[JSC] Implement String#isWellFormed and String#toWellFormed

https://bugs.webkit.org/show_bug.cgi?id=248588
rdar://problem/102849210

Reviewed by Ross Kirsling.

This patch implements String#isWellFormed and String#toWellFormed proposal[1].
String#isWellFormed returns true if String does not have wrong surrogate pairs (e.g. non-paired surrogate).
String#toWellFormed returns a string, which replaces non well-formed characters with unicode replacement character (\ufffd).

[1]: https://github.com/tc39/proposal-is-usv-string

* JSTests/stress/string-well-formed.js: Added.
(shouldBe):
* LayoutTests/js/Object-getOwnPropertyNames-expected.txt:
* LayoutTests/js/script-tests/Object-getOwnPropertyNames.js:
* Source/JavaScriptCore/runtime/CommonIdentifiers.h:
* Source/JavaScriptCore/runtime/OptionsList.h:
* Source/JavaScriptCore/runtime/StringPrototype.cpp:
(JSC::StringPrototype::finishCreation):
(JSC::nonWellFormedIndex):
(JSC::JSC_DEFINE_HOST_FUNCTION):

Canonical link: https://commits.webkit.org/257250@main

JSTests/stress/string-well-formed.js

@@ -0,0 +1,32 @@
+//@ requireOptions("--useStringWellFormed=1")
+function shouldBe(actual, expected) {
+    if (actual !== expected)
+        throw new Error('bad value: ' + actual);
+}
+
+shouldBe("".isWellFormed(), true);
+shouldBe("".toWellFormed(), "");
+
+shouldBe("Hello World".isWellFormed(), true);
+shouldBe("Hello World".toWellFormed(), "Hello World");
+
+shouldBe("こんにちわ".isWellFormed(), true);
+shouldBe("こんにちわ".toWellFormed(), "こんにちわ");
+
+shouldBe("𠮷野家".isWellFormed(), true);
+shouldBe("𠮷野家".toWellFormed(), "𠮷野家");
+
+shouldBe("A\uD842".isWellFormed(), false);
+shouldBe("A\uD842".toWellFormed(), "A\uFFFD");
+
+shouldBe("A\uD842A".isWellFormed(), false);
+shouldBe("A\uD842A".toWellFormed(), "A\uFFFDA");
+
+shouldBe("A\uD842\uDFB7".isWellFormed(), true);
+shouldBe("A\uD842\uDFB7".toWellFormed(), "A\uD842\uDFB7");
+
+shouldBe("A\uDFB7".isWellFormed(), false);
+shouldBe("A\uDFB7".toWellFormed(), "A\uFFFD");
+
+shouldBe("A\uDFB7\uD842".isWellFormed(), false);
+shouldBe("A\uDFB7\uD842".toWellFormed(), "A\uFFFD\uFFFD");

LayoutTests/js/Object-getOwnPropertyNames-expected.txt

@@ -49,7 +49,7 @@ PASS getSortedOwnPropertyNames(Function.prototype) is ['apply', 'arguments', 'bi
 PASS getSortedOwnPropertyNames(Array) is ['from', 'fromAsync', 'isArray', 'length', 'name', 'of', 'prototype']
 PASS getSortedOwnPropertyNames(Array.prototype) is ['at', 'concat', 'constructor', 'copyWithin', 'entries', 'every', 'fill', 'filter', 'find', 'findIndex', 'findLast', 'findLastIndex', 'flat', 'flatMap', 'forEach', 'group', 'groupToMap', 'includes', 'indexOf', 'join', 'keys', 'lastIndexOf', 'length', 'map', 'pop', 'push', 'reduce', 'reduceRight', 'reverse', 'shift', 'slice', 'some', 'sort', 'splice', 'toLocaleString', 'toReversed', 'toSorted', 'toSpliced', 'toString', 'unshift', 'values', 'with']
 PASS getSortedOwnPropertyNames(String) is ['fromCharCode', 'fromCodePoint', 'length', 'name', 'prototype', 'raw']
-PASS getSortedOwnPropertyNames(String.prototype) is ['anchor', 'at', 'big', 'blink', 'bold', 'charAt', 'charCodeAt', 'codePointAt', 'concat', 'constructor', 'endsWith', 'fixed', 'fontcolor', 'fontsize', 'includes', 'indexOf', 'italics', 'lastIndexOf', 'length', 'link', 'localeCompare', 'match', 'matchAll', 'normalize', 'padEnd', 'padStart', 'repeat', 'replace', 'replaceAll', 'search', 'slice', 'small', 'split', 'startsWith', 'strike', 'sub', 'substr', 'substring', 'sup', 'toLocaleLowerCase', 'toLocaleUpperCase', 'toLowerCase', 'toString', 'toUpperCase', 'trim', 'trimEnd', 'trimLeft', 'trimRight', 'trimStart', 'valueOf']
+PASS getSortedOwnPropertyNames(String.prototype) is ['anchor', 'at', 'big', 'blink', 'bold', 'charAt', 'charCodeAt', 'codePointAt', 'concat', 'constructor', 'endsWith', 'fixed', 'fontcolor', 'fontsize', 'includes', 'indexOf', 'isWellFormed', 'italics', 'lastIndexOf', 'length', 'link', 'localeCompare', 'match', 'matchAll', 'normalize', 'padEnd', 'padStart', 'repeat', 'replace', 'replaceAll', 'search', 'slice', 'small', 'split', 'startsWith', 'strike', 'sub', 'substr', 'substring', 'sup', 'toLocaleLowerCase', 'toLocaleUpperCase', 'toLowerCase', 'toString', 'toUpperCase', 'toWellFormed', 'trim', 'trimEnd', 'trimLeft', 'trimRight', 'trimStart', 'valueOf']
 PASS getSortedOwnPropertyNames(Boolean) is ['length', 'name', 'prototype']
 PASS getSortedOwnPropertyNames(Boolean.prototype) is ['constructor', 'toString', 'valueOf']
 PASS getSortedOwnPropertyNames(Number) is ['EPSILON', 'MAX_SAFE_INTEGER', 'MAX_VALUE', 'MIN_SAFE_INTEGER', 'MIN_VALUE', 'NEGATIVE_INFINITY', 'NaN', 'POSITIVE_INFINITY', 'isFinite', 'isInteger', 'isNaN', 'isSafeInteger', 'length', 'name', 'parseFloat', 'parseInt', 'prototype']

LayoutTests/js/script-tests/Object-getOwnPropertyNames.js

@@ -58,7 +58,7 @@ var expectedPropertyNamesSet = {
     "Array": "['from', 'fromAsync', 'isArray', 'length', 'name', 'of', 'prototype']",
     "Array.prototype": "['at', 'concat', 'constructor', 'copyWithin', 'entries', 'every', 'fill', 'filter', 'find', 'findIndex', 'findLast', 'findLastIndex', 'flat', 'flatMap', 'forEach', 'group', 'groupToMap', 'includes', 'indexOf', 'join', 'keys', 'lastIndexOf', 'length', 'map', 'pop', 'push', 'reduce', 'reduceRight', 'reverse', 'shift', 'slice', 'some', 'sort', 'splice', 'toLocaleString', 'toReversed', 'toSorted', 'toSpliced', 'toString', 'unshift', 'values', 'with']",
     "String": "['fromCharCode', 'fromCodePoint', 'length', 'name', 'prototype', 'raw']",
-    "String.prototype": "['anchor', 'at', 'big', 'blink', 'bold', 'charAt', 'charCodeAt', 'codePointAt', 'concat', 'constructor', 'endsWith', 'fixed', 'fontcolor', 'fontsize', 'includes', 'indexOf', 'italics', 'lastIndexOf', 'length', 'link', 'localeCompare', 'match', 'matchAll', 'normalize', 'padEnd', 'padStart', 'repeat', 'replace', 'replaceAll', 'search', 'slice', 'small', 'split', 'startsWith', 'strike', 'sub', 'substr', 'substring', 'sup', 'toLocaleLowerCase', 'toLocaleUpperCase', 'toLowerCase', 'toString', 'toUpperCase', 'trim', 'trimEnd', 'trimLeft', 'trimRight', 'trimStart', 'valueOf']",
+    "String.prototype": "['anchor', 'at', 'big', 'blink', 'bold', 'charAt', 'charCodeAt', 'codePointAt', 'concat', 'constructor', 'endsWith', 'fixed', 'fontcolor', 'fontsize', 'includes', 'indexOf', 'isWellFormed', 'italics', 'lastIndexOf', 'length', 'link', 'localeCompare', 'match', 'matchAll', 'normalize', 'padEnd', 'padStart', 'repeat', 'replace', 'replaceAll', 'search', 'slice', 'small', 'split', 'startsWith', 'strike', 'sub', 'substr', 'substring', 'sup', 'toLocaleLowerCase', 'toLocaleUpperCase', 'toLowerCase', 'toString', 'toUpperCase', 'toWellFormed', 'trim', 'trimEnd', 'trimLeft', 'trimRight', 'trimStart', 'valueOf']",
     "Boolean": "['length', 'name', 'prototype']",
     "Boolean.prototype": "['constructor', 'toString', 'valueOf']",
     "Number": "['EPSILON', 'MAX_SAFE_INTEGER', 'MAX_VALUE', 'MIN_SAFE_INTEGER', 'MIN_VALUE', 'NEGATIVE_INFINITY', 'NaN', 'POSITIVE_INFINITY', 'isFinite', 'isInteger', 'isNaN', 'isSafeInteger', 'length', 'name', 'parseFloat', 'parseInt', 'prototype']",

Source/JavaScriptCore/runtime/CommonIdentifiers.h

@@ -171,6 +171,7 @@
     macro(isPrototypeOf) \
     macro(isView) \
     macro(isWatchpoint) \
+    macro(isWellFormed) \
     macro(isWordLike) \
     macro(jettisonReason) \
     macro(join) \
@@ -270,6 +271,7 @@
     macro(toPrecision) \
     macro(toString) \
     macro(toTemporalInstant) \
+    macro(toWellFormed) \
     macro(trailingZeroDisplay) \
     macro(transfer) \
     macro(type) \

Source/JavaScriptCore/runtime/OptionsList.h

@@ -553,6 +553,7 @@ bool canUseWebAssemblyFastMemory();
     v(Bool, useResizableArrayBuffer, true, Normal, "Expose ResizableArrayBuffer feature.") \
     v(Bool, useSharedArrayBuffer, false, Normal, nullptr) \
     v(Bool, useShadowRealm, false, Normal, "Expose the ShadowRealm object.") \
+    v(Bool, useStringWellFormed, true, Normal, "Expose the String well-formed methods.") \
     v(Bool, useTemporal, false, Normal, "Expose the Temporal object.") \
     v(Bool, useWebAssemblyThreading, true, Normal, "Allow instructions from the wasm threading spec.") \
     v(Bool, useWebAssemblyTypedFunctionReferences, false, Normal, "Allow function types from the wasm typed function references spec.") \

Source/JavaScriptCore/runtime/StringPrototype.cpp

@@ -77,6 +77,8 @@ static JSC_DECLARE_HOST_FUNCTION(stringProtoFuncEndsWith);
 static JSC_DECLARE_HOST_FUNCTION(stringProtoFuncIncludes);
 static JSC_DECLARE_HOST_FUNCTION(stringProtoFuncNormalize);
 static JSC_DECLARE_HOST_FUNCTION(stringProtoFuncIterator);
+static JSC_DECLARE_HOST_FUNCTION(stringProtoFuncIsWellFormed);
+static JSC_DECLARE_HOST_FUNCTION(stringProtoFuncToWellFormed);
 
 }
 
@@ -164,6 +166,11 @@ void StringPrototype::finishCreation(VM& vm, JSGlobalObject* globalObject)
     JSC_NATIVE_FUNCTION_WITHOUT_TRANSITION(vm.propertyNames->builtinNames().substrPrivateName(), stringProtoFuncSubstr, static_cast<unsigned>(PropertyAttribute::DontEnum), 2, ImplementationVisibility::Public);
     JSC_NATIVE_FUNCTION_WITHOUT_TRANSITION(vm.propertyNames->builtinNames().endsWithPrivateName(), stringProtoFuncEndsWith, static_cast<unsigned>(PropertyAttribute::DontEnum), 2, ImplementationVisibility::Public);
 
+    if (Options::useStringWellFormed()) {
+        JSC_NATIVE_FUNCTION_WITHOUT_TRANSITION(vm.propertyNames->isWellFormed, stringProtoFuncIsWellFormed, static_cast<unsigned>(PropertyAttribute::DontEnum), 0, ImplementationVisibility::Public);
+        JSC_NATIVE_FUNCTION_WITHOUT_TRANSITION(vm.propertyNames->toWellFormed, stringProtoFuncToWellFormed, static_cast<unsigned>(PropertyAttribute::DontEnum), 0, ImplementationVisibility::Public);
+    }
+
     // The constructor will be added later, after StringConstructor has been built
 }
 
@@ -1757,4 +1764,117 @@ JSC_DEFINE_HOST_FUNCTION(stringProtoFuncNormalize, (JSGlobalObject* globalObject
     RELEASE_AND_RETURN(scope, JSValue::encode(normalize(globalObject, string, form)));
 }
 
+static inline std::optional<unsigned> illFormedIndex(const UChar* characters, unsigned length)
+{
+    for (unsigned index = 0; index < length; ++index) {
+        UChar character = characters[index];
+        if (!U16_IS_SURROGATE(character))
+            continue;
+
+        if (U16_IS_SURROGATE_TRAIL(character))
+            return index;
+
+        ASSERT(U16_IS_SURROGATE_LEAD(character));
+        if ((index + 1) == length)
+            return index;
+        UChar nextCharacter = characters[index + 1];
+
+        if (!U16_IS_SURROGATE(nextCharacter))
+            return index;
+
+        if (!U16_IS_SURROGATE_TRAIL(nextCharacter))
+            return index;
+
+        ++index; // Increment additionally.
+    }
+    return std::nullopt;
+}
+
+JSC_DEFINE_HOST_FUNCTION(stringProtoFuncIsWellFormed, (JSGlobalObject* globalObject, CallFrame* callFrame))
+{
+    VM& vm = globalObject->vm();
+    auto scope = DECLARE_THROW_SCOPE(vm);
+
+    JSValue thisValue = callFrame->thisValue();
+    if (!checkObjectCoercible(thisValue))
+        return throwVMTypeError(globalObject, scope);
+
+    // Latin-1 characters do not have surrogates.
+    if (thisValue.isString() && asString(thisValue)->is8Bit())
+        return JSValue::encode(jsBoolean(true));
+
+    String string = thisValue.toWTFString(globalObject);
+    RETURN_IF_EXCEPTION(scope, { });
+
+    if (string.is8Bit())
+        return JSValue::encode(jsBoolean(true));
+    return JSValue::encode(jsBoolean(!illFormedIndex(string.characters16(), string.length())));
+}
+
+JSC_DEFINE_HOST_FUNCTION(stringProtoFuncToWellFormed, (JSGlobalObject* globalObject, CallFrame* callFrame))
+{
+    VM& vm = globalObject->vm();
+    auto scope = DECLARE_THROW_SCOPE(vm);
+
+    JSValue thisValue = callFrame->thisValue();
+    if (!checkObjectCoercible(thisValue))
+        return throwVMTypeError(globalObject, scope);
+
+    // Latin-1 characters do not have surrogates.
+    if (thisValue.isString() && asString(thisValue)->is8Bit())
+        return JSValue::encode(thisValue);
+
+    String string = thisValue.toWTFString(globalObject);
+    RETURN_IF_EXCEPTION(scope, { });
+
+    if (string.is8Bit())
+        return JSValue::encode(thisValue);
+
+    const UChar* characters = string.characters16();
+    unsigned length = string.length();
+    auto firstIllFormedIndex = illFormedIndex(characters, length);
+    if (!firstIllFormedIndex)
+        return JSValue::encode(thisValue);
+
+    Vector<UChar> buffer;
+    buffer.reserveInitialCapacity(length);
+    buffer.append(characters, firstIllFormedIndex.value());
+    for (unsigned index = firstIllFormedIndex.value(); index < length; ++index) {
+        UChar character = characters[index];
+
+        if (!U16_IS_SURROGATE(character)) {
+            buffer.append(character);
+            continue;
+        }
+
+        if (U16_IS_SURROGATE_TRAIL(character)) {
+            buffer.append(replacementCharacter);
+            continue;
+        }
+
+        ASSERT(U16_IS_SURROGATE_LEAD(character));
+        if ((index + 1) == length) {
+            buffer.append(replacementCharacter);
+            continue;
+        }
+        UChar nextCharacter = characters[index + 1];
+
+        if (!U16_IS_SURROGATE(nextCharacter)) {
+            buffer.append(replacementCharacter);
+            continue;
+        }
+
+        if (!U16_IS_SURROGATE_TRAIL(nextCharacter)) {
+            buffer.append(replacementCharacter);
+            continue;
+        }
+
+        buffer.append(character);
+        buffer.append(nextCharacter);
+        index += 1;
+    }
+
+    return JSValue::encode(jsString(vm, String::adopt(WTFMove(buffer))));
+}
+
 } // namespace JSC