|
|
<html><head><title>pcre2pattern specification</title></head><body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB"><h1>pcre2pattern man page</h1><p>Return to the <a href="index.html">PCRE2 index page</a>.</p><p>This page is part of the PCRE2 HTML documentation. It was generatedautomatically from the original man page. If there is any nonsense in it,please consult the man page, in case the conversion went wrong.<br><ul><li><a name="TOC1" href="#SEC1">PCRE2 REGULAR EXPRESSION DETAILS</a><li><a name="TOC2" href="#SEC2">SPECIAL START-OF-PATTERN ITEMS</a><li><a name="TOC3" href="#SEC3">EBCDIC CHARACTER CODES</a><li><a name="TOC4" href="#SEC4">CHARACTERS AND METACHARACTERS</a><li><a name="TOC5" href="#SEC5">BACKSLASH</a><li><a name="TOC6" href="#SEC6">CIRCUMFLEX AND DOLLAR</a><li><a name="TOC7" href="#SEC7">FULL STOP (PERIOD, DOT) AND \N</a><li><a name="TOC8" href="#SEC8">MATCHING A SINGLE CODE UNIT</a><li><a name="TOC9" href="#SEC9">SQUARE BRACKETS AND CHARACTER CLASSES</a><li><a name="TOC10" href="#SEC10">POSIX CHARACTER CLASSES</a><li><a name="TOC11" href="#SEC11">COMPATIBILITY FEATURE FOR WORD BOUNDARIES</a><li><a name="TOC12" href="#SEC12">VERTICAL BAR</a><li><a name="TOC13" href="#SEC13">INTERNAL OPTION SETTING</a><li><a name="TOC14" href="#SEC14">GROUPS</a><li><a name="TOC15" href="#SEC15">DUPLICATE GROUP NUMBERS</a><li><a name="TOC16" href="#SEC16">NAMED CAPTURE GROUPS</a><li><a name="TOC17" href="#SEC17">REPETITION</a><li><a name="TOC18" href="#SEC18">ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS</a><li><a name="TOC19" href="#SEC19">BACKREFERENCES</a><li><a name="TOC20" href="#SEC20">ASSERTIONS</a><li><a name="TOC21" href="#SEC21">NON-ATOMIC ASSERTIONS</a><li><a name="TOC22" href="#SEC22">SCRIPT RUNS</a><li><a name="TOC23" href="#SEC23">CONDITIONAL GROUPS</a><li><a name="TOC24" href="#SEC24">COMMENTS</a><li><a name="TOC25" href="#SEC25">RECURSIVE PATTERNS</a><li><a name="TOC26" href="#SEC26">GROUPS AS SUBROUTINES</a><li><a name="TOC27" href="#SEC27">ONIGURUMA SUBROUTINE SYNTAX</a><li><a name="TOC28" href="#SEC28">CALLOUTS</a><li><a name="TOC29" href="#SEC29">BACKTRACKING CONTROL</a><li><a name="TOC30" href="#SEC30">SEE ALSO</a><li><a name="TOC31" href="#SEC31">AUTHOR</a><li><a name="TOC32" href="#SEC32">REVISION</a></ul><br><a name="SEC1" href="#TOC1">PCRE2 REGULAR EXPRESSION DETAILS</a><br><P>The syntax and semantics of the regular expressions that are supported by PCRE2are described in detail below. There is a quick-reference syntax summary in the<a href="pcre2syntax.html"><b>pcre2syntax</b></a>page. PCRE2 tries to match Perl syntax and semantics as closely as it can.PCRE2 also supports some alternative regular expression syntax (which does notconflict with the Perl syntax) in order to provide some compatibility withregular expressions in Python, .NET, and Oniguruma.</P><P>Perl's regular expressions are described in its own documentation, and regularexpressions in general are covered in a number of books, some of which havecopious examples. Jeffrey Friedl's "Mastering Regular Expressions", publishedby O'Reilly, covers regular expressions in great detail. This description ofPCRE2's regular expressions is intended as reference material.</P><P>This document discusses the regular expression patterns that are supported byPCRE2 when its main matching function, <b>pcre2_match()</b>, is used. PCRE2 alsohas an alternative matching function, <b>pcre2_dfa_match()</b>, which matchesusing a different algorithm that is not Perl-compatible. Some of the featuresdiscussed below are not available when DFA matching is used. The advantages anddisadvantages of the alternative function, and how it differs from the normalfunction, are discussed in the<a href="pcre2matching.html"><b>pcre2matching</b></a>page.</P><br><a name="SEC2" href="#TOC1">SPECIAL START-OF-PATTERN ITEMS</a><br><P>A number of options that can be passed to <b>pcre2_compile()</b> can also be setby special items at the start of a pattern. These are not Perl-compatible, butare provided to make these options accessible to pattern writers who are notable to change the program that processes the pattern. Any number of theseitems may appear, but they must all be together right at the start of thepattern string, and the letters must be in upper case.</P><br><b>UTF support</b><br><P>In the 8-bit and 16-bit PCRE2 libraries, characters may be coded either assingle code units, or as multiple UTF-8 or UTF-16 code units. UTF-32 can bespecified for the 32-bit library, in which case it constrains the charactervalues to valid Unicode code points. To process UTF strings, PCRE2 must bebuilt to include Unicode support (which is the default). When using UTF stringsyou must either call the compiling function with one or both of the PCRE2_UTFor PCRE2_MATCH_INVALID_UTF options, or the pattern must start with the specialsequence (*UTF), which is equivalent to setting the relevant PCRE2_UTF. Howsetting a UTF mode affects pattern matching is mentioned in several placesbelow. There is also a summary of features in the<a href="pcre2unicode.html"><b>pcre2unicode</b></a>page.</P><P>Some applications that allow their users to supply patterns may wish torestrict them to non-UTF data for security reasons. If the PCRE2_NEVER_UTFoption is passed to <b>pcre2_compile()</b>, (*UTF) is not allowed, and itsappearance in a pattern causes an error.</P><br><b>Unicode property support</b><br><P>Another special sequence that may appear at the start of a pattern is (*UCP).This has the same effect as setting the PCRE2_UCP option: it causes sequencessuch as \d and \w to use Unicode properties to determine character types,instead of recognizing only characters with codes less than 256 via a lookuptable. If also causes upper/lower casing operations to use Unicode propertiesfor characters with code points greater than 127, even when UTF is not set.These behaviours can be changed within the pattern; see the section entitled<a href="#internaloptions">"Internal Option Setting"</a>below.</P><P>Some applications that allow their users to supply patterns may wish torestrict them for security reasons. If the PCRE2_NEVER_UCP option is passed to<b>pcre2_compile()</b>, (*UCP) is not allowed, and its appearance in a patterncauses an error.</P><br><b>Locking out empty string matching</b><br><P>Starting a pattern with (*NOTEMPTY) or (*NOTEMPTY_ATSTART) has the same effectas passing the PCRE2_NOTEMPTY or PCRE2_NOTEMPTY_ATSTART option to whichevermatching function is subsequently called to match the pattern. These optionslock out the matching of empty strings, either entirely, or only at the startof the subject.</P><br><b>Disabling auto-possessification</b><br><P>If a pattern starts with (*NO_AUTO_POSSESS), it has the same effect as settingthe PCRE2_NO_AUTO_POSSESS option. This stops PCRE2 from making quantifierspossessive when what follows cannot match the repeated item. For example, bydefault a+b is treated as a++b. For more details, see the<a href="pcre2api.html"><b>pcre2api</b></a>documentation.</P><br><b>Disabling start-up optimizations</b><br><P>If a pattern starts with (*NO_START_OPT), it has the same effect as setting thePCRE2_NO_START_OPTIMIZE option. This disables several optimizations for quicklyreaching "no match" results. For more details, see the<a href="pcre2api.html"><b>pcre2api</b></a>documentation.</P><br><b>Disabling automatic anchoring</b><br><P>If a pattern starts with (*NO_DOTSTAR_ANCHOR), it has the same effect assetting the PCRE2_NO_DOTSTAR_ANCHOR option. This disables optimizations thatapply to patterns whose top-level branches all start with .* (match any numberof arbitrary characters). For more details, see the<a href="pcre2api.html"><b>pcre2api</b></a>documentation.</P><br><b>Disabling JIT compilation</b><br><P>If a pattern that starts with (*NO_JIT) is successfully compiled, an attempt bythe application to apply the JIT optimization by calling<b>pcre2_jit_compile()</b> is ignored.</P><br><b>Setting match resource limits</b><br><P>The <b>pcre2_match()</b> function contains a counter that is incremented everytime it goes round its main loop. The caller of <b>pcre2_match()</b> can set alimit on this counter, which therefore limits the amount of computing resourceused for a match. The maximum depth of nested backtracking can also be limited;this indirectly restricts the amount of heap memory that is used, but there isalso an explicit memory limit that can be set.</P><P>These facilities are provided to catch runaway matches that are provoked bypatterns with huge matching trees. A common example is a pattern with nestedunlimited repeats applied to a long string that does not match. When one ofthese limits is reached, <b>pcre2_match()</b> gives an error return. The limitscan also be set by items at the start of the pattern of the form<pre> (*LIMIT_HEAP=d) (*LIMIT_MATCH=d) (*LIMIT_DEPTH=d)</pre>where d is any number of decimal digits. However, the value of the setting mustbe less than the value set (or defaulted) by the caller of <b>pcre2_match()</b>for it to have any effect. In other words, the pattern writer can lower thelimits set by the programmer, but not raise them. If there is more than onesetting of one of these limits, the lower value is used. The heap limit isspecified in kibibytes (units of 1024 bytes).</P><P>Prior to release 10.30, LIMIT_DEPTH was called LIMIT_RECURSION. This name isstill recognized for backwards compatibility.</P><P>The heap limit applies only when the <b>pcre2_match()</b> or<b>pcre2_dfa_match()</b> interpreters are used for matching. It does not applyto JIT. The match limit is used (but in a different way) when JIT is beingused, or when <b>pcre2_dfa_match()</b> is called, to limit computing resourceusage by those matching functions. The depth limit is ignored by JIT but isrelevant for DFA matching, which uses function recursion for recursions withinthe pattern and for lookaround assertions and atomic groups. In this case, thedepth limit controls the depth of such recursion.<a name="newlines"></a></P><br><b>Newline conventions</b><br><P>PCRE2 supports six different conventions for indicating line breaks instrings: a single CR (carriage return) character, a single LF (linefeed)character, the two-character sequence CRLF, any of the three preceding, anyUnicode newline sequence, or the NUL character (binary zero). The<a href="pcre2api.html"><b>pcre2api</b></a>page has<a href="pcre2api.html#newlines">further discussion</a>about newlines, and shows how to set the newline convention when calling<b>pcre2_compile()</b>.</P><P>It is also possible to specify a newline convention by starting a patternstring with one of the following sequences:<pre> (*CR) carriage return (*LF) linefeed (*CRLF) carriage return, followed by linefeed (*ANYCRLF) any of the three above (*ANY) all Unicode newline sequences (*NUL) the NUL character (binary zero)</pre>These override the default and the options given to the compiling function. Forexample, on a Unix system where LF is the default newline sequence, the pattern<pre> (*CR)a.b</pre>changes the convention to CR. That pattern matches "a\nb" because LF is nolonger a newline. If more than one of these settings is present, the last oneis used.</P><P>The newline convention affects where the circumflex and dollar assertions aretrue. It also affects the interpretation of the dot metacharacter whenPCRE2_DOTALL is not set, and the behaviour of \N when not followed by anopening brace. However, it does not affect what the \R escape sequencematches. By default, this is any Unicode newline sequence, for Perlcompatibility. However, this can be changed; see the next section and thedescription of \R in the section entitled<a href="#newlineseq">"Newline sequences"</a>below. A change of \R setting can be combined with a change of newlineconvention.</P><br><b>Specifying what \R matches</b><br><P>It is possible to restrict \R to match only CR, LF, or CRLF (instead of thecomplete set of Unicode line endings) by setting the option PCRE2_BSR_ANYCRLFat compile time. This effect can also be achieved by starting a pattern with(*BSR_ANYCRLF). For completeness, (*BSR_UNICODE) is also recognized,corresponding to PCRE2_BSR_UNICODE.</P><br><a name="SEC3" href="#TOC1">EBCDIC CHARACTER CODES</a><br><P>PCRE2 can be compiled to run in an environment that uses EBCDIC as itscharacter code instead of ASCII or Unicode (typically a mainframe system). Inthe sections below, character code values are ASCII or Unicode; in an EBCDICenvironment these characters may have different code values, and there are nocode points greater than 255.</P><br><a name="SEC4" href="#TOC1">CHARACTERS AND METACHARACTERS</a><br><P>A regular expression is a pattern that is matched against a subject string fromleft to right. Most characters stand for themselves in a pattern, and match thecorresponding characters in the subject. As a trivial example, the pattern<pre> The quick brown fox</pre>matches a portion of a subject string that is identical to itself. Whencaseless matching is specified (the PCRE2_CASELESS option or (?i) within thepattern), letters are matched independently of case. Note that there are twoASCII characters, K and S, that, in addition to their lower case ASCIIequivalents, are case-equivalent with Unicode U+212A (Kelvin sign) and U+017F(long S) respectively when either PCRE2_UTF or PCRE2_UCP is set, unless thePCRE2_EXTRA_CASELESS_RESTRICT option is in force (either passed to<b>pcre2_compile()</b> or set by (?r) within the pattern).</P><P>The power of regular expressions comes from the ability to include wild cards,character classes, alternatives, and repetitions in the pattern. These areencoded in the pattern by the use of <i>metacharacters</i>, which do not standfor themselves but instead are interpreted in some special way.</P><P>There are two different sets of metacharacters: those that are recognizedanywhere in the pattern except within square brackets, and those that arerecognized within square brackets. Outside square brackets, the metacharactersare as follows:<pre> \ general escape character with several uses ^ assert start of string (or line, in multiline mode) $ assert end of string (or line, in multiline mode) . match any character except newline (by default) [ start character class definition | start of alternative branch ( start group or control verb ) end group or control verb * 0 or more quantifier + 1 or more quantifier; also "possessive quantifier" ? 0 or 1 quantifier; also quantifier minimizer { potential start of min/max quantifier</pre>Brace characters { and } are also used to enclose data for constructions suchas \g{2} or \k{name}. In almost all uses of braces, space and/or horizontaltab characters that follow { or precede } are allowed and are ignored. In thecase of quantifiers, they may also appear before or after the comma. Theexception to this is \u{...} which is an ECMAScript compatibility featurethat is recognized only when the PCRE2_EXTRA_ALT_BSUX option is set. ECMAScriptdoes not ignore such white space; it causes the item to be interpreted asliteral.</P><P>Part of a pattern that is in square brackets is called a "character class". Ina character class the only metacharacters are:<pre> \ general escape character ^ negate the class, but only if the first character - indicates character range [ POSIX character class (if followed by POSIX syntax) ] terminates the character class</pre>If a pattern is compiled with the PCRE2_EXTENDED option, most white space inthe pattern, other than in a character class, within a \Q...\E sequence, orbetween a # outside a character class and the next newline, inclusive, areignored. An escaping backslash can be used to include a white space or a #character as part of the pattern. If the PCRE2_EXTENDED_MORE option is set, thesame applies, but in addition unescaped space and horizontal tab characters areignored inside a character class. Note: only these two characters are ignored,not the full set of pattern white space characters that are ignored outside acharacter class. Option settings can be changed within a pattern; see thesection entitled<a href="#internaloptions">"Internal Option Setting"</a>below.</P><P>The following sections describe the use of each of the metacharacters.</P><br><a name="SEC5" href="#TOC1">BACKSLASH</a><br><P>The backslash character has several uses. Firstly, if it is followed by acharacter that is not a digit or a letter, it takes away any special meaningthat character may have. This use of backslash as an escape character appliesboth inside and outside character classes.</P><P>For example, if you want to match a * character, you must write \* in thepattern. This escaping action applies whether or not the following characterwould otherwise be interpreted as a metacharacter, so it is always safe toprecede a non-alphanumeric with backslash to specify that it stands for itself.In particular, if you want to match a backslash, you write \\.</P><P>Only ASCII digits and letters have any special meaning after a backslash. Allother characters (in particular, those whose code points are greater than 127)are treated as literals.</P><P>If you want to treat all characters in a sequence as literals, you can do so byputting them between \Q and \E. Note that this includes white space even whenthe PCRE2_EXTENDED option is set so that most other white space is ignored. Thebehaviour is different from Perl in that $ and @ are handled as literals in\Q...\E sequences in PCRE2, whereas in Perl, $ and @ cause variableinterpolation. Also, Perl does "double-quotish backslash interpolation" on anybackslashes between \Q and \E which, its documentation says, "may lead toconfusing results". PCRE2 treats a backslash between \Q and \E just like anyother character. Note the following examples:<pre> Pattern PCRE2 matches Perl matches
\Qabc$xyz\E abc$xyz abc followed by the contents of $xyz \Qabc\$xyz\E abc\$xyz abc\$xyz \Qabc\E\$\Qxyz\E abc$xyz abc$xyz \QA\B\E A\B A\B \Q\\E \ \\E</pre>The \Q...\E sequence is recognized both inside and outside character classes.An isolated \E that is not preceded by \Q is ignored. If \Q is not followedby \E later in the pattern, the literal interpretation continues to the end ofthe pattern (that is, \E is assumed at the end). If the isolated \Q is insidea character class, this causes an error, because the character class is thennot terminated by a closing square bracket.<a name="digitsafterbackslash"></a></P><br><b>Non-printing characters</b><br><P>A second use of backslash provides a way of encoding non-printing charactersin patterns in a visible manner. There is no restriction on the appearance ofnon-printing characters in a pattern, but when a pattern is being prepared bytext editing, it is often easier to use one of the following escape sequencesinstead of the binary character it represents. In an ASCII or Unicodeenvironment, these escapes are as follows:<pre> \a alarm, that is, the BEL character (hex 07) \cx "control-x", where x is a non-control ASCII character \e escape (hex 1B) \f form feed (hex 0C) \n linefeed (hex 0A) \r carriage return (hex 0D) (but see below) \t tab (hex 09) \0dd character with octal code 0dd \ddd character with octal code ddd, or backreference \o{ddd..} character with octal code ddd.. \xhh character with hex code hh \x{hhh..} character with hex code hhh.. \N{U+hhh..} character with Unicode hex code point hhh..</pre>By default, after \x that is not followed by {, from zero to two hexadecimaldigits are read (letters can be in upper or lower case). Any number ofhexadecimal digits may appear between \x{ and }. If a character other than ahexadecimal digit appears between \x{ and }, or if there is no terminating },an error occurs.</P><P>Characters whose code points are less than 256 can be defined by either of thetwo syntaxes for \x or by an octal sequence. There is no difference in the waythey are handled. For example, \xdc is exactly the same as \x{dc} or \334.However, using the braced versions does make such sequences easier to read.</P><P>Support is available for some ECMAScript (aka JavaScript) escape sequences viatwo compile-time options. If PCRE2_ALT_BSUX is set, the sequence \x followedby { is not recognized. Only if \x is followed by two hexadecimal digits is itrecognized as a character escape. Otherwise it is interpreted as a literal "x"character. In this mode, support for code points greater than 256 is providedby \u, which must be followed by four hexadecimal digits; otherwise it isinterpreted as a literal "u" character.</P><P>PCRE2_EXTRA_ALT_BSUX has the same effect as PCRE2_ALT_BSUX and, in addition,\u{hhh..} is recognized as the character specified by hexadecimal code point.There may be any number of hexadecimal digits, but unlike other places thatalso use curly brackets, spaces are not allowed and would result in the stringbeing interpreted as a literal. This syntax is from ECMAScript 6.</P><P>The \N{U+hhh..} escape sequence is recognized only when PCRE2 is operating inUTF mode. Perl also uses \N{name} to specify characters by Unicode name; PCRE2does not support this. Note that when \N is not followed by an opening brace(curly bracket) it has an entirely different meaning, matching any characterthat is not a newline.</P><P>There are some legacy applications where the escape sequence \r is expected tomatch a newline. If the PCRE2_EXTRA_ESCAPED_CR_IS_LF option is set, \r in apattern is converted to \n so that it matches a LF (linefeed) instead of a CR(carriage return) character.</P><P>An error occurs if \c is not followed by a character whose ASCII code pointis in the range 32 to 126. The precise effect of \cx is as follows: if x is alower case letter, it is converted to upper case. Then bit 6 of the character(hex 40) is inverted. Thus \cA to \cZ become hex 01 to hex 1A (A is 41, Z is5A), but \c{ becomes hex 3B ({ is 7B), and \c; becomes hex 7B (; is 3B). Ifthe code unit following \c has a code point less than 32 or greater than 126,a compile-time error occurs.</P><P>When PCRE2 is compiled in EBCDIC mode, \N{U+hhh..} is not supported. \a, \e,\f, \n, \r, and \t generate the appropriate EBCDIC code values. The \cescape is processed as specified for Perl in the <b>perlebcdic</b> document. Theonly characters that are allowed after \c are A-Z, a-z, or one of @, [, \, ],^, _, or ?. Any other character provokes a compile-time error. The sequence\c@ encodes character code 0; after \c the letters (in either case) encodecharacters 1-26 (hex 01 to hex 1A); [, \, ], ^, and _ encode characters 27-31(hex 1B to hex 1F), and \c? becomes either 255 (hex FF) or 95 (hex 5F).</P><P>Thus, apart from \c?, these escapes generate the same character code values asthey do in an ASCII environment, though the meanings of the values mostlydiffer. For example, \cG always generates code value 7, which is BEL in ASCIIbut DEL in EBCDIC.</P><P>The sequence \c? generates DEL (127, hex 7F) in an ASCII environment, butbecause 127 is not a control character in EBCDIC, Perl makes it generate theAPC character. Unfortunately, there are several variants of EBCDIC. In most ofthem the APC character has the value 255 (hex FF), but in the one Perl callsPOSIX-BC its value is 95 (hex 5F). If certain other characters have POSIX-BCvalues, PCRE2 makes \c? generate 95; otherwise it generates 255.</P><P>After \0 up to two further octal digits are read. If there are fewer than twodigits, just those that are present are used. Thus the sequence \0\x\015specifies two binary zeros followed by a CR character (code value 13). Makesure you supply two digits after the initial zero if the pattern character thatfollows is itself an octal digit.</P><P>The escape \o must be followed by a sequence of octal digits, enclosed inbraces. An error occurs if this is not the case. This escape is a recentaddition to Perl; it provides way of specifying character code points as octalnumbers greater than 0777, and it also allows octal numbers and backreferencesto be unambiguously specified.</P><P>For greater clarity and unambiguity, it is best to avoid following \ by adigit greater than zero. Instead, use \o{...} or \x{...} to specify numericalcharacter code points, and \g{...} to specify backreferences. The followingparagraphs describe the old, ambiguous syntax.</P><P>The handling of a backslash followed by a digit other than 0 is complicated,and Perl has changed over time, causing PCRE2 also to change.</P><P>Outside a character class, PCRE2 reads the digit and any following digits as adecimal number. If the number is less than 10, begins with the digit 8 or 9, orif there are at least that many previous capture groups in the expression, theentire sequence is taken as a <i>backreference</i>. A description of how thisworks is given<a href="#backreferences">later,</a>following the discussion of<a href="#group">parenthesized groups.</a>Otherwise, up to three octal digits are read to form a character code.</P><P>Inside a character class, PCRE2 handles \8 and \9 as the literal characters"8" and "9", and otherwise reads up to three octal digits following thebackslash, using them to generate a data character. Any subsequent digits standfor themselves. For example, outside a character class:<pre> \040 is another way of writing an ASCII space \40 is the same, provided there are fewer than 40 previous capture groups \7 is always a backreference \11 might be a backreference, or another way of writing a tab \011 is always a tab \0113 is a tab followed by the character "3" \113 might be a backreference, otherwise the character with octal code 113 \377 might be a backreference, otherwise the value 255 (decimal) \81 is always a backreference</pre>Note that octal values of 100 or greater that are specified using this syntaxmust not be introduced by a leading zero, because no more than three octaldigits are ever read.</P><br><b>Constraints on character values</b><br><P>Characters that are specified using octal or hexadecimal numbers arelimited to certain values, as follows:<pre> 8-bit non-UTF mode no greater than 0xff 16-bit non-UTF mode no greater than 0xffff 32-bit non-UTF mode no greater than 0xffffffff All UTF modes no greater than 0x10ffff and a valid code point</pre>Invalid Unicode code points are all those in the range 0xd800 to 0xdfff (theso-called "surrogate" code points). The check for these can be disabled by thecaller of <b>pcre2_compile()</b> by setting the optionPCRE2_EXTRA_ALLOW_SURROGATE_ESCAPES. However, this is possible only in UTF-8and UTF-32 modes, because these values are not representable in UTF-16.</P><br><b>Escape sequences in character classes</b><br><P>All the sequences that define a single character value can be used both insideand outside character classes. In addition, inside a character class, \b isinterpreted as the backspace character (hex 08).</P><P>When not followed by an opening brace, \N is not allowed in a character class.\B, \R, and \X are not special inside a character class. Like otherunrecognized alphabetic escape sequences, they cause an error. Outside acharacter class, these sequences have different meanings.</P><br><b>Unsupported escape sequences</b><br><P>In Perl, the sequences \F, \l, \L, \u, and \U are recognized by its stringhandler and used to modify the case of following characters. By default, PCRE2does not support these escape sequences in patterns. However, if either of thePCRE2_ALT_BSUX or PCRE2_EXTRA_ALT_BSUX options is set, \U matches a "U"character, and \u can be used to define a character by code point, asdescribed above.</P><br><b>Absolute and relative backreferences</b><br><P>The sequence \g followed by a signed or unsigned number, optionally enclosedin braces, is an absolute or relative backreference. A named backreferencecan be coded as \g{name}. Backreferences are discussed<a href="#backreferences">later,</a>following the discussion of<a href="#group">parenthesized groups.</a></P><br><b>Absolute and relative subroutine calls</b><br><P>For compatibility with Oniguruma, the non-Perl syntax \g followed by a name ora number enclosed either in angle brackets or single quotes, is an alternativesyntax for referencing a capture group as a subroutine. Details are discussed<a href="#onigurumasubroutines">later.</a>Note that \g{...} (Perl syntax) and \g<...> (Oniguruma syntax) are <i>not</i>synonymous. The former is a backreference; the latter is a<a href="#groupsassubroutines">subroutine</a>call.<a name="genericchartypes"></a></P><br><b>Generic character types</b><br><P>Another use of backslash is for specifying generic character types:<pre> \d any decimal digit \D any character that is not a decimal digit \h any horizontal white space character \H any character that is not a horizontal white space character \N any character that is not a newline \s any white space character \S any character that is not a white space character \v any vertical white space character \V any character that is not a vertical white space character \w any "word" character \W any "non-word" character</pre>The \N escape sequence has the same meaning as<a href="#fullstopdot">the "." metacharacter</a>when PCRE2_DOTALL is not set, but setting PCRE2_DOTALL does not change themeaning of \N. Note that when \N is followed by an opening brace it has adifferent meaning. See the section entitled<a href="#digitsafterbackslash">"Non-printing characters"</a>above for details. Perl also uses \N{name} to specify characters by Unicodename; PCRE2 does not support this.</P><P>Each pair of lower and upper case escape sequences partitions the complete setof characters into two disjoint sets. Any given character matches one, and onlyone, of each pair. The sequences can appear both inside and outside characterclasses. They each match one character of the appropriate type. If the currentmatching point is at the end of the subject string, all of them fail, becausethere is no character to match.</P><P>The default \s characters are HT (9), LF (10), VT (11), FF (12), CR (13), andspace (32), which are defined as white space in the "C" locale. This list mayvary if locale-specific matching is taking place. For example, in some localesthe "non-breaking space" character (\xA0) is recognized as white space, and inothers the VT character is not.</P><P>A "word" character is an underscore or any character that is a letter or digit.By default, the definition of letters and digits is controlled by PCRE2'slow-valued character tables, and may vary if locale-specific matching is takingplace (see<a href="pcre2api.html#localesupport">"Locale support"</a>in the<a href="pcre2api.html"><b>pcre2api</b></a>page). For example, in a French locale such as "fr_FR" in Unix-like systems,or "french" in Windows, some character codes greater than 127 are used foraccented letters, and these are then matched by \w. The use of locales withUnicode is discouraged.</P><P>By default, characters whose code points are greater than 127 never match \d,\s, or \w, and always match \D, \S, and \W, although this may be differentfor characters in the range 128-255 when locale-specific matching is happening.These escape sequences retain their original meanings from before Unicodesupport was available, mainly for efficiency reasons. If the PCRE2_UCP optionis set, the behaviour is changed so that Unicode properties are used todetermine character types, as follows:<pre> \d any character that matches \p{Nd} (decimal digit) \s any character that matches \p{Z} or \h or \v \w any character that matches \p{L}, \p{N}, \p{Mn}, or \p{Pc}</pre>The addition of \p{Mn} (non-spacing mark) and the replacement of an explicittest for underscore with a test for \p{Pc} (connector punctuation) happened inPCRE2 release 10.43. This brings PCRE2 into line with Perl.</P><P>The upper case escapes match the inverse sets of characters. Note that \dmatches only decimal digits, whereas \w matches any Unicode digit, as well asother character categories. Note also that PCRE2_UCP affects \b, and\B because they are defined in terms of \w and \W. Matching these sequencesis noticeably slower when PCRE2_UCP is set.</P><P>The effect of PCRE2_UCP on any one of these escape sequences can be negated bythe options PCRE2_EXTRA_ASCII_BSD, PCRE2_EXTRA_ASCII_BSS, andPCRE2_EXTRA_ASCII_BSW, respectively. These options can be set and reset withina pattern by means of an internal option setting<a href="#internaloptions">(see below).</a></P><P>The sequences \h, \H, \v, and \V, in contrast to the other sequences, whichmatch only ASCII characters by default, always match a specific list of codepoints, whether or not PCRE2_UCP is set. The horizontal space characters are:<pre> U+0009 Horizontal tab (HT) U+0020 Space U+00A0 Non-break space U+1680 Ogham space mark U+180E Mongolian vowel separator U+2000 En quad U+2001 Em quad U+2002 En space U+2003 Em space U+2004 Three-per-em space U+2005 Four-per-em space U+2006 Six-per-em space U+2007 Figure space U+2008 Punctuation space U+2009 Thin space U+200A Hair space U+202F Narrow no-break space U+205F Medium mathematical space U+3000 Ideographic space</pre>The vertical space characters are:<pre> U+000A Linefeed (LF) U+000B Vertical tab (VT) U+000C Form feed (FF) U+000D Carriage return (CR) U+0085 Next line (NEL) U+2028 Line separator U+2029 Paragraph separator</pre>In 8-bit, non-UTF-8 mode, only the characters with code points less than 256are relevant.<a name="newlineseq"></a></P><br><b>Newline sequences</b><br><P>Outside a character class, by default, the escape sequence \R matches anyUnicode newline sequence. In 8-bit non-UTF-8 mode \R is equivalent to thefollowing:<pre> (?>\r\n|\n|\x0b|\f|\r|\x85)</pre>This is an example of an "atomic group", details of which are given<a href="#atomicgroup">below.</a>This particular group matches either the two-character sequence CR followed byLF, or one of the single characters LF (linefeed, U+000A), VT (vertical tab,U+000B), FF (form feed, U+000C), CR (carriage return, U+000D), or NEL (nextline, U+0085). Because this is an atomic group, the two-character sequence istreated as a single unit that cannot be split.</P><P>In other modes, two additional characters whose code points are greater than 255are added: LS (line separator, U+2028) and PS (paragraph separator, U+2029).Unicode support is not needed for these characters to be recognized.</P><P>It is possible to restrict \R to match only CR, LF, or CRLF (instead of thecomplete set of Unicode line endings) by setting the option PCRE2_BSR_ANYCRLFat compile time. (BSR is an abbreviation for "backslash R".) This can be madethe default when PCRE2 is built; if this is the case, the other behaviour canbe requested via the PCRE2_BSR_UNICODE option. It is also possible to specifythese settings by starting a pattern string with one of the followingsequences:<pre> (*BSR_ANYCRLF) CR, LF, or CRLF only (*BSR_UNICODE) any Unicode newline sequence</pre>These override the default and the options given to the compiling function.Note that these special settings, which are not Perl-compatible, are recognizedonly at the very start of a pattern, and that they must be in upper case. Ifmore than one of them is present, the last one is used. They can be combinedwith a change of newline convention; for example, a pattern can start with:<pre> (*ANY)(*BSR_ANYCRLF)</pre>They can also be combined with the (*UTF) or (*UCP) special sequences. Inside acharacter class, \R is treated as an unrecognized escape sequence, and causesan error.<a name="uniextseq"></a></P><br><b>Unicode character properties</b><br><P>When PCRE2 is built with Unicode support (the default), three additional escapesequences that match characters with specific properties are available. Theycan be used in any mode, though in 8-bit and 16-bit non-UTF modes thesesequences are of course limited to testing characters whose code points areless than U+0100 and U+10000, respectively. In 32-bit non-UTF mode, code pointsgreater than 0x10ffff (the Unicode limit) may be encountered. These are alltreated as being in the Unknown script and with an unassigned type.</P><P>Matching characters by Unicode property is not fast, because PCRE2 has to do amultistage table lookup in order to find a character's property. That is whythe traditional escape sequences such as \d and \w do not use Unicodeproperties in PCRE2 by default, though you can make them do so by setting thePCRE2_UCP option or by starting the pattern with (*UCP).</P><P>The extra escape sequences that provide property support are:<pre> \p{<i>xx</i>} a character with the <i>xx</i> property \P{<i>xx</i>} a character without the <i>xx</i> property \X a Unicode extended grapheme cluster</pre>The property names represented by <i>xx</i> above are not case-sensitive, and inaccordance with Unicode's "loose matching" rules, spaces, hyphens, andunderscores are ignored. There is support for Unicode script names, Unicodegeneral category properties, "Any", which matches any character (includingnewline), Bidi_Class, a number of binary (yes/no) properties, and some specialPCRE2 properties (described<a href="#extraprops">below).</a>Certain other Perl properties such as "InMusicalSymbols" are not supported byPCRE2. Note that \P{Any} does not match any characters, so always causes amatch failure.</P><br><b>Script properties for \p and \P</b><br><P>There are three different syntax forms for matching a script. Each Unicodecharacter has a basic script and, optionally, a list of other scripts ("ScriptExtensions") with which it is commonly used. Using the Adlam script as anexample, \p{sc:Adlam} matches characters whose basic script is Adlam, whereas\p{scx:Adlam} matches, in addition, characters that have Adlam in theirextensions list. The full names "script" and "script extensions" for theproperty types are recognized, and a equals sign is an alternative to thecolon. If a script name is given without a property type, for example,\p{Adlam}, it is treated as \p{scx:Adlam}. Perl changed to thisinterpretation at release 5.26 and PCRE2 changed at release 10.40.</P><P>Unassigned characters (and in non-UTF 32-bit mode, characters with code pointsgreater than 0x10FFFF) are assigned the "Unknown" script. Others that are notpart of an identified script are lumped together as "Common". The current listof recognized script names and their 4-character abbreviations can be obtainedby running this command:<pre> pcre2test -LS
</PRE></P><br><b>The general category property for \p and \P</b><br><P>Each character has exactly one Unicode general category property, specified bya two-letter abbreviation. For compatibility with Perl, negation can bespecified by including a circumflex between the opening brace and the propertyname. For example, \p{^Lu} is the same as \P{Lu}.</P><P>If only one letter is specified with \p or \P, it includes all the generalcategory properties that start with that letter. In this case, in the absenceof negation, the curly brackets in the escape sequence are optional; these twoexamples have the same effect:<pre> \p{L} \pL</pre>The following general category property codes are supported:<pre> C Other Cc Control Cf Format Cn Unassigned Co Private use Cs Surrogate
L Letter Ll Lower case letter Lm Modifier letter Lo Other letter Lt Title case letter Lu Upper case letter
M Mark Mc Spacing mark Me Enclosing mark Mn Non-spacing mark
N Number Nd Decimal number Nl Letter number No Other number
P Punctuation Pc Connector punctuation Pd Dash punctuation Pe Close punctuation Pf Final punctuation Pi Initial punctuation Po Other punctuation Ps Open punctuation
S Symbol Sc Currency symbol Sk Modifier symbol Sm Mathematical symbol So Other symbol
Z Separator Zl Line separator Zp Paragraph separator Zs Space separator</pre>The special property LC, which has the synonym L&, is also supported: itmatches a character that has the Lu, Ll, or Lt property, in other words, aletter that is not classified as a modifier or "other".</P><P>The Cs (Surrogate) property applies only to characters whose code points are inthe range U+D800 to U+DFFF. These characters are no different to any othercharacter when PCRE2 is not in UTF mode (using the 16-bit or 32-bit library).However, they are not valid in Unicode strings and so cannot be tested by PCRE2in UTF mode, unless UTF validity checking has been turned off (see thediscussion of PCRE2_NO_UTF_CHECK in the<a href="pcre2api.html"><b>pcre2api</b></a>page).</P><P>The long synonyms for property names that Perl supports (such as \p{Letter})are not supported by PCRE2, nor is it permitted to prefix any of theseproperties with "Is".</P><P>No character that is in the Unicode table has the Cn (unassigned) property.Instead, this property is assumed for any code point that is not in theUnicode table.</P><P>Specifying caseless matching does not affect these escape sequences. Forexample, \p{Lu} always matches only upper case letters. This is different fromthe behaviour of current versions of Perl.</P><br><b>Binary (yes/no) properties for \p and \P</b><br><P>Unicode defines a number of binary properties, that is, properties whose onlyvalues are true or false. You can obtain a list of those that are recognized by\p and \P, along with their abbreviations, by running this command:<pre> pcre2test -LP
</PRE></P><br><b>The Bidi_Class property for \p and \P</b><br><P><pre> \p{Bidi_Class:<class>} matches a character with the given class \p{BC:<class>} matches a character with the given class</pre>The recognized classes are:<pre> AL Arabic letter AN Arabic number B paragraph separator BN boundary neutral CS common separator EN European number ES European separator ET European terminator FSI first strong isolate L left-to-right LRE left-to-right embedding LRI left-to-right isolate LRO left-to-right override NSM non-spacing mark ON other neutral PDF pop directional format PDI pop directional isolate R right-to-left RLE right-to-left embedding RLI right-to-left isolate RLO right-to-left override S segment separator WS which space</pre>An equals sign may be used instead of a colon. The class names arecase-insensitive; only the short names listed above are recognized.</P><br><b>Extended grapheme clusters</b><br><P>The \X escape matches any number of Unicode characters that form an "extendedgrapheme cluster", and treats the sequence as an atomic group<a href="#atomicgroup">(see below).</a>Unicode supports various kinds of composite character by giving each charactera grapheme breaking property, and having rules that use these properties todefine the boundaries of extended grapheme clusters. The rules are defined inUnicode Standard Annex 29, "Unicode Text Segmentation". Unicode 11.0.0abandoned the use of some previous properties that had been used for emojis.Instead it introduced various emoji-specific properties. PCRE2 uses only theExtended Pictographic property.</P><P>\X always matches at least one character. Then it decides whether to addadditional characters according to the following rules for ending a cluster:</P><P>1. End at the end of the subject string.</P><P>2. Do not end between CR and LF; otherwise end after any control character.</P><P>3. Do not break Hangul (a Korean script) syllable sequences. Hangul charactersare of five types: L, V, T, LV, and LVT. An L character may be followed by anL, V, LV, or LVT character; an LV or V character may be followed by a V or Tcharacter; an LVT or T character may be followed only by a T character.</P><P>4. Do not end before extending characters or spacing marks or the zero-widthjoiner (ZWJ) character. Characters with the "mark" property always have the"extend" grapheme breaking property.</P><P>5. Do not end after prepend characters.</P><P>6. Do not end within emoji modifier sequences or emoji ZWJ (zero-widthjoiner) sequences. An emoji ZWJ sequence consists of a character with theExtended_Pictographic property, optionally followed by one or more characterswith the Extend property, followed by the ZWJ character, followed by anotherExtended_Pictographic character.</P><P>7. Do not break within emoji flag sequences. That is, do not break betweenregional indicator (RI) characters if there are an odd number of RI charactersbefore the break point.</P><P>8. Otherwise, end the cluster.<a name="extraprops"></a></P><br><b>PCRE2's additional properties</b><br><P>As well as the standard Unicode properties described above, PCRE2 supports fourmore that make it possible to convert traditional escape sequences such as \wand \s to use Unicode properties. PCRE2 uses these non-standard, non-Perlproperties internally when PCRE2_UCP is set. However, they may also be usedexplicitly. These properties are:<pre> Xan Any alphanumeric character Xps Any POSIX space character Xsp Any Perl space character Xwd Any Perl "word" character</pre>Xan matches characters that have either the L (letter) or the N (number)property. Xps matches the characters tab, linefeed, vertical tab, form feed, orcarriage return, and any other character that has the Z (separator) property.Xsp is the same as Xps; in PCRE1 it used to exclude vertical tab, for Perlcompatibility, but Perl changed. Xwd matches the same characters as Xan, plusthose that match Mn (non-spacing mark) or Pc (connector punctuation, whichincludes underscore).</P><P>There is another non-standard property, Xuc, which matches any character thatcan be represented by a Universal Character Name in C++ and other programminglanguages. These are the characters $, @, ` (grave accent), and all characterswith Unicode code points greater than or equal to U+00A0, except for thesurrogates U+D800 to U+DFFF. Note that most base (ASCII) characters areexcluded. (Universal Character Names are of the form \uHHHH or \UHHHHHHHHwhere H is a hexadecimal digit. Note that the Xuc property does not match thesesequences but the characters that they represent.)<a name="resetmatchstart"></a></P><br><b>Resetting the match start</b><br><P>In normal use, the escape sequence \K causes any previously matched charactersnot to be included in the final matched sequence that is returned. For example,the pattern:<pre> foo\Kbar</pre>matches "foobar", but reports that it has matched "bar". \K does not interactwith anchoring in any way. The pattern:<pre> ^foo\Kbar</pre>matches only when the subject begins with "foobar" (in single line mode),though it again reports the matched string as "bar". This feature is similar toa lookbehind assertion<a href="#lookbehind">(described below),</a>but the part of the pattern that precedes \K is not constrained to match alimited number of characters, as is required for a lookbehind assertion. Theuse of \K does not interfere with the setting of<a href="#group">captured substrings.</a>For example, when the pattern<pre> (foo)\Kbar</pre>matches "foobar", the first substring is still set to "foo".</P><P>From version 5.32.0 Perl forbids the use of \K in lookaround assertions. Fromrelease 10.38 PCRE2 also forbids this by default. However, thePCRE2_EXTRA_ALLOW_LOOKAROUND_BSK option can be used when calling<b>pcre2_compile()</b> to re-enable the previous behaviour. When this option isset, \K is acted upon when it occurs inside positive assertions, but isignored in negative assertions. Note that when a pattern such as (?=ab\K)matches, the reported start of the match can be greater than the end of thematch. Using \K in a lookbehind assertion at the start of a pattern can alsolead to odd effects. For example, consider this pattern:<pre> (?<=\Kfoo)bar</pre>If the subject is "foobar", a call to <b>pcre2_match()</b> with a startingoffset of 3 succeeds and reports the matching string as "foobar", that is, thestart of the reported match is earlier than where the match started.<a name="smallassertions"></a></P><br><b>Simple assertions</b><br><P>The final use of backslash is for certain simple assertions. An assertionspecifies a condition that has to be met at a particular point in a match,without consuming any characters from the subject string. The use ofgroups for more complicated assertions is described<a href="#bigassertions">below.</a>The backslashed assertions are:<pre> \b matches at a word boundary \B matches when not at a word boundary \A matches at the start of the subject \Z matches at the end of the subject also matches before a newline at the end of the subject \z matches only at the end of the subject \G matches at the first matching position in the subject</pre>Inside a character class, \b has a different meaning; it matches the backspacecharacter. If any other of these assertions appears in a character class, an"invalid escape sequence" error is generated.</P><P>A word boundary is a position in the subject string where the current characterand the previous character do not both match \w or \W (i.e. one matches\w and the other matches \W), or the start or end of the string if thefirst or last character matches \w, respectively. When PCRE2 is built withUnicode support, the meanings of \w and \W can be changed by setting thePCRE2_UCP option. When this is done, it also affects \b and \B. Neither PCRE2nor Perl has a separate "start of word" or "end of word" metasequence. However,whatever follows \b normally determines which it is. For example, the fragment\ba matches "a" at the start of a word.</P><P>The \A, \Z, and \z assertions differ from the traditional circumflex anddollar (described in the next section) in that they only ever match at the verystart and end of the subject string, whatever options are set. Thus, they areindependent of multiline mode. These three assertions are not affected by thePCRE2_NOTBOL or PCRE2_NOTEOL options, which affect only the behaviour of thecircumflex and dollar metacharacters. However, if the <i>startoffset</i>argument of <b>pcre2_match()</b> is non-zero, indicating that matching is tostart at a point other than the beginning of the subject, \A can never match.The difference between \Z and \z is that \Z matches before a newline at theend of the string as well as at the very end, whereas \z matches only at theend.</P><P>The \G assertion is true only when the current matching position is at thestart point of the matching process, as specified by the <i>startoffset</i>argument of <b>pcre2_match()</b>. It differs from \A when the value of<i>startoffset</i> is non-zero. By calling <b>pcre2_match()</b> multiple timeswith appropriate arguments, you can mimic Perl's /g option, and it is in thiskind of implementation where \G can be useful.</P><P>Note, however, that PCRE2's implementation of \G, being true at the startingcharacter of the matching process, is subtly different from Perl's, whichdefines it as true at the end of the previous match. In Perl, these can bedifferent when the previously matched string was empty. Because PCRE2 does justone match at a time, it cannot reproduce this behaviour.</P><P>If all the alternatives of a pattern begin with \G, the expression is anchoredto the starting match position, and the "anchored" flag is set in the compiledregular expression.</P><br><a name="SEC6" href="#TOC1">CIRCUMFLEX AND DOLLAR</a><br><P>The circumflex and dollar metacharacters are zero-width assertions. That is,they test for a particular condition being true without consuming anycharacters from the subject string. These two metacharacters are concerned withmatching the starts and ends of lines. If the newline convention is set so thatonly the two-character sequence CRLF is recognized as a newline, isolated CRand LF characters are treated as ordinary data characters, and are notrecognized as newlines.</P><P>Outside a character class, in the default matching mode, the circumflexcharacter is an assertion that is true only if the current matching point is atthe start of the subject string. If the <i>startoffset</i> argument of<b>pcre2_match()</b> is non-zero, or if PCRE2_NOTBOL is set, circumflex cannever match if the PCRE2_MULTILINE option is unset. Inside a character class,circumflex has an entirely different meaning<a href="#characterclass">(see below).</a></P><P>Circumflex need not be the first character of the pattern if a number ofalternatives are involved, but it should be the first thing in each alternativein which it appears if the pattern is ever to match that branch. If allpossible alternatives start with a circumflex, that is, if the pattern isconstrained to match only at the start of the subject, it is said to be an"anchored" pattern. (There are also other constructs that can cause a patternto be anchored.)</P><P>The dollar character is an assertion that is true only if the current matchingpoint is at the end of the subject string, or immediately before a newline atthe end of the string (by default), unless PCRE2_NOTEOL is set. Note, however,that it does not actually match the newline. Dollar need not be the lastcharacter of the pattern if a number of alternatives are involved, but itshould be the last item in any branch in which it appears. Dollar has nospecial meaning in a character class.</P><P>The meaning of dollar can be changed so that it matches only at the very end ofthe string, by setting the PCRE2_DOLLAR_ENDONLY option at compile time. Thisdoes not affect the \Z assertion.</P><P>The meanings of the circumflex and dollar metacharacters are changed if thePCRE2_MULTILINE option is set. When this is the case, a dollar charactermatches before any newlines in the string, as well as at the very end, and acircumflex matches immediately after internal newlines as well as at the startof the subject string. It does not match after a newline that ends the string,for compatibility with Perl. However, this can be changed by setting thePCRE2_ALT_CIRCUMFLEX option.</P><P>For example, the pattern /^abc$/ matches the subject string "def\nabc" (where\n represents a newline) in multiline mode, but not otherwise. Consequently,patterns that are anchored in single line mode because all branches start with^ are not anchored in multiline mode, and a match for circumflex is possiblewhen the <i>startoffset</i> argument of <b>pcre2_match()</b> is non-zero. ThePCRE2_DOLLAR_ENDONLY option is ignored if PCRE2_MULTILINE is set.</P><P>When the newline convention (see<a href="#newlines">"Newline conventions"</a>below) recognizes the two-character sequence CRLF as a newline, this ispreferred, even if the single characters CR and LF are also recognized asnewlines. For example, if the newline convention is "any", a multiline modecircumflex matches before "xyz" in the string "abc\r\nxyz" rather than afterCR, even though CR on its own is a valid newline. (It also matches at the verystart of the string, of course.)</P><P>Note that the sequences \A, \Z, and \z can be used to match the start andend of the subject in both modes, and if all branches of a pattern start with\A it is always anchored, whether or not PCRE2_MULTILINE is set.<a name="fullstopdot"></a></P><br><a name="SEC7" href="#TOC1">FULL STOP (PERIOD, DOT) AND \N</a><br><P>Outside a character class, a dot in the pattern matches any one character inthe subject string except (by default) a character that signifies the end of aline. One or more characters may be specified as line terminators (see<a href="#newlines">"Newline conventions"</a>above).</P><P>Dot never matches a single line-ending character. When the two-charactersequence CRLF is the only line ending, dot does not match CR if it isimmediately followed by LF, but otherwise it matches all characters (includingisolated CRs and LFs). When ANYCRLF is selected for line endings, no occurrencesof CR of LF match dot. When all Unicode line endings are being recognized, dotdoes not match CR or LF or any of the other line ending characters.</P><P>The behaviour of dot with regard to newlines can be changed. If thePCRE2_DOTALL option is set, a dot matches any one character, without exception.If the two-character sequence CRLF is present in the subject string, it takestwo dots to match it.</P><P>The handling of dot is entirely independent of the handling of circumflex anddollar, the only relationship being that they both involve newlines. Dot has nospecial meaning in a character class.</P><P>The escape sequence \N when not followed by an opening brace behaves like adot, except that it is not affected by the PCRE2_DOTALL option. In other words,it matches any character except one that signifies the end of a line.</P><P>When \N is followed by an opening brace it has a different meaning. See thesection entitled<a href="digitsafterbackslash">"Non-printing characters"</a>above for details. Perl also uses \N{name} to specify characters by Unicodename; PCRE2 does not support this.</P><br><a name="SEC8" href="#TOC1">MATCHING A SINGLE CODE UNIT</a><br><P>Outside a character class, the escape sequence \C matches any one code unit,whether or not a UTF mode is set. In the 8-bit library, one code unit is onebyte; in the 16-bit library it is a 16-bit unit; in the 32-bit library it is a32-bit unit. Unlike a dot, \C always matches line-ending characters. Thefeature is provided in Perl in order to match individual bytes in UTF-8 mode,but it is unclear how it can usefully be used.</P><P>Because \C breaks up characters into individual code units, matching one unitwith \C in UTF-8 or UTF-16 mode means that the rest of the string may startwith a malformed UTF character. This has undefined results, because PCRE2assumes that it is matching character by character in a valid UTF string (bydefault it checks the subject string's validity at the start of processingunless the PCRE2_NO_UTF_CHECK or PCRE2_MATCH_INVALID_UTF option is used).</P><P>An application can lock out the use of \C by setting thePCRE2_NEVER_BACKSLASH_C option when compiling a pattern. It is also possible tobuild PCRE2 with the use of \C permanently disabled.</P><P>PCRE2 does not allow \C to appear in lookbehind assertions<a href="#lookbehind">(described below)</a>in UTF-8 or UTF-16 modes, because this would make it impossible to calculatethe length of the lookbehind. Neither the alternative matching function<b>pcre2_dfa_match()</b> nor the JIT optimizer support \C in these UTF modes.The former gives a match-time error; the latter fails to optimize and so thematch is always run using the interpreter.</P><P>In the 32-bit library, however, \C is always supported (when not explicitlylocked out) because it always matches a single code unit, whether or not UTF-32is specified.</P><P>In general, the \C escape sequence is best avoided. However, one way of usingit that avoids the problem of malformed UTF-8 or UTF-16 characters is to use alookahead to check the length of the next character, as in this pattern, whichcould be used with a UTF-8 string (ignore white space and line breaks):<pre> (?| (?=[\x00-\x7f])(\C) | (?=[\x80-\x{7ff}])(\C)(\C) | (?=[\x{800}-\x{ffff}])(\C)(\C)(\C) | (?=[\x{10000}-\x{1fffff}])(\C)(\C)(\C)(\C))</pre>In this example, a group that starts with (?| resets the capturing parenthesesnumbers in each alternative (see<a href="#dupgroupnumber">"Duplicate Group Numbers"</a>below). The assertions at the start of each branch check the next UTF-8character for values whose encoding uses 1, 2, 3, or 4 bytes, respectively. Thecharacter's individual bytes are then captured by the appropriate number of\C groups.<a name="characterclass"></a></P><br><a name="SEC9" href="#TOC1">SQUARE BRACKETS AND CHARACTER CLASSES</a><br><P>An opening square bracket introduces a character class, terminated by a closingsquare bracket. A closing square bracket on its own is not special by default.If a closing square bracket is required as a member of the class, it should bethe first data character in the class (after an initial circumflex, if present)or escaped with a backslash. This means that, by default, an empty class cannotbe defined. However, if the PCRE2_ALLOW_EMPTY_CLASS option is set, a closingsquare bracket at the start does end the (empty) class.</P><P>A character class matches a single character in the subject. A matchedcharacter must be in the set of characters defined by the class, unless thefirst character in the class definition is a circumflex, in which case thesubject character must not be in the set defined by the class. If a circumflexis actually required as a member of the class, ensure it is not the firstcharacter, or escape it with a backslash.</P><P>For example, the character class [aeiou] matches any lower case vowel, while[^aeiou] matches any character that is not a lower case vowel. Note that acircumflex is just a convenient notation for specifying the characters thatare in the class by enumerating those that are not. A class that starts with acircumflex is not an assertion; it still consumes a character from the subjectstring, and therefore it fails if the current pointer is at the end of thestring.</P><P>Characters in a class may be specified by their code points using \o, \x, or\N{U+hh..} in the usual way. When caseless matching is set, any letters in aclass represent both their upper case and lower case versions, so for example,a caseless [aeiou] matches "A" as well as "a", and a caseless [^aeiou] does notmatch "A", whereas a caseful version would. Note that there are two ASCIIcharacters, K and S, that, in addition to their lower case ASCII equivalents,are case-equivalent with Unicode U+212A (Kelvin sign) and U+017F (long S)respectively when either PCRE2_UTF or PCRE2_UCP is set.</P><P>Characters that might indicate line breaks are never treated in any special waywhen matching character classes, whatever line-ending sequence is in use, andwhatever setting of the PCRE2_DOTALL and PCRE2_MULTILINE options is used. Aclass such as [^a] always matches one of these characters.</P><P>The generic character type escape sequences \d, \D, \h, \H, \p, \P, \s,\S, \v, \V, \w, and \W may appear in a character class, and add thecharacters that they match to the class. For example, [\dABCDEF] matches anyhexadecimal digit. In UTF modes, the PCRE2_UCP option affects the meanings of\d, \s, \w and their upper case partners, just as it does when they appearoutside a character class, as described in the section entitled<a href="#genericchartypes">"Generic character types"</a>above. The escape sequence \b has a different meaning inside a characterclass; it matches the backspace character. The sequences \B, \R, and \X arenot special inside a character class. Like any other unrecognized escapesequences, they cause an error. The same is true for \N when not followed byan opening brace.</P><P>The minus (hyphen) character can be used to specify a range of characters in acharacter class. For example, [d-m] matches any letter between d and m,inclusive. If a minus character is required in a class, it must be escaped witha backslash or appear in a position where it cannot be interpreted asindicating a range, typically as the first or last character in the class,or immediately after a range. For example, [b-d-z] matches letters in the rangeb to d, a hyphen character, or z.</P><P>Perl treats a hyphen as a literal if it appears before or after a POSIX class(see below) or before or after a character type escape such as \d or \H.However, unless the hyphen is the last character in the class, Perl outputs awarning in its warning mode, as this is most likely a user error. As PCRE2 hasno facility for warning, an error is given in these cases.</P><P>It is not possible to have the literal character "]" as the end character of arange. A pattern such as [W-]46] is interpreted as a class of two characters("W" and "-") followed by a literal string "46]", so it would match "W46]" or"-46]". However, if the "]" is escaped with a backslash it is interpreted asthe end of range, so [W-\]46] is interpreted as a class containing a rangefollowed by two other characters. The octal or hexadecimal representation of"]" can also be used to end a range.</P><P>Ranges normally include all code points between the start and end characters,inclusive. They can also be used for code points specified numerically, forexample [\000-\037]. Ranges can include any characters that are valid for thecurrent mode. In any UTF mode, the so-called "surrogate" characters (thosewhose code points lie between 0xd800 and 0xdfff inclusive) may not be specifiedexplicitly by default (the PCRE2_EXTRA_ALLOW_SURROGATE_ESCAPES option disablesthis check). However, ranges such as [\x{d7ff}-\x{e000}], which include thesurrogates, are always permitted.</P><P>There is a special case in EBCDIC environments for ranges whose end points areboth specified as literal letters in the same case. For compatibility withPerl, EBCDIC code points within the range that are not letters are omitted. Forexample, [h-k] matches only four characters, even though the codes for h and kare 0x88 and 0x92, a range of 11 code points. However, if the range isspecified numerically, for example, [\x88-\x92] or [h-\x92], all code pointsare included.</P><P>If a range that includes letters is used when caseless matching is set, itmatches the letters in either case. For example, [W-c] is equivalent to[][\\^_`wxyzabc], matched caselessly, and in a non-UTF mode, if charactertables for a French locale are in use, [\xc8-\xcb] matches accented Echaracters in both cases.</P><P>A circumflex can conveniently be used with the upper case character types tospecify a more restricted set of characters than the matching lower case type.For example, the class [^\W_] matches any letter or digit, but not underscore,whereas [\w] includes underscore. A positive character class should be read as"something OR something OR ..." and a negative class as "NOT something AND NOTsomething AND NOT ...".</P><P>The only metacharacters that are recognized in character classes are backslash,hyphen (only where it can be interpreted as specifying a range), circumflex(only at the start), opening square bracket (only when it can be interpreted asintroducing a POSIX class name, or for a special compatibility feature - seethe next two sections), and the terminating closing square bracket. However,escaping other non-alphanumeric characters does no harm.</P><br><a name="SEC10" href="#TOC1">POSIX CHARACTER CLASSES</a><br><P>Perl supports the POSIX notation for character classes. This uses namesenclosed by [: and :] within the enclosing square brackets. PCRE2 also supportsthis notation. For example,<pre> [01[:alpha:]%]</pre>matches "0", "1", any alphabetic character, or "%". The supported class namesare:<pre> alnum letters and digits alpha letters ascii character codes 0 - 127 blank space or tab only cntrl control characters digit decimal digits (same as \d) graph printing characters, excluding space lower lower case letters print printing characters, including space punct printing characters, excluding letters and digits and space space white space (the same as \s from PCRE2 8.34) upper upper case letters word "word" characters (same as \w) xdigit hexadecimal digits</pre>The default "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),and space (32). If locale-specific matching is taking place, the list of spacecharacters may be different; there may be fewer or more of them. "Space" and\s match the same set of characters, as do "word" and \w.</P><P>The name "word" is a Perl extension, and "blank" is a GNU extension from Perl5.8. Another Perl extension is negation, which is indicated by a ^ characterafter the colon. For example,<pre> [12[:^digit:]]</pre>matches "1", "2", or any non-digit. PCRE2 (and Perl) also recognize the POSIXsyntax [.ch.] and [=ch=] where "ch" is a "collating element", but these are notsupported, and an error is given if they are encountered.</P><P>By default, characters with values greater than 127 do not match any of thePOSIX character classes, although this may be different for characters in therange 128-255 when locale-specific matching is happening. However, in UCP mode,unless certain options are set (see below), some of the classes are changed sothat Unicode character properties are used. This is achieved by replacingPOSIX classes with other sequences, as follows:<pre> [:alnum:] becomes \p{Xan} [:alpha:] becomes \p{L} [:blank:] becomes \h [:cntrl:] becomes \p{Cc} [:digit:] becomes \p{Nd} [:lower:] becomes \p{Ll} [:space:] becomes \p{Xps} [:upper:] becomes \p{Lu} [:word:] becomes \p{Xwd}</pre>Negated versions, such as [:^alpha:] use \P instead of \p. Four other POSIXclasses are handled specially in UCP mode:</P><P>[:graph:]This matches characters that have glyphs that mark the page when printed. InUnicode property terms, it matches all characters with the L, M, N, P, S, or Cfproperties, except for:<pre> U+061C Arabic Letter Mark U+180E Mongolian Vowel Separator U+2066 - U+2069 Various "isolate"s
</PRE></P><P>[:print:]This matches the same characters as [:graph:] plus space characters that arenot controls, that is, characters with the Zs property.</P><P>[:punct:]This matches all characters that have the Unicode P (punctuation) property,plus those characters with code points less than 256 that have the S (Symbol)property.</P><P>[:xdigit:]In addition to the ASCII hexadecimal digits, this also matches the "fullwidth"versions of those characters, whose Unicode code points start at U+FF10. Thisis a change that was made in PCRE release 10.43 for Perl compatibility.</P><P>The other POSIX classes are unchanged by PCRE2_UCP, and match only characterswith code points less than 256.</P><P>There are two options that can be used to restrict the POSIX classes to ASCIIcharacters when PCRE2_UCP is set. The option PCRE2_EXTRA_ASCII_DIGIT affectsjust [:digit:] and [:xdigit:]. Within a pattern, this can be set and unset by(?aT) and (?-aT). The PCRE2_EXTRA_ASCII_POSIX option disables UCP processingfor all POSIX classes, including [:digit:] and [:xdigit:]. Within a pattern,(?aP) and (?-aP) set and unset both these options for consistency.</P><br><a name="SEC11" href="#TOC1">COMPATIBILITY FEATURE FOR WORD BOUNDARIES</a><br><P>In the POSIX.2 compliant library that was included in 4.4BSD Unix, the uglysyntax [[:<:]] and [[:>:]] is used for matching "start of word" and "end ofword". PCRE2 treats these items as follows:<pre> [[:<:]] is converted to \b(?=\w) [[:>:]] is converted to \b(?<=\w)</pre>Only these exact character sequences are recognized. A sequence such as[a[:<:]b] provokes error for an unrecognized POSIX class name. This support isnot compatible with Perl. It is provided to help migrations from otherenvironments, and is best not used in any new patterns. Note that \b matchesat the start and the end of a word (see<a href="#smallassertions">"Simple assertions"</a>above), and in a Perl-style pattern the preceding or following characternormally shows which is wanted, without the need for the assertions that areused above in order to give exactly the POSIX behaviour. Note also that thePCRE2_UCP option changes the meaning of \w (and therefore \b) by default, soit also affects these POSIX sequences.</P><br><a name="SEC12" href="#TOC1">VERTICAL BAR</a><br><P>Vertical bar characters are used to separate alternative patterns. For example,the pattern<pre> gilbert|sullivan</pre>matches either "gilbert" or "sullivan". Any number of alternatives may appear,and an empty alternative is permitted (matching the empty string). The matchingprocess tries each alternative in turn, from left to right, and the first onethat succeeds is used. If the alternatives are within a group<a href="#group">(defined below),</a>"succeeds" means matching the rest of the main pattern as well as thealternative in the group.<a name="internaloptions"></a></P><br><a name="SEC13" href="#TOC1">INTERNAL OPTION SETTING</a><br><P>The settings of several options can be changed within a pattern by a sequenceof letters enclosed between "(?" and ")". The following are Perl-compatible,and are described in detail in the<a href="pcre2api.html"><b>pcre2api</b></a>documentation. The option letters are:<pre> i for PCRE2_CASELESS m for PCRE2_MULTILINE n for PCRE2_NO_AUTO_CAPTURE s for PCRE2_DOTALL x for PCRE2_EXTENDED xx for PCRE2_EXTENDED_MORE</pre>For example, (?im) sets caseless, multiline matching. It is also possible tounset these options by preceding the relevant letters with a hyphen, forexample (?-im). The two "extended" options are not independent; unsettingeither one cancels the effects of both of them.</P><P>A combined setting and unsetting such as (?im-sx), which sets PCRE2_CASELESSand PCRE2_MULTILINE while unsetting PCRE2_DOTALL and PCRE2_EXTENDED, is alsopermitted. Only one hyphen may appear in the options string. If a letterappears both before and after the hyphen, the option is unset. An empty optionssetting "(?)" is allowed. Needless to say, it has no effect.</P><P>If the first character following (? is a circumflex, it causes all of the aboveoptions to be unset. Letters may follow the circumflex to cause some options tobe re-instated, but a hyphen may not appear.</P><P>Some PCRE2-specific options can be changed by the same mechanism using thesepairs or individual letters:<pre> aD for PCRE2_EXTRA_ASCII_BSD aS for PCRE2_EXTRA_ASCII_BSS aW for PCRE2_EXTRA_ASCII_BSW aP for PCRE2_EXTRA_ASCII_POSIX and PCRE2_EXTRA_ASCII_DIGIT aT for PCRE2_EXTRA_ASCII_DIGIT r for PCRE2_EXTRA_CASELESS_RESTRICT J for PCRE2_DUPNAMES U for PCRE2_UNGREEDY</pre>However, except for 'r', these are not unset by (?^), which is equivalent to(?-imnrsx). If 'a' is not followed by any of the upper case letters shownabove, it sets (or unsets) all the ASCII options.</P><P>PCRE2_EXTRA_ASCII_DIGIT has no additional effect when PCRE2_EXTRA_ASCII_POSIXis set, but including it in (?aP) means that (?-aP) suppresses all ASCIIrestrictions for POSIX classes.</P><P>When one of these option changes occurs at top level (that is, not inside groupparentheses), the change applies until a subsequent change, or the end of thepattern. An option change within a group (see below for a description ofgroups) affects only that part of the group that follows it. At the end of thegroup these options are reset to the state they were before the group. Forexample,<pre> (a(?i)b)c</pre>matches abc and aBc and no other strings (assuming PCRE2_CASELESS is not setexternally). Any changes made in one alternative do carry on into subsequentbranches within the same group. For example,<pre> (a(?i)b|c)</pre>matches "ab", "aB", "c", and "C", even though when matching "C" the firstbranch is abandoned before the option setting. This is because the effects ofoption settings happen at compile time. There would be some very weirdbehaviour otherwise.</P><P>As a convenient shorthand, if any option settings are required at the start ofa non-capturing group (see the next section), the option letters mayappear between the "?" and the ":". Thus the two patterns<pre> (?i:saturday|sunday) (?:(?i)saturday|sunday)</pre>match exactly the same set of strings.</P><P><b>Note:</b> There are other PCRE2-specific options, applying to the wholepattern, which can be set by the application when the compiling function iscalled. In addition, the pattern can contain special leading sequences such as(*CRLF) to override what the application has set or what has been defaulted.Details are given in the section entitled<a href="#newlineseq">"Newline sequences"</a>above. There are also the (*UTF) and (*UCP) leading sequences that can be usedto set UTF and Unicode property modes; they are equivalent to setting thePCRE2_UTF and PCRE2_UCP options, respectively. However, the application can setthe PCRE2_NEVER_UTF or PCRE2_NEVER_UCP options, which lock out the use of the(*UTF) and (*UCP) sequences.<a name="group"></a></P><br><a name="SEC14" href="#TOC1">GROUPS</a><br><P>Groups are delimited by parentheses (round brackets), which can be nested.Turning part of a pattern into a group does two things:<br><br>1. It localizes a set of alternatives. For example, the pattern<pre> cat(aract|erpillar|)</pre>matches "cataract", "caterpillar", or "cat". Without the parentheses, it wouldmatch "cataract", "erpillar" or an empty string.<br><br>2. It creates a "capture group". This means that, when the whole patternmatches, the portion of the subject string that matched the group is passedback to the caller, separately from the portion that matched the whole pattern.(This applies only to the traditional matching function; the DFA matchingfunction does not support capturing.)</P><P>Opening parentheses are counted from left to right (starting from 1) to obtainnumbers for capture groups. For example, if the string "the red king" ismatched against the pattern<pre> the ((red|white) (king|queen))</pre>the captured substrings are "red king", "red", and "king", and are numbered 1,2, and 3, respectively.</P><P>The fact that plain parentheses fulfil two functions is not always helpful.There are often times when grouping is required without capturing. If anopening parenthesis is followed by a question mark and a colon, the groupdoes not do any capturing, and is not counted when computing the number of anysubsequent capture groups. For example, if the string "the white queen"is matched against the pattern<pre> the ((?:red|white) (king|queen))</pre>the captured substrings are "white queen" and "queen", and are numbered 1 and2. The maximum number of capture groups is 65535.</P><P>As a convenient shorthand, if any option settings are required at the start ofa non-capturing group, the option letters may appear between the "?" and the":". Thus the two patterns<pre> (?i:saturday|sunday) (?:(?i)saturday|sunday)</pre>match exactly the same set of strings. Because alternative branches are triedfrom left to right, and options are not reset until the end of the group isreached, an option setting in one branch does affect subsequent branches, sothe above patterns match "SUNDAY" as well as "Saturday".<a name="dupgroupnumber"></a></P><br><a name="SEC15" href="#TOC1">DUPLICATE GROUP NUMBERS</a><br><P>Perl 5.10 introduced a feature whereby each alternative in a group uses thesame numbers for its capturing parentheses. Such a group starts with (?| and isitself a non-capturing group. For example, consider this pattern:<pre> (?|(Sat)ur|(Sun))day</pre>Because the two alternatives are inside a (?| group, both sets of capturingparentheses are numbered one. Thus, when the pattern matches, you can lookat captured substring number one, whichever alternative matched. This constructis useful when you want to capture part, but not all, of one of a number ofalternatives. Inside a (?| group, parentheses are numbered as usual, but thenumber is reset at the start of each branch. The numbers of any capturingparentheses that follow the whole group start after the highest number used inany branch. The following example is taken from the Perl documentation. Thenumbers underneath show in which buffer the captured content will be stored.<pre> # before ---------------branch-reset----------- after / ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x # 1 2 2 3 2 3 4</pre>A backreference to a capture group uses the most recent value that is set forthe group. The following pattern matches "abcabc" or "defdef":<pre> /(?|(abc)|(def))\1/</pre>In contrast, a subroutine call to a capture group always refers to thefirst one in the pattern with the given number. The following pattern matches"abcabc" or "defabc":<pre> /(?|(abc)|(def))(?1)/</pre>A relative reference such as (?-1) is no different: it is just a convenient wayof computing an absolute group number.</P><P>If a<a href="#conditions">condition test</a>for a group's having matched refers to a non-unique number, the test istrue if any group with that number has matched.</P><P>An alternative approach to using this "branch reset" feature is to useduplicate named groups, as described in the next section.</P><br><a name="SEC16" href="#TOC1">NAMED CAPTURE GROUPS</a><br><P>Identifying capture groups by number is simple, but it can be very hard to keeptrack of the numbers in complicated patterns. Furthermore, if an expression ismodified, the numbers may change. To help with this difficulty, PCRE2 supportsthe naming of capture groups. This feature was not added to Perl until release5.10. Python had the feature earlier, and PCRE1 introduced it at release 4.0,using the Python syntax. PCRE2 supports both the Perl and the Python syntax.</P><P>In PCRE2, a capture group can be named in one of three ways: (?<name>...) or(?'name'...) as in Perl, or (?P<name>...) as in Python. Names may be up to 128code units long. When PCRE2_UTF is not set, they may contain only ASCIIalphanumeric characters and underscores, but must start with a non-digit. WhenPCRE2_UTF is set, the syntax of group names is extended to allow any Unicodeletter or Unicode decimal digit. In other words, group names must match one ofthese patterns:<pre> ^[_A-Za-z][_A-Za-z0-9]*\z when PCRE2_UTF is not set ^[_\p{L}][_\p{L}\p{Nd}]*\z when PCRE2_UTF is set</pre>References to capture groups from other parts of the pattern, such as<a href="#backreferences">backreferences,</a><a href="#recursion">recursion,</a>and<a href="#conditions">conditions,</a>can all be made by name as well as by number.</P><P>Named capture groups are allocated numbers as well as names, exactly asif the names were not present. In both PCRE2 and Perl, capture groupsare primarily identified by numbers; any names are just aliases for thesenumbers. The PCRE2 API provides function calls for extracting the completename-to-number translation table from a compiled pattern, as well asconvenience functions for extracting captured substrings by name.</P><P><b>Warning:</b> When more than one capture group has the same number, asdescribed in the previous section, a name given to one of them applies to allof them. Perl allows identically numbered groups to have different names.Consider this pattern, where there are two capture groups, both numbered 1:<pre> (?|(?<AA>aa)|(?<BB>bb))</pre>Perl allows this, with both names AA and BB as aliases of group 1. Thus, aftera successful match, both names yield the same value (either "aa" or "bb").</P><P>In an attempt to reduce confusion, PCRE2 does not allow the same group numberto be associated with more than one name. The example above provokes acompile-time error. However, there is still scope for confusion. Consider thispattern:<pre> (?|(?<AA>aa)|(bb))</pre>Although the second group number 1 is not explicitly named, the name AA isstill an alias for any group 1. Whether the pattern matches "aa" or "bb", areference by name to group AA yields the matched string.</P><P>By default, a name must be unique within a pattern, except that duplicate namesare permitted for groups with the same number, for example:<pre> (?|(?<AA>aa)|(?<AA>bb))</pre>The duplicate name constraint can be disabled by setting the PCRE2_DUPNAMESoption at compile time, or by the use of (?J) within the pattern, as describedin the section entitled<a href="#internaloptions">"Internal Option Setting"</a>above.</P><P>Duplicate names can be useful for patterns where only one instance of the namedcapture group can match. Suppose you want to match the name of a weekday,either as a 3-letter abbreviation or as the full name, and in both cases youwant to extract the abbreviation. This pattern (ignoring the line breaks) doesthe job:<pre> (?J) (?<DN>Mon|Fri|Sun)(?:day)?| (?<DN>Tue)(?:sday)?| (?<DN>Wed)(?:nesday)?| (?<DN>Thu)(?:rsday)?| (?<DN>Sat)(?:urday)?</pre>There are five capture groups, but only one is ever set after a match. Theconvenience functions for extracting the data by name returns the substring forthe first (and in this example, the only) group of that name that matched. Thissaves searching to find which numbered group it was. (An alternative way ofsolving this problem is to use a "branch reset" group, as described in theprevious section.)</P><P>If you make a backreference to a non-unique named group from elsewhere in thepattern, the groups to which the name refers are checked in the order in whichthey appear in the overall pattern. The first one that is set is used for thereference. For example, this pattern matches both "foofoo" and "barbar" but not"foobar" or "barfoo":<pre> (?J)(?:(?<n>foo)|(?<n>bar))\k<n>
</PRE></P><P>If you make a subroutine call to a non-unique named group, the one thatcorresponds to the first occurrence of the name is used. In the absence ofduplicate numbers this is the one with the lowest number.</P><P>If you use a named reference in a conditiontest (see the<a href="#conditions">section about conditions</a>below), either to check whether a capture group has matched, or to check forrecursion, all groups with the same name are tested. If the condition is truefor any one of them, the overall condition is true. This is the same behaviouras testing by number. For further details of the interfaces for handling namedcapture groups, see the<a href="pcre2api.html"><b>pcre2api</b></a>documentation.</P><br><a name="SEC17" href="#TOC1">REPETITION</a><br><P>Repetition is specified by quantifiers, which may follow any one of theseitems:<pre> a literal data character the dot metacharacter the \C escape sequence the \R escape sequence the \X escape sequence any escape sequence that matches a single character a character class a backreference a parenthesized group (including lookaround assertions) a subroutine call (recursive or otherwise)</pre>If a quantifier does not follow a repeatable item, an error occurs. Thegeneral repetition quantifier specifies a minimum and maximum number ofpermitted matches by giving two numbers in curly brackets (braces), separatedby a comma. The numbers must be less than 65536, and the first must be lessthan or equal to the second. For example,<pre> z{2,4}</pre>matches "zz", "zzz", or "zzzz". A closing brace on its own is not a specialcharacter. If the second number is omitted, but the comma is present, there isno upper limit; if the second number and the comma are both omitted, thequantifier specifies an exact number of required matches. Thus<pre> [aeiou]{3,}</pre>matches at least 3 successive vowels, but may match many more, whereas<pre> \d{8}</pre>matches exactly 8 digits. If the first number is omitted, the lower limit istaken as zero; in this case the upper limit must be present.<pre> X{,4} is interpreted as X{0,4}</pre>This is a change in behaviour that happened in Perl 5.34.0 and PCRE2 10.43. Inearlier versions such a sequence was not interpreted as a quantifier. Otherregular expression engines may behave either way.</P><P>If the characters that follow an opening brace do not match the syntax of aquantifier, the brace is taken as a literal character. In particular, thismeans that {,} is a literal string of three characters.</P><P>Note that not every opening brace is potentially the start of a quantifierbecause braces are used in other items such as \N{U+345} or \k{name}.</P><P>In UTF modes, quantifiers apply to characters rather than to individual codeunits. Thus, for example, \x{100}{2} matches two characters, each ofwhich is represented by a two-byte sequence in a UTF-8 string. Similarly,\X{3} matches three Unicode extended grapheme clusters, each of which may beseveral code units long (and they may be of different lengths).</P><P>The quantifier {0} is permitted, causing the expression to behave as if theprevious item and the quantifier were not present. This may be useful forcapture groups that are referenced as<a href="#groupsassubroutines">subroutines</a>from elsewhere in the pattern (but see also the section entitled<a href="#subdefine">"Defining capture groups for use by reference only"</a>below). Except for parenthesized groups, items that have a {0} quantifier areomitted from the compiled pattern.</P><P>For convenience, the three most common quantifiers have single-characterabbreviations:<pre> * is equivalent to {0,} + is equivalent to {1,} ? is equivalent to {0,1}</pre>It is possible to construct infinite loops by following a group that can matchno characters with a quantifier that has no upper limit, for example:<pre> (a?)*</pre>Earlier versions of Perl and PCRE1 used to give an error at compile time forsuch patterns. However, because there are cases where this can be useful, suchpatterns are now accepted, but whenever an iteration of such a group matches nocharacters, matching moves on to the next item in the pattern instead ofrepeatedly matching an empty string. This does not prevent backtracking intoany of the iterations if a subsequent item fails to match.</P><P>By default, quantifiers are "greedy", that is, they match as much as possible(up to the maximum number of permitted repetitions), without causing the restof the pattern to fail. The classic example of where this gives problems is intrying to match comments in C programs. These appear between /* and */ andwithin the comment, individual * and / characters may appear. An attempt tomatch C comments by applying the pattern<pre> /\*.*\*/</pre>to the string<pre> /* first comment */ not comment /* second comment */</pre>fails, because it matches the entire string owing to the greediness of the .*item. However, if a quantifier is followed by a question mark, it ceases to begreedy, and instead matches the minimum number of times possible, so thepattern<pre> /\*.*?\*/</pre>does the right thing with C comments. The meaning of the various quantifiers isnot otherwise changed, just the preferred number of matches. Do not confusethis use of question mark with its use as a quantifier in its own right.Because it has two uses, it can sometimes appear doubled, as in<pre> \d??\d</pre>which matches one digit by preference, but can match two if that is the onlyway the rest of the pattern matches.</P><P>If the PCRE2_UNGREEDY option is set (an option that is not available in Perl),the quantifiers are not greedy by default, but individual ones can be madegreedy by following them with a question mark. In other words, it inverts thedefault behaviour.</P><P>When a parenthesized group is quantified with a minimum repeat count thatis greater than 1 or with a limited maximum, more memory is required for thecompiled pattern, in proportion to the size of the minimum or maximum.</P><P>If a pattern starts with .* or .{0,} and the PCRE2_DOTALL option (equivalentto Perl's /s) is set, thus allowing the dot to match newlines, the pattern isimplicitly anchored, because whatever follows will be tried against everycharacter position in the subject string, so there is no point in retrying theoverall match at any position after the first. PCRE2 normally treats such apattern as though it were preceded by \A.</P><P>In cases where it is known that the subject string contains no newlines, it isworth setting PCRE2_DOTALL in order to obtain this optimization, oralternatively, using ^ to indicate anchoring explicitly.</P><P>However, there are some cases where the optimization cannot be used. When .*is inside capturing parentheses that are the subject of a backreferenceelsewhere in the pattern, a match at the start may fail where a later onesucceeds. Consider, for example:<pre> (.*)abc\1</pre>If the subject is "xyz123abc123" the match point is the fourth character. Forthis reason, such a pattern is not implicitly anchored.</P><P>Another case where implicit anchoring is not applied is when the leading .* isinside an atomic group. Once again, a match at the start may fail where a laterone succeeds. Consider this pattern:<pre> (?>.*?a)b</pre>It matches "ab" in the subject "aab". The use of the backtracking control verbs(*PRUNE) and (*SKIP) also disable this optimization, and there is an option,PCRE2_NO_DOTSTAR_ANCHOR, to do so explicitly.</P><P>When a capture group is repeated, the value captured is the substring thatmatched the final iteration. For example, after<pre> (tweedle[dume]{3}\s*)+</pre>has matched "tweedledum tweedledee" the value of the captured substring is"tweedledee". However, if there are nested capture groups, the correspondingcaptured values may have been set in previous iterations. For example, after<pre> (a|(b))+</pre>matches "aba" the value of the second captured substring is "b".<a name="atomicgroup"></a></P><br><a name="SEC18" href="#TOC1">ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS</a><br><P>With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")repetition, failure of what follows normally causes the repeated item to bere-evaluated to see if a different number of repeats allows the rest of thepattern to match. Sometimes it is useful to prevent this, either to change thenature of the match, or to cause it fail earlier than it otherwise might, whenthe author of the pattern knows there is no point in carrying on.</P><P>Consider, for example, the pattern \d+foo when applied to the subject line<pre> 123456bar</pre>After matching all 6 digits and then failing to match "foo", the normalaction of the matcher is to try again with only 5 digits matching the \d+item, and then with 4, and so on, before ultimately failing. "Atomic grouping"(a term taken from Jeffrey Friedl's book) provides the means for specifyingthat once a group has matched, it is not to be re-evaluated in this way.</P><P>If we use atomic grouping for the previous example, the matcher gives upimmediately on failing to match "foo" the first time. The notation is a kind ofspecial parenthesis, starting with (?> as in this example:<pre> (?>\d+)foo</pre>Perl 5.28 introduced an experimental alphabetic form starting with (* which maybe easier to remember:<pre> (*atomic:\d+)foo</pre>This kind of parenthesized group "locks up" the part of the pattern it containsonce it has matched, and a failure further into the pattern is prevented frombacktracking into it. Backtracking past it to previous items, however, works asnormal.</P><P>An alternative description is that a group of this type matches exactly thestring of characters that an identical standalone pattern would match, ifanchored at the current point in the subject string.</P><P>Atomic groups are not capture groups. Simple cases such as the above examplecan be thought of as a maximizing repeat that must swallow everything it can.So, while both \d+ and \d+? are prepared to adjust the number of digits theymatch in order to make the rest of the pattern match, (?>\d+) can only matchan entire sequence of digits.</P><P>Atomic groups in general can of course contain arbitrarily complicatedexpressions, and can be nested. However, when the contents of an atomicgroup is just a single repeated item, as in the example above, a simplernotation, called a "possessive quantifier" can be used. This consists of anadditional + character following a quantifier. Using this notation, theprevious example can be rewritten as<pre> \d++foo</pre>Note that a possessive quantifier can be used with an entire group, forexample:<pre> (abc|xyz){2,3}+</pre>Possessive quantifiers are always greedy; the setting of the PCRE2_UNGREEDYoption is ignored. They are a convenient notation for the simpler forms ofatomic group. However, there is no difference in the meaning of a possessivequantifier and the equivalent atomic group, though there may be a performancedifference; possessive quantifiers should be slightly faster.</P><P>The possessive quantifier syntax is an extension to the Perl 5.8 syntax.Jeffrey Friedl originated the idea (and the name) in the first edition of hisbook. Mike McCloskey liked it, so implemented it when he built Sun's Javapackage, and PCRE1 copied it from there. It found its way into Perl at release5.10.</P><P>PCRE2 has an optimization that automatically "possessifies" certain simplepattern constructs. For example, the sequence A+B is treated as A++B becausethere is no point in backtracking into a sequence of A's when B must follow.This feature can be disabled by the PCRE2_NO_AUTOPOSSESS option, or startingthe pattern with (*NO_AUTO_POSSESS).</P><P>When a pattern contains an unlimited repeat inside a group that can itself berepeated an unlimited number of times, the use of an atomic group is the onlyway to avoid some failing matches taking a very long time indeed. The pattern<pre> (\D+|<\d+>)*[!?]</pre>matches an unlimited number of substrings that either consist of non-digits, ordigits enclosed in <>, followed by either ! or ?. When it matches, it runsquickly. However, if it is applied to<pre> aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa</pre>it takes a long time before reporting failure. This is because the string canbe divided between the internal \D+ repeat and the external * repeat in alarge number of ways, and all have to be tried. (The example uses [!?] ratherthan a single character at the end, because both PCRE2 and Perl have anoptimization that allows for fast failure when a single character is used. Theyremember the last single character that is required for a match, and fail earlyif it is not present in the string.) If the pattern is changed so that it usesan atomic group, like this:<pre> ((?>\D+)|<\d+>)*[!?]</pre>sequences of non-digits cannot be broken, and failure happens quickly.<a name="backreferences"></a></P><br><a name="SEC19" href="#TOC1">BACKREFERENCES</a><br><P>Outside a character class, a backslash followed by a digit greater than 0 (andpossibly further digits) is a backreference to a capture group earlier (thatis, to its left) in the pattern, provided there have been that many previouscapture groups.</P><P>However, if the decimal number following the backslash is less than 8, it isalways taken as a backreference, and causes an error only if there are not thatmany capture groups in the entire pattern. In other words, the group that isreferenced need not be to the left of the reference for numbers less than 8. A"forward backreference" of this type can make sense when a repetition isinvolved and the group to the right has participated in an earlier iteration.</P><P>It is not possible to have a numerical "forward backreference" to a group whosenumber is 8 or more using this syntax because a sequence such as \50 isinterpreted as a character defined in octal. See the subsection entitled"Non-printing characters"<a href="#digitsafterbackslash">above</a>for further details of the handling of digits following a backslash. Otherforms of backreferencing do not suffer from this restriction. In particular,there is no problem when named capture groups are used (see below).</P><P>Another way of avoiding the ambiguity inherent in the use of digits following abackslash is to use the \g escape sequence. This escape must be followed by asigned or unsigned number, optionally enclosed in braces. These examples areall identical:<pre> (ring), \1 (ring), \g1 (ring), \g{1}</pre>An unsigned number specifies an absolute reference without the ambiguity thatis present in the older syntax. It is also useful when literal digits followthe reference. A signed number is a relative reference. Consider this example:<pre> (abc(def)ghi)\g{-1}</pre>The sequence \g{-1} is a reference to the capture group whose number is oneless than the number of the next group to be started, so in this example (wherethe next group would be numbered 3) is it equivalent to \2, and \g{-2} wouldbe equivalent to \1. Note that if this construct is inside a capture group,that group is included in the count, so in this example \g{-2} also refers togroup 1:<pre> (A)(\g{-2}B)</pre>The use of relative references can be helpful in long patterns, and also inpatterns that are created by joining together fragments that contain referenceswithin themselves.</P><P>The sequence \g{+1} is a reference to the next capture group that is startedafter this item, and \g{+2} refers to the one after that, and so on. This kindof forward reference can be useful in patterns that repeat. Perl does notsupport the use of + in this way.</P><P>A backreference matches whatever actually most recently matched the capturegroup in the current subject string, rather than anything at all that matchesthe group (see<a href="#groupsassubroutines">"Groups as subroutines"</a>below for a way of doing that). So the pattern<pre> (sens|respons)e and \1ibility</pre>matches "sense and sensibility" and "response and responsibility", but not"sense and responsibility". If caseful matching is in force at the time of thebackreference, the case of letters is relevant. For example,<pre> ((?i)rah)\s+\1</pre>matches "rah rah" and "RAH RAH", but not "RAH rah", even though the originalcapture group is matched caselessly.</P><P>There are several different ways of writing backreferences to named capturegroups. The .NET syntax is \k{name}, the Python syntax is (?=name), and theoriginal Perl syntax is \k<name> or \k'name'. All of these are now supportedby both Perl and PCRE2. Perl 5.10's unified backreference syntax, in which \gcan be used for both numeric and named references, is also supported by PCRE2.We could rewrite the above example in any of the following ways:<pre> (?<p1>(?i)rah)\s+\k<p1> (?'p1'(?i)rah)\s+\k{p1} (?P<p1>(?i)rah)\s+(?P=p1) (?<p1>(?i)rah)\s+\g{p1}</pre>A capture group that is referenced by name may appear in the pattern before orafter the reference.</P><P>There may be more than one backreference to the same group. If a group has notactually been used in a particular match, backreferences to it always fail bydefault. For example, the pattern<pre> (a|(bc))\2</pre>always fails if it starts to match "a" rather than "bc". However, if thePCRE2_MATCH_UNSET_BACKREF option is set at compile time, a backreference to anunset value matches an empty string.</P><P>Because there may be many capture groups in a pattern, all digits following abackslash are taken as part of a potential backreference number. If the patterncontinues with a digit character, some delimiter must be used to terminate thebackreference. If the PCRE2_EXTENDED or PCRE2_EXTENDED_MORE option is set, thiscan be white space. Otherwise, the \g{} syntax or an empty comment (see<a href="#comments">"Comments"</a>below) can be used.</P><br><b>Recursive backreferences</b><br><P>A backreference that occurs inside the group to which it refers fails when thegroup is first used, so, for example, (a\1) never matches. However, suchreferences can be useful inside repeated groups. For example, the pattern<pre> (a|b\1)+</pre>matches any number of "a"s and also "aba", "ababbaa" etc. At each iteration ofthe group, the backreference matches the character string corresponding to theprevious iteration. In order for this to work, the pattern must be such thatthe first iteration does not need to match the backreference. This can be doneusing alternation, as in the example above, or by a quantifier with a minimumof zero.</P><P>For versions of PCRE2 less than 10.25, backreferences of this type used tocause the group that they reference to be treated as an<a href="#atomicgroup">atomic group.</a>This restriction no longer applies, and backtracking into such groups can occuras normal.<a name="bigassertions"></a></P><br><a name="SEC20" href="#TOC1">ASSERTIONS</a><br><P>An assertion is a test on the characters following or preceding the currentmatching point that does not consume any characters. The simple assertionscoded as \b, \B, \A, \G, \Z, \z, ^ and $ are described<a href="#smallassertions">above.</a></P><P>More complicated assertions are coded as parenthesized groups. There are twokinds: those that look ahead of the current position in the subject string, andthose that look behind it, and in each case an assertion may be positive (mustmatch for the assertion to be true) or negative (must not match for theassertion to be true). An assertion group is matched in the normal way,and if it is true, matching continues after it, but with the matching positionin the subject string reset to what it was before the assertion was processed.</P><P>The Perl-compatible lookaround assertions are atomic. If an assertion is true,but there is a subsequent matching failure, there is no backtracking into theassertion. However, there are some cases where non-atomic assertions can beuseful. PCRE2 has some support for these, described in the section entitled<a href="#nonatomicassertions">"Non-atomic assertions"</a>below, but they are not Perl-compatible.</P><P>A lookaround assertion may appear as the condition in a<a href="#conditions">conditional group</a>(see below). In this case, the result of matching the assertion determineswhich branch of the condition is followed.</P><P>Assertion groups are not capture groups. If an assertion contains capturegroups within it, these are counted for the purposes of numbering the capturegroups in the whole pattern. Within each branch of an assertion, locallycaptured substrings may be referenced in the usual way. For example, a sequencesuch as (.)\g{-1} can be used to check that two adjacent characters are thesame.</P><P>When a branch within an assertion fails to match, any substrings that werecaptured are discarded (as happens with any pattern branch that fails tomatch). A negative assertion is true only when all its branches fail to match;this means that no captured substrings are ever retained after a successfulnegative assertion. When an assertion contains a matching branch, what happensdepends on the type of assertion.</P><P>For a positive assertion, internally captured substrings in the successfulbranch are retained, and matching continues with the next pattern item afterthe assertion. For a negative assertion, a matching branch means that theassertion is not true. If such an assertion is being used as a condition in a<a href="#conditions">conditional group</a>(see below), captured substrings are retained, because matching continues withthe "no" branch of the condition. For other failing negative assertions,control passes to the previous backtracking point, thus discarding any capturedstrings within the assertion.</P><P>Most assertion groups may be repeated; though it makes no sense to assert thesame thing several times, the side effect of capturing in positive assertionsmay occasionally be useful. However, an assertion that forms the condition fora conditional group may not be quantified. PCRE2 used to restrict therepetition of assertions, but from release 10.35 the only restriction is thatan unlimited maximum repetition is changed to be one more than the minimum. Forexample, {3,} is treated as {3,4}.</P><br><b>Alphabetic assertion names</b><br><P>Traditionally, symbolic sequences such as (?= and (?<= have been used tospecify lookaround assertions. Perl 5.28 introduced some experimentalalphabetic alternatives which might be easier to remember. They all start with(* instead of (? and must be written using lower case letters. PCRE2 supportsthe following synonyms:<pre> (*positive_lookahead: or (*pla: is the same as (?= (*negative_lookahead: or (*nla: is the same as (?! (*positive_lookbehind: or (*plb: is the same as (?<= (*negative_lookbehind: or (*nlb: is the same as (?<!</pre>For example, (*pla:foo) is the same assertion as (?=foo). In the followingsections, the various assertions are described using the original symbolicforms.</P><br><b>Lookahead assertions</b><br><P>Lookahead assertions start with (?= for positive assertions and (?! fornegative assertions. For example,<pre> \w+(?=;)</pre>matches a word followed by a semicolon, but does not include the semicolon inthe match, and<pre> foo(?!bar)</pre>matches any occurrence of "foo" that is not followed by "bar". Note that theapparently similar pattern<pre> (?!foo)bar</pre>does not find an occurrence of "bar" that is preceded by something other than"foo"; it finds any occurrence of "bar" whatsoever, because the assertion(?!foo) is always true when the next three characters are "bar". Alookbehind assertion is needed to achieve the other effect.</P><P>If you want to force a matching failure at some point in a pattern, the mostconvenient way to do it is with (?!) because an empty string always matches, soan assertion that requires there not to be an empty string must always fail.The backtracking control verb (*FAIL) or (*F) is a synonym for (?!).<a name="lookbehind"></a></P><br><b>Lookbehind assertions</b><br><P>Lookbehind assertions start with (?<= for positive assertions and (?<! fornegative assertions. For example,<pre> (?<!foo)bar</pre>does find an occurrence of "bar" that is not preceded by "foo". The contents ofa lookbehind assertion are restricted such that there must be a known maximumto the lengths of all the strings it matches. There are two cases:</P><P>If every top-level alternative matches a fixed length, for example<pre> (?<=colour|color)</pre>there is a limit of 65535 characters to the lengths, which do not have to bethe same, as this example demonstrates. This is the only kind of lookbehindsupported by PCRE2 versions earlier than 10.43 and by the alternative matchingfunction <b>pcre2_dfa_match()</b>.</P><P>In PCRE2 10.43 and later, <b>pcre2_match()</b> supports lookbehind assertions inwhich one or more top-level alternatives can match more than one string length,for example<pre> (?<=colou?r)</pre>The maximum matching length for any branch of the lookbehind is limited to avalue set by the calling program (default 255 characters). Unlimited repetition(for example \d*) is not supported. In some cases, the escape sequence \K<a href="#resetmatchstart">(see above)</a>can be used instead of a lookbehind assertion at the start of a pattern to getround the length limit restriction.</P><P>In UTF-8 and UTF-16 modes, PCRE2 does not allow the \C escape (which matches asingle code unit even in a UTF mode) to appear in lookbehind assertions,because it makes it impossible to calculate the length of the lookbehind. The\X and \R escapes, which can match different numbers of code units, are neverpermitted in lookbehinds.</P><P><a href="#groupsassubroutines">"Subroutine"</a>calls (see below) such as (?2) or (?&X) are permitted in lookbehinds, as longas the called capture group matches a limited-length string. However,<a href="#recursion">recursion,</a>that is, a "subroutine" call into a group that is already active,is not supported.</P><P>PCRE2 supports backreferences in lookbehinds, but only if certain conditionsare met. The PCRE2_MATCH_UNSET_BACKREF option must not be set, there must be nouse of (?| in the pattern (it creates duplicate group numbers), and if thebackreference is by name, the name must be unique. Of course, the referencedgroup must itself match a limited length substring. The following patternmatches words containing at least two characters that begin and end with thesame character:<pre> \b(\w)\w++(?<=\1)</PRE></P><P>Possessive quantifiers can be used in conjunction with lookbehind assertions tospecify efficient matching at the end of subject strings. Consider a simplepattern such as<pre> abcd$</pre>when applied to a long string that does not match. Because matching proceedsfrom left to right, PCRE2 will look for each "a" in the subject and then see ifwhat follows matches the rest of the pattern. If the pattern is specified as<pre> ^.*abcd$</pre>the initial .* matches the entire string at first, but when this fails (becausethere is no following "a"), it backtracks to match all but the last character,then all but the last two characters, and so on. Once again the search for "a"covers the entire string, from right to left, so we are no better off. However,if the pattern is written as<pre> ^.*+(?<=abcd)</pre>there can be no backtracking for the .*+ item because of the possessivequantifier; it can match only the entire string. The subsequent lookbehindassertion does a single test on the last four characters. If it fails, thematch fails immediately. For long strings, this approach makes a significantdifference to the processing time.</P><br><b>Using multiple assertions</b><br><P>Several assertions (of any sort) may occur in succession. For example,<pre> (?<=\d{3})(?<!999)foo</pre>matches "foo" preceded by three digits that are not "999". Notice that each ofthe assertions is applied independently at the same point in the subjectstring. First there is a check that the previous three characters are alldigits, and then there is a check that the same three characters are not "999".This pattern does <i>not</i> match "foo" preceded by six characters, the firstof which are digits and the last three of which are not "999". For example, itdoesn't match "123abcfoo". A pattern to do that is<pre> (?<=\d{3}...)(?<!999)foo</pre>This time the first assertion looks at the preceding six characters, checkingthat the first three are digits, and then the second assertion checks that thepreceding three characters are not "999".</P><P>Assertions can be nested in any combination. For example,<pre> (?<=(?<!foo)bar)baz</pre>matches an occurrence of "baz" that is preceded by "bar" which in turn is notpreceded by "foo", while<pre> (?<=\d{3}(?!999)...)foo</pre>is another pattern that matches "foo" preceded by three digits and any threecharacters that are not "999".<a name="nonatomicassertions"></a></P><br><a name="SEC21" href="#TOC1">NON-ATOMIC ASSERTIONS</a><br><P>Traditional lookaround assertions are atomic. That is, if an assertion is true,but there is a subsequent matching failure, there is no backtracking into theassertion. However, there are some cases where non-atomic positive assertionscan be useful. PCRE2 provides these using the following syntax:<pre> (*non_atomic_positive_lookahead: or (*napla: or (?* (*non_atomic_positive_lookbehind: or (*naplb: or (?<*</pre>Consider the problem of finding the right-most word in a string that alsoappears earlier in the string, that is, it must appear at least twice in total.This pattern returns the required result as captured substring 1:<pre> ^(?x)(*napla: .* \b(\w++)) (?> .*? \b\1\b ){2}</pre>For a subject such as "word1 word2 word3 word2 word3 word4" the result is"word3". How does it work? At the start, ^(?x) anchors the pattern and sets the"x" option, which causes white space (introduced for readability) to beignored. Inside the assertion, the greedy .* at first consumes the entirestring, but then has to backtrack until the rest of the assertion can match aword, which is captured by group 1. In other words, when the assertion firstsucceeds, it captures the right-most word in the string.</P><P>The current matching point is then reset to the start of the subject, and therest of the pattern match checks for two occurrences of the captured word,using an ungreedy .*? to scan from the left. If this succeeds, we are done, butif the last word in the string does not occur twice, this part of the patternfails. If a traditional atomic lookahead (?= or (*pla: had been used, theassertion could not be re-entered, and the whole match would fail. The patternwould succeed only if the very last word in the subject was found twice.</P><P>Using a non-atomic lookahead, however, means that when the last word does notoccur twice in the string, the lookahead can backtrack and find the second-lastword, and so on, until either the match succeeds, or all words have beentested.</P><P>Two conditions must be met for a non-atomic assertion to be useful: thecontents of one or more capturing groups must change after a backtrack into theassertion, and there must be a backreference to a changed group later in thepattern. If this is not the case, the rest of the pattern match fails exactlyas before because nothing has changed, so using a non-atomic assertion justwastes resources.</P><P>There is one exception to backtracking into a non-atomic assertion. If an(*ACCEPT) control verb is triggered, the assertion succeeds atomically. Thatis, a subsequent match failure cannot backtrack into the assertion.</P><P>Non-atomic assertions are not supported by the alternative matching function<b>pcre2_dfa_match()</b>. They are supported by JIT, but only if they do notcontain any control verbs such as (*ACCEPT). (This may change in future). Notethat assertions that appear as conditions for<a href="#conditions">conditional groups</a>(see below) must be atomic.</P><br><a name="SEC22" href="#TOC1">SCRIPT RUNS</a><br><P>In concept, a script run is a sequence of characters that are all from the sameUnicode script such as Latin or Greek. However, because some scripts arecommonly used together, and because some diacritical and other marks are usedwith multiple scripts, it is not that simple. There is a full description ofthe rules that PCRE2 uses in the section entitled<a href="pcre2unicode.html#scriptruns">"Script Runs"</a>in the<a href="pcre2unicode.html"><b>pcre2unicode</b></a>documentation.</P><P>If part of a pattern is enclosed between (*script_run: or (*sr: and a closingparenthesis, it fails if the sequence of characters that it matches are not ascript run. After a failure, normal backtracking occurs. Script runs can beused to detect spoofing attacks using characters that look the same, but arefrom different scripts. The string "paypal.com" is an infamous example, wherethe letters could be a mixture of Latin and Cyrillic. This pattern ensures thatthe matched characters in a sequence of non-spaces that follow white space area script run:<pre> \s+(*sr:\S+)</pre>To be sure that they are all from the Latin script (for example), a lookaheadcan be used:<pre> \s+(?=\p{Latin})(*sr:\S+)</pre>This works as long as the first character is expected to be a character in thatscript, and not (for example) punctuation, which is allowed with any script. Ifthis is not the case, a more creative lookahead is needed. For example, ifdigits, underscore, and dots are permitted at the start:<pre> \s+(?=[0-9_.]*\p{Latin})(*sr:\S+)
</PRE></P><P>In many cases, backtracking into a script run pattern fragment is notdesirable. The script run can employ an atomic group to prevent this. Becausethis is a common requirement, a shorthand notation is provided by(*atomic_script_run: or (*asr:<pre> (*asr:...) is the same as (*sr:(?>...))</pre>Note that the atomic group is inside the script run. Putting it outside wouldnot prevent backtracking into the script run pattern.</P><P>Support for script runs is not available if PCRE2 is compiled without Unicodesupport. A compile-time error is given if any of the above constructs isencountered. Script runs are not supported by the alternate matching function,<b>pcre2_dfa_match()</b> because they use the same mechanism as capturingparentheses.</P><P><b>Warning:</b> The (*ACCEPT) control verb<a href="#acceptverb">(see below)</a>should not be used within a script run group, because it causes an immediateexit from the group, bypassing the script run checking.<a name="conditions"></a></P><br><a name="SEC23" href="#TOC1">CONDITIONAL GROUPS</a><br><P>It is possible to cause the matching process to obey a pattern fragmentconditionally or to choose between two alternative fragments, depending onthe result of an assertion, or whether a specific capture group hasalready been matched. The two possible forms of conditional group are:<pre> (?(condition)yes-pattern) (?(condition)yes-pattern|no-pattern)</pre>If the condition is satisfied, the yes-pattern is used; otherwise theno-pattern (if present) is used. An absent no-pattern is equivalent to an emptystring (it always matches). If there are more than two alternatives in thegroup, a compile-time error occurs. Each of the two alternatives may itselfcontain nested groups of any form, including conditional groups; therestriction to two alternatives applies only at the level of the conditionitself. This pattern fragment is an example where the alternatives are complex:<pre> (?(1) (A|B|C) | (D | (?(2)E|F) | E) )
</PRE></P><P>There are five kinds of condition: references to capture groups, references torecursion, two pseudo-conditions called DEFINE and VERSION, and assertions.</P><br><b>Checking for a used capture group by number</b><br><P>If the text between the parentheses consists of a sequence of digits, thecondition is true if a capture group of that number has previously matched. Ifthere is more than one capture group with the same number (see the earlier<a href="#recursion">section about duplicate group numbers),</a>the condition is true if any of them have matched. An alternative notation,which is a PCRE2 extension, not supported by Perl, is to precede the digitswith a plus or minus sign. In this case, the group number is relative ratherthan absolute. The most recently opened capture group (which could be enclosingthis condition) can be referenced by (?(-1), the next most recent by (?(-2),and so on. Inside loops it can also make sense to refer to subsequent groups.The next capture group to be opened can be referenced as (?(+1), and so on. Thevalue zero in any of these forms is not used; it provokes a compile-time error.</P><P>Consider the following pattern, which contains non-significant white space tomake it more readable (assume the PCRE2_EXTENDED option) and to divide it intothree parts for ease of discussion:<pre> ( \( )? [^()]+ (?(1) \) )</pre>The first part matches an optional opening parenthesis, and if thatcharacter is present, sets it as the first captured substring. The second partmatches one or more characters that are not parentheses. The third part is aconditional group that tests whether or not the first capture groupmatched. If it did, that is, if subject started with an opening parenthesis,the condition is true, and so the yes-pattern is executed and a closingparenthesis is required. Otherwise, since no-pattern is not present, theconditional group matches nothing. In other words, this pattern matches asequence of non-parentheses, optionally enclosed in parentheses.</P><P>If you were embedding this pattern in a larger one, you could use a relativereference:<pre> ...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...</pre>This makes the fragment independent of the parentheses in the larger pattern.</P><br><b>Checking for a used capture group by name</b><br><P>Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a usedcapture group by name. For compatibility with earlier versions of PCRE1, whichhad this facility before Perl, the syntax (?(name)...) is also recognized.Note, however, that undelimited names consisting of the letter R followed bydigits are ambiguous (see the following section). Rewriting the above exampleto use a named group gives this:<pre> (?<OPEN> \( )? [^()]+ (?(<OPEN>) \) )</pre>If the name used in a condition of this kind is a duplicate, the test isapplied to all groups of the same name, and is true if any one of them hasmatched.</P><br><b>Checking for pattern recursion</b><br><P>"Recursion" in this sense refers to any subroutine-like call from one part ofthe pattern to another, whether or not it is actually recursive. See thesections entitled<a href="#recursion">"Recursive patterns"</a>and<a href="#groupsassubroutines">"Groups as subroutines"</a>below for details of recursion and subroutine calls.</P><P>If a condition is the string (R), and there is no capture group with the nameR, the condition is true if matching is currently in a recursion or subroutinecall to the whole pattern or any capture group. If digits follow the letter R,and there is no group with that name, the condition is true if the most recentcall is into a group with the given number, which must exist somewhere in theoverall pattern. This is a contrived example that is equivalent to a+b:<pre> ((?(R1)a+|(?1)b))</pre>However, in both cases, if there is a capture group with a matching name, thecondition tests for its being set, as described in the section above, insteadof testing for recursion. For example, creating a group with the name R1 byadding (?<R1>) to the above pattern completely changes its meaning.</P><P>If a name preceded by ampersand follows the letter R, for example:<pre> (?(R&name)...)</pre>the condition is true if the most recent recursion is into a group of that name(which must exist within the pattern).</P><P>This condition does not check the entire recursion stack. It tests only thecurrent level. If the name used in a condition of this kind is a duplicate, thetest is applied to all groups of the same name, and is true if any one ofthem is the most recent recursion.</P><P>At "top level", all these recursion test conditions are false.<a name="subdefine"></a></P><br><b>Defining capture groups for use by reference only</b><br><P>If the condition is the string (DEFINE), the condition is always false, even ifthere is a group with the name DEFINE. In this case, there may be only onealternative in the rest of the conditional group. It is always skipped ifcontrol reaches this point in the pattern; the idea of DEFINE is that it can beused to define subroutines that can be referenced from elsewhere. (The use of<a href="#groupsassubroutines">subroutines</a>is described below.) For example, a pattern to match an IPv4 address such as"192.168.23.245" could be written like this (ignore white space and linebreaks):<pre> (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) ) \b (?&byte) (\.(?&byte)){3} \b</pre>The first part of the pattern is a DEFINE group inside which another groupnamed "byte" is defined. This matches an individual component of an IPv4address (a number less than 256). When matching takes place, this part of thepattern is skipped because DEFINE acts like a false condition. The rest of thepattern uses references to the named group to match the four dot-separatedcomponents of an IPv4 address, insisting on a word boundary at each end.</P><br><b>Checking the PCRE2 version</b><br><P>Programs that link with a PCRE2 library can check the version by calling<b>pcre2_config()</b> with appropriate arguments. Users of applications that donot have access to the underlying code cannot do this. A special "condition"called VERSION exists to allow such users to discover which version of PCRE2they are dealing with by using this condition to match a string such as"yesno". VERSION must be followed either by "=" or ">=" and a version number.For example:<pre> (?(VERSION>=10.4)yes|no)</pre>This pattern matches "yes" if the PCRE2 version is greater or equal to 10.4, or"no" otherwise. The fractional part of the version number may not contain morethan two digits.</P><br><b>Assertion conditions</b><br><P>If the condition is not in any of the above formats, it must be a parenthesizedassertion. This may be a positive or negative lookahead or lookbehindassertion. However, it must be a traditional atomic assertion, not one of the<a href="#nonatomicassertions">non-atomic assertions.</a></P><P>Consider this pattern, again containing non-significant white space, and withthe two alternatives on the second line:<pre> (?(?=[^a-z]*[a-z]) \d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )</pre>The condition is a positive lookahead assertion that matches an optionalsequence of non-letters followed by a letter. In other words, it tests for thepresence of at least one letter in the subject. If a letter is found, thesubject is matched against the first alternative; otherwise it is matchedagainst the second. This pattern matches strings in one of the two formsdd-aaa-dd or dd-dd-dd, where aaa are letters and dd are digits.</P><P>When an assertion that is a condition contains capture groups, anycapturing that occurs in a matching branch is retained afterwards, for bothpositive and negative assertions, because matching always continues after theassertion, whether it succeeds or fails. (Compare non-conditional assertions,for which captures are retained only for positive assertions that succeed.)<a name="comments"></a></P><br><a name="SEC24" href="#TOC1">COMMENTS</a><br><P>There are two ways of including comments in patterns that are processed byPCRE2. In both cases, the start of the comment must not be in a characterclass, nor in the middle of any other sequence of related characters such as(?: or a group name or number. The characters that make up a comment playno part in the pattern matching.</P><P>The sequence (?# marks the start of a comment that continues up to the nextclosing parenthesis. Nested parentheses are not permitted. If thePCRE2_EXTENDED or PCRE2_EXTENDED_MORE option is set, an unescaped # characteralso introduces a comment, which in this case continues to immediately afterthe next newline character or character sequence in the pattern. Whichcharacters are interpreted as newlines is controlled by an option passed to thecompiling function or by a special sequence at the start of the pattern, asdescribed in the section entitled<a href="#newlines">"Newline conventions"</a>above. Note that the end of this type of comment is a literal newline sequencein the pattern; escape sequences that happen to represent a newline do notcount. For example, consider this pattern when PCRE2_EXTENDED is set, and thedefault newline convention (a single linefeed character) is in force:<pre> abc #comment \n still comment</pre>On encountering the # character, <b>pcre2_compile()</b> skips along, looking fora newline in the pattern. The sequence \n is still literal at this stage, soit does not terminate the comment. Only an actual character with the code value0x0a (the default newline) does so.<a name="recursion"></a></P><br><a name="SEC25" href="#TOC1">RECURSIVE PATTERNS</a><br><P>Consider the problem of matching a string in parentheses, allowing forunlimited nested parentheses. Without the use of recursion, the best that canbe done is to use a pattern that matches up to some fixed depth of nesting. Itis not possible to handle an arbitrary nesting depth.</P><P>For some time, Perl has provided a facility that allows regular expressions torecurse (amongst other things). It does this by interpolating Perl code in theexpression at run time, and the code can refer to the expression itself. A Perlpattern using code interpolation to solve the parentheses problem can becreated like this:<pre> $re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;</pre>The (?p{...}) item interpolates Perl code at run time, and in this case refersrecursively to the pattern in which it appears.</P><P>Obviously, PCRE2 cannot support the interpolation of Perl code. Instead, itsupports special syntax for recursion of the entire pattern, and also forindividual capture group recursion. After its introduction in PCRE1 and Python,this kind of recursion was subsequently introduced into Perl at release 5.10.</P><P>A special item that consists of (? followed by a number greater than zero and aclosing parenthesis is a recursive subroutine call of the capture group of thegiven number, provided that it occurs inside that group. (If not, it is a<a href="#groupsassubroutines">non-recursive subroutine</a>call, which is described in the next section.) The special item (?R) or (?0) isa recursive call of the entire regular expression.</P><P>This PCRE2 pattern solves the nested parentheses problem (assume thePCRE2_EXTENDED option is set so that white space is ignored):<pre> \( ( [^()]++ | (?R) )* \)</pre>First it matches an opening parenthesis. Then it matches any number ofsubstrings which can either be a sequence of non-parentheses, or a recursivematch of the pattern itself (that is, a correctly parenthesized substring).Finally there is a closing parenthesis. Note the use of a possessive quantifierto avoid backtracking into sequences of non-parentheses.</P><P>If this were part of a larger pattern, you would not want to recurse the entirepattern, so instead you could use this:<pre> ( \( ( [^()]++ | (?1) )* \) )</pre>We have put the pattern into parentheses, and caused the recursion to refer tothem instead of the whole pattern.</P><P>In a larger pattern, keeping track of parenthesis numbers can be tricky. Thisis made easier by the use of relative references. Instead of (?1) in thepattern above you can write (?-2) to refer to the second most recently openedparentheses preceding the recursion. In other words, a negative number countscapturing parentheses leftwards from the point at which it is encountered.</P><P>Be aware however, that if<a href="#dupgroupnumber">duplicate capture group numbers</a>are in use, relative references refer to the earliest group with theappropriate number. Consider, for example:<pre> (?|(a)|(b)) (c) (?-2)</pre>The first two capture groups (a) and (b) are both numbered 1, and group (c)is number 2. When the reference (?-2) is encountered, the second most recentlyopened parentheses has the number 1, but it is the first such group (the (a)group) to which the recursion refers. This would be the same if an absolutereference (?1) was used. In other words, relative references are just ashorthand for computing a group number.</P><P>It is also possible to refer to subsequent capture groups, by writingreferences such as (?+2). However, these cannot be recursive because thereference is not inside the parentheses that are referenced. They are always<a href="#groupsassubroutines">non-recursive subroutine</a>calls, as described in the next section.</P><P>An alternative approach is to use named parentheses. The Perl syntax for thisis (?&name); PCRE1's earlier syntax (?P>name) is also supported. We couldrewrite the above example as follows:<pre> (?<pn> \( ( [^()]++ | (?&pn) )* \) )</pre>If there is more than one group with the same name, the earliest one isused.</P><P>The example pattern that we have been looking at contains nested unlimitedrepeats, and so the use of a possessive quantifier for matching strings ofnon-parentheses is important when applying the pattern to strings that do notmatch. For example, when this pattern is applied to<pre> (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()</pre>it yields "no match" quickly. However, if a possessive quantifier is not used,the match runs for a very long time indeed because there are so many differentways the + and * repeats can carve up the subject, and all have to be testedbefore failure can be reported.</P><P>At the end of a match, the values of capturing parentheses are those fromthe outermost level. If you want to obtain intermediate values, a calloutfunction can be used (see below and the<a href="pcre2callout.html"><b>pcre2callout</b></a>documentation). If the pattern above is matched against<pre> (ab(cd)ef)</pre>the value for the inner capturing parentheses (numbered 2) is "ef", which isthe last value taken on at the top level. If a capture group is not matched atthe top level, its final captured value is unset, even if it was (temporarily)set at a deeper level during the matching process.</P><P>Do not confuse the (?R) item with the condition (R), which tests for recursion.Consider this pattern, which matches text in angle brackets, allowing forarbitrary nesting. Only digits are allowed in nested brackets (that is, whenrecursing), whereas any characters are permitted at the outer level.<pre> < (?: (?(R) \d++ | [^<>]*+) | (?R)) * ></pre>In this pattern, (?(R) is the start of a conditional group, with two differentalternatives for the recursive and non-recursive cases. The (?R) item is theactual recursive call.<a name="recursiondifference"></a></P><br><b>Differences in recursion processing between PCRE2 and Perl</b><br><P>Some former differences between PCRE2 and Perl no longer exist.</P><P>Before release 10.30, recursion processing in PCRE2 differed from Perl in thata recursive subroutine call was always treated as an atomic group. That is,once it had matched some of the subject string, it was never re-entered, evenif it contained untried alternatives and there was a subsequent matchingfailure. (Historical note: PCRE implemented recursion before Perl did.)</P><P>Starting with release 10.30, recursive subroutine calls are no longer treatedas atomic. That is, they can be re-entered to try unused alternatives if thereis a matching failure later in the pattern. This is now compatible with the wayPerl works. If you want a subroutine call to be atomic, you must explicitlyenclose it in an atomic group.</P><P>Supporting backtracking into recursions simplifies certain types of recursivepattern. For example, this pattern matches palindromic strings:<pre> ^((.)(?1)\2|.?)$</pre>The second branch in the group matches a single central character in thepalindrome when there are an odd number of characters, or nothing when thereare an even number of characters, but in order to work it has to be able to trythe second case when the rest of the pattern match fails. If you want to matchtypical palindromic phrases, the pattern has to ignore all non-word characters,which can be done like this:<pre> ^\W*+((.)\W*+(?1)\W*+\2|\W*+.?)\W*+$</pre>If run with the PCRE2_CASELESS option, this pattern matches phrases such as "Aman, a plan, a canal: Panama!". Note the use of the possessive quantifier *+ toavoid backtracking into sequences of non-word characters. Without this, PCRE2takes a great deal longer (ten times or more) to match typical phrases, andPerl takes so long that you think it has gone into a loop.</P><P>Another way in which PCRE2 and Perl used to differ in their recursionprocessing is in the handling of captured values. Formerly in Perl, when agroup was called recursively or as a subroutine (see the next section), ithad no access to any values that were captured outside the recursion, whereasin PCRE2 these values can be referenced. Consider this pattern:<pre> ^(.)(\1|a(?2))</pre>This pattern matches "bab". The first capturing parentheses match "b", then inthe second group, when the backreference \1 fails to match "b", the secondalternative matches "a" and then recurses. In the recursion, \1 does now match"b" and so the whole match succeeds. This match used to fail in Perl, but inlater versions (I tried 5.024) it now works.<a name="groupsassubroutines"></a></P><br><a name="SEC26" href="#TOC1">GROUPS AS SUBROUTINES</a><br><P>If the syntax for a recursive group call (either by number or by name) is usedoutside the parentheses to which it refers, it operates a bit like a subroutinein a programming language. More accurately, PCRE2 treats the referenced groupas an independent subpattern which it tries to match at the current matchingposition. The called group may be defined before or after the reference. Anumbered reference can be absolute or relative, as in these examples:<pre> (...(absolute)...)...(?2)... (...(relative)...)...(?-1)... (...(?+1)...(relative)...</pre>An earlier example pointed out that the pattern<pre> (sens|respons)e and \1ibility</pre>matches "sense and sensibility" and "response and responsibility", but not"sense and responsibility". If instead the pattern<pre> (sens|respons)e and (?1)ibility</pre>is used, it does match "sense and responsibility" as well as the other twostrings. Another example is given in the discussion of DEFINE above.</P><P>Like recursions, subroutine calls used to be treated as atomic, but thischanged at PCRE2 release 10.30, so backtracking into subroutine calls can nowoccur. However, any capturing parentheses that are set during the subroutinecall revert to their previous values afterwards.</P><P>Processing options such as case-independence are fixed when a group isdefined, so if it is used as a subroutine, such options cannot be changed fordifferent calls. For example, consider this pattern:<pre> (abc)(?i:(?-1))</pre>It matches "abcabc". It does not match "abcABC" because the change ofprocessing option does not affect the called group.</P><P>The behaviour of<a href="#backtrackcontrol">backtracking control verbs</a>in groups when called as subroutines is described in the section entitled<a href="#btsub">"Backtracking verbs in subroutines"</a>below.<a name="onigurumasubroutines"></a></P><br><a name="SEC27" href="#TOC1">ONIGURUMA SUBROUTINE SYNTAX</a><br><P>For compatibility with Oniguruma, the non-Perl syntax \g followed by a name ora number enclosed either in angle brackets or single quotes, is an alternativesyntax for calling a group as a subroutine, possibly recursively. Here are twoof the examples used above, rewritten using this syntax:<pre> (?<pn> \( ( (?>[^()]+) | \g<pn> )* \) ) (sens|respons)e and \g'1'ibility</pre>PCRE2 supports an extension to Oniguruma: if a number is preceded by aplus or a minus sign it is taken as a relative reference. For example:<pre> (abc)(?i:\g<-1>)</pre>Note that \g{...} (Perl syntax) and \g<...> (Oniguruma syntax) are <i>not</i>synonymous. The former is a backreference; the latter is a subroutine call.</P><br><a name="SEC28" href="#TOC1">CALLOUTS</a><br><P>Perl has a feature whereby using the sequence (?{...}) causes arbitrary Perlcode to be obeyed in the middle of matching a regular expression. This makes itpossible, amongst other things, to extract different substrings that match thesame pair of parentheses when there is a repetition.</P><P>PCRE2 provides a similar feature, but of course it cannot obey arbitrary Perlcode. The feature is called "callout". The caller of PCRE2 provides an externalfunction by putting its entry point in a match context using the function<b>pcre2_set_callout()</b>, and then passing that context to <b>pcre2_match()</b>or <b>pcre2_dfa_match()</b>. If no match context is passed, or if the calloutentry point is set to NULL, callouts are disabled.</P><P>Within a regular expression, (?C<arg>) indicates a point at which the externalfunction is to be called. There are two kinds of callout: those with anumerical argument and those with a string argument. (?C) on its own with noargument is treated as (?C0). A numerical argument allows the application todistinguish between different callouts. String arguments were added for release10.20 to make it possible for script languages that use PCRE2 to embed shortscripts within patterns in a similar way to Perl.</P><P>During matching, when PCRE2 reaches a callout point, the external function iscalled. It is provided with the number or string argument of the callout, theposition in the pattern, and one item of data that is also set in the matchblock. The callout function may cause matching to proceed, to backtrack, or tofail.</P><P>By default, PCRE2 implements a number of optimizations at matching time, andone side-effect is that sometimes callouts are skipped. If you need allpossible callouts to happen, you need to set options that disable the relevantoptimizations. More details, including a complete description of theprogramming interface to the callout function, are given in the<a href="pcre2callout.html"><b>pcre2callout</b></a>documentation.</P><br><b>Callouts with numerical arguments</b><br><P>If you just want to have a means of identifying different callout points, put anumber less than 256 after the letter C. For example, this pattern has twocallout points:<pre> (?C1)abc(?C2)def</pre>If the PCRE2_AUTO_CALLOUT flag is passed to <b>pcre2_compile()</b>, numericalcallouts are automatically installed before each item in the pattern. They areall numbered 255. If there is a conditional group in the pattern whosecondition is an assertion, an additional callout is inserted just before thecondition. An explicit callout may also be set at this position, as in thisexample:<pre> (?(?C9)(?=a)abc|def)</pre>Note that this applies only to assertion conditions, not to other types ofcondition.</P><br><b>Callouts with string arguments</b><br><P>A delimited string may be used instead of a number as a callout argument. Thestarting delimiter must be one of ` ' " ^ % # $ { and the ending delimiter isthe same as the start, except for {, where the ending delimiter is }. If theending delimiter is needed within the string, it must be doubled. Forexample:<pre> (?C'ab ''c'' d')xyz(?C{any text})pqr</pre>The doubling is removed before the string is passed to the callout function.<a name="backtrackcontrol"></a></P><br><a name="SEC29" href="#TOC1">BACKTRACKING CONTROL</a><br><P>There are a number of special "Backtracking Control Verbs" (to use Perl'sterminology) that modify the behaviour of backtracking during matching. Theyare generally of the form (*VERB) or (*VERB:NAME). Some verbs take either form,and may behave differently depending on whether or not a name argument ispresent. The names are not required to be unique within the pattern.</P><P>By default, for compatibility with Perl, a name is any sequence of charactersthat does not include a closing parenthesis. The name is not processed inany way, and it is not possible to include a closing parenthesis in the name.This can be changed by setting the PCRE2_ALT_VERBNAMES option, but the resultis no longer Perl-compatible.</P><P>When PCRE2_ALT_VERBNAMES is set, backslash processing is applied to verb namesand only an unescaped closing parenthesis terminates the name. However, theonly backslash items that are permitted are \Q, \E, and sequences such as\x{100} that define character code points. Character type escapes such as \dare faulted.</P><P>A closing parenthesis can be included in a name either as \) or between \Qand \E. In addition to backslash processing, if the PCRE2_EXTENDED orPCRE2_EXTENDED_MORE option is also set, unescaped whitespace in verb names isskipped, and #-comments are recognized, exactly as in the rest of the pattern.PCRE2_EXTENDED and PCRE2_EXTENDED_MORE do not affect verb names unlessPCRE2_ALT_VERBNAMES is also set.</P><P>The maximum length of a name is 255 in the 8-bit library and 65535 in the16-bit and 32-bit libraries. If the name is empty, that is, if the closingparenthesis immediately follows the colon, the effect is as if the colon werenot there. Any number of these verbs may occur in a pattern. Except for(*ACCEPT), they may not be quantified.</P><P>Since these verbs are specifically related to backtracking, most of them can beused only when the pattern is to be matched using the traditional matchingfunction, because that uses a backtracking algorithm. With the exception of(*FAIL), which behaves like a failing negative assertion, the backtrackingcontrol verbs cause an error if encountered by the DFA matching function.</P><P>The behaviour of these verbs in<a href="#btrepeat">repeated groups,</a><a href="#btassert">assertions,</a>and in<a href="#btsub">capture groups called as subroutines</a>(whether or not recursively) is documented below.<a name="nooptimize"></a></P><br><b>Optimizations that affect backtracking verbs</b><br><P>PCRE2 contains some optimizations that are used to speed up matching by runningsome checks at the start of each match attempt. For example, it may know theminimum length of matching subject, or that a particular character must bepresent. When one of these optimizations bypasses the running of a match, anyincluded backtracking verbs will not, of course, be processed. You can suppressthe start-of-match optimizations by setting the PCRE2_NO_START_OPTIMIZE optionwhen calling <b>pcre2_compile()</b>, or by starting the pattern with(*NO_START_OPT). There is more discussion of this option in the sectionentitled<a href="pcre2api.html#compiling">"Compiling a pattern"</a>in the<a href="pcre2api.html"><b>pcre2api</b></a>documentation.</P><P>Experiments with Perl suggest that it too has similar optimizations, and likePCRE2, turning them off can change the result of a match.<a name="acceptverb"></a></P><br><b>Verbs that act immediately</b><br><P>The following verbs act as soon as they are encountered.<pre> (*ACCEPT) or (*ACCEPT:NAME)</pre>This verb causes the match to end successfully, skipping the remainder of thepattern. However, when it is inside a capture group that is called as asubroutine, only that group is ended successfully. Matching then continuesat the outer level. If (*ACCEPT) in triggered in a positive assertion, theassertion succeeds; in a negative assertion, the assertion fails.</P><P>If (*ACCEPT) is inside capturing parentheses, the data so far is captured. Forexample:<pre> A((?:A|B(*ACCEPT)|C)D)</pre>This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is captured bythe outer parentheses.</P><P>(*ACCEPT) is the only backtracking verb that is allowed to be quantifiedbecause an ungreedy quantification with a minimum of zero acts only when abacktrack happens. Consider, for example,<pre> (A(*ACCEPT)??B)C</pre>where A, B, and C may be complex expressions. After matching "A", the matcherprocesses "BC"; if that fails, causing a backtrack, (*ACCEPT) is triggered andthe match succeeds. In both cases, all but C is captured. Whereas (*COMMIT)(see below) means "fail on backtrack", a repeated (*ACCEPT) of this type means"succeed on backtrack".</P><P><b>Warning:</b> (*ACCEPT) should not be used within a script run group, becauseit causes an immediate exit from the group, bypassing the script run checking.<pre> (*FAIL) or (*FAIL:NAME)</pre>This verb causes a matching failure, forcing backtracking to occur. It may beabbreviated to (*F). It is equivalent to (?!) but easier to read. The Perldocumentation notes that it is probably useful only when combined with (?{}) or(??{}). Those are, of course, Perl features that are not present in PCRE2. Thenearest equivalent is the callout feature, as for example in this pattern:<pre> a+(?C)(*FAIL)</pre>A match with the string "aaaa" always fails, but the callout is taken beforeeach backtrack happens (in this example, 10 times).</P><P>(*ACCEPT:NAME) and (*FAIL:NAME) behave the same as (*MARK:NAME)(*ACCEPT) and(*MARK:NAME)(*FAIL), respectively, that is, a (*MARK) is recorded just beforethe verb acts.</P><br><b>Recording which path was taken</b><br><P>There is one verb whose main purpose is to track how a match was arrived at,though it also has a secondary use in conjunction with advancing the matchstarting point (see (*SKIP) below).<pre> (*MARK:NAME) or (*:NAME)</pre>A name is always required with this verb. For all the other backtrackingcontrol verbs, a NAME argument is optional.</P><P>When a match succeeds, the name of the last-encountered mark name on thematching path is passed back to the caller as described in the section entitled<a href="pcre2api.html#matchotherdata">"Other information about the match"</a>in the<a href="pcre2api.html"><b>pcre2api</b></a>documentation. This applies to all instances of (*MARK) and other verbs,including those inside assertions and atomic groups. However, there aredifferences in those cases when (*MARK) is used in conjunction with (*SKIP) asdescribed below.</P><P>The mark name that was last encountered on the matching path is passed back. Averb without a NAME argument is ignored for this purpose. Here is an example of<b>pcre2test</b> output, where the "mark" modifier requests the retrieval andoutputting of (*MARK) data:<pre> re> /X(*MARK:A)Y|X(*MARK:B)Z/mark data> XY 0: XY MK: A XZ 0: XZ MK: B</pre>The (*MARK) name is tagged with "MK:" in this output, and in this example itindicates which of the two alternatives matched. This is a more efficient wayof obtaining this information than putting each alternative in its owncapturing parentheses.</P><P>If a verb with a name is encountered in a positive assertion that is true, thename is recorded and passed back if it is the last-encountered. This does nothappen for negative assertions or failing positive assertions.</P><P>After a partial match or a failed match, the last encountered name in theentire match process is returned. For example:<pre> re> /X(*MARK:A)Y|X(*MARK:B)Z/mark data> XP No match, mark = B</pre>Note that in this unanchored example the mark is retained from the matchattempt that started at the letter "X" in the subject. Subsequent matchattempts starting at "P" and then with an empty string do not get as far as the(*MARK) item, but nevertheless do not reset it.</P><P>If you are interested in (*MARK) values after failed matches, you shouldprobably set the PCRE2_NO_START_OPTIMIZE option<a href="#nooptimize">(see above)</a>to ensure that the match is always attempted.</P><br><b>Verbs that act after backtracking</b><br><P>The following verbs do nothing when they are encountered. Matching continueswith what follows, but if there is a subsequent match failure, causing abacktrack to the verb, a failure is forced. That is, backtracking cannot passto the left of the verb. However, when one of these verbs appears inside anatomic group or in a lookaround assertion that is true, its effect is confinedto that group, because once the group has been matched, there is never anybacktracking into it. Backtracking from beyond an assertion or an atomic groupignores the entire group, and seeks a preceding backtracking point.</P><P>These verbs differ in exactly what kind of failure occurs when backtrackingreaches them. The behaviour described below is what happens when the verb isnot in a subroutine or an assertion. Subsequent sections cover these specialcases.<pre> (*COMMIT) or (*COMMIT:NAME)</pre>This verb causes the whole match to fail outright if there is a later matchingfailure that causes backtracking to reach it. Even if the pattern isunanchored, no further attempts to find a match by advancing the starting pointtake place. If (*COMMIT) is the only backtracking verb that is encountered,once it has been passed <b>pcre2_match()</b> is committed to finding a match atthe current starting point, or not at all. For example:<pre> a+(*COMMIT)b</pre>This matches "xxaab" but not "aacaab". It can be thought of as a kind ofdynamic anchor, or "I've started, so I must finish."</P><P>The behaviour of (*COMMIT:NAME) is not the same as (*MARK:NAME)(*COMMIT). It islike (*MARK:NAME) in that the name is remembered for passing back to thecaller. However, (*SKIP:NAME) searches only for names that are set with(*MARK), ignoring those set by any of the other backtracking verbs.</P><P>If there is more than one backtracking verb in a pattern, a different one thatfollows (*COMMIT) may be triggered first, so merely passing (*COMMIT) during amatch does not always guarantee that a match must be at this starting point.</P><P>Note that (*COMMIT) at the start of a pattern is not the same as an anchor,unless PCRE2's start-of-match optimizations are turned off, as shown in thisoutput from <b>pcre2test</b>:<pre> re> /(*COMMIT)abc/ data> xyzabc 0: abc data> re> /(*COMMIT)abc/no_start_optimize data> xyzabc No match</pre>For the first pattern, PCRE2 knows that any match must start with "a", so theoptimization skips along the subject to "a" before applying the pattern to thefirst set of data. The match attempt then succeeds. The second pattern disablesthe optimization that skips along to the first character. The pattern is nowapplied starting at "x", and so the (*COMMIT) causes the match to fail withouttrying any other starting points.<pre> (*PRUNE) or (*PRUNE:NAME)</pre>This verb causes the match to fail at the current starting position in thesubject if there is a later matching failure that causes backtracking to reachit. If the pattern is unanchored, the normal "bumpalong" advance to the nextstarting character then happens. Backtracking can occur as usual to the left of(*PRUNE), before it is reached, or when matching to the right of (*PRUNE), butif there is no match to the right, backtracking cannot cross (*PRUNE). Insimple cases, the use of (*PRUNE) is just an alternative to an atomic group orpossessive quantifier, but there are some uses of (*PRUNE) that cannot beexpressed in any other way. In an anchored pattern (*PRUNE) has the same effectas (*COMMIT).</P><P>The behaviour of (*PRUNE:NAME) is not the same as (*MARK:NAME)(*PRUNE). It islike (*MARK:NAME) in that the name is remembered for passing back to thecaller. However, (*SKIP:NAME) searches only for names set with (*MARK),ignoring those set by other backtracking verbs.<pre> (*SKIP)</pre>This verb, when given without a name, is like (*PRUNE), except that if thepattern is unanchored, the "bumpalong" advance is not to the next character,but to the position in the subject where (*SKIP) was encountered. (*SKIP)signifies that whatever text was matched leading up to it cannot be part of asuccessful match if there is a later mismatch. Consider:<pre> a+(*SKIP)b</pre>If the subject is "aaaac...", after the first match attempt fails (starting atthe first character in the string), the starting point skips on to start thenext attempt at "c". Note that a possessive quantifier does not have the sameeffect as this example; although it would suppress backtracking during thefirst match attempt, the second attempt would start at the second characterinstead of skipping on to "c".</P><P>If (*SKIP) is used to specify a new starting position that is the same as thestarting position of the current match, or (by being inside a lookbehind)earlier, the position specified by (*SKIP) is ignored, and instead the normal"bumpalong" occurs.<pre> (*SKIP:NAME)</pre>When (*SKIP) has an associated name, its behaviour is modified. When such a(*SKIP) is triggered, the previous path through the pattern is searched for themost recent (*MARK) that has the same name. If one is found, the "bumpalong"advance is to the subject position that corresponds to that (*MARK) instead ofto where (*SKIP) was encountered. If no (*MARK) with a matching name is found,the (*SKIP) is ignored.</P><P>The search for a (*MARK) name uses the normal backtracking mechanism, whichmeans that it does not see (*MARK) settings that are inside atomic groups orassertions, because they are never re-entered by backtracking. Compare thefollowing <b>pcre2test</b> examples:<pre> re> /a(?>(*MARK:X))(*SKIP:X)(*F)|(.)/ data: abc 0: a 1: a data: re> /a(?:(*MARK:X))(*SKIP:X)(*F)|(.)/ data: abc 0: b 1: b</pre>In the first example, the (*MARK) setting is in an atomic group, so it is notseen when (*SKIP:X) triggers, causing the (*SKIP) to be ignored. This allowsthe second branch of the pattern to be tried at the first character position.In the second example, the (*MARK) setting is not in an atomic group. Thisallows (*SKIP:X) to find the (*MARK) when it backtracks, and this causes a newmatching attempt to start at the second character. This time, the (*MARK) isnever seen because "a" does not match "b", so the matcher immediately jumps tothe second branch of the pattern.</P><P>Note that (*SKIP:NAME) searches only for names set by (*MARK:NAME). It ignoresnames that are set by other backtracking verbs.<pre> (*THEN) or (*THEN:NAME)</pre>This verb causes a skip to the next innermost alternative when backtrackingreaches it. That is, it cancels any further backtracking within the currentalternative. Its name comes from the observation that it can be used for apattern-based if-then-else block:<pre> ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...</pre>If the COND1 pattern matches, FOO is tried (and possibly further items afterthe end of the group if FOO succeeds); on failure, the matcher skips to thesecond alternative and tries COND2, without backtracking into COND1. If thatsucceeds and BAR fails, COND3 is tried. If subsequently BAZ fails, there are nomore alternatives, so there is a backtrack to whatever came before the entiregroup. If (*THEN) is not inside an alternation, it acts like (*PRUNE).</P><P>The behaviour of (*THEN:NAME) is not the same as (*MARK:NAME)(*THEN). It islike (*MARK:NAME) in that the name is remembered for passing back to thecaller. However, (*SKIP:NAME) searches only for names set with (*MARK),ignoring those set by other backtracking verbs.</P><P>A group that does not contain a | character is just a part of the enclosingalternative; it is not a nested alternation with only one alternative. Theeffect of (*THEN) extends beyond such a group to the enclosing alternative.Consider this pattern, where A, B, etc. are complex pattern fragments that donot contain any | characters at this level:<pre> A (B(*THEN)C) | D</pre>If A and B are matched, but there is a failure in C, matching does notbacktrack into A; instead it moves to the next alternative, that is, D.However, if the group containing (*THEN) is given an alternative, itbehaves differently:<pre> A (B(*THEN)C | (*FAIL)) | D</pre>The effect of (*THEN) is now confined to the inner group. After a failure in C,matching moves to (*FAIL), which causes the whole group to fail because thereare no more alternatives to try. In this case, matching does backtrack into A.</P><P>Note that a conditional group is not considered as having two alternatives,because only one is ever used. In other words, the | character in a conditionalgroup has a different meaning. Ignoring white space, consider:<pre> ^.*? (?(?=a) a | b(*THEN)c )</pre>If the subject is "ba", this pattern does not match. Because .*? is ungreedy,it initially matches zero characters. The condition (?=a) then fails, thecharacter "b" is matched, but "c" is not. At this point, matching does notbacktrack to .*? as might perhaps be expected from the presence of the |character. The conditional group is part of the single alternative thatcomprises the whole pattern, and so the match fails. (If there was a backtrackinto .*?, allowing it to match "b", the match would succeed.)</P><P>The verbs just described provide four different "strengths" of control whensubsequent matching fails. (*THEN) is the weakest, carrying on the match at thenext alternative. (*PRUNE) comes next, failing the match at the currentstarting position, but allowing an advance to the next character (for anunanchored pattern). (*SKIP) is similar, except that the advance may be morethan one character. (*COMMIT) is the strongest, causing the entire match tofail.</P><br><b>More than one backtracking verb</b><br><P>If more than one backtracking verb is present in a pattern, the one that isbacktracked onto first acts. For example, consider this pattern, where A, B,etc. are complex pattern fragments:<pre> (A(*COMMIT)B(*THEN)C|ABD)</pre>If A matches but B fails, the backtrack to (*COMMIT) causes the entire match tofail. However, if A and B match, but C fails, the backtrack to (*THEN) causesthe next alternative (ABD) to be tried. This behaviour is consistent, but isnot always the same as Perl's. It means that if two or more backtracking verbsappear in succession, all but the last of them has no effect. Consider thisexample:<pre> ...(*COMMIT)(*PRUNE)...</pre>If there is a matching failure to the right, backtracking onto (*PRUNE) causesit to be triggered, and its action is taken. There can never be a backtrackonto (*COMMIT).<a name="btrepeat"></a></P><br><b>Backtracking verbs in repeated groups</b><br><P>PCRE2 sometimes differs from Perl in its handling of backtracking verbs inrepeated groups. For example, consider:<pre> /(a(*COMMIT)b)+ac/</pre>If the subject is "abac", Perl matches unless its optimizations are disabled,but PCRE2 always fails because the (*COMMIT) in the second repeat of the groupacts.<a name="btassert"></a></P><br><b>Backtracking verbs in assertions</b><br><P>(*FAIL) in any assertion has its normal effect: it forces an immediatebacktrack. The behaviour of the other backtracking verbs depends on whether ornot the assertion is standalone or acting as the condition in a conditionalgroup.</P><P>(*ACCEPT) in a standalone positive assertion causes the assertion to succeedwithout any further processing; captured strings and a mark name (if set) areretained. In a standalone negative assertion, (*ACCEPT) causes the assertion tofail without any further processing; captured substrings and any mark name arediscarded.</P><P>If the assertion is a condition, (*ACCEPT) causes the condition to be true fora positive assertion and false for a negative one; captured substrings areretained in both cases.</P><P>The remaining verbs act only when a later failure causes a backtrack toreach them. This means that, for the Perl-compatible assertions, their effectis confined to the assertion, because Perl lookaround assertions are atomic. Abacktrack that occurs after such an assertion is complete does not jump backinto the assertion. Note in particular that a (*MARK) name that is set in anassertion is not "seen" by an instance of (*SKIP:NAME) later in the pattern.</P><P>PCRE2 now supports non-atomic positive assertions, as described in the sectionentitled<a href="#nonatomicassertions">"Non-atomic assertions"</a>above. These assertions must be standalone (not used as conditions). They arenot Perl-compatible. For these assertions, a later backtrack does jump backinto the assertion, and therefore verbs such as (*COMMIT) can be triggered bybacktracks from later in the pattern.</P><P>The effect of (*THEN) is not allowed to escape beyond an assertion. If thereare no more branches to try, (*THEN) causes a positive assertion to be false,and a negative assertion to be true.</P><P>The other backtracking verbs are not treated specially if they appear in astandalone positive assertion. In a conditional positive assertion,backtracking (from within the assertion) into (*COMMIT), (*SKIP), or (*PRUNE)causes the condition to be false. However, for both standalone and conditionalnegative assertions, backtracking into (*COMMIT), (*SKIP), or (*PRUNE) causesthe assertion to be true, without considering any further alternative branches.<a name="btsub"></a></P><br><b>Backtracking verbs in subroutines</b><br><P>These behaviours occur whether or not the group is called recursively.</P><P>(*ACCEPT) in a group called as a subroutine causes the subroutine match tosucceed without any further processing. Matching then continues after thesubroutine call. Perl documents this behaviour. Perl's treatment of the otherverbs in subroutines is different in some cases.</P><P>(*FAIL) in a group called as a subroutine has its normal effect: it forcesan immediate backtrack.</P><P>(*COMMIT), (*SKIP), and (*PRUNE) cause the subroutine match to fail whentriggered by being backtracked to in a group called as a subroutine. There isthen a backtrack at the outer level.</P><P>(*THEN), when triggered, skips to the next alternative in the innermostenclosing group that has alternatives (its normal behaviour). However, if thereis no such group within the subroutine's group, the subroutine match fails andthere is a backtrack at the outer level.</P><br><a name="SEC30" href="#TOC1">SEE ALSO</a><br><P><b>pcre2api</b>(3), <b>pcre2callout</b>(3), <b>pcre2matching</b>(3),<b>pcre2syntax</b>(3), <b>pcre2</b>(3).</P><br><a name="SEC31" href="#TOC1">AUTHOR</a><br><P>Philip Hazel<br>Retired from University Computing Service<br>Cambridge, England.<br></P><br><a name="SEC32" href="#TOC1">REVISION</a><br><P>Last updated: 04 June 2024<br>Copyright © 1997-2024 University of Cambridge.<br><p>Return to the <a href="index.html">PCRE2 index page</a>.</p>
|
