src/kjs/string_object.cpp

/*
 *  This file is part of the KDE libraries
 *  Copyright (C) 1999-2001 Harri Porten (porten@kde.org)
 *  Copyright (C) 2004 Apple Computer, Inc.
 *
 *  This library is free software; you can redistribute it and/or
 *  modify it under the terms of the GNU Lesser General Public
 *  License as published by the Free Software Foundation; either
 *  version 2 of the License, or (at your option) any later version.
 *
 *  This library is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 *  Lesser General Public License for more details.
 *
 *  You should have received a copy of the GNU Lesser General Public
 *  License along with this library; if not, write to the Free Software
 *  Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
 *
 */

#include "string_object.h"
#include "string_object.lut.h"

#include "error_object.h"
#include "operations.h"
#include "PropertyNameArray.h"
#include "regexp_object.h"
#include "commonunicode.h"
#include <wtf/unicode/libc/UnicodeLibC.h>

#if PLATFORM(WIN_OS)
#include <windows.h>
#endif

using namespace WTF;

namespace KJS
{

// ------------------------------ StringInstance ----------------------------

const ClassInfo StringInstance::info = {"String", nullptr, nullptr, nullptr};

StringInstance::StringInstance(JSObject *proto)
    : JSWrapperObject(proto), m_conversionsCustomized(false)
{
    setInternalValue(jsString(""));
}

StringInstance::StringInstance(JSObject *proto, StringImp *string)
    : JSWrapperObject(proto), m_conversionsCustomized(false)
{
    setInternalValue(string);
}

StringInstance::StringInstance(JSObject *proto, const UString &string)
    : JSWrapperObject(proto), m_conversionsCustomized(false)
{
    setInternalValue(jsString(string));
}

JSValue *StringInstance::lengthGetter(ExecState *, JSObject *, const Identifier &, const PropertySlot &slot)
{
    return jsNumber(static_cast<StringInstance *>(slot.slotBase())->internalValue()->value().size());
}

JSValue *StringInstance::indexGetter(ExecState *, JSObject *, unsigned, const PropertySlot &slot)
{
    const UChar c = static_cast<StringInstance *>(slot.slotBase())->internalValue()->value()[slot.index()];
    return jsString(UString(&c, 1));
}

bool StringInstance::getOwnPropertySlot(ExecState *exec, const Identifier &propertyName, PropertySlot &slot)
{
    if (propertyName == exec->propertyNames().length) {
        slot.setCustom(this, lengthGetter);
        return true;
    }

    bool isStrictUInt32;
    unsigned i = propertyName.toStrictUInt32(&isStrictUInt32);
    unsigned length = internalValue()->value().size();
    if (isStrictUInt32 && i < length) {
        slot.setCustomIndex(this, i, indexGetter);
        return true;
    }

    return JSObject::getOwnPropertySlot(exec, propertyName, slot);
}

bool StringInstance::getOwnPropertySlot(ExecState *exec, unsigned propertyName, PropertySlot &slot)
{
    unsigned length = internalValue()->value().size();
    if (propertyName < length) {
        slot.setCustomIndex(this, propertyName, indexGetter);
        return true;
    }

    return JSObject::getOwnPropertySlot(exec, Identifier::from(propertyName), slot);
}

bool StringInstance::getOwnPropertyDescriptor(ExecState *exec, const Identifier &propertyName, PropertyDescriptor &desc)
{
    if (propertyName == exec->propertyNames().length) {
        desc.setPropertyDescriptorValues(exec, jsNumber(internalValue()->value().size()), ReadOnly | DontDelete | DontEnum);
        return true;
    }

    return KJS::JSObject::getOwnPropertyDescriptor(exec, propertyName, desc);
}

void StringInstance::put(ExecState *exec, const Identifier &propertyName, JSValue *value, int attr)
{
    if (propertyName == exec->propertyNames().length) {
        return;
    }

    if (propertyName == exec->propertyNames().valueOf || propertyName == exec->propertyNames().toString) {
        m_conversionsCustomized = true;
    }

    JSObject::put(exec, propertyName, value, attr);
}

bool StringInstance::deleteProperty(ExecState *exec, const Identifier &propertyName)
{
    if (propertyName == exec->propertyNames().length) {
        return false;
    }
    return JSObject::deleteProperty(exec, propertyName);
}

void StringInstance::getOwnPropertyNames(ExecState *exec, PropertyNameArray &propertyNames, PropertyMap::PropertyMode mode)
{
    int size = internalValue()->getString().size();
    for (int i = 0; i < size; i++) {
        propertyNames.add(Identifier(UString::from(i)));
    }

    if (mode == PropertyMap::IncludeDontEnumProperties) {
        propertyNames.add(exec->propertyNames().length);
    }

    return JSObject::getOwnPropertyNames(exec, propertyNames, mode);
}

UString StringInstance::toString(ExecState *exec) const
{
    if (prototype() == originalProto() && !conversionsCustomized() &&
            !static_cast<StringPrototype *>(prototype())->conversionsCustomized()) {
        return internalValue()->value();
    } else {
        return JSObject::toString(exec);
    }
}

JSObject *StringInstance::valueClone(Interpreter *targetCtx) const
{
    return new StringInstance(targetCtx->builtinStringPrototype(), internalValue());
}

// ------------------------------ StringPrototype ---------------------------
const ClassInfo StringPrototype::info = {"String", &StringInstance::info, &stringTable, nullptr};
/* Source for string_object.lut.h
@begin stringTable 26
  toString              StringProtoFunc::ToString       DontEnum|Function       0
  valueOf               StringProtoFunc::ValueOf        DontEnum|Function       0
  charAt                StringProtoFunc::CharAt         DontEnum|Function       1
  charCodeAt            StringProtoFunc::CharCodeAt     DontEnum|Function       1
  concat                StringProtoFunc::Concat         DontEnum|Function       1
  endsWith		StringProtoFunc::EndsWith	DontEnum|Function	1
  includes		StringProtoFunc::Includes	DontEnum|Function	1
  indexOf               StringProtoFunc::IndexOf        DontEnum|Function       1
  lastIndexOf           StringProtoFunc::LastIndexOf    DontEnum|Function       1
  match                 StringProtoFunc::Match          DontEnum|Function       1
  replace               StringProtoFunc::Replace        DontEnum|Function       2
  search                StringProtoFunc::Search         DontEnum|Function       1
  slice                 StringProtoFunc::Slice          DontEnum|Function       2
  split                 StringProtoFunc::Split          DontEnum|Function       2
  startsWith		StringProtoFunc::StartsWith	DontEnum|Function	1
  substr                StringProtoFunc::Substr         DontEnum|Function       2
  substring             StringProtoFunc::Substring      DontEnum|Function       2
  toLowerCase           StringProtoFunc::ToLowerCase    DontEnum|Function       0
  toUpperCase           StringProtoFunc::ToUpperCase    DontEnum|Function       0
  toLocaleLowerCase     StringProtoFunc::ToLocaleLowerCase DontEnum|Function    0
  toLocaleUpperCase     StringProtoFunc::ToLocaleUpperCase DontEnum|Function    0
  trim                  StringProtoFunc::Trim           DontEnum|Function       0
  localeCompare         StringProtoFunc::LocaleCompare  DontEnum|Function       1
  repeat                StringProtoFunc::Repeat         DontEnum|Function       1

#
# Under here: html extension, should only exist if KJS_PURE_ECMA is not defined
# I guess we need to generate two hashtables in the .lut.h file, and use #ifdef
# to select the right one... TODO. #####
  big                   StringProtoFunc::Big            DontEnum|Function       0
  small                 StringProtoFunc::Small          DontEnum|Function       0
  blink                 StringProtoFunc::Blink          DontEnum|Function       0
  bold                  StringProtoFunc::Bold           DontEnum|Function       0
  fixed                 StringProtoFunc::Fixed          DontEnum|Function       0
  italics               StringProtoFunc::Italics        DontEnum|Function       0
  strike                StringProtoFunc::Strike         DontEnum|Function       0
  sub                   StringProtoFunc::Sub            DontEnum|Function       0
  sup                   StringProtoFunc::Sup            DontEnum|Function       0
  fontcolor             StringProtoFunc::Fontcolor      DontEnum|Function       1
  fontsize              StringProtoFunc::Fontsize       DontEnum|Function       1
  anchor                StringProtoFunc::Anchor         DontEnum|Function       1
  link                  StringProtoFunc::Link           DontEnum|Function       1
  trimLeft              StringProtoFunc::TrimLeft       DontEnum|Function       0
  trimRight             StringProtoFunc::TrimRight      DontEnum|Function       0
@end
*/
// ECMA 15.5.4
StringPrototype::StringPrototype(ExecState *exec, ObjectPrototype *objProto)
    : StringInstance(objProto)
{
    // The constructor will be added later, after StringObjectImp has been built
    putDirect(exec->propertyNames().length, jsNumber(0), DontDelete | ReadOnly | DontEnum);
}

bool StringPrototype::getOwnPropertySlot(ExecState *exec, const Identifier &propertyName, PropertySlot &slot)
{
    return getStaticFunctionSlot<StringProtoFunc, StringInstance>(exec, &stringTable, this, propertyName, slot);
}

// ------------------------------ StringProtoFunc ---------------------------

StringProtoFunc::StringProtoFunc(ExecState *exec, int i, int len, const Identifier &name)
    : InternalFunctionImp(static_cast<FunctionPrototype *>(exec->lexicalInterpreter()->builtinFunctionPrototype()), name)
    , id(i)
{
    putDirect(exec->propertyNames().length, len, DontDelete | ReadOnly | DontEnum);
}

static inline void expandSourceRanges(UString::Range *&array, int &count, int &capacity)
{
    int newCapacity;
    if (capacity == 0) {
        newCapacity = 16;
    } else {
        newCapacity = capacity * 2;
    }

    UString::Range *newArray = new UString::Range[newCapacity];
    for (int i = 0; i < count; i++) {
        newArray[i] = array[i];
    }

    delete [] array;

    capacity = newCapacity;
    array = newArray;
}

static void pushSourceRange(UString::Range *&array, int &count, int &capacity, UString::Range range)
{
    if (count + 1 > capacity) {
        expandSourceRanges(array, count, capacity);
    }

    array[count] = range;
    count++;
}

static inline void expandReplacements(UString *&array, int &count, int &capacity)
{
    int newCapacity;
    if (capacity == 0) {
        newCapacity = 16;
    } else {
        newCapacity = capacity * 2;
    }

    UString *newArray = new UString[newCapacity];
    for (int i = 0; i < count; i++) {
        newArray[i] = array[i];
    }

    delete [] array;

    capacity = newCapacity;
    array = newArray;
}

static void pushReplacement(UString *&array, int &count, int &capacity, UString replacement)
{
    if (count + 1 > capacity) {
        expandReplacements(array, count, capacity);
    }

    array[count] = replacement;
    count++;
}

static inline UString substituteBackreferences(const UString &replacement, const UString &source, int *ovector, RegExp *reg)
{
    UString substitutedReplacement = replacement;

    int i = -1;
    while ((i = substitutedReplacement.find(UString("$"), i + 1)) != -1) {
        if (i + 1 == substitutedReplacement.size()) {
            break;
        }

        unsigned short ref = substitutedReplacement[i + 1].unicode();
        int backrefStart = 0;
        int backrefLength = 0;
        int advance = 0;

        if (ref == '$') {  // "$$" -> "$"
            substitutedReplacement = substitutedReplacement.substr(0, i + 1) + substitutedReplacement.substr(i + 2);
            continue;
        } else if (ref == '&') {
            backrefStart = ovector[0];
            backrefLength = ovector[1] - backrefStart;
        } else if (ref == '`') {
            backrefStart = 0;
            backrefLength = ovector[0];
        } else if (ref == '\'') {
            backrefStart = ovector[1];
            backrefLength = source.size() - backrefStart;
        } else if (ref >= '0' && ref <= '9') {
            // 1- and 2-digit back references are allowed
            unsigned backrefIndex = ref - '0';
            if (backrefIndex > (unsigned)reg->subPatterns()) {
                continue;
            }
            if (substitutedReplacement.size() > i + 2) {
                ref = substitutedReplacement[i + 2].unicode();
                if (ref >= '0' && ref <= '9') {
                    backrefIndex = 10 * backrefIndex + ref - '0';
                    if (backrefIndex > (unsigned)reg->subPatterns()) {
                        backrefIndex = backrefIndex / 10;    // Fall back to the 1-digit reference
                    } else {
                        advance = 1;
                    }
                }
            }
            backrefStart = ovector[2 * backrefIndex];
            backrefLength = ovector[2 * backrefIndex + 1] - backrefStart;
        } else {
            continue;
        }

        substitutedReplacement = substitutedReplacement.substr(0, i) + source.substr(backrefStart, backrefLength) + substitutedReplacement.substr(i + 2 + advance);
        i += backrefLength - 1; // - 1 offsets 'i + 1'
    }

    return substitutedReplacement;
}

static inline int localeCompare(const UString &a, const UString &b)
{
#if PLATFORM(WIN_OS)
    int retval = CompareStringW(LOCALE_USER_DEFAULT, 0,
                                reinterpret_cast<LPCWSTR>(a.data()), a.size(),
                                reinterpret_cast<LPCWSTR>(b.data()), b.size());
    return !retval ? retval : retval - 2;
#elif PLATFORM(CF)
    CFStringRef sa = CFStringCreateWithCharactersNoCopy(kCFAllocatorDefault, reinterpret_cast<const UniChar *>(a.data()), a.size(), kCFAllocatorNull);
    CFStringRef sb = CFStringCreateWithCharactersNoCopy(kCFAllocatorDefault, reinterpret_cast<const UniChar *>(b.data()), b.size(), kCFAllocatorNull);

    int retval = CFStringCompare(sa, sb, kCFCompareLocalized);

    CFRelease(sa);
    CFRelease(sb);

    return retval;
#else
    // ### use as fallback only. implement locale aware version.
    // ### other browsers have more detailed return values than -1, 0 and 1
    return compare(a, b);
#endif
}

static JSValue *replace(ExecState *exec, const UString &source, JSValue *pattern, JSValue *replacement)
{
    JSObject *replacementFunction = nullptr;
    UString replacementString;

    if (JSValue::isObject(replacement) && JSValue::toObject(replacement, exec)->implementsCall()) {
        replacementFunction = JSValue::toObject(replacement, exec);
    } else {
        replacementString = JSValue::toString(replacement, exec);
    }

    if (JSValue::isObject(pattern) && static_cast<JSObject *>(pattern)->inherits(&RegExpImp::info)) {
        RegExp *reg = static_cast<RegExpImp *>(pattern)->regExp();
        bool global = reg->flags() & RegExp::Global;

        RegExpObjectImp *regExpObj = static_cast<RegExpObjectImp *>(exec->lexicalInterpreter()->builtinRegExp());

        int matchIndex = 0;
        int lastIndex = 0;
        int startPosition = 0;

        UString::Range *sourceRanges = nullptr;
        int sourceRangeCount = 0;
        int sourceRangeCapacity = 0;
        UString *replacements = nullptr;
        int replacementCount = 0;
        int replacementCapacity = 0;

        // This is either a loop (if global is set) or a one-way (if not).
        RegExpStringContext ctx(source);
        do {
            int *ovector;
            UString matchString = regExpObj->performMatch(reg, exec, ctx, source, startPosition, &matchIndex, &ovector);
            if (matchIndex == -1) {
                break;
            }
            int matchLen = matchString.size();

            pushSourceRange(sourceRanges, sourceRangeCount, sourceRangeCapacity, UString::Range(lastIndex, matchIndex - lastIndex));
            UString substitutedReplacement;
            if (replacementFunction) {
                int completeMatchStart = ovector[0];
                List args;

                args.append(jsString(matchString));

                for (unsigned i = 0; i < reg->subPatterns(); i++) {
                    int matchStart = ovector[(i + 1) * 2];
                    int matchLen = ovector[(i + 1) * 2 + 1] - matchStart;

                    args.append(jsString(source.substr(matchStart, matchLen)));
                }

                args.append(jsNumber(completeMatchStart));
                args.append(jsString(source));

                substitutedReplacement = JSValue::toString(replacementFunction->call(exec, exec->dynamicInterpreter()->globalObject(), args), exec);
            } else {
                substitutedReplacement = substituteBackreferences(replacementString, source, ovector, reg);
            }
            pushReplacement(replacements, replacementCount, replacementCapacity, substitutedReplacement);

            lastIndex = matchIndex + matchLen;
            startPosition = lastIndex;

            // special case of empty match
            if (matchLen == 0) {
                startPosition++;
                if (startPosition > source.size()) {
                    break;
                }
            }
        } while (global);

        if (lastIndex < source.size()) {
            pushSourceRange(sourceRanges, sourceRangeCount, sourceRangeCapacity, UString::Range(lastIndex, source.size() - lastIndex));
        }

        UString result;
        if (sourceRanges) {
            result = source.spliceSubstringsWithSeparators(sourceRanges, sourceRangeCount, replacements, replacementCount);
        }

        delete [] sourceRanges;
        delete [] replacements;

        return jsString(result);
    }

    // First arg is a string
    UString patternString = JSValue::toString(pattern, exec);
    int matchPos = source.find(patternString);
    int matchLen = patternString.size();
    // Do the replacement
    if (matchPos == -1) {
        return jsString(source);
    }

    if (replacementFunction) {
        List args;

        args.append(jsString(source.substr(matchPos, matchLen)));
        args.append(jsNumber(matchPos));
        args.append(jsString(source));

        replacementString = JSValue::toString(replacementFunction->call(exec, exec->dynamicInterpreter()->globalObject(), args), exec);
    }

    return jsString(source.substr(0, matchPos) + replacementString + source.substr(matchPos + matchLen));
}

static UnicodeSupport::StringConversionFunction toUpperF = Unicode::toUpper;
static UnicodeSupport::StringConversionFunction toLowerF = Unicode::toLower;

void StringProtoFunc::setToLowerFunction(UnicodeSupport::StringConversionFunction f)
{
    toLowerF = f;
}

void StringProtoFunc::setToUpperFunction(UnicodeSupport::StringConversionFunction f)
{
    toUpperF = f;
}

// ECMA 15.5.4.2 - 15.5.4.20
JSValue *StringProtoFunc::callAsFunction(ExecState *exec, JSObject *thisObj, const List &args)
{
    JSValue *result = nullptr;

    // toString and valueOf are no generic function.
    if (id == ToString || id == ValueOf) {
        if (!thisObj || !thisObj->inherits(&StringInstance::info)) {
            return throwError(exec, TypeError);
        }

        return jsString(static_cast<StringInstance *>(thisObj)->internalValue()->toString(exec));
    }

    UString u, u2, u3;
    int pos, p0, i;
    double dpos;
    double d = 0.0;

    UString s = thisObj->toString(exec);

    int len = s.size();
    JSValue *a0 = args[0];
    JSValue *a1 = args[1];

    switch (id) {
    case ToString:
    case ValueOf:
        // handled above
        break;
    case CharAt:
        // Other browsers treat an omitted parameter as 0 rather than NaN.
        // That doesn't match the ECMA standard, but is needed for site compatibility.
        dpos = JSValue::isUndefined(a0) ? 0 : JSValue::toInteger(a0, exec);
        if (dpos >= 0 && dpos < len) { // false for NaN
            u = s.substr(static_cast<int>(dpos), 1);
        } else {
            u = "";
        }
        result = jsString(u);
        break;
    case CharCodeAt:
        // Other browsers treat an omitted parameter as 0 rather than NaN.
        // That doesn't match the ECMA standard, but is needed for site compatibility.
        dpos = JSValue::isUndefined(a0) ? 0 : JSValue::toInteger(a0, exec);
        if (dpos >= 0 && dpos < len) { // false for NaN
            result = jsNumber(s[static_cast<int>(dpos)].unicode());
        } else {
            result = jsNaN();
        }
        break;
    case Concat: {
        ListIterator it = args.begin();
        for (; it != args.end(); ++it) {
            s += JSValue::toString(*it, exec);
        }
        result = jsString(s);
        break;
    }
    case EndsWith:
	if (JSValue::isObject(a0) && static_cast<JSObject*>(a0)->inherits(&RegExpImp::info)) {
	    return throwError(exec, TypeError, "RegExp reserved for future");
	}
	u2 = JSValue::toString(a0, exec);
	if (JSValue::isUndefined(a1))
	    i = s.size();
	else
	    i = minInt(s.size(), JSValue::toInteger(a1, exec));
	pos = i - u2.size();
	result = jsBoolean(pos >= 0 && s.substr(pos, u2.size()) == u2);
	break;
    case Includes:
	if (JSValue::isObject(a0) && static_cast<JSObject*>(a0)->inherits(&RegExpImp::info)) {
	    return throwError(exec, TypeError, "RegExp reserved for future");
	}
	u2 = JSValue::toString(a0, exec);
	pos = JSValue::toInteger(a1, exec);
	i = s.find(u2, pos);
	result = jsBoolean(i >= 0);
	break;
    case IndexOf:
        u2 = JSValue::toString(a0, exec);
        if (JSValue::isUndefined(a1)) {
            dpos = 0;
        } else {
            dpos = JSValue::toInteger(a1, exec);
            if (dpos >= 0) { // false for NaN
                if (dpos > len) {
                    dpos = len;
                }
            } else {
                dpos = 0;
            }
        }
        result = jsNumber(s.find(u2, static_cast<int>(dpos)));
        break;
    case LastIndexOf:
        u2 = JSValue::toString(a0, exec);
        d = JSValue::toNumber(a1, exec);
        if (JSValue::isUndefined(a1) || KJS::isNaN(d)) {
            dpos = len;
        } else {
            dpos = JSValue::toInteger(a1, exec);
            if (dpos >= 0) { // false for NaN
                if (dpos > len) {
                    dpos = len;
                }
            } else {
                dpos = 0;
            }
        }
        result = jsNumber(s.rfind(u2, static_cast<int>(dpos)));
        break;
    case Match:
    case Search: {
        u = s;
        RegExp *reg, *tmpReg = nullptr;
        RegExpImp *imp = nullptr;
        if (JSValue::isObject(a0) && static_cast<JSObject *>(a0)->inherits(&RegExpImp::info)) {
            reg = static_cast<RegExpImp *>(a0)->regExp();
        } else {
            /*
             *  ECMA 15.5.4.12 String.prototype.search (regexp)
             *  If regexp is not an object whose [[Class]] property is "RegExp", it is
             *  replaced with the result of the expression new RegExp(regexp).
             */
            reg = tmpReg = new RegExp(JSValue::toString(a0, exec), RegExp::None);
        }
        if (!reg->isValid()) {
            delete tmpReg;
            return throwError(exec, SyntaxError, "Invalid regular expression");
        }
        RegExpObjectImp *regExpObj = static_cast<RegExpObjectImp *>(exec->lexicalInterpreter()->builtinRegExp());

        RegExpStringContext ctx(u);
        UString mstr = regExpObj->performMatch(reg, exec, ctx, u, 0, &pos);

        if (id == Search) {
            result = jsNumber(pos);
        } else {
            // Exec
            if ((reg->flags() & RegExp::Global) == 0) {
                // case without 'g' flag is handled like RegExp.prototype.exec
                if (mstr.isNull()) {
                    result = jsNull();
                } else {
                    result = regExpObj->arrayOfMatches(exec, mstr);
                }
            } else {
                // return array of matches
                List list;
                int lastIndex = 0;
                while (pos >= 0) {
                    if (mstr.isNull()) {
                        list.append(jsUndefined());
                    } else {
                        list.append(jsString(mstr));
                    }
                    lastIndex = pos;
                    pos += mstr.isEmpty() ? 1 : mstr.size();
                    mstr = regExpObj->performMatch(reg, exec, ctx, u, pos, &pos);
                }
                if (imp) {
                    imp->put(exec, "lastIndex", jsNumber(lastIndex), DontDelete | DontEnum);
                }
                if (list.isEmpty()) {
                    // if there are no matches at all, it's important to return
                    // Null instead of an empty array, because this matches
                    // other browsers and because Null is a false value.
                    result = jsNull();
                } else {
                    result = exec->lexicalInterpreter()->builtinArray()->construct(exec, list);
                }
            }
        }

        delete tmpReg;
        break;
    }
    case Replace:
        result = replace(exec, s, a0, a1);
        break;
    case Slice: {
        // The arg processing is very much like ArrayProtoFunc::Slice
        double start = JSValue::toInteger(a0, exec);
        double end = JSValue::isUndefined(a1) ? len : JSValue::toInteger(a1, exec);
        double from = start < 0 ? len + start : start;
        double to = end < 0 ? len + end : end;
        if (to > from && to > 0 && from < len) {
            if (from < 0) {
                from = 0;
            }
            if (to > len) {
                to = len;
            }
            result = jsString(s.substr(static_cast<int>(from), static_cast<int>(to - from)));
        } else {
            result = jsString("");
        }
        break;
    }
    case Split: {
        JSObject *constructor = exec->lexicalInterpreter()->builtinArray();
        JSObject *res = static_cast<JSObject *>(constructor->construct(exec, List::empty()));
        result = res;
        u = s;
        i = p0 = 0;
        uint32_t limit = JSValue::isUndefined(a1) ? 0xFFFFFFFFU : JSValue::toUInt32(a1, exec);
        if (JSValue::isObject(a0) && static_cast<JSObject *>(a0)->inherits(&RegExpImp::info)) {
            RegExp *reg = static_cast<RegExpImp *>(a0)->regExp();

            RegExpStringContext ctx(u);
            bool error = false;
            if (u.isEmpty() && !reg->match(ctx, u, &error, 0).isNull()) {

                // empty string matched by regexp -> empty array
                res->put(exec, exec->propertyNames().length, jsNumber(0));
                break;
            }
            pos = 0;
            while (!error && static_cast<uint32_t>(i) != limit && pos < u.size()) {
                // TODO: back references
                int mpos;
                int *ovector = nullptr;
                UString mstr = reg->match(ctx, u, &error, pos, &mpos, &ovector);
                delete [] ovector; ovector = nullptr;
                if (mpos < 0) {
                    break;
                }
                pos = mpos + (mstr.isEmpty() ? 1 : mstr.size());
                if (mpos != p0 || !mstr.isEmpty()) {
                    res->put(exec, i, jsString(u.substr(p0, mpos - p0)));
                    p0 = mpos + mstr.size();
                    i++;
                }
            }

            if (error) {
                RegExpObjectImp::throwRegExpError(exec);
            }

        } else {
            u2 = JSValue::toString(a0, exec);
            if (u2.isEmpty()) {
                if (u.isEmpty()) {
                    // empty separator matches empty string -> empty array
                    put(exec, exec->propertyNames().length, jsNumber(0));
                    break;
                } else {
                    while (static_cast<uint32_t>(i) != limit && i < u.size() - 1) {
                        res->put(exec, i++, jsString(u.substr(p0++, 1)));
                    }
                }
            } else {
                while (static_cast<uint32_t>(i) != limit && (pos = u.find(u2, p0)) >= 0) {
                    res->put(exec, i, jsString(u.substr(p0, pos - p0)));
                    p0 = pos + u2.size();
                    i++;
                }
            }
        }
        // add remaining string, if any
        if (static_cast<uint32_t>(i) != limit) {
            res->put(exec, i++, jsString(u.substr(p0)));
        }
        res->put(exec, exec->propertyNames().length, jsNumber(i));
    }
    break;
    case StartsWith:
	if (JSValue::isObject(a0) && static_cast<JSObject*>(a0)->inherits(&RegExpImp::info)) {
	    return throwError(exec, TypeError, "RegExp reserved for future");
	}
	u2 = JSValue::toString(a0, exec);
	pos = maxInt(0, JSValue::toInteger(a1, exec));
	result = jsBoolean(s.substr(pos, u2.size()) == u2);
	break;
    case Substr: { //B.2.3
        // Note: NaN is effectively handled as 0 here for both length
        // and start, hence toInteger does fine, and removes worries
        // about weird comparison results below.
        int len = s.size();
        double start  = JSValue::toInteger(a0, exec);
        double length = JSValue::isUndefined(a1) ? len : JSValue::toInteger(a1, exec);

        if (start >= len) {
            return jsString("");
        }
        if (length < 0) {
            return jsString("");
        }
        if (start < 0) {
            start += s.size();
            if (start < 0) {
                start = 0;
            }
        }

        if (length > len - start) {
            length = len - start;
        }

        result = jsString(s.substr(static_cast<int>(start), static_cast<int>(length)));
        break;
    }
    case Substring: {
        double start = JSValue::toNumber(a0, exec);
        double end = JSValue::toNumber(a1, exec);
        if (isNaN(start)) {
            start = 0;
        }
        if (isNaN(end)) {
            end = 0;
        }
        if (start < 0) {
            start = 0;
        }
        if (end < 0) {
            end = 0;
        }
        if (start > len) {
            start = len;
        }
        if (end > len) {
            end = len;
        }
        if (JSValue::isUndefined(a1)) {
            end = len;
        }
        if (start > end) {
            double temp = end;
            end = start;
            start = temp;
        }
        result = jsString(s.substr((int)start, (int)end - (int)start));
    }
    break;
    case ToLowerCase:
    case ToLocaleLowerCase: { // FIXME: See http://www.unicode.org/Public/UNIDATA/SpecialCasing.txt for locale-sensitive mappings that aren't implemented.
        u = s;
        u.copyForWriting();
        uint16_t *dataPtr = reinterpret_cast<uint16_t *>(u.rep()->data());
        uint16_t *destIfNeeded;

        int len = toLowerF(dataPtr, u.size(), destIfNeeded);
        if (len >= 0) {
            result = jsString(UString(reinterpret_cast<UChar *>(destIfNeeded ? destIfNeeded : dataPtr), len));
        } else {
            result = jsString(s);
        }

        free(destIfNeeded);
        break;
    }
    case ToUpperCase:
    case ToLocaleUpperCase: { // FIXME: See http://www.unicode.org/Public/UNIDATA/SpecialCasing.txt for locale-sensitive mappings that aren't implemented.
        u = s;
        u.copyForWriting();
        uint16_t *dataPtr = reinterpret_cast<uint16_t *>(u.rep()->data());
        uint16_t *destIfNeeded;

        int len = toUpperF(dataPtr, u.size(), destIfNeeded);
        if (len >= 0) {
            result = jsString(UString(reinterpret_cast<UChar *>(destIfNeeded ? destIfNeeded : dataPtr), len));
        } else {
            result = jsString(s);
        }

        free(destIfNeeded);
        break;
    }
    case LocaleCompare:
        if (args.size() < 1) {
            return jsNumber(0);
        }
        return jsNumber(localeCompare(s, JSValue::toString(a0, exec)));
    case Repeat: {
        double n = JSValue::toInteger(args[0], exec);
        if (exec->hadException())
            return jsUndefined();
        if (n < 0 || KJS::isPosInf(n))
            return throwError(exec, RangeError, "Invalid repeat count");

        UString ret;
        for (int i = 0; i < n; ++i)
        {
            ret += s;
        }
        return jsString(ret);
    }
    case Trim:
    case TrimRight:
    case TrimLeft: {
        const uint16_t *dataPtr = reinterpret_cast<uint16_t *>(s.rep()->data());

        const int size = s.size();
        int left = 0;
        if (id != TrimRight)
            while (left < size && CommonUnicode::isStrWhiteSpace(dataPtr[left])) {
                left++;
            }

        int right = size;
        if (id != TrimLeft)
            while (right > left && CommonUnicode::isStrWhiteSpace(dataPtr[right - 1])) {
                right--;
            }

        result = jsString(s.substr(left, right - left));
        break;
    }
#ifndef KJS_PURE_ECMA
    case Big:
        result = jsString("<big>" + s + "</big>");
        break;
    case Small:
        result = jsString("<small>" + s + "</small>");
        break;
    case Blink:
        result = jsString("<blink>" + s + "</blink>");
        break;
    case Bold:
        result = jsString("<b>" + s + "</b>");
        break;
    case Fixed:
        result = jsString("<tt>" + s + "</tt>");
        break;
    case Italics:
        result = jsString("<i>" + s + "</i>");
        break;
    case Strike:
        result = jsString("<strike>" + s + "</strike>");
        break;
    case Sub:
        result = jsString("<sub>" + s + "</sub>");
        break;
    case Sup:
        result = jsString("<sup>" + s + "</sup>");
        break;
    case Fontcolor:
        result = jsString("<font color=\"" + JSValue::toString(a0, exec) + "\">" + s + "</font>");
        break;
    case Fontsize:
        result = jsString("<font size=\"" + JSValue::toString(a0, exec) + "\">" + s + "</font>");
        break;
    case Anchor:
        result = jsString("<a name=\"" + JSValue::toString(a0, exec) + "\">" + s + "</a>");
        break;
    case Link:
        result = jsString("<a href=\"" + JSValue::toString(a0, exec) + "\">" + s + "</a>");
        break;
#endif
    }

    return result;
}

// ------------------------------ StringObjectImp ------------------------------

StringObjectImp::StringObjectImp(ExecState *exec,
                                 FunctionPrototype *funcProto,
                                 StringPrototype *stringProto)
    : InternalFunctionImp(funcProto)
{
    // ECMA 15.5.3.1 String.prototype
    putDirect(exec->propertyNames().prototype, stringProto, DontEnum | DontDelete | ReadOnly);

    putDirectFunction(new StringObjectFuncImp(exec, funcProto, exec->propertyNames().fromCharCode), DontEnum);

    // no. of arguments for constructor
    putDirect(exec->propertyNames().length, jsNumber(1), ReadOnly | DontDelete | DontEnum);
}

bool StringObjectImp::implementsConstruct() const
{
    return true;
}

// ECMA 15.5.2
JSObject *StringObjectImp::construct(ExecState *exec, const List &args)
{
    JSObject *proto = exec->lexicalInterpreter()->builtinStringPrototype();
    if (args.size() == 0) {
        return new StringInstance(proto);
    }
    return new StringInstance(proto, JSValue::toString(*args.begin(), exec));
}

// ECMA 15.5.1
JSValue *StringObjectImp::callAsFunction(ExecState *exec, JSObject * /*thisObj*/, const List &args)
{
    if (args.isEmpty()) {
        return jsString("");
    } else {
        JSValue *v = args[0];
        return jsString(JSValue::toString(v, exec));
    }
}

// ------------------------------ StringObjectFuncImp --------------------------

// ECMA 15.5.3.2 fromCharCode()
StringObjectFuncImp::StringObjectFuncImp(ExecState *exec, FunctionPrototype *funcProto, const Identifier &name)
    : InternalFunctionImp(funcProto, name)
{
    putDirect(exec->propertyNames().length, jsNumber(1), DontDelete | ReadOnly | DontEnum);
}

JSValue *StringObjectFuncImp::callAsFunction(ExecState *exec, JSObject * /*thisObj*/, const List &args)
{
    UString s;
    if (args.size()) {
        UChar *buf = static_cast<UChar *>(fastMalloc(args.size() * sizeof(UChar)));
        UChar *p = buf;
        ListIterator it = args.begin();
        while (it != args.end()) {
            unsigned short u = JSValue::toUInt16(*it, exec);
            *p++ = UChar(u);
            it++;
        }
        s = UString(buf, args.size(), false);
    } else {
        s = "";
    }

    return jsString(s);
}

}