src: move ToUSVString() to node_util.cc

Since `toUSVString()` was exposed in `util` as a public API, not only
for internal `url` any more.

PR-URL: https://github.com/nodejs/node/pull/40204
Reviewed-By: James M Snell <jasnell@gmail.com>
Reviewed-By: Colin Ihrig <cjihrig@gmail.com>
Reviewed-By: Zijian Liu <lxxyxzj@gmail.com>

lib/internal/util.js

@@ -28,10 +28,6 @@ const {
   SymbolFor,
 } = primordials;
 
-const {
-  toUSVString: _toUSVString,
-} = internalBinding('url');
-
 const {
   hideStackFrames,
   codes: {
@@ -47,7 +43,8 @@ const {
   setHiddenValue,
   arrow_message_private_symbol: kArrowMessagePrivateSymbolIndex,
   decorated_private_symbol: kDecoratedPrivateSymbolIndex,
-  sleep: _sleep
+  sleep: _sleep,
+  toUSVString: _toUSVString,
 } = internalBinding('util');
 const { isNativeError } = internalBinding('types');
 

src/node_url.cc

@@ -52,9 +52,6 @@ namespace {
 // https://url.spec.whatwg.org/#eof-code-point
 constexpr char kEOL = -1;
 
-// Used in ToUSVString().
-constexpr char16_t kUnicodeReplacementCharacter = 0xFFFD;
-
 // https://url.spec.whatwg.org/#concept-host
 class URLHost {
  public:
@@ -160,14 +157,6 @@ enum url_error_cb_args {
 #undef XX
 };
 
-#define CHAR_TEST(bits, name, expr)                                           \
-  template <typename T>                                                       \
-  bool name(const T ch) {                                              \
-    static_assert(sizeof(ch) >= (bits) / 8,                                   \
-                  "Character must be wider than " #bits " bits");             \
-    return (expr);                                                            \
-  }
-
 #define TWO_CHAR_STRING_TEST(bits, name, expr)                                \
   template <typename T>                                                       \
   bool name(const T ch1, const T ch2) {                                \
@@ -225,19 +214,8 @@ TWO_CHAR_STRING_TEST(8, IsWindowsDriveLetter,
 TWO_CHAR_STRING_TEST(8, IsNormalizedWindowsDriveLetter,
                      (IsASCIIAlpha(ch1) && ch2 == ':'))
 
-// If a UTF-16 character is a low/trailing surrogate.
-CHAR_TEST(16, IsUnicodeTrail, (ch & 0xFC00) == 0xDC00)
-
-// If a UTF-16 character is a surrogate.
-CHAR_TEST(16, IsUnicodeSurrogate, (ch & 0xF800) == 0xD800)
-
-// If a UTF-16 surrogate is a low/trailing one.
-CHAR_TEST(16, IsUnicodeSurrogateTrail, (ch & 0x400) != 0)
-
-#undef CHAR_TEST
 #undef TWO_CHAR_STRING_TEST
 
-
 bool BitAt(const uint8_t a[], const uint8_t i) {
   return !!(a[i >> 3] & (1 << (i & 7)));
 }
@@ -1736,40 +1714,6 @@ void EncodeAuthSet(const FunctionCallbackInfo<Value>& args) {
       String::NewFromUtf8(env->isolate(), output.c_str()).ToLocalChecked());
 }
 
-void ToUSVString(const FunctionCallbackInfo<Value>& args) {
-  Environment* env = Environment::GetCurrent(args);
-  CHECK_GE(args.Length(), 2);
-  CHECK(args[0]->IsString());
-  CHECK(args[1]->IsNumber());
-
-  TwoByteValue value(env->isolate(), args[0]);
-
-  int64_t start = args[1]->IntegerValue(env->context()).FromJust();
-  CHECK_GE(start, 0);
-
-  for (size_t i = start; i < value.length(); i++) {
-    char16_t c = value[i];
-    if (!IsUnicodeSurrogate(c)) {
-      continue;
-    } else if (IsUnicodeSurrogateTrail(c) || i == value.length() - 1) {
-      value[i] = kUnicodeReplacementCharacter;
-    } else {
-      char16_t d = value[i + 1];
-      if (IsUnicodeTrail(d)) {
-        i++;
-      } else {
-        value[i] = kUnicodeReplacementCharacter;
-      }
-    }
-  }
-
-  args.GetReturnValue().Set(
-      String::NewFromTwoByte(env->isolate(),
-                             *value,
-                             NewStringType::kNormal,
-                             value.length()).ToLocalChecked());
-}
-
 void DomainToASCII(const FunctionCallbackInfo<Value>& args) {
   Environment* env = Environment::GetCurrent(args);
   CHECK_GE(args.Length(), 1);
@@ -1820,7 +1764,6 @@ void Initialize(Local<Object> target,
   Environment* env = Environment::GetCurrent(context);
   env->SetMethod(target, "parse", Parse);
   env->SetMethodNoSideEffect(target, "encodeAuth", EncodeAuthSet);
-  env->SetMethodNoSideEffect(target, "toUSVString", ToUSVString);
   env->SetMethodNoSideEffect(target, "domainToASCII", DomainToASCII);
   env->SetMethodNoSideEffect(target, "domainToUnicode", DomainToUnicode);
   env->SetMethod(target, "setURLConstructor", SetURLConstructor);
@@ -1838,7 +1781,6 @@ void Initialize(Local<Object> target,
 void RegisterExternalReferences(ExternalReferenceRegistry* registry) {
   registry->Register(Parse);
   registry->Register(EncodeAuthSet);
-  registry->Register(ToUSVString);
   registry->Register(DomainToASCII);
   registry->Register(DomainToUnicode);
   registry->Register(SetURLConstructor);

src/node_util.cc

@@ -35,6 +35,18 @@ using v8::String;
 using v8::Uint32;
 using v8::Value;
 
+// Used in ToUSVString().
+constexpr char16_t kUnicodeReplacementCharacter = 0xFFFD;
+
+// If a UTF-16 character is a low/trailing surrogate.
+CHAR_TEST(16, IsUnicodeTrail, (ch & 0xFC00) == 0xDC00)
+
+// If a UTF-16 character is a surrogate.
+CHAR_TEST(16, IsUnicodeSurrogate, (ch & 0xF800) == 0xD800)
+
+// If a UTF-16 surrogate is a low/trailing one.
+CHAR_TEST(16, IsUnicodeSurrogateTrail, (ch & 0x400) != 0)
+
 static void GetOwnNonIndexProperties(
     const FunctionCallbackInfo<Value>& args) {
   Environment* env = Environment::GetCurrent(args);
@@ -282,6 +294,40 @@ static void IsConstructor(const FunctionCallbackInfo<Value>& args) {
   args.GetReturnValue().Set(args[0].As<v8::Function>()->IsConstructor());
 }
 
+static void ToUSVString(const FunctionCallbackInfo<Value>& args) {
+  Environment* env = Environment::GetCurrent(args);
+  CHECK_GE(args.Length(), 2);
+  CHECK(args[0]->IsString());
+  CHECK(args[1]->IsNumber());
+
+  TwoByteValue value(env->isolate(), args[0]);
+
+  int64_t start = args[1]->IntegerValue(env->context()).FromJust();
+  CHECK_GE(start, 0);
+
+  for (size_t i = start; i < value.length(); i++) {
+    char16_t c = value[i];
+    if (!IsUnicodeSurrogate(c)) {
+      continue;
+    } else if (IsUnicodeSurrogateTrail(c) || i == value.length() - 1) {
+      value[i] = kUnicodeReplacementCharacter;
+    } else {
+      char16_t d = value[i + 1];
+      if (IsUnicodeTrail(d)) {
+        i++;
+      } else {
+        value[i] = kUnicodeReplacementCharacter;
+      }
+    }
+  }
+
+  args.GetReturnValue().Set(
+      String::NewFromTwoByte(env->isolate(),
+                             *value,
+                             v8::NewStringType::kNormal,
+                             value.length()).ToLocalChecked());
+}
+
 void RegisterExternalReferences(ExternalReferenceRegistry* registry) {
   registry->Register(GetHiddenValue);
   registry->Register(SetHiddenValue);
@@ -299,6 +345,7 @@ void RegisterExternalReferences(ExternalReferenceRegistry* registry) {
   registry->Register(WeakReference::DecRef);
   registry->Register(GuessHandleType);
   registry->Register(IsConstructor);
+  registry->Register(ToUSVString);
 }
 
 void Initialize(Local<Object> target,
@@ -370,6 +417,8 @@ void Initialize(Local<Object> target,
   env->SetConstructorFunction(target, "WeakReference", weak_ref);
 
   env->SetMethod(target, "guessHandleType", GuessHandleType);
+
+  env->SetMethodNoSideEffect(target, "toUSVString", ToUSVString);
 }
 
 }  // namespace util

src/util-inl.h

@@ -64,6 +64,14 @@
   (((x) & 0x00000000000000FFull) << 56)
 #endif
 
+#define CHAR_TEST(bits, name, expr)                                           \
+  template <typename T>                                                       \
+  bool name(const T ch) {                                                     \
+    static_assert(sizeof(ch) >= (bits) / 8,                                   \
+                  "Character must be wider than " #bits " bits");             \
+    return (expr);                                                            \
+  }
+
 namespace node {
 
 template <typename T>