osx regrouping

[wxWidgets.git] / include / wx / string.h

diff --git a/include/wx/string.h b/include/wx/string.h
index 346316a4928fe647a55f002f30f5bda8d69b1421..0a72bfe3206dacfb33d94b37044935577a9cfc1f 100644 (file)
--- a/include/wx/string.h
+++ b/include/wx/string.h
@@ -24,6 +24,7 @@
 
 #include "wx/defs.h"        // everybody should include this
 
 
 #include "wx/defs.h"        // everybody should include this
 

+#ifndef __WXPALMOS5__

 #if defined(__WXMAC__) || defined(__VISAGECPP__)
     #include <ctype.h>
 #endif
 #if defined(__WXMAC__) || defined(__VISAGECPP__)
     #include <ctype.h>
 #endif


@@ -46,12 +47,9 @@
 #ifdef HAVE_STRCASECMP_IN_STRINGS_H
     #include <strings.h>    // for strcasecmp()
 #endif // HAVE_STRCASECMP_IN_STRINGS_H
 #ifdef HAVE_STRCASECMP_IN_STRINGS_H
     #include <strings.h>    // for strcasecmp()
 #endif // HAVE_STRCASECMP_IN_STRINGS_H

+#endif // ! __WXPALMOS5__

 
 

-#ifdef __WXPALMOS__
-    #include <StringMgr.h>
-#endif
-
-#include "wx/wxcrt.h"       // for wxChar, wxStrlen() etc.
+#include "wx/wxcrtbase.h"   // for wxChar, wxStrlen() etc.

 #include "wx/strvararg.h"
 #include "wx/buffer.h"      // for wxCharBuffer
 #include "wx/strconv.h"     // for wxConvertXXX() macros and wxMBConv classes
 #include "wx/strvararg.h"
 #include "wx/buffer.h"      // for wxCharBuffer
 #include "wx/strconv.h"     // for wxConvertXXX() macros and wxMBConv classes


@@ -59,7 +57,7 @@
 #include "wx/stringops.h"
 #include "wx/unichar.h"
 
 #include "wx/stringops.h"
 #include "wx/unichar.h"
 

-class WXDLLIMPEXP_BASE wxString;
+class WXDLLIMPEXP_FWD_BASE wxString;

 
 // unless this symbol is predefined to disable the compatibility functions, do
 // use them
 
 // unless this symbol is predefined to disable the compatibility functions, do
 // use them


@@ -67,6 +65,11 @@ class WXDLLIMPEXP_BASE wxString;
     #define WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER 1
 #endif
 
     #define WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER 1
 #endif
 

+namespace wxPrivate
+{
+    template <typename T> struct wxStringAsBufHelper;
+}
+

 // ---------------------------------------------------------------------------
 // macros
 // ---------------------------------------------------------------------------
 // ---------------------------------------------------------------------------
 // macros
 // ---------------------------------------------------------------------------


@@ -125,8 +128,6 @@ inline int Stricmp(const char *psz1, const char *psz2)
   return _stricmp(psz1, psz2);
 #elif defined(__SC__)
   return _stricmp(psz1, psz2);
   return _stricmp(psz1, psz2);
 #elif defined(__SC__)
   return _stricmp(psz1, psz2);

-#elif defined(__SALFORDC__)
-  return stricmp(psz1, psz2);

 #elif defined(__BORLANDC__)
   return stricmp(psz1, psz2);
 #elif defined(__WATCOMC__)
 #elif defined(__BORLANDC__)
   return stricmp(psz1, psz2);
 #elif defined(__WATCOMC__)


@@ -201,7 +202,7 @@ public:
     const wchar_t* AsWChar() const;
     operator const wchar_t*() const { return AsWChar(); }
 
     const wchar_t* AsWChar() const;
     operator const wchar_t*() const { return AsWChar(); }
 

-#if !wxUSE_UNICODE
+#if !wxUSE_UNICODE || wxUSE_UTF8_LOCALE_ONLY

     inline
 #endif
     const char* AsChar() const;
     inline
 #endif
     const char* AsChar() const;


@@ -239,22 +240,12 @@ public:
     wxCStrData operator+(size_t n) const
         { return wxCStrData(m_str, m_offset + n, m_owned); }
 
     wxCStrData operator+(size_t n) const
         { return wxCStrData(m_str, m_offset + n, m_owned); }
 

-    // and these for "str.c_str() + n - 2":
-    wxCStrData operator-(int n) const
+    // and these for "str.c_str() + (p2 - p1)" (it also works for any integer
+    // expression but it must be ptrdiff_t and not e.g. int to work in this
+    // example):
+    wxCStrData operator-(ptrdiff_t n) const

     {
     {

-        wxASSERT_MSG( n <= (int)m_offset,
-                      _T("attempt to construct address before the beginning of the string") );
-        return wxCStrData(m_str, m_offset - n, m_owned);
-    }
-    wxCStrData operator-(long n) const
-    {
-        wxASSERT_MSG( n <= (int)m_offset,
-                      _T("attempt to construct address before the beginning of the string") );
-        return wxCStrData(m_str, m_offset - n, m_owned);
-    }
-    wxCStrData operator-(size_t n) const
-    {
-        wxASSERT_MSG( n <= m_offset,
+        wxASSERT_MSG( n <= (ptrdiff_t)m_offset,

                       _T("attempt to construct address before the beginning of the string") );
         return wxCStrData(m_str, m_offset - n, m_owned);
     }
                       _T("attempt to construct address before the beginning of the string") );
         return wxCStrData(m_str, m_offset - n, m_owned);
     }


@@ -268,7 +259,7 @@ private:
     size_t m_offset;
     bool m_owned;
 
     size_t m_offset;
     bool m_owned;
 

-    friend class WXDLLIMPEXP_BASE wxString;
+    friend class WXDLLIMPEXP_FWD_BASE wxString;

 };
 
 // ----------------------------------------------------------------------------
 };
 
 // ----------------------------------------------------------------------------


@@ -329,26 +320,28 @@ public:
 
     // static wxString Format(const wString& format, ...) ATTRIBUTE_PRINTF_1;
     WX_DEFINE_VARARG_FUNC_SANS_N0(static typename StringReturnType<T1>::type,
 
     // static wxString Format(const wString& format, ...) ATTRIBUTE_PRINTF_1;
     WX_DEFINE_VARARG_FUNC_SANS_N0(static typename StringReturnType<T1>::type,

-                                  Format, 1, (const wxString&),
+                                  Format, 1, (const wxFormatString&),

                                   DoFormatWchar, DoFormatUtf8)
     // We have to implement the version without template arguments manually
     // because of the StringReturnType<> hack, although WX_DEFINE_VARARG_FUNC
     // normally does it itself. It has to be a template so that we can use
                                   DoFormatWchar, DoFormatUtf8)
     // We have to implement the version without template arguments manually
     // because of the StringReturnType<> hack, although WX_DEFINE_VARARG_FUNC
     // normally does it itself. It has to be a template so that we can use

-    // the hack, even though there's no real template parameter:
-    struct FormatDummyArg {} ;
-
+    // the hack, even though there's no real template parameter. We can't move
+    // it to wxStrig, because it would shadow these versions of Format() then.

     template<typename T>
     inline static typename StringReturnType<T>::type
     template<typename T>
     inline static typename StringReturnType<T>::type

-    Format(const wxString& fmt, FormatDummyArg dummy = FormatDummyArg())
+    Format(const T& fmt)

     {
     {

-        return DoFormat(fmt);
+        // NB: this doesn't compile if T is not (some form of) a string;
+        //     this makes Format's prototype equivalent to
+        //     Format(const wxFormatString& fmt)
+        return DoFormatWchar(wxFormatString(fmt));

     }
 
     // int Printf(const wxString& format, ...);
     }
 
     // int Printf(const wxString& format, ...);

-    WX_DEFINE_VARARG_FUNC(int, Printf, 1, (const wxString&),
+    WX_DEFINE_VARARG_FUNC(int, Printf, 1, (const wxFormatString&),

                           DoPrintfWchar, DoPrintfUtf8)
     // int sprintf(const wxString& format, ...) ATTRIBUTE_PRINTF_2;
                           DoPrintfWchar, DoPrintfUtf8)
     // int sprintf(const wxString& format, ...) ATTRIBUTE_PRINTF_2;

-    WX_DEFINE_VARARG_FUNC(int, sprintf, 1, (const wxString&),
+    WX_DEFINE_VARARG_FUNC(int, sprintf, 1, (const wxFormatString&),

                           DoPrintfWchar, DoPrintfUtf8)
 
 protected:
                           DoPrintfWchar, DoPrintfUtf8)
 
 protected:


@@ -368,6 +361,43 @@ protected:
     #pragma warning (disable:4275)
 #endif
 
     #pragma warning (disable:4275)
 #endif
 

+#if wxUSE_UNICODE_UTF8
+// see the comment near wxString::iterator for why we need this
+class WXDLLIMPEXP_BASE wxStringIteratorNode
+{
+public:
+    wxStringIteratorNode()
+        : m_str(NULL), m_citer(NULL), m_iter(NULL), m_prev(NULL), m_next(NULL) {}
+    wxStringIteratorNode(const wxString *str,
+                          wxStringImpl::const_iterator *citer)
+        { DoSet(str, citer, NULL); }
+    wxStringIteratorNode(const wxString *str, wxStringImpl::iterator *iter)
+        { DoSet(str, NULL, iter); }
+    ~wxStringIteratorNode()
+        { clear(); }
+
+    inline void set(const wxString *str, wxStringImpl::const_iterator *citer)
+        { clear(); DoSet(str, citer, NULL); }
+    inline void set(const wxString *str, wxStringImpl::iterator *iter)
+        { clear(); DoSet(str, NULL, iter); }
+
+    const wxString *m_str;
+    wxStringImpl::const_iterator *m_citer;
+    wxStringImpl::iterator *m_iter;
+    wxStringIteratorNode *m_prev, *m_next;
+
+private:
+    inline void clear();
+    inline void DoSet(const wxString *str,
+                      wxStringImpl::const_iterator *citer,
+                      wxStringImpl::iterator *iter);
+
+    // the node belongs to a particular iterator instance, it's not copied
+    // when a copy of the iterator is made
+    DECLARE_NO_COPY_CLASS(wxStringIteratorNode)
+};
+#endif // wxUSE_UNICODE_UTF8
+

 class WXDLLIMPEXP_BASE wxString
 #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN
                                 : public wxStringPrintfMixin
 class WXDLLIMPEXP_BASE wxString
 #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN
                                 : public wxStringPrintfMixin


@@ -406,7 +436,10 @@ private:
       size_t len;
 
       SubstrBufFromType(const T& data_, size_t len_)
       size_t len;
 
       SubstrBufFromType(const T& data_, size_t len_)

-          : data(data_), len(len_) {}
+          : data(data_), len(len_)
+      {
+          wxASSERT_MSG( len != npos, "must have real length" );
+      }

   };
 
 #if wxUSE_UNICODE_UTF8
   };
 
 #if wxUSE_UNICODE_UTF8


@@ -477,9 +510,6 @@ private:
 
 #else // wxUSE_UNICODE_UTF8
 
 
 #else // wxUSE_UNICODE_UTF8
 

-  // FIXME-UTF8: return as-is without copying under UTF8 locale, return
-  //             converted string under other locales - needs wxCharBuffer
-  //             changes

   static wxCharBuffer ImplStr(const char* str,
                               const wxMBConv& conv = wxConvLibc)
     { return ConvertStr(str, npos, conv).data; }
   static wxCharBuffer ImplStr(const char* str,
                               const wxMBConv& conv = wxConvLibc)
     { return ConvertStr(str, npos, conv).data; }


@@ -488,9 +518,9 @@ private:
     { return ConvertStr(str, n, conv); }
 
   static wxCharBuffer ImplStr(const wchar_t* str)
     { return ConvertStr(str, n, conv); }
 
   static wxCharBuffer ImplStr(const wchar_t* str)

-    { return ConvertStr(str, npos, wxConvUTF8).data; }
+    { return ConvertStr(str, npos, wxMBConvUTF8()).data; }

   static SubstrBufFromWC ImplStr(const wchar_t* str, size_t n)
   static SubstrBufFromWC ImplStr(const wchar_t* str, size_t n)

-    { return ConvertStr(str, n, wxConvUTF8); }
+    { return ConvertStr(str, n, wxMBConvUTF8()); }

 
   size_t PosToImpl(size_t pos) const
   {
 
   size_t PosToImpl(size_t pos) const
   {


@@ -514,7 +544,7 @@ private:
       if ( pos == 0 || pos == npos )
           return pos;
       else
       if ( pos == 0 || pos == npos )
           return pos;
       else

-          return const_iterator(m_impl.begin() + pos) - begin();
+          return const_iterator(this, m_impl.begin() + pos) - begin();

   }
 #endif // !wxUSE_UNICODE_UTF8/wxUSE_UNICODE_UTF8
 
   }
 #endif // !wxUSE_UNICODE_UTF8/wxUSE_UNICODE_UTF8
 


@@ -540,8 +570,7 @@ public:
   #define WX_STR_ITERATOR_TAG void /* dummy type */
 #endif
 
   #define WX_STR_ITERATOR_TAG void /* dummy type */
 #endif
 

-  #define WX_STR_ITERATOR_IMPL(iterator_name, pointer_type,                 \
-                               reference_type, reference_ctor)              \
+  #define WX_STR_ITERATOR_IMPL(iterator_name, pointer_type, reference_type) \

       private:                                                              \
           typedef wxStringImpl::iterator_name underlying_iterator;          \
       public:                                                               \
       private:                                                              \
           typedef wxStringImpl::iterator_name underlying_iterator;          \
       public:                                                               \


@@ -551,7 +580,6 @@ public:
           typedef reference_type reference;                                 \
           typedef pointer_type pointer;                                     \
                                                                             \
           typedef reference_type reference;                                 \
           typedef pointer_type pointer;                                     \
                                                                             \

-          reference operator*() const { return reference_ctor; }            \

           reference operator[](size_t n) const { return *(*this + n); }     \
                                                                             \
           iterator_name& operator++()                                       \
           reference operator[](size_t n) const { return *(*this + n); }     \
                                                                             \
           iterator_name& operator++()                                       \


@@ -571,25 +599,15 @@ public:
               return tmp;                                                   \
           }                                                                 \
                                                                             \
               return tmp;                                                   \
           }                                                                 \
                                                                             \

-          iterator_name& operator+=(int n)                                  \
+          iterator_name& operator+=(ptrdiff_t n)                            \

           {                                                                 \
               m_cur = wxStringOperations::AddToIter(m_cur, n);              \
               return *this;                                                 \
           }                                                                 \
           {                                                                 \
               m_cur = wxStringOperations::AddToIter(m_cur, n);              \
               return *this;                                                 \
           }                                                                 \

-          iterator_name& operator+=(size_t n)                               \
-          {                                                                 \
-              m_cur = wxStringOperations::AddToIter(m_cur, (int)n);         \
-              return *this;                                                 \
-          }                                                                 \
-          iterator_name& operator-=(int n)                                  \
+          iterator_name& operator-=(ptrdiff_t n)                            \

           {                                                                 \
               m_cur = wxStringOperations::AddToIter(m_cur, -n);             \
               return *this;                                                 \
           {                                                                 \
               m_cur = wxStringOperations::AddToIter(m_cur, -n);             \
               return *this;                                                 \

-          }                                                                 \
-          iterator_name& operator-=(size_t n)                               \
-          {                                                                 \
-              m_cur = wxStringOperations::AddToIter(m_cur, -(int)n);        \
-              return *this;                                                 \

           }                                                                 \
                                                                             \
           difference_type operator-(const iterator_name& i) const           \
           }                                                                 \
                                                                             \
           difference_type operator-(const iterator_name& i) const           \


@@ -613,70 +631,129 @@ public:
           /* for internal wxString use only: */                             \
           underlying_iterator impl() const { return m_cur; }                \
                                                                             \
           /* for internal wxString use only: */                             \
           underlying_iterator impl() const { return m_cur; }                \
                                                                             \

-          friend class WXDLLIMPEXP_BASE wxString;                           \
-          friend class WXDLLIMPEXP_BASE wxCStrData;                         \
+          friend class wxString;                                            \
+          friend class wxCStrData;                                          \

                                                                             \
       private:                                                              \
           underlying_iterator m_cur
 
                                                                             \
       private:                                                              \
           underlying_iterator m_cur
 

-  class const_iterator;
+  class WXDLLIMPEXP_FWD_BASE const_iterator;

 
 #if wxUSE_UNICODE_UTF8
 
 #if wxUSE_UNICODE_UTF8

-  class iterator
-  {
-      // NB: In UTF-8 build, (non-const) iterator needs to keep reference
-      //     to the underlying wxStringImpl, because UTF-8 is variable-length
-      //     encoding and changing the value pointer to by an iterator using
-      //     its operator* requires calling wxStringImpl::replace() if the old
-      //     and new values differ in their encoding's length.
+  // NB: In UTF-8 build, (non-const) iterator needs to keep reference
+  //     to the underlying wxStringImpl, because UTF-8 is variable-length
+  //     encoding and changing the value pointer to by an iterator (using
+  //     its operator*) requires calling wxStringImpl::replace() if the old
+  //     and new values differ in their encoding's length.
+  //
+  //     Furthermore, the replace() call may invalid all iterators for the
+  //     string, so we have to keep track of outstanding iterators and update
+  //     them if replace() happens.
+  //
+  //     This is implemented by maintaining linked list of iterators for every
+  //     string and traversing it in wxUniCharRef::operator=(). Head of the
+  //     list is stored in wxString. (FIXME-UTF8)

 
 

-      WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef,
-                           wxUniCharRef::CreateForString(m_str, m_cur));
+  class WXDLLIMPEXP_BASE iterator
+  {
+      WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef);

 
   public:
 
   public:

-      iterator(const iterator& i) : m_cur(i.m_cur), m_str(i.m_str) {}
+      iterator() {}
+      iterator(const iterator& i)
+          : m_cur(i.m_cur), m_node(i.str(), &m_cur) {}
+      iterator& operator=(const iterator& i)
+      {
+          if (&i != this)
+          {
+              m_cur = i.m_cur;
+              m_node.set(i.str(), &m_cur);
+          }
+          return *this;
+      }
+
+      reference operator*()
+        { return wxUniCharRef::CreateForString(m_node, m_cur); }

 
 

-      iterator operator+(int n) const
-        { return iterator(m_str, wxStringOperations::AddToIter(m_cur, n)); }
-      iterator operator+(size_t n) const
-        { return iterator(m_str, wxStringOperations::AddToIter(m_cur, (int)n)); }
-      iterator operator-(int n) const
-        { return iterator(m_str, wxStringOperations::AddToIter(m_cur, -n)); }
-      iterator operator-(size_t n) const
-        { return iterator(m_str, wxStringOperations::AddToIter(m_cur, -(int)n)); }
+      iterator operator+(ptrdiff_t n) const
+        { return iterator(str(), wxStringOperations::AddToIter(m_cur, n)); }
+      iterator operator-(ptrdiff_t n) const
+        { return iterator(str(), wxStringOperations::AddToIter(m_cur, -n)); }

 
   private:
       iterator(wxString *str, underlying_iterator ptr)
 
   private:
       iterator(wxString *str, underlying_iterator ptr)

-          : m_cur(ptr), m_str(str->m_impl) {}
-      iterator(wxStringImpl& str, underlying_iterator ptr)
-          : m_cur(ptr), m_str(str) {}
+          : m_cur(ptr), m_node(str, &m_cur) {}

 
 

-      wxStringImpl& m_str;
+      wxString* str() const { return wx_const_cast(wxString*, m_node.m_str); }
+
+      wxStringIteratorNode m_node;

 
       friend class const_iterator;
   };
 
 
       friend class const_iterator;
   };
 

+  class WXDLLIMPEXP_BASE const_iterator
+  {
+      // NB: reference_type is intentionally value, not reference, the character
+      //     may be encoded differently in wxString data:
+      WX_STR_ITERATOR_IMPL(const_iterator, const wxChar*, wxUniChar);
+
+  public:
+      const_iterator() {}
+      const_iterator(const const_iterator& i)
+          : m_cur(i.m_cur), m_node(i.str(), &m_cur) {}
+      const_iterator(const iterator& i)
+          : m_cur(i.m_cur), m_node(i.str(), &m_cur) {}
+
+      const_iterator& operator=(const const_iterator& i)
+      {
+          if (&i != this)
+          {
+              m_cur = i.m_cur;
+              m_node.set(i.str(), &m_cur);
+          }
+          return *this;
+      }
+      const_iterator& operator=(const iterator& i)
+        { m_cur = i.m_cur; m_node.set(i.str(), &m_cur); return *this; }
+
+      reference operator*() const
+        { return wxStringOperations::DecodeChar(m_cur); }
+
+      const_iterator operator+(ptrdiff_t n) const
+        { return const_iterator(str(), wxStringOperations::AddToIter(m_cur, n)); }
+      const_iterator operator-(ptrdiff_t n) const
+        { return const_iterator(str(), wxStringOperations::AddToIter(m_cur, -n)); }
+
+  private:
+      // for internal wxString use only:
+      const_iterator(const wxString *str, underlying_iterator ptr)
+          : m_cur(ptr), m_node(str, &m_cur) {}
+
+      const wxString* str() const { return m_node.m_str; }
+
+      wxStringIteratorNode m_node;
+  };
+

   size_t IterToImplPos(wxString::iterator i) const
     { return wxStringImpl::const_iterator(i.impl()) - m_impl.begin(); }
 
 #else // !wxUSE_UNICODE_UTF8
 
   size_t IterToImplPos(wxString::iterator i) const
     { return wxStringImpl::const_iterator(i.impl()) - m_impl.begin(); }
 
 #else // !wxUSE_UNICODE_UTF8
 

-  class iterator
+  class WXDLLIMPEXP_BASE iterator

   {
   {

-      WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef,
-                           wxUniCharRef::CreateForString(m_cur));
+      WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef);

 
   public:
 
   public:

+      iterator() {}

       iterator(const iterator& i) : m_cur(i.m_cur) {}
 
       iterator(const iterator& i) : m_cur(i.m_cur) {}
 

-      iterator operator+(int n) const
+      reference operator*()
+        { return wxUniCharRef::CreateForString(m_cur); }
+
+      iterator operator+(ptrdiff_t n) const

         { return iterator(wxStringOperations::AddToIter(m_cur, n)); }
         { return iterator(wxStringOperations::AddToIter(m_cur, n)); }

-      iterator operator+(size_t n) const
-        { return iterator(wxStringOperations::AddToIter(m_cur, (int)n)); }
-      iterator operator-(int n) const
+      iterator operator-(ptrdiff_t n) const

         { return iterator(wxStringOperations::AddToIter(m_cur, -n)); }
         { return iterator(wxStringOperations::AddToIter(m_cur, -n)); }

-      iterator operator-(size_t n) const
-        { return iterator(wxStringOperations::AddToIter(m_cur, -(int)n)); }

 
   private:
       // for internal wxString use only:
 
   private:
       // for internal wxString use only:


@@ -685,32 +762,33 @@ public:
 
       friend class const_iterator;
   };
 
       friend class const_iterator;
   };

-#endif // wxUSE_UNICODE_UTF8/!wxUSE_UNICODE_UTF8

 
 

-  class const_iterator
+  class WXDLLIMPEXP_BASE const_iterator

   {
       // NB: reference_type is intentionally value, not reference, the character
       //     may be encoded differently in wxString data:
   {
       // NB: reference_type is intentionally value, not reference, the character
       //     may be encoded differently in wxString data:

-      WX_STR_ITERATOR_IMPL(const_iterator, const wxChar*, wxUniChar,
-                           wxStringOperations::DecodeChar(m_cur));
+      WX_STR_ITERATOR_IMPL(const_iterator, const wxChar*, wxUniChar);

 
   public:
 
   public:

+      const_iterator() {}

       const_iterator(const const_iterator& i) : m_cur(i.m_cur) {}
       const_iterator(const iterator& i) : m_cur(i.m_cur) {}
 
       const_iterator(const const_iterator& i) : m_cur(i.m_cur) {}
       const_iterator(const iterator& i) : m_cur(i.m_cur) {}
 

-      const_iterator operator+(int n) const
+      reference operator*() const
+        { return wxStringOperations::DecodeChar(m_cur); }
+
+      const_iterator operator+(ptrdiff_t n) const

         { return const_iterator(wxStringOperations::AddToIter(m_cur, n)); }
         { return const_iterator(wxStringOperations::AddToIter(m_cur, n)); }

-      const_iterator operator+(size_t n) const
-        { return const_iterator(wxStringOperations::AddToIter(m_cur, (int)n)); }
-      const_iterator operator-(int n) const
+      const_iterator operator-(ptrdiff_t n) const

         { return const_iterator(wxStringOperations::AddToIter(m_cur, -n)); }
         { return const_iterator(wxStringOperations::AddToIter(m_cur, -n)); }

-      const_iterator operator-(size_t n) const
-        { return const_iterator(wxStringOperations::AddToIter(m_cur, -(int)n)); }

 
   private:
       // for internal wxString use only:
       const_iterator(underlying_iterator ptr) : m_cur(ptr) {}
 
   private:
       // for internal wxString use only:
       const_iterator(underlying_iterator ptr) : m_cur(ptr) {}

+      const_iterator(const wxString *WXUNUSED(str), underlying_iterator ptr)
+          : m_cur(ptr) {}

   };
   };

+#endif // wxUSE_UNICODE_UTF8/!wxUSE_UNICODE_UTF8

 
   #undef WX_STR_ITERATOR_TAG
   #undef WX_STR_ITERATOR_IMPL
 
   #undef WX_STR_ITERATOR_TAG
   #undef WX_STR_ITERATOR_IMPL


@@ -730,6 +808,7 @@ public:
       typedef typename T::reference reference;
       typedef typename T::pointer *pointer;
 
       typedef typename T::reference reference;
       typedef typename T::pointer *pointer;
 

+      reverse_iterator_impl() {}

       reverse_iterator_impl(iterator_type i) : m_cur(i) {}
       reverse_iterator_impl(const reverse_iterator_impl& ri)
           : m_cur(ri.m_cur) {}
       reverse_iterator_impl(iterator_type i) : m_cur(i) {}
       reverse_iterator_impl(const reverse_iterator_impl& ri)
           : m_cur(ri.m_cur) {}


@@ -749,21 +828,13 @@ public:
         { reverse_iterator_impl tmp = *this; ++m_cur; return tmp; }
 
       // NB: explicit <T> in the functions below is to keep BCC 5.5 happy
         { reverse_iterator_impl tmp = *this; ++m_cur; return tmp; }
 
       // NB: explicit <T> in the functions below is to keep BCC 5.5 happy

-      reverse_iterator_impl operator+(int n) const
+      reverse_iterator_impl operator+(ptrdiff_t n) const

         { return reverse_iterator_impl<T>(m_cur - n); }
         { return reverse_iterator_impl<T>(m_cur - n); }

-      reverse_iterator_impl operator+(size_t n) const
-        { return reverse_iterator_impl<T>(m_cur - n); }
-      reverse_iterator_impl operator-(int n) const
-        { return reverse_iterator_impl<T>(m_cur + n); }
-      reverse_iterator_impl operator-(size_t n) const
+      reverse_iterator_impl operator-(ptrdiff_t n) const

         { return reverse_iterator_impl<T>(m_cur + n); }
         { return reverse_iterator_impl<T>(m_cur + n); }

-      reverse_iterator_impl operator+=(int n)
+      reverse_iterator_impl operator+=(ptrdiff_t n)

         { m_cur -= n; return *this; }
         { m_cur -= n; return *this; }

-      reverse_iterator_impl operator+=(size_t n)
-        { m_cur -= n; return *this; }
-      reverse_iterator_impl operator-=(int n)
-        { m_cur += n; return *this; }
-      reverse_iterator_impl operator-=(size_t n)
+      reverse_iterator_impl operator-=(ptrdiff_t n)

         { m_cur += n; return *this; }
 
       unsigned operator-(const reverse_iterator_impl& i) const
         { m_cur += n; return *this; }
 
       unsigned operator-(const reverse_iterator_impl& i) const


@@ -795,7 +866,8 @@ private:
   // wxCStrData) to an iterator into the string
   static const_iterator CreateConstIterator(const wxCStrData& data)
   {
   // wxCStrData) to an iterator into the string
   static const_iterator CreateConstIterator(const wxCStrData& data)
   {

-      return const_iterator(data.m_str->begin() + data.m_offset);
+      return const_iterator(data.m_str,
+                            (data.m_str->begin() + data.m_offset).impl());

   }
 
   // in UTF-8 STL build, creation from std::string requires conversion under
   }
 
   // in UTF-8 STL build, creation from std::string requires conversion under


@@ -828,7 +900,7 @@ public:
   wxString(const wxString& stringSrc) : m_impl(stringSrc.m_impl) { }
 
     // string containing nRepeat copies of ch
   wxString(const wxString& stringSrc) : m_impl(stringSrc.m_impl) { }
 
     // string containing nRepeat copies of ch

-  wxString(wxUniChar ch, size_t nRepeat = 1)
+  wxString(wxUniChar ch, size_t nRepeat = 1 )

     { assign(nRepeat, ch); }
   wxString(size_t nRepeat, wxUniChar ch)
     { assign(nRepeat, ch); }
     { assign(nRepeat, ch); }
   wxString(size_t nRepeat, wxUniChar ch)
     { assign(nRepeat, ch); }


@@ -886,8 +958,11 @@ public:
   wxString(const wxWCharBuffer& buf)
     { assign(buf.data()); } // FIXME-UTF8: fix for embedded NUL and buffer length
 
   wxString(const wxWCharBuffer& buf)
     { assign(buf.data()); } // FIXME-UTF8: fix for embedded NUL and buffer length
 

+    // NB: this version uses m_impl.c_str() to force making a copy of the
+    //     string, so that "wxString(str.c_str())" idiom for passing strings
+    //     between threads works

   wxString(const wxCStrData& cstr)
   wxString(const wxCStrData& cstr)

-      : m_impl(cstr.AsString().m_impl) { }
+      : m_impl(cstr.AsString().m_impl.c_str()) { }

 
     // as we provide both ctors with this signature for both char and unsigned
     // char string, we need to provide one for wxCStrData to resolve ambiguity
 
     // as we provide both ctors with this signature for both char and unsigned
     // char string, we need to provide one for wxCStrData to resolve ambiguity


@@ -897,7 +972,7 @@ public:
     // and because wxString is convertible to wxCStrData and const wxChar *
     // we also need to provide this one
   wxString(const wxString& str, size_t nLength)
     // and because wxString is convertible to wxCStrData and const wxChar *
     // we also need to provide this one
   wxString(const wxString& str, size_t nLength)

-      : m_impl(str.Mid(0, nLength).m_impl) {}
+    { assign(str, nLength); }

 
   // even if we're not built with wxUSE_STL == 1 it is very convenient to allow
   // implicit conversions from std::string to wxString and vice verse as this
 
   // even if we're not built with wxUSE_STL == 1 it is very convenient to allow
   // implicit conversions from std::string to wxString and vice verse as this


@@ -920,7 +995,12 @@ public:
     wxString(const std::string& str)
         { assign(str.c_str(), str.length()); }
   #endif
     wxString(const std::string& str)
         { assign(str.c_str(), str.length()); }
   #endif

+#endif // wxUSE_STD_STRING

 
 

+  // Unlike ctor from std::string, we provide conversion to std::string only
+  // if wxUSE_STL and not merely wxUSE_STD_STRING (which is on by default),
+  // because it conflicts with operator const char/wchar_t*:
+#if wxUSE_STL

   #if wxUSE_UNICODE_WCHAR && wxUSE_STL_BASED_WXSTRING
     // wxStringImpl is std::string in the encoding we want
     operator const wxStdWideString&() const { return m_impl; }
   #if wxUSE_UNICODE_WCHAR && wxUSE_STL_BASED_WXSTRING
     // wxStringImpl is std::string in the encoding we want
     operator const wxStdWideString&() const { return m_impl; }


@@ -931,8 +1011,7 @@ public:
         { return wxStdWideString(wc_str()); }
   #endif
 
         { return wxStdWideString(wc_str()); }
   #endif
 

-  #if !wxUSE_UNICODE && wxUSE_STL_BASED_WXSTRING
-    // FIXME-UTF8: do this in UTF8 build #if wxUSE_UTF8_LOCALE_ONLY, too
+  #if (!wxUSE_UNICODE || wxUSE_UTF8_LOCALE_ONLY) && wxUSE_STL_BASED_WXSTRING

     // wxStringImpl is std::string in the encoding we want
     operator const std::string&() const { return m_impl; }
   #else
     // wxStringImpl is std::string in the encoding we want
     operator const std::string&() const { return m_impl; }
   #else


@@ -941,14 +1020,20 @@ public:
         // FIXME-UTF8: broken for embedded NULs
         { return std::string(mb_str()); }
   #endif
         // FIXME-UTF8: broken for embedded NULs
         { return std::string(mb_str()); }
   #endif

+#endif // wxUSE_STL

 
 

-#endif // wxUSE_STD_STRING
+  wxString Clone() const
+  {
+      // make a deep copy of the string, i.e. the returned string will have
+      // ref count = 1 with refcounted implementation
+      return wxString::FromImpl(wxStringImpl(m_impl.c_str(), m_impl.length()));
+  }

 
   // first valid index position
 
   // first valid index position

-  const_iterator begin() const { return const_iterator(m_impl.begin()); }
+  const_iterator begin() const { return const_iterator(this, m_impl.begin()); }

   iterator begin() { return iterator(this, m_impl.begin()); }
   // position one after the last valid one
   iterator begin() { return iterator(this, m_impl.begin()); }
   // position one after the last valid one

-  const_iterator end() const { return const_iterator(m_impl.end()); }
+  const_iterator end() const { return const_iterator(this, m_impl.end()); }

   iterator end() { return iterator(this, m_impl.end()); }
 
   // first element of the reversed string
   iterator end() { return iterator(this, m_impl.end()); }
 
   // first element of the reversed string


@@ -979,19 +1064,25 @@ public:
 
   void resize(size_t nSize, wxUniChar ch = wxT('\0'))
   {
 
   void resize(size_t nSize, wxUniChar ch = wxT('\0'))
   {

+    const size_t len = length();
+    if ( nSize == len)
+        return;
+

 #if wxUSE_UNICODE_UTF8
 #if wxUSE_UNICODE_UTF8

-    if ( !ch.IsAscii() )
+    if ( nSize < len )

     {
     {

-        size_t len = length();
-        if ( nSize == len)
-            return;
-        else if ( nSize < len )
-            erase(nSize);
-        else
-            append(nSize - len, ch);
+        // we can't use wxStringImpl::resize() for truncating the string as it
+        // counts in bytes, not characters
+        erase(nSize);
+        return;

     }
     }

+
+    // we also can't use (presumably more efficient) resize() if we have to
+    // append characters taking more than one byte
+    if ( !ch.IsAscii() )
+        append(nSize - len, ch);

     else
     else

-#endif
+#endif // wxUSE_UNICODE_UTF8

         m_impl.resize(nSize, (wxStringCharType)ch);
   }
 
         m_impl.resize(nSize, (wxStringCharType)ch);
   }
 


@@ -1050,18 +1141,17 @@ public:
 
     // get last character
     wxUniChar Last() const
 
     // get last character
     wxUniChar Last() const

-      {
-          wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") );
-
-          return at(length() - 1);
-      }
+    {
+      wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") );
+      return *rbegin();
+    }

 
     // get writable last character
     wxUniCharRef Last()
 
     // get writable last character
     wxUniCharRef Last()

-      {
-          wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") );
-          return at(length() - 1);
-      }
+    {
+      wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") );
+      return *rbegin();
+    }

 
     /*
        Note that we we must define all of the overloads below to avoid
 
     /*
        Note that we we must define all of the overloads below to avoid


@@ -1096,6 +1186,13 @@ public:
 
     // implicit conversion to C string
     operator wxCStrData() const { return c_str(); }
 
     // implicit conversion to C string
     operator wxCStrData() const { return c_str(); }

+
+    // the first two operators conflict with operators for conversion to
+    // std::string and they must be disabled in STL build; the next one only
+    // makes sense if conversions to char* are also defined and not defining it
+    // in STL build also helps us to get more clear error messages for the code
+    // which relies on implicit conversion to char* in STL build
+#if !wxUSE_STL

     operator const char*() const { return c_str(); }
     operator const wchar_t*() const { return c_str(); }
 
     operator const char*() const { return c_str(); }
     operator const wchar_t*() const { return c_str(); }
 


@@ -1103,6 +1200,7 @@ public:
     // wxWidgets versions: this is the same as conversion to const char * so it
     // may fail!
     operator const void*() const { return c_str(); }
     // wxWidgets versions: this is the same as conversion to const char * so it
     // may fail!
     operator const void*() const { return c_str(); }

+#endif // wxUSE_STL

 
     // identical to c_str(), for MFC compatibility
     const wxCStrData GetData() const { return c_str(); }
 
     // identical to c_str(), for MFC compatibility
     const wxCStrData GetData() const { return c_str(); }


@@ -1118,6 +1216,30 @@ public:
         { return mb_str(conv); }
     wxWritableWCharBuffer wchar_str() const { return wc_str(); }
 
         { return mb_str(conv); }
     wxWritableWCharBuffer wchar_str() const { return wc_str(); }
 

+    // conversion to the buffer of the given type T (= char or wchar_t) and
+    // also optionally return the buffer length
+    //
+    // this is mostly/only useful for the template functions
+    //
+    // FIXME-VC6: the second argument only exists for VC6 which doesn't support
+    //            explicit template function selection, do not use it unless
+    //            you must support VC6!
+    template <typename T>
+    wxCharTypeBuffer<T> tchar_str(size_t *len = NULL,
+                                  T * WXUNUSED(dummy) = NULL) const
+    {
+#if wxUSE_UNICODE
+        // we need a helper dispatcher depending on type
+        return wxPrivate::wxStringAsBufHelper<T>::Get(*this, len);
+#else // ANSI
+        // T can only be char in ANSI build
+        if ( len )
+            *len = length();
+
+        return wxCharTypeBuffer<T>::CreateNonOwned(wx_str());
+#endif // Unicode build kind
+    }
+

     // conversion to/from plain (i.e. 7 bit) ASCII: this is useful for
     // converting numbers or strings which are certain not to contain special
     // chars (typically system functions, X atoms, environment variables etc.)
     // conversion to/from plain (i.e. 7 bit) ASCII: this is useful for
     // converting numbers or strings which are certain not to contain special
     // chars (typically system functions, X atoms, environment variables etc.)


@@ -1125,15 +1247,107 @@ public:
     // the behaviour of these functions with the strings containing anything
     // else than 7 bit ASCII characters is undefined, use at your own risk.
 #if wxUSE_UNICODE
     // the behaviour of these functions with the strings containing anything
     // else than 7 bit ASCII characters is undefined, use at your own risk.
 #if wxUSE_UNICODE

-    static wxString FromAscii(const char *ascii);  // string
-    static wxString FromAscii(const char ascii);   // char
+    static wxString FromAscii(const char *ascii, size_t len);
+    static wxString FromAscii(const char *ascii);
+    static wxString FromAscii(char ascii);

     const wxCharBuffer ToAscii() const;
 #else // ANSI
     static wxString FromAscii(const char *ascii) { return wxString( ascii ); }
     const wxCharBuffer ToAscii() const;
 #else // ANSI
     static wxString FromAscii(const char *ascii) { return wxString( ascii ); }

-    static wxString FromAscii(const char ascii) { return wxString( ascii ); }
+    static wxString FromAscii(const char *ascii, size_t len)
+        { return wxString( ascii, len ); }
+    static wxString FromAscii(char ascii) { return wxString( ascii ); }

     const char *ToAscii() const { return c_str(); }
 #endif // Unicode/!Unicode
 
     const char *ToAscii() const { return c_str(); }
 #endif // Unicode/!Unicode
 

+    // also provide unsigned char overloads as signed/unsigned doesn't matter
+    // for 7 bit ASCII characters
+    static wxString FromAscii(const unsigned char *ascii)
+        { return FromAscii((const char *)ascii); }
+    static wxString FromAscii(const unsigned char *ascii, size_t len)
+        { return FromAscii((const char *)ascii, len); }
+
+    // conversion to/from UTF-8:
+#if wxUSE_UNICODE_UTF8
+    static wxString FromUTF8Unchecked(const char *utf8)
+    {
+      if ( !utf8 )
+          return wxEmptyString;
+
+      wxASSERT( wxStringOperations::IsValidUtf8String(utf8) );
+      return FromImpl(wxStringImpl(utf8));
+    }
+    static wxString FromUTF8Unchecked(const char *utf8, size_t len)
+    {
+      if ( !utf8 )
+          return wxEmptyString;
+      if ( len == npos )
+          return FromUTF8Unchecked(utf8);
+
+      wxASSERT( wxStringOperations::IsValidUtf8String(utf8, len) );
+      return FromImpl(wxStringImpl(utf8, len));
+    }
+
+    static wxString FromUTF8(const char *utf8)
+    {
+        if ( !utf8 || !wxStringOperations::IsValidUtf8String(utf8) )
+            return "";
+
+        return FromImpl(wxStringImpl(utf8));
+    }
+    static wxString FromUTF8(const char *utf8, size_t len)
+    {
+        if ( len == npos )
+            return FromUTF8(utf8);
+
+        if ( !utf8 || !wxStringOperations::IsValidUtf8String(utf8, len) )
+            return "";
+
+        return FromImpl(wxStringImpl(utf8, len));
+    }
+
+    const char* utf8_str() const { return wx_str(); }
+    const char* ToUTF8() const { return wx_str(); }
+
+    // this function exists in UTF-8 build only and returns the length of the
+    // internal UTF-8 representation
+    size_t utf8_length() const { return m_impl.length(); }
+#elif wxUSE_UNICODE_WCHAR
+    static wxString FromUTF8(const char *utf8, size_t len = npos)
+      { return wxString(utf8, wxMBConvUTF8(), len); }
+    static wxString FromUTF8Unchecked(const char *utf8, size_t len = npos)
+    {
+        const wxString s(utf8, wxMBConvUTF8(), len);
+        wxASSERT_MSG( !utf8 || !*utf8 || !s.empty(),
+                      "string must be valid UTF-8" );
+        return s;
+    }
+    const wxCharBuffer utf8_str() const { return mb_str(wxMBConvUTF8()); }
+    const wxCharBuffer ToUTF8() const { return utf8_str(); }
+#else // ANSI
+    static wxString FromUTF8(const char *utf8)
+      { return wxString(wxMBConvUTF8().cMB2WC(utf8)); }
+    static wxString FromUTF8(const char *utf8, size_t len)
+    {
+        size_t wlen;
+        wxWCharBuffer buf(wxMBConvUTF8().cMB2WC(utf8, len == npos ? wxNO_LEN : len, &wlen));
+        return wxString(buf.data(), wlen);
+    }
+    static wxString FromUTF8Unchecked(const char *utf8, size_t len = npos)
+    {
+        size_t wlen;
+        wxWCharBuffer buf(wxMBConvUTF8().cMB2WC(utf8,
+                                                len == npos ? wxNO_LEN : len,
+                                                &wlen));
+        wxASSERT_MSG( !utf8 || !*utf8 || wlen,
+                      "string must be valid UTF-8" );
+
+        return wxString(buf.data(), wlen);
+    }
+    const wxCharBuffer utf8_str() const
+      { return wxMBConvUTF8().cWC2MB(wc_str()); }
+    const wxCharBuffer ToUTF8() const { return utf8_str(); }
+#endif
+

     // functions for storing binary data in wxString:
 #if wxUSE_UNICODE
     static wxString From8BitData(const char *data, size_t len)
     // functions for storing binary data in wxString:
 #if wxUSE_UNICODE
     static wxString From8BitData(const char *data, size_t len)


@@ -1161,12 +1375,18 @@ public:
     // type differs because a function may either return pointer to the buffer
     // directly or have to use intermediate buffer for translation.
 #if wxUSE_UNICODE
     // type differs because a function may either return pointer to the buffer
     // directly or have to use intermediate buffer for translation.
 #if wxUSE_UNICODE

+
+#if wxUSE_UTF8_LOCALE_ONLY
+    const char* mb_str() const { return wx_str(); }
+    const wxCharBuffer mb_str(const wxMBConv& conv) const;
+#else

     const wxCharBuffer mb_str(const wxMBConv& conv = wxConvLibc) const;
     const wxCharBuffer mb_str(const wxMBConv& conv = wxConvLibc) const;

+#endif

 
     const wxWX2MBbuf mbc_str() const { return mb_str(*wxConvCurrent); }
 
 #if wxUSE_UNICODE_WCHAR
 
     const wxWX2MBbuf mbc_str() const { return mb_str(*wxConvCurrent); }
 
 #if wxUSE_UNICODE_WCHAR

-    const wxChar* wc_str() const { return wx_str(); }
+    const wchar_t* wc_str() const { return wx_str(); }

 #elif wxUSE_UNICODE_UTF8
     const wxWCharBuffer wc_str() const;
 #endif
 #elif wxUSE_UNICODE_UTF8
     const wxWCharBuffer wc_str() const;
 #endif


@@ -1184,29 +1404,42 @@ public:
     const wxChar* mb_str() const { return wx_str(); }
 
     // for compatibility with wxUSE_UNICODE version
     const wxChar* mb_str() const { return wx_str(); }
 
     // for compatibility with wxUSE_UNICODE version

-    const wxChar* mb_str(const wxMBConv& WXUNUSED(conv)) const { return wx_str(); }
+    const char* mb_str(const wxMBConv& WXUNUSED(conv)) const { return wx_str(); }

 
     const wxWX2MBbuf mbc_str() const { return mb_str(); }
 
 #if wxUSE_WCHAR_T
     const wxWCharBuffer wc_str(const wxMBConv& conv = wxConvLibc) const;
 #endif // wxUSE_WCHAR_T
 
     const wxWX2MBbuf mbc_str() const { return mb_str(); }
 
 #if wxUSE_WCHAR_T
     const wxWCharBuffer wc_str(const wxMBConv& conv = wxConvLibc) const;
 #endif // wxUSE_WCHAR_T

-#ifdef __WXOSX__
-    const wxCharBuffer fn_str() const { return wxConvFile.cWC2WX( wc_str( wxConvLocal ) ); }
-#else
-    const wxChar* fn_str() const { return c_str(); }
-#endif
+    const wxCharBuffer fn_str() const { return wxConvFile.cWC2WX( wc_str( wxConvLibc ) ); }

 #endif // Unicode/ANSI
 
 #endif // Unicode/ANSI
 

+#if wxUSE_UNICODE_UTF8
+    const wxWCharBuffer t_str() const { return wc_str(); }
+#elif wxUSE_UNICODE_WCHAR
+    const wchar_t* t_str() const { return wx_str(); }
+#else
+    const char* t_str() const { return wx_str(); }
+#endif 
+
+

   // overloaded assignment
     // from another wxString
   wxString& operator=(const wxString& stringSrc)
   // overloaded assignment
     // from another wxString
   wxString& operator=(const wxString& stringSrc)

-    { m_impl = stringSrc.m_impl; return *this; }
+    { if (&stringSrc != this) m_impl = stringSrc.m_impl; return *this; }

   wxString& operator=(const wxCStrData& cstr)
     { return *this = cstr.AsString(); }
     // from a character
   wxString& operator=(wxUniChar ch)
   wxString& operator=(const wxCStrData& cstr)
     { return *this = cstr.AsString(); }
     // from a character
   wxString& operator=(wxUniChar ch)

-    { m_impl = wxStringOperations::EncodeChar(ch); return *this; }
+  {
+#if wxUSE_UNICODE_UTF8
+    if ( !ch.IsAscii() )
+        m_impl = wxStringOperations::EncodeChar(ch);
+    else
+#endif
+        m_impl = (wxStringCharType)ch;
+    return *this;
+  }

   wxString& operator=(wxUniCharRef ch)
     { return operator=((wxUniChar)ch); }
   wxString& operator=(char ch)
   wxString& operator=(wxUniCharRef ch)
     { return operator=((wxUniChar)ch); }
   wxString& operator=(char ch)


@@ -1296,6 +1529,16 @@ public:
     { append(psz); return *this; }
   wxString& Append(const wxWCharBuffer& psz)
     { append(psz); return *this; }
     { append(psz); return *this; }
   wxString& Append(const wxWCharBuffer& psz)
     { append(psz); return *this; }

+  wxString& Append(const char* psz, size_t nLen)
+    { append(psz, nLen); return *this; }
+  wxString& Append(const wchar_t* pwz, size_t nLen)
+    { append(pwz, nLen); return *this; }
+  wxString& Append(const wxCStrData& psz, size_t nLen)
+    { append(psz, nLen); return *this; }
+  wxString& Append(const wxCharBuffer& psz, size_t nLen)
+    { append(psz, nLen); return *this; }
+  wxString& Append(const wxWCharBuffer& psz, size_t nLen)
+    { append(psz, nLen); return *this; }

     // append count copies of given character
   wxString& Append(wxUniChar ch, size_t count = 1u)
     { append(count, ch); return *this; }
     // append count copies of given character
   wxString& Append(wxUniChar ch, size_t count = 1u)
     { append(count, ch); return *this; }


@@ -1307,10 +1550,6 @@ public:
     { append(count, ch); return *this; }
   wxString& Append(wchar_t ch, size_t count = 1u)
     { append(count, ch); return *this; }
     { append(count, ch); return *this; }
   wxString& Append(wchar_t ch, size_t count = 1u)
     { append(count, ch); return *this; }

-  wxString& Append(const char* psz, size_t nLen)
-    { append(psz, nLen); return *this; }
-  wxString& Append(const wchar_t* pwz, size_t nLen)
-    { append(pwz, nLen); return *this; }

 
     // prepend a string, return the string itself
   wxString& Prepend(const wxString& str)
 
     // prepend a string, return the string itself
   wxString& Prepend(const wxString& str)


@@ -1385,14 +1624,23 @@ public:
     { return compare(s); }
     // same as Cmp() but not case-sensitive
   int CmpNoCase(const wxString& s) const;
     { return compare(s); }
     // same as Cmp() but not case-sensitive
   int CmpNoCase(const wxString& s) const;

+

     // test for the string equality, either considering case or not
     // (if compareWithCase then the case matters)
   bool IsSameAs(const wxString& str, bool compareWithCase = true) const
     // test for the string equality, either considering case or not
     // (if compareWithCase then the case matters)
   bool IsSameAs(const wxString& str, bool compareWithCase = true) const

-    { return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; }
+  {
+#if !wxUSE_UNICODE_UTF8
+      // in UTF-8 build, length() is O(n) and doing this would be _slower_
+      if ( length() != str.length() )
+          return false;
+#endif
+      return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0;
+  }

   bool IsSameAs(const char *str, bool compareWithCase = true) const
     { return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; }
   bool IsSameAs(const wchar_t *str, bool compareWithCase = true) const
     { return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; }
   bool IsSameAs(const char *str, bool compareWithCase = true) const
     { return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; }
   bool IsSameAs(const wchar_t *str, bool compareWithCase = true) const
     { return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; }

+

   bool IsSameAs(const wxCStrData& str, bool compareWithCase = true) const
     { return IsSameAs(str.AsString(), compareWithCase); }
   bool IsSameAs(const wxCharBuffer& str, bool compareWithCase = true) const
   bool IsSameAs(const wxCStrData& str, bool compareWithCase = true) const
     { return IsSameAs(str.AsString(), compareWithCase); }
   bool IsSameAs(const wxCharBuffer& str, bool compareWithCase = true) const


@@ -1400,11 +1648,7 @@ public:
   bool IsSameAs(const wxWCharBuffer& str, bool compareWithCase = true) const
     { return IsSameAs(str.data(), compareWithCase); }
     // comparison with a single character: returns true if equal
   bool IsSameAs(const wxWCharBuffer& str, bool compareWithCase = true) const
     { return IsSameAs(str.data(), compareWithCase); }
     // comparison with a single character: returns true if equal

-  bool IsSameAs(wxUniChar c, bool compareWithCase = true) const
-    {
-      return (length() == 1) && (compareWithCase ? GetChar(0u) == c
-                              : wxToupper(GetChar(0u)) == wxToupper(c));
-    }
+  bool IsSameAs(wxUniChar c, bool compareWithCase = true) const;

   // FIXME-UTF8: remove these overloads
   bool IsSameAs(wxUniCharRef c, bool compareWithCase = true) const
     { return IsSameAs(wxUniChar(c), compareWithCase); }
   // FIXME-UTF8: remove these overloads
   bool IsSameAs(wxUniCharRef c, bool compareWithCase = true) const
     { return IsSameAs(wxUniChar(c), compareWithCase); }


@@ -1429,11 +1673,11 @@ public:
       // check if the string starts with the given prefix and return the rest
       // of the string in the provided pointer if it is not NULL; otherwise
       // return false
       // check if the string starts with the given prefix and return the rest
       // of the string in the provided pointer if it is not NULL; otherwise
       // return false

-  bool StartsWith(const wxChar *prefix, wxString *rest = NULL) const;
+  bool StartsWith(const wxString& prefix, wxString *rest = NULL) const;

       // check if the string ends with the given suffix and return the
       // beginning of the string before the suffix in the provided pointer if
       // it is not NULL; otherwise return false
       // check if the string ends with the given suffix and return the
       // beginning of the string before the suffix in the provided pointer if
       // it is not NULL; otherwise return false

-  bool EndsWith(const wxChar *suffix, wxString *rest = NULL) const;
+  bool EndsWith(const wxString& suffix, wxString *rest = NULL) const;

 
       // get first nCount characters
   wxString Left(size_t nCount) const;
 
       // get first nCount characters
   wxString Left(size_t nCount) const;


@@ -1541,15 +1785,18 @@ public:
     // as sprintf(), returns the number of characters written or < 0 on error
     // (take 'this' into account in attribute parameter count)
   // int Printf(const wxString& format, ...);
     // as sprintf(), returns the number of characters written or < 0 on error
     // (take 'this' into account in attribute parameter count)
   // int Printf(const wxString& format, ...);

-  WX_DEFINE_VARARG_FUNC(int, Printf, 1, (const wxString&),
+  WX_DEFINE_VARARG_FUNC(int, Printf, 1, (const wxFormatString&),

                         DoPrintfWchar, DoPrintfUtf8)
 #ifdef __WATCOMC__
                         DoPrintfWchar, DoPrintfUtf8)
 #ifdef __WATCOMC__

-  WX_DEFINE_VARARG_FUNC(int, Printf, 1, (const char*)
-                        DoPrintfWchar, DoPrintfUtf8)
-  WX_DEFINE_VARARG_FUNC(int, Printf, 1, (const wchar_t*)
-                        DoPrintfWchar, DoPrintfUtf8)
-  WX_DEFINE_VARARG_FUNC(int, Printf, 1, (const wxCStrData&)
-                        DoPrintfWchar, DoPrintfUtf8)
+  // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351
+  WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wxString&),
+                              (wxFormatString(f1)));
+  WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wxCStrData&),
+                              (wxFormatString(f1)));
+  WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const char*),
+                              (wxFormatString(f1)));
+  WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wchar_t*),
+                              (wxFormatString(f1)));

 #endif
 #endif // !wxNEEDS_WXSTRING_PRINTF_MIXIN
     // as vprintf(), returns the number of characters written or < 0 on error
 #endif
 #endif // !wxNEEDS_WXSTRING_PRINTF_MIXIN
     // as vprintf(), returns the number of characters written or < 0 on error


@@ -1558,16 +1805,18 @@ public:
 #ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN
     // returns the string containing the result of Printf() to it
   // static wxString Format(const wxString& format, ...) ATTRIBUTE_PRINTF_1;
 #ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN
     // returns the string containing the result of Printf() to it
   // static wxString Format(const wxString& format, ...) ATTRIBUTE_PRINTF_1;

-  WX_DEFINE_VARARG_FUNC(static wxString, Format, 1, (const wxString&),
+  WX_DEFINE_VARARG_FUNC(static wxString, Format, 1, (const wxFormatString&),

                         DoFormatWchar, DoFormatUtf8)
 #ifdef __WATCOMC__
   // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351
                         DoFormatWchar, DoFormatUtf8)
 #ifdef __WATCOMC__
   // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351

-  WX_DEFINE_VARARG_FUNC(static wxString, Format, 1, (const char*),
-                        DoFormatWchar, DoFormatUtf8)
-  WX_DEFINE_VARARG_FUNC(static wxString, Format, 1, (const wchar_t*),
-                        DoFormatWchar, DoFormatUtf8)
-  WX_DEFINE_VARARG_FUNC(static wxString, Format, 1, (const wxCStrData&),
-                        DoFormatWchar, DoFormatUtf8)
+  WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wxString&),
+                              (wxFormatString(f1)));
+  WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wxCStrData&),
+                              (wxFormatString(f1)));
+  WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const char*),
+                              (wxFormatString(f1)));
+  WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wchar_t*),
+                              (wxFormatString(f1)));

 #endif
 #endif
     // the same as above, but takes a va_list
 #endif
 #endif
     // the same as above, but takes a va_list


@@ -1576,7 +1825,7 @@ public:
   // raw access to string memory
     // ensure that string has space for at least nLen characters
     // only works if the data of this string is not shared
   // raw access to string memory
     // ensure that string has space for at least nLen characters
     // only works if the data of this string is not shared

-  bool Alloc(size_t nLen) { reserve(nLen); /*return capacity() >= nLen;*/ return true; }
+  bool Alloc(size_t nLen) { reserve(nLen); return capacity() >= nLen; }

     // minimize the string's memory
     // only works if the data of this string is not shared
   bool Shrink();
     // minimize the string's memory
     // only works if the data of this string is not shared
   bool Shrink();


@@ -1605,16 +1854,18 @@ public:
   // use Printf()
   // (take 'this' into account in attribute parameter count)
   // int sprintf(const wxString& format, ...) ATTRIBUTE_PRINTF_2;
   // use Printf()
   // (take 'this' into account in attribute parameter count)
   // int sprintf(const wxString& format, ...) ATTRIBUTE_PRINTF_2;

-  WX_DEFINE_VARARG_FUNC(int, sprintf, 1, (const wxString&),
+  WX_DEFINE_VARARG_FUNC(int, sprintf, 1, (const wxFormatString&),

                         DoPrintfWchar, DoPrintfUtf8)
 #ifdef __WATCOMC__
   // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351
                         DoPrintfWchar, DoPrintfUtf8)
 #ifdef __WATCOMC__
   // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351

-  WX_DEFINE_VARARG_FUNC(int, sprintf, 1, (const char*),
-                        DoPrintfWchar, DoPrintfUtf8)
-  WX_DEFINE_VARARG_FUNC(int, sprintf, 1, (const wchar_t*),
-                        DoPrintfWchar, DoPrintfUtf8)
-  WX_DEFINE_VARARG_FUNC(int, sprintf, 1, (const wxCStrData&),
-                        DoPrintfWchar, DoPrintfUtf8)
+  WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wxString&),
+                              (wxFormatString(f1)));
+  WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wxCStrData&),
+                              (wxFormatString(f1)));
+  WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const char*),
+                              (wxFormatString(f1)));
+  WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wchar_t*),
+                              (wxFormatString(f1)));

 #endif
 #endif // wxNEEDS_WXSTRING_PRINTF_MIXIN
 
 #endif
 #endif // wxNEEDS_WXSTRING_PRINTF_MIXIN
 


@@ -1660,7 +1911,7 @@ public:
 
     // take nLen chars starting at nPos
   wxString(const wxString& str, size_t nPos, size_t nLen)
 
     // take nLen chars starting at nPos
   wxString(const wxString& str, size_t nPos, size_t nLen)

-      : m_impl(str.m_impl, nPos, nLen) { }
+      { assign(str, nPos, nLen); }

     // take all characters from first to last
   wxString(const_iterator first, const_iterator last)
       : m_impl(first.impl(), last.impl()) { }
     // take all characters from first to last
   wxString(const_iterator first, const_iterator last)
       : m_impl(first.impl(), last.impl()) { }


@@ -1723,6 +1974,12 @@ public:
     { return append(str.data()); }
   wxString& append(const wxWCharBuffer& str)
     { return append(str.data()); }
     { return append(str.data()); }
   wxString& append(const wxWCharBuffer& str)
     { return append(str.data()); }

+  wxString& append(const wxCStrData& str, size_t n)
+    { return append(str.AsString(), 0, n); }
+  wxString& append(const wxCharBuffer& str, size_t n)
+    { return append(str.data(), n); }
+  wxString& append(const wxWCharBuffer& str, size_t n)
+    { return append(str.data(), n); }

 
     // append n copies of ch
   wxString& append(size_t n, wxUniChar ch)
 
     // append n copies of ch
   wxString& append(size_t n, wxUniChar ch)


@@ -1735,6 +1992,15 @@ public:
         m_impl.append(n, (wxStringCharType)ch);
     return *this;
   }
         m_impl.append(n, (wxStringCharType)ch);
     return *this;
   }

+  wxString& append(size_t n, wxUniCharRef ch)
+    { return append(n, wxUniChar(ch)); }
+  wxString& append(size_t n, char ch)
+    { return append(n, wxUniChar(ch)); }
+  wxString& append(size_t n, unsigned char ch)
+    { return append(n, wxUniChar(ch)); }
+  wxString& append(size_t n, wchar_t ch)
+    { return append(n, wxUniChar(ch)); }
+

     // append from first to last
   wxString& append(const_iterator first, const_iterator last)
     { m_impl.append(first.impl(), last.impl()); return *this; }
     // append from first to last
   wxString& append(const_iterator first, const_iterator last)
     { m_impl.append(first.impl(), last.impl()); return *this; }


@@ -2093,8 +2359,15 @@ public:
     // find the first occurence of character ch after nStart
   size_t find(wxUniChar ch, size_t nStart = 0) const
   {
     // find the first occurence of character ch after nStart
   size_t find(wxUniChar ch, size_t nStart = 0) const
   {

-    return PosFromImpl(m_impl.find(wxStringOperations::EncodeChar(ch),
-                                   PosToImpl(nStart)));
+#if wxUSE_UNICODE_UTF8
+    if ( !ch.IsAscii() )
+        return PosFromImpl(m_impl.find(wxStringOperations::EncodeChar(ch),
+                                       PosToImpl(nStart)));
+    else
+#endif
+        return PosFromImpl(m_impl.find((wxStringCharType)ch,
+                                       PosToImpl(nStart)));
+

   }
   size_t find(wxUniCharRef ch, size_t nStart = 0) const
     {  return find(wxUniChar(ch), nStart); }
   }
   size_t find(wxUniCharRef ch, size_t nStart = 0) const
     {  return find(wxUniChar(ch), nStart); }


@@ -2131,8 +2404,14 @@ public:
     // as find, but from the end
   size_t rfind(wxUniChar ch, size_t nStart = npos) const
   {
     // as find, but from the end
   size_t rfind(wxUniChar ch, size_t nStart = npos) const
   {

-    return PosFromImpl(m_impl.rfind(wxStringOperations::EncodeChar(ch),
-                                    PosToImpl(nStart)));
+#if wxUSE_UNICODE_UTF8
+    if ( !ch.IsAscii() )
+        return PosFromImpl(m_impl.rfind(wxStringOperations::EncodeChar(ch),
+                                        PosToImpl(nStart)));
+    else
+#endif
+        return PosFromImpl(m_impl.rfind((wxStringCharType)ch,
+                                        PosToImpl(nStart)));

   }
   size_t rfind(wxUniCharRef ch, size_t nStart = npos) const
     {  return rfind(wxUniChar(ch), nStart); }
   }
   size_t rfind(wxUniCharRef ch, size_t nStart = npos) const
     {  return rfind(wxUniChar(ch), nStart); }


@@ -2370,7 +2649,15 @@ public:
     { return operator+=(s.data()); }
       // string += char
   wxString& operator+=(wxUniChar ch)
     { return operator+=(s.data()); }
       // string += char
   wxString& operator+=(wxUniChar ch)

-    { m_impl += wxStringOperations::EncodeChar(ch); return *this; }
+  {
+#if wxUSE_UNICODE_UTF8
+    if ( !ch.IsAscii() )
+        m_impl += wxStringOperations::EncodeChar(ch);
+    else
+#endif
+        m_impl += (wxStringCharType)ch;
+    return *this;
+  }

   wxString& operator+=(wxUniCharRef ch) { return *this += wxUniChar(ch); }
   wxString& operator+=(int ch) { return *this += wxUniChar(ch); }
   wxString& operator+=(char ch) { return *this += wxUniChar(ch); }
   wxString& operator+=(wxUniCharRef ch) { return *this += wxUniChar(ch); }
   wxString& operator+=(int ch) { return *this += wxUniChar(ch); }
   wxString& operator+=(char ch) { return *this += wxUniChar(ch); }


@@ -2428,16 +2715,37 @@ private:
 
       T *m_buf;
   };
 
       T *m_buf;
   };

-#if wxUSE_UNICODE
+#if wxUSE_UNICODE && !wxUSE_UTF8_LOCALE_ONLY

   ConvertedBuffer<char> m_convertedToChar;
 #endif
 #if !wxUSE_UNICODE_WCHAR
   ConvertedBuffer<wchar_t> m_convertedToWChar;
 #endif
 
   ConvertedBuffer<char> m_convertedToChar;
 #endif
 #if !wxUSE_UNICODE_WCHAR
   ConvertedBuffer<wchar_t> m_convertedToWChar;
 #endif
 

-  friend class WXDLLIMPEXP_BASE wxCStrData;
-  friend class wxImplStringBuffer;
-  friend class wxImplStringBufferLength;
+#if wxUSE_UNICODE_UTF8
+  // FIXME-UTF8: (try to) move this elsewhere (TLS) or solve differently
+  //             assigning to character pointer to by wxString::interator may
+  //             change the underlying wxStringImpl iterator, so we have to
+  //             keep track of all iterators and update them as necessary:
+  struct wxStringIteratorNodeHead
+  {
+      wxStringIteratorNodeHead() : ptr(NULL) {}
+      wxStringIteratorNode *ptr;
+
+      // copying is disallowed as it would result in more than one pointer into
+      // the same linked list
+      DECLARE_NO_COPY_CLASS(wxStringIteratorNodeHead)
+  };
+
+  wxStringIteratorNodeHead m_iterators;
+
+  friend class WXDLLIMPEXP_FWD_BASE wxStringIteratorNode;
+  friend class WXDLLIMPEXP_FWD_BASE wxUniCharRef;
+#endif // wxUSE_UNICODE_UTF8
+
+  friend class WXDLLIMPEXP_FWD_BASE wxCStrData;
+  friend class wxStringInternalBuffer;
+  friend class wxStringInternalBufferLength;

 };
 
 #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN
 };
 
 #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN


@@ -2446,21 +2754,13 @@ private:
 
 // string iterator operators that satisfy STL Random Access Iterator
 // requirements:
 
 // string iterator operators that satisfy STL Random Access Iterator
 // requirements:

-inline wxString::iterator operator+(int n, wxString::iterator i)
-  { return i + n; }
-inline wxString::iterator operator+(size_t n, wxString::iterator i)
-  { return i + n; }
-inline wxString::const_iterator operator+(int n, wxString::const_iterator i)
-  { return i + n; }
-inline wxString::const_iterator operator+(size_t n, wxString::const_iterator i)
-  { return i + n; }
-inline wxString::reverse_iterator operator+(int n, wxString::reverse_iterator i)
+inline wxString::iterator operator+(ptrdiff_t n, wxString::iterator i)

   { return i + n; }
   { return i + n; }

-inline wxString::reverse_iterator operator+(size_t n, wxString::reverse_iterator i)
+inline wxString::const_iterator operator+(ptrdiff_t n, wxString::const_iterator i)

   { return i + n; }
   { return i + n; }

-inline wxString::const_reverse_iterator operator+(int n, wxString::const_reverse_iterator i)
+inline wxString::reverse_iterator operator+(ptrdiff_t n, wxString::reverse_iterator i)

   { return i + n; }
   { return i + n; }

-inline wxString::const_reverse_iterator operator+(size_t n, wxString::const_reverse_iterator i)
+inline wxString::const_reverse_iterator operator+(ptrdiff_t n, wxString::const_reverse_iterator i)

   { return i + n; }
 
 // notice that even though for many compilers the friend declarations above are
   { return i + n; }
 
 // notice that even though for many compilers the friend declarations above are


@@ -2490,16 +2790,68 @@ inline wxString operator+(wchar_t ch, const wxString& string)
     { return wxUniChar(ch) + string; }
 
 
     { return wxUniChar(ch) + string; }
 
 

-#if wxUSE_STL_BASED_WXSTRING
-    // return an empty wxString (not very useful with wxUSE_STL == 1)
-    inline const wxString wxGetEmptyString() { return wxString(); }
-#else // !wxUSE_STL_BASED_WXSTRING
-    // return an empty wxString (more efficient than wxString() here)
-    inline const wxString& wxGetEmptyString()
+#define wxGetEmptyString() wxString()
+
+// ----------------------------------------------------------------------------
+// helper functions which couldn't be defined inline
+// ----------------------------------------------------------------------------
+
+namespace wxPrivate
+{
+
+#if wxUSE_UNICODE_WCHAR
+
+template <>
+struct wxStringAsBufHelper<char>
+{
+    static wxCharBuffer Get(const wxString& s, size_t *len)
+    {
+        wxCharBuffer buf(s.mb_str());
+        if ( len )
+            *len = buf ? strlen(buf) : 0;
+        return buf;
+    }
+};
+
+template <>
+struct wxStringAsBufHelper<wchar_t>
+{
+    static wxWCharBuffer Get(const wxString& s, size_t *len)
+    {
+        if ( len )
+            *len = s.length();
+        return wxWCharBuffer::CreateNonOwned(s.wx_str());
+    }
+};
+
+#elif wxUSE_UNICODE_UTF8
+
+template <>
+struct wxStringAsBufHelper<char>
+{
+    static wxCharBuffer Get(const wxString& s, size_t *len)

     {
     {

-        return *(wxString *)&wxEmptyString;
+        if ( len )
+            *len = s.utf8_length();
+        return wxCharBuffer::CreateNonOwned(s.wx_str());

     }
     }

-#endif // wxUSE_STL_BASED_WXSTRING/!wxUSE_STL_BASED_WXSTRING
+};
+
+template <>
+struct wxStringAsBufHelper<wchar_t>
+{
+    static wxWCharBuffer Get(const wxString& s, size_t *len)
+    {
+        wxWCharBuffer wbuf(s.wc_str());
+        if ( len )
+            *len = wxWcslen(wbuf);
+        return wbuf;
+    }
+};
+
+#endif // Unicode build kind
+
+} // namespace wxPrivate

 
 // ----------------------------------------------------------------------------
 // wxStringBuffer: a tiny class allowing to get a writable pointer into string
 
 // ----------------------------------------------------------------------------
 // wxStringBuffer: a tiny class allowing to get a writable pointer into string


@@ -2508,16 +2860,16 @@ inline wxString operator+(wchar_t ch, const wxString& string)
 #if !wxUSE_STL_BASED_WXSTRING
 // string buffer for direct access to string data in their native
 // representation:
 #if !wxUSE_STL_BASED_WXSTRING
 // string buffer for direct access to string data in their native
 // representation:

-class wxImplStringBuffer
+class wxStringInternalBuffer

 {
 public:
     typedef wxStringCharType CharType;
 
 {
 public:
     typedef wxStringCharType CharType;
 

-    wxImplStringBuffer(wxString& str, size_t lenWanted = 1024)
+    wxStringInternalBuffer(wxString& str, size_t lenWanted = 1024)

         : m_str(str), m_buf(NULL)
         { m_buf = m_str.DoGetWriteBuf(lenWanted); }
 
         : m_str(str), m_buf(NULL)
         { m_buf = m_str.DoGetWriteBuf(lenWanted); }
 

-    ~wxImplStringBuffer() { m_str.DoUngetWriteBuf(); }
+    ~wxStringInternalBuffer() { m_str.DoUngetWriteBuf(); }

 
     operator wxStringCharType*() const { return m_buf; }
 
 
     operator wxStringCharType*() const { return m_buf; }
 


@@ -2525,22 +2877,22 @@ private:
     wxString&         m_str;
     wxStringCharType *m_buf;
 
     wxString&         m_str;
     wxStringCharType *m_buf;
 

-    DECLARE_NO_COPY_CLASS(wxImplStringBuffer)
+    DECLARE_NO_COPY_CLASS(wxStringInternalBuffer)

 };
 
 };
 

-class wxImplStringBufferLength
+class wxStringInternalBufferLength

 {
 public:
     typedef wxStringCharType CharType;
 
 {
 public:
     typedef wxStringCharType CharType;
 

-    wxImplStringBufferLength(wxString& str, size_t lenWanted = 1024)
+    wxStringInternalBufferLength(wxString& str, size_t lenWanted = 1024)

         : m_str(str), m_buf(NULL), m_len(0), m_lenSet(false)
     {
         m_buf = m_str.DoGetWriteBuf(lenWanted);
         wxASSERT(m_buf != NULL);
     }
 
         : m_str(str), m_buf(NULL), m_len(0), m_lenSet(false)
     {
         m_buf = m_str.DoGetWriteBuf(lenWanted);
         wxASSERT(m_buf != NULL);
     }
 

-    ~wxImplStringBufferLength()
+    ~wxStringInternalBufferLength()

     {
         wxASSERT(m_lenSet);
         m_str.DoUngetWriteBuf(m_len);
     {
         wxASSERT(m_lenSet);
         m_str.DoUngetWriteBuf(m_len);


@@ -2555,21 +2907,43 @@ private:
     size_t            m_len;
     bool              m_lenSet;
 
     size_t            m_len;
     bool              m_lenSet;
 

-    DECLARE_NO_COPY_CLASS(wxImplStringBufferLength)
+    DECLARE_NO_COPY_CLASS(wxStringInternalBufferLength)

 };
 
 #endif // !wxUSE_STL_BASED_WXSTRING
 
 template<typename T>
 };
 
 #endif // !wxUSE_STL_BASED_WXSTRING
 
 template<typename T>

-class wxStringTypeBufferBase
+class WXDLLIMPEXP_BASE wxStringTypeBufferBase

 {
 public:
     typedef T CharType;
 
     wxStringTypeBufferBase(wxString& str, size_t lenWanted = 1024)
         : m_str(str), m_buf(lenWanted)
 {
 public:
     typedef T CharType;
 
     wxStringTypeBufferBase(wxString& str, size_t lenWanted = 1024)
         : m_str(str), m_buf(lenWanted)

-        { }
-
+    {
+        // for compatibility with old wxStringBuffer which provided direct
+        // access to wxString internal buffer, initialize ourselves with the
+        // string initial contents
+
+        // FIXME-VC6: remove the ugly (CharType *)NULL and use normal
+        //            tchar_str<CharType>
+        size_t len;
+        const wxCharTypeBuffer<CharType> buf(str.tchar_str(&len, (CharType *)NULL));
+        if ( buf )
+        {
+            if ( len > lenWanted )
+            {
+                // in this case there is not enough space for terminating NUL,
+                // ensure that we still put it there
+                m_buf.data()[lenWanted] = 0;
+                len = lenWanted - 1;
+            }
+
+            memcpy(m_buf.data(), buf, (len + 1)*sizeof(CharType));
+        }
+        //else: conversion failed, this can happen when trying to get Unicode
+        //      string contents into a char string
+    }

 
     operator CharType*() { return m_buf.data(); }
 
 
     operator CharType*() { return m_buf.data(); }
 


@@ -2579,29 +2953,26 @@ protected:
 };
 
 template<typename T>
 };
 
 template<typename T>

-class wxStringTypeBufferLengthBase
+class WXDLLIMPEXP_BASE wxStringTypeBufferLengthBase
+    : public wxStringTypeBufferBase<T>

 {
 public:
 {
 public:

-    typedef T CharType;
-

     wxStringTypeBufferLengthBase(wxString& str, size_t lenWanted = 1024)
     wxStringTypeBufferLengthBase(wxString& str, size_t lenWanted = 1024)

-        : m_str(str), m_buf(lenWanted), m_len(0), m_lenSet(false)
+        : wxStringTypeBufferBase<T>(str, lenWanted),
+          m_len(0),
+          m_lenSet(false)

         { }
 
     ~wxStringTypeBufferLengthBase()
     {
         { }
 
     ~wxStringTypeBufferLengthBase()
     {

-        wxASSERT(m_lenSet);
-        m_str.assign(m_buf.data(), m_len);
+        wxASSERT_MSG( this->m_lenSet, "forgot to call SetLength()" );

     }
 
     }
 

-    operator CharType*() { return m_buf.data(); }

     void SetLength(size_t length) { m_len = length; m_lenSet = true; }
 
 protected:
     void SetLength(size_t length) { m_len = length; m_lenSet = true; }
 
 protected:

-    wxString& m_str;
-    wxCharTypeBuffer<CharType> m_buf;
-    size_t        m_len;
-    bool          m_lenSet;
+    size_t m_len;
+    bool m_lenSet;

 };
 
 template<typename T>
 };
 
 template<typename T>


@@ -2609,7 +2980,9 @@ class wxStringTypeBuffer : public wxStringTypeBufferBase<T>
 {
 public:
     wxStringTypeBuffer(wxString& str, size_t lenWanted = 1024)
 {
 public:
     wxStringTypeBuffer(wxString& str, size_t lenWanted = 1024)

-        : wxStringTypeBufferBase<T>(str, lenWanted) {}
+        : wxStringTypeBufferBase<T>(str, lenWanted)
+        { }
+

     ~wxStringTypeBuffer()
     {
         this->m_str.assign(this->m_buf.data());
     ~wxStringTypeBuffer()
     {
         this->m_str.assign(this->m_buf.data());


@@ -2623,11 +2996,11 @@ class wxStringTypeBufferLength : public wxStringTypeBufferLengthBase<T>
 {
 public:
     wxStringTypeBufferLength(wxString& str, size_t lenWanted = 1024)
 {
 public:
     wxStringTypeBufferLength(wxString& str, size_t lenWanted = 1024)

-        : wxStringTypeBufferLengthBase<T>(str, lenWanted) {}
+        : wxStringTypeBufferLengthBase<T>(str, lenWanted)
+        { }

 
     ~wxStringTypeBufferLength()
     {
 
     ~wxStringTypeBufferLength()
     {

-        wxASSERT(this->m_lenSet);

         this->m_str.assign(this->m_buf.data(), this->m_len);
     }
 
         this->m_str.assign(this->m_buf.data(), this->m_len);
     }
 


@@ -2635,31 +3008,38 @@ public:
 };
 
 #if wxUSE_STL_BASED_WXSTRING
 };
 
 #if wxUSE_STL_BASED_WXSTRING

-class wxImplStringBuffer : public wxStringTypeBufferBase<wxStringCharType>
+
+WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferBase<wxStringCharType> )
+
+class wxStringInternalBuffer : public wxStringTypeBufferBase<wxStringCharType>

 {
 public:
 {
 public:

-    wxImplStringBuffer(wxString& str, size_t lenWanted = 1024)
+    wxStringInternalBuffer(wxString& str, size_t lenWanted = 1024)

         : wxStringTypeBufferBase<wxStringCharType>(str, lenWanted) {}
         : wxStringTypeBufferBase<wxStringCharType>(str, lenWanted) {}

-    ~wxImplStringBuffer()
+    ~wxStringInternalBuffer()

         { m_str.m_impl.assign(m_buf.data()); }
 
         { m_str.m_impl.assign(m_buf.data()); }
 

-    DECLARE_NO_COPY_CLASS(wxImplStringBuffer)
+    DECLARE_NO_COPY_CLASS(wxStringInternalBuffer)

 };
 
 };
 

-class wxImplStringBufferLength : public wxStringTypeBufferLengthBase<wxStringCharType>
+WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE(
+    wxStringTypeBufferLengthBase<wxStringCharType> )
+
+class wxStringInternalBufferLength
+    : public wxStringTypeBufferLengthBase<wxStringCharType>

 {
 public:
 {
 public:

-    wxImplStringBufferLength(wxString& str, size_t lenWanted = 1024)
+    wxStringInternalBufferLength(wxString& str, size_t lenWanted = 1024)

         : wxStringTypeBufferLengthBase<wxStringCharType>(str, lenWanted) {}
 
         : wxStringTypeBufferLengthBase<wxStringCharType>(str, lenWanted) {}
 

-    ~wxImplStringBufferLength()
+    ~wxStringInternalBufferLength()

     {
     {

-        wxASSERT(m_lenSet);

         m_str.m_impl.assign(m_buf.data(), m_len);
     }
 
         m_str.m_impl.assign(m_buf.data(), m_len);
     }
 

-    DECLARE_NO_COPY_CLASS(wxImplStringBufferLength)
+    DECLARE_NO_COPY_CLASS(wxStringInternalBufferLength)

 };
 };

+

 #endif // wxUSE_STL_BASED_WXSTRING
 
 
 #endif // wxUSE_STL_BASED_WXSTRING
 
 


@@ -2667,10 +3047,42 @@ public:
 typedef wxStringTypeBuffer<wxChar>        wxStringBuffer;
 typedef wxStringTypeBufferLength<wxChar>  wxStringBufferLength;
 #else // if !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
 typedef wxStringTypeBuffer<wxChar>        wxStringBuffer;
 typedef wxStringTypeBufferLength<wxChar>  wxStringBufferLength;
 #else // if !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8

-typedef wxImplStringBuffer                wxStringBuffer;
-typedef wxImplStringBufferLength          wxStringBufferLength;
+typedef wxStringInternalBuffer                wxStringBuffer;
+typedef wxStringInternalBufferLength          wxStringBufferLength;

 #endif // !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
 
 #endif // !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8
 

+#if wxUSE_UNICODE_UTF8
+typedef wxStringInternalBuffer                wxUTF8StringBuffer;
+typedef wxStringInternalBufferLength          wxUTF8StringBufferLength;
+#elif wxUSE_UNICODE_WCHAR
+
+WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferBase<char> )
+
+class WXDLLIMPEXP_BASE wxUTF8StringBuffer : public wxStringTypeBufferBase<char>
+{
+public:
+    wxUTF8StringBuffer(wxString& str, size_t lenWanted = 1024)
+        : wxStringTypeBufferBase<char>(str, lenWanted) {}
+    ~wxUTF8StringBuffer();
+
+    DECLARE_NO_COPY_CLASS(wxUTF8StringBuffer)
+};
+
+WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferLengthBase<char> )
+
+class WXDLLIMPEXP_BASE wxUTF8StringBufferLength
+    : public wxStringTypeBufferLengthBase<char>
+{
+public:
+    wxUTF8StringBufferLength(wxString& str, size_t lenWanted = 1024)
+        : wxStringTypeBufferLengthBase<char>(str, lenWanted) {}
+    ~wxUTF8StringBufferLength();
+
+    DECLARE_NO_COPY_CLASS(wxUTF8StringBufferLength)
+};
+#endif // wxUSE_UNICODE_UTF8/wxUSE_UNICODE_WCHAR
+
+

 // ---------------------------------------------------------------------------
 // wxString comparison functions: operator versions are always case sensitive
 // ---------------------------------------------------------------------------
 // ---------------------------------------------------------------------------
 // wxString comparison functions: operator versions are always case sensitive
 // ---------------------------------------------------------------------------


@@ -2681,12 +3093,10 @@ wxDEFINE_ALL_COMPARISONS(const wxChar *, const wxString&, wxCMP_WXCHAR_STRING)
 
 #undef wxCMP_WXCHAR_STRING
 
 
 #undef wxCMP_WXCHAR_STRING
 

-// note that there is an optimization in operator==() and !=(): we (quickly)
-// checks the strings length first, before comparing their data

 inline bool operator==(const wxString& s1, const wxString& s2)
 inline bool operator==(const wxString& s1, const wxString& s2)

-    { return (s1.Len() == s2.Len()) && (s1.Cmp(s2) == 0); }
+    { return s1.IsSameAs(s2); }

 inline bool operator!=(const wxString& s1, const wxString& s2)
 inline bool operator!=(const wxString& s1, const wxString& s2)

-    { return (s1.Len() != s2.Len()) || (s1.Cmp(s2) != 0); }
+    { return !s1.IsSameAs(s2); }

 inline bool operator< (const wxString& s1, const wxString& s2)
     { return s1.Cmp(s2) < 0; }
 inline bool operator> (const wxString& s1, const wxString& s2)
 inline bool operator< (const wxString& s1, const wxString& s2)
     { return s1.Cmp(s2) < 0; }
 inline bool operator> (const wxString& s1, const wxString& s2)


@@ -2696,7 +3106,15 @@ inline bool operator<=(const wxString& s1, const wxString& s2)
 inline bool operator>=(const wxString& s1, const wxString& s2)
     { return s1.Cmp(s2) >= 0; }
 
 inline bool operator>=(const wxString& s1, const wxString& s2)
     { return s1.Cmp(s2) >= 0; }
 

-#if wxUSE_UNICODE
+inline bool operator==(const wxString& s1, const wxCStrData& s2)
+    { return s1 == s2.AsString(); }
+inline bool operator==(const wxCStrData& s1, const wxString& s2)
+    { return s1.AsString() == s2; }
+inline bool operator!=(const wxString& s1, const wxCStrData& s2)
+    { return s1 != s2.AsString(); }
+inline bool operator!=(const wxCStrData& s1, const wxString& s2)
+    { return s1.AsString() != s2; }
+

 inline bool operator==(const wxString& s1, const wxWCharBuffer& s2)
     { return (s1.Cmp((const wchar_t *)s2) == 0); }
 inline bool operator==(const wxWCharBuffer& s1, const wxString& s2)
 inline bool operator==(const wxString& s1, const wxWCharBuffer& s2)
     { return (s1.Cmp((const wchar_t *)s2) == 0); }
 inline bool operator==(const wxWCharBuffer& s1, const wxString& s2)


@@ -2705,7 +3123,7 @@ inline bool operator!=(const wxString& s1, const wxWCharBuffer& s2)
     { return (s1.Cmp((const wchar_t *)s2) != 0); }
 inline bool operator!=(const wxWCharBuffer& s1, const wxString& s2)
     { return (s2.Cmp((const wchar_t *)s1) != 0); }
     { return (s1.Cmp((const wchar_t *)s2) != 0); }
 inline bool operator!=(const wxWCharBuffer& s1, const wxString& s2)
     { return (s2.Cmp((const wchar_t *)s1) != 0); }

-#else // !wxUSE_UNICODE
+

 inline bool operator==(const wxString& s1, const wxCharBuffer& s2)
     { return (s1.Cmp((const char *)s2) == 0); }
 inline bool operator==(const wxCharBuffer& s1, const wxString& s2)
 inline bool operator==(const wxString& s1, const wxCharBuffer& s2)
     { return (s1.Cmp((const char *)s2) == 0); }
 inline bool operator==(const wxCharBuffer& s1, const wxString& s2)


@@ -2714,19 +3132,16 @@ inline bool operator!=(const wxString& s1, const wxCharBuffer& s2)
     { return (s1.Cmp((const char *)s2) != 0); }
 inline bool operator!=(const wxCharBuffer& s1, const wxString& s2)
     { return (s2.Cmp((const char *)s1) != 0); }
     { return (s1.Cmp((const char *)s2) != 0); }
 inline bool operator!=(const wxCharBuffer& s1, const wxString& s2)
     { return (s2.Cmp((const char *)s1) != 0); }

-#endif // wxUSE_UNICODE/!wxUSE_UNICODE

 
 

-#if wxUSE_UNICODE

 inline wxString operator+(const wxString& string, const wxWCharBuffer& buf)
     { return string + (const wchar_t *)buf; }
 inline wxString operator+(const wxWCharBuffer& buf, const wxString& string)
     { return (const wchar_t *)buf + string; }
 inline wxString operator+(const wxString& string, const wxWCharBuffer& buf)
     { return string + (const wchar_t *)buf; }
 inline wxString operator+(const wxWCharBuffer& buf, const wxString& string)
     { return (const wchar_t *)buf + string; }

-#else // !wxUSE_UNICODE
+

 inline wxString operator+(const wxString& string, const wxCharBuffer& buf)
     { return string + (const char *)buf; }
 inline wxString operator+(const wxCharBuffer& buf, const wxString& string)
     { return (const char *)buf + string; }
 inline wxString operator+(const wxString& string, const wxCharBuffer& buf)
     { return string + (const char *)buf; }
 inline wxString operator+(const wxCharBuffer& buf, const wxString& string)
     { return (const char *)buf + string; }

-#endif // wxUSE_UNICODE/!wxUSE_UNICODE

 
 // comparison with char
 inline bool operator==(const wxUniChar& c, const wxString& s) { return s.IsSameAs(c); }
 
 // comparison with char
 inline bool operator==(const wxUniChar& c, const wxString& s) { return s.IsSameAs(c); }


@@ -2788,7 +3203,15 @@ WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxCharBuffer&);
 WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxWCharBuffer&);
 #endif
 
 WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxWCharBuffer&);
 #endif
 

-#endif  // wxSTD_STRING_COMPATIBILITY
+#if wxUSE_UNICODE && defined(HAVE_WOSTREAM)
+
+WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxString&);
+WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxCStrData&);
+WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxWCharBuffer&);
+
+#endif  // wxUSE_UNICODE && defined(HAVE_WOSTREAM)
+
+#endif  // wxUSE_STD_IOSTREAM

 
 // ---------------------------------------------------------------------------
 // wxCStrData implementation
 
 // ---------------------------------------------------------------------------
 // wxCStrData implementation


@@ -2809,7 +3232,7 @@ inline wxCStrData::wxCStrData(const wxCStrData& data)
 inline wxCStrData::~wxCStrData()
 {
     if ( m_owned )
 inline wxCStrData::~wxCStrData()
 {
     if ( m_owned )

-        delete m_str;
+        delete wx_const_cast(wxString*, m_str); // cast to silence warnings

 }
 
 // simple cases for AsChar() and AsWChar(), the complicated ones are
 }
 
 // simple cases for AsChar() and AsWChar(), the complicated ones are


@@ -2828,6 +3251,13 @@ inline const char* wxCStrData::AsChar() const
 }
 #endif // !wxUSE_UNICODE
 
 }
 #endif // !wxUSE_UNICODE
 

+#if wxUSE_UTF8_LOCALE_ONLY
+inline const char* wxCStrData::AsChar() const
+{
+    return wxStringOperations::AddToIter(m_str->wx_str(), m_offset);
+}
+#endif // wxUSE_UTF8_LOCALE_ONLY
+

 inline const wxCharBuffer wxCStrData::AsCharBuf() const
 {
 #if !wxUSE_UNICODE
 inline const wxCharBuffer wxCStrData::AsCharBuf() const
 {
 #if !wxUSE_UNICODE


@@ -2856,7 +3286,11 @@ inline wxString wxCStrData::AsString() const
 
 inline const wxStringCharType *wxCStrData::AsInternal() const
 {
 
 inline const wxStringCharType *wxCStrData::AsInternal() const
 {

+#if wxUSE_UNICODE_UTF8

     return wxStringOperations::AddToIter(m_str->wx_str(), m_offset);
     return wxStringOperations::AddToIter(m_str->wx_str(), m_offset);

+#else
+    return m_str->wx_str() + m_offset;
+#endif

 }
 
 inline wxUniChar wxCStrData::operator*() const
 }
 
 inline wxUniChar wxCStrData::operator*() const


@@ -2905,12 +3339,54 @@ inline wxWCharBuffer::wxWCharBuffer(const wxCStrData& cstr)
 {
 }
 
 {
 }
 

+#if wxUSE_UNICODE_UTF8
+// ----------------------------------------------------------------------------
+// implementation of wxStringIteratorNode inline methods
+// ----------------------------------------------------------------------------
+
+void wxStringIteratorNode::DoSet(const wxString *str,
+                                 wxStringImpl::const_iterator *citer,
+                                 wxStringImpl::iterator *iter)
+{
+    m_prev = NULL;
+    m_iter = iter;
+    m_citer = citer;
+    m_str = str;
+    if ( str )
+    {
+        m_next = str->m_iterators.ptr;
+        wx_const_cast(wxString*, m_str)->m_iterators.ptr = this;
+        if ( m_next )
+            m_next->m_prev = this;
+    }
+    else
+    {
+        m_next = NULL;
+    }
+}
+
+void wxStringIteratorNode::clear()
+{
+    if ( m_next )
+        m_next->m_prev = m_prev;
+    if ( m_prev )
+        m_prev->m_next = m_next;
+    else if ( m_str ) // first in the list
+        wx_const_cast(wxString*, m_str)->m_iterators.ptr = m_next;
+
+    m_next = m_prev = NULL;
+    m_citer = NULL;
+    m_iter = NULL;
+    m_str = NULL;
+}
+#endif // wxUSE_UNICODE_UTF8
+

 #if WXWIN_COMPATIBILITY_2_8
 #if WXWIN_COMPATIBILITY_2_8

-    // lot of code out there doesn't explicitly include wx/wxchar.h, but uses
+    // lot of code out there doesn't explicitly include wx/crt.h, but uses

     // CRT wrappers that are now declared in wx/wxcrt.h and wx/wxcrtvararg.h,
     // so let's include this header now that wxString is defined and it's safe
     // to do it:
     // CRT wrappers that are now declared in wx/wxcrt.h and wx/wxcrtvararg.h,
     // so let's include this header now that wxString is defined and it's safe
     // to do it:

-    #include "wx/wxchar.h"
+    #include "wx/crt.h"

 #endif
 
 #endif  // _WX_WXSTRING_H_
 #endif
 
 #endif  // _WX_WXSTRING_H_