Version 3.9.18

Ensure consistency of Math.sqrt on Intel platforms.

Remove static initializers in v8. (issue 1859)

Add explicit dependency on v8_base in the GYP-based build.

Performance and stability improvements on all platforms.

git-svn-id: http://v8.googlecode.com/svn/trunk@11017 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
diff --git a/test/cctest/test-api.cc b/test/cctest/test-api.cc
index 5137c65..79818ac 100644
--- a/test/cctest/test-api.cc
+++ b/test/cctest/test-api.cc
@@ -5526,6 +5526,17 @@
 }
 
 
+int GetUtf8Length(Handle<String> str) {
+  int len = str->Utf8Length();
+  if (len < 0) {
+    i::Handle<i::String> istr(v8::Utils::OpenHandle(*str));
+    i::FlattenString(istr);
+    len = str->Utf8Length();
+  }
+  return len;
+}
+
+
 THREADED_TEST(StringWrite) {
   LocalContext context;
   v8::HandleScope scope;
@@ -5606,7 +5617,7 @@
   CHECK_EQ(0, strncmp(utf8buf, "ab\1", 3));
 
   memset(utf8buf, 0x1, sizeof(utf8buf));
-  len = left_tree->Utf8Length();
+  len = GetUtf8Length(left_tree);
   int utf8_expected =
       (0x80 + (0x800 - 0x80) * 2 + (0xd800 - 0x800) * 3) / kStride;
   CHECK_EQ(utf8_expected, len);
@@ -5620,7 +5631,7 @@
   CHECK_EQ(1, utf8buf[utf8_expected]);
 
   memset(utf8buf, 0x1, sizeof(utf8buf));
-  len = right_tree->Utf8Length();
+  len = GetUtf8Length(right_tree);
   CHECK_EQ(utf8_expected, len);
   len = right_tree->WriteUtf8(utf8buf, utf8_expected, &charlen);
   CHECK_EQ(utf8_expected, len);
@@ -5745,6 +5756,217 @@
 }
 
 
+static void Utf16Helper(
+    LocalContext& context,
+    const char* name,
+    const char* lengths_name,
+    int len) {
+  Local<v8::Array> a =
+      Local<v8::Array>::Cast(context->Global()->Get(v8_str(name)));
+  Local<v8::Array> alens =
+      Local<v8::Array>::Cast(context->Global()->Get(v8_str(lengths_name)));
+  for (int i = 0; i < len; i++) {
+    Local<v8::String> string =
+      Local<v8::String>::Cast(a->Get(i));
+    Local<v8::Number> expected_len =
+      Local<v8::Number>::Cast(alens->Get(i));
+    int length = GetUtf8Length(string);
+    CHECK_EQ(static_cast<int>(expected_len->Value()), length);
+  }
+}
+
+
+static uint16_t StringGet(Handle<String> str, int index) {
+  i::Handle<i::String> istring =
+      v8::Utils::OpenHandle(String::Cast(*str));
+  return istring->Get(index);
+}
+
+
+static void WriteUtf8Helper(
+    LocalContext& context,
+    const char* name,
+    const char* lengths_name,
+    int len) {
+  Local<v8::Array> b =
+      Local<v8::Array>::Cast(context->Global()->Get(v8_str(name)));
+  Local<v8::Array> alens =
+      Local<v8::Array>::Cast(context->Global()->Get(v8_str(lengths_name)));
+  char buffer[1000];
+  char buffer2[1000];
+  for (int i = 0; i < len; i++) {
+    Local<v8::String> string =
+      Local<v8::String>::Cast(b->Get(i));
+    Local<v8::Number> expected_len =
+      Local<v8::Number>::Cast(alens->Get(i));
+    int utf8_length = static_cast<int>(expected_len->Value());
+    for (int j = utf8_length + 1; j >= 0; j--) {
+      memset(reinterpret_cast<void*>(&buffer), 42, sizeof(buffer));
+      memset(reinterpret_cast<void*>(&buffer2), 42, sizeof(buffer2));
+      int nchars;
+      int utf8_written =
+          string->WriteUtf8(buffer, j, &nchars, String::NO_OPTIONS);
+      int utf8_written2 =
+          string->WriteUtf8(buffer2, j, &nchars, String::NO_NULL_TERMINATION);
+      CHECK_GE(utf8_length + 1, utf8_written);
+      CHECK_GE(utf8_length, utf8_written2);
+      for (int k = 0; k < utf8_written2; k++) {
+        CHECK_EQ(buffer[k], buffer2[k]);
+      }
+      CHECK(nchars * 3 >= utf8_written - 1);
+      CHECK(nchars <= utf8_written);
+      if (j == utf8_length + 1) {
+        CHECK_EQ(utf8_written2, utf8_length);
+        CHECK_EQ(utf8_written2 + 1, utf8_written);
+      }
+      CHECK_EQ(buffer[utf8_written], 42);
+      if (j > utf8_length) {
+        if (utf8_written != 0) CHECK_EQ(buffer[utf8_written - 1], 0);
+        if (utf8_written > 1) CHECK_NE(buffer[utf8_written - 2], 42);
+        Handle<String> roundtrip = v8_str(buffer);
+        CHECK(roundtrip->Equals(string));
+      } else {
+        if (utf8_written != 0) CHECK_NE(buffer[utf8_written - 1], 42);
+      }
+      if (utf8_written2 != 0) CHECK_NE(buffer[utf8_written - 1], 42);
+      if (nchars >= 2) {
+        uint16_t trail = StringGet(string, nchars - 1);
+        uint16_t lead = StringGet(string, nchars - 2);
+        if (((lead & 0xfc00) == 0xd800) &&
+            ((trail & 0xfc00) == 0xdc00)) {
+          unsigned char u1 = buffer2[utf8_written2 - 4];
+          unsigned char u2 = buffer2[utf8_written2 - 3];
+          unsigned char u3 = buffer2[utf8_written2 - 2];
+          unsigned char u4 = buffer2[utf8_written2 - 1];
+          CHECK_EQ((u1 & 0xf8), 0xf0);
+          CHECK_EQ((u2 & 0xc0), 0x80);
+          CHECK_EQ((u3 & 0xc0), 0x80);
+          CHECK_EQ((u4 & 0xc0), 0x80);
+          uint32_t c = 0x10000 + ((lead & 0x3ff) << 10) + (trail & 0x3ff);
+          CHECK_EQ((u4 & 0x3f), (c & 0x3f));
+          CHECK_EQ((u3 & 0x3f), ((c >> 6) & 0x3f));
+          CHECK_EQ((u2 & 0x3f), ((c >> 12) & 0x3f));
+          CHECK_EQ((u1 & 0x3), c >> 18);
+        }
+      }
+    }
+  }
+}
+
+
+THREADED_TEST(Utf16) {
+  LocalContext context;
+  v8::HandleScope scope;
+  CompileRun(
+      "var pad = '01234567890123456789';"
+      "var p = [];"
+      "var plens = [20, 3, 3];"
+      "p.push('01234567890123456789');"
+      "var lead = 0xd800;"
+      "var trail = 0xdc00;"
+      "p.push(String.fromCharCode(0xd800));"
+      "p.push(String.fromCharCode(0xdc00));"
+      "var a = [];"
+      "var b = [];"
+      "var alens = [];"
+      "for (var i = 0; i < 3; i++) {"
+      "  p[1] = String.fromCharCode(lead++);"
+      "  for (var j = 0; j < 3; j++) {"
+      "    p[2] = String.fromCharCode(trail++);"
+      "    a.push(p[i] + p[j]);"
+      "    b.push(p[i] + p[j]);"
+      "    alens.push(plens[i] + plens[j]);"
+      "  }"
+      "}"
+      "alens[5] -= 2;"  // Here the surrogate pairs match up.
+      "var a2 = [];"
+      "var b2 = [];"
+      "var a2lens = [];"
+      "for (var m = 0; m < 9; m++) {"
+      "  for (var n = 0; n < 9; n++) {"
+      "    a2.push(a[m] + a[n]);"
+      "    b2.push(b[m] + b[n]);"
+      "    var utf = alens[m] + alens[n];"  // And here.
+           // The 'n's that start with 0xdc.. are 6-8
+           // The 'm's that end with 0xd8.. are 1, 4 and 7
+      "    if ((m % 3) == 1 && n >= 6) utf -= 2;"
+      "    a2lens.push(utf);"
+      "  }"
+      "}");
+  Utf16Helper(context, "a", "alens", 9);
+  Utf16Helper(context, "a2", "a2lens", 81);
+  WriteUtf8Helper(context, "b", "alens", 9);
+  WriteUtf8Helper(context, "b2", "a2lens", 81);
+}
+
+
+static bool SameSymbol(Handle<String> s1, Handle<String> s2) {
+  i::Handle<i::String> is1(v8::Utils::OpenHandle(*s1));
+  i::Handle<i::String> is2(v8::Utils::OpenHandle(*s2));
+  return *is1 == *is2;
+}
+
+
+static void SameSymbolHelper(const char* a, const char* b) {
+  Handle<String> symbol1 = v8::String::NewSymbol(a);
+  Handle<String> symbol2 = v8::String::NewSymbol(b);
+  CHECK(SameSymbol(symbol1, symbol2));
+}
+
+
+THREADED_TEST(Utf16Symbol) {
+  LocalContext context;
+  v8::HandleScope scope;
+
+  Handle<String> symbol1 = v8::String::NewSymbol("abc");
+  Handle<String> symbol2 = v8::String::NewSymbol("abc");
+  CHECK(SameSymbol(symbol1, symbol2));
+
+  SameSymbolHelper("\360\220\220\205",  // 4 byte encoding.
+                   "\355\240\201\355\260\205");  // 2 3-byte surrogates.
+  SameSymbolHelper("\355\240\201\355\260\206",  // 2 3-byte surrogates.
+                   "\360\220\220\206");  // 4 byte encoding.
+  SameSymbolHelper("x\360\220\220\205",  // 4 byte encoding.
+                   "x\355\240\201\355\260\205");  // 2 3-byte surrogates.
+  SameSymbolHelper("x\355\240\201\355\260\206",  // 2 3-byte surrogates.
+                   "x\360\220\220\206");  // 4 byte encoding.
+  CompileRun(
+      "var sym0 = 'benedictus';"
+      "var sym0b = 'S\303\270ren';"
+      "var sym1 = '\355\240\201\355\260\207';"
+      "var sym2 = '\360\220\220\210';"
+      "var sym3 = 'x\355\240\201\355\260\207';"
+      "var sym4 = 'x\360\220\220\210';"
+      "if (sym1.length != 2) throw sym1;"
+      "if (sym1.charCodeAt(1) != 0xdc07) throw sym1.charCodeAt(1);"
+      "if (sym2.length != 2) throw sym2;"
+      "if (sym2.charCodeAt(1) != 0xdc08) throw sym2.charCodeAt(2);"
+      "if (sym3.length != 3) throw sym3;"
+      "if (sym3.charCodeAt(2) != 0xdc07) throw sym1.charCodeAt(2);"
+      "if (sym4.length != 3) throw sym4;"
+      "if (sym4.charCodeAt(2) != 0xdc08) throw sym2.charCodeAt(2);");
+  Handle<String> sym0 = v8::String::NewSymbol("benedictus");
+  Handle<String> sym0b = v8::String::NewSymbol("S\303\270ren");
+  Handle<String> sym1 = v8::String::NewSymbol("\355\240\201\355\260\207");
+  Handle<String> sym2 = v8::String::NewSymbol("\360\220\220\210");
+  Handle<String> sym3 = v8::String::NewSymbol("x\355\240\201\355\260\207");
+  Handle<String> sym4 = v8::String::NewSymbol("x\360\220\220\210");
+  v8::Local<v8::Object> global = context->Global();
+  Local<Value> s0 = global->Get(v8_str("sym0"));
+  Local<Value> s0b = global->Get(v8_str("sym0b"));
+  Local<Value> s1 = global->Get(v8_str("sym1"));
+  Local<Value> s2 = global->Get(v8_str("sym2"));
+  Local<Value> s3 = global->Get(v8_str("sym3"));
+  Local<Value> s4 = global->Get(v8_str("sym4"));
+  CHECK(SameSymbol(sym0, Handle<String>(String::Cast(*s0))));
+  CHECK(SameSymbol(sym0b, Handle<String>(String::Cast(*s0b))));
+  CHECK(SameSymbol(sym1, Handle<String>(String::Cast(*s1))));
+  CHECK(SameSymbol(sym2, Handle<String>(String::Cast(*s2))));
+  CHECK(SameSymbol(sym3, Handle<String>(String::Cast(*s3))));
+  CHECK(SameSymbol(sym4, Handle<String>(String::Cast(*s4))));
+}
+
+
 THREADED_TEST(ToArrayIndex) {
   v8::HandleScope scope;
   LocalContext context;
diff --git a/test/cctest/test-parsing.cc b/test/cctest/test-parsing.cc
index cd8a6af..6bcae7c 100755
--- a/test/cctest/test-parsing.cc
+++ b/test/cctest/test-parsing.cc
@@ -63,7 +63,7 @@
     int length = i::StrLength(key_token.keyword);
     CHECK(static_cast<int>(sizeof(buffer)) >= length);
     {
-      i::Utf8ToUC16CharacterStream stream(keyword, length);
+      i::Utf8ToUtf16CharacterStream stream(keyword, length);
       i::Scanner scanner(&unicode_cache);
       // The scanner should parse Harmony keywords for this test.
       scanner.SetHarmonyScoping(true);
@@ -74,7 +74,7 @@
     }
     // Removing characters will make keyword matching fail.
     {
-      i::Utf8ToUC16CharacterStream stream(keyword, length - 1);
+      i::Utf8ToUtf16CharacterStream stream(keyword, length - 1);
       i::Scanner scanner(&unicode_cache);
       scanner.Initialize(&stream);
       CHECK_EQ(i::Token::IDENTIFIER, scanner.Next());
@@ -85,7 +85,7 @@
     for (int j = 0; j < static_cast<int>(ARRAY_SIZE(chars_to_append)); ++j) {
       memmove(buffer, keyword, length);
       buffer[length] = chars_to_append[j];
-      i::Utf8ToUC16CharacterStream stream(buffer, length + 1);
+      i::Utf8ToUtf16CharacterStream stream(buffer, length + 1);
       i::Scanner scanner(&unicode_cache);
       scanner.Initialize(&stream);
       CHECK_EQ(i::Token::IDENTIFIER, scanner.Next());
@@ -95,7 +95,7 @@
     {
       memmove(buffer, keyword, length);
       buffer[length - 1] = '_';
-      i::Utf8ToUC16CharacterStream stream(buffer, length);
+      i::Utf8ToUtf16CharacterStream stream(buffer, length);
       i::Scanner scanner(&unicode_cache);
       scanner.Initialize(&stream);
       CHECK_EQ(i::Token::IDENTIFIER, scanner.Next());
@@ -255,7 +255,7 @@
   uintptr_t stack_limit = i::Isolate::Current()->stack_guard()->real_climit();
   for (int i = 0; programs[i]; i++) {
     const char* program = programs[i];
-    i::Utf8ToUC16CharacterStream stream(
+    i::Utf8ToUtf16CharacterStream stream(
         reinterpret_cast<const i::byte*>(program),
         static_cast<unsigned>(strlen(program)));
     i::CompleteParserRecorder log;
@@ -291,7 +291,7 @@
   uintptr_t stack_limit = i::Isolate::Current()->stack_guard()->real_climit();
   for (int i = 0; programs[i]; i++) {
     const char* program = programs[i];
-    i::Utf8ToUC16CharacterStream stream(
+    i::Utf8ToUtf16CharacterStream stream(
         reinterpret_cast<const i::byte*>(program),
         static_cast<unsigned>(strlen(program)));
     i::CompleteParserRecorder log;
@@ -326,8 +326,9 @@
   // and then used the invalid currently scanned literal. This always
   // failed in debug mode, and sometimes crashed in release mode.
 
-  i::Utf8ToUC16CharacterStream stream(reinterpret_cast<const i::byte*>(program),
-                                      static_cast<unsigned>(strlen(program)));
+  i::Utf8ToUtf16CharacterStream stream(
+      reinterpret_cast<const i::byte*>(program),
+      static_cast<unsigned>(strlen(program)));
   i::ScriptDataImpl* data =
       i::ParserApi::PreParse(&stream, NULL, false);
   CHECK(data->HasError());
@@ -360,7 +361,7 @@
 
   int first_function =
       static_cast<int>(strstr(program, "function") - program);
-  int first_lbrace = first_function + static_cast<int>(strlen("function () "));
+  int first_lbrace = first_function + i::StrLength("function () ");
   CHECK_EQ('{', program[first_lbrace]);
   i::FunctionEntry entry1 = data->GetFunctionEntry(first_lbrace);
   CHECK(!entry1.is_valid());
@@ -368,7 +369,7 @@
   int second_function =
       static_cast<int>(strstr(program + first_lbrace, "function") - program);
   int second_lbrace =
-      second_function + static_cast<int>(strlen("function () "));
+      second_function + i::StrLength("function () ");
   CHECK_EQ('{', program[second_lbrace]);
   i::FunctionEntry entry2 = data->GetFunctionEntry(second_lbrace);
   CHECK(entry2.is_valid());
@@ -392,7 +393,7 @@
 
   uintptr_t stack_limit = i::Isolate::Current()->stack_guard()->real_climit();
 
-  i::Utf8ToUC16CharacterStream stream(
+  i::Utf8ToUtf16CharacterStream stream(
       reinterpret_cast<const i::byte*>(*program),
       static_cast<unsigned>(kProgramSize));
   i::CompleteParserRecorder log;
@@ -449,10 +450,10 @@
   i::Handle<i::String> uc16_string(
       FACTORY->NewExternalStringFromTwoByte(&resource));
 
-  i::ExternalTwoByteStringUC16CharacterStream uc16_stream(
+  i::ExternalTwoByteStringUtf16CharacterStream uc16_stream(
       i::Handle<i::ExternalTwoByteString>::cast(uc16_string), start, end);
-  i::GenericStringUC16CharacterStream string_stream(ascii_string, start, end);
-  i::Utf8ToUC16CharacterStream utf8_stream(
+  i::GenericStringUtf16CharacterStream string_stream(ascii_string, start, end);
+  i::Utf8ToUtf16CharacterStream utf8_stream(
       reinterpret_cast<const i::byte*>(ascii_source), end);
   utf8_stream.SeekForward(start);
 
@@ -575,12 +576,14 @@
   char buffer[kAllUtf8CharsSizeU];
   unsigned cursor = 0;
   for (int i = 0; i <= kMaxUC16Char; i++) {
-    cursor += unibrow::Utf8::Encode(buffer + cursor, i);
+    cursor += unibrow::Utf8::Encode(buffer + cursor,
+                                    i,
+                                    unibrow::Utf16::kNoPreviousCharacter);
   }
   ASSERT(cursor == kAllUtf8CharsSizeU);
 
-  i::Utf8ToUC16CharacterStream stream(reinterpret_cast<const i::byte*>(buffer),
-                                      kAllUtf8CharsSizeU);
+  i::Utf8ToUtf16CharacterStream stream(reinterpret_cast<const i::byte*>(buffer),
+                                       kAllUtf8CharsSizeU);
   for (int i = 0; i <= kMaxUC16Char; i++) {
     CHECK_EQU(i, stream.pos());
     int32_t c = stream.Advance();
@@ -610,7 +613,7 @@
 
 #undef CHECK_EQU
 
-void TestStreamScanner(i::UC16CharacterStream* stream,
+void TestStreamScanner(i::Utf16CharacterStream* stream,
                        i::Token::Value* expected_tokens,
                        int skip_pos = 0,  // Zero means not skipping.
                        int skip_to = 0) {
@@ -633,8 +636,8 @@
   v8::V8::Initialize();
 
   const char* str1 = "{ foo get for : */ <- \n\n /*foo*/ bib";
-  i::Utf8ToUC16CharacterStream stream1(reinterpret_cast<const i::byte*>(str1),
-                                       static_cast<unsigned>(strlen(str1)));
+  i::Utf8ToUtf16CharacterStream stream1(reinterpret_cast<const i::byte*>(str1),
+                                        static_cast<unsigned>(strlen(str1)));
   i::Token::Value expectations1[] = {
       i::Token::LBRACE,
       i::Token::IDENTIFIER,
@@ -652,8 +655,8 @@
   TestStreamScanner(&stream1, expectations1, 0, 0);
 
   const char* str2 = "case default const {THIS\nPART\nSKIPPED} do";
-  i::Utf8ToUC16CharacterStream stream2(reinterpret_cast<const i::byte*>(str2),
-                                       static_cast<unsigned>(strlen(str2)));
+  i::Utf8ToUtf16CharacterStream stream2(reinterpret_cast<const i::byte*>(str2),
+                                        static_cast<unsigned>(strlen(str2)));
   i::Token::Value expectations2[] = {
       i::Token::CASE,
       i::Token::DEFAULT,
@@ -683,7 +686,7 @@
   for (int i = 0; i <= 4; i++) {
      expectations3[6 - i] = i::Token::ILLEGAL;
      expectations3[5 - i] = i::Token::EOS;
-     i::Utf8ToUC16CharacterStream stream3(
+     i::Utf8ToUtf16CharacterStream stream3(
          reinterpret_cast<const i::byte*>(str3),
          static_cast<unsigned>(strlen(str3)));
      TestStreamScanner(&stream3, expectations3, 1, 1 + i);
@@ -692,7 +695,7 @@
 
 
 void TestScanRegExp(const char* re_source, const char* expected) {
-  i::Utf8ToUC16CharacterStream stream(
+  i::Utf8ToUtf16CharacterStream stream(
        reinterpret_cast<const i::byte*>(re_source),
        static_cast<unsigned>(strlen(re_source)));
   i::Scanner scanner(i::Isolate::Current()->unicode_cache());
@@ -748,6 +751,67 @@
 }
 
 
+static int Utf8LengthHelper(const char* s) {
+  int len = i::StrLength(s);
+  int character_length = len;
+  for (int i = 0; i < len; i++) {
+    unsigned char c = s[i];
+    int input_offset = 0;
+    int output_adjust = 0;
+    if (c > 0x7f) {
+      if (c < 0xc0) continue;
+      if (c >= 0xf0) {
+        if (c >= 0xf8) {
+          // 5 and 6 byte UTF-8 sequences turn into a kBadChar for each UTF-8
+          // byte.
+          continue;  // Handle first UTF-8 byte.
+        }
+        if ((c & 7) == 0 && ((s[i + 1] & 0x30) == 0)) {
+          // This 4 byte sequence could have been coded as a 3 byte sequence.
+          // Record a single kBadChar for the first byte and continue.
+          continue;
+        }
+        input_offset = 3;
+        // 4 bytes of UTF-8 turn into 2 UTF-16 code units.
+        character_length -= 2;
+      } else if (c >= 0xe0) {
+        if ((c & 0xf) == 0 && ((s[i + 1] & 0x20) == 0)) {
+          // This 3 byte sequence could have been coded as a 2 byte sequence.
+          // Record a single kBadChar for the first byte and continue.
+          continue;
+        }
+        input_offset = 2;
+        // 3 bytes of UTF-8 turn into 1 UTF-16 code unit.
+        output_adjust = 2;
+      } else {
+        if ((c & 0x1e) == 0) {
+          // This 2 byte sequence could have been coded as a 1 byte sequence.
+          // Record a single kBadChar for the first byte and continue.
+          continue;
+        }
+        input_offset = 1;
+        // 2 bytes of UTF-8 turn into 1 UTF-16 code unit.
+        output_adjust = 1;
+      }
+      bool bad = false;
+      for (int j = 1; j <= input_offset; j++) {
+        if ((s[i + j] & 0xc0) != 0x80) {
+          // Bad UTF-8 sequence turns the first in the sequence into kBadChar,
+          // which is a single UTF-16 code unit.
+          bad = true;
+          break;
+        }
+      }
+      if (!bad) {
+        i += input_offset;
+        character_length -= output_adjust;
+      }
+    }
+  }
+  return character_length;
+}
+
+
 TEST(ScopePositions) {
   // Test the parser for correctly setting the start and end positions
   // of a scope. We check the scope positions of exactly one scope
@@ -835,6 +899,91 @@
     { "  for ", "(let x in {})\n"
       "    statement;", "\n"
       "  more;", i::BLOCK_SCOPE, i::EXTENDED_MODE },
+    // Check that 6-byte and 4-byte encodings of UTF-8 strings do not throw
+    // the preparser off in terms of byte offsets.
+    // 6 byte encoding.
+    { "  'foo\355\240\201\355\260\211';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // 4 byte encoding.
+    { "  'foo\360\220\220\212';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // 3 byte encoding of \u0fff.
+    { "  'foo\340\277\277';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Broken 6 byte encoding with missing last byte.
+    { "  'foo\355\240\201\355\211';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Broken 3 byte encoding of \u0fff with missing last byte.
+    { "  'foo\340\277';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Broken 3 byte encoding of \u0fff with missing 2 last bytes.
+    { "  'foo\340';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Broken 3 byte encoding of \u00ff should be a 2 byte encoding.
+    { "  'foo\340\203\277';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Broken 3 byte encoding of \u007f should be a 2 byte encoding.
+    { "  'foo\340\201\277';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Unpaired lead surrogate.
+    { "  'foo\355\240\201';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Unpaired lead surrogate where following code point is a 3 byte sequence.
+    { "  'foo\355\240\201\340\277\277';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Unpaired lead surrogate where following code point is a 4 byte encoding
+    // of a trail surrogate.
+    { "  'foo\355\240\201\360\215\260\211';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Unpaired trail surrogate.
+    { "  'foo\355\260\211';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // 2 byte encoding of \u00ff.
+    { "  'foo\303\277';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Broken 2 byte encoding of \u00ff with missing last byte.
+    { "  'foo\303';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Broken 2 byte encoding of \u007f should be a 1 byte encoding.
+    { "  'foo\301\277';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Illegal 5 byte encoding.
+    { "  'foo\370\277\277\277\277';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Illegal 6 byte encoding.
+    { "  'foo\374\277\277\277\277\277';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Illegal 0xfe byte
+    { "  'foo\376\277\277\277\277\277\277';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    // Illegal 0xff byte
+    { "  'foo\377\277\277\277\277\277\277\277';\n"
+      "  (function fun", "(a,b) { infunction; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    { "  'foo';\n"
+      "  (function fun", "(a,b) { 'bar\355\240\201\355\260\213'; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
+    { "  'foo';\n"
+      "  (function fun", "(a,b) { 'bar\360\220\220\214'; }", ")();",
+      i::FUNCTION_SCOPE, i::CLASSIC_MODE },
     { NULL, NULL, NULL, i::EVAL_SCOPE, i::CLASSIC_MODE }
   };
 
@@ -848,20 +997,24 @@
   i::FLAG_harmony_scoping = true;
 
   for (int i = 0; source_data[i].outer_prefix; i++) {
-    int kPrefixLen = i::StrLength(source_data[i].outer_prefix);
-    int kInnerLen = i::StrLength(source_data[i].inner_source);
-    int kSuffixLen = i::StrLength(source_data[i].outer_suffix);
+    int kPrefixLen = Utf8LengthHelper(source_data[i].outer_prefix);
+    int kInnerLen = Utf8LengthHelper(source_data[i].inner_source);
+    int kSuffixLen = Utf8LengthHelper(source_data[i].outer_suffix);
+    int kPrefixByteLen = i::StrLength(source_data[i].outer_prefix);
+    int kInnerByteLen = i::StrLength(source_data[i].inner_source);
+    int kSuffixByteLen = i::StrLength(source_data[i].outer_suffix);
     int kProgramSize = kPrefixLen + kInnerLen + kSuffixLen;
-    i::Vector<char> program = i::Vector<char>::New(kProgramSize + 1);
-    int length = i::OS::SNPrintF(program, "%s%s%s",
-                                 source_data[i].outer_prefix,
-                                 source_data[i].inner_source,
-                                 source_data[i].outer_suffix);
-    CHECK(length == kProgramSize);
+    int kProgramByteSize = kPrefixByteLen + kInnerByteLen + kSuffixByteLen;
+    i::Vector<char> program = i::Vector<char>::New(kProgramByteSize + 1);
+    i::OS::SNPrintF(program, "%s%s%s",
+                             source_data[i].outer_prefix,
+                             source_data[i].inner_source,
+                             source_data[i].outer_suffix);
 
     // Parse program source.
     i::Handle<i::String> source(
-        FACTORY->NewStringFromAscii(i::CStrVector(program.start())));
+        FACTORY->NewStringFromUtf8(i::CStrVector(program.start())));
+    CHECK_EQ(source->length(), kProgramSize);
     i::Handle<i::Script> script = FACTORY->NewScript(source);
     i::Parser parser(script, i::kAllowLazy | i::EXTENDED_MODE, NULL, NULL);
     i::CompilationInfo info(script);
@@ -894,7 +1047,7 @@
   // Preparse the data.
   i::CompleteParserRecorder log;
   i::Scanner scanner(i::Isolate::Current()->unicode_cache());
-  i::GenericStringUC16CharacterStream stream(source, 0, source->length());
+  i::GenericStringUtf16CharacterStream stream(source, 0, source->length());
   scanner.SetHarmonyScoping(harmony_scoping);
   scanner.Initialize(&stream);
   v8::preparser::PreParser::PreParseResult result =
diff --git a/test/mjsunit/compiler/inline-arguments.js b/test/mjsunit/compiler/inline-arguments.js
index 532fc26..b79f59c 100644
--- a/test/mjsunit/compiler/inline-arguments.js
+++ b/test/mjsunit/compiler/inline-arguments.js
@@ -27,11 +27,30 @@
 
 // Flags: --allow-natives-syntax
 
-// Test inlining functions that use arguments.
-function f() { return g(1, 2, 3); }
+function A() {
+}
 
-function g(x, y, z) { return %_ArgumentsLength(); }
+A.prototype.X = function (a, b, c) {
+  assertTrue(this instanceof A);
+  assertEquals(1, a);
+  assertEquals(2, b);
+  assertEquals(3, c);
+};
 
-for (var i = 0; i < 5; ++i) f();
-%OptimizeFunctionOnNextCall(f);
-assertEquals(3, f());
+A.prototype.Y = function () {
+  this.X.apply(this, arguments);
+};
+
+A.prototype.Z = function () {
+  this.Y(1,2,3);
+};
+
+var a = new A();
+a.Z(4,5,6);
+a.Z(4,5,6);
+%OptimizeFunctionOnNextCall(a.Z);
+a.Z(4,5,6);
+A.prototype.X.apply = function (receiver, args) {
+  return Function.prototype.apply.call(this, receiver, args);
+};
+a.Z(4,5,6);
diff --git a/test/mjsunit/regress/regress-sqrt.js b/test/mjsunit/regress/regress-sqrt.js
new file mode 100644
index 0000000..f2a7e55
--- /dev/null
+++ b/test/mjsunit/regress/regress-sqrt.js
@@ -0,0 +1,47 @@
+// Copyright 2012 the V8 project authors. All rights reserved.
+// Redistribution and use in source and binary forms, with or without
+// modification, are permitted provided that the following conditions are
+// met:
+//
+//     * Redistributions of source code must retain the above copyright
+//       notice, this list of conditions and the following disclaimer.
+//     * Redistributions in binary form must reproduce the above
+//       copyright notice, this list of conditions and the following
+//       disclaimer in the documentation and/or other materials provided
+//       with the distribution.
+//     * Neither the name of Google Inc. nor the names of its
+//       contributors may be used to endorse or promote products derived
+//       from this software without specific prior written permission.
+//
+// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+// Flags: --allow-natives-syntax
+
+// Check that Math.sqrt returns the same value regardless of being
+// optimized or not.
+
+function f(x) {
+  return Math.sqrt(x);
+}
+
+var x = 7.0506280066499245e-233;
+
+var a = f(x);
+
+f(0.1);
+f(0.2);
+%OptimizeFunctionOnNextCall(f);
+
+var b = f(x);
+
+assertEquals(a, b);