stringimpl.cpp

linux下开源浏览器WebKit的源码,市面上的很多商用浏览器都是移植自WebKit

}

static bool equal(const UChar* a, const char* b, int length)
{
    ASSERT(length >= 0);
    while (length--) {
        unsigned char bc = *b++;
        if (*a++ != bc)
            return false;
    }
    return true;
}

static bool equalIgnoringCase(const UChar* a, const char* b, int length)
{
    ASSERT(length >= 0);
    while (length--) {
        unsigned char bc = *b++;
        if (foldCase(*a++) != foldCase(bc))
            return false;
    }
    return true;
}

static inline bool equalIgnoringCase(const UChar* a, const UChar* b, int length)
{
    ASSERT(length >= 0);
    return umemcasecmp(a, b, length) == 0;
}

int StringImpl::find(const char* chs, int index, bool caseSensitive)
{
    if (!chs || index < 0)
        return -1;

    int chsLength = strlen(chs);
    int n = m_length - index;
    if (n < 0)
        return -1;
    n -= chsLength - 1;
    if (n <= 0)
        return -1;

    const char* chsPlusOne = chs + 1;
    int chsLengthMinusOne = chsLength - 1;
    
    const UChar* ptr = m_data + index - 1;
    if (caseSensitive) {
        UChar c = *chs;
        do {
            if (*++ptr == c && equal(ptr + 1, chsPlusOne, chsLengthMinusOne))
                return m_length - chsLength - n + 1;
        } while (--n);
    } else {
        UChar lc = Unicode::foldCase(*chs);
        do {
            if (Unicode::foldCase(*++ptr) == lc && equalIgnoringCase(ptr + 1, chsPlusOne, chsLengthMinusOne))
                return m_length - chsLength - n + 1;
        } while (--n);
    }

    return -1;
}

int StringImpl::find(UChar c, int start)
{
    return WebCore::find(m_data, m_length, c, start);
}

int StringImpl::find(CharacterMatchFunctionPtr matchFunction, int start)
{
    return WebCore::find(m_data, m_length, matchFunction, start);
}

int StringImpl::find(StringImpl* str, int index, bool caseSensitive)
{
    /*
      We use a simple trick for efficiency's sake. Instead of
      comparing strings, we compare the sum of str with that of
      a part of this string. Only if that matches, we call memcmp
      or ucstrnicmp.
    */
    ASSERT(str);
    if (index < 0)
        index += m_length;
    int lstr = str->m_length;
    int lthis = m_length - index;
    if ((unsigned)lthis > m_length)
        return -1;
    int delta = lthis - lstr;
    if (delta < 0)
        return -1;

    const UChar* uthis = m_data + index;
    const UChar* ustr = str->m_data;
    unsigned hthis = 0;
    unsigned hstr = 0;
    if (caseSensitive) {
        for (int i = 0; i < lstr; i++) {
            hthis += uthis[i];
            hstr += ustr[i];
        }
        int i = 0;
        while (1) {
            if (hthis == hstr && memcmp(uthis + i, ustr, lstr * sizeof(UChar)) == 0)
                return index + i;
            if (i == delta)
                return -1;
            hthis += uthis[i + lstr];
            hthis -= uthis[i];
            i++;
        }
    } else {
        for (int i = 0; i < lstr; i++ ) {
            hthis += toASCIILower(uthis[i]);
            hstr += toASCIILower(ustr[i]);
        }
        int i = 0;
        while (1) {
            if (hthis == hstr && equalIgnoringCase(uthis + i, ustr, lstr))
                return index + i;
            if (i == delta)
                return -1;
            hthis += toASCIILower(uthis[i + lstr]);
            hthis -= toASCIILower(uthis[i]);
            i++;
        }
    }
}

int StringImpl::reverseFind(UChar c, int index)
{
    return WebCore::reverseFind(m_data, m_length, c, index);
}

int StringImpl::reverseFind(StringImpl* str, int index, bool caseSensitive)
{
    /*
     See StringImpl::find() for explanations.
     */
    ASSERT(str);
    int lthis = m_length;
    if (index < 0)
        index += lthis;
    
    int lstr = str->m_length;
    int delta = lthis - lstr;
    if ( index < 0 || index > lthis || delta < 0 )
        return -1;
    if ( index > delta )
        index = delta;
    
    const UChar *uthis = m_data;
    const UChar *ustr = str->m_data;
    unsigned hthis = 0;
    unsigned hstr = 0;
    int i;
    if (caseSensitive) {
        for ( i = 0; i < lstr; i++ ) {
            hthis += uthis[index + i];
            hstr += ustr[i];
        }
        i = index;
        while (1) {
            if (hthis == hstr && memcmp(uthis + i, ustr, lstr * sizeof(UChar)) == 0)
                return i;
            if (i == 0)
                return -1;
            i--;
            hthis -= uthis[i + lstr];
            hthis += uthis[i];
        }
    } else {
        for (i = 0; i < lstr; i++) {
            hthis += toASCIILower(uthis[index + i]);
            hstr += toASCIILower(ustr[i]);
        }
        i = index;
        while (1) {
            if (hthis == hstr && equalIgnoringCase(uthis + i, ustr, lstr) )
                return i;
            if (i == 0)
                return -1;
            i--;
            hthis -= toASCIILower(uthis[i + lstr]);
            hthis += toASCIILower(uthis[i]);
        }
    }
    
    // Should never get here.
    return -1;
}

bool StringImpl::endsWith(StringImpl* m_data, bool caseSensitive)
{
    ASSERT(m_data);
    int start = m_length - m_data->m_length;
    if (start >= 0)
        return (find(m_data, start, caseSensitive) == start);
    return false;
}

PassRefPtr<StringImpl> StringImpl::replace(UChar oldC, UChar newC)
{
    if (oldC == newC)
        return this;
    unsigned i;
    for (i = 0; i != m_length; ++i)
        if (m_data[i] == oldC)
            break;
    if (i == m_length)
        return this;

    StringBuffer data(m_length);
    for (i = 0; i != m_length; ++i) {
        UChar ch = m_data[i];
        if (ch == oldC)
            ch = newC;
        data[i] = ch;
    }
    return adopt(data);
}

PassRefPtr<StringImpl> StringImpl::replace(unsigned position, unsigned lengthToReplace, StringImpl* str)
{
    position = min(position, length());
    lengthToReplace = min(lengthToReplace, length() - position);
    unsigned lengthToInsert = str ? str->length() : 0;
    if (!lengthToReplace && !lengthToInsert)
        return this;
    StringBuffer buffer(length() - lengthToReplace + lengthToInsert);
    memcpy(buffer.characters(), characters(), position * sizeof(UChar));
    if (str)
        memcpy(buffer.characters() + position, str->characters(), lengthToInsert * sizeof(UChar));
    memcpy(buffer.characters() + position + lengthToInsert, characters() + position + lengthToReplace,
        (length() - position - lengthToReplace) * sizeof(UChar));
    return adopt(buffer);
}

PassRefPtr<StringImpl> StringImpl::replace(UChar pattern, StringImpl* replacement)
{
    if (!replacement)
        return this;
        
    int repStrLength = replacement->length();
    int srcSegmentStart = 0;
    int matchCount = 0;
    
    // Count the matches
    while ((srcSegmentStart = find(pattern, srcSegmentStart)) >= 0) {
        ++matchCount;
        ++srcSegmentStart;
    }
    
    // If we have 0 matches, we don't have to do any more work
    if (!matchCount)
        return this;
    
    StringBuffer data(m_length - matchCount + (matchCount * repStrLength));
    
    // Construct the new data
    int srcSegmentEnd;
    int srcSegmentLength;
    srcSegmentStart = 0;
    int dstOffset = 0;
    
    while ((srcSegmentEnd = find(pattern, srcSegmentStart)) >= 0) {
        srcSegmentLength = srcSegmentEnd - srcSegmentStart;
        memcpy(data.characters() + dstOffset, m_data + srcSegmentStart, srcSegmentLength * sizeof(UChar));
        dstOffset += srcSegmentLength;
        memcpy(data.characters() + dstOffset, replacement->m_data, repStrLength * sizeof(UChar));
        dstOffset += repStrLength;
        srcSegmentStart = srcSegmentEnd + 1;
    }

    srcSegmentLength = m_length - srcSegmentStart;
    memcpy(data.characters() + dstOffset, m_data + srcSegmentStart, srcSegmentLength * sizeof(UChar));

    ASSERT(dstOffset + srcSegmentLength == static_cast<int>(data.length()));

    return adopt(data);
}

PassRefPtr<StringImpl> StringImpl::replace(StringImpl* pattern, StringImpl* replacement)
{
    if (!pattern || !replacement)
        return this;

    int patternLength = pattern->length();
    if (!patternLength)
        return this;
        
    int repStrLength = replacement->length();
    int srcSegmentStart = 0;
    int matchCount = 0;
    
    // Count the matches
    while ((srcSegmentStart = find(pattern, srcSegmentStart)) >= 0) {
        ++matchCount;
        srcSegmentStart += patternLength;
    }
    
    // If we have 0 matches, we don't have to do any more work
    if (!matchCount)
        return this;
    
    StringBuffer data(m_length + matchCount * (repStrLength - patternLength));
    
    // Construct the new data
    int srcSegmentEnd;
    int srcSegmentLength;
    srcSegmentStart = 0;
    int dstOffset = 0;
    
    while ((srcSegmentEnd = find(pattern, srcSegmentStart)) >= 0) {
        srcSegmentLength = srcSegmentEnd - srcSegmentStart;
        memcpy(data.characters() + dstOffset, m_data + srcSegmentStart, srcSegmentLength * sizeof(UChar));
        dstOffset += srcSegmentLength;
        memcpy(data.characters() + dstOffset, replacement->m_data, repStrLength * sizeof(UChar));
        dstOffset += repStrLength;
        srcSegmentStart = srcSegmentEnd + patternLength;
    }

    srcSegmentLength = m_length - srcSegmentStart;
    memcpy(data.characters() + dstOffset, m_data + srcSegmentStart, srcSegmentLength * sizeof(UChar));

    ASSERT(dstOffset + srcSegmentLength == static_cast<int>(data.length()));

    return adopt(data);
}

bool equal(StringImpl* a, StringImpl* b)
{
    return StringHash::equal(a, b);
}

bool equal(StringImpl* a, const char* b)
{
    if (!a)
        return !b;
    if (!b)
        return !a;

    unsigned length = a->length();
    const UChar* as = a->characters();
    for (unsigned i = 0; i != length; ++i) {
        unsigned char bc = b[i];
        if (!bc)
            return false;
        if (as[i] != bc)
            return false;
    }

    return !b[length];
}

bool equalIgnoringCase(StringImpl* a, StringImpl* b)
{
    return CaseFoldingHash::equal(a, b);
}

bool equalIgnoringCase(StringImpl* a, const char* b)
{
    if (!a)
        return !b;
    if (!b)
        return !a;

    unsigned length = a->length();
    const UChar* as = a->characters();

    // Do a faster loop for the case where all the characters are ASCII.
    UChar ored = 0;
    bool equal = true;
    for (unsigned i = 0; i != length; ++i) {
        char bc = b[i];
        if (!bc)
            return false;
        UChar ac = as[i];
        ored |= ac;
        equal = equal && (toASCIILower(ac) == toASCIILower(bc));
    }

    // Do a slower implementation for cases that include non-ASCII characters.
    if (ored & ~0x7F) {
        equal = true;
        for (unsigned i = 0; i != length; ++i) {
            unsigned char bc = b[i];
            equal = equal && (foldCase(as[i]) == foldCase(bc));
        }
    }

    return equal && !b[length];
}

Vector<char> StringImpl::ascii()
{
    Vector<char> buffer(m_length + 1);
    for (unsigned i = 0; i != m_length; ++i) {
        UChar c = m_data[i];
        if ((c >= 0x20 && c < 0x7F) || c == 0x00)
            buffer[i] = c;
        else
            buffer[i] = '?';
    }
    buffer[m_length] = '\0';
    return buffer;
}

WTF::Unicode::Direction StringImpl::defaultWritingDirection()
{
    for (unsigned i = 0; i < m_length; ++i) {
        WTF::Unicode::Direction charDirection = WTF::Unicode::direction(m_data[i]);
        if (charDirection == WTF::Unicode::LeftToRight)
            return WTF::Unicode::LeftToRight;
        if (charDirection == WTF::Unicode::RightToLeft || charDirection == WTF::Unicode::RightToLeftArabic)
            return WTF::Unicode::RightToLeft;
    }
    return WTF::Unicode::LeftToRight;
}

// This is a hot function because it's used when parsing HTML.
PassRefPtr<StringImpl> StringImpl::createStrippingNullCharactersSlowCase(const UChar* characters, unsigned length)
{
    StringBuffer strippedCopy(length);
    unsigned strippedLength = 0;
    for (unsigned i = 0; i < length; i++) {
        if (int c = characters[i])
            strippedCopy[strippedLength++] = c;
    }
    ASSERT(strippedLength < length);  // Only take the slow case when stripping.
    strippedCopy.shrink(strippedLength);
    return adopt(strippedCopy);
}

PassRefPtr<StringImpl> StringImpl::adopt(StringBuffer& buffer)
{
    unsigned length = buffer.length();
    if (length == 0)
        return empty();
    return adoptRef(new StringImpl(buffer.release(), length, AdoptBuffer()));
}

PassRefPtr<StringImpl> StringImpl::adopt(Vector<UChar>& vector)
{
    size_t size = vector.size();
    if (size == 0)
        return empty();
    return adoptRef(new StringImpl(vector.releaseBuffer(), size, AdoptBuffer()));
}

PassRefPtr<StringImpl> StringImpl::create(const UChar* characters, unsigned length)
{
    if (!characters || !length)
        return empty();
    return adoptRef(new StringImpl(characters, length));
}

PassRefPtr<StringImpl> StringImpl::create(const char* characters, unsigned length)
{
    if (!characters || !length)
        return empty();
    return adoptRef(new StringImpl(characters, length));
}

PassRefPtr<StringImpl> StringImpl::create(const char* string)
{
    if (!string)
        return empty();
    unsigned length = strlen(string);
    if (!length)
        return empty();
    return adoptRef(new StringImpl(string, length));
}

PassRefPtr<StringImpl> StringImpl::createWithTerminatingNullCharacter(const StringImpl& string)
{
    return adoptRef(new StringImpl(string, WithTerminatingNullCharacter()));
}

PassRefPtr<StringImpl> StringImpl::copy()
{
    return adoptRef(new StringImpl(m_data, m_length));
}

} // namespace WebCore