QRegExp Class Reference

The QRegExp class provides pattern matching using regular expressions or wildcards. More...

#include <qregexp.h>

Public Member Functions
	QRegExp ()
	QRegExp (const QCString &, bool caseSensitive=TRUE, bool wildcard=FALSE)
	QRegExp (const QRegExp &)
	~QRegExp ()
QRegExp &	operator= (const QRegExp &)
QRegExp &	operator= (const QCString &pattern)
bool	operator== (const QRegExp &) const
bool	operator!= (const QRegExp &r) const
bool	isEmpty () const
bool	isValid () const
bool	caseSensitive () const
void	setCaseSensitive (bool)
bool	wildcard () const
void	setWildcard (bool)
QCString	pattern () const
void	setPattern (const QCString &pattern)
int	match (const QCString &str, int index=0, int *len=0, bool indexIsStart=TRUE) const
int	find (const QCString &str, int index)

Protected Member Functions
void	compile ()
const char *	matchstr (uint *, const char *, uint, const char *) const

Private Attributes
QCString	rxstring
uint *	rxdata
int	error
bool	cs
bool	wc

Detailed Description

The QRegExp class provides pattern matching using regular expressions or wildcards.

QRegExp knows these regexp primitives:

• c matches the character 'c'
• . matches any character
• ^ matches start of input
• $ matches end of input
• [] matches a defined set of characters - see below.
• a* matches a sequence of zero or more a's
• a+ matches a sequence of one or more a's
• a? matches an optional a
• escape code for matching special characters such as \, [, *, +, . etc.
• matches the TAB character (9)
• 
matches newline (10)
• matches return (13)
• matches a white space (defined as any character for which QChar::isSpace() returns TRUE. This includes at least ASCII characters 9 (TAB), 10 (LF), 11 (VT), 12(FF), 13 (CR) and 32 (Space)).
• matches a digit (defined as any character for which QChar::isDigit() returns TRUE. This includes at least ASCII characters '0'-'9').
• matches the character with unicode point U1f6b (hexadecimal 1f6b). will match the ASCII/Latin1 character 0x12 (18 decimal, 12 hexadecimal).
• \022 matches the ASCII/Latin1 character 022 (18 decimal, 22 octal).

In wildcard mode, it only knows four primitives:

• c matches the character 'c'
• ? matches any character
• * matches any sequence of characters
• [] matches a defined set of characters - see below.

QRegExp supports Unicode both in the pattern strings and in the strings to be matched.

When writing regular expressions in C++ code, remember that C++ processes \ characters. So in order to match e.g. a "." character, you must write "\\." in C++ source, not "\.".

A character set matches a defined set of characters. For example, [BSD] matches any of 'B', 'D' and 'S'. Within a character set, the special characters '.', '*', '?', '^', '$', '+' and '[' lose their special meanings. The following special characters apply:

• ^ When placed first in the list, changes the character set to match any character not in the list. To include the character '^' itself in the set, escape it or place it anywhere but first.
• - Defines a range of characters. To include the character '-' itself in the set, escape it or place it last.
• ] Ends the character set definition. To include the character ']' itself in the set, escape it or place it first (but after the negation operator '^', if present)

Thus, [a-zA-Z0-9.] matches upper and lower case ASCII letters, digits and dot; and [^] matches everything except white space.

Bug:

Case insensitive matching is not supported for non-ASCII/Latin1 (non-8bit) characters. Any character with a non-zero QChar.row() is matched case sensitively even if the QRegExp is in case insensitive mode.

Note

In Qt 3.0, the language of regular expressions will contain five more special characters, namely '(', ')', '{', '|' and '}'. To ease porting, it's a good idea to escape these characters with a backslash in all the regular expressions you'll write from now on.

Definition at line 46 of file qregexp.h.

Constructor & Destructor Documentation

QRegExp::QRegExp

(

)

Constructs an empty regular expression.

Definition at line 174 of file qregexp.cpp.

{
    rxdata = 0;
    cs = TRUE;
    wc = FALSE;
    error = PatOk;
}

QRegExp::QRegExp	(	const QCString &	pattern,
		bool	caseSensitive = TRUE,
		bool	wildcard = FALSE
	)

Constructs a regular expression.

• pattern is the regular expression pattern string.
• caseSensitive specifies whether or not to use case sensitive matching.
• wildcard specifies whether the pattern string should be used for wildcard matching (also called globbing expression), normally used for matching file names.

See also

setWildcard()

Definition at line 195 of file qregexp.cpp.

{
    rxstring = pattern;
    rxdata = 0;
    cs = caseSensitive;
    wc = wildcard;
    compile();
}

QRegExp::QRegExp

(

const QRegExp &

r

)

Constructs a regular expression which is a copy of r.

See also

operator=(const QRegExp&)

Definition at line 209 of file qregexp.cpp.

{
    rxstring = r.pattern();
    rxdata = 0;
    cs = r.caseSensitive();
    wc = r.wildcard();
    compile();
}

QRegExp::~QRegExp

(

)

Destructs the regular expression and cleans up its internal data.

Definition at line 222 of file qregexp.cpp.

{
    if ( rxdata )                      // Avoid purify complaints
        delete [] rxdata;
}

Member Function Documentation

bool QRegExp::caseSensitive ( ) const

inline

Returns TRUE if case sensitivity is enabled, otherwise FALSE. The default is TRUE.

See also

setCaseSensitive()

Definition at line 63 of file qregexp.h.

{ return cs; }

void QRegExp::compile ( )

protected

Definition at line 914 of file qregexp.cpp.

{
    if ( rxdata ) {                             // delete old data
        delete [] rxdata;
        rxdata = 0;
    }
    if ( rxstring.isEmpty() ) {                 // no regexp pattern set
        error = PatNull;
        return;
    }

    error = PatOk;                              // assume pattern is ok

    QCString pattern;
    if ( wc )
        pattern = wc2rx(rxstring);
    else
        pattern = rxstring;
    const char *start = pattern.data();         // pattern pointer
    const char *p = start;                      // pattern pointer
    uint pl = pattern.length();
    uint *d = rxarray;                          // data pointer
    uint *prev_d = 0;

#define GEN(x)  *d++ = (x)

    while ( pl ) {
        char ch = (char)*p;
        switch ( ch ) {

            case '^':                           // beginning of line
                prev_d = d;
                GEN( p == start ? BOL : (CHR | ch) );
                p++;
                pl--;
                break;

            case '$':                           // end of line
                prev_d = d;
                GEN( pl == 1 ? EOL : (CHR | ch) );
                p++;
                pl--;
                break;

            case '.':                           // any char
                prev_d = d;
                GEN( ANY );
                p++;
                pl--;
                break;

            case '[':                           // character class
                {
                prev_d = d;
                p++;
                pl--;
                if ( !pl ) {
                    error = PatSyntax;
                    return;
                }
                bool firstIsEscaped = ( (char)*p == '\\' );
                uint cch = char_val( &p, &pl );
                if ( cch == '^' && !firstIsEscaped ) {  // negate!
                    GEN( CCN );
                    if ( !pl ) {
                        error = PatSyntax;
                        return;
                    }
                    cch = char_val( &p, &pl );
                } else {
                    GEN( CCL );
                }
                uint numFields = 0;
                while ( pl ) {
                    if ((pl>2) && ((char)*p == '-') && ((char)*(p+1) != ']')) {
                        // Found a range
                        char_val( &p, &pl ); // Read the '-'
                        uint cch2 = char_val( &p, &pl ); // Read the range end
                        if ( cch > cch2 ) {             // swap start and stop
                            int tmp = cch;
                            cch = cch2;
                            cch2 = tmp;
                        }
                        GEN( (cch << 16) | cch2 );      // from < to
                        numFields++;
                    }
                    else {
                        // Found a single character
                        if ( cch & MCD ) // It's a code; will not be mistaken
                            GEN( cch );  // for a range, since from > to
                        else
                            GEN( (cch << 16) | cch ); // from == to range
                        numFields++;
                    }
                    if ( d >= rxarray + maxlen ) {      // pattern too long
                        error = PatOverflow;            
                        return;
                    }
                    if ( !pl ) {                // At least ']' should be left
                        error = PatSyntax;
                        return;
                    }
                    bool nextIsEscaped = ( (char)*p == '\\' );
                    cch = char_val( &p, &pl );
                    if ( cch == (uint)']' && !nextIsEscaped )
                        break;
                    if ( !pl ) {                // End, should have seen ']'
                        error = PatSyntax;
                        return;
                    }
                }
                *prev_d |= numFields;           // Store number of fields
                }
                break;

            case '*':                           // Kleene closure, or
            case '+':                           // positive closure, or
            case '?':                           // optional closure
                {
                if ( prev_d == 0 ) {            // no previous expression
                    error = PatSyntax;          // empty closure
                    return;
                }
                switch ( *prev_d ) {            // test if invalid closure
                    case BOL:
                    case BOW:
                    case EOW:
                    case CLO:
                    case OPT:
                        error = PatSyntax;
                        return;
                }
                int ddiff = (int)(d - prev_d);
                if ( *p == '+' ) {              // convert to Kleene closure
                    if ( d + ddiff >= rxarray + maxlen ) {
                        error = PatOverflow;    // pattern too long
                        return;
                    }
                    memcpy( d, prev_d, ddiff*sizeof(uint) );
                    d += ddiff;
                    prev_d += ddiff;
                }
                memmove( prev_d+1, prev_d, ddiff*sizeof(uint) );
                *prev_d = ch == '?' ? OPT : CLO;
                d++;
                GEN( END );
                p++;
                pl--;
                }
                break;

            default:
                {
                prev_d = d;
                uint cv = char_val( &p, &pl );
                if ( cv & MCD ) {                       // It's a code
                    GEN( cv );
                }
                else {
                    if ( !cs && cv <= 0xff )            // #only 8bit support
                        cv = tolower( cv );
                    GEN( CHR | cv );
                }
                }
        }
        if ( d >= rxarray + maxlen ) {          // oops!
            error = PatOverflow;                // pattern too long
            return;
        }
    }
    GEN( END );
    int len = (int)(d - rxarray);
    rxdata = new uint[ len ];                   // copy from rxarray to rxdata
    CHECK_PTR( rxdata );
    memcpy( rxdata, rxarray, len*sizeof(uint) );
#if defined(DEBUG)
    //dump( rxdata );   // uncomment this line for debugging
#endif
}

int QRegExp::find ( const QCString &  str, int  index  )

inline

Attempts to match in str, starting from position index. Returns the position of the match, or -1 if there was no match.

See also

match()

Definition at line 76 of file qregexp.h.

                                        { return match( str, index ); }

bool QRegExp::isEmpty ( ) const

inline

Returns TRUE if the regexp is empty.

Definition at line 60 of file qregexp.h.

{ return rxdata == 0; }

bool QRegExp::isValid ( ) const

inline

Returns TRUE if the regexp is valid, or FALSE if it is invalid.

The pattern "[a-z" is an example of an invalid pattern, since it lacks a closing bracket.

Definition at line 61 of file qregexp.h.

{ return error == 0; }

int QRegExp::match	(	const QCString &	str,
		int	index = 0,
		int *	len = 0,
		bool	indexIsStart = TRUE
	)		const

Attempts to match in str, starting from position index. Returns the position of the match, or -1 if there was no match.

If len is not a null pointer, the length of the match is stored in *len.

If indexIsStart is TRUE (the default), the position index in the string will match the start-of-input primitive (^) in the regexp, if present. Otherwise, position 0 in str will match.

Example:

QRegExp r("[0-9]*\\.[0-9]+");         // matches floating point
int len;
r.match("pi = 3.1416", 0, &len);              // returns 5, len == 6

Note

In Qt 3.0, this function will be replaced by find().

Definition at line 649 of file qregexp.cpp.

{
    if ( !isValid() || isEmpty() )
        return -1;
    if ( str.length() < (uint)index )
        return -1;
    const char *start = str.data();
    const char *p = start + index;
    uint pl = str.length() - index;
    uint *d  = rxdata;
    int ep = -1;

    if ( *d == BOL ) {                          // match from beginning of line
        ep = matchstring( d, p, pl, indexIsStart ? p : start, cs );
    } else {
        if ( *d & CHR ) {
            char c = *d;
            if ( !cs /*&& !c.row()*/ ) {                // case sensitive, # only 8bit
                while ( pl && ( /*p->row() ||*/ tolower(*p) != c ) ) {
                    p++;
                    pl--;
                }
            } else {                            // case insensitive
                while ( pl && *p != c ) {
                    p++;
                    pl--;
                }
            }
        }
        while( 1 ) {                            // regular match
            ep = matchstring( d, p, pl, indexIsStart ? start+index : start, cs );
            if ( ep >= 0 )
                break;
            if ( !pl )
                break;
            p++;
            pl--;
        }
    }
    if ( len )
        *len = ep >= 0 ? ep : 0;      // No match -> 0, for historical reasons
    return ep >= 0 ? (int)(p - start) : -1;             // return index;
}

const char* QRegExp::matchstr ( uint *  , const char *  , uint  , const char *    ) const

protected

bool QRegExp::operator!= ( const QRegExp &  r ) const

inline

Returns TRUE if this regexp is not equal to r.

See also

operator==()

Definition at line 57 of file qregexp.h.

                                        { return !(this->operator==(r)); }

QRegExp & QRegExp::operator=

(

const QRegExp &

r

)

Copies the regexp r and returns a reference to this regexp. The case sensitivity and wildcard options are copied, as well.

Definition at line 233 of file qregexp.cpp.

{
    rxstring = r.rxstring;
    cs = r.cs;
    wc = r.wc;
    compile();
    return *this;
}

QRegExp & QRegExp::operator=

(

const QCString &

pattern

)

Consider using setPattern() instead of this method.

Sets the pattern string to pattern and returns a reference to this regexp. The case sensitivity or wildcard options do not change.

Definition at line 250 of file qregexp.cpp.

{
    rxstring = pattern;
    compile();
    return *this;
}

bool QRegExp::operator==

(

const QRegExp &

r

)

const

Returns TRUE if this regexp is equal to r.

Two regexp objects are equal if they have equal pattern strings, case sensitivity options and wildcard options.

Definition at line 265 of file qregexp.cpp.

{
    return rxstring == r.rxstring && cs == r.cs && wc == r.wc;
}

QCString QRegExp::pattern ( ) const

inline

Returns the pattern string of the regexp.

Definition at line 69 of file qregexp.h.

{ return rxstring; }

void QRegExp::setCaseSensitive

(

bool

enable

)

Enables or disables case sensitive matching.

In case sensitive mode, "a.e" matches "axe" but not "Axe".

See also: caseSensitive()

Definition at line 335 of file qregexp.cpp.

{
    if ( cs != enable ) {
        cs = enable;
        compile();
    }
}

void QRegExp::setPattern ( const QCString &  pattern )

inline

Sets the pattern string to pattern and returns a reference to this regexp. The case sensitivity or wildcard options do not change.

Definition at line 71 of file qregexp.h.

                                        { operator=( pattern ); }

void QRegExp::setWildcard

(

bool

wildcard

)

Sets the wildcard option for the regular expression. The default is FALSE.

Setting wildcard to TRUE makes it convenient to match filenames instead of plain text.

For example, "qr*.cpp" matches the string "qregexp.cpp" in wildcard mode, but not "qicpp" (which would be matched in normal mode).

See also

wildcard()

Definition at line 310 of file qregexp.cpp.

{
    if ( wildcard != wc ) {
        wc = wildcard;
        compile();
    }
}

bool QRegExp::wildcard ( ) const

inline

Returns TRUE if wildcard mode is on, otherwise FALSE.

See also

setWildcard().

Definition at line 66 of file qregexp.h.

{ return wc; }

Member Data Documentation

bool QRegExp::cs

private

Definition at line 87 of file qregexp.h.

int QRegExp::error

private

Definition at line 86 of file qregexp.h.

uint* QRegExp::rxdata

private

Definition at line 85 of file qregexp.h.

QCString QRegExp::rxstring

private

Definition at line 84 of file qregexp.h.

bool QRegExp::wc

private

Definition at line 88 of file qregexp.h.

---

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

• doxygen-1.8.11/qtools/qregexp.h
• doxygen-1.8.11/qtools/qregexp.cpp