1 files changed, 24 insertions, 5 deletions

diff --git a/src/utfstring.h b/src/utfstring.h
index 3bdf51c..6f85e93 100644
--- a/src/utfstring.h
+++ b/src/utfstring.h
@@ -12,21 +12,40 @@
 
 namespace Bu
 {
-        class String;
+        /**
+         * UtfChar isn't actually a character, unicode specifies "code points" not
+         * characters.  The main reason for this is that not all code points define
+         * usable characters.  Some control text directionality, some apply
+         * properties to other code points which are characters.  However, most of
+         * these distinctions are only important when implementing displays that
+         * comply with the Unicode standard fully.
+         */
+        typedef uint32_t UtfChar;
 
         class UtfString
         {
         public:
+                enum Encoding
+                {
+                        Utf8,
+                        Utf16,
+                        Utf16be,
+                        Utf16le,
+                        Utf32,
+                        Ucs16,
+                        GuessEncoding
+                };
+
                 UtfString();
+                UtfString( const Bu::String &sInput, Encoding eEnc=Utf8 );
                 virtual ~UtfString();
 
-                typedef uint32_t point;
-
                 static void debugUtf8( const Bu::String &sUtf8 );
 
         private:
-//              typedef BasicString<uint16_t> RawString;
-//              RawString rsStore;
+                uint16_t *pData;
+                int iRawLen;
+                int iCharLen;
         };
 };