtual void Visit(const Boolean& boolean);
+   virtual void Visit(const Null& null);
+
+   std::ostream& m_ostr;
+   int m_nTabDepth;
+};
+
+
+} // End namespace
+
+
+#include "writer.inl"
\ No newline at end of file
diff --git a/src/cajun/writer.inl b/src/cajun/writer.inl
new file mode 100644
index 0000000..71145da
--- /dev/null
+++ b/src/cajun/writer.inl
@@ -0,0 +1,178 @@
+/******************************************************************************
+
+Copyright (c) 2009-2010, Terry Caton
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+    * Redistributions of source code must retain the above copyright 
+      notice, this list of conditions and the following disclaimer.
+    * Redistributions in binary form must reproduce the above copyright
+      notice, this list of conditions and the following disclaimer in the
+      documentation and/or other materials provided with the distribution.
+    * Neither the name of the projecct nor the names of its contributors 
+      may be used to endorse or promote products derived from this software 
+      without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE 
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
+ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+#include "writer.h"
+#include <iostream>
+#include <iomanip>
+
+/*  
+
+TODO:
+* better documentation
+* unicode character encoding
+
+*/
+
+namespace json
+{
+
+
+inline void Writer::Write(const UnknownElement& elementRoot, std::ostream& ostr) { Write_i(elementRoot, ostr); }
+inline void Writer::Write(const Object& object, std::ostream& ostr)              { Write_i(object, ostr); }
+inline void Writer::Write(const Array& array, std::ostream& ostr)                { Write_i(array, ostr); }
+inline void Writer::Write(const Number& number, std::ostream& ostr)              { Write_i(number, ostr); }
+inline void Writer::Write(const String& string, std::ostream& ostr)              { Write_i(string, ostr); }
+inline void Writer::Write(const Boolean& boolean, std::ostream& ostr)            { Write_i(boolean, ostr); }
+inline void Writer::Write(const Null& null, std::ostream& ostr)                  { Write_i(null, ostr); }
+
+
+inline Writer::Writer(std::ostream& ostr) :
+   m_ostr(ostr),
+   m_nTabDepth(0)
+{}
+
+template <typename ElementTypeT>
+void Writer::Write_i(const ElementTypeT& element, std::ostream& ostr)
+{
+   Writer writer(ostr);
+   writer.Write_i(element);
+   ostr.flush(); // all done
+}
+
+inline void Writer::Write_i(const Array& array)
+{
+   if (array.Empty())
+      m_ostr << "[]";
+   else
+   {
+      m_ostr << '[' << std::endl;
+      ++m_nTabDepth;
+
+      Array::const_iterator it(array.Begin()),
+                            itEnd(array.End());
+      while (it != itEnd) {
+         m_ostr << std::string(m_nTabDepth, '\t');
+         
+         Write_i(*it);
+
+         if (++it != itEnd)
+            m_ostr << ',';
+         m_ostr << std::endl;
+      }
+
+      --m_nTabDepth;
+      m_ostr << std::string(m_nTabDepth, '\t') << ']';
+   }
+}
+
+inline void Writer::Write_i(const Object& object)
+{
+   if (object.Empty())
+      m_ostr << "{}";
+   else
+   {
+      m_ostr << '{' << std::endl;
+      ++m_nTabDepth;
+
+      Object::const_iterator it(object.Begin()),
+                             itEnd(object.End());
+      while (it != itEnd) {
+         m_ostr << std::string(m_nTabDepth, '\t');
+         
+         Write_i(it->name);
+
+         m_ostr << " : ";
+         Write_i(it->element); 
+
+         if (++it != itEnd)
+            m_ostr << ',';
+         m_ostr << std::endl;
+      }
+
+      --m_nTabDepth;
+      m_ostr << std::string(m_nTabDepth, '\t') << '}';
+   }
+}
+
+inline void Writer::Write_i(const Number& numberElement)
+{
+   m_ostr << std::setprecision(20) << numberElement.Value();
+}
+
+inline void Writer::Write_i(const Boolean& booleanElement)
+{
+   m_ostr << (booleanElement.Value() ? "true" : "false");
+}
+
+inline void Writer::Write_i(const String& stringElement)
+{
+   m_ostr << '"';
+
+   const std::string& s = stringElement.Value();
+   std::string::const_iterator it(s.begin()),
+                               itEnd(s.end());
+   for (; it != itEnd; ++it)
+   {
+      switch (*it)
+      {
+         case '"':         m_ostr << "\\\"";   break;
+         case '\\':        m_ostr << "\\\\";   break;
+         case '\b':        m_ostr << "\\b";    break;
+         case '\f':        m_ostr << "\\f";    break;
+         case '\n':        m_ostr << "\\n";    break;
+         case '\r':        m_ostr << "\\r";    break;
+         case '\t':        m_ostr << "\\t";    break;
+		 //case '\u':        m_ostr << "\\u";    break; // uh...
+         default:          m_ostr << *it;      break;
+      }
+   }
+
+   m_ostr << '"';   
+}
+
+inline void Writer::Write_i(const Null& )
+{
+   m_ostr << "null";
+}
+
+inline void Writer::Write_i(const UnknownElement& unknown)
+{
+   unknown.Accept(*this); 
+}
+
+inline void Writer::Visit(const Array& array)       { Write_i(array); }
+inline void Writer::Visit(const Object& object)     { Write_i(object); }
+inline void Writer::Visit(const Number& number)     { Write_i(number); }
+inline void Writer::Visit(const String& string)     { Write_i(string); }
+inline void Writer::Visit(const Boolean& boolean)   { Write_i(boolean); }
+inline void Writer::Visit(const Null& null)         { Write_i(null); }
+
+
+
+} // End namespace