QRegExpEngine Class Reference

Classes
class	Box

Public Functions
void	addAnchors (int from, int to, int a)
void	addCatTransitions (const QVector< int > &from, const QVector< int > &to)
void	addPlusTransitions (const QVector< int > &from, const QVector< int > &to, int atom)
int	anchorAlternation (int a, int b)
int	anchorConcatenation (int a, int b)
int	captureCount () const
int	createState (QChar ch)
int	createState (const QRegExpCharClass &cc)
int	createState (int bref)
void	dump () const
const QString &	errorString () const
void	heuristicallyChooseHeuristic ()
bool	isValid () const
	QRegExpEngine (Qt::CaseSensitivity cs, bool greedyQuantifiers)
	QRegExpEngine (const QRegExpEngineKey &key)
	~QRegExpEngine ()

Public Variables
QAtomicInt	ref

Private Types
enum	{ CharClassBit = 0x10000, BackRefBit = 0x20000 }
enum	{ InitialState = 0, FinalState = 1 }
enum	{ MaxLookaheads = 13, MaxBackRefs = 14 }
enum	{   Anchor_Dollar = 0x00000001, Anchor_Caret = 0x00000002, Anchor_Word = 0x00000004, Anchor_NonWord = 0x00000008,   Anchor_FirstLookahead = 0x00000010, Anchor_BackRef1Empty = Anchor_FirstLookahead << MaxLookaheads, Anchor_BackRef0Empty = Anchor_BackRef1Empty >> 1, Anchor_Alternation = unsigned(Anchor_BackRef1Empty) << MaxBackRefs,   Anchor_LookaheadMask }
enum	{   Tok_Eos, Tok_Dollar, Tok_LeftParen, Tok_MagicLeftParen,   Tok_PosLookahead, Tok_NegLookahead, Tok_RightParen, Tok_CharClass,   Tok_Caret, Tok_Quantifier, Tok_Bar, Tok_Word,   Tok_NonWord, Tok_Char = 0x10000, Tok_BackRef = 0x20000 }

Private Functions
int	addLookahead (QRegExpEngine *eng, bool negative)
bool	badCharMatch (QRegExpMatchState &matchState) const
void	error (const char *msg)
void	finishAtom (int atom, bool needCapture)
int	getChar ()
int	getEscape ()
int	getRep (int def)
int	getToken ()
bool	goodStringMatch (QRegExpMatchState &matchState) const
int	parse (const QChar *rx, int len)
void	parseAtom (Box *box)
void	parseExpression (Box *box)
void	parseFactor (Box *box)
void	parseTerm (Box *box)
void	setup ()
void	setupCategoriesRangeMap ()
int	setupState (int match)
void	skipChars (int n)
int	startAtom (bool officialCapture)
void	startTokenizer (const QChar *rx, int len)

Properties
QVector< QRegExpAnchorAlternation >	aa
QVector< QRegExpLookahead * >	ahead
QVector< int >	captureForOfficialCapture
bool	caretAnchored
QHash< QByteArray, QPair< int, int > >	categoriesRangeMap
int	cf
QVector< QRegExpCharClass >	cl
Qt::CaseSensitivity	cs
QVector< QRegExpAtom >	f
int	goodEarlyStart
int	goodLateStart
QString	goodStr
bool	greedyQuantifiers
int	minl
int	nbrefs
int	ncap
int	nf
QVector< int >	occ1
int	officialncap
QVector< QRegExpAutomatonState >	s
bool	trivial
bool	useGoodStringHeuristic
bool	valid
bool	xmlSchemaExtensions
int	yyCh
QScopedPointer< QRegExpCharClass >	yyCharClass
QString	yyError
const QChar *	yyIn
int	yyLen
int	yyMaxRep
bool	yyMayCapture
int	yyMinRep
int	yyPos
int	yyPos0
int	yyTok

Friends
class	Box
struct	QRegExpMatchState

Detailed Description

Definition at line 1074 of file qregexp.cpp.

Enumerations

anonymous enum

anonymous enum

private

Enumerator
CharClassBit
BackRefBit

Definition at line 1118 of file qregexp.cpp.

{ CharClassBit = 0x10000, BackRefBit = 0x20000 };

anonymous enum

anonymous enum

private

Enumerator
InitialState
FinalState

Definition at line 1119 of file qregexp.cpp.

{ InitialState = 0, FinalState = 1 };

anonymous enum

anonymous enum

private

Enumerator
MaxLookaheads
MaxBackRefs

Definition at line 1128 of file qregexp.cpp.

{ MaxLookaheads = 13, MaxBackRefs = 14 };

anonymous enum

anonymous enum

private

Enumerator
Anchor_Dollar
Anchor_Caret
Anchor_Word
Anchor_NonWord
Anchor_FirstLookahead
Anchor_BackRef1Empty
Anchor_BackRef0Empty
Anchor_Alternation
Anchor_LookaheadMask

Definition at line 1129 of file qregexp.cpp.

         { Anchor_Dollar = 0x00000001, Anchor_Caret = 0x00000002, Anchor_Word = 0x00000004,
           Anchor_NonWord = 0x00000008, Anchor_FirstLookahead = 0x00000010,
           Anchor_BackRef1Empty = Anchor_FirstLookahead << MaxLookaheads,
           Anchor_BackRef0Empty = Anchor_BackRef1Empty >> 1,
           Anchor_Alternation = unsigned(Anchor_BackRef1Empty) << MaxBackRefs,

           Anchor_LookaheadMask = (Anchor_FirstLookahead - 1) ^
                   ((Anchor_FirstLookahead << MaxLookaheads) - 1) };

anonymous enum

anonymous enum

private

Enumerator
Tok_Eos
Tok_Dollar
Tok_LeftParen
Tok_MagicLeftParen
Tok_PosLookahead
Tok_NegLookahead
Tok_RightParen
Tok_CharClass
Tok_Caret
Tok_Quantifier
Tok_Bar
Tok_Word
Tok_NonWord
Tok_Char
Tok_BackRef

Definition at line 1261 of file qregexp.cpp.

         { Tok_Eos, Tok_Dollar, Tok_LeftParen, Tok_MagicLeftParen, Tok_PosLookahead,
           Tok_NegLookahead, Tok_RightParen, Tok_CharClass, Tok_Caret, Tok_Quantifier, Tok_Bar,
           Tok_Word, Tok_NonWord, Tok_Char = 0x10000, Tok_BackRef = 0x20000 };

Constructors and Destructors

QRegExpEngine() [1/2]

QRegExpEngine::QRegExpEngine ( Qt::CaseSensitivity  cs, bool  greedyQuantifiers  )

inline

Definition at line 1077 of file qregexp.cpp.

Referenced by parseAtom(), and prepareEngine_helper().

        : cs(cs), greedyQuantifiers(greedyQuantifiers) { setup(); }

QRegExpEngine() [2/2]

QRegExpEngine::QRegExpEngine

(

const QRegExpEngineKey &

key

)

Definition at line 1343 of file qregexp.cpp.

    : cs(key.cs), greedyQuantifiers(key.patternSyntax == QRegExp::RegExp2),
      xmlSchemaExtensions(key.patternSyntax == QRegExp::W3CXmlSchema11)
{
    setup();

    QString rx = qt_regexp_toCanonical(key.pattern, key.patternSyntax);

    valid = (parse(rx.unicode(), rx.length()) == rx.length());
    if (!valid) {
#ifndef QT_NO_REGEXP_OPTIM
        trivial = false;
#endif
        error(RXERR_LEFTDELIM);
    }
}

~QRegExpEngine()

QRegExpEngine::~QRegExpEngine

(

)

Definition at line 1360 of file qregexp.cpp.

{
#ifndef QT_NO_REGEXP_LOOKAHEAD
    qDeleteAll(ahead);
#endif
}

Functions

addAnchors()

void QRegExpEngine::addAnchors	(	int	from,
		int	to,
		int	a
	)

Definition at line 1593 of file qregexp.cpp.

Referenced by QRegExpEngine::Box::addAnchorsToEngine().

{
    QRegExpAutomatonState &st = s[from];
    if (st.anchors.contains(to))
        a = anchorAlternation(st.anchors.value(to), a);
    st.anchors.insert(to, a);
}

addCatTransitions()

void QRegExpEngine::addCatTransitions	(	const QVector< int > &	from,
		const QVector< int > &	to
	)

Definition at line 1528 of file qregexp.cpp.

Referenced by QRegExpEngine::Box::cat(), and QRegExpEngine::Box::plus().

{
    for (int i = 0; i < from.size(); i++)
        mergeInto(&s[from.at(i)].outs, to);
}

addLookahead()

int QRegExpEngine::addLookahead ( QRegExpEngine *  eng, bool  negative  )

private

Definition at line 1769 of file qregexp.cpp.

Referenced by parseAtom().

{
    int n = ahead.size();
    if (n == MaxLookaheads) {
        error(RXERR_LIMIT);
        return 0;
    }
    ahead += new QRegExpLookahead(eng, negative);
    return Anchor_FirstLookahead << n;
}

addPlusTransitions()

void QRegExpEngine::addPlusTransitions	(	const QVector< int > &	from,
		const QVector< int > &	to,
		int	atom
	)

Definition at line 1535 of file qregexp.cpp.

Referenced by QRegExpEngine::Box::plus().

{
    for (int i = 0; i < from.size(); i++) {
        QRegExpAutomatonState &st = s[from.at(i)];
        const QVector<int> oldOuts = st.outs;
        mergeInto(&st.outs, to);
        if (f.at(atom).capture != QRegExpAtom::NoCapture) {
            for (int j = 0; j < to.size(); j++) {
                // ### st.reenter.contains(to.at(j)) check looks suspicious
                if (!st.reenter.contains(to.at(j)) &&
                     qBinaryFind(oldOuts.constBegin(), oldOuts.constEnd(), to.at(j)) == oldOuts.end())
                    st.reenter.insert(to.at(j), atom);
            }
        }
    }
}

anchorAlternation()

int QRegExpEngine::anchorAlternation	(	int	a,
		int	b
	)

Definition at line 1557 of file qregexp.cpp.

Referenced by addAnchors(), anchorConcatenation(), and QRegExpEngine::Box::orx().

{
    if (((a & b) == a || (a & b) == b) && ((a | b) & Anchor_Alternation) == 0)
        return a & b;

    int n = aa.size();
#ifndef QT_NO_REGEXP_OPTIM
    if (n > 0 && aa.at(n - 1).a == a && aa.at(n - 1).b == b)
        return Anchor_Alternation | (n - 1);
#endif

    QRegExpAnchorAlternation element = {a, b};
    aa.append(element);
    return Anchor_Alternation | n;
}

anchorConcatenation()

int QRegExpEngine::anchorConcatenation	(	int	a,
		int	b
	)

Definition at line 1576 of file qregexp.cpp.

Referenced by QRegExpEngine::Box::addAnchorsToEngine(), QRegExpEngine::Box::cat(), and QRegExpEngine::Box::catAnchor().

{
    if (((a | b) & Anchor_Alternation) == 0)
        return a | b;
    if ((b & Anchor_Alternation) != 0)
        qSwap(a, b);

    int aprime = anchorConcatenation(aa.at(a ^ Anchor_Alternation).a, b);
    int bprime = anchorConcatenation(aa.at(a ^ Anchor_Alternation).b, b);
    return anchorAlternation(aprime, bprime);
}

badCharMatch()

bool QRegExpEngine::badCharMatch ( QRegExpMatchState &  matchState ) const

private

Definition at line 1892 of file qregexp.cpp.

{
    int slideHead = 0;
    int slideNext = 0;
    int i;
    int lastPos = matchState.len - minl;
    memset(matchState.slideTab, 0, matchState.slideTabSize * sizeof(int));

    /*
      Set up the slide table, used for the bad-character heuristic,
      using the table of first occurrence of each character.
    */
    for (i = 0; i < minl; i++) {
        int sk = occ1[BadChar(matchState.in[matchState.pos + i])];
        if (sk == NoOccurrence)
            sk = i + 1;
        if (sk > 0) {
            int k = i + 1 - sk;
            if (k < 0) {
                sk = i + 1;
                k = 0;
            }
            if (sk > matchState.slideTab[k])
                matchState.slideTab[k] = sk;
        }
    }

    if (matchState.pos > lastPos)
        return false;

    for (;;) {
        if (++slideNext >= matchState.slideTabSize)
            slideNext = 0;
        if (matchState.slideTab[slideHead] > 0) {
            if (matchState.slideTab[slideHead] - 1 > matchState.slideTab[slideNext])
                matchState.slideTab[slideNext] = matchState.slideTab[slideHead] - 1;
            matchState.slideTab[slideHead] = 0;
        } else {
            if (matchState.matchHere())
                return true;
        }

        if (matchState.pos == lastPos)
            break;

        /*
          Update the slide table. This code has much in common with
          the initialization code.
        */
        int sk = occ1[BadChar(matchState.in[matchState.pos + minl])];
        if (sk == NoOccurrence) {
            matchState.slideTab[slideNext] = minl;
        } else if (sk > 0) {
            int k = slideNext + minl - sk;
            if (k >= matchState.slideTabSize)
                k -= matchState.slideTabSize;
            if (sk > matchState.slideTab[k])
                matchState.slideTab[k] = sk;
        }
        slideHead = slideNext;
        ++matchState.pos;
    }
    return false;
}

captureCount()

int QRegExpEngine::captureCount ( ) const

inline

Definition at line 1085 of file qregexp.cpp.

Referenced by QRegExp::numCaptures(), and QRegExpMatchState::prepareForMatch().

{ return officialncap; }

createState() [1/3]

int QRegExpEngine::createState

(

QChar

ch

)

Definition at line 1489 of file qregexp.cpp.

Referenced by QRegExpEngine::Box::set().

{
    return setupState(ch.unicode());
}

createState() [2/3]

int QRegExpEngine::createState

(

const QRegExpCharClass &

cc

)

Definition at line 1494 of file qregexp.cpp.

{
#ifndef QT_NO_REGEXP_CCLASS
    int n = cl.size();
    cl += QRegExpCharClass(cc);
    return setupState(CharClassBit | n);
#else
    Q_UNUSED(cc);
    return setupState(CharClassBit);
#endif
}

createState() [3/3]

int QRegExpEngine::createState

(

int

bref

)

Definition at line 1507 of file qregexp.cpp.

{
    if (bref > nbrefs) {
        nbrefs = bref;
        if (nbrefs > MaxBackRefs) {
            error(RXERR_LIMIT);
            return 0;
        }
    }
    return setupState(BackRefBit | bref);
}

dump()

void QRegExpEngine::dump

(

)

const

Definition at line 1649 of file qregexp.cpp.

{
    int i, j;
    qDebug("Case %ssensitive engine", cs ? "" : "in");
    qDebug("  States");
    for (i = 0; i < s.size(); i++) {
        qDebug("  %d%s", i, i == InitialState ? " (initial)" : i == FinalState ? " (final)" : "");
#ifndef QT_NO_REGEXP_CAPTURE
        if (nf > 0)
            qDebug("    in atom %d", s[i].atom);
#endif
        int m = s[i].match;
        if ((m & CharClassBit) != 0) {
            qDebug("    match character class %d", m ^ CharClassBit);
#ifndef QT_NO_REGEXP_CCLASS
            cl[m ^ CharClassBit].dump();
#else
            qDebug("    negative character class");
#endif
        } else if ((m & BackRefBit) != 0) {
            qDebug("    match back-reference %d", m ^ BackRefBit);
        } else if (m >= 0x20 && m <= 0x7e) {
            qDebug("    match 0x%.4x (%c)", m, m);
        } else {
            qDebug("    match 0x%.4x", m);
        }
        for (j = 0; j < s[i].outs.size(); j++) {
            int next = s[i].outs[j];
            qDebug("    -> %d", next);
            if (s[i].reenter.contains(next))
                qDebug("       [reenter %d]", s[i].reenter[next]);
            if (s[i].anchors.value(next) != 0)
                qDebug("       [anchors 0x%.8x]", s[i].anchors[next]);
        }
    }
#ifndef QT_NO_REGEXP_CAPTURE
    if (nf > 0) {
        qDebug("  Atom    Parent  Capture");
        for (i = 0; i < nf; i++) {
            if (f[i].capture == QRegExpAtom::NoCapture) {
                qDebug("  %6d  %6d     nil", i, f[i].parent);
            } else {
                int cap = f[i].capture;
                bool official = captureForOfficialCapture.contains(cap);
                qDebug("  %6d  %6d  %6d  %s", i, f[i].parent, f[i].capture,
                       official ? "official" : "");
            }
        }
    }
#endif
#ifndef QT_NO_REGEXP_ANCHOR_ALT
    for (i = 0; i < aa.size(); i++)
        qDebug("  Anchor alternation 0x%.8x: 0x%.8x 0x%.9x", i, aa[i].a, aa[i].b);
#endif
}

error()

void QRegExpEngine::error ( const char *  msg )

private

Definition at line 3181 of file qregexp.cpp.

Referenced by addLookahead(), createState(), getEscape(), getRep(), getToken(), parseAtom(), and QRegExpEngine().

{
    if (yyError.isEmpty())
        yyError = QLatin1String(msg);
}

errorString()

const QString& QRegExpEngine::errorString ( ) const

inline

Definition at line 1084 of file qregexp.cpp.

Referenced by QRegExp::errorString().

{ return yyError; }

finishAtom()

void QRegExpEngine::finishAtom ( int  atom, bool  needCapture  )

private

Definition at line 1757 of file qregexp.cpp.

Referenced by parse(), and parseFactor().

{
    if (greedyQuantifiers && needCapture && f[atom].capture == QRegExpAtom::NoCapture)
        f[atom].capture = QRegExpAtom::UnofficialCapture;
    cf = f.at(atom).parent;
}

getChar()

int QRegExpEngine::getChar ( )

private

Definition at line 2855 of file qregexp.cpp.

Referenced by getEscape(), getRep(), getToken(), skipChars(), and startTokenizer().

{
    return (yyPos == yyLen) ? EOS : yyIn[yyPos++].unicode();
}

getEscape()

int QRegExpEngine::getEscape ( )

private

Definition at line 2860 of file qregexp.cpp.

Referenced by getToken().

{
#ifndef QT_NO_REGEXP_ESCAPE
    const char tab[] = "afnrtv"; // no b, as \b means word boundary
    const char backTab[] = "\a\f\n\r\t\v";
    ushort low;
    int i;
#endif
    ushort val;
    int prevCh = yyCh;

    if (prevCh == EOS) {
        error(RXERR_END);
        return Tok_Char | '\\';
    }
    yyCh = getChar();
#ifndef QT_NO_REGEXP_ESCAPE
    if ((prevCh & ~0xff) == 0) {
        const char *p = strchr(tab, prevCh);
        if (p != 0)
            return Tok_Char | backTab[p - tab];
    }
#endif

    switch (prevCh) {
#ifndef QT_NO_REGEXP_ESCAPE
    case '0':
        val = 0;
        for (i = 0; i < 3; i++) {
            if (yyCh >= '0' && yyCh <= '7')
                val = (val << 3) | (yyCh - '0');
            else
                break;
            yyCh = getChar();
        }
        if ((val & ~0377) != 0)
            error(RXERR_OCTAL);
        return Tok_Char | val;
#endif
#ifndef QT_NO_REGEXP_ESCAPE
    case 'B':
        return Tok_NonWord;
#endif
#ifndef QT_NO_REGEXP_CCLASS
    case 'D':
        // see QChar::isDigit()
        yyCharClass->addCategories(0x7fffffef);
        return Tok_CharClass;
    case 'S':
        // see QChar::isSpace()
        yyCharClass->addCategories(0x7ffff87f);
        yyCharClass->addRange(0x0000, 0x0008);
        yyCharClass->addRange(0x000e, 0x001f);
        yyCharClass->addRange(0x007f, 0x009f);
        return Tok_CharClass;
    case 'W':
        // see QChar::isLetterOrNumber() and QChar::isMark()
        yyCharClass->addCategories(0x7fe07f81);
        yyCharClass->addRange(0x203f, 0x2040);
        yyCharClass->addSingleton(0x2040);
        yyCharClass->addSingleton(0x2054);
        yyCharClass->addSingleton(0x30fb);
        yyCharClass->addRange(0xfe33, 0xfe34);
        yyCharClass->addRange(0xfe4d, 0xfe4f);
        yyCharClass->addSingleton(0xff3f);
        yyCharClass->addSingleton(0xff65);
        return Tok_CharClass;
#endif
#ifndef QT_NO_REGEXP_ESCAPE
    case 'b':
        return Tok_Word;
#endif
#ifndef QT_NO_REGEXP_CCLASS
    case 'd':
        // see QChar::isDigit()
        yyCharClass->addCategories(0x00000010);
        return Tok_CharClass;
    case 's':
        // see QChar::isSpace()
        yyCharClass->addCategories(0x00000380);
        yyCharClass->addRange(0x0009, 0x000d);
        return Tok_CharClass;
    case 'w':
        // see QChar::isLetterOrNumber() and QChar::isMark()
        yyCharClass->addCategories(0x000f807e);
        yyCharClass->addSingleton(0x005f); // '_'
        return Tok_CharClass;
    case 'I':
        if (xmlSchemaExtensions) {
            yyCharClass->setNegative(!yyCharClass->negative());
            // fall through
        }
    case 'i':
        if (xmlSchemaExtensions) {
            yyCharClass->addCategories(0x000f807e);
            yyCharClass->addSingleton(0x003a); // ':'
            yyCharClass->addSingleton(0x005f); // '_'
            yyCharClass->addRange(0x0041, 0x005a); // [A-Z]
            yyCharClass->addRange(0x0061, 0x007a); // [a-z]
            yyCharClass->addRange(0xc0, 0xd6);
            yyCharClass->addRange(0xd8, 0xf6);
            yyCharClass->addRange(0xf8, 0x2ff);
            yyCharClass->addRange(0x370, 0x37d);
            yyCharClass->addRange(0x37f, 0x1fff);
            yyCharClass->addRange(0x200c, 0x200d);
            yyCharClass->addRange(0x2070, 0x218f);
            yyCharClass->addRange(0x2c00, 0x2fef);
            yyCharClass->addRange(0x3001, 0xd7ff);
            yyCharClass->addRange(0xf900, 0xfdcf);
            yyCharClass->addRange(0xfdf0, 0xfffd);
            yyCharClass->addRange((ushort)0x10000, (ushort)0xeffff);
        }
        return Tok_CharClass;
    case 'C':
        if (xmlSchemaExtensions) {
            yyCharClass->setNegative(!yyCharClass->negative());
            // fall through
        }
    case 'c':
        if (xmlSchemaExtensions) {
            yyCharClass->addCategories(0x000f807e);
            yyCharClass->addSingleton(0x002d); // '-'
            yyCharClass->addSingleton(0x002e); // '.'
            yyCharClass->addSingleton(0x003a); // ':'
            yyCharClass->addSingleton(0x005f); // '_'
            yyCharClass->addSingleton(0xb7);
            yyCharClass->addRange(0x0030, 0x0039); // [0-9]
            yyCharClass->addRange(0x0041, 0x005a); // [A-Z]
            yyCharClass->addRange(0x0061, 0x007a); // [a-z]
            yyCharClass->addRange(0xc0, 0xd6);
            yyCharClass->addRange(0xd8, 0xf6);
            yyCharClass->addRange(0xf8, 0x2ff);
            yyCharClass->addRange(0x370, 0x37d);
            yyCharClass->addRange(0x37f, 0x1fff);
            yyCharClass->addRange(0x200c, 0x200d);
            yyCharClass->addRange(0x2070, 0x218f);
            yyCharClass->addRange(0x2c00, 0x2fef);
            yyCharClass->addRange(0x3001, 0xd7ff);
            yyCharClass->addRange(0xf900, 0xfdcf);
            yyCharClass->addRange(0xfdf0, 0xfffd);
            yyCharClass->addRange((ushort)0x10000, (ushort)0xeffff);
            yyCharClass->addRange(0x0300, 0x036f);
            yyCharClass->addRange(0x203f, 0x2040);
        }
        return Tok_CharClass;
    case 'P':
        if (xmlSchemaExtensions) {
            yyCharClass->setNegative(!yyCharClass->negative());
            // fall through
        }
    case 'p':
        if (xmlSchemaExtensions) {
            if (yyCh != '{') {
                error(RXERR_CHARCLASS);
                return Tok_CharClass;
            }

            QByteArray category;
            yyCh = getChar();
            while (yyCh != '}') {
                if (yyCh == EOS) {
                    error(RXERR_END);
                    return Tok_CharClass;
                }
                category.append(yyCh);
                yyCh = getChar();
            }
            yyCh = getChar(); // skip closing '}'

            if (category == "M") {
                yyCharClass->addCategories(0x0000000e);
            } else if (category == "Mn") {
                yyCharClass->addCategories(0x00000002);
            } else if (category == "Mc") {
                yyCharClass->addCategories(0x00000004);
            } else if (category == "Me") {
                yyCharClass->addCategories(0x00000008);
            } else if (category == "N") {
                yyCharClass->addCategories(0x00000070);
            } else if (category == "Nd") {
                yyCharClass->addCategories(0x00000010);
            } else if (category == "Nl") {
                yyCharClass->addCategories(0x00000020);
            } else if (category == "No") {
                yyCharClass->addCategories(0x00000040);
            } else if (category == "Z") {
                yyCharClass->addCategories(0x00000380);
            } else if (category == "Zs") {
                yyCharClass->addCategories(0x00000080);
            } else if (category == "Zl") {
                yyCharClass->addCategories(0x00000100);
            } else if (category == "Zp") {
                yyCharClass->addCategories(0x00000200);
            } else if (category == "C") {
                yyCharClass->addCategories(0x00006c00);
            } else if (category == "Cc") {
                yyCharClass->addCategories(0x00000400);
            } else if (category == "Cf") {
                yyCharClass->addCategories(0x00000800);
            } else if (category == "Cs") {
                yyCharClass->addCategories(0x00001000);
            } else if (category == "Co") {
                yyCharClass->addCategories(0x00002000);
            } else if (category == "Cn") {
                yyCharClass->addCategories(0x00004000);
            } else if (category == "L") {
                yyCharClass->addCategories(0x000f8000);
            } else if (category == "Lu") {
                yyCharClass->addCategories(0x00008000);
            } else if (category == "Ll") {
                yyCharClass->addCategories(0x00010000);
            } else if (category == "Lt") {
                yyCharClass->addCategories(0x00020000);
            } else if (category == "Lm") {
                yyCharClass->addCategories(0x00040000);
            } else if (category == "Lo") {
                yyCharClass->addCategories(0x00080000);
            } else if (category == "P") {
                yyCharClass->addCategories(0x4f580780);
            } else if (category == "Pc") {
                yyCharClass->addCategories(0x00100000);
            } else if (category == "Pd") {
                yyCharClass->addCategories(0x00200000);
            } else if (category == "Ps") {
                yyCharClass->addCategories(0x00400000);
            } else if (category == "Pe") {
                yyCharClass->addCategories(0x00800000);
            } else if (category == "Pi") {
                yyCharClass->addCategories(0x01000000);
            } else if (category == "Pf") {
                yyCharClass->addCategories(0x02000000);
            } else if (category == "Po") {
                yyCharClass->addCategories(0x04000000);
            } else if (category == "S") {
                yyCharClass->addCategories(0x78000000);
            } else if (category == "Sm") {
                yyCharClass->addCategories(0x08000000);
            } else if (category == "Sc") {
                yyCharClass->addCategories(0x10000000);
            } else if (category == "Sk") {
                yyCharClass->addCategories(0x20000000);
            } else if (category == "So") {
                yyCharClass->addCategories(0x40000000);
            } else if (category.startsWith("Is")) {
                if (categoriesRangeMap.isEmpty())
                    setupCategoriesRangeMap();

                if (categoriesRangeMap.contains(category)) {
                    const QPair<int, int> range = categoriesRangeMap.value(category);
                    yyCharClass->addRange(range.first, range.second);
                } else {
                    error(RXERR_CATEGORY);
                }
            } else {
                error(RXERR_CATEGORY);
            }
        }
        return Tok_CharClass;
#endif
#ifndef QT_NO_REGEXP_ESCAPE
    case 'x':
        val = 0;
        for (i = 0; i < 4; i++) {
            low = QChar(yyCh).toLower().unicode();
            if (low >= '0' && low <= '9')
                val = (val << 4) | (low - '0');
            else if (low >= 'a' && low <= 'f')
                val = (val << 4) | (low - 'a' + 10);
            else
                break;
            yyCh = getChar();
        }
        return Tok_Char | val;
#endif
    default:
        if (prevCh >= '1' && prevCh <= '9') {
#ifndef QT_NO_REGEXP_BACKREF
            val = prevCh - '0';
            while (yyCh >= '0' && yyCh <= '9') {
                val = (val * 10) + (yyCh - '0');
                yyCh = getChar();
            }
            return Tok_BackRef | val;
#else
            error(RXERR_DISABLED);
#endif
        }
        return Tok_Char | prevCh;
    }
}

getRep()

int QRegExpEngine::getRep ( int  def )

private

Definition at line 3152 of file qregexp.cpp.

Referenced by getToken().

{
    if (yyCh >= '0' && yyCh <= '9') {
        int rep = 0;
        do {
            rep = 10 * rep + yyCh - '0';
            if (rep >= InftyRep) {
                error(RXERR_REPETITION);
                rep = def;
            }
            yyCh = getChar();
        } while (yyCh >= '0' && yyCh <= '9');
        return rep;
    } else {
        return def;
    }
}

getToken()

int QRegExpEngine::getToken ( )

private

Definition at line 3200 of file qregexp.cpp.

Referenced by parse(), parseAtom(), parseExpression(), and parseFactor().

{
#ifndef QT_NO_REGEXP_CCLASS
    ushort pendingCh = 0;
    bool charPending;
    bool rangePending;
    int tok;
#endif
    int prevCh = yyCh;

    yyPos0 = yyPos - 1;
#ifndef QT_NO_REGEXP_CCLASS
    yyCharClass->clear();
#endif
    yyMinRep = 0;
    yyMaxRep = 0;
    yyCh = getChar();

    switch (prevCh) {
    case EOS:
        yyPos0 = yyPos;
        return Tok_Eos;
    case '$':
        return Tok_Dollar;
    case '(':
        if (yyCh == '?') {
            prevCh = getChar();
            yyCh = getChar();
            switch (prevCh) {
#ifndef QT_NO_REGEXP_LOOKAHEAD
            case '!':
                return Tok_NegLookahead;
            case '=':
                return Tok_PosLookahead;
#endif
            case ':':
                return Tok_MagicLeftParen;
            case '<':
                error(RXERR_LOOKBEHIND);
                return Tok_MagicLeftParen;
            default:
                error(RXERR_LOOKAHEAD);
                return Tok_MagicLeftParen;
            }
        } else {
            return Tok_LeftParen;
        }
    case ')':
        return Tok_RightParen;
    case '*':
        yyMinRep = 0;
        yyMaxRep = InftyRep;
        return Tok_Quantifier;
    case '+':
        yyMinRep = 1;
        yyMaxRep = InftyRep;
        return Tok_Quantifier;
    case '.':
#ifndef QT_NO_REGEXP_CCLASS
        yyCharClass->setNegative(true);
#endif
        return Tok_CharClass;
    case '?':
        yyMinRep = 0;
        yyMaxRep = 1;
        return Tok_Quantifier;
    case '[':
#ifndef QT_NO_REGEXP_CCLASS
        if (yyCh == '^') {
            yyCharClass->setNegative(true);
            yyCh = getChar();
        }
        charPending = false;
        rangePending = false;
        do {
            if (yyCh == '-' && charPending && !rangePending) {
                rangePending = true;
                yyCh = getChar();
            } else {
                if (charPending && !rangePending) {
                    yyCharClass->addSingleton(pendingCh);
                    charPending = false;
                }
                if (yyCh == '\\') {
                    yyCh = getChar();
                    tok = getEscape();
                    if (tok == Tok_Word)
                        tok = '\b';
                } else {
                    tok = Tok_Char | yyCh;
                    yyCh = getChar();
                }
                if (tok == Tok_CharClass) {
                    if (rangePending) {
                        yyCharClass->addSingleton('-');
                        yyCharClass->addSingleton(pendingCh);
                        charPending = false;
                        rangePending = false;
                    }
                } else if ((tok & Tok_Char) != 0) {
                    if (rangePending) {
                        yyCharClass->addRange(pendingCh, tok ^ Tok_Char);
                        charPending = false;
                        rangePending = false;
                    } else {
                        pendingCh = tok ^ Tok_Char;
                        charPending = true;
                    }
                } else {
                    error(RXERR_CHARCLASS);
                }
            }
        }  while (yyCh != ']' && yyCh != EOS);
        if (rangePending)
            yyCharClass->addSingleton('-');
        if (charPending)
            yyCharClass->addSingleton(pendingCh);
        if (yyCh == EOS)
            error(RXERR_END);
        else
            yyCh = getChar();
        return Tok_CharClass;
#else
        error(RXERR_END);
        return Tok_Char | '[';
#endif
    case '\\':
        return getEscape();
    case ']':
        error(RXERR_LEFTDELIM);
        return Tok_Char | ']';
    case '^':
        return Tok_Caret;
    case '{':
#ifndef QT_NO_REGEXP_INTERVAL
        yyMinRep = getRep(0);
        yyMaxRep = yyMinRep;
        if (yyCh == ',') {
            yyCh = getChar();
            yyMaxRep = getRep(InftyRep);
        }
        if (yyMaxRep < yyMinRep)
            error(RXERR_INTERVAL);
        if (yyCh != '}')
            error(RXERR_REPETITION);
        yyCh = getChar();
        return Tok_Quantifier;
#else
        error(RXERR_DISABLED);
        return Tok_Char | '{';
#endif
    case '|':
        return Tok_Bar;
    case '}':
        error(RXERR_LEFTDELIM);
        return Tok_Char | '}';
    default:
        return Tok_Char | prevCh;
    }
}

goodStringMatch()

bool QRegExpEngine::goodStringMatch ( QRegExpMatchState &  matchState ) const

private

Definition at line 1872 of file qregexp.cpp.

{
    int k = matchState.pos + goodEarlyStart;
    QStringMatcher matcher(goodStr.unicode(), goodStr.length(), cs);
    while ((k = matcher.indexIn(matchState.in, matchState.len, k)) != -1) {
        int from = k - goodLateStart;
        int to = k - goodEarlyStart;
        if (from > matchState.pos)
            matchState.pos = from;

        while (matchState.pos <= to) {
            if (matchState.matchHere())
                return true;
            ++matchState.pos;
        }
        ++k;
    }
    return false;
}

heuristicallyChooseHeuristic()

void QRegExpEngine::heuristicallyChooseHeuristic

(

)

Definition at line 1616 of file qregexp.cpp.

Referenced by QRegExpEngine::Box::setupHeuristics().

{
    if (minl == 0) {
        useGoodStringHeuristic = false;
    } else if (trivial) {
        useGoodStringHeuristic = true;
    } else {
        /*
          Magic formula: The good string has to constitute a good
          proportion of the minimum-length string, and appear at a
          more-or-less known index.
        */
        int goodStringScore = (64 * goodStr.length() / minl) -
                              (goodLateStart - goodEarlyStart);
        /*
          Less magic formula: We pick some characters at random, and
          check whether they are good or bad.
        */
        int badCharScore = 0;
        int step = qMax(1, NumBadChars / 32);
        for (int i = 1; i < NumBadChars; i += step) {
            if (occ1.at(i) == NoOccurrence)
                badCharScore += minl;
            else
                badCharScore += occ1.at(i);
        }
        badCharScore /= minl;
        useGoodStringHeuristic = (goodStringScore > badCharScore);
    }
}

isValid()

bool QRegExpEngine::isValid ( ) const

inline

Definition at line 1083 of file qregexp.cpp.

Referenced by QRegExp::errorString().

{ return valid; }

parse()

int QRegExpEngine::parse ( const QChar *  rx, int  len  )

private

Definition at line 3361 of file qregexp.cpp.

Referenced by parseAtom(), and QRegExpEngine().

{
    valid = true;
    startTokenizer(pattern, len);
    yyTok = getToken();
#ifndef QT_NO_REGEXP_CAPTURE
    yyMayCapture = true;
#else
    yyMayCapture = false;
#endif

#ifndef QT_NO_REGEXP_CAPTURE
    int atom = startAtom(false);
#endif
    QRegExpCharClass anything;
    Box box(this); // create InitialState
    box.set(anything);
    Box rightBox(this); // create FinalState
    rightBox.set(anything);

    Box middleBox(this);
    parseExpression(&middleBox);
#ifndef QT_NO_REGEXP_CAPTURE
    finishAtom(atom, false);
#endif
#ifndef QT_NO_REGEXP_OPTIM
    middleBox.setupHeuristics();
#endif
    box.cat(middleBox);
    box.cat(rightBox);
    yyCharClass.reset(0);

#ifndef QT_NO_REGEXP_CAPTURE
    for (int i = 0; i < nf; ++i) {
        switch (f[i].capture) {
        case QRegExpAtom::NoCapture:
            break;
        case QRegExpAtom::OfficialCapture:
            f[i].capture = ncap;
            captureForOfficialCapture.append(ncap);
            ++ncap;
            ++officialncap;
            break;
        case QRegExpAtom::UnofficialCapture:
            f[i].capture = greedyQuantifiers ? ncap++ : QRegExpAtom::NoCapture;
        }
    }

#ifndef QT_NO_REGEXP_BACKREF
#ifndef QT_NO_REGEXP_OPTIM
    if (officialncap == 0 && nbrefs == 0) {
        ncap = nf = 0;
        f.clear();
    }
#endif
    // handle the case where there's a \5 with no corresponding capture
    // (captureForOfficialCapture.size() != officialncap)
    for (int i = 0; i < nbrefs - officialncap; ++i) {
        captureForOfficialCapture.append(ncap);
        ++ncap;
    }
#endif
#endif

    if (!yyError.isEmpty())
        return -1;

#ifndef QT_NO_REGEXP_OPTIM
    const QRegExpAutomatonState &sinit = s.at(InitialState);
    caretAnchored = !sinit.anchors.isEmpty();
    if (caretAnchored) {
        const QMap<int, int> &anchors = sinit.anchors;
        QMap<int, int>::const_iterator a;
        for (a = anchors.constBegin(); a != anchors.constEnd(); ++a) {
            if (
#ifndef QT_NO_REGEXP_ANCHOR_ALT
                (*a & Anchor_Alternation) != 0 ||
#endif
                (*a & Anchor_Caret) == 0)
            {
                caretAnchored = false;
                break;
            }
        }
    }
#endif

    // cleanup anchors
    int numStates = s.count();
    for (int i = 0; i < numStates; ++i) {
        QRegExpAutomatonState &state = s[i];
        if (!state.anchors.isEmpty()) {
            QMap<int, int>::iterator a = state.anchors.begin();
            while (a != state.anchors.end()) {
                if (a.value() == 0)
                    a = state.anchors.erase(a);
                else
                    ++a;
            }
        }
    }

    return yyPos0;
}

parseAtom()

void QRegExpEngine::parseAtom ( Box *  box )

private

Definition at line 3466 of file qregexp.cpp.

Referenced by parseFactor().

{
#ifndef QT_NO_REGEXP_LOOKAHEAD
    QRegExpEngine *eng = 0;
    bool neg;
    int len;
#endif

    if ((yyTok & Tok_Char) != 0) {
        box->set(QChar(yyTok ^ Tok_Char));
    } else {
#ifndef QT_NO_REGEXP_OPTIM
        trivial = false;
#endif
        switch (yyTok) {
        case Tok_Dollar:
            box->catAnchor(Anchor_Dollar);
            break;
        case Tok_Caret:
            box->catAnchor(Anchor_Caret);
            break;
#ifndef QT_NO_REGEXP_LOOKAHEAD
        case Tok_PosLookahead:
        case Tok_NegLookahead:
            neg = (yyTok == Tok_NegLookahead);
            eng = new QRegExpEngine(cs, greedyQuantifiers);
            len = eng->parse(yyIn + yyPos - 1, yyLen - yyPos + 1);
            if (len >= 0)
                skipChars(len);
            else
                error(RXERR_LOOKAHEAD);
            box->catAnchor(addLookahead(eng, neg));
            yyTok = getToken();
            if (yyTok != Tok_RightParen)
                error(RXERR_LOOKAHEAD);
            break;
#endif
#ifndef QT_NO_REGEXP_ESCAPE
        case Tok_Word:
            box->catAnchor(Anchor_Word);
            break;
        case Tok_NonWord:
            box->catAnchor(Anchor_NonWord);
            break;
#endif
        case Tok_LeftParen:
        case Tok_MagicLeftParen:
            yyTok = getToken();
            parseExpression(box);
            if (yyTok != Tok_RightParen)
                error(RXERR_END);
            break;
        case Tok_CharClass:
            box->set(*yyCharClass);
            break;
        case Tok_Quantifier:
            error(RXERR_REPETITION);
            break;
        default:
#ifndef QT_NO_REGEXP_BACKREF
            if ((yyTok & Tok_BackRef) != 0)
                box->set(yyTok ^ Tok_BackRef);
            else
#endif
                error(RXERR_DISABLED);
        }
    }
    yyTok = getToken();
}

parseExpression()

void QRegExpEngine::parseExpression ( Box *  box )

private

Definition at line 3634 of file qregexp.cpp.

Referenced by parse(), and parseAtom().

{
    parseTerm(box);
    while (yyTok == Tok_Bar) {
#ifndef QT_NO_REGEXP_OPTIM
        trivial = false;
#endif
        Box rightBox(this);
        yyTok = getToken();
        parseTerm(&rightBox);
        box->orx(rightBox);
    }
}

parseFactor()

void QRegExpEngine::parseFactor ( Box *  box )

private

Definition at line 3536 of file qregexp.cpp.

Referenced by parseTerm().

{
#ifndef QT_NO_REGEXP_CAPTURE
    int outerAtom = greedyQuantifiers ? startAtom(false) : -1;
    int innerAtom = startAtom(yyMayCapture && yyTok == Tok_LeftParen);
    bool magicLeftParen = (yyTok == Tok_MagicLeftParen);
#else
    const int innerAtom = -1;
#endif

#ifndef QT_NO_REGEXP_INTERVAL
#define YYREDO() \
        yyIn = in, yyPos0 = pos0, yyPos = pos, yyLen = len, yyCh = ch, \
        *yyCharClass = charClass, yyMinRep = 0, yyMaxRep = 0, yyTok = tok

    const QChar *in = yyIn;
    int pos0 = yyPos0;
    int pos = yyPos;
    int len = yyLen;
    int ch = yyCh;
    QRegExpCharClass charClass;
    if (yyTok == Tok_CharClass)
        charClass = *yyCharClass;
    int tok = yyTok;
    bool mayCapture = yyMayCapture;
#endif

    parseAtom(box);
#ifndef QT_NO_REGEXP_CAPTURE
    finishAtom(innerAtom, magicLeftParen);
#endif

    bool hasQuantifier = (yyTok == Tok_Quantifier);
    if (hasQuantifier) {
#ifndef QT_NO_REGEXP_OPTIM
        trivial = false;
#endif
        if (yyMaxRep == InftyRep) {
            box->plus(innerAtom);
#ifndef QT_NO_REGEXP_INTERVAL
        } else if (yyMaxRep == 0) {
            box->clear();
#endif
        }
        if (yyMinRep == 0)
            box->opt();

#ifndef QT_NO_REGEXP_INTERVAL
        yyMayCapture = false;
        int alpha = (yyMinRep == 0) ? 0 : yyMinRep - 1;
        int beta = (yyMaxRep == InftyRep) ? 0 : yyMaxRep - (alpha + 1);

        Box rightBox(this);
        int i;

        for (i = 0; i < beta; i++) {
            YYREDO();
            Box leftBox(this);
            parseAtom(&leftBox);
            leftBox.cat(rightBox);
            leftBox.opt();
            rightBox = leftBox;
        }
        for (i = 0; i < alpha; i++) {
            YYREDO();
            Box leftBox(this);
            parseAtom(&leftBox);
            leftBox.cat(rightBox);
            rightBox = leftBox;
        }
        rightBox.cat(*box);
        *box = rightBox;
#endif
        yyTok = getToken();
#ifndef QT_NO_REGEXP_INTERVAL
        yyMayCapture = mayCapture;
#endif
    }
#undef YYREDO
#ifndef QT_NO_REGEXP_CAPTURE
    if (greedyQuantifiers)
        finishAtom(outerAtom, hasQuantifier);
#endif
}

parseTerm()

void QRegExpEngine::parseTerm ( Box *  box )

private

Definition at line 3621 of file qregexp.cpp.

Referenced by parseExpression().

{
#ifndef QT_NO_REGEXP_OPTIM
    if (yyTok != Tok_Eos && yyTok != Tok_RightParen && yyTok != Tok_Bar)
        parseFactor(box);
#endif
    while (yyTok != Tok_Eos && yyTok != Tok_RightParen && yyTok != Tok_Bar) {
        Box rightBox(this);
        parseFactor(&rightBox);
        box->cat(rightBox);
    }
}

setup()

void QRegExpEngine::setup ( )

private

Definition at line 1706 of file qregexp.cpp.

Referenced by QRegExpEngine().

{
    ref = 1;
#ifndef QT_NO_REGEXP_CAPTURE
    f.resize(32);
    nf = 0;
    cf = -1;
#endif
    officialncap = 0;
    ncap = 0;
#ifndef QT_NO_REGEXP_OPTIM
    caretAnchored = true;
    trivial = true;
#endif
    valid = false;
#ifndef QT_NO_REGEXP_BACKREF
    nbrefs = 0;
#endif
#ifndef QT_NO_REGEXP_OPTIM
    useGoodStringHeuristic = true;
    minl = 0;
    occ1.fill(0, NumBadChars);
#endif
}

setupCategoriesRangeMap()

void QRegExpEngine::setupCategoriesRangeMap ( )

private

Definition at line 2709 of file qregexp.cpp.

Referenced by getEscape().

{
   categoriesRangeMap.insert("IsBasicLatin",                           qMakePair(0x0000, 0x007F));
   categoriesRangeMap.insert("IsLatin-1Supplement",                    qMakePair(0x0080, 0x00FF));
   categoriesRangeMap.insert("IsLatinExtended-A",                      qMakePair(0x0100, 0x017F));
   categoriesRangeMap.insert("IsLatinExtended-B",                      qMakePair(0x0180, 0x024F));
   categoriesRangeMap.insert("IsIPAExtensions",                        qMakePair(0x0250, 0x02AF));
   categoriesRangeMap.insert("IsSpacingModifierLetters",               qMakePair(0x02B0, 0x02FF));
   categoriesRangeMap.insert("IsCombiningDiacriticalMarks",            qMakePair(0x0300, 0x036F));
   categoriesRangeMap.insert("IsGreek",                                qMakePair(0x0370, 0x03FF));
   categoriesRangeMap.insert("IsCyrillic",                             qMakePair(0x0400, 0x04FF));
   categoriesRangeMap.insert("IsCyrillicSupplement",                   qMakePair(0x0500, 0x052F));
   categoriesRangeMap.insert("IsArmenian",                             qMakePair(0x0530, 0x058F));
   categoriesRangeMap.insert("IsHebrew",                               qMakePair(0x0590, 0x05FF));
   categoriesRangeMap.insert("IsArabic",                               qMakePair(0x0600, 0x06FF));
   categoriesRangeMap.insert("IsSyriac",                               qMakePair(0x0700, 0x074F));
   categoriesRangeMap.insert("IsArabicSupplement",                     qMakePair(0x0750, 0x077F));
   categoriesRangeMap.insert("IsThaana",                               qMakePair(0x0780, 0x07BF));
   categoriesRangeMap.insert("IsDevanagari",                           qMakePair(0x0900, 0x097F));
   categoriesRangeMap.insert("IsBengali",                              qMakePair(0x0980, 0x09FF));
   categoriesRangeMap.insert("IsGurmukhi",                             qMakePair(0x0A00, 0x0A7F));
   categoriesRangeMap.insert("IsGujarati",                             qMakePair(0x0A80, 0x0AFF));
   categoriesRangeMap.insert("IsOriya",                                qMakePair(0x0B00, 0x0B7F));
   categoriesRangeMap.insert("IsTamil",                                qMakePair(0x0B80, 0x0BFF));
   categoriesRangeMap.insert("IsTelugu",                               qMakePair(0x0C00, 0x0C7F));
   categoriesRangeMap.insert("IsKannada",                              qMakePair(0x0C80, 0x0CFF));
   categoriesRangeMap.insert("IsMalayalam",                            qMakePair(0x0D00, 0x0D7F));
   categoriesRangeMap.insert("IsSinhala",                              qMakePair(0x0D80, 0x0DFF));
   categoriesRangeMap.insert("IsThai",                                 qMakePair(0x0E00, 0x0E7F));
   categoriesRangeMap.insert("IsLao",                                  qMakePair(0x0E80, 0x0EFF));
   categoriesRangeMap.insert("IsTibetan",                              qMakePair(0x0F00, 0x0FFF));
   categoriesRangeMap.insert("IsMyanmar",                              qMakePair(0x1000, 0x109F));
   categoriesRangeMap.insert("IsGeorgian",                             qMakePair(0x10A0, 0x10FF));
   categoriesRangeMap.insert("IsHangulJamo",                           qMakePair(0x1100, 0x11FF));
   categoriesRangeMap.insert("IsEthiopic",                             qMakePair(0x1200, 0x137F));
   categoriesRangeMap.insert("IsEthiopicSupplement",                   qMakePair(0x1380, 0x139F));
   categoriesRangeMap.insert("IsCherokee",                             qMakePair(0x13A0, 0x13FF));
   categoriesRangeMap.insert("IsUnifiedCanadianAboriginalSyllabics",   qMakePair(0x1400, 0x167F));
   categoriesRangeMap.insert("IsOgham",                                qMakePair(0x1680, 0x169F));
   categoriesRangeMap.insert("IsRunic",                                qMakePair(0x16A0, 0x16FF));
   categoriesRangeMap.insert("IsTagalog",                              qMakePair(0x1700, 0x171F));
   categoriesRangeMap.insert("IsHanunoo",                              qMakePair(0x1720, 0x173F));
   categoriesRangeMap.insert("IsBuhid",                                qMakePair(0x1740, 0x175F));
   categoriesRangeMap.insert("IsTagbanwa",                             qMakePair(0x1760, 0x177F));
   categoriesRangeMap.insert("IsKhmer",                                qMakePair(0x1780, 0x17FF));
   categoriesRangeMap.insert("IsMongolian",                            qMakePair(0x1800, 0x18AF));
   categoriesRangeMap.insert("IsLimbu",                                qMakePair(0x1900, 0x194F));
   categoriesRangeMap.insert("IsTaiLe",                                qMakePair(0x1950, 0x197F));
   categoriesRangeMap.insert("IsNewTaiLue",                            qMakePair(0x1980, 0x19DF));
   categoriesRangeMap.insert("IsKhmerSymbols",                         qMakePair(0x19E0, 0x19FF));
   categoriesRangeMap.insert("IsBuginese",                             qMakePair(0x1A00, 0x1A1F));
   categoriesRangeMap.insert("IsPhoneticExtensions",                   qMakePair(0x1D00, 0x1D7F));
   categoriesRangeMap.insert("IsPhoneticExtensionsSupplement",         qMakePair(0x1D80, 0x1DBF));
   categoriesRangeMap.insert("IsCombiningDiacriticalMarksSupplement",  qMakePair(0x1DC0, 0x1DFF));
   categoriesRangeMap.insert("IsLatinExtendedAdditional",              qMakePair(0x1E00, 0x1EFF));
   categoriesRangeMap.insert("IsGreekExtended",                        qMakePair(0x1F00, 0x1FFF));
   categoriesRangeMap.insert("IsGeneralPunctuation",                   qMakePair(0x2000, 0x206F));
   categoriesRangeMap.insert("IsSuperscriptsandSubscripts",            qMakePair(0x2070, 0x209F));
   categoriesRangeMap.insert("IsCurrencySymbols",                      qMakePair(0x20A0, 0x20CF));
   categoriesRangeMap.insert("IsCombiningMarksforSymbols",             qMakePair(0x20D0, 0x20FF));
   categoriesRangeMap.insert("IsLetterlikeSymbols",                    qMakePair(0x2100, 0x214F));
   categoriesRangeMap.insert("IsNumberForms",                          qMakePair(0x2150, 0x218F));
   categoriesRangeMap.insert("IsArrows",                               qMakePair(0x2190, 0x21FF));
   categoriesRangeMap.insert("IsMathematicalOperators",                qMakePair(0x2200, 0x22FF));
   categoriesRangeMap.insert("IsMiscellaneousTechnical",               qMakePair(0x2300, 0x23FF));
   categoriesRangeMap.insert("IsControlPictures",                      qMakePair(0x2400, 0x243F));
   categoriesRangeMap.insert("IsOpticalCharacterRecognition",          qMakePair(0x2440, 0x245F));
   categoriesRangeMap.insert("IsEnclosedAlphanumerics",                qMakePair(0x2460, 0x24FF));
   categoriesRangeMap.insert("IsBoxDrawing",                           qMakePair(0x2500, 0x257F));
   categoriesRangeMap.insert("IsBlockElements",                        qMakePair(0x2580, 0x259F));
   categoriesRangeMap.insert("IsGeometricShapes",                      qMakePair(0x25A0, 0x25FF));
   categoriesRangeMap.insert("IsMiscellaneousSymbols",                 qMakePair(0x2600, 0x26FF));
   categoriesRangeMap.insert("IsDingbats",                             qMakePair(0x2700, 0x27BF));
   categoriesRangeMap.insert("IsMiscellaneousMathematicalSymbols-A",   qMakePair(0x27C0, 0x27EF));
   categoriesRangeMap.insert("IsSupplementalArrows-A",                 qMakePair(0x27F0, 0x27FF));
   categoriesRangeMap.insert("IsBraillePatterns",                      qMakePair(0x2800, 0x28FF));
   categoriesRangeMap.insert("IsSupplementalArrows-B",                 qMakePair(0x2900, 0x297F));
   categoriesRangeMap.insert("IsMiscellaneousMathematicalSymbols-B",   qMakePair(0x2980, 0x29FF));
   categoriesRangeMap.insert("IsSupplementalMathematicalOperators",    qMakePair(0x2A00, 0x2AFF));
   categoriesRangeMap.insert("IsMiscellaneousSymbolsandArrows",        qMakePair(0x2B00, 0x2BFF));
   categoriesRangeMap.insert("IsGlagolitic",                           qMakePair(0x2C00, 0x2C5F));
   categoriesRangeMap.insert("IsCoptic",                               qMakePair(0x2C80, 0x2CFF));
   categoriesRangeMap.insert("IsGeorgianSupplement",                   qMakePair(0x2D00, 0x2D2F));
   categoriesRangeMap.insert("IsTifinagh",                             qMakePair(0x2D30, 0x2D7F));
   categoriesRangeMap.insert("IsEthiopicExtended",                     qMakePair(0x2D80, 0x2DDF));
   categoriesRangeMap.insert("IsSupplementalPunctuation",              qMakePair(0x2E00, 0x2E7F));
   categoriesRangeMap.insert("IsCJKRadicalsSupplement",                qMakePair(0x2E80, 0x2EFF));
   categoriesRangeMap.insert("IsKangxiRadicals",                       qMakePair(0x2F00, 0x2FDF));
   categoriesRangeMap.insert("IsIdeographicDescriptionCharacters",     qMakePair(0x2FF0, 0x2FFF));
   categoriesRangeMap.insert("IsCJKSymbolsandPunctuation",             qMakePair(0x3000, 0x303F));
   categoriesRangeMap.insert("IsHiragana",                             qMakePair(0x3040, 0x309F));
   categoriesRangeMap.insert("IsKatakana",                             qMakePair(0x30A0, 0x30FF));
   categoriesRangeMap.insert("IsBopomofo",                             qMakePair(0x3100, 0x312F));
   categoriesRangeMap.insert("IsHangulCompatibilityJamo",              qMakePair(0x3130, 0x318F));
   categoriesRangeMap.insert("IsKanbun",                               qMakePair(0x3190, 0x319F));
   categoriesRangeMap.insert("IsBopomofoExtended",                     qMakePair(0x31A0, 0x31BF));
   categoriesRangeMap.insert("IsCJKStrokes",                           qMakePair(0x31C0, 0x31EF));
   categoriesRangeMap.insert("IsKatakanaPhoneticExtensions",           qMakePair(0x31F0, 0x31FF));
   categoriesRangeMap.insert("IsEnclosedCJKLettersandMonths",          qMakePair(0x3200, 0x32FF));
   categoriesRangeMap.insert("IsCJKCompatibility",                     qMakePair(0x3300, 0x33FF));
   categoriesRangeMap.insert("IsCJKUnifiedIdeographsExtensionA",       qMakePair(0x3400, 0x4DB5));
   categoriesRangeMap.insert("IsYijingHexagramSymbols",                qMakePair(0x4DC0, 0x4DFF));
   categoriesRangeMap.insert("IsCJKUnifiedIdeographs",                 qMakePair(0x4E00, 0x9FFF));
   categoriesRangeMap.insert("IsYiSyllables",                          qMakePair(0xA000, 0xA48F));
   categoriesRangeMap.insert("IsYiRadicals",                           qMakePair(0xA490, 0xA4CF));
   categoriesRangeMap.insert("IsModifierToneLetters",                  qMakePair(0xA700, 0xA71F));
   categoriesRangeMap.insert("IsSylotiNagri",                          qMakePair(0xA800, 0xA82F));
   categoriesRangeMap.insert("IsHangulSyllables",                      qMakePair(0xAC00, 0xD7A3));
   categoriesRangeMap.insert("IsPrivateUse",                           qMakePair(0xE000, 0xF8FF));
   categoriesRangeMap.insert("IsCJKCompatibilityIdeographs",           qMakePair(0xF900, 0xFAFF));
   categoriesRangeMap.insert("IsAlphabeticPresentationForms",          qMakePair(0xFB00, 0xFB4F));
   categoriesRangeMap.insert("IsArabicPresentationForms-A",            qMakePair(0xFB50, 0xFDFF));
   categoriesRangeMap.insert("IsVariationSelectors",                   qMakePair(0xFE00, 0xFE0F));
   categoriesRangeMap.insert("IsVerticalForms",                        qMakePair(0xFE10, 0xFE1F));
   categoriesRangeMap.insert("IsCombiningHalfMarks",                   qMakePair(0xFE20, 0xFE2F));
   categoriesRangeMap.insert("IsCJKCompatibilityForms",                qMakePair(0xFE30, 0xFE4F));
   categoriesRangeMap.insert("IsSmallFormVariants",                    qMakePair(0xFE50, 0xFE6F));
   categoriesRangeMap.insert("IsArabicPresentationForms-B",            qMakePair(0xFE70, 0xFEFF));
   categoriesRangeMap.insert("IsHalfwidthandFullwidthForms",           qMakePair(0xFF00, 0xFFEF));
   categoriesRangeMap.insert("IsSpecials",                             qMakePair(0xFFF0, 0xFFFF));
   categoriesRangeMap.insert("IsLinearBSyllabary",                     qMakePair(0x10000, 0x1007F));
   categoriesRangeMap.insert("IsLinearBIdeograms",                     qMakePair(0x10080, 0x100FF));
   categoriesRangeMap.insert("IsAegeanNumbers",                        qMakePair(0x10100, 0x1013F));
   categoriesRangeMap.insert("IsAncientGreekNumbers",                  qMakePair(0x10140, 0x1018F));
   categoriesRangeMap.insert("IsOldItalic",                            qMakePair(0x10300, 0x1032F));
   categoriesRangeMap.insert("IsGothic",                               qMakePair(0x10330, 0x1034F));
   categoriesRangeMap.insert("IsUgaritic",                             qMakePair(0x10380, 0x1039F));
   categoriesRangeMap.insert("IsOldPersian",                           qMakePair(0x103A0, 0x103DF));
   categoriesRangeMap.insert("IsDeseret",                              qMakePair(0x10400, 0x1044F));
   categoriesRangeMap.insert("IsShavian",                              qMakePair(0x10450, 0x1047F));
   categoriesRangeMap.insert("IsOsmanya",                              qMakePair(0x10480, 0x104AF));
   categoriesRangeMap.insert("IsCypriotSyllabary",                     qMakePair(0x10800, 0x1083F));
   categoriesRangeMap.insert("IsKharoshthi",                           qMakePair(0x10A00, 0x10A5F));
   categoriesRangeMap.insert("IsByzantineMusicalSymbols",              qMakePair(0x1D000, 0x1D0FF));
   categoriesRangeMap.insert("IsMusicalSymbols",                       qMakePair(0x1D100, 0x1D1FF));
   categoriesRangeMap.insert("IsAncientGreekMusicalNotation",          qMakePair(0x1D200, 0x1D24F));
   categoriesRangeMap.insert("IsTaiXuanJingSymbols",                   qMakePair(0x1D300, 0x1D35F));
   categoriesRangeMap.insert("IsMathematicalAlphanumericSymbols",      qMakePair(0x1D400, 0x1D7FF));
   categoriesRangeMap.insert("IsCJKUnifiedIdeographsExtensionB",       qMakePair(0x20000, 0x2A6DF));
   categoriesRangeMap.insert("IsCJKCompatibilityIdeographsSupplement", qMakePair(0x2F800, 0x2FA1F));
   categoriesRangeMap.insert("IsTags",                                 qMakePair(0xE0000, 0xE007F));
   categoriesRangeMap.insert("IsVariationSelectorsSupplement",         qMakePair(0xE0100, 0xE01EF));
   categoriesRangeMap.insert("IsSupplementaryPrivateUseArea-A",        qMakePair(0xF0000, 0xFFFFF));
   categoriesRangeMap.insert("IsSupplementaryPrivateUseArea-B",        qMakePair(0x100000, 0x10FFFF));
}

setupState()

int QRegExpEngine::setupState ( int  match )

private

Definition at line 1731 of file qregexp.cpp.

Referenced by createState().

{
#ifndef QT_NO_REGEXP_CAPTURE
    s += QRegExpAutomatonState(cf, match);
#else
    s += QRegExpAutomatonState(match);
#endif
    return s.size() - 1;
}

skipChars()

void QRegExpEngine::skipChars ( int  n )

private

Definition at line 3172 of file qregexp.cpp.

Referenced by parseAtom().

{
    if (n > 0) {
        yyPos += n - 1;
        yyCh = getChar();
    }
}

startAtom()

int QRegExpEngine::startAtom ( bool  officialCapture )

private

Definition at line 1747 of file qregexp.cpp.

Referenced by parse(), and parseFactor().

{
    if ((nf & (nf + 1)) == 0 && nf + 1 >= f.size())
        f.resize((nf + 1) << 1);
    f[nf].parent = cf;
    cf = nf++;
    f[cf].capture = officialCapture ? QRegExpAtom::OfficialCapture : QRegExpAtom::NoCapture;
    return cf;
}

startTokenizer()

void QRegExpEngine::startTokenizer ( const QChar *  rx, int  len  )

private

Definition at line 3187 of file qregexp.cpp.

Referenced by parse().

{
    yyIn = rx;
    yyPos0 = 0;
    yyPos = 0;
    yyLen = len;
    yyCh = getChar();
    yyCharClass.reset(new QRegExpCharClass);
    yyMinRep = 0;
    yyMaxRep = 0;
    yyError = QString();
}

Friends and Related Functions

Box

friend class Box

friend

Definition at line 1254 of file qregexp.cpp.

QRegExpMatchState

friend struct QRegExpMatchState

friend

Definition at line 1299 of file qregexp.cpp.

Properties

aa

QVector<QRegExpAnchorAlternation> QRegExpEngine::aa

private

Definition at line 1169 of file qregexp.cpp.

Referenced by anchorAlternation(), anchorConcatenation(), and dump().

ahead

QVector<QRegExpLookahead *> QRegExpEngine::ahead

private

Definition at line 1166 of file qregexp.cpp.

Referenced by addLookahead(), QRegExpMatchState::testAnchor(), and ~QRegExpEngine().

captureForOfficialCapture

QVector<int> QRegExpEngine::captureForOfficialCapture

private

Definition at line 1158 of file qregexp.cpp.

Referenced by dump(), and parse().

caretAnchored

bool QRegExpEngine::caretAnchored

private

Definition at line 1172 of file qregexp.cpp.

Referenced by parse(), and setup().

categoriesRangeMap

QHash<QByteArray, QPair<int, int> > QRegExpEngine::categoriesRangeMap

private

Definition at line 1297 of file qregexp.cpp.

Referenced by getEscape(), and setupCategoriesRangeMap().

cf

int QRegExpEngine::cf

private

Definition at line 1157 of file qregexp.cpp.

Referenced by finishAtom(), setup(), setupState(), and startAtom().

cl

QVector<QRegExpCharClass> QRegExpEngine::cl

private

Definition at line 1163 of file qregexp.cpp.

Referenced by createState(), and dump().

cs

Qt::CaseSensitivity QRegExpEngine::cs

private

Definition at line 1176 of file qregexp.cpp.

Referenced by dump(), goodStringMatch(), QRegExp::operator>>(), parseAtom(), QRegExp::setCaseSensitivity(), and QRegExpEngine::Box::setupHeuristics().

f

QVector<QRegExpAtom> QRegExpEngine::f

private

Definition at line 1155 of file qregexp.cpp.

Referenced by addPlusTransitions(), dump(), finishAtom(), parse(), setup(), and startAtom().

goodEarlyStart

int QRegExpEngine::goodEarlyStart

private

Definition at line 1186 of file qregexp.cpp.

Referenced by goodStringMatch(), heuristicallyChooseHeuristic(), and QRegExpEngine::Box::setupHeuristics().

goodLateStart

int QRegExpEngine::goodLateStart

private

Definition at line 1187 of file qregexp.cpp.

Referenced by goodStringMatch(), heuristicallyChooseHeuristic(), and QRegExpEngine::Box::setupHeuristics().

goodStr

QString QRegExpEngine::goodStr

private

Definition at line 1188 of file qregexp.cpp.

Referenced by goodStringMatch(), heuristicallyChooseHeuristic(), and QRegExpEngine::Box::setupHeuristics().

greedyQuantifiers

bool QRegExpEngine::greedyQuantifiers

private

Definition at line 1177 of file qregexp.cpp.

Referenced by finishAtom(), parse(), parseAtom(), and parseFactor().

minl

int QRegExpEngine::minl

private

Definition at line 1190 of file qregexp.cpp.

Referenced by badCharMatch(), heuristicallyChooseHeuristic(), QRegExpMatchState::match(), QRegExpMatchState::prepareForMatch(), setup(), and QRegExpEngine::Box::setupHeuristics().

nbrefs

int QRegExpEngine::nbrefs

private

Definition at line 1180 of file qregexp.cpp.

Referenced by createState(), parse(), and setup().

ncap

int QRegExpEngine::ncap

private

Definition at line 1161 of file qregexp.cpp.

Referenced by isBetterCapture(), QRegExpMatchState::matchHere(), parse(), QRegExpMatchState::prepareForMatch(), and setup().

nf

int QRegExpEngine::nf

private

Definition at line 1156 of file qregexp.cpp.

Referenced by dump(), parse(), setup(), and startAtom().

occ1

QVector<int> QRegExpEngine::occ1

private

Definition at line 1191 of file qregexp.cpp.

Referenced by badCharMatch(), heuristicallyChooseHeuristic(), setup(), and QRegExpEngine::Box::setupHeuristics().

officialncap

int QRegExpEngine::officialncap

private

Definition at line 1160 of file qregexp.cpp.

Referenced by parse(), and setup().

ref

QAtomicInt QRegExpEngine::ref

Definition at line 1115 of file qregexp.cpp.

Referenced by derefEngine(), QRegExp::operator=(), and prepareEngine_helper().

s

QVector<QRegExpAutomatonState> QRegExpEngine::s

private

Definition at line 1153 of file qregexp.cpp.

Referenced by addAnchors(), addCatTransitions(), addPlusTransitions(), dump(), parse(), QRegExpMatchState::prepareForMatch(), and setupState().

trivial

bool QRegExpEngine::trivial

private

Definition at line 1173 of file qregexp.cpp.

Referenced by heuristicallyChooseHeuristic(), parseAtom(), parseExpression(), parseFactor(), QRegExpEngine(), and setup().

useGoodStringHeuristic

bool QRegExpEngine::useGoodStringHeuristic

private

Definition at line 1184 of file qregexp.cpp.

Referenced by heuristicallyChooseHeuristic(), and setup().

valid

bool QRegExpEngine::valid

private

Definition at line 1175 of file qregexp.cpp.

Referenced by parse(), QRegExpEngine(), and setup().

xmlSchemaExtensions

bool QRegExpEngine::xmlSchemaExtensions

private

Definition at line 1178 of file qregexp.cpp.

Referenced by getEscape().

yyCh

int QRegExpEngine::yyCh

private

Definition at line 1280 of file qregexp.cpp.

Referenced by getEscape(), getRep(), getToken(), parseFactor(), skipChars(), and startTokenizer().

yyCharClass

QScopedPointer<QRegExpCharClass> QRegExpEngine::yyCharClass

private

Definition at line 1281 of file qregexp.cpp.

Referenced by getEscape(), getToken(), parse(), parseAtom(), parseFactor(), and startTokenizer().

yyError

QString QRegExpEngine::yyError

private

Definition at line 1284 of file qregexp.cpp.

Referenced by error(), parse(), and startTokenizer().

yyIn

const QChar* QRegExpEngine::yyIn

private

Definition at line 1276 of file qregexp.cpp.

Referenced by getChar(), parseAtom(), parseFactor(), and startTokenizer().

yyLen

int QRegExpEngine::yyLen

private

Definition at line 1279 of file qregexp.cpp.

Referenced by getChar(), parseAtom(), parseFactor(), and startTokenizer().

yyMaxRep

int QRegExpEngine::yyMaxRep

private

Definition at line 1283 of file qregexp.cpp.

Referenced by getToken(), parseFactor(), and startTokenizer().

yyMayCapture

bool QRegExpEngine::yyMayCapture

private

Definition at line 1296 of file qregexp.cpp.

Referenced by parse(), and parseFactor().

yyMinRep

int QRegExpEngine::yyMinRep

private

Definition at line 1282 of file qregexp.cpp.

Referenced by getToken(), parseFactor(), and startTokenizer().

yyPos

int QRegExpEngine::yyPos

private

Definition at line 1278 of file qregexp.cpp.

Referenced by getChar(), getToken(), parseAtom(), parseFactor(), skipChars(), and startTokenizer().

yyPos0

int QRegExpEngine::yyPos0

private

Definition at line 1277 of file qregexp.cpp.

Referenced by getToken(), parse(), parseFactor(), and startTokenizer().

yyTok

int QRegExpEngine::yyTok

private

Definition at line 1295 of file qregexp.cpp.

Referenced by parse(), parseAtom(), parseExpression(), parseFactor(), and parseTerm().

---

The documentation for this class was generated from the following file:

• /src/corelib/tools/qregexp.cpp