C++ RegExpConstructor类代码示例

// Shared implementation used by test and exec.
bool RegExpObject::match(ExecState* exec, const ArgList& args)
{
    RegExpConstructor* regExpConstructor = exec->lexicalGlobalObject()->regExpConstructor();

    UString input = args.isEmpty() ? regExpConstructor->input() : args.at(0).toString(exec);
    if (input.isNull()) {
        throwError(exec, GeneralError, "No input to " + toString(exec) + ".");
        return false;
    }

    if (!regExp()->global()) {
        int position;
        int length;
        regExpConstructor->performMatch(d->regExp.get(), input, 0, position, length);
        return position >= 0;
    }

    if (d->lastIndex < 0 || d->lastIndex > input.size()) {
        d->lastIndex = 0;
        return false;
    }

    int position;
    int length = 0;
    regExpConstructor->performMatch(d->regExp.get(), input, static_cast<int>(d->lastIndex), position, length);
    if (position < 0) {
        d->lastIndex = 0;
        return false;
    }

    d->lastIndex = position + length;
    return true;
}

// Shared implementation used by test and exec.
bool RegExpObject::match(ExecState* exec)
{
    RegExpConstructor* regExpConstructor = exec->lexicalGlobalObject()->regExpConstructor();

    UString input = !exec->argumentCount() ? regExpConstructor->input() : exec->argument(0).toString(exec);
    if (input.isNull()) {
        throwError(exec, createError(exec, makeString("No input to ", toString(exec), ".")));
        return false;
    }

    if (!regExp()->global()) {
        int position;
        int length;
        regExpConstructor->performMatch(d->regExp.get(), input, 0, position, length);
        return position >= 0;
    }

    if (d->lastIndex < 0 || d->lastIndex > input.length()) {
        d->lastIndex = 0;
        return false;
    }

    int position;
    int length = 0;
    regExpConstructor->performMatch(d->regExp.get(), input, static_cast<int>(d->lastIndex), position, length);
    if (position < 0) {
        d->lastIndex = 0;
        return false;
    }

    d->lastIndex = position + length;
    return true;
}

void RegExpConstructor::visitChildren(JSCell* cell, SlotVisitor& visitor)
{
    RegExpConstructor* thisObject = jsCast<RegExpConstructor*>(cell);
    ASSERT_GC_OBJECT_INHERITS(thisObject, &s_info);
    COMPILE_ASSERT(StructureFlags & OverridesVisitChildren, OverridesVisitChildrenWithoutSettingFlag);
    ASSERT(thisObject->structure()->typeInfo().overridesVisitChildren());

    Base::visitChildren(thisObject, visitor);
    thisObject->m_cachedResult.visitChildren(visitor);
}

JSValue JSC_HOST_CALL stringProtoFuncMatch(ExecState* exec, JSObject*, JSValue thisValue, const ArgList& args)
{
    UString s = thisValue.toThisString(exec);

    JSValue a0 = args.at(0);

    UString u = s;
    RefPtr<RegExp> reg;
    RegExpObject* imp = 0;
    if (a0.isObject(&RegExpObject::info))
        reg = asRegExpObject(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 = RegExp::create(&exec->globalData(), a0.toString(exec));
    }
    RegExpConstructor* regExpConstructor = exec->lexicalGlobalObject()->regExpConstructor();
    int pos;
    int matchLength;
    regExpConstructor->performMatch(reg.get(), u, 0, pos, matchLength);
    if (!(reg->global())) {
        // case without 'g' flag is handled like RegExp.prototype.exec
        if (pos < 0)
            return jsNull();
        return regExpConstructor->arrayOfMatches(exec);
    }

    // return array of matches
    MarkedArgumentBuffer list;
    int lastIndex = 0;
    while (pos >= 0) {
        list.append(jsSubstring(exec, u, pos, matchLength));
        lastIndex = pos;
        pos += matchLength == 0 ? 1 : matchLength;
        regExpConstructor->performMatch(reg.get(), u, pos, pos, matchLength);
    }
    if (imp)
        imp->setLastIndex(lastIndex);
    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.
        return jsNull();
    }

    return constructArray(exec, list);
}

EncodedJSValue JSC_HOST_CALL regExpProtoFuncSearch(ExecState* exec)
{
    JSValue thisValue = exec->thisValue();
    if (!thisValue.inherits(RegExpObject::info()))
        return throwVMTypeError(exec);
    RegExp* regExp = asRegExpObject(thisValue)->regExp();

    JSString* string = exec->argument(0).toString(exec);
    String s = string->value(exec);
    if (exec->hadException())
        return JSValue::encode(jsUndefined());

    RegExpConstructor* regExpConstructor = exec->lexicalGlobalObject()->regExpConstructor();
    MatchResult result = regExpConstructor->performMatch(exec->vm(), regExp, string, s, 0);
    return JSValue::encode(result ? jsNumber(result.start) : jsNumber(-1));
}

// Shared implementation used by test and exec.
bool RegExpObject::match(ExecState* exec)
{
    RegExpConstructor* regExpConstructor = exec->lexicalGlobalObject()->regExpConstructor();
    UString input = exec->argument(0).toString(exec);
    JSGlobalData* globalData = &exec->globalData();
    if (!regExp()->global()) {
        int position;
        int length;
        regExpConstructor->performMatch(*globalData, d->regExp.get(), input, 0, position, length);
        return position >= 0;
    }

    JSValue jsLastIndex = getLastIndex();
    unsigned lastIndex;
    if (LIKELY(jsLastIndex.isUInt32())) {
        lastIndex = jsLastIndex.asUInt32();
        if (lastIndex > input.length()) {
            setLastIndex(0);
            return false;
        }
    } else {
        double doubleLastIndex = jsLastIndex.toInteger(exec);
        if (doubleLastIndex < 0 || doubleLastIndex > input.length()) {
            setLastIndex(0);
            return false;
        }
        lastIndex = static_cast<unsigned>(doubleLastIndex);
    }

    int position;
    int length = 0;
    regExpConstructor->performMatch(*globalData, d->regExp.get(), input, lastIndex, position, length);
    if (position < 0) {
        setLastIndex(0);
        return false;
    }

    setLastIndex(position + length);
    return true;
}

JSValue JSC_HOST_CALL stringProtoFuncSearch(ExecState* exec, JSObject*, JSValue thisValue, const ArgList& args)
{
    UString s = thisValue.toThisString(exec);

    JSValue a0 = args.at(0);

    UString u = s;
    RefPtr<RegExp> reg;
    if (a0.isObject(&RegExpObject::info))
        reg = asRegExpObject(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 = RegExp::create(&exec->globalData(), a0.toString(exec));
    }
    RegExpConstructor* regExpConstructor = exec->lexicalGlobalObject()->regExpConstructor();
    int pos;
    int matchLength;
    regExpConstructor->performMatch(reg.get(), u, 0, pos, matchLength);
    return jsNumber(exec, pos);
}

JSValue* stringProtoFuncSearch(ExecState* exec, JSObject*, JSValue* thisValue, const ArgList& args)
{
    UString s = thisValue->toThisString(exec);

    JSValue* a0 = args.at(exec, 0);

    UString u = s;
    RefPtr<RegExp> reg;
    if (a0->isObject() && static_cast<JSObject*>(a0)->inherits(&RegExpObject::info))
        reg = static_cast<RegExpObject*>(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 = RegExp::create(exec, a0->toString(exec));
    }
    RegExpConstructor* regExpObj = exec->lexicalGlobalObject()->regExpConstructor();
    int pos;
    int matchLength;
    regExpObj->performMatch(reg.get(), u, 0, pos, matchLength);
    return jsNumber(exec, pos);
}

bool RegExpObject::match(ExecState* exec, const ArgList& args)
{
    RegExpConstructor* regExpObj = exec->lexicalGlobalObject()->regExpConstructor();

    UString input;
    if (!args.isEmpty())
        input = args.at(exec, 0)->toString(exec);
    else {
        input = regExpObj->input();
        if (input.isNull()) {
            throwError(exec, GeneralError, "No input.");
            return false;
        }
    }

    bool global = get(exec, exec->propertyNames().global)->toBoolean(exec);
    int lastIndex = 0;
    if (global) {
        if (d->lastIndex < 0 || d->lastIndex > input.size()) {
            d->lastIndex = 0;
            return false;
        }
        lastIndex = static_cast<int>(d->lastIndex);
    }

    int foundIndex;
    int foundLength;
    regExpObj->performMatch(d->regExp.get(), input, lastIndex, foundIndex, foundLength);

    if (global) {
        lastIndex = foundIndex < 0 ? 0 : foundIndex + foundLength;
        d->lastIndex = lastIndex;
    }

    return foundIndex >= 0;
}

JSValue JSC_HOST_CALL stringProtoFuncReplace(ExecState* exec, JSObject*, JSValue thisValue, const ArgList& args)
{
    JSString* sourceVal = thisValue.toThisJSString(exec);
    const UString& source = sourceVal->value();

    JSValue pattern = args.at(0);

    JSValue replacement = args.at(1);
    UString replacementString;
    CallData callData;
    CallType callType = replacement.getCallData(callData);
    if (callType == CallTypeNone)
        replacementString = replacement.toString(exec);

    if (pattern.isObject(&RegExpObject::info)) {
        RegExp* reg = asRegExpObject(pattern)->regExp();
        bool global = reg->global();

        RegExpConstructor* regExpConstructor = exec->lexicalGlobalObject()->regExpConstructor();

        int lastIndex = 0;
        int startPosition = 0;

        Vector<UString::Range, 16> sourceRanges;
        Vector<UString, 16> replacements;

        // This is either a loop (if global is set) or a one-way (if not).
        if (global && callType == CallTypeJS) {
            // reg->numSubpatterns() + 1 for pattern args, + 2 for match start and sourceValue
            int argCount = reg->numSubpatterns() + 1 + 2;
            JSFunction* func = asFunction(replacement);
            CachedCall cachedCall(exec, func, argCount, exec->exceptionSlot());
            if (exec->hadException())
                return jsNull();
            while (true) {
                int matchIndex;
                int matchLen;
                int* ovector;
                regExpConstructor->performMatch(reg, source, startPosition, matchIndex, matchLen, &ovector);
                if (matchIndex < 0)
                    break;
                
                sourceRanges.append(UString::Range(lastIndex, matchIndex - lastIndex));

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

                    if (matchStart < 0)
                        cachedCall.setArgument(i, jsUndefined());
                    else
                        cachedCall.setArgument(i, jsSubstring(exec, source, matchStart, matchLen));
                }

                cachedCall.setArgument(i++, jsNumber(exec, completeMatchStart));
                cachedCall.setArgument(i++, sourceVal);
                
                cachedCall.setThis(exec->globalThisValue());
                replacements.append(cachedCall.call().toString(cachedCall.newCallFrame()));
                if (exec->hadException())
                    break;

                lastIndex = matchIndex + matchLen;
                startPosition = lastIndex;

                // special case of empty match
                if (matchLen == 0) {
                    startPosition++;
                    if (startPosition > source.size())
                        break;
                }
            }            
        } else {
            do {
                int matchIndex;
                int matchLen;
                int* ovector;
                regExpConstructor->performMatch(reg, source, startPosition, matchIndex, matchLen, &ovector);
                if (matchIndex < 0)
                    break;

                sourceRanges.append(UString::Range(lastIndex, matchIndex - lastIndex));

                if (callType != CallTypeNone) {
                    int completeMatchStart = ovector[0];
                    MarkedArgumentBuffer args;

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

                        if (matchStart < 0)
                            args.append(jsUndefined());
                        else
                            args.append(jsSubstring(exec, source, matchStart, matchLen));
                    }

                    args.append(jsNumber(exec, completeMatchStart));
//.........这里部分代码省略.........

JSValue RegExpObject::matchGlobal(ExecState* exec, JSGlobalObject* globalObject, JSString* string)
{
    RegExp* regExp = this->regExp();

    ASSERT(regExp->global());

    VM* vm = &globalObject->vm();

    setLastIndex(exec, 0);
    if (exec->hadException())
        return jsUndefined();

    String s = string->value(exec);
    RegExpConstructor* regExpConstructor = globalObject->regExpConstructor();
    MatchResult result = regExpConstructor->performMatch(*vm, regExp, string, s, 0);

    // return array of matches
    MarkedArgumentBuffer list;
    // We defend ourselves from crazy.
    const size_t maximumReasonableMatchSize = 1000000000;

    if (regExp->unicode()) {
        unsigned stringLength = s.length();
        while (result) {
            if (list.size() > maximumReasonableMatchSize) {
                throwOutOfMemoryError(exec);
                return jsUndefined();
            }

            size_t end = result.end;
            size_t length = end - result.start;
            list.append(jsSubstring(exec, s, result.start, length));
            if (!length)
                end = advanceStringUnicode(s, stringLength, end);
            result = regExpConstructor->performMatch(*vm, regExp, string, s, end);
        }
    } else {
        while (result) {
            if (list.size() > maximumReasonableMatchSize) {
                throwOutOfMemoryError(exec);
                return jsUndefined();
            }

            size_t end = result.end;
            size_t length = end - result.start;
            list.append(jsSubstring(exec, s, result.start, length));
            if (!length)
                ++end;
            result = regExpConstructor->performMatch(*vm, regExp, string, s, end);
        }
    }

    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.
        return jsNull();
    }

    return constructArray(exec, static_cast<ArrayAllocationProfile*>(0), list);
}

JSValue* stringProtoFuncReplace(ExecState* exec, JSObject*, JSValue* thisValue, const ArgList& args)
{
    JSString* sourceVal = thisValue->toThisJSString(exec);
    const UString& source = sourceVal->value();

    JSValue* pattern = args.at(exec, 0);

    JSValue* replacement = args.at(exec, 1);
    UString replacementString;
    CallData callData;
    CallType callType = replacement->getCallData(callData);
    if (callType == CallTypeNone)
        replacementString = replacement->toString(exec);

    if (pattern->isObject(&RegExpObject::info)) {
        RegExp* reg = static_cast<RegExpObject*>(pattern)->regExp();
        bool global = reg->global();

        RegExpConstructor* regExpObj = exec->lexicalGlobalObject()->regExpConstructor();

        int lastIndex = 0;
        int startPosition = 0;

        Vector<UString::Range, 16> sourceRanges;
        Vector<UString, 16> replacements;

        // This is either a loop (if global is set) or a one-way (if not).
        do {
            int matchIndex;
            int matchLen;
            int* ovector;
            regExpObj->performMatch(reg, source, startPosition, matchIndex, matchLen, &ovector);
            if (matchIndex < 0)
                break;

            sourceRanges.append(UString::Range(lastIndex, matchIndex - lastIndex));

            if (callType != CallTypeNone) {
                int completeMatchStart = ovector[0];
                ArgList args;

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

                    if (matchStart < 0)
                        args.append(jsUndefined());
                    else
                        args.append(jsSubstring(exec, source, matchStart, matchLen));
                }

                args.append(jsNumber(exec, completeMatchStart));
                args.append(sourceVal);

                replacements.append(call(exec, replacement, callType, callData, exec->globalThisValue(), args)->toString(exec));
            } else
                replacements.append(substituteBackreferences(replacementString, source, ovector, reg));

            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())
            sourceRanges.append(UString::Range(lastIndex, source.size() - lastIndex));

        UString result = source.spliceSubstringsWithSeparators(sourceRanges.data(), sourceRanges.size(), replacements.data(), replacements.size());

        if (result == source)
            return sourceVal;

        return jsString(exec, result);
    }

    // First arg is a string
    UString patternString = pattern->toString(exec);
    int matchPos = source.find(patternString);
    int matchLen = patternString.size();
    // Do the replacement
    if (matchPos == -1)
        return sourceVal;

    if (callType != CallTypeNone) {
        ArgList args;
        args.append(jsSubstring(exec, source, matchPos, matchLen));
        args.append(jsNumber(exec, matchPos));
        args.append(sourceVal);

        replacementString = call(exec, replacement, callType, callData, exec->globalThisValue(), args)->toString(exec);
    }

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