Merge remote-tracking branch 'origin/3.6'

18 files changed, 1229 insertions, 264 deletions

diff --git a/cpp/src/IceGrid/.gitignore b/cpp/src/IceGrid/.gitignore
index dd791e621d9..41d881d4e8a 100644
--- a/cpp/src/IceGrid/.gitignore
+++ b/cpp/src/IceGrid/.gitignore
@@ -1,6 +1,8 @@
 // Generated by makegitignore.py
 
 // IMPORTANT: Do not edit this file -- any edits made here will be lost!
+DBTypes.cpp
+DBTypes.h
 IceLocatorDiscovery.cpp
 IceLocatorDiscovery.h
 Internal.cpp
diff --git a/cpp/src/IceStorm/.gitignore b/cpp/src/IceStorm/.gitignore
index 6ede462d77f..4fad7b4af9d 100644
--- a/cpp/src/IceStorm/.gitignore
+++ b/cpp/src/IceStorm/.gitignore
@@ -1,6 +1,8 @@
 // Generated by makegitignore.py
 
 // IMPORTANT: Do not edit this file -- any edits made here will be lost!
+DBTypes.h
+DBTypes.cpp
 Instrumentation.cpp
 Election.cpp
 IceStormInternal.cpp
diff --git a/cpp/src/IceUtil/Unicode.cpp b/cpp/src/IceUtil/Unicode.cpp
index cae3476e277..7bad1d67c17 100644
--- a/cpp/src/IceUtil/Unicode.cpp
+++ b/cpp/src/IceUtil/Unicode.cpp
@@ -128,4 +128,42 @@ IceUtilInternal::convertUTF8ToUTFWstring(const Byte*& sourceStart, const Byte* s
     return result;
 }
 
+ConversionResult
+IceUtilInternal::convertUTF8ToUTF16(const vector<unsigned char>& source, vector<unsigned short>& target, ConversionFlags flags)
+{
+    target.resize(source.size());
+    const unsigned char* sourceStart = &source[0];
+    const unsigned char* sourceEnd = &source[0] + source.size();
+    
+    unsigned short* targetStart = &target[0]; 
+    unsigned short* targetEnd = &target[0] + target.size();
+    ConversionResult result = ConvertUTF8toUTF16(&sourceStart, sourceEnd, &targetStart, targetEnd, flags);
+    
+    if(result == conversionOK)
+    {
+        target.resize(targetStart - &target[0]);
+    }
+    return result;
+}
+
+ConversionResult
+IceUtilInternal::convertUTF32ToUTF8(const vector<unsigned int>& source, vector<unsigned char>& target, ConversionFlags flags)
+{
+    target.resize(source.size() * 4);
+
+    const unsigned int* sourceStart = &source[0];
+    const unsigned int* sourceEnd = &source[0] + source.size();
+    
+    unsigned char* targetStart = &target[0]; 
+    unsigned char* targetEnd = &target[0] + target.size();
+    ConversionResult result = ConvertUTF32toUTF8(&sourceStart, sourceEnd, &targetStart, targetEnd, flags);
+    
+    if(result == conversionOK)
+    {
+        target.resize(targetStart - &target[0]);
+    }
+    return result;
+}
+
+
 
diff --git a/cpp/src/IceUtil/Unicode.h b/cpp/src/IceUtil/Unicode.h
index 00333ce8a44..2c96d6c6448 100644
--- a/cpp/src/IceUtil/Unicode.h
+++ b/cpp/src/IceUtil/Unicode.h
@@ -44,6 +44,15 @@ ConversionResult
 convertUTF8ToUTFWstring(const IceUtil::Byte*& sourceStart, const IceUtil::Byte* sourceEnd, 
                         std::wstring& target, IceUtil::ConversionFlags flags);
 
+
+ICE_UTIL_API ConversionResult
+convertUTF8ToUTF16(const std::vector<unsigned char>&, std::vector<unsigned short>&,
+                   IceUtil::ConversionFlags);
+
+ICE_UTIL_API ConversionResult
+convertUTF32ToUTF8(const std::vector<unsigned int>&, std::vector<unsigned char>&,
+                   IceUtil::ConversionFlags);
+
 }
 
 #endif
diff --git a/cpp/src/Slice/PythonUtil.cpp b/cpp/src/Slice/PythonUtil.cpp
index a3abf0de03a..23b95898317 100644
--- a/cpp/src/Slice/PythonUtil.cpp
+++ b/cpp/src/Slice/PythonUtil.cpp
@@ -13,6 +13,7 @@
 #include <IceUtil/IceUtil.h>
 #include <IceUtil/StringUtil.h>
 #include <IceUtil/InputUtil.h>
+#include <IceUtil/Unicode.h>
 #include <climits>
 #include <iterator>
 
@@ -1881,68 +1882,138 @@ Slice::Python::CodeVisitor::writeConstantValue(const TypePtr& type, const Syntax
 
                 _out << "\"";                                       // Opening "
 
-                for(string::const_iterator c = value.begin(); c != value.end(); ++c)
+                for(size_t i = 0; i < value.size();)
                 {
-                    switch(*c)
+                    char c = value[i];
+                    switch(c)
                     {
-                    case '"':
-                    {
-                        _out << "\\\"";
-                        break;
-                    }
-                    case '\\':
-                    {
-                        _out << "\\\\";
-                        break;
-                    }
-                    case '\r':
-                    {
-                        _out << "\\r";
-                        break;
-                    }
-                    case '\n':
-                    {
-                        _out << "\\n";
-                        break;
-                    }
-                    case '\t':
-                    {
-                        _out << "\\t";
-                        break;
-                    }
-                    case '\b':
-                    {
-                        _out << "\\b";
-                        break;
-                    }
-                    case '\f':
-                    {
-                        _out << "\\f";
-                        break;
-                    }
-                    default:
-                    {
-                        if(charSet.find(*c) == charSet.end())
+                        case '"':
                         {
-                            unsigned char uc = *c;              // Char may be signed, so make it positive.
-                            stringstream s;
-                            s << "\\";                          // Print as octal if not in basic source character set.
-                            s.flags(ios_base::oct);
-                            s.width(3);
-                            s.fill('0');
-                            s << static_cast<unsigned>(uc);
-                            _out << s.str();
+                            _out << "\\\"";
+                            break;
                         }
-                        else
+                        case '\\':
                         {
-                            _out << *c;                         // Print normally if in basic source character set.
+                            string s = "\\";
+                            size_t j = i + 1;
+                            for(; j < value.size(); ++j)
+                            {
+                                if(value[j] != '\\')
+                                {
+                                    break;
+                                }
+                                s += "\\";
+                            }
+
+                            //
+                            // An even number of slash \ will escape the backslash and
+                            // the codepoint will be interpreted as its charaters
+                            //
+                            // \\u00000041  - ['\\', 'u', '0', '0', '0', '0', '0', '0', '4', '1']
+                            // \\\u00000041 - ['\\', 'A'] (41 is the codepoint for 'A')
+                            //
+                            if(s.size() % 2 != 0 && (value[j] == 'U' || value[j] == 'u'))
+                            {
+                                //
+                                // Convert codepoint to UTF8 bytes and write the escaped bytes
+                                //
+                                _out << s.substr(0, s.size() - 1);
+
+                                size_t sz = value[j] == 'U' ? 8 : 4;
+                                string codepoint = value.substr(j + 1, sz);
+                                assert(codepoint.size() ==  sz);
+
+                                IceUtil::Int64 v = IceUtilInternal::strToInt64(codepoint.c_str(), 0, 16);
+
+                                vector<unsigned int> u32buffer;
+                                u32buffer.push_back(static_cast<unsigned int>(v));
+
+                                vector<unsigned char> u8buffer;
+                                IceUtilInternal::ConversionResult result = convertUTF32ToUTF8(u32buffer, u8buffer, IceUtil::lenientConversion);
+                                switch(result)
+                                {
+                                    case conversionOK:
+                                        break;
+                                    case sourceExhausted:
+                                        throw IceUtil::IllegalConversionException(__FILE__, __LINE__, "string source exhausted");
+                                    case sourceIllegal:
+                                        throw IceUtil::IllegalConversionException(__FILE__, __LINE__, "string source illegal");
+                                    default:
+                                    {
+                                        assert(0);
+                                        throw IceUtil::IllegalConversionException(__FILE__, __LINE__);
+                                    }
+                                }
+
+                                ostringstream s;
+                                for(vector<unsigned char>::const_iterator q = u8buffer.begin(); q != u8buffer.end(); ++q)
+                                {
+                                    s << "\\";
+                                    s.fill('0');
+                                    s.width(3);
+                                    s << oct;
+                                    s << static_cast<unsigned int>(*q);
+                                }
+                                _out << s.str();
+
+                                i = j + 1 + sz;
+                            }
+                            else
+                            {
+                                _out << s;
+                                i = j;
+                            }
+                            continue;
+                        }
+                        case '\r':
+                        {
+                            _out << "\\r";
+                            break;
+                        }
+                        case '\n':
+                        {
+                            _out << "\\n";
+                            break;
+                        }
+                        case '\t':
+                        {
+                            _out << "\\t";
+                            break;
+                        }
+                        case '\b':
+                        {
+                            _out << "\\b";
+                            break;
+                        }
+                        case '\f':
+                        {
+                            _out << "\\f";
+                            break;
+                        }
+                        default:
+                        {
+                            if(charSet.find(c) == charSet.end())
+                            {
+                                unsigned char uc = c;               // Char may be signed, so make it positive.
+                                stringstream s;
+                                s << "\\";                          // Print as octal if not in basic source character set.
+                                s.flags(ios_base::oct);
+                                s.width(3);
+                                s.fill('0');
+                                s << static_cast<unsigned>(uc);
+                                _out << s.str();
+                            }
+                            else
+                            {
+                                _out << c;                          // Print normally if in basic source character set.
+                            }
+                            break;
                         }
-                        break;
-                    }
                     }
+                    ++i;
                 }
 
-                _out << "\"";                                   // Closing "
+                _out << "\"";                                       // Closing "
                 break;
             }
             case Slice::Builtin::KindValue:
diff --git a/cpp/src/Slice/RubyUtil.cpp b/cpp/src/Slice/RubyUtil.cpp
index 0c9cfd49bd5..2c066a8efbe 100644
--- a/cpp/src/Slice/RubyUtil.cpp
+++ b/cpp/src/Slice/RubyUtil.cpp
@@ -12,6 +12,7 @@
 #include <Slice/Util.h>
 #include <IceUtil/Functional.h>
 #include <IceUtil/InputUtil.h>
+#include <IceUtil/Unicode.h>
 #include <iterator>
 
 using namespace std;
@@ -1472,68 +1473,137 @@ Slice::Ruby::CodeVisitor::writeConstantValue(const TypePtr& type, const SyntaxTr
 
                 _out << "\"";                                      // Opening "
 
-                for(string::const_iterator c = value.begin(); c != value.end(); ++c)
+                for(size_t i = 0; i < value.size();)
                 {
-                    switch(*c)
+                    char c = value[i];
+                    switch(c)
                     {
-                    case '"':
-                    {
-                        _out << "\\\"";
-                        break;
-                    }
-                    case '\\':
-                    {
-                        _out << "\\\\";
-                        break;
-                    }
-                    case '\r':
-                    {
-                        _out << "\\r";
-                        break;
-                    }
-                    case '\n':
-                    {
-                        _out << "\\n";
-                        break;
-                    }
-                    case '\t':
-                    {
-                        _out << "\\t";
-                        break;
-                    }
-                    case '\b':
-                    {
-                        _out << "\\b";
-                        break;
-                    }
-                    case '\f':
-                    {
-                        _out << "\\f";
-                        break;
-                    }
-                    default:
-                    {
-                        if(charSet.find(*c) == charSet.end())
+                        case '"':
                         {
-                            unsigned char uc = *c;              // Char may be signed, so make it positive.
-                            stringstream s;
-                            s << "\\";                          // Print as octal if not in basic source character set.
-                            s.flags(ios_base::oct);
-                            s.width(3);
-                            s.fill('0');
-                            s << static_cast<unsigned>(uc);
-                            _out << s.str();
+                            _out << "\\\"";
+                            break;
                         }
-                        else
+                        case '\\':
                         {
-                            _out << *c;                         // Print normally if in basic source character set.
+                            string s = "\\";
+                            size_t j = i + 1;
+                            for(; j < value.size(); ++j)
+                            {
+                                if(value[j] != '\\')
+                                {
+                                    break;
+                                }
+                                s += "\\";
+                            }
+
+                            //
+                            // An even number of slash \ will escape the backslash and
+                            // the codepoint will be interpreted as its charaters
+                            //
+                            // \\u00000041  - ['\\', 'u', '0', '0', '0', '0', '0', '0', '4', '1']
+                            // \\\u00000041 - ['\\', 'A'] (41 is the codepoint for 'A')
+                            //
+                            if(s.size() % 2 != 0 && (value[j] == 'U' || value[j] == 'u'))
+                            {
+                                //
+                                // Convert codepoint to UTF8 bytes and write the escaped bytes
+                                //
+                                _out << s.substr(0, s.size() - 1);
+
+                                size_t sz = value[j] == 'U' ? 8 : 4;
+                                string codepoint = value.substr(j + 1, sz);
+                                assert(codepoint.size() == sz);
+                                IceUtil::Int64 v = IceUtilInternal::strToInt64(codepoint.c_str(), 0, 16);
+
+                                vector<unsigned int> u32buffer;
+                                u32buffer.push_back(static_cast<unsigned int>(v));
+
+                                vector<unsigned char> u8buffer;
+                                IceUtilInternal::ConversionResult result = convertUTF32ToUTF8(u32buffer, u8buffer, IceUtil::lenientConversion);
+                                switch(result)
+                                {
+                                    case conversionOK:
+                                        break;
+                                    case sourceExhausted:
+                                        throw IceUtil::IllegalConversionException(__FILE__, __LINE__, "string source exhausted");
+                                    case sourceIllegal:
+                                        throw IceUtil::IllegalConversionException(__FILE__, __LINE__, "string source illegal");
+                                    default:
+                                    {
+                                        assert(0);
+                                        throw IceUtil::IllegalConversionException(__FILE__, __LINE__);
+                                    }
+                                }
+
+                                ostringstream s;
+                                for(vector<unsigned char>::const_iterator q = u8buffer.begin(); q != u8buffer.end(); ++q)
+                                {
+                                    s << "\\";
+                                    s.fill('0');
+                                    s.width(3);
+                                    s << oct;
+                                    s << static_cast<unsigned int>(*q);
+                                }
+                                _out << s.str();
+
+                                i = j + 1 + sz;
+                            }
+                            else
+                            {
+                                _out << s;
+                                i = j;
+                            }
+                            continue;
+                        }
+                        case '\r':
+                        {
+                            _out << "\\r";
+                            break;
+                        }
+                        case '\n':
+                        {
+                            _out << "\\n";
+                            break;
+                        }
+                        case '\t':
+                        {
+                            _out << "\\t";
+                            break;
+                        }
+                        case '\b':
+                        {
+                            _out << "\\b";
+                            break;
+                        }
+                        case '\f':
+                        {
+                            _out << "\\f";
+                            break;
+                        }
+                        default:
+                        {
+                            if(charSet.find(c) == charSet.end())
+                            {
+                                unsigned char uc = c;              // Char may be signed, so make it positive.
+                                stringstream s;
+                                s << "\\";                         // Print as octal if not in basic source character set.
+                                s.flags(ios_base::oct);
+                                s.width(3);
+                                s.fill('0');
+                                s << static_cast<unsigned>(uc);
+                                _out << s.str();
+                            }
+                            else
+                            {
+                                _out << c;                         // Print normally if in basic source character set.
+                            }
+                            break;
                         }
-                        break;
-                    }
                     }
+                    ++i;
                 }
 
-                _out << "\"";                                   // Closing "
+                _out << "\".force_encoding(\"utf-8\")";                                   // Closing "
                 break;
             }
 
diff --git a/cpp/src/Slice/Scanner.cpp b/cpp/src/Slice/Scanner.cpp
index a49c6c0433a..27d362e04c9 100644
--- a/cpp/src/Slice/Scanner.cpp
+++ b/cpp/src/Slice/Scanner.cpp
@@ -568,6 +568,8 @@ char *slice_text;
 #include <Slice/Grammar.h>
 #include <IceUtil/InputUtil.h>
 
+#include <iomanip>
+
 #include <stdlib.h>
 #include <math.h>
 
@@ -630,7 +632,7 @@ int checkKeyword(string&);
 
 
 
-#line 633 "lex.yy.c"
+#line 635 "lex.yy.c"
 
 #define INITIAL 0
 #define BOMSCAN 1
@@ -849,10 +851,10 @@ YY_DECL
 		}
 
 	{
-#line 92 "Scanner.l"
+#line 94 "Scanner.l"
 
 
-#line 855 "lex.yy.c"
+#line 857 "lex.yy.c"
 
 	while ( 1 )		/* loops until end-of-file is reached */
 		{
@@ -911,7 +913,7 @@ case 1:
 (yy_c_buf_p) = yy_cp -= 1;
 YY_DO_BEFORE_ACTION; /* set up slice_text again */
 YY_RULE_SETUP
-#line 94 "Scanner.l"
+#line 96 "Scanner.l"
 {
     if(unit->scanPosition(slice_text))
     {
@@ -926,7 +928,7 @@ YY_LINENO_REWIND_TO(yy_cp - 1);
 (yy_c_buf_p) = yy_cp -= 1;
 YY_DO_BEFORE_ACTION; /* set up slice_text again */
 YY_RULE_SETUP
-#line 101 "Scanner.l"
+#line 103 "Scanner.l"
 {
     if(unit->scanPosition(slice_text))
     {
@@ -939,7 +941,7 @@ case 3:
 (yy_c_buf_p) = yy_cp -= 1;
 YY_DO_BEFORE_ACTION; /* set up slice_text again */
 YY_RULE_SETUP
-#line 108 "Scanner.l"
+#line 110 "Scanner.l"
 {
     if(unit->scanPosition(slice_text))
     {
@@ -954,7 +956,7 @@ YY_LINENO_REWIND_TO(yy_cp - 1);
 (yy_c_buf_p) = yy_cp -= 1;
 YY_DO_BEFORE_ACTION; /* set up slice_text again */
 YY_RULE_SETUP
-#line 115 "Scanner.l"
+#line 117 "Scanner.l"
 {
     if(unit->scanPosition(slice_text))
     {
@@ -964,7 +966,7 @@ YY_RULE_SETUP
 	YY_BREAK
 case 5:
 YY_RULE_SETUP
-#line 122 "Scanner.l"
+#line 124 "Scanner.l"
 {
     // C++-style comment
     BEGIN(MAINSCAN);
@@ -982,7 +984,7 @@ YY_RULE_SETUP
 	YY_BREAK
 case 6:
 YY_RULE_SETUP
-#line 137 "Scanner.l"
+#line 139 "Scanner.l"
 {
     // C-style comment
     BEGIN(MAINSCAN);
@@ -1026,7 +1028,7 @@ YY_RULE_SETUP
 	YY_BREAK
 case 7:
 YY_RULE_SETUP
-#line 178 "Scanner.l"
+#line 180 "Scanner.l"
 {
     BEGIN(MAINSCAN);
     return ICE_SCOPE_DELIMITER;
@@ -1034,7 +1036,7 @@ YY_RULE_SETUP
 	YY_BREAK
 case 8:
 YY_RULE_SETUP
-#line 183 "Scanner.l"
+#line 185 "Scanner.l"
 {
     BEGIN(MAINSCAN);
     return ICE_METADATA_OPEN;
@@ -1042,7 +1044,7 @@ YY_RULE_SETUP
 	YY_BREAK
 case 9:
 YY_RULE_SETUP
-#line 188 "Scanner.l"
+#line 190 "Scanner.l"
 {
     BEGIN(MAINSCAN);
     return ICE_METADATA_CLOSE;
@@ -1050,7 +1052,7 @@ YY_RULE_SETUP
 	YY_BREAK
 case 10:
 YY_RULE_SETUP
-#line 193 "Scanner.l"
+#line 195 "Scanner.l"
 {
     BEGIN(MAINSCAN);
     return ICE_GLOBAL_METADATA_OPEN;
@@ -1058,7 +1060,7 @@ YY_RULE_SETUP
 	YY_BREAK
 case 11:
 YY_RULE_SETUP
-#line 198 "Scanner.l"
+#line 200 "Scanner.l"
 {
     BEGIN(MAINSCAN);
     return ICE_GLOBAL_METADATA_CLOSE;
@@ -1067,7 +1069,7 @@ YY_RULE_SETUP
 case 12:
 /* rule 12 can match eol */
 YY_RULE_SETUP
-#line 203 "Scanner.l"
+#line 205 "Scanner.l"
 {
     BEGIN(MAINSCAN);
     StringTokPtr ident = new StringTok;
@@ -1095,7 +1097,7 @@ YY_RULE_SETUP
 	YY_BREAK
 case 13:
 YY_RULE_SETUP
-#line 228 "Scanner.l"
+#line 230 "Scanner.l"
 {
     BEGIN(MAINSCAN);
     StringTokPtr ident = new StringTok;
@@ -1106,7 +1108,7 @@ YY_RULE_SETUP
 	YY_BREAK
 case 14:
 YY_RULE_SETUP
-#line 236 "Scanner.l"
+#line 238 "Scanner.l"
 {
     BEGIN(MAINSCAN);
     StringTokPtr str = new StringTok;
@@ -1135,32 +1137,42 @@ YY_RULE_SETUP
             switch(next)
             {
                 case '\\':
+                {
+                    str->v += '\\';
+                    str->v += '\\';
+                    break;
+                }
                 case '"':
                 case '\'':
                 {
                     str->v += next;
                     break;
                 }
+            
                 case 'n':
                 {
                     str->v += '\n';
                     break;
                 }
+            
                 case 'r':
                 {
                     str->v += '\r';
                     break;
                 }
+
                 case 't':
                 {
                     str->v += '\t';
                     break;
                 }
+            
                 case 'v':
                 {
                     str->v += '\v';
                     break;
                 }
+            
                 case 'f':
                 {
                     str->v += '\f';
@@ -1185,68 +1197,181 @@ YY_RULE_SETUP
                     break;
                 }
 
+                //
+                // Octal value \nnn limited to three octal digits but terminate at the first character 
+                // that is not a valid octal digit if encountered sooner.
+                //
                 case '0':
                 case '1':
                 case '2':
                 case '3':
+                case '4':
+                case '5':
+                case '7':
                 {
                     static string octalDigits = "01234567";
-                    unsigned short us = next - '0';
-                    if(octalDigits.find_first_of(next = static_cast<char>(yyinput())) != string::npos)
+                    IceUtil::Int64 value = 0;
+                    string escape;
+                    escape += next;
+                    for(int i = 0; i < 2; ++i)
                     {
-                        str->literal += next;
-                        us = us * 8 + next - '0';
-                        if(octalDigits.find_first_of(next = static_cast<char>(yyinput())) != string::npos)
-                        {
-                            us = us * 8 + next - '0';
-                        }
-                        else
+                        next = static_cast<char>(yyinput());
+                        if(octalDigits.find_first_of(next) == string::npos)
                         {
                             unput(next);
+                            break;
                         }
+                        escape += next;                        
                     }
-                    else
+                    str->literal += escape;
+                    value = IceUtilInternal::strToInt64(escape.c_str(), 0, 8);
+                    
+                    if(value == 0)
                     {
-                        unput(next);
+                        unit->error("illegal NUL character in string constant");
                     }
-                    if(us == 0)
+                    else if(value > 255)
                     {
-                        unit->error("illegal NUL character in string constant");
+                        ostringstream os;
+                        os << "octal escape sequence out of range: '\\" << oct << value << "'";
+                        unit->warning(os.str());
                     }
-                    str->v += static_cast<char>(us);
+                    str->v += static_cast<char>(value);
                     break;
                 }
                 case 'x':
                 {
-                    IceUtil::Int64 ull = 0;
+                    IceUtil::Int64 value = 0;
+                    string escape = "";
                     while(isxdigit(static_cast<unsigned char>(next = static_cast<char>(yyinput()))))
                     {
+                        escape += next;
+                    }
+                    unput(next);
+                    
+                    str->literal += escape;
+                    value = IceUtilInternal::strToInt64(escape.c_str(), 0, 16);
+
+                    if(value == 0)
+                    {
+                        unit->error("illegal NUL character in string constant");
+                    }
+                    else if(value > 255)
+                    {
+                        ostringstream os;
+                        os << "hex escape sequence out of range: '\\x" << hex << value << "'";
+                        unit->warning(os.str());
+                    }
+                    str->v += static_cast<char>(value);
+                    break;
+                }
+        
+                //
+                // Universal character name \unnnn code point U+nnnn
+                //
+                case 'u':
+                {
+                    IceUtil::Int64 value = 0;
+                    string escape = "";
+
+                    for(int i = 0; i < 4; ++i)
+                    {
+                        next = static_cast<char>(yyinput());
                         str->literal += next;
-                        ull *= 16;
-                        if(isdigit(static_cast<unsigned char>(next)))
+                        if(!isxdigit(static_cast<unsigned char>(next)))
                         {
-                            ull += next - '0';
+                            unit->error("unknown escape sequence in string constant: " + str->literal);
+                            break;
                         }
-                        else if(islower(static_cast<unsigned char>(next)))
-                        {
-                            ull += next - 'a' + 10;
-                        }
-                        else
+                        escape += next;
+                    }
+                    
+                    value = escape.size() == 4 ? IceUtilInternal::strToInt64(escape.c_str(), 0, 16) : -1;
+
+                    ostringstream os;
+                    os << '\\' << 'u';
+                    os.fill('0');
+                    os.width(4);
+                    os << hex << value;
+                    
+                    if(value == 0)
+                    {
+                        unit->error("illegal NUL character in string constant");
+                    }
+                    
+                    
+                    //
+                    // Determine if a character is a surrogate:
+                    // 
+                    // * High surrogate code point, ranging from 0xd800 to 0xdbff, inclusive
+                    // * Low surrogate code point, ranging from 0xdc00 to 0xdfff, inclusive.
+                    //
+                    else if((value >= 0xd800 && value <= 0xdbff) || (value >= 0xdc00 && value <= 0xdfff))
+                    {
+                        unit->error("unknown escape sequence in string constant: '" + os.str() + "'");
+                    }
+
+                    str->v += os.str();
+                    
+                    break;
+                }
+                
+                case 'U':
+                {
+                    IceUtil::Int64 value = 0;
+                    string escape = "";
+                    
+                    for(int i = 0; i < 8; ++i)
+                    {
+                        next = static_cast<char>(yyinput());
+                        str->literal += next;
+                        if(!isxdigit(static_cast<unsigned char>(next)))
                         {
-                            ull += next - 'A' + 10;
+
+                            unit->error("unknown escape sequence in string constant: " + str->literal);
+                            break;
                         }
+                        escape += next;
                     }
-                    unput(next);
-                    if(ull == 0)
+                    
+                    value = escape.size() == 8 ? IceUtilInternal::strToInt64(escape.c_str(), 0, 16) : -1;
+
+                    ostringstream os;
+                    os << '\\' << 'U';
+                    os.fill('0');
+                    os.width(8);
+                    os << hex << value;
+                    
+                    if(value == 0)
                     {
                         unit->error("illegal NUL character in string constant");
                     }
-                    str->v += static_cast<char>(ull);
+                    
+                    //
+                    // Determine if a character is a surrogate:
+                    // 
+                    // * High surrogate code point, ranging from 0xd800 to 0xdbff, inclusive
+                    // * Low surrogate code point, ranging from 0xdc00 to 0xdfff, inclusive.
+                    //
+                    else if((value >= 0xd800 && value <= 0xdbff) || (value >= 0xdc00 && value <= 0xdfff))
+                    {
+                        unit->error("unknown escape sequence in string constant: '" + os.str() + "'");
+                    }
+
+                    str->v += os.str();
                     break;
                 }
-                // TODO: add universal character names
+
                 default:
                 {
+                    ostringstream os;
+                    os << "unknown escape sequence '\\" << next << "'";
+                    unit->warning(os.str());
+                    //
+                    // We escape the backslack in a unknown escape sequence 
+                    // to keep compativility with 3.6"
+                    //
+                    str->v += '\\';
                     str->v += c;
                     unput(next);
                 }
@@ -1263,7 +1388,7 @@ YY_RULE_SETUP
 	YY_BREAK
 case 15:
 YY_RULE_SETUP
-#line 390 "Scanner.l"
+#line 515 "Scanner.l"
 {
     BEGIN(MAINSCAN);
     IntegerTokPtr itp = new IntegerTok;
@@ -1282,7 +1407,7 @@ YY_RULE_SETUP
 	YY_BREAK
 case 16:
 YY_RULE_SETUP
-#line 406 "Scanner.l"
+#line 531 "Scanner.l"
 {
     BEGIN(MAINSCAN);
     errno = 0;
@@ -1316,7 +1441,7 @@ YY_RULE_SETUP
 case 17:
 /* rule 17 can match eol */
 YY_RULE_SETUP
-#line 436 "Scanner.l"
+#line 561 "Scanner.l"
 {
     // Ignore white-space
 
@@ -1332,7 +1457,7 @@ YY_RULE_SETUP
 	YY_BREAK
 case 18:
 YY_RULE_SETUP
-#line 449 "Scanner.l"
+#line 574 "Scanner.l"
 {
     // Ignore UTF-8 BOM, rule only active when parsing start of file.
 
@@ -1341,7 +1466,7 @@ YY_RULE_SETUP
 	YY_BREAK
 case 19:
 YY_RULE_SETUP
-#line 455 "Scanner.l"
+#line 580 "Scanner.l"
 {
     BEGIN(MAINSCAN);
     if(slice_text[0] < 32 || slice_text[0] > 126)
@@ -1360,10 +1485,10 @@ YY_RULE_SETUP
 	YY_BREAK
 case 20:
 YY_RULE_SETUP
-#line 471 "Scanner.l"
+#line 596 "Scanner.l"
 ECHO;
 	YY_BREAK
-#line 1366 "lex.yy.c"
+#line 1491 "lex.yy.c"
 case YY_STATE_EOF(INITIAL):
 case YY_STATE_EOF(BOMSCAN):
 case YY_STATE_EOF(MAINSCAN):
@@ -2364,7 +2489,7 @@ void slice_free (void * ptr )
 
 #define YYTABLES_NAME "yytables"
 
-#line 470 "Scanner.l"
+#line 595 "Scanner.l"
 
 
 
diff --git a/cpp/src/Slice/Scanner.l b/cpp/src/Slice/Scanner.l
index 190c00bcf5c..a9c381b7260 100644
--- a/cpp/src/Slice/Scanner.l
+++ b/cpp/src/Slice/Scanner.l
@@ -13,6 +13,8 @@
 #include <Slice/Grammar.h>
 #include <IceUtil/InputUtil.h>
 
+#include <iomanip>
+
 #include <stdlib.h>
 #include <math.h>
 
@@ -261,32 +263,42 @@ floating_literal    (({fractional_constant}{exponent_part}?)|((\+|-)?[[:digit:]]
             switch(next)
             {
                 case '\\':
+                {
+                    str->v += '\\';
+                    str->v += '\\';
+                    break;
+                }
                 case '"':
                 case '\'':
                 {
                     str->v += next;
                     break;
                 }
+            
                 case 'n':
                 {
                     str->v += '\n';
                     break;
                 }
+            
                 case 'r':
                 {
                     str->v += '\r';
                     break;
                 }
+
                 case 't':
                 {
                     str->v += '\t';
                     break;
                 }
+            
                 case 'v':
                 {
                     str->v += '\v';
                     break;
                 }
+            
                 case 'f':
                 {
                     str->v += '\f';
@@ -311,68 +323,181 @@ floating_literal    (({fractional_constant}{exponent_part}?)|((\+|-)?[[:digit:]]
                     break;
                 }
 
+                //
+                // Octal value \nnn limited to three octal digits but terminate at the first character 
+                // that is not a valid octal digit if encountered sooner.
+                //
                 case '0':
                 case '1':
                 case '2':
                 case '3':
+                case '4':
+                case '5':
+                case '7':
                 {
                     static string octalDigits = "01234567";
-                    unsigned short us = next - '0';
-                    if(octalDigits.find_first_of(next = static_cast<char>(yyinput())) != string::npos)
+                    IceUtil::Int64 value = 0;
+                    string escape;
+                    escape += next;
+                    for(int i = 0; i < 2; ++i)
                     {
-                        str->literal += next;
-                        us = us * 8 + next - '0';
-                        if(octalDigits.find_first_of(next = static_cast<char>(yyinput())) != string::npos)
-                        {
-                            us = us * 8 + next - '0';
-                        }
-                        else
+                        next = static_cast<char>(yyinput());
+                        if(octalDigits.find_first_of(next) == string::npos)
                         {
                             unput(next);
+                            break;
                         }
+                        escape += next;                        
                     }
-                    else
+                    str->literal += escape;
+                    value = IceUtilInternal::strToInt64(escape.c_str(), 0, 8);
+                    
+                    if(value == 0)
                     {
-                        unput(next);
+                        unit->error("illegal NUL character in string constant");
                     }
-                    if(us == 0)
+                    else if(value > 255)
                     {
-                        unit->error("illegal NUL character in string constant");
+                        ostringstream os;
+                        os << "octal escape sequence out of range: '\\" << oct << value << "'";
+                        unit->warning(os.str());
                     }
-                    str->v += static_cast<char>(us);
+                    str->v += static_cast<char>(value);
                     break;
                 }
                 case 'x':
                 {
-                    IceUtil::Int64 ull = 0;
+                    IceUtil::Int64 value = 0;
+                    string escape = "";
                     while(isxdigit(static_cast<unsigned char>(next = static_cast<char>(yyinput()))))
                     {
+                        escape += next;
+                    }
+                    unput(next);
+                    
+                    str->literal += escape;
+                    value = IceUtilInternal::strToInt64(escape.c_str(), 0, 16);
+
+                    if(value == 0)
+                    {
+                        unit->error("illegal NUL character in string constant");
+                    }
+                    else if(value > 255)
+                    {
+                        ostringstream os;
+                        os << "hex escape sequence out of range: '\\x" << hex << value << "'";
+                        unit->warning(os.str());
+                    }
+                    str->v += static_cast<char>(value);
+                    break;
+                }
+        
+                //
+                // Universal character name \unnnn code point U+nnnn
+                //
+                case 'u':
+                {
+                    IceUtil::Int64 value = 0;
+                    string escape = "";
+
+                    for(int i = 0; i < 4; ++i)
+                    {
+                        next = static_cast<char>(yyinput());
                         str->literal += next;
-                        ull *= 16;
-                        if(isdigit(static_cast<unsigned char>(next)))
-                        {
-                            ull += next - '0';
-                        }
-                        else if(islower(static_cast<unsigned char>(next)))
+                        if(!isxdigit(static_cast<unsigned char>(next)))
                         {
-                            ull += next - 'a' + 10;
+                            unit->error("unknown escape sequence in string constant: " + str->literal);
+                            break;
                         }
-                        else
+                        escape += next;
+                    }
+                    
+                    value = escape.size() == 4 ? IceUtilInternal::strToInt64(escape.c_str(), 0, 16) : -1;
+
+                    ostringstream os;
+                    os << '\\' << 'u';
+                    os.fill('0');
+                    os.width(4);
+                    os << hex << value;
+                    
+                    if(value == 0)
+                    {
+                        unit->error("illegal NUL character in string constant");
+                    }
+                    
+                    
+                    //
+                    // Determine if a character is a surrogate:
+                    // 
+                    // * High surrogate code point, ranging from 0xd800 to 0xdbff, inclusive
+                    // * Low surrogate code point, ranging from 0xdc00 to 0xdfff, inclusive.
+                    //
+                    else if((value >= 0xd800 && value <= 0xdbff) || (value >= 0xdc00 && value <= 0xdfff))
+                    {
+                        unit->error("unknown escape sequence in string constant: '" + os.str() + "'");
+                    }
+
+                    str->v += os.str();
+                    
+                    break;
+                }
+                
+                case 'U':
+                {
+                    IceUtil::Int64 value = 0;
+                    string escape = "";
+                    
+                    for(int i = 0; i < 8; ++i)
+                    {
+                        next = static_cast<char>(yyinput());
+                        str->literal += next;
+                        if(!isxdigit(static_cast<unsigned char>(next)))
                         {
-                            ull += next - 'A' + 10;
+
+                            unit->error("unknown escape sequence in string constant: " + str->literal);
+                            break;
                         }
+                        escape += next;
                     }
-                    unput(next);
-                    if(ull == 0)
+                    
+                    value = escape.size() == 8 ? IceUtilInternal::strToInt64(escape.c_str(), 0, 16) : -1;
+
+                    ostringstream os;
+                    os << '\\' << 'U';
+                    os.fill('0');
+                    os.width(8);
+                    os << hex << value;
+                    
+                    if(value == 0)
                     {
                         unit->error("illegal NUL character in string constant");
                     }
-                    str->v += static_cast<char>(ull);
+                    
+                    //
+                    // Determine if a character is a surrogate:
+                    // 
+                    // * High surrogate code point, ranging from 0xd800 to 0xdbff, inclusive
+                    // * Low surrogate code point, ranging from 0xdc00 to 0xdfff, inclusive.
+                    //
+                    else if((value >= 0xd800 && value <= 0xdbff) || (value >= 0xdc00 && value <= 0xdfff))
+                    {
+                        unit->error("unknown escape sequence in string constant: '" + os.str() + "'");
+                    }
+
+                    str->v += os.str();
                     break;
                 }
-                // TODO: add universal character names
+
                 default:
                 {
+                    ostringstream os;
+                    os << "unknown escape sequence '\\" << next << "'";
+                    unit->warning(os.str());
+                    //
+                    // We escape the backslack in a unknown escape sequence 
+                    // to keep compativility with 3.6"
+                    //
+                    str->v += '\\';
                     str->v += c;
                     unput(next);
                 }
diff --git a/cpp/src/slice2cpp/Gen.cpp b/cpp/src/slice2cpp/Gen.cpp
index 98dc43e28aa..c90dc87088b 100644
--- a/cpp/src/slice2cpp/Gen.cpp
+++ b/cpp/src/slice2cpp/Gen.cpp
@@ -13,6 +13,8 @@
 #include <Slice/CPlusPlusUtil.h>
 #include <IceUtil/Functional.h>
 #include <IceUtil/Iterator.h>
+#include <IceUtil/InputUtil.h>
+#include <IceUtil/Unicode.h>
 #include <Slice/Checksum.h>
 #include <Slice/FileTracker.h>
 
@@ -126,11 +128,11 @@ writeConstantValue(IceUtilInternal::Output& out, const TypePtr& type, const Synt
             }
             out << "\"";                                    // Opening "
 
-            for(string::const_iterator c = value.begin(); c != value.end(); ++c)
+            for(size_t i = 0; i < value.size();)
             {
-                if(charSet.find(*c) == charSet.end())
+                if(charSet.find(value[i]) == charSet.end())
                 {
-                    unsigned char uc = *c;                  // char may be signed, so make it positive
+                    unsigned char uc = value[i];                  // char may be signed, so make it positive
                     ostringstream s;
                     s << "\\";                              // Print as octal if not in basic source character set
                     s.width(3);
@@ -141,17 +143,93 @@ writeConstantValue(IceUtilInternal::Output& out, const TypePtr& type, const Synt
                 }
                 else
                 {
-                    switch(*c)
+                    switch(value[i])
                     {
                         case '\\':
+                        {
+                            string s = "\\";
+                            size_t j = i + 1;
+                            for(; j < value.size(); ++j)
+                            {
+                                if(value[j] != '\\')
+                                {
+                                    break;
+                                }
+                                s += "\\";
+                            }
+
+                            //
+                            // An even number of slash \ will escape the backslash and
+                            // the codepoint will be interpreted as its charaters
+                            //
+                            // \\U00000041  - ['\\', 'U', '0', '0', '0', '0', '0', '0', '4', '1']
+                            // \\\U00000041 - ['\\', 'A'] (41 is the codepoint for 'A')
+                            //
+                            if(s.size() % 2 != 0 && (value[j] == 'U' || value[j] == 'u'))
+                            {
+                                //
+                                // Convert codepoint to UTF8 bytes and write the escaped bytes
+                                //
+                                out << s.substr(0, s.size() - 1);
+
+                                size_t sz = value[j] == 'U' ? 8 : 4;
+                                string codepoint = value.substr(j + 1, sz);
+                                assert(codepoint.size() ==  sz);
+
+                                IceUtil::Int64 v = IceUtilInternal::strToInt64(codepoint.c_str(), 0, 16);
+
+
+                                vector<unsigned int> u32buffer;
+                                u32buffer.push_back(static_cast<unsigned int>(v));
+
+                                vector<unsigned char> u8buffer;
+
+                                IceUtilInternal::ConversionResult result = convertUTF32ToUTF8(u32buffer, u8buffer, IceUtil::lenientConversion);
+                                switch(result)
+                                {
+                                    case conversionOK:
+                                        break;
+                                    case sourceExhausted:
+                                        throw IceUtil::IllegalConversionException(__FILE__, __LINE__, "string source exhausted");
+                                    case sourceIllegal:
+                                        throw IceUtil::IllegalConversionException(__FILE__, __LINE__, "string source illegal");
+                                    default:
+                                    {
+                                        assert(0);
+                                        throw IceUtil::IllegalConversionException(__FILE__, __LINE__);
+                                    }
+                                }
+
+                                ostringstream s;
+                                for(vector<unsigned char>::const_iterator q = u8buffer.begin(); q != u8buffer.end(); ++q)
+                                {
+                                    s << "\\";
+                                    s.fill('0');
+                                    s.width(3);
+                                    s << oct;
+                                    s << static_cast<unsigned int>(*q);
+                                }
+                                out << s.str();
+
+                                i = j + 1 + sz;
+                            }
+                            else
+                            {
+                                out << s;
+                                i = j;
+                            }
+                            continue;
+                        }
                         case '"':
                         {
                             out << "\\";
                             break;
                         }
                     }
-                    out << *c;                              // Print normally if in basic source character set
+                    
+                    out << value[i];                              // Print normally if in basic source character set
                 }
+                ++i;
             }
 
             out << "\"";                                    // Closing "
diff --git a/cpp/src/slice2cpp/Makefile b/cpp/src/slice2cpp/Makefile
index 790ad0aecc3..3af754207f8 100644
--- a/cpp/src/slice2cpp/Makefile
+++ b/cpp/src/slice2cpp/Makefile
@@ -20,7 +20,7 @@ RPATH_DIR	= $(LOADER_PATH)/../$(libsubdir)
 
 include $(top_srcdir)/config/Make.rules
 
-CPPFLAGS	:= -I. $(CPPFLAGS)
+CPPFLAGS	:= -I. -I.. $(CPPFLAGS)
 
 $(NAME): $(OBJS)
 	rm -f $@
diff --git a/cpp/src/slice2cs/Gen.cpp b/cpp/src/slice2cs/Gen.cpp
index 50076dc087c..85f10179740 100644
--- a/cpp/src/slice2cs/Gen.cpp
+++ b/cpp/src/slice2cs/Gen.cpp
@@ -10,7 +10,9 @@
 #include <IceUtil/DisableWarnings.h>
 #include <IceUtil/Functional.h>
 #include <IceUtil/StringUtil.h>
-#include "Gen.h"
+#include <IceUtil/InputUtil.h>
+#include <Gen.h>
+
 #include <limits>
 #include <sys/stat.h>
 #ifndef _WIN32
@@ -20,6 +22,7 @@
 #endif
 #include <IceUtil/Iterator.h>
 #include <IceUtil/UUID.h>
+#include <IceUtil/Unicode.h>
 #include <Slice/Checksum.h>
 #include <Slice/DotNetNames.h>
 #include <Slice/FileTracker.h>
@@ -35,6 +38,45 @@ namespace
 {
 
 string
+u16CodePoint(unsigned short value)
+{
+    ostringstream s;
+    s << "\\u";
+    s << hex;
+    s.width(4);
+    s.fill('0');
+    s << value;
+    return s.str();
+}
+
+
+void
+writeU8Buffer(const vector<unsigned char>& u8buffer, ::IceUtilInternal::Output& out)
+{
+    vector<unsigned short> u16buffer;
+    IceUtilInternal::ConversionResult result = convertUTF8ToUTF16(u8buffer, u16buffer, IceUtil::lenientConversion);
+    switch(result)
+    {
+        case conversionOK:
+            break;
+        case sourceExhausted:
+            throw IceUtil::IllegalConversionException(__FILE__, __LINE__, "string source exhausted");
+        case sourceIllegal:
+            throw IceUtil::IllegalConversionException(__FILE__, __LINE__, "string source illegal");
+        default:
+        {
+            assert(0);
+            throw IceUtil::IllegalConversionException(__FILE__, __LINE__);
+        }
+    }
+
+    for(vector<unsigned short>::const_iterator c = u16buffer.begin(); c != u16buffer.end(); ++c)
+    {
+        out << u16CodePoint(*c);
+    }
+}
+
+string
 sliceModeToIceMode(Operation::Mode opMode)
 {
     string mode;
@@ -1544,39 +1586,118 @@ Slice::CsVisitor::writeConstantValue(const TypePtr& type, const SyntaxTreeBasePt
             // here because they are sensitive to the current locale.
             //
             static const string basicSourceChars = "abcdefghijklmnopqrstuvwxyz"
-                                                   "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
-                                                   "0123456789"
-                                                   "_{}[]#()<>%:;.?*+-/^&|~!=,\\\"' ";
+                                                    "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
+                                                    "0123456789"
+                                                    "_{}[]#()<>%:;.?*+-/^&|~!=,\\\"' ";
+
             static const set<char> charSet(basicSourceChars.begin(), basicSourceChars.end());
 
             _out << "\"";                                    // Opening "
 
-            for(string::const_iterator c = value.begin(); c != value.end(); ++c)
+            vector<unsigned char> u8buffer;                  // Buffer to convert multibyte characters
+
+            for(size_t i = 0; i < value.size();)
             {
-                if(charSet.find(*c) == charSet.end())
+                if(charSet.find(value[i]) == charSet.end())
                 {
-                    unsigned char uc = *c;                   // char may be signed, so make it positive
-                    ostringstream s;
-                    s << "\\u";                      // Print as unicode if not in basic source character set
-                    s << hex;
-                    s.width(4);
-                    s.fill('0');
-                    s << static_cast<unsigned>(uc);
-                    _out << s.str();
+                    if(static_cast<unsigned char>(value[i]) < 128) // Single byte character
+                    {
+                        //
+                        // Print as unicode if not in basic source character set
+                        //
+                        _out << u16CodePoint(static_cast<unsigned int>(value[i]));
+                    }
+                    else
+                    {
+                        u8buffer.push_back(value[i]);
+                    }
                 }
                 else
                 {
-                   switch(*c)
+                    //
+                    // Write any pedding characters in the utf8 buffer
+                    //
+                    if(!u8buffer.empty())
+                    {
+                        writeU8Buffer(u8buffer, _out);
+                        u8buffer.clear();
+                    }
+                    switch(value[i])
                     {
                         case '\\':
+                        {
+                            string s = "\\";
+                            size_t j = i + 1;
+                            for(; j < value.size(); ++j)
+                            {
+                                if(value[j] != '\\')
+                                {
+                                    break;
+                                }
+                                s += "\\";
+                            }
+
+                            //
+                            // An even number of slash \ will escape the backslash and
+                            // the codepoint will be interpreted as its charaters
+                            //
+                            // \\U00000041  - ['\\', 'U', '0', '0', '0', '0', '0', '0', '4', '1']
+                            // \\\U00000041 - ['\\', 'A'] (41 is the codepoint for 'A')
+                            //
+                            if(s.size() % 2 != 0 && value[j] == 'U')
+                            {
+                                _out << s.substr(0, s.size() - 1);
+                                i = j + 1;
+
+                                string codepoint = value.substr(j + 1, 8);
+                                assert(codepoint.size() ==  8);
+
+                                IceUtil::Int64 v = IceUtilInternal::strToInt64(codepoint.c_str(), 0, 16);
+
+
+                                //
+                                // Unicode character in the range U+10000 to U+10FFFF is not permitted in a character literal
+                                // and is represented using a Unicode surrogate pair.
+                                //
+                                if(v > 0xFFFF)
+                                {
+                                    unsigned int high = ((static_cast<unsigned int>(v) - 0x10000) / 0x400) + 0xD800;
+                                    unsigned int low = ((static_cast<unsigned int>(v) - 0x10000) % 0x400) + 0xDC00;
+                                    _out << u16CodePoint(high);
+                                    _out << u16CodePoint(low);
+                                }
+                                else
+                                {
+                                    _out << "\\U" << codepoint;
+                                }
+
+                                i = j + 1 + 8;
+                            }
+                            else
+                            {
+                                _out << s;
+                                i = j;
+                            }
+                            continue;
+                        }
                         case '"':
                         {
                             _out << "\\";
                             break;
                         }
                     }
-                    _out << *c;                              // Print normally if in basic source character set
+                    _out << value[i];                        // Print normally if in basic source character set
                 }
+                i++;
+            }
+
+            //
+            // Write any pedding characters in the utf8 buffer
+            //
+            if(!u8buffer.empty())
+            {
+                writeU8Buffer(u8buffer, _out);
+                u8buffer.clear();
             }
 
             _out << "\"";                                    // Closing "
diff --git a/cpp/src/slice2cs/Makefile b/cpp/src/slice2cs/Makefile
index e51e24c0445..e46c1005dd6 100644
--- a/cpp/src/slice2cs/Makefile
+++ b/cpp/src/slice2cs/Makefile
@@ -20,7 +20,7 @@ RPATH_DIR	= $(LOADER_PATH)/../$(libsubdir)
 
 include $(top_srcdir)/config/Make.rules
 
-CPPFLAGS	:= -I. $(CPPFLAGS)
+CPPFLAGS	:= -I. -I.. $(CPPFLAGS)
 
 $(NAME): $(OBJS)
 	rm -f $@
diff --git a/cpp/src/slice2java/Gen.cpp b/cpp/src/slice2java/Gen.cpp
index 4c5cf6f2477..038568cf64c 100644
--- a/cpp/src/slice2java/Gen.cpp
+++ b/cpp/src/slice2java/Gen.cpp
@@ -14,6 +14,7 @@
 #include <IceUtil/Iterator.h>
 #include <IceUtil/StringUtil.h>
 #include <IceUtil/InputUtil.h>
+#include <IceUtil/Unicode.h>
 #include <cstring>
 
 #include <limits>
@@ -23,6 +24,44 @@ using namespace Slice;
 using namespace IceUtil;
 using namespace IceUtilInternal;
 
+string
+u16CodePoint(unsigned short value)
+{
+    ostringstream s;
+    s << "\\u";
+    s << hex;
+    s.width(4);
+    s.fill('0');
+    s << value;
+    return s.str();
+}
+
+void
+writeU8Buffer(const vector<unsigned char>& u8buffer, ::IceUtilInternal::Output& out)
+{
+    vector<unsigned short> u16buffer;
+    IceUtilInternal::ConversionResult result = convertUTF8ToUTF16(u8buffer, u16buffer, IceUtil::lenientConversion);
+    switch(result)
+    {
+        case conversionOK:
+            break;
+        case sourceExhausted:
+            throw IceUtil::IllegalConversionException(__FILE__, __LINE__, "string source exhausted");
+        case sourceIllegal:
+            throw IceUtil::IllegalConversionException(__FILE__, __LINE__, "string source illegal");
+        default:
+        {
+            assert(0);
+            throw IceUtil::IllegalConversionException(__FILE__, __LINE__);
+        }
+    }
+
+    for(vector<unsigned short>::const_iterator c = u16buffer.begin(); c != u16buffer.end(); ++c)
+    {
+        out << u16CodePoint(*c);
+    }
+}
+
 static string
 sliceModeToIceMode(Operation::Mode opMode)
 {
@@ -1729,53 +1768,152 @@ Slice::JavaVisitor::writeConstantValue(Output& out, const TypePtr& type, const S
                     static const set<char> charSet(basicSourceChars.begin(), basicSourceChars.end());
                     out << "\"";
 
-                    for(string::const_iterator c = value.begin(); c != value.end(); ++c)
+                    vector<unsigned char> u8buffer;                  // Buffer to convert multibyte characters
+
+                    for(size_t i = 0; i < value.size();)
                     {
-                        if(charSet.find(*c) == charSet.end())
+                        if(charSet.find(value[i]) == charSet.end())
                         {
-                            switch(*c)
+                            char c = value[i];
+                            if(static_cast<unsigned char>(c) < 128) // Single byte character
                             {
                                 //
-                                // Java doesn't want '\n' or '\r\n' encoded as universal
-                                // characters, that gives an error "unclosed string literal"
+                                // Print as unicode if not in basic source character set
                                 //
-                                case '\r':
+                                switch(c)
                                 {
-                                    out << "\\r";
-                                    break;
-                                }
-                                case '\n':
-                                {
-                                    out << "\\n";
-                                    break;
-                                }
-                                default:
-                                {
-                                    unsigned char uc = *c;
-                                    ostringstream s;
-                                    s << "\\u";
-                                    s.flags(ios_base::hex);
-                                    s.width(4);
-                                    s.fill('0');
-                                    s << static_cast<unsigned>(uc);
-                                    out << s.str();
-                                    break;
+                                    //
+                                    // Java doesn't want '\n' or '\r\n' encoded as universal
+                                    // characters, that gives an error "unclosed string literal"
+                                    //
+                                    case '\r':
+                                    {
+                                        out << "\\r";
+                                        break;
+                                    }
+                                    case '\n':
+                                    {
+                                        out << "\\n";
+                                        break;
+                                    }
+                                    default:
+                                    {
+                                        out << u16CodePoint(c);
+                                        break;
+                                    }
                                 }
                             }
+                            else
+                            {
+                                u8buffer.push_back(value[i]);
+                            }
                         }
                         else
                         {
-                            switch(*c)
+                            //
+                            // Write any pedding characters in the utf8 buffer
+                            //
+                            if(!u8buffer.empty())
+                            {
+                                writeU8Buffer(u8buffer, out);
+                                u8buffer.clear();
+                            }
+                            switch(value[i])
                             {
                                 case '\\':
+                                {
+                                    string s = "\\";
+                                    size_t j = i + 1;
+                                    for(; j < value.size(); ++j)
+                                    {
+                                        if(value[j] != '\\')
+                                        {
+                                            break;
+                                        }
+                                        s += "\\";
+                                    }
+
+                                    //
+                                    // An even number of slash \ will escape the backslash and
+                                    // the codepoint will be interpreted as its charaters
+                                    //
+                                    // \\U00000041  - ['\\', 'U', '0', '0', '0', '0', '0', '0', '4', '1']
+                                    // \\\U00000041 - ['\\', 'A'] (41 is the codepoint for 'A')
+                                    //
+                                    if(s.size() % 2 != 0 && (value[j] == 'U' || value[j] == 'u'))
+                                    {
+                                        size_t sz = value[j] == 'U' ? 8 : 4;
+                                        out << s.substr(0, s.size() - 1);
+                                        i = j + 1;
+
+                                        string codepoint = value.substr(j + 1, sz);
+                                        assert(codepoint.size() ==  sz);
+
+                                        IceUtil::Int64 v = IceUtilInternal::strToInt64(codepoint.c_str(), 0, 16);
+
+
+                                        //
+                                        // Java doesn't like this special characters encoded as universal characters
+                                        //
+                                        if(v == 0x5c)
+                                        {
+                                            out << "\\\\";
+                                        }
+                                        else if(v == 0xa)
+                                        {
+                                            out << "\\n";
+                                        }
+                                        else if(v == 0xd)
+                                        {
+                                            out << "\\r";
+                                        }
+                                        else if(v == 0x22)
+                                        {
+                                            out << "\\\"";
+                                        }
+                                        //
+                                        // Unicode character in the range U+10000 to U+10FFFF is not permitted in a character literal
+                                        // and is represented using a Unicode surrogate pair.
+                                        //
+                                        else if(v > 0xFFFF)
+                                        {
+                                            unsigned int high = ((static_cast<unsigned int>(v) - 0x10000) / 0x400) + 0xD800;
+                                            unsigned int low = ((static_cast<unsigned int>(v) - 0x10000) % 0x400) + 0xDC00;
+                                            out << u16CodePoint(high);
+                                            out << u16CodePoint(low);
+                                        }
+                                        else
+                                        {
+                                            out << u16CodePoint(static_cast<unsigned int>(v));
+                                        }
+
+                                        i = j + 1 + sz;
+                                    }
+                                    else
+                                    {
+                                        out << s;
+                                        i = j;
+                                    }
+                                    continue;
+                                }
                                 case '"':
                                 {
                                     out << "\\";
                                     break;
                                 }
                             }
-                            out << *c;
+                            out << value[i];                        // Print normally if in basic source character set
                         }
+                        i++;
+                    }
+
+                    //
+                    // Write any pedding characters in the utf8 buffer
+                    //
+                    if(!u8buffer.empty())
+                    {
+                        writeU8Buffer(u8buffer, out);
+                        u8buffer.clear();
                     }
 
                     out << "\"";
diff --git a/cpp/src/slice2java/Makefile b/cpp/src/slice2java/Makefile
index 010554d8e12..e2a882bfb6f 100644
--- a/cpp/src/slice2java/Makefile
+++ b/cpp/src/slice2java/Makefile
@@ -20,7 +20,7 @@ RPATH_DIR	= $(LOADER_PATH)/../$(libsubdir)
 
 include $(top_srcdir)/config/Make.rules
 
-CPPFLAGS	:= -I. $(CPPFLAGS)
+CPPFLAGS	:= -I. -I.. $(CPPFLAGS)
 
 $(NAME): $(OBJS)
 	rm -f $@
diff --git a/cpp/src/slice2js/Gen.cpp b/cpp/src/slice2js/Gen.cpp
index 11bd0f608e5..573d0180c67 100644
--- a/cpp/src/slice2js/Gen.cpp
+++ b/cpp/src/slice2js/Gen.cpp
@@ -20,6 +20,7 @@
 #include <direct.h>
 #endif
 #include <IceUtil/Iterator.h>
+#include <IceUtil/Unicode.h>
 #include <IceUtil/UUID.h>
 #include <Slice/Checksum.h>
 #include <Slice/FileTracker.h>
@@ -35,6 +36,44 @@ namespace
 {
 
 string
+u16CodePoint(unsigned short value)
+{
+    ostringstream s;
+    s << "\\u";
+    s << hex;
+    s.width(4);
+    s.fill('0');
+    s << value;
+    return s.str();
+}
+
+void
+writeU8Buffer(const vector<unsigned char>& u8buffer, ::IceUtilInternal::Output& out)
+{
+    vector<unsigned short> u16buffer;
+    IceUtilInternal::ConversionResult result = convertUTF8ToUTF16(u8buffer, u16buffer, IceUtil::lenientConversion);
+    switch(result)
+    {
+        case conversionOK:
+            break;
+        case sourceExhausted:
+            throw IceUtil::IllegalConversionException(__FILE__, __LINE__, "string source exhausted");
+        case sourceIllegal:
+            throw IceUtil::IllegalConversionException(__FILE__, __LINE__, "string source illegal");
+        default:
+        {
+            assert(0);
+            throw IceUtil::IllegalConversionException(__FILE__, __LINE__);
+        }
+    }
+
+    for(vector<unsigned short>::const_iterator c = u16buffer.begin(); c != u16buffer.end(); ++c)
+    {
+        out << u16CodePoint(*c);
+    }
+}
+
+string
 sliceModeToIceMode(Operation::Mode opMode)
 {
     switch(opMode)
@@ -479,32 +518,110 @@ Slice::JsVisitor::writeConstantValue(const string& scope, const TypePtr& type, c
 
             _out << "\"";                                    // Opening "
 
-            for(string::const_iterator c = value.begin(); c != value.end(); ++c)
+            vector<unsigned char> u8buffer;                  // Buffer to convert multibyte characters
+
+            for(size_t i = 0; i < value.size();)
             {
-                if(charSet.find(*c) == charSet.end())
+                if(charSet.find(value[i]) == charSet.end())
                 {
-                    unsigned char uc = *c;                   // char may be signed, so make it positive
-                    ostringstream s;
-                    s << "\\u";                      // Print as unicode if not in basic source character set
-                    s << hex;
-                    s.width(4);
-                    s.fill('0');
-                    s << static_cast<unsigned>(uc);
-                    _out << s.str();
+                    if(static_cast<unsigned char>(value[i]) < 128) // Single byte character
+                    {
+                        //
+                        // Print as unicode if not in basic source character set
+                        //
+                        _out << u16CodePoint(static_cast<unsigned int>(value[i]));
+                    }
+                    else
+                    {
+                        u8buffer.push_back(value[i]);
+                    }
                 }
                 else
                 {
-                   switch(*c)
+                    //
+                    // Write any pedding characters in the utf8 buffer
+                    //
+                    if(!u8buffer.empty())
+                    {
+                        writeU8Buffer(u8buffer, _out);
+                        u8buffer.clear();
+                    }
+                    switch(value[i])
                     {
                         case '\\':
+                        {
+                            string s = "\\";
+                            size_t j = i + 1;
+                            for(; j < value.size(); ++j)
+                            {
+                                if(value[j] != '\\')
+                                {
+                                    break;
+                                }
+                                s += "\\";
+                            }
+
+                            //
+                            // An even number of slash \ will escape the backslash and
+                            // the codepoint will be interpreted as its charaters
+                            //
+                            // \\U00000041  - ['\\', 'U', '0', '0', '0', '0', '0', '0', '4', '1']
+                            // \\\U00000041 - ['\\', 'A'] (41 is the codepoint for 'A')
+                            //
+                            if(s.size() % 2 != 0 && value[j] == 'U')
+                            {
+                                _out << s.substr(0, s.size() - 1);
+                                i = j + 1;
+
+                                string codepoint = value.substr(j + 1, 8);
+                                assert(codepoint.size() ==  8);
+
+                                IceUtil::Int64 v = IceUtilInternal::strToInt64(codepoint.c_str(), 0, 16);
+
+
+                                //
+                                // Unicode character in the range U+10000 to U+10FFFF is not permitted in a character literal
+                                // and is represented using a Unicode surrogate pair.
+                                //
+                                if(v > 0xFFFF)
+                                {
+                                    unsigned int high = ((static_cast<unsigned int>(v) - 0x10000) / 0x400) + 0xD800;
+                                    unsigned int low = ((static_cast<unsigned int>(v) - 0x10000) % 0x400) + 0xDC00;
+                                    _out << u16CodePoint(high);
+                                    _out << u16CodePoint(low);
+                                }
+                                else
+                                {
+                                    _out << u16CodePoint(static_cast<unsigned int>(v));
+                                }
+
+                                i = j + 1 + 8;
+                            }
+                            else
+                            {
+                                _out << s;
+                                i = j;
+                            }
+                            continue;
+                        }
                         case '"':
                         {
                             _out << "\\";
                             break;
                         }
                     }
-                    _out << *c;                              // Print normally if in basic source character set
+                    _out << value[i];                        // Print normally if in basic source character set
                 }
+                i++;
+            }
+
+            //
+            // Write any pedding characters in the utf8 buffer
+            //
+            if(!u8buffer.empty())
+            {
+                writeU8Buffer(u8buffer, _out);
+                u8buffer.clear();
             }
 
             _out << "\"";                                    // Closing "
diff --git a/cpp/src/slice2js/Makefile b/cpp/src/slice2js/Makefile
index bd1bbe967f8..0aaf14a7a0d 100644
--- a/cpp/src/slice2js/Makefile
+++ b/cpp/src/slice2js/Makefile
@@ -21,7 +21,7 @@ RPATH_DIR	= $(LOADER_PATH)/../$(libsubdir)
 
 include $(top_srcdir)/config/Make.rules
 
-CPPFLAGS	:= -I. $(CPPFLAGS)
+CPPFLAGS	:= -I. -I.. $(CPPFLAGS)
 
 $(NAME): $(OBJS)
 	rm -f $@
diff --git a/cpp/src/slice2php/Main.cpp b/cpp/src/slice2php/Main.cpp
index 2838863454a..35117e68f36 100644
--- a/cpp/src/slice2php/Main.cpp
+++ b/cpp/src/slice2php/Main.cpp
@@ -16,6 +16,7 @@
 #include <IceUtil/StringUtil.h>
 #include <IceUtil/Mutex.h>
 #include <IceUtil/MutexPtrLock.h>
+#include <IceUtil/Unicode.h>
 #include <Slice/Checksum.h>
 #include <Slice/Preprocessor.h>
 #include <Slice/FileTracker.h>
@@ -1273,9 +1274,10 @@ CodeVisitor::writeConstantValue(const TypePtr& type, const SyntaxTreeBasePtr& va
 
                 _out << "\"";                                       // Opening "
 
-                for(string::const_iterator c = value.begin(); c != value.end(); ++c)
+                for(size_t i = 0; i < value.size();)
                 {
-                    switch(*c)
+                    char c = value[i];
+                    switch(c)
                     {
                     case '$':
                     {
@@ -1289,8 +1291,79 @@ CodeVisitor::writeConstantValue(const TypePtr& type, const SyntaxTreeBasePtr& va
                     }
                     case '\\':
                     {
-                        _out << "\\\\";
-                        break;
+
+                        string s = "\\";
+                        size_t j = i + 1;
+                        for(; j < value.size(); ++j)
+                        {
+                            if(value[j] != '\\')
+                            {
+                                break;
+                            }
+                            s += "\\";
+                        }
+
+                        //
+                        // An even number of slash \ will escape the backslash and
+                        // the codepoint will be interpreted as its charaters
+                        //
+                        // \\u00000041  - ['\\', 'u', '0', '0', '0', '0', '0', '0', '4', '1']
+                        // \\\u00000041 - ['\\', 'A'] (41 is the codepoint for 'A')
+                        //
+                        if(s.size() % 2 != 0 && (value[j] == 'U' || value[j] == 'u'))
+                        {
+                            //
+                            // Convert codepoint to UTF8 bytes and write the escaped bytes
+                            //
+                            _out << s.substr(0, s.size() - 1);
+
+                            size_t sz = value[j] == 'U' ? 8 : 4;
+                            string codepoint = value.substr(j + 1, sz);
+                            assert(codepoint.size() ==  sz);
+
+                            IceUtil::Int64 v = IceUtilInternal::strToInt64(codepoint.c_str(), 0, 16);
+
+
+                            vector<unsigned int> u32buffer;
+                            u32buffer.push_back(static_cast<unsigned int>(v));
+
+                            vector<unsigned char> u8buffer;
+
+                            IceUtilInternal::ConversionResult result = convertUTF32ToUTF8(u32buffer, u8buffer, IceUtil::lenientConversion);
+                            switch(result)
+                            {
+                                case conversionOK:
+                                    break;
+                                case sourceExhausted:
+                                    throw IceUtil::IllegalConversionException(__FILE__, __LINE__, "string source exhausted");
+                                case sourceIllegal:
+                                    throw IceUtil::IllegalConversionException(__FILE__, __LINE__, "string source illegal");
+                                default:
+                                {
+                                    assert(0);
+                                    throw IceUtil::IllegalConversionException(__FILE__, __LINE__);
+                                }
+                            }
+
+                            ostringstream s;
+                            for(vector<unsigned char>::const_iterator q = u8buffer.begin(); q != u8buffer.end(); ++q)
+                            {
+                                s << "\\";
+                                s.fill('0');
+                                s.width(3);
+                                s << oct;
+                                s << static_cast<unsigned int>(*q);
+                            }
+                            _out << s.str();
+
+                            i = j + 1 + sz;
+                        }
+                        else
+                        {
+                            _out << s;
+                            i = j;
+                        }
+                        continue;
                     }
                     case '\r':
                     {
@@ -1307,11 +1380,6 @@ CodeVisitor::writeConstantValue(const TypePtr& type, const SyntaxTreeBasePtr& va
                         _out << "\\t";
                         break;
                     }
-                    case '\b':
-                    {
-                        _out << "\\b";
-                        break;
-                    }
                     case '\f':
                     {
                         _out << "\\f";
@@ -1319,9 +1387,9 @@ CodeVisitor::writeConstantValue(const TypePtr& type, const SyntaxTreeBasePtr& va
                     }
                     default:
                     {
-                        if(charSet.find(*c) == charSet.end())
+                        if(charSet.find(c) == charSet.end())
                         {
-                            unsigned char uc = *c;              // Char may be signed, so make it positive.
+                            unsigned char uc = c;              // Char may be signed, so make it positive.
                             stringstream s;
                             s << "\\";                          // Print as octal if not in basic source character set.
                             s.flags(ios_base::oct);
@@ -1332,11 +1400,12 @@ CodeVisitor::writeConstantValue(const TypePtr& type, const SyntaxTreeBasePtr& va
                         }
                         else
                         {
-                            _out << *c;                         // Print normally if in basic source character set.
+                            _out << c;                          // Print normally if in basic source character set.
                         }
                         break;
                     }
                     }
+                    ++i;
                 }
 
                 _out << "\"";                                   // Closing "
diff --git a/cpp/src/slice2php/Makefile b/cpp/src/slice2php/Makefile
index 8bc91a8cd66..62f82531201 100644
--- a/cpp/src/slice2php/Makefile
+++ b/cpp/src/slice2php/Makefile
@@ -19,7 +19,7 @@ RPATH_DIR	= $(LOADER_PATH)/../$(libsubdir)
 
 include $(top_srcdir)/config/Make.rules
 
-CPPFLAGS	:= -I. $(CPPFLAGS)
+CPPFLAGS	:= -I. -I.. $(CPPFLAGS)
 
 $(NAME): $(OBJS)
 	rm -f $@