libraries/spcre/libpcre/pcre/doc/html/pcrepattern.html
changeset 0 7f656887cf89
equal deleted inserted replaced
-1:000000000000 0:7f656887cf89
       
     1 <html>
       
     2 <head>
       
     3 <title>pcrepattern specification</title>
       
     4 </head>
       
     5 <body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB">
       
     6 <h1>pcrepattern man page</h1>
       
     7 <p>
       
     8 Return to the <a href="index.html">PCRE index page</a>.
       
     9 </p>
       
    10 <p>
       
    11 This page is part of the PCRE HTML documentation. It was generated automatically
       
    12 from the original man page. If there is any nonsense in it, please consult the
       
    13 man page, in case the conversion went wrong.
       
    14 <br>
       
    15 <ul>
       
    16 <li><a name="TOC1" href="#SEC1">PCRE REGULAR EXPRESSION DETAILS</a>
       
    17 <li><a name="TOC2" href="#SEC2">NEWLINE CONVENTIONS</a>
       
    18 <li><a name="TOC3" href="#SEC3">CHARACTERS AND METACHARACTERS</a>
       
    19 <li><a name="TOC4" href="#SEC4">BACKSLASH</a>
       
    20 <li><a name="TOC5" href="#SEC5">CIRCUMFLEX AND DOLLAR</a>
       
    21 <li><a name="TOC6" href="#SEC6">FULL STOP (PERIOD, DOT)</a>
       
    22 <li><a name="TOC7" href="#SEC7">MATCHING A SINGLE BYTE</a>
       
    23 <li><a name="TOC8" href="#SEC8">SQUARE BRACKETS AND CHARACTER CLASSES</a>
       
    24 <li><a name="TOC9" href="#SEC9">POSIX CHARACTER CLASSES</a>
       
    25 <li><a name="TOC10" href="#SEC10">VERTICAL BAR</a>
       
    26 <li><a name="TOC11" href="#SEC11">INTERNAL OPTION SETTING</a>
       
    27 <li><a name="TOC12" href="#SEC12">SUBPATTERNS</a>
       
    28 <li><a name="TOC13" href="#SEC13">DUPLICATE SUBPATTERN NUMBERS</a>
       
    29 <li><a name="TOC14" href="#SEC14">NAMED SUBPATTERNS</a>
       
    30 <li><a name="TOC15" href="#SEC15">REPETITION</a>
       
    31 <li><a name="TOC16" href="#SEC16">ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS</a>
       
    32 <li><a name="TOC17" href="#SEC17">BACK REFERENCES</a>
       
    33 <li><a name="TOC18" href="#SEC18">ASSERTIONS</a>
       
    34 <li><a name="TOC19" href="#SEC19">CONDITIONAL SUBPATTERNS</a>
       
    35 <li><a name="TOC20" href="#SEC20">COMMENTS</a>
       
    36 <li><a name="TOC21" href="#SEC21">RECURSIVE PATTERNS</a>
       
    37 <li><a name="TOC22" href="#SEC22">SUBPATTERNS AS SUBROUTINES</a>
       
    38 <li><a name="TOC23" href="#SEC23">ONIGURUMA SUBROUTINE SYNTAX</a>
       
    39 <li><a name="TOC24" href="#SEC24">CALLOUTS</a>
       
    40 <li><a name="TOC25" href="#SEC25">BACKTRACKING CONTROL</a>
       
    41 <li><a name="TOC26" href="#SEC26">SEE ALSO</a>
       
    42 <li><a name="TOC27" href="#SEC27">AUTHOR</a>
       
    43 <li><a name="TOC28" href="#SEC28">REVISION</a>
       
    44 </ul>
       
    45 <br><a name="SEC1" href="#TOC1">PCRE REGULAR EXPRESSION DETAILS</a><br>
       
    46 <P>
       
    47 The syntax and semantics of the regular expressions that are supported by PCRE
       
    48 are described in detail below. There is a quick-reference syntax summary in the
       
    49 <a href="pcresyntax.html"><b>pcresyntax</b></a>
       
    50 page. PCRE tries to match Perl syntax and semantics as closely as it can. PCRE
       
    51 also supports some alternative regular expression syntax (which does not
       
    52 conflict with the Perl syntax) in order to provide some compatibility with
       
    53 regular expressions in Python, .NET, and Oniguruma.
       
    54 </P>
       
    55 <P>
       
    56 Perl's regular expressions are described in its own documentation, and
       
    57 regular expressions in general are covered in a number of books, some of which
       
    58 have copious examples. Jeffrey Friedl's "Mastering Regular Expressions",
       
    59 published by O'Reilly, covers regular expressions in great detail. This
       
    60 description of PCRE's regular expressions is intended as reference material.
       
    61 </P>
       
    62 <P>
       
    63 The original operation of PCRE was on strings of one-byte characters. However,
       
    64 there is now also support for UTF-8 character strings. To use this, you must
       
    65 build PCRE to include UTF-8 support, and then call <b>pcre_compile()</b> with
       
    66 the PCRE_UTF8 option. How this affects pattern matching is mentioned in several
       
    67 places below. There is also a summary of UTF-8 features in the
       
    68 <a href="pcre.html#utf8support">section on UTF-8 support</a>
       
    69 in the main
       
    70 <a href="pcre.html"><b>pcre</b></a>
       
    71 page.
       
    72 </P>
       
    73 <P>
       
    74 The remainder of this document discusses the patterns that are supported by
       
    75 PCRE when its main matching function, <b>pcre_exec()</b>, is used.
       
    76 From release 6.0, PCRE offers a second matching function,
       
    77 <b>pcre_dfa_exec()</b>, which matches using a different algorithm that is not
       
    78 Perl-compatible. Some of the features discussed below are not available when
       
    79 <b>pcre_dfa_exec()</b> is used. The advantages and disadvantages of the
       
    80 alternative function, and how it differs from the normal function, are
       
    81 discussed in the
       
    82 <a href="pcrematching.html"><b>pcrematching</b></a>
       
    83 page.
       
    84 </P>
       
    85 <br><a name="SEC2" href="#TOC1">NEWLINE CONVENTIONS</a><br>
       
    86 <P>
       
    87 PCRE supports five different conventions for indicating line breaks in
       
    88 strings: a single CR (carriage return) character, a single LF (linefeed)
       
    89 character, the two-character sequence CRLF, any of the three preceding, or any
       
    90 Unicode newline sequence. The
       
    91 <a href="pcreapi.html"><b>pcreapi</b></a>
       
    92 page has
       
    93 <a href="pcreapi.html#newlines">further discussion</a>
       
    94 about newlines, and shows how to set the newline convention in the
       
    95 <i>options</i> arguments for the compiling and matching functions.
       
    96 </P>
       
    97 <P>
       
    98 It is also possible to specify a newline convention by starting a pattern
       
    99 string with one of the following five sequences:
       
   100 <pre>
       
   101   (*CR)        carriage return
       
   102   (*LF)        linefeed
       
   103   (*CRLF)      carriage return, followed by linefeed
       
   104   (*ANYCRLF)   any of the three above
       
   105   (*ANY)       all Unicode newline sequences
       
   106 </pre>
       
   107 These override the default and the options given to <b>pcre_compile()</b>. For
       
   108 example, on a Unix system where LF is the default newline sequence, the pattern
       
   109 <pre>
       
   110   (*CR)a.b
       
   111 </pre>
       
   112 changes the convention to CR. That pattern matches "a\nb" because LF is no
       
   113 longer a newline. Note that these special settings, which are not
       
   114 Perl-compatible, are recognized only at the very start of a pattern, and that
       
   115 they must be in upper case. If more than one of them is present, the last one
       
   116 is used.
       
   117 </P>
       
   118 <P>
       
   119 The newline convention does not affect what the \R escape sequence matches. By
       
   120 default, this is any Unicode newline sequence, for Perl compatibility. However,
       
   121 this can be changed; see the description of \R in the section entitled
       
   122 <a href="#newlineseq">"Newline sequences"</a>
       
   123 below. A change of \R setting can be combined with a change of newline
       
   124 convention.
       
   125 </P>
       
   126 <br><a name="SEC3" href="#TOC1">CHARACTERS AND METACHARACTERS</a><br>
       
   127 <P>
       
   128 A regular expression is a pattern that is matched against a subject string from
       
   129 left to right. Most characters stand for themselves in a pattern, and match the
       
   130 corresponding characters in the subject. As a trivial example, the pattern
       
   131 <pre>
       
   132   The quick brown fox
       
   133 </pre>
       
   134 matches a portion of a subject string that is identical to itself. When
       
   135 caseless matching is specified (the PCRE_CASELESS option), letters are matched
       
   136 independently of case. In UTF-8 mode, PCRE always understands the concept of
       
   137 case for characters whose values are less than 128, so caseless matching is
       
   138 always possible. For characters with higher values, the concept of case is
       
   139 supported if PCRE is compiled with Unicode property support, but not otherwise.
       
   140 If you want to use caseless matching for characters 128 and above, you must
       
   141 ensure that PCRE is compiled with Unicode property support as well as with
       
   142 UTF-8 support.
       
   143 </P>
       
   144 <P>
       
   145 The power of regular expressions comes from the ability to include alternatives
       
   146 and repetitions in the pattern. These are encoded in the pattern by the use of
       
   147 <i>metacharacters</i>, which do not stand for themselves but instead are
       
   148 interpreted in some special way.
       
   149 </P>
       
   150 <P>
       
   151 There are two different sets of metacharacters: those that are recognized
       
   152 anywhere in the pattern except within square brackets, and those that are
       
   153 recognized within square brackets. Outside square brackets, the metacharacters
       
   154 are as follows:
       
   155 <pre>
       
   156   \      general escape character with several uses
       
   157   ^      assert start of string (or line, in multiline mode)
       
   158   $      assert end of string (or line, in multiline mode)
       
   159   .      match any character except newline (by default)
       
   160   [      start character class definition
       
   161   |      start of alternative branch
       
   162   (      start subpattern
       
   163   )      end subpattern
       
   164   ?      extends the meaning of (
       
   165          also 0 or 1 quantifier
       
   166          also quantifier minimizer
       
   167   *      0 or more quantifier
       
   168   +      1 or more quantifier
       
   169          also "possessive quantifier"
       
   170   {      start min/max quantifier
       
   171 </pre>
       
   172 Part of a pattern that is in square brackets is called a "character class". In
       
   173 a character class the only metacharacters are:
       
   174 <pre>
       
   175   \      general escape character
       
   176   ^      negate the class, but only if the first character
       
   177   -      indicates character range
       
   178   [      POSIX character class (only if followed by POSIX syntax)
       
   179   ]      terminates the character class
       
   180 </pre>
       
   181 The following sections describe the use of each of the metacharacters.
       
   182 </P>
       
   183 <br><a name="SEC4" href="#TOC1">BACKSLASH</a><br>
       
   184 <P>
       
   185 The backslash character has several uses. Firstly, if it is followed by a
       
   186 non-alphanumeric character, it takes away any special meaning that character
       
   187 may have. This use of backslash as an escape character applies both inside and
       
   188 outside character classes.
       
   189 </P>
       
   190 <P>
       
   191 For example, if you want to match a * character, you write \* in the pattern.
       
   192 This escaping action applies whether or not the following character would
       
   193 otherwise be interpreted as a metacharacter, so it is always safe to precede a
       
   194 non-alphanumeric with backslash to specify that it stands for itself. In
       
   195 particular, if you want to match a backslash, you write \\.
       
   196 </P>
       
   197 <P>
       
   198 If a pattern is compiled with the PCRE_EXTENDED option, whitespace in the
       
   199 pattern (other than in a character class) and characters between a # outside
       
   200 a character class and the next newline are ignored. An escaping backslash can
       
   201 be used to include a whitespace or # character as part of the pattern.
       
   202 </P>
       
   203 <P>
       
   204 If you want to remove the special meaning from a sequence of characters, you
       
   205 can do so by putting them between \Q and \E. This is different from Perl in
       
   206 that $ and @ are handled as literals in \Q...\E sequences in PCRE, whereas in
       
   207 Perl, $ and @ cause variable interpolation. Note the following examples:
       
   208 <pre>
       
   209   Pattern            PCRE matches   Perl matches
       
   210 
       
   211   \Qabc$xyz\E        abc$xyz        abc followed by the contents of $xyz
       
   212   \Qabc\$xyz\E       abc\$xyz       abc\$xyz
       
   213   \Qabc\E\$\Qxyz\E   abc$xyz        abc$xyz
       
   214 </pre>
       
   215 The \Q...\E sequence is recognized both inside and outside character classes.
       
   216 <a name="digitsafterbackslash"></a></P>
       
   217 <br><b>
       
   218 Non-printing characters
       
   219 </b><br>
       
   220 <P>
       
   221 A second use of backslash provides a way of encoding non-printing characters
       
   222 in patterns in a visible manner. There is no restriction on the appearance of
       
   223 non-printing characters, apart from the binary zero that terminates a pattern,
       
   224 but when a pattern is being prepared by text editing, it is usually easier to
       
   225 use one of the following escape sequences than the binary character it
       
   226 represents:
       
   227 <pre>
       
   228   \a        alarm, that is, the BEL character (hex 07)
       
   229   \cx       "control-x", where x is any character
       
   230   \e        escape (hex 1B)
       
   231   \f        formfeed (hex 0C)
       
   232   \n        linefeed (hex 0A)
       
   233   \r        carriage return (hex 0D)
       
   234   \t        tab (hex 09)
       
   235   \ddd      character with octal code ddd, or backreference
       
   236   \xhh      character with hex code hh
       
   237   \x{hhh..} character with hex code hhh..
       
   238 </pre>
       
   239 The precise effect of \cx is as follows: if x is a lower case letter, it
       
   240 is converted to upper case. Then bit 6 of the character (hex 40) is inverted.
       
   241 Thus \cz becomes hex 1A, but \c{ becomes hex 3B, while \c; becomes hex
       
   242 7B.
       
   243 </P>
       
   244 <P>
       
   245 After \x, from zero to two hexadecimal digits are read (letters can be in
       
   246 upper or lower case). Any number of hexadecimal digits may appear between \x{
       
   247 and }, but the value of the character code must be less than 256 in non-UTF-8
       
   248 mode, and less than 2**31 in UTF-8 mode. That is, the maximum value in
       
   249 hexadecimal is 7FFFFFFF. Note that this is bigger than the largest Unicode code
       
   250 point, which is 10FFFF.
       
   251 </P>
       
   252 <P>
       
   253 If characters other than hexadecimal digits appear between \x{ and }, or if
       
   254 there is no terminating }, this form of escape is not recognized. Instead, the
       
   255 initial \x will be interpreted as a basic hexadecimal escape, with no
       
   256 following digits, giving a character whose value is zero.
       
   257 </P>
       
   258 <P>
       
   259 Characters whose value is less than 256 can be defined by either of the two
       
   260 syntaxes for \x. There is no difference in the way they are handled. For
       
   261 example, \xdc is exactly the same as \x{dc}.
       
   262 </P>
       
   263 <P>
       
   264 After \0 up to two further octal digits are read. If there are fewer than two
       
   265 digits, just those that are present are used. Thus the sequence \0\x\07
       
   266 specifies two binary zeros followed by a BEL character (code value 7). Make
       
   267 sure you supply two digits after the initial zero if the pattern character that
       
   268 follows is itself an octal digit.
       
   269 </P>
       
   270 <P>
       
   271 The handling of a backslash followed by a digit other than 0 is complicated.
       
   272 Outside a character class, PCRE reads it and any following digits as a decimal
       
   273 number. If the number is less than 10, or if there have been at least that many
       
   274 previous capturing left parentheses in the expression, the entire sequence is
       
   275 taken as a <i>back reference</i>. A description of how this works is given
       
   276 <a href="#backreferences">later,</a>
       
   277 following the discussion of
       
   278 <a href="#subpattern">parenthesized subpatterns.</a>
       
   279 </P>
       
   280 <P>
       
   281 Inside a character class, or if the decimal number is greater than 9 and there
       
   282 have not been that many capturing subpatterns, PCRE re-reads up to three octal
       
   283 digits following the backslash, and uses them to generate a data character. Any
       
   284 subsequent digits stand for themselves. In non-UTF-8 mode, the value of a
       
   285 character specified in octal must be less than \400. In UTF-8 mode, values up
       
   286 to \777 are permitted. For example:
       
   287 <pre>
       
   288   \040   is another way of writing a space
       
   289   \40    is the same, provided there are fewer than 40 previous capturing subpatterns
       
   290   \7     is always a back reference
       
   291   \11    might be a back reference, or another way of writing a tab
       
   292   \011   is always a tab
       
   293   \0113  is a tab followed by the character "3"
       
   294   \113   might be a back reference, otherwise the character with octal code 113
       
   295   \377   might be a back reference, otherwise the byte consisting entirely of 1 bits
       
   296   \81    is either a back reference, or a binary zero followed by the two characters "8" and "1"
       
   297 </pre>
       
   298 Note that octal values of 100 or greater must not be introduced by a leading
       
   299 zero, because no more than three octal digits are ever read.
       
   300 </P>
       
   301 <P>
       
   302 All the sequences that define a single character value can be used both inside
       
   303 and outside character classes. In addition, inside a character class, the
       
   304 sequence \b is interpreted as the backspace character (hex 08), and the
       
   305 sequences \R and \X are interpreted as the characters "R" and "X",
       
   306 respectively. Outside a character class, these sequences have different
       
   307 meanings
       
   308 <a href="#uniextseq">(see below).</a>
       
   309 </P>
       
   310 <br><b>
       
   311 Absolute and relative back references
       
   312 </b><br>
       
   313 <P>
       
   314 The sequence \g followed by an unsigned or a negative number, optionally
       
   315 enclosed in braces, is an absolute or relative back reference. A named back
       
   316 reference can be coded as \g{name}. Back references are discussed
       
   317 <a href="#backreferences">later,</a>
       
   318 following the discussion of
       
   319 <a href="#subpattern">parenthesized subpatterns.</a>
       
   320 </P>
       
   321 <br><b>
       
   322 Absolute and relative subroutine calls
       
   323 </b><br>
       
   324 <P>
       
   325 For compatibility with Oniguruma, the non-Perl syntax \g followed by a name or
       
   326 a number enclosed either in angle brackets or single quotes, is an alternative
       
   327 syntax for referencing a subpattern as a "subroutine". Details are discussed
       
   328 <a href="#onigurumasubroutines">later.</a>
       
   329 Note that \g{...} (Perl syntax) and \g&#60;...&#62; (Oniguruma syntax) are <i>not</i>
       
   330 synonymous. The former is a back reference; the latter is a subroutine call.
       
   331 </P>
       
   332 <br><b>
       
   333 Generic character types
       
   334 </b><br>
       
   335 <P>
       
   336 Another use of backslash is for specifying generic character types. The
       
   337 following are always recognized:
       
   338 <pre>
       
   339   \d     any decimal digit
       
   340   \D     any character that is not a decimal digit
       
   341   \h     any horizontal whitespace character
       
   342   \H     any character that is not a horizontal whitespace character
       
   343   \s     any whitespace character
       
   344   \S     any character that is not a whitespace character
       
   345   \v     any vertical whitespace character
       
   346   \V     any character that is not a vertical whitespace character
       
   347   \w     any "word" character
       
   348   \W     any "non-word" character
       
   349 </pre>
       
   350 Each pair of escape sequences partitions the complete set of characters into
       
   351 two disjoint sets. Any given character matches one, and only one, of each pair.
       
   352 </P>
       
   353 <P>
       
   354 These character type sequences can appear both inside and outside character
       
   355 classes. They each match one character of the appropriate type. If the current
       
   356 matching point is at the end of the subject string, all of them fail, since
       
   357 there is no character to match.
       
   358 </P>
       
   359 <P>
       
   360 For compatibility with Perl, \s does not match the VT character (code 11).
       
   361 This makes it different from the the POSIX "space" class. The \s characters
       
   362 are HT (9), LF (10), FF (12), CR (13), and space (32). If "use locale;" is
       
   363 included in a Perl script, \s may match the VT character. In PCRE, it never
       
   364 does.
       
   365 </P>
       
   366 <P>
       
   367 In UTF-8 mode, characters with values greater than 128 never match \d, \s, or
       
   368 \w, and always match \D, \S, and \W. This is true even when Unicode
       
   369 character property support is available. These sequences retain their original
       
   370 meanings from before UTF-8 support was available, mainly for efficiency
       
   371 reasons.
       
   372 </P>
       
   373 <P>
       
   374 The sequences \h, \H, \v, and \V are Perl 5.10 features. In contrast to the
       
   375 other sequences, these do match certain high-valued codepoints in UTF-8 mode.
       
   376 The horizontal space characters are:
       
   377 <pre>
       
   378   U+0009     Horizontal tab
       
   379   U+0020     Space
       
   380   U+00A0     Non-break space
       
   381   U+1680     Ogham space mark
       
   382   U+180E     Mongolian vowel separator
       
   383   U+2000     En quad
       
   384   U+2001     Em quad
       
   385   U+2002     En space
       
   386   U+2003     Em space
       
   387   U+2004     Three-per-em space
       
   388   U+2005     Four-per-em space
       
   389   U+2006     Six-per-em space
       
   390   U+2007     Figure space
       
   391   U+2008     Punctuation space
       
   392   U+2009     Thin space
       
   393   U+200A     Hair space
       
   394   U+202F     Narrow no-break space
       
   395   U+205F     Medium mathematical space
       
   396   U+3000     Ideographic space
       
   397 </pre>
       
   398 The vertical space characters are:
       
   399 <pre>
       
   400   U+000A     Linefeed
       
   401   U+000B     Vertical tab
       
   402   U+000C     Formfeed
       
   403   U+000D     Carriage return
       
   404   U+0085     Next line
       
   405   U+2028     Line separator
       
   406   U+2029     Paragraph separator
       
   407 </PRE>
       
   408 </P>
       
   409 <P>
       
   410 A "word" character is an underscore or any character less than 256 that is a
       
   411 letter or digit. The definition of letters and digits is controlled by PCRE's
       
   412 low-valued character tables, and may vary if locale-specific matching is taking
       
   413 place (see
       
   414 <a href="pcreapi.html#localesupport">"Locale support"</a>
       
   415 in the
       
   416 <a href="pcreapi.html"><b>pcreapi</b></a>
       
   417 page). For example, in a French locale such as "fr_FR" in Unix-like systems,
       
   418 or "french" in Windows, some character codes greater than 128 are used for
       
   419 accented letters, and these are matched by \w. The use of locales with Unicode
       
   420 is discouraged.
       
   421 <a name="newlineseq"></a></P>
       
   422 <br><b>
       
   423 Newline sequences
       
   424 </b><br>
       
   425 <P>
       
   426 Outside a character class, by default, the escape sequence \R matches any
       
   427 Unicode newline sequence. This is a Perl 5.10 feature. In non-UTF-8 mode \R is
       
   428 equivalent to the following:
       
   429 <pre>
       
   430   (?&#62;\r\n|\n|\x0b|\f|\r|\x85)
       
   431 </pre>
       
   432 This is an example of an "atomic group", details of which are given
       
   433 <a href="#atomicgroup">below.</a>
       
   434 This particular group matches either the two-character sequence CR followed by
       
   435 LF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,
       
   436 U+000B), FF (formfeed, U+000C), CR (carriage return, U+000D), or NEL (next
       
   437 line, U+0085). The two-character sequence is treated as a single unit that
       
   438 cannot be split.
       
   439 </P>
       
   440 <P>
       
   441 In UTF-8 mode, two additional characters whose codepoints are greater than 255
       
   442 are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).
       
   443 Unicode character property support is not needed for these characters to be
       
   444 recognized.
       
   445 </P>
       
   446 <P>
       
   447 It is possible to restrict \R to match only CR, LF, or CRLF (instead of the
       
   448 complete set of Unicode line endings) by setting the option PCRE_BSR_ANYCRLF
       
   449 either at compile time or when the pattern is matched. (BSR is an abbrevation
       
   450 for "backslash R".) This can be made the default when PCRE is built; if this is
       
   451 the case, the other behaviour can be requested via the PCRE_BSR_UNICODE option.
       
   452 It is also possible to specify these settings by starting a pattern string with
       
   453 one of the following sequences:
       
   454 <pre>
       
   455   (*BSR_ANYCRLF)   CR, LF, or CRLF only
       
   456   (*BSR_UNICODE)   any Unicode newline sequence
       
   457 </pre>
       
   458 These override the default and the options given to <b>pcre_compile()</b>, but
       
   459 they can be overridden by options given to <b>pcre_exec()</b>. Note that these
       
   460 special settings, which are not Perl-compatible, are recognized only at the
       
   461 very start of a pattern, and that they must be in upper case. If more than one
       
   462 of them is present, the last one is used. They can be combined with a change of
       
   463 newline convention, for example, a pattern can start with:
       
   464 <pre>
       
   465   (*ANY)(*BSR_ANYCRLF)
       
   466 </pre>
       
   467 Inside a character class, \R matches the letter "R".
       
   468 <a name="uniextseq"></a></P>
       
   469 <br><b>
       
   470 Unicode character properties
       
   471 </b><br>
       
   472 <P>
       
   473 When PCRE is built with Unicode character property support, three additional
       
   474 escape sequences that match characters with specific properties are available.
       
   475 When not in UTF-8 mode, these sequences are of course limited to testing
       
   476 characters whose codepoints are less than 256, but they do work in this mode.
       
   477 The extra escape sequences are:
       
   478 <pre>
       
   479   \p{<i>xx</i>}   a character with the <i>xx</i> property
       
   480   \P{<i>xx</i>}   a character without the <i>xx</i> property
       
   481   \X       an extended Unicode sequence
       
   482 </pre>
       
   483 The property names represented by <i>xx</i> above are limited to the Unicode
       
   484 script names, the general category properties, and "Any", which matches any
       
   485 character (including newline). Other properties such as "InMusicalSymbols" are
       
   486 not currently supported by PCRE. Note that \P{Any} does not match any
       
   487 characters, so always causes a match failure.
       
   488 </P>
       
   489 <P>
       
   490 Sets of Unicode characters are defined as belonging to certain scripts. A
       
   491 character from one of these sets can be matched using a script name. For
       
   492 example:
       
   493 <pre>
       
   494   \p{Greek}
       
   495   \P{Han}
       
   496 </pre>
       
   497 Those that are not part of an identified script are lumped together as
       
   498 "Common". The current list of scripts is:
       
   499 </P>
       
   500 <P>
       
   501 Arabic,
       
   502 Armenian,
       
   503 Balinese,
       
   504 Bengali,
       
   505 Bopomofo,
       
   506 Braille,
       
   507 Buginese,
       
   508 Buhid,
       
   509 Canadian_Aboriginal,
       
   510 Cherokee,
       
   511 Common,
       
   512 Coptic,
       
   513 Cuneiform,
       
   514 Cypriot,
       
   515 Cyrillic,
       
   516 Deseret,
       
   517 Devanagari,
       
   518 Ethiopic,
       
   519 Georgian,
       
   520 Glagolitic,
       
   521 Gothic,
       
   522 Greek,
       
   523 Gujarati,
       
   524 Gurmukhi,
       
   525 Han,
       
   526 Hangul,
       
   527 Hanunoo,
       
   528 Hebrew,
       
   529 Hiragana,
       
   530 Inherited,
       
   531 Kannada,
       
   532 Katakana,
       
   533 Kharoshthi,
       
   534 Khmer,
       
   535 Lao,
       
   536 Latin,
       
   537 Limbu,
       
   538 Linear_B,
       
   539 Malayalam,
       
   540 Mongolian,
       
   541 Myanmar,
       
   542 New_Tai_Lue,
       
   543 Nko,
       
   544 Ogham,
       
   545 Old_Italic,
       
   546 Old_Persian,
       
   547 Oriya,
       
   548 Osmanya,
       
   549 Phags_Pa,
       
   550 Phoenician,
       
   551 Runic,
       
   552 Shavian,
       
   553 Sinhala,
       
   554 Syloti_Nagri,
       
   555 Syriac,
       
   556 Tagalog,
       
   557 Tagbanwa,
       
   558 Tai_Le,
       
   559 Tamil,
       
   560 Telugu,
       
   561 Thaana,
       
   562 Thai,
       
   563 Tibetan,
       
   564 Tifinagh,
       
   565 Ugaritic,
       
   566 Yi.
       
   567 </P>
       
   568 <P>
       
   569 Each character has exactly one general category property, specified by a
       
   570 two-letter abbreviation. For compatibility with Perl, negation can be specified
       
   571 by including a circumflex between the opening brace and the property name. For
       
   572 example, \p{^Lu} is the same as \P{Lu}.
       
   573 </P>
       
   574 <P>
       
   575 If only one letter is specified with \p or \P, it includes all the general
       
   576 category properties that start with that letter. In this case, in the absence
       
   577 of negation, the curly brackets in the escape sequence are optional; these two
       
   578 examples have the same effect:
       
   579 <pre>
       
   580   \p{L}
       
   581   \pL
       
   582 </pre>
       
   583 The following general category property codes are supported:
       
   584 <pre>
       
   585   C     Other
       
   586   Cc    Control
       
   587   Cf    Format
       
   588   Cn    Unassigned
       
   589   Co    Private use
       
   590   Cs    Surrogate
       
   591 
       
   592   L     Letter
       
   593   Ll    Lower case letter
       
   594   Lm    Modifier letter
       
   595   Lo    Other letter
       
   596   Lt    Title case letter
       
   597   Lu    Upper case letter
       
   598 
       
   599   M     Mark
       
   600   Mc    Spacing mark
       
   601   Me    Enclosing mark
       
   602   Mn    Non-spacing mark
       
   603 
       
   604   N     Number
       
   605   Nd    Decimal number
       
   606   Nl    Letter number
       
   607   No    Other number
       
   608 
       
   609   P     Punctuation
       
   610   Pc    Connector punctuation
       
   611   Pd    Dash punctuation
       
   612   Pe    Close punctuation
       
   613   Pf    Final punctuation
       
   614   Pi    Initial punctuation
       
   615   Po    Other punctuation
       
   616   Ps    Open punctuation
       
   617 
       
   618   S     Symbol
       
   619   Sc    Currency symbol
       
   620   Sk    Modifier symbol
       
   621   Sm    Mathematical symbol
       
   622   So    Other symbol
       
   623 
       
   624   Z     Separator
       
   625   Zl    Line separator
       
   626   Zp    Paragraph separator
       
   627   Zs    Space separator
       
   628 </pre>
       
   629 The special property L& is also supported: it matches a character that has
       
   630 the Lu, Ll, or Lt property, in other words, a letter that is not classified as
       
   631 a modifier or "other".
       
   632 </P>
       
   633 <P>
       
   634 The Cs (Surrogate) property applies only to characters in the range U+D800 to
       
   635 U+DFFF. Such characters are not valid in UTF-8 strings (see RFC 3629) and so
       
   636 cannot be tested by PCRE, unless UTF-8 validity checking has been turned off
       
   637 (see the discussion of PCRE_NO_UTF8_CHECK in the
       
   638 <a href="pcreapi.html"><b>pcreapi</b></a>
       
   639 page).
       
   640 </P>
       
   641 <P>
       
   642 The long synonyms for these properties that Perl supports (such as \p{Letter})
       
   643 are not supported by PCRE, nor is it permitted to prefix any of these
       
   644 properties with "Is".
       
   645 </P>
       
   646 <P>
       
   647 No character that is in the Unicode table has the Cn (unassigned) property.
       
   648 Instead, this property is assumed for any code point that is not in the
       
   649 Unicode table.
       
   650 </P>
       
   651 <P>
       
   652 Specifying caseless matching does not affect these escape sequences. For
       
   653 example, \p{Lu} always matches only upper case letters.
       
   654 </P>
       
   655 <P>
       
   656 The \X escape matches any number of Unicode characters that form an extended
       
   657 Unicode sequence. \X is equivalent to
       
   658 <pre>
       
   659   (?&#62;\PM\pM*)
       
   660 </pre>
       
   661 That is, it matches a character without the "mark" property, followed by zero
       
   662 or more characters with the "mark" property, and treats the sequence as an
       
   663 atomic group
       
   664 <a href="#atomicgroup">(see below).</a>
       
   665 Characters with the "mark" property are typically accents that affect the
       
   666 preceding character. None of them have codepoints less than 256, so in
       
   667 non-UTF-8 mode \X matches any one character.
       
   668 </P>
       
   669 <P>
       
   670 Matching characters by Unicode property is not fast, because PCRE has to search
       
   671 a structure that contains data for over fifteen thousand characters. That is
       
   672 why the traditional escape sequences such as \d and \w do not use Unicode
       
   673 properties in PCRE.
       
   674 <a name="resetmatchstart"></a></P>
       
   675 <br><b>
       
   676 Resetting the match start
       
   677 </b><br>
       
   678 <P>
       
   679 The escape sequence \K, which is a Perl 5.10 feature, causes any previously
       
   680 matched characters not to be included in the final matched sequence. For
       
   681 example, the pattern:
       
   682 <pre>
       
   683   foo\Kbar
       
   684 </pre>
       
   685 matches "foobar", but reports that it has matched "bar". This feature is
       
   686 similar to a lookbehind assertion
       
   687 <a href="#lookbehind">(described below).</a>
       
   688 However, in this case, the part of the subject before the real match does not
       
   689 have to be of fixed length, as lookbehind assertions do. The use of \K does
       
   690 not interfere with the setting of
       
   691 <a href="#subpattern">captured substrings.</a>
       
   692 For example, when the pattern
       
   693 <pre>
       
   694   (foo)\Kbar
       
   695 </pre>
       
   696 matches "foobar", the first substring is still set to "foo".
       
   697 <a name="smallassertions"></a></P>
       
   698 <br><b>
       
   699 Simple assertions
       
   700 </b><br>
       
   701 <P>
       
   702 The final use of backslash is for certain simple assertions. An assertion
       
   703 specifies a condition that has to be met at a particular point in a match,
       
   704 without consuming any characters from the subject string. The use of
       
   705 subpatterns for more complicated assertions is described
       
   706 <a href="#bigassertions">below.</a>
       
   707 The backslashed assertions are:
       
   708 <pre>
       
   709   \b     matches at a word boundary
       
   710   \B     matches when not at a word boundary
       
   711   \A     matches at the start of the subject
       
   712   \Z     matches at the end of the subject
       
   713           also matches before a newline at the end of the subject
       
   714   \z     matches only at the end of the subject
       
   715   \G     matches at the first matching position in the subject
       
   716 </pre>
       
   717 These assertions may not appear in character classes (but note that \b has a
       
   718 different meaning, namely the backspace character, inside a character class).
       
   719 </P>
       
   720 <P>
       
   721 A word boundary is a position in the subject string where the current character
       
   722 and the previous character do not both match \w or \W (i.e. one matches
       
   723 \w and the other matches \W), or the start or end of the string if the
       
   724 first or last character matches \w, respectively.
       
   725 </P>
       
   726 <P>
       
   727 The \A, \Z, and \z assertions differ from the traditional circumflex and
       
   728 dollar (described in the next section) in that they only ever match at the very
       
   729 start and end of the subject string, whatever options are set. Thus, they are
       
   730 independent of multiline mode. These three assertions are not affected by the
       
   731 PCRE_NOTBOL or PCRE_NOTEOL options, which affect only the behaviour of the
       
   732 circumflex and dollar metacharacters. However, if the <i>startoffset</i>
       
   733 argument of <b>pcre_exec()</b> is non-zero, indicating that matching is to start
       
   734 at a point other than the beginning of the subject, \A can never match. The
       
   735 difference between \Z and \z is that \Z matches before a newline at the end
       
   736 of the string as well as at the very end, whereas \z matches only at the end.
       
   737 </P>
       
   738 <P>
       
   739 The \G assertion is true only when the current matching position is at the
       
   740 start point of the match, as specified by the <i>startoffset</i> argument of
       
   741 <b>pcre_exec()</b>. It differs from \A when the value of <i>startoffset</i> is
       
   742 non-zero. By calling <b>pcre_exec()</b> multiple times with appropriate
       
   743 arguments, you can mimic Perl's /g option, and it is in this kind of
       
   744 implementation where \G can be useful.
       
   745 </P>
       
   746 <P>
       
   747 Note, however, that PCRE's interpretation of \G, as the start of the current
       
   748 match, is subtly different from Perl's, which defines it as the end of the
       
   749 previous match. In Perl, these can be different when the previously matched
       
   750 string was empty. Because PCRE does just one match at a time, it cannot
       
   751 reproduce this behaviour.
       
   752 </P>
       
   753 <P>
       
   754 If all the alternatives of a pattern begin with \G, the expression is anchored
       
   755 to the starting match position, and the "anchored" flag is set in the compiled
       
   756 regular expression.
       
   757 </P>
       
   758 <br><a name="SEC5" href="#TOC1">CIRCUMFLEX AND DOLLAR</a><br>
       
   759 <P>
       
   760 Outside a character class, in the default matching mode, the circumflex
       
   761 character is an assertion that is true only if the current matching point is
       
   762 at the start of the subject string. If the <i>startoffset</i> argument of
       
   763 <b>pcre_exec()</b> is non-zero, circumflex can never match if the PCRE_MULTILINE
       
   764 option is unset. Inside a character class, circumflex has an entirely different
       
   765 meaning
       
   766 <a href="#characterclass">(see below).</a>
       
   767 </P>
       
   768 <P>
       
   769 Circumflex need not be the first character of the pattern if a number of
       
   770 alternatives are involved, but it should be the first thing in each alternative
       
   771 in which it appears if the pattern is ever to match that branch. If all
       
   772 possible alternatives start with a circumflex, that is, if the pattern is
       
   773 constrained to match only at the start of the subject, it is said to be an
       
   774 "anchored" pattern. (There are also other constructs that can cause a pattern
       
   775 to be anchored.)
       
   776 </P>
       
   777 <P>
       
   778 A dollar character is an assertion that is true only if the current matching
       
   779 point is at the end of the subject string, or immediately before a newline
       
   780 at the end of the string (by default). Dollar need not be the last character of
       
   781 the pattern if a number of alternatives are involved, but it should be the last
       
   782 item in any branch in which it appears. Dollar has no special meaning in a
       
   783 character class.
       
   784 </P>
       
   785 <P>
       
   786 The meaning of dollar can be changed so that it matches only at the very end of
       
   787 the string, by setting the PCRE_DOLLAR_ENDONLY option at compile time. This
       
   788 does not affect the \Z assertion.
       
   789 </P>
       
   790 <P>
       
   791 The meanings of the circumflex and dollar characters are changed if the
       
   792 PCRE_MULTILINE option is set. When this is the case, a circumflex matches
       
   793 immediately after internal newlines as well as at the start of the subject
       
   794 string. It does not match after a newline that ends the string. A dollar
       
   795 matches before any newlines in the string, as well as at the very end, when
       
   796 PCRE_MULTILINE is set. When newline is specified as the two-character
       
   797 sequence CRLF, isolated CR and LF characters do not indicate newlines.
       
   798 </P>
       
   799 <P>
       
   800 For example, the pattern /^abc$/ matches the subject string "def\nabc" (where
       
   801 \n represents a newline) in multiline mode, but not otherwise. Consequently,
       
   802 patterns that are anchored in single line mode because all branches start with
       
   803 ^ are not anchored in multiline mode, and a match for circumflex is possible
       
   804 when the <i>startoffset</i> argument of <b>pcre_exec()</b> is non-zero. The
       
   805 PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
       
   806 </P>
       
   807 <P>
       
   808 Note that the sequences \A, \Z, and \z can be used to match the start and
       
   809 end of the subject in both modes, and if all branches of a pattern start with
       
   810 \A it is always anchored, whether or not PCRE_MULTILINE is set.
       
   811 </P>
       
   812 <br><a name="SEC6" href="#TOC1">FULL STOP (PERIOD, DOT)</a><br>
       
   813 <P>
       
   814 Outside a character class, a dot in the pattern matches any one character in
       
   815 the subject string except (by default) a character that signifies the end of a
       
   816 line. In UTF-8 mode, the matched character may be more than one byte long.
       
   817 </P>
       
   818 <P>
       
   819 When a line ending is defined as a single character, dot never matches that
       
   820 character; when the two-character sequence CRLF is used, dot does not match CR
       
   821 if it is immediately followed by LF, but otherwise it matches all characters
       
   822 (including isolated CRs and LFs). When any Unicode line endings are being
       
   823 recognized, dot does not match CR or LF or any of the other line ending
       
   824 characters.
       
   825 </P>
       
   826 <P>
       
   827 The behaviour of dot with regard to newlines can be changed. If the PCRE_DOTALL
       
   828 option is set, a dot matches any one character, without exception. If the
       
   829 two-character sequence CRLF is present in the subject string, it takes two dots
       
   830 to match it.
       
   831 </P>
       
   832 <P>
       
   833 The handling of dot is entirely independent of the handling of circumflex and
       
   834 dollar, the only relationship being that they both involve newlines. Dot has no
       
   835 special meaning in a character class.
       
   836 </P>
       
   837 <br><a name="SEC7" href="#TOC1">MATCHING A SINGLE BYTE</a><br>
       
   838 <P>
       
   839 Outside a character class, the escape sequence \C matches any one byte, both
       
   840 in and out of UTF-8 mode. Unlike a dot, it always matches any line-ending
       
   841 characters. The feature is provided in Perl in order to match individual bytes
       
   842 in UTF-8 mode. Because it breaks up UTF-8 characters into individual bytes,
       
   843 what remains in the string may be a malformed UTF-8 string. For this reason,
       
   844 the \C escape sequence is best avoided.
       
   845 </P>
       
   846 <P>
       
   847 PCRE does not allow \C to appear in lookbehind assertions
       
   848 <a href="#lookbehind">(described below),</a>
       
   849 because in UTF-8 mode this would make it impossible to calculate the length of
       
   850 the lookbehind.
       
   851 <a name="characterclass"></a></P>
       
   852 <br><a name="SEC8" href="#TOC1">SQUARE BRACKETS AND CHARACTER CLASSES</a><br>
       
   853 <P>
       
   854 An opening square bracket introduces a character class, terminated by a closing
       
   855 square bracket. A closing square bracket on its own is not special. If a
       
   856 closing square bracket is required as a member of the class, it should be the
       
   857 first data character in the class (after an initial circumflex, if present) or
       
   858 escaped with a backslash.
       
   859 </P>
       
   860 <P>
       
   861 A character class matches a single character in the subject. In UTF-8 mode, the
       
   862 character may occupy more than one byte. A matched character must be in the set
       
   863 of characters defined by the class, unless the first character in the class
       
   864 definition is a circumflex, in which case the subject character must not be in
       
   865 the set defined by the class. If a circumflex is actually required as a member
       
   866 of the class, ensure it is not the first character, or escape it with a
       
   867 backslash.
       
   868 </P>
       
   869 <P>
       
   870 For example, the character class [aeiou] matches any lower case vowel, while
       
   871 [^aeiou] matches any character that is not a lower case vowel. Note that a
       
   872 circumflex is just a convenient notation for specifying the characters that
       
   873 are in the class by enumerating those that are not. A class that starts with a
       
   874 circumflex is not an assertion: it still consumes a character from the subject
       
   875 string, and therefore it fails if the current pointer is at the end of the
       
   876 string.
       
   877 </P>
       
   878 <P>
       
   879 In UTF-8 mode, characters with values greater than 255 can be included in a
       
   880 class as a literal string of bytes, or by using the \x{ escaping mechanism.
       
   881 </P>
       
   882 <P>
       
   883 When caseless matching is set, any letters in a class represent both their
       
   884 upper case and lower case versions, so for example, a caseless [aeiou] matches
       
   885 "A" as well as "a", and a caseless [^aeiou] does not match "A", whereas a
       
   886 caseful version would. In UTF-8 mode, PCRE always understands the concept of
       
   887 case for characters whose values are less than 128, so caseless matching is
       
   888 always possible. For characters with higher values, the concept of case is
       
   889 supported if PCRE is compiled with Unicode property support, but not otherwise.
       
   890 If you want to use caseless matching for characters 128 and above, you must
       
   891 ensure that PCRE is compiled with Unicode property support as well as with
       
   892 UTF-8 support.
       
   893 </P>
       
   894 <P>
       
   895 Characters that might indicate line breaks are never treated in any special way
       
   896 when matching character classes, whatever line-ending sequence is in use, and
       
   897 whatever setting of the PCRE_DOTALL and PCRE_MULTILINE options is used. A class
       
   898 such as [^a] always matches one of these characters.
       
   899 </P>
       
   900 <P>
       
   901 The minus (hyphen) character can be used to specify a range of characters in a
       
   902 character class. For example, [d-m] matches any letter between d and m,
       
   903 inclusive. If a minus character is required in a class, it must be escaped with
       
   904 a backslash or appear in a position where it cannot be interpreted as
       
   905 indicating a range, typically as the first or last character in the class.
       
   906 </P>
       
   907 <P>
       
   908 It is not possible to have the literal character "]" as the end character of a
       
   909 range. A pattern such as [W-]46] is interpreted as a class of two characters
       
   910 ("W" and "-") followed by a literal string "46]", so it would match "W46]" or
       
   911 "-46]". However, if the "]" is escaped with a backslash it is interpreted as
       
   912 the end of range, so [W-\]46] is interpreted as a class containing a range
       
   913 followed by two other characters. The octal or hexadecimal representation of
       
   914 "]" can also be used to end a range.
       
   915 </P>
       
   916 <P>
       
   917 Ranges operate in the collating sequence of character values. They can also be
       
   918 used for characters specified numerically, for example [\000-\037]. In UTF-8
       
   919 mode, ranges can include characters whose values are greater than 255, for
       
   920 example [\x{100}-\x{2ff}].
       
   921 </P>
       
   922 <P>
       
   923 If a range that includes letters is used when caseless matching is set, it
       
   924 matches the letters in either case. For example, [W-c] is equivalent to
       
   925 [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if character
       
   926 tables for a French locale are in use, [\xc8-\xcb] matches accented E
       
   927 characters in both cases. In UTF-8 mode, PCRE supports the concept of case for
       
   928 characters with values greater than 128 only when it is compiled with Unicode
       
   929 property support.
       
   930 </P>
       
   931 <P>
       
   932 The character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear
       
   933 in a character class, and add the characters that they match to the class. For
       
   934 example, [\dABCDEF] matches any hexadecimal digit. A circumflex can
       
   935 conveniently be used with the upper case character types to specify a more
       
   936 restricted set of characters than the matching lower case type. For example,
       
   937 the class [^\W_] matches any letter or digit, but not underscore.
       
   938 </P>
       
   939 <P>
       
   940 The only metacharacters that are recognized in character classes are backslash,
       
   941 hyphen (only where it can be interpreted as specifying a range), circumflex
       
   942 (only at the start), opening square bracket (only when it can be interpreted as
       
   943 introducing a POSIX class name - see the next section), and the terminating
       
   944 closing square bracket. However, escaping other non-alphanumeric characters
       
   945 does no harm.
       
   946 </P>
       
   947 <br><a name="SEC9" href="#TOC1">POSIX CHARACTER CLASSES</a><br>
       
   948 <P>
       
   949 Perl supports the POSIX notation for character classes. This uses names
       
   950 enclosed by [: and :] within the enclosing square brackets. PCRE also supports
       
   951 this notation. For example,
       
   952 <pre>
       
   953   [01[:alpha:]%]
       
   954 </pre>
       
   955 matches "0", "1", any alphabetic character, or "%". The supported class names
       
   956 are
       
   957 <pre>
       
   958   alnum    letters and digits
       
   959   alpha    letters
       
   960   ascii    character codes 0 - 127
       
   961   blank    space or tab only
       
   962   cntrl    control characters
       
   963   digit    decimal digits (same as \d)
       
   964   graph    printing characters, excluding space
       
   965   lower    lower case letters
       
   966   print    printing characters, including space
       
   967   punct    printing characters, excluding letters and digits
       
   968   space    white space (not quite the same as \s)
       
   969   upper    upper case letters
       
   970   word     "word" characters (same as \w)
       
   971   xdigit   hexadecimal digits
       
   972 </pre>
       
   973 The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13), and
       
   974 space (32). Notice that this list includes the VT character (code 11). This
       
   975 makes "space" different to \s, which does not include VT (for Perl
       
   976 compatibility).
       
   977 </P>
       
   978 <P>
       
   979 The name "word" is a Perl extension, and "blank" is a GNU extension from Perl
       
   980 5.8. Another Perl extension is negation, which is indicated by a ^ character
       
   981 after the colon. For example,
       
   982 <pre>
       
   983   [12[:^digit:]]
       
   984 </pre>
       
   985 matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the POSIX
       
   986 syntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are not
       
   987 supported, and an error is given if they are encountered.
       
   988 </P>
       
   989 <P>
       
   990 In UTF-8 mode, characters with values greater than 128 do not match any of
       
   991 the POSIX character classes.
       
   992 </P>
       
   993 <br><a name="SEC10" href="#TOC1">VERTICAL BAR</a><br>
       
   994 <P>
       
   995 Vertical bar characters are used to separate alternative patterns. For example,
       
   996 the pattern
       
   997 <pre>
       
   998   gilbert|sullivan
       
   999 </pre>
       
  1000 matches either "gilbert" or "sullivan". Any number of alternatives may appear,
       
  1001 and an empty alternative is permitted (matching the empty string). The matching
       
  1002 process tries each alternative in turn, from left to right, and the first one
       
  1003 that succeeds is used. If the alternatives are within a subpattern
       
  1004 <a href="#subpattern">(defined below),</a>
       
  1005 "succeeds" means matching the rest of the main pattern as well as the
       
  1006 alternative in the subpattern.
       
  1007 </P>
       
  1008 <br><a name="SEC11" href="#TOC1">INTERNAL OPTION SETTING</a><br>
       
  1009 <P>
       
  1010 The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
       
  1011 PCRE_EXTENDED options (which are Perl-compatible) can be changed from within
       
  1012 the pattern by a sequence of Perl option letters enclosed between "(?" and ")".
       
  1013 The option letters are
       
  1014 <pre>
       
  1015   i  for PCRE_CASELESS
       
  1016   m  for PCRE_MULTILINE
       
  1017   s  for PCRE_DOTALL
       
  1018   x  for PCRE_EXTENDED
       
  1019 </pre>
       
  1020 For example, (?im) sets caseless, multiline matching. It is also possible to
       
  1021 unset these options by preceding the letter with a hyphen, and a combined
       
  1022 setting and unsetting such as (?im-sx), which sets PCRE_CASELESS and
       
  1023 PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED, is also
       
  1024 permitted. If a letter appears both before and after the hyphen, the option is
       
  1025 unset.
       
  1026 </P>
       
  1027 <P>
       
  1028 The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA can be
       
  1029 changed in the same way as the Perl-compatible options by using the characters
       
  1030 J, U and X respectively.
       
  1031 </P>
       
  1032 <P>
       
  1033 When an option change occurs at top level (that is, not inside subpattern
       
  1034 parentheses), the change applies to the remainder of the pattern that follows.
       
  1035 If the change is placed right at the start of a pattern, PCRE extracts it into
       
  1036 the global options (and it will therefore show up in data extracted by the
       
  1037 <b>pcre_fullinfo()</b> function).
       
  1038 </P>
       
  1039 <P>
       
  1040 An option change within a subpattern (see below for a description of
       
  1041 subpatterns) affects only that part of the current pattern that follows it, so
       
  1042 <pre>
       
  1043   (a(?i)b)c
       
  1044 </pre>
       
  1045 matches abc and aBc and no other strings (assuming PCRE_CASELESS is not used).
       
  1046 By this means, options can be made to have different settings in different
       
  1047 parts of the pattern. Any changes made in one alternative do carry on
       
  1048 into subsequent branches within the same subpattern. For example,
       
  1049 <pre>
       
  1050   (a(?i)b|c)
       
  1051 </pre>
       
  1052 matches "ab", "aB", "c", and "C", even though when matching "C" the first
       
  1053 branch is abandoned before the option setting. This is because the effects of
       
  1054 option settings happen at compile time. There would be some very weird
       
  1055 behaviour otherwise.
       
  1056 </P>
       
  1057 <P>
       
  1058 <b>Note:</b> There are other PCRE-specific options that can be set by the
       
  1059 application when the compile or match functions are called. In some cases the
       
  1060 pattern can contain special leading sequences to override what the application
       
  1061 has set or what has been defaulted. Details are given in the section entitled
       
  1062 <a href="#newlineseq">"Newline sequences"</a>
       
  1063 above.
       
  1064 <a name="subpattern"></a></P>
       
  1065 <br><a name="SEC12" href="#TOC1">SUBPATTERNS</a><br>
       
  1066 <P>
       
  1067 Subpatterns are delimited by parentheses (round brackets), which can be nested.
       
  1068 Turning part of a pattern into a subpattern does two things:
       
  1069 <br>
       
  1070 <br>
       
  1071 1. It localizes a set of alternatives. For example, the pattern
       
  1072 <pre>
       
  1073   cat(aract|erpillar|)
       
  1074 </pre>
       
  1075 matches one of the words "cat", "cataract", or "caterpillar". Without the
       
  1076 parentheses, it would match "cataract", "erpillar" or an empty string.
       
  1077 <br>
       
  1078 <br>
       
  1079 2. It sets up the subpattern as a capturing subpattern. This means that, when
       
  1080 the whole pattern matches, that portion of the subject string that matched the
       
  1081 subpattern is passed back to the caller via the <i>ovector</i> argument of
       
  1082 <b>pcre_exec()</b>. Opening parentheses are counted from left to right (starting
       
  1083 from 1) to obtain numbers for the capturing subpatterns.
       
  1084 </P>
       
  1085 <P>
       
  1086 For example, if the string "the red king" is matched against the pattern
       
  1087 <pre>
       
  1088   the ((red|white) (king|queen))
       
  1089 </pre>
       
  1090 the captured substrings are "red king", "red", and "king", and are numbered 1,
       
  1091 2, and 3, respectively.
       
  1092 </P>
       
  1093 <P>
       
  1094 The fact that plain parentheses fulfil two functions is not always helpful.
       
  1095 There are often times when a grouping subpattern is required without a
       
  1096 capturing requirement. If an opening parenthesis is followed by a question mark
       
  1097 and a colon, the subpattern does not do any capturing, and is not counted when
       
  1098 computing the number of any subsequent capturing subpatterns. For example, if
       
  1099 the string "the white queen" is matched against the pattern
       
  1100 <pre>
       
  1101   the ((?:red|white) (king|queen))
       
  1102 </pre>
       
  1103 the captured substrings are "white queen" and "queen", and are numbered 1 and
       
  1104 2. The maximum number of capturing subpatterns is 65535.
       
  1105 </P>
       
  1106 <P>
       
  1107 As a convenient shorthand, if any option settings are required at the start of
       
  1108 a non-capturing subpattern, the option letters may appear between the "?" and
       
  1109 the ":". Thus the two patterns
       
  1110 <pre>
       
  1111   (?i:saturday|sunday)
       
  1112   (?:(?i)saturday|sunday)
       
  1113 </pre>
       
  1114 match exactly the same set of strings. Because alternative branches are tried
       
  1115 from left to right, and options are not reset until the end of the subpattern
       
  1116 is reached, an option setting in one branch does affect subsequent branches, so
       
  1117 the above patterns match "SUNDAY" as well as "Saturday".
       
  1118 </P>
       
  1119 <br><a name="SEC13" href="#TOC1">DUPLICATE SUBPATTERN NUMBERS</a><br>
       
  1120 <P>
       
  1121 Perl 5.10 introduced a feature whereby each alternative in a subpattern uses
       
  1122 the same numbers for its capturing parentheses. Such a subpattern starts with
       
  1123 (?| and is itself a non-capturing subpattern. For example, consider this
       
  1124 pattern:
       
  1125 <pre>
       
  1126   (?|(Sat)ur|(Sun))day
       
  1127 </pre>
       
  1128 Because the two alternatives are inside a (?| group, both sets of capturing
       
  1129 parentheses are numbered one. Thus, when the pattern matches, you can look
       
  1130 at captured substring number one, whichever alternative matched. This construct
       
  1131 is useful when you want to capture part, but not all, of one of a number of
       
  1132 alternatives. Inside a (?| group, parentheses are numbered as usual, but the
       
  1133 number is reset at the start of each branch. The numbers of any capturing
       
  1134 buffers that follow the subpattern start after the highest number used in any
       
  1135 branch. The following example is taken from the Perl documentation.
       
  1136 The numbers underneath show in which buffer the captured content will be
       
  1137 stored.
       
  1138 <pre>
       
  1139   # before  ---------------branch-reset----------- after
       
  1140   / ( a )  (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
       
  1141   # 1            2         2  3        2     3     4
       
  1142 </pre>
       
  1143 A backreference or a recursive call to a numbered subpattern always refers to
       
  1144 the first one in the pattern with the given number.
       
  1145 </P>
       
  1146 <P>
       
  1147 An alternative approach to using this "branch reset" feature is to use
       
  1148 duplicate named subpatterns, as described in the next section.
       
  1149 </P>
       
  1150 <br><a name="SEC14" href="#TOC1">NAMED SUBPATTERNS</a><br>
       
  1151 <P>
       
  1152 Identifying capturing parentheses by number is simple, but it can be very hard
       
  1153 to keep track of the numbers in complicated regular expressions. Furthermore,
       
  1154 if an expression is modified, the numbers may change. To help with this
       
  1155 difficulty, PCRE supports the naming of subpatterns. This feature was not
       
  1156 added to Perl until release 5.10. Python had the feature earlier, and PCRE
       
  1157 introduced it at release 4.0, using the Python syntax. PCRE now supports both
       
  1158 the Perl and the Python syntax.
       
  1159 </P>
       
  1160 <P>
       
  1161 In PCRE, a subpattern can be named in one of three ways: (?&#60;name&#62;...) or
       
  1162 (?'name'...) as in Perl, or (?P&#60;name&#62;...) as in Python. References to capturing
       
  1163 parentheses from other parts of the pattern, such as
       
  1164 <a href="#backreferences">backreferences,</a>
       
  1165 <a href="#recursion">recursion,</a>
       
  1166 and
       
  1167 <a href="#conditions">conditions,</a>
       
  1168 can be made by name as well as by number.
       
  1169 </P>
       
  1170 <P>
       
  1171 Names consist of up to 32 alphanumeric characters and underscores. Named
       
  1172 capturing parentheses are still allocated numbers as well as names, exactly as
       
  1173 if the names were not present. The PCRE API provides function calls for
       
  1174 extracting the name-to-number translation table from a compiled pattern. There
       
  1175 is also a convenience function for extracting a captured substring by name.
       
  1176 </P>
       
  1177 <P>
       
  1178 By default, a name must be unique within a pattern, but it is possible to relax
       
  1179 this constraint by setting the PCRE_DUPNAMES option at compile time. This can
       
  1180 be useful for patterns where only one instance of the named parentheses can
       
  1181 match. Suppose you want to match the name of a weekday, either as a 3-letter
       
  1182 abbreviation or as the full name, and in both cases you want to extract the
       
  1183 abbreviation. This pattern (ignoring the line breaks) does the job:
       
  1184 <pre>
       
  1185   (?&#60;DN&#62;Mon|Fri|Sun)(?:day)?|
       
  1186   (?&#60;DN&#62;Tue)(?:sday)?|
       
  1187   (?&#60;DN&#62;Wed)(?:nesday)?|
       
  1188   (?&#60;DN&#62;Thu)(?:rsday)?|
       
  1189   (?&#60;DN&#62;Sat)(?:urday)?
       
  1190 </pre>
       
  1191 There are five capturing substrings, but only one is ever set after a match.
       
  1192 (An alternative way of solving this problem is to use a "branch reset"
       
  1193 subpattern, as described in the previous section.)
       
  1194 </P>
       
  1195 <P>
       
  1196 The convenience function for extracting the data by name returns the substring
       
  1197 for the first (and in this example, the only) subpattern of that name that
       
  1198 matched. This saves searching to find which numbered subpattern it was. If you
       
  1199 make a reference to a non-unique named subpattern from elsewhere in the
       
  1200 pattern, the one that corresponds to the lowest number is used. For further
       
  1201 details of the interfaces for handling named subpatterns, see the
       
  1202 <a href="pcreapi.html"><b>pcreapi</b></a>
       
  1203 documentation.
       
  1204 </P>
       
  1205 <br><a name="SEC15" href="#TOC1">REPETITION</a><br>
       
  1206 <P>
       
  1207 Repetition is specified by quantifiers, which can follow any of the following
       
  1208 items:
       
  1209 <pre>
       
  1210   a literal data character
       
  1211   the dot metacharacter
       
  1212   the \C escape sequence
       
  1213   the \X escape sequence (in UTF-8 mode with Unicode properties)
       
  1214   the \R escape sequence
       
  1215   an escape such as \d that matches a single character
       
  1216   a character class
       
  1217   a back reference (see next section)
       
  1218   a parenthesized subpattern (unless it is an assertion)
       
  1219 </pre>
       
  1220 The general repetition quantifier specifies a minimum and maximum number of
       
  1221 permitted matches, by giving the two numbers in curly brackets (braces),
       
  1222 separated by a comma. The numbers must be less than 65536, and the first must
       
  1223 be less than or equal to the second. For example:
       
  1224 <pre>
       
  1225   z{2,4}
       
  1226 </pre>
       
  1227 matches "zz", "zzz", or "zzzz". A closing brace on its own is not a special
       
  1228 character. If the second number is omitted, but the comma is present, there is
       
  1229 no upper limit; if the second number and the comma are both omitted, the
       
  1230 quantifier specifies an exact number of required matches. Thus
       
  1231 <pre>
       
  1232   [aeiou]{3,}
       
  1233 </pre>
       
  1234 matches at least 3 successive vowels, but may match many more, while
       
  1235 <pre>
       
  1236   \d{8}
       
  1237 </pre>
       
  1238 matches exactly 8 digits. An opening curly bracket that appears in a position
       
  1239 where a quantifier is not allowed, or one that does not match the syntax of a
       
  1240 quantifier, is taken as a literal character. For example, {,6} is not a
       
  1241 quantifier, but a literal string of four characters.
       
  1242 </P>
       
  1243 <P>
       
  1244 In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to individual
       
  1245 bytes. Thus, for example, \x{100}{2} matches two UTF-8 characters, each of
       
  1246 which is represented by a two-byte sequence. Similarly, when Unicode property
       
  1247 support is available, \X{3} matches three Unicode extended sequences, each of
       
  1248 which may be several bytes long (and they may be of different lengths).
       
  1249 </P>
       
  1250 <P>
       
  1251 The quantifier {0} is permitted, causing the expression to behave as if the
       
  1252 previous item and the quantifier were not present. This may be useful for
       
  1253 subpatterns that are referenced as
       
  1254 <a href="#subpatternsassubroutines">subroutines</a>
       
  1255 from elsewhere in the pattern. Items other than subpatterns that have a {0}
       
  1256 quantifier are omitted from the compiled pattern.
       
  1257 </P>
       
  1258 <P>
       
  1259 For convenience, the three most common quantifiers have single-character
       
  1260 abbreviations:
       
  1261 <pre>
       
  1262   *    is equivalent to {0,}
       
  1263   +    is equivalent to {1,}
       
  1264   ?    is equivalent to {0,1}
       
  1265 </pre>
       
  1266 It is possible to construct infinite loops by following a subpattern that can
       
  1267 match no characters with a quantifier that has no upper limit, for example:
       
  1268 <pre>
       
  1269   (a?)*
       
  1270 </pre>
       
  1271 Earlier versions of Perl and PCRE used to give an error at compile time for
       
  1272 such patterns. However, because there are cases where this can be useful, such
       
  1273 patterns are now accepted, but if any repetition of the subpattern does in fact
       
  1274 match no characters, the loop is forcibly broken.
       
  1275 </P>
       
  1276 <P>
       
  1277 By default, the quantifiers are "greedy", that is, they match as much as
       
  1278 possible (up to the maximum number of permitted times), without causing the
       
  1279 rest of the pattern to fail. The classic example of where this gives problems
       
  1280 is in trying to match comments in C programs. These appear between /* and */
       
  1281 and within the comment, individual * and / characters may appear. An attempt to
       
  1282 match C comments by applying the pattern
       
  1283 <pre>
       
  1284   /\*.*\*/
       
  1285 </pre>
       
  1286 to the string
       
  1287 <pre>
       
  1288   /* first comment */  not comment  /* second comment */
       
  1289 </pre>
       
  1290 fails, because it matches the entire string owing to the greediness of the .*
       
  1291 item.
       
  1292 </P>
       
  1293 <P>
       
  1294 However, if a quantifier is followed by a question mark, it ceases to be
       
  1295 greedy, and instead matches the minimum number of times possible, so the
       
  1296 pattern
       
  1297 <pre>
       
  1298   /\*.*?\*/
       
  1299 </pre>
       
  1300 does the right thing with the C comments. The meaning of the various
       
  1301 quantifiers is not otherwise changed, just the preferred number of matches.
       
  1302 Do not confuse this use of question mark with its use as a quantifier in its
       
  1303 own right. Because it has two uses, it can sometimes appear doubled, as in
       
  1304 <pre>
       
  1305   \d??\d
       
  1306 </pre>
       
  1307 which matches one digit by preference, but can match two if that is the only
       
  1308 way the rest of the pattern matches.
       
  1309 </P>
       
  1310 <P>
       
  1311 If the PCRE_UNGREEDY option is set (an option that is not available in Perl),
       
  1312 the quantifiers are not greedy by default, but individual ones can be made
       
  1313 greedy by following them with a question mark. In other words, it inverts the
       
  1314 default behaviour.
       
  1315 </P>
       
  1316 <P>
       
  1317 When a parenthesized subpattern is quantified with a minimum repeat count that
       
  1318 is greater than 1 or with a limited maximum, more memory is required for the
       
  1319 compiled pattern, in proportion to the size of the minimum or maximum.
       
  1320 </P>
       
  1321 <P>
       
  1322 If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equivalent
       
  1323 to Perl's /s) is set, thus allowing the dot to match newlines, the pattern is
       
  1324 implicitly anchored, because whatever follows will be tried against every
       
  1325 character position in the subject string, so there is no point in retrying the
       
  1326 overall match at any position after the first. PCRE normally treats such a
       
  1327 pattern as though it were preceded by \A.
       
  1328 </P>
       
  1329 <P>
       
  1330 In cases where it is known that the subject string contains no newlines, it is
       
  1331 worth setting PCRE_DOTALL in order to obtain this optimization, or
       
  1332 alternatively using ^ to indicate anchoring explicitly.
       
  1333 </P>
       
  1334 <P>
       
  1335 However, there is one situation where the optimization cannot be used. When .*
       
  1336 is inside capturing parentheses that are the subject of a backreference
       
  1337 elsewhere in the pattern, a match at the start may fail where a later one
       
  1338 succeeds. Consider, for example:
       
  1339 <pre>
       
  1340   (.*)abc\1
       
  1341 </pre>
       
  1342 If the subject is "xyz123abc123" the match point is the fourth character. For
       
  1343 this reason, such a pattern is not implicitly anchored.
       
  1344 </P>
       
  1345 <P>
       
  1346 When a capturing subpattern is repeated, the value captured is the substring
       
  1347 that matched the final iteration. For example, after
       
  1348 <pre>
       
  1349   (tweedle[dume]{3}\s*)+
       
  1350 </pre>
       
  1351 has matched "tweedledum tweedledee" the value of the captured substring is
       
  1352 "tweedledee". However, if there are nested capturing subpatterns, the
       
  1353 corresponding captured values may have been set in previous iterations. For
       
  1354 example, after
       
  1355 <pre>
       
  1356   /(a|(b))+/
       
  1357 </pre>
       
  1358 matches "aba" the value of the second captured substring is "b".
       
  1359 <a name="atomicgroup"></a></P>
       
  1360 <br><a name="SEC16" href="#TOC1">ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS</a><br>
       
  1361 <P>
       
  1362 With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
       
  1363 repetition, failure of what follows normally causes the repeated item to be
       
  1364 re-evaluated to see if a different number of repeats allows the rest of the
       
  1365 pattern to match. Sometimes it is useful to prevent this, either to change the
       
  1366 nature of the match, or to cause it fail earlier than it otherwise might, when
       
  1367 the author of the pattern knows there is no point in carrying on.
       
  1368 </P>
       
  1369 <P>
       
  1370 Consider, for example, the pattern \d+foo when applied to the subject line
       
  1371 <pre>
       
  1372   123456bar
       
  1373 </pre>
       
  1374 After matching all 6 digits and then failing to match "foo", the normal
       
  1375 action of the matcher is to try again with only 5 digits matching the \d+
       
  1376 item, and then with 4, and so on, before ultimately failing. "Atomic grouping"
       
  1377 (a term taken from Jeffrey Friedl's book) provides the means for specifying
       
  1378 that once a subpattern has matched, it is not to be re-evaluated in this way.
       
  1379 </P>
       
  1380 <P>
       
  1381 If we use atomic grouping for the previous example, the matcher gives up
       
  1382 immediately on failing to match "foo" the first time. The notation is a kind of
       
  1383 special parenthesis, starting with (?&#62; as in this example:
       
  1384 <pre>
       
  1385   (?&#62;\d+)foo
       
  1386 </pre>
       
  1387 This kind of parenthesis "locks up" the  part of the pattern it contains once
       
  1388 it has matched, and a failure further into the pattern is prevented from
       
  1389 backtracking into it. Backtracking past it to previous items, however, works as
       
  1390 normal.
       
  1391 </P>
       
  1392 <P>
       
  1393 An alternative description is that a subpattern of this type matches the string
       
  1394 of characters that an identical standalone pattern would match, if anchored at
       
  1395 the current point in the subject string.
       
  1396 </P>
       
  1397 <P>
       
  1398 Atomic grouping subpatterns are not capturing subpatterns. Simple cases such as
       
  1399 the above example can be thought of as a maximizing repeat that must swallow
       
  1400 everything it can. So, while both \d+ and \d+? are prepared to adjust the
       
  1401 number of digits they match in order to make the rest of the pattern match,
       
  1402 (?&#62;\d+) can only match an entire sequence of digits.
       
  1403 </P>
       
  1404 <P>
       
  1405 Atomic groups in general can of course contain arbitrarily complicated
       
  1406 subpatterns, and can be nested. However, when the subpattern for an atomic
       
  1407 group is just a single repeated item, as in the example above, a simpler
       
  1408 notation, called a "possessive quantifier" can be used. This consists of an
       
  1409 additional + character following a quantifier. Using this notation, the
       
  1410 previous example can be rewritten as
       
  1411 <pre>
       
  1412   \d++foo
       
  1413 </pre>
       
  1414 Note that a possessive quantifier can be used with an entire group, for
       
  1415 example:
       
  1416 <pre>
       
  1417   (abc|xyz){2,3}+
       
  1418 </pre>
       
  1419 Possessive quantifiers are always greedy; the setting of the PCRE_UNGREEDY
       
  1420 option is ignored. They are a convenient notation for the simpler forms of
       
  1421 atomic group. However, there is no difference in the meaning of a possessive
       
  1422 quantifier and the equivalent atomic group, though there may be a performance
       
  1423 difference; possessive quantifiers should be slightly faster.
       
  1424 </P>
       
  1425 <P>
       
  1426 The possessive quantifier syntax is an extension to the Perl 5.8 syntax.
       
  1427 Jeffrey Friedl originated the idea (and the name) in the first edition of his
       
  1428 book. Mike McCloskey liked it, so implemented it when he built Sun's Java
       
  1429 package, and PCRE copied it from there. It ultimately found its way into Perl
       
  1430 at release 5.10.
       
  1431 </P>
       
  1432 <P>
       
  1433 PCRE has an optimization that automatically "possessifies" certain simple
       
  1434 pattern constructs. For example, the sequence A+B is treated as A++B because
       
  1435 there is no point in backtracking into a sequence of A's when B must follow.
       
  1436 </P>
       
  1437 <P>
       
  1438 When a pattern contains an unlimited repeat inside a subpattern that can itself
       
  1439 be repeated an unlimited number of times, the use of an atomic group is the
       
  1440 only way to avoid some failing matches taking a very long time indeed. The
       
  1441 pattern
       
  1442 <pre>
       
  1443   (\D+|&#60;\d+&#62;)*[!?]
       
  1444 </pre>
       
  1445 matches an unlimited number of substrings that either consist of non-digits, or
       
  1446 digits enclosed in &#60;&#62;, followed by either ! or ?. When it matches, it runs
       
  1447 quickly. However, if it is applied to
       
  1448 <pre>
       
  1449   aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
       
  1450 </pre>
       
  1451 it takes a long time before reporting failure. This is because the string can
       
  1452 be divided between the internal \D+ repeat and the external * repeat in a
       
  1453 large number of ways, and all have to be tried. (The example uses [!?] rather
       
  1454 than a single character at the end, because both PCRE and Perl have an
       
  1455 optimization that allows for fast failure when a single character is used. They
       
  1456 remember the last single character that is required for a match, and fail early
       
  1457 if it is not present in the string.) If the pattern is changed so that it uses
       
  1458 an atomic group, like this:
       
  1459 <pre>
       
  1460   ((?&#62;\D+)|&#60;\d+&#62;)*[!?]
       
  1461 </pre>
       
  1462 sequences of non-digits cannot be broken, and failure happens quickly.
       
  1463 <a name="backreferences"></a></P>
       
  1464 <br><a name="SEC17" href="#TOC1">BACK REFERENCES</a><br>
       
  1465 <P>
       
  1466 Outside a character class, a backslash followed by a digit greater than 0 (and
       
  1467 possibly further digits) is a back reference to a capturing subpattern earlier
       
  1468 (that is, to its left) in the pattern, provided there have been that many
       
  1469 previous capturing left parentheses.
       
  1470 </P>
       
  1471 <P>
       
  1472 However, if the decimal number following the backslash is less than 10, it is
       
  1473 always taken as a back reference, and causes an error only if there are not
       
  1474 that many capturing left parentheses in the entire pattern. In other words, the
       
  1475 parentheses that are referenced need not be to the left of the reference for
       
  1476 numbers less than 10. A "forward back reference" of this type can make sense
       
  1477 when a repetition is involved and the subpattern to the right has participated
       
  1478 in an earlier iteration.
       
  1479 </P>
       
  1480 <P>
       
  1481 It is not possible to have a numerical "forward back reference" to a subpattern
       
  1482 whose number is 10 or more using this syntax because a sequence such as \50 is
       
  1483 interpreted as a character defined in octal. See the subsection entitled
       
  1484 "Non-printing characters"
       
  1485 <a href="#digitsafterbackslash">above</a>
       
  1486 for further details of the handling of digits following a backslash. There is
       
  1487 no such problem when named parentheses are used. A back reference to any
       
  1488 subpattern is possible using named parentheses (see below).
       
  1489 </P>
       
  1490 <P>
       
  1491 Another way of avoiding the ambiguity inherent in the use of digits following a
       
  1492 backslash is to use the \g escape sequence, which is a feature introduced in
       
  1493 Perl 5.10. This escape must be followed by an unsigned number or a negative
       
  1494 number, optionally enclosed in braces. These examples are all identical:
       
  1495 <pre>
       
  1496   (ring), \1
       
  1497   (ring), \g1
       
  1498   (ring), \g{1}
       
  1499 </pre>
       
  1500 An unsigned number specifies an absolute reference without the ambiguity that
       
  1501 is present in the older syntax. It is also useful when literal digits follow
       
  1502 the reference. A negative number is a relative reference. Consider this
       
  1503 example:
       
  1504 <pre>
       
  1505   (abc(def)ghi)\g{-1}
       
  1506 </pre>
       
  1507 The sequence \g{-1} is a reference to the most recently started capturing
       
  1508 subpattern before \g, that is, is it equivalent to \2. Similarly, \g{-2}
       
  1509 would be equivalent to \1. The use of relative references can be helpful in
       
  1510 long patterns, and also in patterns that are created by joining together
       
  1511 fragments that contain references within themselves.
       
  1512 </P>
       
  1513 <P>
       
  1514 A back reference matches whatever actually matched the capturing subpattern in
       
  1515 the current subject string, rather than anything matching the subpattern
       
  1516 itself (see
       
  1517 <a href="#subpatternsassubroutines">"Subpatterns as subroutines"</a>
       
  1518 below for a way of doing that). So the pattern
       
  1519 <pre>
       
  1520   (sens|respons)e and \1ibility
       
  1521 </pre>
       
  1522 matches "sense and sensibility" and "response and responsibility", but not
       
  1523 "sense and responsibility". If caseful matching is in force at the time of the
       
  1524 back reference, the case of letters is relevant. For example,
       
  1525 <pre>
       
  1526   ((?i)rah)\s+\1
       
  1527 </pre>
       
  1528 matches "rah rah" and "RAH RAH", but not "RAH rah", even though the original
       
  1529 capturing subpattern is matched caselessly.
       
  1530 </P>
       
  1531 <P>
       
  1532 There are several different ways of writing back references to named
       
  1533 subpatterns. The .NET syntax \k{name} and the Perl syntax \k&#60;name&#62; or
       
  1534 \k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's unified
       
  1535 back reference syntax, in which \g can be used for both numeric and named
       
  1536 references, is also supported. We could rewrite the above example in any of
       
  1537 the following ways:
       
  1538 <pre>
       
  1539   (?&#60;p1&#62;(?i)rah)\s+\k&#60;p1&#62;
       
  1540   (?'p1'(?i)rah)\s+\k{p1}
       
  1541   (?P&#60;p1&#62;(?i)rah)\s+(?P=p1)
       
  1542   (?&#60;p1&#62;(?i)rah)\s+\g{p1}
       
  1543 </pre>
       
  1544 A subpattern that is referenced by name may appear in the pattern before or
       
  1545 after the reference.
       
  1546 </P>
       
  1547 <P>
       
  1548 There may be more than one back reference to the same subpattern. If a
       
  1549 subpattern has not actually been used in a particular match, any back
       
  1550 references to it always fail. For example, the pattern
       
  1551 <pre>
       
  1552   (a|(bc))\2
       
  1553 </pre>
       
  1554 always fails if it starts to match "a" rather than "bc". Because there may be
       
  1555 many capturing parentheses in a pattern, all digits following the backslash are
       
  1556 taken as part of a potential back reference number. If the pattern continues
       
  1557 with a digit character, some delimiter must be used to terminate the back
       
  1558 reference. If the PCRE_EXTENDED option is set, this can be whitespace.
       
  1559 Otherwise an empty comment (see
       
  1560 <a href="#comments">"Comments"</a>
       
  1561 below) can be used.
       
  1562 </P>
       
  1563 <P>
       
  1564 A back reference that occurs inside the parentheses to which it refers fails
       
  1565 when the subpattern is first used, so, for example, (a\1) never matches.
       
  1566 However, such references can be useful inside repeated subpatterns. For
       
  1567 example, the pattern
       
  1568 <pre>
       
  1569   (a|b\1)+
       
  1570 </pre>
       
  1571 matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration of
       
  1572 the subpattern, the back reference matches the character string corresponding
       
  1573 to the previous iteration. In order for this to work, the pattern must be such
       
  1574 that the first iteration does not need to match the back reference. This can be
       
  1575 done using alternation, as in the example above, or by a quantifier with a
       
  1576 minimum of zero.
       
  1577 <a name="bigassertions"></a></P>
       
  1578 <br><a name="SEC18" href="#TOC1">ASSERTIONS</a><br>
       
  1579 <P>
       
  1580 An assertion is a test on the characters following or preceding the current
       
  1581 matching point that does not actually consume any characters. The simple
       
  1582 assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are described
       
  1583 <a href="#smallassertions">above.</a>
       
  1584 </P>
       
  1585 <P>
       
  1586 More complicated assertions are coded as subpatterns. There are two kinds:
       
  1587 those that look ahead of the current position in the subject string, and those
       
  1588 that look behind it. An assertion subpattern is matched in the normal way,
       
  1589 except that it does not cause the current matching position to be changed.
       
  1590 </P>
       
  1591 <P>
       
  1592 Assertion subpatterns are not capturing subpatterns, and may not be repeated,
       
  1593 because it makes no sense to assert the same thing several times. If any kind
       
  1594 of assertion contains capturing subpatterns within it, these are counted for
       
  1595 the purposes of numbering the capturing subpatterns in the whole pattern.
       
  1596 However, substring capturing is carried out only for positive assertions,
       
  1597 because it does not make sense for negative assertions.
       
  1598 </P>
       
  1599 <br><b>
       
  1600 Lookahead assertions
       
  1601 </b><br>
       
  1602 <P>
       
  1603 Lookahead assertions start with (?= for positive assertions and (?! for
       
  1604 negative assertions. For example,
       
  1605 <pre>
       
  1606   \w+(?=;)
       
  1607 </pre>
       
  1608 matches a word followed by a semicolon, but does not include the semicolon in
       
  1609 the match, and
       
  1610 <pre>
       
  1611   foo(?!bar)
       
  1612 </pre>
       
  1613 matches any occurrence of "foo" that is not followed by "bar". Note that the
       
  1614 apparently similar pattern
       
  1615 <pre>
       
  1616   (?!foo)bar
       
  1617 </pre>
       
  1618 does not find an occurrence of "bar" that is preceded by something other than
       
  1619 "foo"; it finds any occurrence of "bar" whatsoever, because the assertion
       
  1620 (?!foo) is always true when the next three characters are "bar". A
       
  1621 lookbehind assertion is needed to achieve the other effect.
       
  1622 </P>
       
  1623 <P>
       
  1624 If you want to force a matching failure at some point in a pattern, the most
       
  1625 convenient way to do it is with (?!) because an empty string always matches, so
       
  1626 an assertion that requires there not to be an empty string must always fail.
       
  1627 <a name="lookbehind"></a></P>
       
  1628 <br><b>
       
  1629 Lookbehind assertions
       
  1630 </b><br>
       
  1631 <P>
       
  1632 Lookbehind assertions start with (?&#60;= for positive assertions and (?&#60;! for
       
  1633 negative assertions. For example,
       
  1634 <pre>
       
  1635   (?&#60;!foo)bar
       
  1636 </pre>
       
  1637 does find an occurrence of "bar" that is not preceded by "foo". The contents of
       
  1638 a lookbehind assertion are restricted such that all the strings it matches must
       
  1639 have a fixed length. However, if there are several top-level alternatives, they
       
  1640 do not all have to have the same fixed length. Thus
       
  1641 <pre>
       
  1642   (?&#60;=bullock|donkey)
       
  1643 </pre>
       
  1644 is permitted, but
       
  1645 <pre>
       
  1646   (?&#60;!dogs?|cats?)
       
  1647 </pre>
       
  1648 causes an error at compile time. Branches that match different length strings
       
  1649 are permitted only at the top level of a lookbehind assertion. This is an
       
  1650 extension compared with Perl (at least for 5.8), which requires all branches to
       
  1651 match the same length of string. An assertion such as
       
  1652 <pre>
       
  1653   (?&#60;=ab(c|de))
       
  1654 </pre>
       
  1655 is not permitted, because its single top-level branch can match two different
       
  1656 lengths, but it is acceptable if rewritten to use two top-level branches:
       
  1657 <pre>
       
  1658   (?&#60;=abc|abde)
       
  1659 </pre>
       
  1660 In some cases, the Perl 5.10 escape sequence \K
       
  1661 <a href="#resetmatchstart">(see above)</a>
       
  1662 can be used instead of a lookbehind assertion; this is not restricted to a
       
  1663 fixed-length.
       
  1664 </P>
       
  1665 <P>
       
  1666 The implementation of lookbehind assertions is, for each alternative, to
       
  1667 temporarily move the current position back by the fixed length and then try to
       
  1668 match. If there are insufficient characters before the current position, the
       
  1669 assertion fails.
       
  1670 </P>
       
  1671 <P>
       
  1672 PCRE does not allow the \C escape (which matches a single byte in UTF-8 mode)
       
  1673 to appear in lookbehind assertions, because it makes it impossible to calculate
       
  1674 the length of the lookbehind. The \X and \R escapes, which can match
       
  1675 different numbers of bytes, are also not permitted.
       
  1676 </P>
       
  1677 <P>
       
  1678 Possessive quantifiers can be used in conjunction with lookbehind assertions to
       
  1679 specify efficient matching at the end of the subject string. Consider a simple
       
  1680 pattern such as
       
  1681 <pre>
       
  1682   abcd$
       
  1683 </pre>
       
  1684 when applied to a long string that does not match. Because matching proceeds
       
  1685 from left to right, PCRE will look for each "a" in the subject and then see if
       
  1686 what follows matches the rest of the pattern. If the pattern is specified as
       
  1687 <pre>
       
  1688   ^.*abcd$
       
  1689 </pre>
       
  1690 the initial .* matches the entire string at first, but when this fails (because
       
  1691 there is no following "a"), it backtracks to match all but the last character,
       
  1692 then all but the last two characters, and so on. Once again the search for "a"
       
  1693 covers the entire string, from right to left, so we are no better off. However,
       
  1694 if the pattern is written as
       
  1695 <pre>
       
  1696   ^.*+(?&#60;=abcd)
       
  1697 </pre>
       
  1698 there can be no backtracking for the .*+ item; it can match only the entire
       
  1699 string. The subsequent lookbehind assertion does a single test on the last four
       
  1700 characters. If it fails, the match fails immediately. For long strings, this
       
  1701 approach makes a significant difference to the processing time.
       
  1702 </P>
       
  1703 <br><b>
       
  1704 Using multiple assertions
       
  1705 </b><br>
       
  1706 <P>
       
  1707 Several assertions (of any sort) may occur in succession. For example,
       
  1708 <pre>
       
  1709   (?&#60;=\d{3})(?&#60;!999)foo
       
  1710 </pre>
       
  1711 matches "foo" preceded by three digits that are not "999". Notice that each of
       
  1712 the assertions is applied independently at the same point in the subject
       
  1713 string. First there is a check that the previous three characters are all
       
  1714 digits, and then there is a check that the same three characters are not "999".
       
  1715 This pattern does <i>not</i> match "foo" preceded by six characters, the first
       
  1716 of which are digits and the last three of which are not "999". For example, it
       
  1717 doesn't match "123abcfoo". A pattern to do that is
       
  1718 <pre>
       
  1719   (?&#60;=\d{3}...)(?&#60;!999)foo
       
  1720 </pre>
       
  1721 This time the first assertion looks at the preceding six characters, checking
       
  1722 that the first three are digits, and then the second assertion checks that the
       
  1723 preceding three characters are not "999".
       
  1724 </P>
       
  1725 <P>
       
  1726 Assertions can be nested in any combination. For example,
       
  1727 <pre>
       
  1728   (?&#60;=(?&#60;!foo)bar)baz
       
  1729 </pre>
       
  1730 matches an occurrence of "baz" that is preceded by "bar" which in turn is not
       
  1731 preceded by "foo", while
       
  1732 <pre>
       
  1733   (?&#60;=\d{3}(?!999)...)foo
       
  1734 </pre>
       
  1735 is another pattern that matches "foo" preceded by three digits and any three
       
  1736 characters that are not "999".
       
  1737 <a name="conditions"></a></P>
       
  1738 <br><a name="SEC19" href="#TOC1">CONDITIONAL SUBPATTERNS</a><br>
       
  1739 <P>
       
  1740 It is possible to cause the matching process to obey a subpattern
       
  1741 conditionally or to choose between two alternative subpatterns, depending on
       
  1742 the result of an assertion, or whether a previous capturing subpattern matched
       
  1743 or not. The two possible forms of conditional subpattern are
       
  1744 <pre>
       
  1745   (?(condition)yes-pattern)
       
  1746   (?(condition)yes-pattern|no-pattern)
       
  1747 </pre>
       
  1748 If the condition is satisfied, the yes-pattern is used; otherwise the
       
  1749 no-pattern (if present) is used. If there are more than two alternatives in the
       
  1750 subpattern, a compile-time error occurs.
       
  1751 </P>
       
  1752 <P>
       
  1753 There are four kinds of condition: references to subpatterns, references to
       
  1754 recursion, a pseudo-condition called DEFINE, and assertions.
       
  1755 </P>
       
  1756 <br><b>
       
  1757 Checking for a used subpattern by number
       
  1758 </b><br>
       
  1759 <P>
       
  1760 If the text between the parentheses consists of a sequence of digits, the
       
  1761 condition is true if the capturing subpattern of that number has previously
       
  1762 matched. An alternative notation is to precede the digits with a plus or minus
       
  1763 sign. In this case, the subpattern number is relative rather than absolute.
       
  1764 The most recently opened parentheses can be referenced by (?(-1), the next most
       
  1765 recent by (?(-2), and so on. In looping constructs it can also make sense to
       
  1766 refer to subsequent groups with constructs such as (?(+2).
       
  1767 </P>
       
  1768 <P>
       
  1769 Consider the following pattern, which contains non-significant white space to
       
  1770 make it more readable (assume the PCRE_EXTENDED option) and to divide it into
       
  1771 three parts for ease of discussion:
       
  1772 <pre>
       
  1773   ( \( )?    [^()]+    (?(1) \) )
       
  1774 </pre>
       
  1775 The first part matches an optional opening parenthesis, and if that
       
  1776 character is present, sets it as the first captured substring. The second part
       
  1777 matches one or more characters that are not parentheses. The third part is a
       
  1778 conditional subpattern that tests whether the first set of parentheses matched
       
  1779 or not. If they did, that is, if subject started with an opening parenthesis,
       
  1780 the condition is true, and so the yes-pattern is executed and a closing
       
  1781 parenthesis is required. Otherwise, since no-pattern is not present, the
       
  1782 subpattern matches nothing. In other words, this pattern matches a sequence of
       
  1783 non-parentheses, optionally enclosed in parentheses.
       
  1784 </P>
       
  1785 <P>
       
  1786 If you were embedding this pattern in a larger one, you could use a relative
       
  1787 reference:
       
  1788 <pre>
       
  1789   ...other stuff... ( \( )?    [^()]+    (?(-1) \) ) ...
       
  1790 </pre>
       
  1791 This makes the fragment independent of the parentheses in the larger pattern.
       
  1792 </P>
       
  1793 <br><b>
       
  1794 Checking for a used subpattern by name
       
  1795 </b><br>
       
  1796 <P>
       
  1797 Perl uses the syntax (?(&#60;name&#62;)...) or (?('name')...) to test for a used
       
  1798 subpattern by name. For compatibility with earlier versions of PCRE, which had
       
  1799 this facility before Perl, the syntax (?(name)...) is also recognized. However,
       
  1800 there is a possible ambiguity with this syntax, because subpattern names may
       
  1801 consist entirely of digits. PCRE looks first for a named subpattern; if it
       
  1802 cannot find one and the name consists entirely of digits, PCRE looks for a
       
  1803 subpattern of that number, which must be greater than zero. Using subpattern
       
  1804 names that consist entirely of digits is not recommended.
       
  1805 </P>
       
  1806 <P>
       
  1807 Rewriting the above example to use a named subpattern gives this:
       
  1808 <pre>
       
  1809   (?&#60;OPEN&#62; \( )?    [^()]+    (?(&#60;OPEN&#62;) \) )
       
  1810 
       
  1811 </PRE>
       
  1812 </P>
       
  1813 <br><b>
       
  1814 Checking for pattern recursion
       
  1815 </b><br>
       
  1816 <P>
       
  1817 If the condition is the string (R), and there is no subpattern with the name R,
       
  1818 the condition is true if a recursive call to the whole pattern or any
       
  1819 subpattern has been made. If digits or a name preceded by ampersand follow the
       
  1820 letter R, for example:
       
  1821 <pre>
       
  1822   (?(R3)...) or (?(R&name)...)
       
  1823 </pre>
       
  1824 the condition is true if the most recent recursion is into the subpattern whose
       
  1825 number or name is given. This condition does not check the entire recursion
       
  1826 stack.
       
  1827 </P>
       
  1828 <P>
       
  1829 At "top level", all these recursion test conditions are false. Recursive
       
  1830 patterns are described below.
       
  1831 </P>
       
  1832 <br><b>
       
  1833 Defining subpatterns for use by reference only
       
  1834 </b><br>
       
  1835 <P>
       
  1836 If the condition is the string (DEFINE), and there is no subpattern with the
       
  1837 name DEFINE, the condition is always false. In this case, there may be only one
       
  1838 alternative in the subpattern. It is always skipped if control reaches this
       
  1839 point in the pattern; the idea of DEFINE is that it can be used to define
       
  1840 "subroutines" that can be referenced from elsewhere. (The use of "subroutines"
       
  1841 is described below.) For example, a pattern to match an IPv4 address could be
       
  1842 written like this (ignore whitespace and line breaks):
       
  1843 <pre>
       
  1844   (?(DEFINE) (?&#60;byte&#62; 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
       
  1845   \b (?&byte) (\.(?&byte)){3} \b
       
  1846 </pre>
       
  1847 The first part of the pattern is a DEFINE group inside which a another group
       
  1848 named "byte" is defined. This matches an individual component of an IPv4
       
  1849 address (a number less than 256). When matching takes place, this part of the
       
  1850 pattern is skipped because DEFINE acts like a false condition.
       
  1851 </P>
       
  1852 <P>
       
  1853 The rest of the pattern uses references to the named group to match the four
       
  1854 dot-separated components of an IPv4 address, insisting on a word boundary at
       
  1855 each end.
       
  1856 </P>
       
  1857 <br><b>
       
  1858 Assertion conditions
       
  1859 </b><br>
       
  1860 <P>
       
  1861 If the condition is not in any of the above formats, it must be an assertion.
       
  1862 This may be a positive or negative lookahead or lookbehind assertion. Consider
       
  1863 this pattern, again containing non-significant white space, and with the two
       
  1864 alternatives on the second line:
       
  1865 <pre>
       
  1866   (?(?=[^a-z]*[a-z])
       
  1867   \d{2}-[a-z]{3}-\d{2}  |  \d{2}-\d{2}-\d{2} )
       
  1868 </pre>
       
  1869 The condition is a positive lookahead assertion that matches an optional
       
  1870 sequence of non-letters followed by a letter. In other words, it tests for the
       
  1871 presence of at least one letter in the subject. If a letter is found, the
       
  1872 subject is matched against the first alternative; otherwise it is matched
       
  1873 against the second. This pattern matches strings in one of the two forms
       
  1874 dd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.
       
  1875 <a name="comments"></a></P>
       
  1876 <br><a name="SEC20" href="#TOC1">COMMENTS</a><br>
       
  1877 <P>
       
  1878 The sequence (?# marks the start of a comment that continues up to the next
       
  1879 closing parenthesis. Nested parentheses are not permitted. The characters
       
  1880 that make up a comment play no part in the pattern matching at all.
       
  1881 </P>
       
  1882 <P>
       
  1883 If the PCRE_EXTENDED option is set, an unescaped # character outside a
       
  1884 character class introduces a comment that continues to immediately after the
       
  1885 next newline in the pattern.
       
  1886 <a name="recursion"></a></P>
       
  1887 <br><a name="SEC21" href="#TOC1">RECURSIVE PATTERNS</a><br>
       
  1888 <P>
       
  1889 Consider the problem of matching a string in parentheses, allowing for
       
  1890 unlimited nested parentheses. Without the use of recursion, the best that can
       
  1891 be done is to use a pattern that matches up to some fixed depth of nesting. It
       
  1892 is not possible to handle an arbitrary nesting depth.
       
  1893 </P>
       
  1894 <P>
       
  1895 For some time, Perl has provided a facility that allows regular expressions to
       
  1896 recurse (amongst other things). It does this by interpolating Perl code in the
       
  1897 expression at run time, and the code can refer to the expression itself. A Perl
       
  1898 pattern using code interpolation to solve the parentheses problem can be
       
  1899 created like this:
       
  1900 <pre>
       
  1901   $re = qr{\( (?: (?&#62;[^()]+) | (?p{$re}) )* \)}x;
       
  1902 </pre>
       
  1903 The (?p{...}) item interpolates Perl code at run time, and in this case refers
       
  1904 recursively to the pattern in which it appears.
       
  1905 </P>
       
  1906 <P>
       
  1907 Obviously, PCRE cannot support the interpolation of Perl code. Instead, it
       
  1908 supports special syntax for recursion of the entire pattern, and also for
       
  1909 individual subpattern recursion. After its introduction in PCRE and Python,
       
  1910 this kind of recursion was introduced into Perl at release 5.10.
       
  1911 </P>
       
  1912 <P>
       
  1913 A special item that consists of (? followed by a number greater than zero and a
       
  1914 closing parenthesis is a recursive call of the subpattern of the given number,
       
  1915 provided that it occurs inside that subpattern. (If not, it is a "subroutine"
       
  1916 call, which is described in the next section.) The special item (?R) or (?0) is
       
  1917 a recursive call of the entire regular expression.
       
  1918 </P>
       
  1919 <P>
       
  1920 In PCRE (like Python, but unlike Perl), a recursive subpattern call is always
       
  1921 treated as an atomic group. That is, once it has matched some of the subject
       
  1922 string, it is never re-entered, even if it contains untried alternatives and
       
  1923 there is a subsequent matching failure.
       
  1924 </P>
       
  1925 <P>
       
  1926 This PCRE pattern solves the nested parentheses problem (assume the
       
  1927 PCRE_EXTENDED option is set so that white space is ignored):
       
  1928 <pre>
       
  1929   \( ( (?&#62;[^()]+) | (?R) )* \)
       
  1930 </pre>
       
  1931 First it matches an opening parenthesis. Then it matches any number of
       
  1932 substrings which can either be a sequence of non-parentheses, or a recursive
       
  1933 match of the pattern itself (that is, a correctly parenthesized substring).
       
  1934 Finally there is a closing parenthesis.
       
  1935 </P>
       
  1936 <P>
       
  1937 If this were part of a larger pattern, you would not want to recurse the entire
       
  1938 pattern, so instead you could use this:
       
  1939 <pre>
       
  1940   ( \( ( (?&#62;[^()]+) | (?1) )* \) )
       
  1941 </pre>
       
  1942 We have put the pattern into parentheses, and caused the recursion to refer to
       
  1943 them instead of the whole pattern.
       
  1944 </P>
       
  1945 <P>
       
  1946 In a larger pattern, keeping track of parenthesis numbers can be tricky. This
       
  1947 is made easier by the use of relative references. (A Perl 5.10 feature.)
       
  1948 Instead of (?1) in the pattern above you can write (?-2) to refer to the second
       
  1949 most recently opened parentheses preceding the recursion. In other words, a
       
  1950 negative number counts capturing parentheses leftwards from the point at which
       
  1951 it is encountered.
       
  1952 </P>
       
  1953 <P>
       
  1954 It is also possible to refer to subsequently opened parentheses, by writing
       
  1955 references such as (?+2). However, these cannot be recursive because the
       
  1956 reference is not inside the parentheses that are referenced. They are always
       
  1957 "subroutine" calls, as described in the next section.
       
  1958 </P>
       
  1959 <P>
       
  1960 An alternative approach is to use named parentheses instead. The Perl syntax
       
  1961 for this is (?&name); PCRE's earlier syntax (?P&#62;name) is also supported. We
       
  1962 could rewrite the above example as follows:
       
  1963 <pre>
       
  1964   (?&#60;pn&#62; \( ( (?&#62;[^()]+) | (?&pn) )* \) )
       
  1965 </pre>
       
  1966 If there is more than one subpattern with the same name, the earliest one is
       
  1967 used.
       
  1968 </P>
       
  1969 <P>
       
  1970 This particular example pattern that we have been looking at contains nested
       
  1971 unlimited repeats, and so the use of atomic grouping for matching strings of
       
  1972 non-parentheses is important when applying the pattern to strings that do not
       
  1973 match. For example, when this pattern is applied to
       
  1974 <pre>
       
  1975   (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
       
  1976 </pre>
       
  1977 it yields "no match" quickly. However, if atomic grouping is not used,
       
  1978 the match runs for a very long time indeed because there are so many different
       
  1979 ways the + and * repeats can carve up the subject, and all have to be tested
       
  1980 before failure can be reported.
       
  1981 </P>
       
  1982 <P>
       
  1983 At the end of a match, the values set for any capturing subpatterns are those
       
  1984 from the outermost level of the recursion at which the subpattern value is set.
       
  1985 If you want to obtain intermediate values, a callout function can be used (see
       
  1986 below and the
       
  1987 <a href="pcrecallout.html"><b>pcrecallout</b></a>
       
  1988 documentation). If the pattern above is matched against
       
  1989 <pre>
       
  1990   (ab(cd)ef)
       
  1991 </pre>
       
  1992 the value for the capturing parentheses is "ef", which is the last value taken
       
  1993 on at the top level. If additional parentheses are added, giving
       
  1994 <pre>
       
  1995   \( ( ( (?&#62;[^()]+) | (?R) )* ) \)
       
  1996      ^                        ^
       
  1997      ^                        ^
       
  1998 </pre>
       
  1999 the string they capture is "ab(cd)ef", the contents of the top level
       
  2000 parentheses. If there are more than 15 capturing parentheses in a pattern, PCRE
       
  2001 has to obtain extra memory to store data during a recursion, which it does by
       
  2002 using <b>pcre_malloc</b>, freeing it via <b>pcre_free</b> afterwards. If no
       
  2003 memory can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
       
  2004 </P>
       
  2005 <P>
       
  2006 Do not confuse the (?R) item with the condition (R), which tests for recursion.
       
  2007 Consider this pattern, which matches text in angle brackets, allowing for
       
  2008 arbitrary nesting. Only digits are allowed in nested brackets (that is, when
       
  2009 recursing), whereas any characters are permitted at the outer level.
       
  2010 <pre>
       
  2011   &#60; (?: (?(R) \d++  | [^&#60;&#62;]*+) | (?R)) * &#62;
       
  2012 </pre>
       
  2013 In this pattern, (?(R) is the start of a conditional subpattern, with two
       
  2014 different alternatives for the recursive and non-recursive cases. The (?R) item
       
  2015 is the actual recursive call.
       
  2016 <a name="subpatternsassubroutines"></a></P>
       
  2017 <br><a name="SEC22" href="#TOC1">SUBPATTERNS AS SUBROUTINES</a><br>
       
  2018 <P>
       
  2019 If the syntax for a recursive subpattern reference (either by number or by
       
  2020 name) is used outside the parentheses to which it refers, it operates like a
       
  2021 subroutine in a programming language. The "called" subpattern may be defined
       
  2022 before or after the reference. A numbered reference can be absolute or
       
  2023 relative, as in these examples:
       
  2024 <pre>
       
  2025   (...(absolute)...)...(?2)...
       
  2026   (...(relative)...)...(?-1)...
       
  2027   (...(?+1)...(relative)...
       
  2028 </pre>
       
  2029 An earlier example pointed out that the pattern
       
  2030 <pre>
       
  2031   (sens|respons)e and \1ibility
       
  2032 </pre>
       
  2033 matches "sense and sensibility" and "response and responsibility", but not
       
  2034 "sense and responsibility". If instead the pattern
       
  2035 <pre>
       
  2036   (sens|respons)e and (?1)ibility
       
  2037 </pre>
       
  2038 is used, it does match "sense and responsibility" as well as the other two
       
  2039 strings. Another example is given in the discussion of DEFINE above.
       
  2040 </P>
       
  2041 <P>
       
  2042 Like recursive subpatterns, a "subroutine" call is always treated as an atomic
       
  2043 group. That is, once it has matched some of the subject string, it is never
       
  2044 re-entered, even if it contains untried alternatives and there is a subsequent
       
  2045 matching failure.
       
  2046 </P>
       
  2047 <P>
       
  2048 When a subpattern is used as a subroutine, processing options such as
       
  2049 case-independence are fixed when the subpattern is defined. They cannot be
       
  2050 changed for different calls. For example, consider this pattern:
       
  2051 <pre>
       
  2052   (abc)(?i:(?-1))
       
  2053 </pre>
       
  2054 It matches "abcabc". It does not match "abcABC" because the change of
       
  2055 processing option does not affect the called subpattern.
       
  2056 <a name="onigurumasubroutines"></a></P>
       
  2057 <br><a name="SEC23" href="#TOC1">ONIGURUMA SUBROUTINE SYNTAX</a><br>
       
  2058 <P>
       
  2059 For compatibility with Oniguruma, the non-Perl syntax \g followed by a name or
       
  2060 a number enclosed either in angle brackets or single quotes, is an alternative
       
  2061 syntax for referencing a subpattern as a subroutine, possibly recursively. Here
       
  2062 are two of the examples used above, rewritten using this syntax:
       
  2063 <pre>
       
  2064   (?&#60;pn&#62; \( ( (?&#62;[^()]+) | \g&#60;pn&#62; )* \) )
       
  2065   (sens|respons)e and \g'1'ibility
       
  2066 </pre>
       
  2067 PCRE supports an extension to Oniguruma: if a number is preceded by a
       
  2068 plus or a minus sign it is taken as a relative reference. For example:
       
  2069 <pre>
       
  2070   (abc)(?i:\g&#60;-1&#62;)
       
  2071 </pre>
       
  2072 Note that \g{...} (Perl syntax) and \g&#60;...&#62; (Oniguruma syntax) are <i>not</i>
       
  2073 synonymous. The former is a back reference; the latter is a subroutine call.
       
  2074 </P>
       
  2075 <br><a name="SEC24" href="#TOC1">CALLOUTS</a><br>
       
  2076 <P>
       
  2077 Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perl
       
  2078 code to be obeyed in the middle of matching a regular expression. This makes it
       
  2079 possible, amongst other things, to extract different substrings that match the
       
  2080 same pair of parentheses when there is a repetition.
       
  2081 </P>
       
  2082 <P>
       
  2083 PCRE provides a similar feature, but of course it cannot obey arbitrary Perl
       
  2084 code. The feature is called "callout". The caller of PCRE provides an external
       
  2085 function by putting its entry point in the global variable <i>pcre_callout</i>.
       
  2086 By default, this variable contains NULL, which disables all calling out.
       
  2087 </P>
       
  2088 <P>
       
  2089 Within a regular expression, (?C) indicates the points at which the external
       
  2090 function is to be called. If you want to identify different callout points, you
       
  2091 can put a number less than 256 after the letter C. The default value is zero.
       
  2092 For example, this pattern has two callout points:
       
  2093 <pre>
       
  2094   (?C1)abc(?C2)def
       
  2095 </pre>
       
  2096 If the PCRE_AUTO_CALLOUT flag is passed to <b>pcre_compile()</b>, callouts are
       
  2097 automatically installed before each item in the pattern. They are all numbered
       
  2098 255.
       
  2099 </P>
       
  2100 <P>
       
  2101 During matching, when PCRE reaches a callout point (and <i>pcre_callout</i> is
       
  2102 set), the external function is called. It is provided with the number of the
       
  2103 callout, the position in the pattern, and, optionally, one item of data
       
  2104 originally supplied by the caller of <b>pcre_exec()</b>. The callout function
       
  2105 may cause matching to proceed, to backtrack, or to fail altogether. A complete
       
  2106 description of the interface to the callout function is given in the
       
  2107 <a href="pcrecallout.html"><b>pcrecallout</b></a>
       
  2108 documentation.
       
  2109 </P>
       
  2110 <br><a name="SEC25" href="#TOC1">BACKTRACKING CONTROL</a><br>
       
  2111 <P>
       
  2112 Perl 5.10 introduced a number of "Special Backtracking Control Verbs", which
       
  2113 are described in the Perl documentation as "experimental and subject to change
       
  2114 or removal in a future version of Perl". It goes on to say: "Their usage in
       
  2115 production code should be noted to avoid problems during upgrades." The same
       
  2116 remarks apply to the PCRE features described in this section.
       
  2117 </P>
       
  2118 <P>
       
  2119 Since these verbs are specifically related to backtracking, most of them can be
       
  2120 used only when the pattern is to be matched using <b>pcre_exec()</b>, which uses
       
  2121 a backtracking algorithm. With the exception of (*FAIL), which behaves like a
       
  2122 failing negative assertion, they cause an error if encountered by
       
  2123 <b>pcre_dfa_exec()</b>.
       
  2124 </P>
       
  2125 <P>
       
  2126 The new verbs make use of what was previously invalid syntax: an opening
       
  2127 parenthesis followed by an asterisk. In Perl, they are generally of the form
       
  2128 (*VERB:ARG) but PCRE does not support the use of arguments, so its general
       
  2129 form is just (*VERB). Any number of these verbs may occur in a pattern. There
       
  2130 are two kinds:
       
  2131 </P>
       
  2132 <br><b>
       
  2133 Verbs that act immediately
       
  2134 </b><br>
       
  2135 <P>
       
  2136 The following verbs act as soon as they are encountered:
       
  2137 <pre>
       
  2138    (*ACCEPT)
       
  2139 </pre>
       
  2140 This verb causes the match to end successfully, skipping the remainder of the
       
  2141 pattern. When inside a recursion, only the innermost pattern is ended
       
  2142 immediately. PCRE differs from Perl in what happens if the (*ACCEPT) is inside
       
  2143 capturing parentheses. In Perl, the data so far is captured: in PCRE no data is
       
  2144 captured. For example:
       
  2145 <pre>
       
  2146   A(A|B(*ACCEPT)|C)D
       
  2147 </pre>
       
  2148 This matches "AB", "AAD", or "ACD", but when it matches "AB", no data is
       
  2149 captured.
       
  2150 <pre>
       
  2151   (*FAIL) or (*F)
       
  2152 </pre>
       
  2153 This verb causes the match to fail, forcing backtracking to occur. It is
       
  2154 equivalent to (?!) but easier to read. The Perl documentation notes that it is
       
  2155 probably useful only when combined with (?{}) or (??{}). Those are, of course,
       
  2156 Perl features that are not present in PCRE. The nearest equivalent is the
       
  2157 callout feature, as for example in this pattern:
       
  2158 <pre>
       
  2159   a+(?C)(*FAIL)
       
  2160 </pre>
       
  2161 A match with the string "aaaa" always fails, but the callout is taken before
       
  2162 each backtrack happens (in this example, 10 times).
       
  2163 </P>
       
  2164 <br><b>
       
  2165 Verbs that act after backtracking
       
  2166 </b><br>
       
  2167 <P>
       
  2168 The following verbs do nothing when they are encountered. Matching continues
       
  2169 with what follows, but if there is no subsequent match, a failure is forced.
       
  2170 The verbs differ in exactly what kind of failure occurs.
       
  2171 <pre>
       
  2172   (*COMMIT)
       
  2173 </pre>
       
  2174 This verb causes the whole match to fail outright if the rest of the pattern
       
  2175 does not match. Even if the pattern is unanchored, no further attempts to find
       
  2176 a match by advancing the start point take place. Once (*COMMIT) has been
       
  2177 passed, <b>pcre_exec()</b> is committed to finding a match at the current
       
  2178 starting point, or not at all. For example:
       
  2179 <pre>
       
  2180   a+(*COMMIT)b
       
  2181 </pre>
       
  2182 This matches "xxaab" but not "aacaab". It can be thought of as a kind of
       
  2183 dynamic anchor, or "I've started, so I must finish."
       
  2184 <pre>
       
  2185   (*PRUNE)
       
  2186 </pre>
       
  2187 This verb causes the match to fail at the current position if the rest of the
       
  2188 pattern does not match. If the pattern is unanchored, the normal "bumpalong"
       
  2189 advance to the next starting character then happens. Backtracking can occur as
       
  2190 usual to the left of (*PRUNE), or when matching to the right of (*PRUNE), but
       
  2191 if there is no match to the right, backtracking cannot cross (*PRUNE).
       
  2192 In simple cases, the use of (*PRUNE) is just an alternative to an atomic
       
  2193 group or possessive quantifier, but there are some uses of (*PRUNE) that cannot
       
  2194 be expressed in any other way.
       
  2195 <pre>
       
  2196   (*SKIP)
       
  2197 </pre>
       
  2198 This verb is like (*PRUNE), except that if the pattern is unanchored, the
       
  2199 "bumpalong" advance is not to the next character, but to the position in the
       
  2200 subject where (*SKIP) was encountered. (*SKIP) signifies that whatever text
       
  2201 was matched leading up to it cannot be part of a successful match. Consider:
       
  2202 <pre>
       
  2203   a+(*SKIP)b
       
  2204 </pre>
       
  2205 If the subject is "aaaac...", after the first match attempt fails (starting at
       
  2206 the first character in the string), the starting point skips on to start the
       
  2207 next attempt at "c". Note that a possessive quantifer does not have the same
       
  2208 effect in this example; although it would suppress backtracking during the
       
  2209 first match attempt, the second attempt would start at the second character
       
  2210 instead of skipping on to "c".
       
  2211 <pre>
       
  2212   (*THEN)
       
  2213 </pre>
       
  2214 This verb causes a skip to the next alternation if the rest of the pattern does
       
  2215 not match. That is, it cancels pending backtracking, but only within the
       
  2216 current alternation. Its name comes from the observation that it can be used
       
  2217 for a pattern-based if-then-else block:
       
  2218 <pre>
       
  2219   ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
       
  2220 </pre>
       
  2221 If the COND1 pattern matches, FOO is tried (and possibly further items after
       
  2222 the end of the group if FOO succeeds); on failure the matcher skips to the
       
  2223 second alternative and tries COND2, without backtracking into COND1. If (*THEN)
       
  2224 is used outside of any alternation, it acts exactly like (*PRUNE).
       
  2225 </P>
       
  2226 <br><a name="SEC26" href="#TOC1">SEE ALSO</a><br>
       
  2227 <P>
       
  2228 <b>pcreapi</b>(3), <b>pcrecallout</b>(3), <b>pcrematching</b>(3), <b>pcre</b>(3).
       
  2229 </P>
       
  2230 <br><a name="SEC27" href="#TOC1">AUTHOR</a><br>
       
  2231 <P>
       
  2232 Philip Hazel
       
  2233 <br>
       
  2234 University Computing Service
       
  2235 <br>
       
  2236 Cambridge CB2 3QH, England.
       
  2237 <br>
       
  2238 </P>
       
  2239 <br><a name="SEC28" href="#TOC1">REVISION</a><br>
       
  2240 <P>
       
  2241 Last updated: 19 April 2008
       
  2242 <br>
       
  2243 Copyright &copy; 1997-2008 University of Cambridge.
       
  2244 <br>
       
  2245 <p>
       
  2246 Return to the <a href="index.html">PCRE index page</a>.
       
  2247 </p>