/[pcre]/code/trunk/doc/pcre.txt
ViewVC logotype

Diff of /code/trunk/doc/pcre.txt

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 506 by ph10, Mon Jan 11 15:29:42 2010 UTC revision 507 by ph10, Wed Mar 10 16:08:01 2010 UTC
# Line 29  INTRODUCTION Line 29  INTRODUCTION
29         5.10, including support for UTF-8 encoded strings and  Unicode  general         5.10, including support for UTF-8 encoded strings and  Unicode  general
30         category  properties.  However,  UTF-8  and  Unicode  support has to be         category  properties.  However,  UTF-8  and  Unicode  support has to be
31         explicitly enabled; it is not the default. The  Unicode  tables  corre-         explicitly enabled; it is not the default. The  Unicode  tables  corre-
32         spond to Unicode release 5.1.         spond to Unicode release 5.2.0.
33    
34         In  addition to the Perl-compatible matching function, PCRE contains an         In  addition to the Perl-compatible matching function, PCRE contains an
35         alternative function that matches the same compiled patterns in a  dif-         alternative function that matches the same compiled patterns in a  dif-
# Line 263  AUTHOR Line 263  AUTHOR
263    
264  REVISION  REVISION
265    
266         Last updated: 28 September 2009         Last updated: 01 March 2010
267         Copyright (c) 1997-2009 University of Cambridge.         Copyright (c) 1997-2010 University of Cambridge.
268  ------------------------------------------------------------------------------  ------------------------------------------------------------------------------
269    
270    
# Line 3488  BACKSLASH Line 3488  BACKSLASH
3488         Those that are not part of an identified script are lumped together  as         Those that are not part of an identified script are lumped together  as
3489         "Common". The current list of scripts is:         "Common". The current list of scripts is:
3490    
3491         Arabic,  Armenian,  Balinese,  Bengali,  Bopomofo,  Braille,  Buginese,         Arabic, Armenian, Avestan, Balinese, Bamum, Bengali, Bopomofo, Braille,
3492         Buhid,  Canadian_Aboriginal,  Cherokee,  Common,   Coptic,   Cuneiform,         Buginese, Buhid, Canadian_Aboriginal, Carian, Cham,  Cherokee,  Common,
3493         Cypriot, Cyrillic, Deseret, Devanagari, Ethiopic, Georgian, Glagolitic,         Coptic,   Cuneiform,  Cypriot,  Cyrillic,  Deseret,  Devanagari,  Egyp-
3494         Gothic, Greek, Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew,  Hira-         tian_Hieroglyphs,  Ethiopic,  Georgian,  Glagolitic,   Gothic,   Greek,
3495         gana,  Inherited,  Kannada,  Katakana,  Kharoshthi,  Khmer, Lao, Latin,         Gujarati,  Gurmukhi,  Han,  Hangul,  Hanunoo,  Hebrew,  Hiragana, Impe-
3496         Limbu,  Linear_B,  Malayalam,  Mongolian,  Myanmar,  New_Tai_Lue,  Nko,         rial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscriptional_Parthian,
3497         Ogham,  Old_Italic,  Old_Persian, Oriya, Osmanya, Phags_Pa, Phoenician,         Javanese,  Kaithi, Kannada, Katakana, Kayah_Li, Kharoshthi, Khmer, Lao,
3498         Runic,  Shavian,  Sinhala,  Syloti_Nagri,  Syriac,  Tagalog,  Tagbanwa,         Latin,  Lepcha,  Limbu,  Linear_B,  Lisu,  Lycian,  Lydian,  Malayalam,
3499         Tai_Le, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh, Ugaritic, Yi.         Meetei_Mayek,  Mongolian, Myanmar, New_Tai_Lue, Nko, Ogham, Old_Italic,
3500           Old_Persian, Old_South_Arabian, Old_Turkic, Ol_Chiki,  Oriya,  Osmanya,
3501           Phags_Pa,  Phoenician,  Rejang,  Runic, Samaritan, Saurashtra, Shavian,
3502           Sinhala, Sundanese, Syloti_Nagri, Syriac,  Tagalog,  Tagbanwa,  Tai_Le,
3503           Tai_Tham,  Tai_Viet,  Tamil,  Telugu,  Thaana, Thai, Tibetan, Tifinagh,
3504           Ugaritic, Vai, Yi.
3505    
3506         Each  character has exactly one general category property, specified by         Each character has exactly one general category property, specified  by
3507         a two-letter abbreviation. For compatibility with Perl, negation can be         a two-letter abbreviation. For compatibility with Perl, negation can be
3508         specified  by  including a circumflex between the opening brace and the         specified by including a circumflex between the opening brace  and  the
3509         property name. For example, \p{^Lu} is the same as \P{Lu}.         property name. For example, \p{^Lu} is the same as \P{Lu}.
3510    
3511         If only one letter is specified with \p or \P, it includes all the gen-         If only one letter is specified with \p or \P, it includes all the gen-
3512         eral  category properties that start with that letter. In this case, in         eral category properties that start with that letter. In this case,  in
3513         the absence of negation, the curly brackets in the escape sequence  are         the  absence of negation, the curly brackets in the escape sequence are
3514         optional; these two examples have the same effect:         optional; these two examples have the same effect:
3515    
3516           \p{L}           \p{L}
# Line 3557  BACKSLASH Line 3562  BACKSLASH
3562           Zp    Paragraph separator           Zp    Paragraph separator
3563           Zs    Space separator           Zs    Space separator
3564    
3565         The  special property L& is also supported: it matches a character that         The special property L& is also supported: it matches a character  that
3566         has the Lu, Ll, or Lt property, in other words, a letter  that  is  not         has  the  Lu,  Ll, or Lt property, in other words, a letter that is not
3567         classified as a modifier or "other".         classified as a modifier or "other".
3568    
3569         The  Cs  (Surrogate)  property  applies only to characters in the range         The Cs (Surrogate) property applies only to  characters  in  the  range
3570         U+D800 to U+DFFF. Such characters are not valid in UTF-8  strings  (see         U+D800  to  U+DFFF. Such characters are not valid in UTF-8 strings (see
3571         RFC 3629) and so cannot be tested by PCRE, unless UTF-8 validity check-         RFC 3629) and so cannot be tested by PCRE, unless UTF-8 validity check-
3572         ing has been turned off (see the discussion  of  PCRE_NO_UTF8_CHECK  in         ing  has  been  turned off (see the discussion of PCRE_NO_UTF8_CHECK in
3573         the pcreapi page). Perl does not support the Cs property.         the pcreapi page). Perl does not support the Cs property.
3574    
3575         The  long  synonyms  for  property  names  that  Perl supports (such as         The long synonyms for  property  names  that  Perl  supports  (such  as
3576         \p{Letter}) are not supported by PCRE, nor is it  permitted  to  prefix         \p{Letter})  are  not  supported by PCRE, nor is it permitted to prefix
3577         any of these properties with "Is".         any of these properties with "Is".
3578    
3579         No character that is in the Unicode table has the Cn (unassigned) prop-         No character that is in the Unicode table has the Cn (unassigned) prop-
3580         erty.  Instead, this property is assumed for any code point that is not         erty.  Instead, this property is assumed for any code point that is not
3581         in the Unicode table.         in the Unicode table.
3582    
3583         Specifying  caseless  matching  does not affect these escape sequences.         Specifying caseless matching does not affect  these  escape  sequences.
3584         For example, \p{Lu} always matches only upper case letters.         For example, \p{Lu} always matches only upper case letters.
3585    
3586         The \X escape matches any number of Unicode  characters  that  form  an         The  \X  escape  matches  any number of Unicode characters that form an
3587         extended Unicode sequence. \X is equivalent to         extended Unicode sequence. \X is equivalent to
3588    
3589           (?>\PM\pM*)           (?>\PM\pM*)
3590    
3591         That  is,  it matches a character without the "mark" property, followed         That is, it matches a character without the "mark"  property,  followed
3592         by zero or more characters with the "mark"  property,  and  treats  the         by  zero  or  more  characters with the "mark" property, and treats the
3593         sequence  as  an  atomic group (see below).  Characters with the "mark"         sequence as an atomic group (see below).  Characters  with  the  "mark"
3594         property are typically accents that  affect  the  preceding  character.         property  are  typically  accents  that affect the preceding character.
3595         None  of  them  have  codepoints less than 256, so in non-UTF-8 mode \X         None of them have codepoints less than 256, so  in  non-UTF-8  mode  \X
3596         matches any one character.         matches any one character.
3597    
3598         Matching characters by Unicode property is not fast, because  PCRE  has         Matching  characters  by Unicode property is not fast, because PCRE has
3599         to  search  a  structure  that  contains data for over fifteen thousand         to search a structure that contains  data  for  over  fifteen  thousand
3600         characters. That is why the traditional escape sequences such as \d and         characters. That is why the traditional escape sequences such as \d and
3601         \w do not use Unicode properties in PCRE.         \w do not use Unicode properties in PCRE.
3602    
3603     Resetting the match start     Resetting the match start
3604    
3605         The escape sequence \K, which is a Perl 5.10 feature, causes any previ-         The escape sequence \K, which is a Perl 5.10 feature, causes any previ-
3606         ously matched characters not  to  be  included  in  the  final  matched         ously  matched  characters  not  to  be  included  in the final matched
3607         sequence. For example, the pattern:         sequence. For example, the pattern:
3608    
3609           foo\Kbar           foo\Kbar
3610    
3611         matches  "foobar",  but reports that it has matched "bar". This feature         matches "foobar", but reports that it has matched "bar".  This  feature
3612         is similar to a lookbehind assertion (described  below).   However,  in         is  similar  to  a lookbehind assertion (described below).  However, in
3613         this  case, the part of the subject before the real match does not have         this case, the part of the subject before the real match does not  have
3614         to be of fixed length, as lookbehind assertions do. The use of \K  does         to  be of fixed length, as lookbehind assertions do. The use of \K does
3615         not  interfere  with  the setting of captured substrings.  For example,         not interfere with the setting of captured  substrings.   For  example,
3616         when the pattern         when the pattern
3617    
3618           (foo)\Kbar           (foo)\Kbar
3619    
3620         matches "foobar", the first substring is still set to "foo".         matches "foobar", the first substring is still set to "foo".
3621    
3622           Perl  documents  that  the  use  of  \K  within assertions is "not well
3623           defined". In PCRE, \K is acted upon  when  it  occurs  inside  positive
3624           assertions, but is ignored in negative assertions.
3625    
3626     Simple assertions     Simple assertions
3627    
3628         The final use of backslash is for certain simple assertions. An  asser-         The  final use of backslash is for certain simple assertions. An asser-
3629         tion  specifies a condition that has to be met at a particular point in         tion specifies a condition that has to be met at a particular point  in
3630         a match, without consuming any characters from the subject string.  The         a  match, without consuming any characters from the subject string. The
3631         use  of subpatterns for more complicated assertions is described below.         use of subpatterns for more complicated assertions is described  below.
3632         The backslashed assertions are:         The backslashed assertions are:
3633    
3634           \b     matches at a word boundary           \b     matches at a word boundary
# Line 3630  BACKSLASH Line 3639  BACKSLASH
3639           \z     matches only at the end of the subject           \z     matches only at the end of the subject
3640           \G     matches at the first matching position in the subject           \G     matches at the first matching position in the subject
3641    
3642         These assertions may not appear in character classes (but note that  \b         These  assertions may not appear in character classes (but note that \b
3643         has a different meaning, namely the backspace character, inside a char-         has a different meaning, namely the backspace character, inside a char-
3644         acter class).         acter class).
3645    
3646         A word boundary is a position in the subject string where  the  current         A  word  boundary is a position in the subject string where the current
3647         character  and  the previous character do not both match \w or \W (i.e.         character and the previous character do not both match \w or  \W  (i.e.
3648         one matches \w and the other matches \W), or the start or  end  of  the         one  matches  \w  and the other matches \W), or the start or end of the
3649         string if the first or last character matches \w, respectively. Neither         string if the first or last character matches \w, respectively. Neither
3650         PCRE nor Perl has a separte "start of word" or "end  of  word"  metase-         PCRE  nor  Perl  has a separte "start of word" or "end of word" metase-
3651         quence.  However,  whatever follows \b normally determines which it is.         quence. However, whatever follows \b normally determines which  it  is.
3652         For example, the fragment \ba matches "a" at the start of a word.         For example, the fragment \ba matches "a" at the start of a word.
3653    
3654         The \A, \Z, and \z assertions differ from  the  traditional  circumflex         The  \A,  \Z,  and \z assertions differ from the traditional circumflex
3655         and dollar (described in the next section) in that they only ever match         and dollar (described in the next section) in that they only ever match
3656         at the very start and end of the subject string, whatever  options  are         at  the  very start and end of the subject string, whatever options are
3657         set.  Thus,  they are independent of multiline mode. These three asser-         set. Thus, they are independent of multiline mode. These  three  asser-
3658         tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which         tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
3659         affect  only the behaviour of the circumflex and dollar metacharacters.         affect only the behaviour of the circumflex and dollar  metacharacters.
3660         However, if the startoffset argument of pcre_exec() is non-zero,  indi-         However,  if the startoffset argument of pcre_exec() is non-zero, indi-
3661         cating that matching is to start at a point other than the beginning of         cating that matching is to start at a point other than the beginning of
3662         the subject, \A can never match. The difference between \Z  and  \z  is         the  subject,  \A  can never match. The difference between \Z and \z is
3663         that \Z matches before a newline at the end of the string as well as at         that \Z matches before a newline at the end of the string as well as at
3664         the very end, whereas \z matches only at the end.         the very end, whereas \z matches only at the end.
3665    
3666         The \G assertion is true only when the current matching position is  at         The  \G assertion is true only when the current matching position is at
3667         the  start point of the match, as specified by the startoffset argument         the start point of the match, as specified by the startoffset  argument
3668         of pcre_exec(). It differs from \A when the  value  of  startoffset  is         of  pcre_exec().  It  differs  from \A when the value of startoffset is
3669         non-zero.  By calling pcre_exec() multiple times with appropriate argu-         non-zero. By calling pcre_exec() multiple times with appropriate  argu-
3670         ments, you can mimic Perl's /g option, and it is in this kind of imple-         ments, you can mimic Perl's /g option, and it is in this kind of imple-
3671         mentation where \G can be useful.         mentation where \G can be useful.
3672    
3673         Note,  however,  that  PCRE's interpretation of \G, as the start of the         Note, however, that PCRE's interpretation of \G, as the  start  of  the
3674         current match, is subtly different from Perl's, which defines it as the         current match, is subtly different from Perl's, which defines it as the
3675         end  of  the  previous  match. In Perl, these can be different when the         end of the previous match. In Perl, these can  be  different  when  the
3676         previously matched string was empty. Because PCRE does just  one  match         previously  matched  string was empty. Because PCRE does just one match
3677         at a time, it cannot reproduce this behaviour.         at a time, it cannot reproduce this behaviour.
3678    
3679         If  all  the alternatives of a pattern begin with \G, the expression is         If all the alternatives of a pattern begin with \G, the  expression  is
3680         anchored to the starting match position, and the "anchored" flag is set         anchored to the starting match position, and the "anchored" flag is set
3681         in the compiled regular expression.         in the compiled regular expression.
3682    
# Line 3675  BACKSLASH Line 3684  BACKSLASH
3684  CIRCUMFLEX AND DOLLAR  CIRCUMFLEX AND DOLLAR
3685    
3686         Outside a character class, in the default matching mode, the circumflex         Outside a character class, in the default matching mode, the circumflex
3687         character is an assertion that is true only  if  the  current  matching         character  is  an  assertion  that is true only if the current matching
3688         point  is  at the start of the subject string. If the startoffset argu-         point is at the start of the subject string. If the  startoffset  argu-
3689         ment of pcre_exec() is non-zero, circumflex  can  never  match  if  the         ment  of  pcre_exec()  is  non-zero,  circumflex can never match if the
3690         PCRE_MULTILINE  option  is  unset. Inside a character class, circumflex         PCRE_MULTILINE option is unset. Inside a  character  class,  circumflex
3691         has an entirely different meaning (see below).         has an entirely different meaning (see below).
3692    
3693         Circumflex need not be the first character of the pattern if  a  number         Circumflex  need  not be the first character of the pattern if a number
3694         of  alternatives are involved, but it should be the first thing in each         of alternatives are involved, but it should be the first thing in  each
3695         alternative in which it appears if the pattern is ever  to  match  that         alternative  in  which  it appears if the pattern is ever to match that
3696         branch.  If all possible alternatives start with a circumflex, that is,         branch. If all possible alternatives start with a circumflex, that  is,
3697         if the pattern is constrained to match only at the start  of  the  sub-         if  the  pattern  is constrained to match only at the start of the sub-
3698         ject,  it  is  said  to be an "anchored" pattern. (There are also other         ject, it is said to be an "anchored" pattern.  (There  are  also  other
3699         constructs that can cause a pattern to be anchored.)         constructs that can cause a pattern to be anchored.)
3700    
3701         A dollar character is an assertion that is true  only  if  the  current         A  dollar  character  is  an assertion that is true only if the current
3702         matching  point  is  at  the  end of the subject string, or immediately         matching point is at the end of  the  subject  string,  or  immediately
3703         before a newline at the end of the string (by default). Dollar need not         before a newline at the end of the string (by default). Dollar need not
3704         be  the  last  character of the pattern if a number of alternatives are         be the last character of the pattern if a number  of  alternatives  are
3705         involved, but it should be the last item in  any  branch  in  which  it         involved,  but  it  should  be  the last item in any branch in which it
3706         appears. Dollar has no special meaning in a character class.         appears. Dollar has no special meaning in a character class.
3707    
3708         The  meaning  of  dollar  can be changed so that it matches only at the         The meaning of dollar can be changed so that it  matches  only  at  the
3709         very end of the string, by setting the  PCRE_DOLLAR_ENDONLY  option  at         very  end  of  the string, by setting the PCRE_DOLLAR_ENDONLY option at
3710         compile time. This does not affect the \Z assertion.         compile time. This does not affect the \Z assertion.
3711    
3712         The meanings of the circumflex and dollar characters are changed if the         The meanings of the circumflex and dollar characters are changed if the
3713         PCRE_MULTILINE option is set. When  this  is  the  case,  a  circumflex         PCRE_MULTILINE  option  is  set.  When  this  is the case, a circumflex
3714         matches  immediately after internal newlines as well as at the start of         matches immediately after internal newlines as well as at the start  of
3715         the subject string. It does not match after a  newline  that  ends  the         the  subject  string.  It  does not match after a newline that ends the
3716         string.  A dollar matches before any newlines in the string, as well as         string. A dollar matches before any newlines in the string, as well  as
3717         at the very end, when PCRE_MULTILINE is set. When newline is  specified         at  the very end, when PCRE_MULTILINE is set. When newline is specified
3718         as  the  two-character  sequence CRLF, isolated CR and LF characters do         as the two-character sequence CRLF, isolated CR and  LF  characters  do
3719         not indicate newlines.         not indicate newlines.
3720    
3721         For example, the pattern /^abc$/ matches the subject string  "def\nabc"         For  example, the pattern /^abc$/ matches the subject string "def\nabc"
3722         (where  \n  represents a newline) in multiline mode, but not otherwise.         (where \n represents a newline) in multiline mode, but  not  otherwise.
3723         Consequently, patterns that are anchored in single  line  mode  because         Consequently,  patterns  that  are anchored in single line mode because
3724         all  branches  start  with  ^ are not anchored in multiline mode, and a         all branches start with ^ are not anchored in  multiline  mode,  and  a
3725         match for circumflex is  possible  when  the  startoffset  argument  of         match  for  circumflex  is  possible  when  the startoffset argument of
3726         pcre_exec()  is  non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if         pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is  ignored  if
3727         PCRE_MULTILINE is set.         PCRE_MULTILINE is set.
3728    
3729         Note that the sequences \A, \Z, and \z can be used to match  the  start         Note  that  the sequences \A, \Z, and \z can be used to match the start
3730         and  end of the subject in both modes, and if all branches of a pattern         and end of the subject in both modes, and if all branches of a  pattern
3731         start with \A it is always anchored, whether or not  PCRE_MULTILINE  is         start  with  \A it is always anchored, whether or not PCRE_MULTILINE is
3732         set.         set.
3733    
3734    
3735  FULL STOP (PERIOD, DOT)  FULL STOP (PERIOD, DOT)
3736    
3737         Outside a character class, a dot in the pattern matches any one charac-         Outside a character class, a dot in the pattern matches any one charac-
3738         ter in the subject string except (by default) a character  that  signi-         ter  in  the subject string except (by default) a character that signi-
3739         fies  the  end  of  a line. In UTF-8 mode, the matched character may be         fies the end of a line. In UTF-8 mode, the  matched  character  may  be
3740         more than one byte long.         more than one byte long.
3741    
3742         When a line ending is defined as a single character, dot never  matches         When  a line ending is defined as a single character, dot never matches
3743         that  character; when the two-character sequence CRLF is used, dot does         that character; when the two-character sequence CRLF is used, dot  does
3744         not match CR if it is immediately followed  by  LF,  but  otherwise  it         not  match  CR  if  it  is immediately followed by LF, but otherwise it
3745         matches  all characters (including isolated CRs and LFs). When any Uni-         matches all characters (including isolated CRs and LFs). When any  Uni-
3746         code line endings are being recognized, dot does not match CR or LF  or         code  line endings are being recognized, dot does not match CR or LF or
3747         any of the other line ending characters.         any of the other line ending characters.
3748    
3749         The  behaviour  of  dot  with regard to newlines can be changed. If the         The behaviour of dot with regard to newlines can  be  changed.  If  the
3750         PCRE_DOTALL option is set, a dot matches  any  one  character,  without         PCRE_DOTALL  option  is  set,  a dot matches any one character, without
3751         exception. If the two-character sequence CRLF is present in the subject         exception. If the two-character sequence CRLF is present in the subject
3752         string, it takes two dots to match it.         string, it takes two dots to match it.
3753    
3754         The handling of dot is entirely independent of the handling of  circum-         The  handling of dot is entirely independent of the handling of circum-
3755         flex  and  dollar,  the  only relationship being that they both involve         flex and dollar, the only relationship being  that  they  both  involve
3756         newlines. Dot has no special meaning in a character class.         newlines. Dot has no special meaning in a character class.
3757    
3758    
3759  MATCHING A SINGLE BYTE  MATCHING A SINGLE BYTE
3760    
3761         Outside a character class, the escape sequence \C matches any one byte,         Outside a character class, the escape sequence \C matches any one byte,
3762         both  in  and  out  of  UTF-8 mode. Unlike a dot, it always matches any         both in and out of UTF-8 mode. Unlike a  dot,  it  always  matches  any
3763         line-ending characters. The feature is provided in  Perl  in  order  to         line-ending  characters.  The  feature  is provided in Perl in order to
3764         match  individual bytes in UTF-8 mode. Because it breaks up UTF-8 char-         match individual bytes in UTF-8 mode. Because it breaks up UTF-8  char-
3765         acters into individual bytes, what remains in the string may be a  mal-         acters  into individual bytes, what remains in the string may be a mal-
3766         formed  UTF-8  string.  For this reason, the \C escape sequence is best         formed UTF-8 string. For this reason, the \C escape  sequence  is  best
3767         avoided.         avoided.
3768    
3769         PCRE does not allow \C to appear in  lookbehind  assertions  (described         PCRE  does  not  allow \C to appear in lookbehind assertions (described
3770         below),  because  in UTF-8 mode this would make it impossible to calcu-         below), because in UTF-8 mode this would make it impossible  to  calcu-
3771         late the length of the lookbehind.         late the length of the lookbehind.
3772    
3773    
# Line 3768  SQUARE BRACKETS AND CHARACTER CLASSES Line 3777  SQUARE BRACKETS AND CHARACTER CLASSES
3777         closing square bracket. A closing square bracket on its own is not spe-         closing square bracket. A closing square bracket on its own is not spe-
3778         cial by default.  However, if the PCRE_JAVASCRIPT_COMPAT option is set,         cial by default.  However, if the PCRE_JAVASCRIPT_COMPAT option is set,
3779         a lone closing square bracket causes a compile-time error. If a closing         a lone closing square bracket causes a compile-time error. If a closing
3780         square bracket is required as a member of the class, it should  be  the         square  bracket  is required as a member of the class, it should be the
3781         first  data  character  in  the  class (after an initial circumflex, if         first data character in the class  (after  an  initial  circumflex,  if
3782         present) or escaped with a backslash.         present) or escaped with a backslash.
3783    
3784         A character class matches a single character in the subject.  In  UTF-8         A  character  class matches a single character in the subject. In UTF-8
3785         mode, the character may be more than one byte long. A matched character         mode, the character may be more than one byte long. A matched character
3786         must be in the set of characters defined by the class, unless the first         must be in the set of characters defined by the class, unless the first
3787         character  in  the  class definition is a circumflex, in which case the         character in the class definition is a circumflex, in  which  case  the
3788         subject character must not be in the set defined by  the  class.  If  a         subject  character  must  not  be in the set defined by the class. If a
3789         circumflex  is actually required as a member of the class, ensure it is         circumflex is actually required as a member of the class, ensure it  is
3790         not the first character, or escape it with a backslash.         not the first character, or escape it with a backslash.
3791    
3792         For example, the character class [aeiou] matches any lower case  vowel,         For  example, the character class [aeiou] matches any lower case vowel,
3793         while  [^aeiou]  matches  any character that is not a lower case vowel.         while [^aeiou] matches any character that is not a  lower  case  vowel.
3794         Note that a circumflex is just a convenient notation for specifying the         Note that a circumflex is just a convenient notation for specifying the
3795         characters  that  are in the class by enumerating those that are not. A         characters that are in the class by enumerating those that are  not.  A
3796         class that starts with a circumflex is not an assertion; it still  con-         class  that starts with a circumflex is not an assertion; it still con-
3797         sumes  a  character  from the subject string, and therefore it fails if         sumes a character from the subject string, and therefore  it  fails  if
3798         the current pointer is at the end of the string.         the current pointer is at the end of the string.
3799    
3800         In UTF-8 mode, characters with values greater than 255 can be  included         In  UTF-8 mode, characters with values greater than 255 can be included
3801         in  a  class as a literal string of bytes, or by using the \x{ escaping         in a class as a literal string of bytes, or by using the  \x{  escaping
3802         mechanism.         mechanism.
3803    
3804         When caseless matching is set, any letters in a  class  represent  both         When  caseless  matching  is set, any letters in a class represent both
3805         their  upper  case  and lower case versions, so for example, a caseless         their upper case and lower case versions, so for  example,  a  caseless
3806         [aeiou] matches "A" as well as "a", and a caseless  [^aeiou]  does  not         [aeiou]  matches  "A"  as well as "a", and a caseless [^aeiou] does not
3807         match  "A", whereas a caseful version would. In UTF-8 mode, PCRE always         match "A", whereas a caseful version would. In UTF-8 mode, PCRE  always
3808         understands the concept of case for characters whose  values  are  less         understands  the  concept  of case for characters whose values are less
3809         than  128, so caseless matching is always possible. For characters with         than 128, so caseless matching is always possible. For characters  with
3810         higher values, the concept of case is supported  if  PCRE  is  compiled         higher  values,  the  concept  of case is supported if PCRE is compiled
3811         with  Unicode  property support, but not otherwise.  If you want to use         with Unicode property support, but not otherwise.  If you want  to  use
3812         caseless matching in UTF8-mode for characters 128 and above,  you  must         caseless  matching  in UTF8-mode for characters 128 and above, you must
3813         ensure  that  PCRE is compiled with Unicode property support as well as         ensure that PCRE is compiled with Unicode property support as  well  as
3814         with UTF-8 support.         with UTF-8 support.
3815    
3816         Characters that might indicate line breaks are  never  treated  in  any         Characters  that  might  indicate  line breaks are never treated in any
3817         special  way  when  matching  character  classes,  whatever line-ending         special way  when  matching  character  classes,  whatever  line-ending
3818         sequence is in  use,  and  whatever  setting  of  the  PCRE_DOTALL  and         sequence  is  in  use,  and  whatever  setting  of  the PCRE_DOTALL and
3819         PCRE_MULTILINE options is used. A class such as [^a] always matches one         PCRE_MULTILINE options is used. A class such as [^a] always matches one
3820         of these characters.         of these characters.
3821    
3822         The minus (hyphen) character can be used to specify a range of  charac-         The  minus (hyphen) character can be used to specify a range of charac-
3823         ters  in  a  character  class.  For  example,  [d-m] matches any letter         ters in a character  class.  For  example,  [d-m]  matches  any  letter
3824         between d and m, inclusive. If a  minus  character  is  required  in  a         between  d  and  m,  inclusive.  If  a minus character is required in a
3825         class,  it  must  be  escaped  with a backslash or appear in a position         class, it must be escaped with a backslash  or  appear  in  a  position
3826         where it cannot be interpreted as indicating a range, typically as  the         where  it cannot be interpreted as indicating a range, typically as the
3827         first or last character in the class.         first or last character in the class.
3828    
3829         It is not possible to have the literal character "]" as the end charac-         It is not possible to have the literal character "]" as the end charac-
3830         ter of a range. A pattern such as [W-]46] is interpreted as a class  of         ter  of a range. A pattern such as [W-]46] is interpreted as a class of
3831         two  characters ("W" and "-") followed by a literal string "46]", so it         two characters ("W" and "-") followed by a literal string "46]", so  it
3832         would match "W46]" or "-46]". However, if the "]"  is  escaped  with  a         would  match  "W46]"  or  "-46]". However, if the "]" is escaped with a
3833         backslash  it is interpreted as the end of range, so [W-\]46] is inter-         backslash it is interpreted as the end of range, so [W-\]46] is  inter-
3834         preted as a class containing a range followed by two other  characters.         preted  as a class containing a range followed by two other characters.
3835         The  octal or hexadecimal representation of "]" can also be used to end         The octal or hexadecimal representation of "]" can also be used to  end
3836         a range.         a range.
3837    
3838         Ranges operate in the collating sequence of character values. They  can         Ranges  operate in the collating sequence of character values. They can
3839         also   be  used  for  characters  specified  numerically,  for  example         also  be  used  for  characters  specified  numerically,  for   example
3840         [\000-\037]. In UTF-8 mode, ranges can include characters whose  values         [\000-\037].  In UTF-8 mode, ranges can include characters whose values
3841         are greater than 255, for example [\x{100}-\x{2ff}].         are greater than 255, for example [\x{100}-\x{2ff}].
3842    
3843         If a range that includes letters is used when caseless matching is set,         If a range that includes letters is used when caseless matching is set,
3844         it matches the letters in either case. For example, [W-c] is equivalent         it matches the letters in either case. For example, [W-c] is equivalent
3845         to  [][\\^_`wxyzabc],  matched  caselessly,  and  in non-UTF-8 mode, if         to [][\\^_`wxyzabc], matched caselessly,  and  in  non-UTF-8  mode,  if
3846         character tables for a French locale are in  use,  [\xc8-\xcb]  matches         character  tables  for  a French locale are in use, [\xc8-\xcb] matches
3847         accented  E  characters in both cases. In UTF-8 mode, PCRE supports the         accented E characters in both cases. In UTF-8 mode, PCRE  supports  the
3848         concept of case for characters with values greater than 128  only  when         concept  of  case for characters with values greater than 128 only when
3849         it is compiled with Unicode property support.         it is compiled with Unicode property support.
3850    
3851         The  character types \d, \D, \p, \P, \s, \S, \w, and \W may also appear         The character types \d, \D, \p, \P, \s, \S, \w, and \W may also  appear
3852         in a character class, and add the characters that  they  match  to  the         in  a  character  class,  and add the characters that they match to the
3853         class. For example, [\dABCDEF] matches any hexadecimal digit. A circum-         class. For example, [\dABCDEF] matches any hexadecimal digit. A circum-
3854         flex can conveniently be used with the upper case  character  types  to         flex  can  conveniently  be used with the upper case character types to
3855         specify  a  more  restricted  set of characters than the matching lower         specify a more restricted set of characters  than  the  matching  lower
3856         case type. For example, the class [^\W_] matches any letter  or  digit,         case  type.  For example, the class [^\W_] matches any letter or digit,
3857         but not underscore.         but not underscore.
3858    
3859         The  only  metacharacters  that are recognized in character classes are         The only metacharacters that are recognized in  character  classes  are
3860         backslash, hyphen (only where it can be  interpreted  as  specifying  a         backslash,  hyphen  (only  where  it can be interpreted as specifying a
3861         range),  circumflex  (only  at the start), opening square bracket (only         range), circumflex (only at the start), opening  square  bracket  (only
3862         when it can be interpreted as introducing a POSIX class name - see  the         when  it can be interpreted as introducing a POSIX class name - see the
3863         next  section),  and  the  terminating closing square bracket. However,         next section), and the terminating  closing  square  bracket.  However,
3864         escaping other non-alphanumeric characters does no harm.         escaping other non-alphanumeric characters does no harm.
3865    
3866    
3867  POSIX CHARACTER CLASSES  POSIX CHARACTER CLASSES
3868    
3869         Perl supports the POSIX notation for character classes. This uses names         Perl supports the POSIX notation for character classes. This uses names
3870         enclosed  by  [: and :] within the enclosing square brackets. PCRE also         enclosed by [: and :] within the enclosing square brackets.  PCRE  also
3871         supports this notation. For example,         supports this notation. For example,
3872    
3873           [01[:alpha:]%]           [01[:alpha:]%]
# Line 3881  POSIX CHARACTER CLASSES Line 3890  POSIX CHARACTER CLASSES
3890           word     "word" characters (same as \w)           word     "word" characters (same as \w)
3891           xdigit   hexadecimal digits           xdigit   hexadecimal digits
3892    
3893         The  "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),         The "space" characters are HT (9), LF (10), VT (11), FF (12), CR  (13),
3894         and space (32). Notice that this list includes the VT  character  (code         and  space  (32). Notice that this list includes the VT character (code
3895         11). This makes "space" different to \s, which does not include VT (for         11). This makes "space" different to \s, which does not include VT (for
3896         Perl compatibility).         Perl compatibility).
3897    
3898         The name "word" is a Perl extension, and "blank"  is  a  GNU  extension         The  name  "word"  is  a Perl extension, and "blank" is a GNU extension
3899         from  Perl  5.8. Another Perl extension is negation, which is indicated         from Perl 5.8. Another Perl extension is negation, which  is  indicated
3900         by a ^ character after the colon. For example,         by a ^ character after the colon. For example,
3901    
3902           [12[:^digit:]]           [12[:^digit:]]
3903    
3904         matches "1", "2", or any non-digit. PCRE (and Perl) also recognize  the         matches  "1", "2", or any non-digit. PCRE (and Perl) also recognize the
3905         POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but         POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
3906         these are not supported, and an error is given if they are encountered.         these are not supported, and an error is given if they are encountered.
3907    
# Line 3902  POSIX CHARACTER CLASSES Line 3911  POSIX CHARACTER CLASSES
3911    
3912  VERTICAL BAR  VERTICAL BAR
3913    
3914         Vertical  bar characters are used to separate alternative patterns. For         Vertical bar characters are used to separate alternative patterns.  For
3915         example, the pattern         example, the pattern
3916    
3917           gilbert|sullivan           gilbert|sullivan
3918    
3919         matches either "gilbert" or "sullivan". Any number of alternatives  may         matches  either "gilbert" or "sullivan". Any number of alternatives may
3920         appear,  and  an  empty  alternative  is  permitted (matching the empty         appear, and an empty  alternative  is  permitted  (matching  the  empty
3921         string). The matching process tries each alternative in turn, from left         string). The matching process tries each alternative in turn, from left
3922         to  right, and the first one that succeeds is used. If the alternatives         to right, and the first one that succeeds is used. If the  alternatives
3923         are within a subpattern (defined below), "succeeds" means matching  the         are  within a subpattern (defined below), "succeeds" means matching the
3924         rest of the main pattern as well as the alternative in the subpattern.         rest of the main pattern as well as the alternative in the subpattern.
3925    
3926    
3927  INTERNAL OPTION SETTING  INTERNAL OPTION SETTING
3928    
3929         The  settings  of  the  PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and         The settings of the  PCRE_CASELESS,  PCRE_MULTILINE,  PCRE_DOTALL,  and
3930         PCRE_EXTENDED options (which are Perl-compatible) can be  changed  from         PCRE_EXTENDED  options  (which are Perl-compatible) can be changed from
3931         within  the  pattern  by  a  sequence  of  Perl option letters enclosed         within the pattern by  a  sequence  of  Perl  option  letters  enclosed
3932         between "(?" and ")".  The option letters are         between "(?" and ")".  The option letters are
3933    
3934           i  for PCRE_CASELESS           i  for PCRE_CASELESS
# Line 3929  INTERNAL OPTION SETTING Line 3938  INTERNAL OPTION SETTING
3938    
3939         For example, (?im) sets caseless, multiline matching. It is also possi-         For example, (?im) sets caseless, multiline matching. It is also possi-
3940         ble to unset these options by preceding the letter with a hyphen, and a         ble to unset these options by preceding the letter with a hyphen, and a
3941         combined setting and unsetting such as (?im-sx), which sets  PCRE_CASE-         combined  setting and unsetting such as (?im-sx), which sets PCRE_CASE-
3942         LESS  and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,         LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and  PCRE_EXTENDED,
3943         is also permitted. If a  letter  appears  both  before  and  after  the         is  also  permitted.  If  a  letter  appears  both before and after the
3944         hyphen, the option is unset.         hyphen, the option is unset.
3945    
3946         The  PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA         The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and  PCRE_EXTRA
3947         can be changed in the same way as the Perl-compatible options by  using         can  be changed in the same way as the Perl-compatible options by using
3948         the characters J, U and X respectively.         the characters J, U and X respectively.
3949    
3950         When  one  of  these  option  changes occurs at top level (that is, not         When one of these option changes occurs at  top  level  (that  is,  not
3951         inside subpattern parentheses), the change applies to the remainder  of         inside  subpattern parentheses), the change applies to the remainder of
3952         the pattern that follows. If the change is placed right at the start of         the pattern that follows. If the change is placed right at the start of
3953         a pattern, PCRE extracts it into the global options (and it will there-         a pattern, PCRE extracts it into the global options (and it will there-
3954         fore show up in data extracted by the pcre_fullinfo() function).         fore show up in data extracted by the pcre_fullinfo() function).
3955    
3956         An  option  change  within a subpattern (see below for a description of         An option change within a subpattern (see below for  a  description  of
3957         subpatterns) affects only that part of the current pattern that follows         subpatterns) affects only that part of the current pattern that follows
3958         it, so         it, so
3959    
3960           (a(?i)b)c           (a(?i)b)c
3961    
3962         matches abc and aBc and no other strings (assuming PCRE_CASELESS is not         matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
3963         used).  By this means, options can be made to have  different  settings         used).   By  this means, options can be made to have different settings
3964         in  different parts of the pattern. Any changes made in one alternative         in different parts of the pattern. Any changes made in one  alternative
3965         do carry on into subsequent branches within the  same  subpattern.  For         do  carry  on  into subsequent branches within the same subpattern. For
3966         example,         example,
3967    
3968           (a(?i)b|c)           (a(?i)b|c)
3969    
3970         matches  "ab",  "aB",  "c",  and "C", even though when matching "C" the         matches "ab", "aB", "c", and "C", even though  when  matching  "C"  the
3971         first branch is abandoned before the option setting.  This  is  because         first  branch  is  abandoned before the option setting. This is because
3972         the  effects  of option settings happen at compile time. There would be         the effects of option settings happen at compile time. There  would  be
3973         some very weird behaviour otherwise.         some very weird behaviour otherwise.
3974    
3975         Note: There are other PCRE-specific options that  can  be  set  by  the         Note:  There  are  other  PCRE-specific  options that can be set by the
3976         application  when  the  compile  or match functions are called. In some         application when the compile or match functions  are  called.  In  some
3977         cases the pattern can contain special leading sequences such as (*CRLF)         cases the pattern can contain special leading sequences such as (*CRLF)
3978         to  override  what  the application has set or what has been defaulted.         to override what the application has set or what  has  been  defaulted.
3979         Details are given in the section entitled  "Newline  sequences"  above.         Details  are  given  in the section entitled "Newline sequences" above.
3980         There  is  also  the  (*UTF8)  leading sequence that can be used to set         There is also the (*UTF8) leading sequence that  can  be  used  to  set
3981         UTF-8 mode; this is equivalent to setting the PCRE_UTF8 option.         UTF-8 mode; this is equivalent to setting the PCRE_UTF8 option.
3982    
3983    
# Line 3981  SUBPATTERNS Line 3990  SUBPATTERNS
3990    
3991           cat(aract|erpillar|)           cat(aract|erpillar|)
3992    
3993         matches  one  of the words "cat", "cataract", or "caterpillar". Without         matches one of the words "cat", "cataract", or  "caterpillar".  Without
3994         the parentheses, it would match  "cataract",  "erpillar"  or  an  empty         the  parentheses,  it  would  match  "cataract", "erpillar" or an empty
3995         string.         string.
3996    
3997         2.  It  sets  up  the  subpattern as a capturing subpattern. This means         2. It sets up the subpattern as  a  capturing  subpattern.  This  means
3998         that, when the whole pattern  matches,  that  portion  of  the  subject         that,  when  the  whole  pattern  matches,  that portion of the subject
3999         string that matched the subpattern is passed back to the caller via the         string that matched the subpattern is passed back to the caller via the
4000         ovector argument of pcre_exec(). Opening parentheses are  counted  from         ovector  argument  of pcre_exec(). Opening parentheses are counted from
4001         left  to  right  (starting  from 1) to obtain numbers for the capturing         left to right (starting from 1) to obtain  numbers  for  the  capturing
4002         subpatterns.         subpatterns.
4003    
4004         For example, if the string "the red king" is matched against  the  pat-         For  example,  if the string "the red king" is matched against the pat-
4005         tern         tern
4006    
4007           the ((red|white) (king|queen))           the ((red|white) (king|queen))
# Line 4000  SUBPATTERNS Line 4009  SUBPATTERNS
4009         the captured substrings are "red king", "red", and "king", and are num-         the captured substrings are "red king", "red", and "king", and are num-
4010         bered 1, 2, and 3, respectively.         bered 1, 2, and 3, respectively.
4011    
4012         The fact that plain parentheses fulfil  two  functions  is  not  always         The  fact  that  plain  parentheses  fulfil two functions is not always
4013         helpful.   There are often times when a grouping subpattern is required         helpful.  There are often times when a grouping subpattern is  required
4014         without a capturing requirement. If an opening parenthesis is  followed         without  a capturing requirement. If an opening parenthesis is followed
4015         by  a question mark and a colon, the subpattern does not do any captur-         by a question mark and a colon, the subpattern does not do any  captur-
4016         ing, and is not counted when computing the  number  of  any  subsequent         ing,  and  is  not  counted when computing the number of any subsequent
4017         capturing  subpatterns. For example, if the string "the white queen" is         capturing subpatterns. For example, if the string "the white queen"  is
4018         matched against the pattern         matched against the pattern
4019    
4020           the ((?:red|white) (king|queen))           the ((?:red|white) (king|queen))
# Line 4013  SUBPATTERNS Line 4022  SUBPATTERNS
4022         the captured substrings are "white queen" and "queen", and are numbered         the captured substrings are "white queen" and "queen", and are numbered
4023         1 and 2. The maximum number of capturing subpatterns is 65535.         1 and 2. The maximum number of capturing subpatterns is 65535.
4024    
4025         As  a  convenient shorthand, if any option settings are required at the         As a convenient shorthand, if any option settings are required  at  the
4026         start of a non-capturing subpattern,  the  option  letters  may  appear         start  of  a  non-capturing  subpattern,  the option letters may appear
4027         between the "?" and the ":". Thus the two patterns         between the "?" and the ":". Thus the two patterns
4028    
4029           (?i:saturday|sunday)           (?i:saturday|sunday)
4030           (?:(?i)saturday|sunday)           (?:(?i)saturday|sunday)
4031    
4032         match exactly the same set of strings. Because alternative branches are         match exactly the same set of strings. Because alternative branches are
4033         tried from left to right, and options are not reset until  the  end  of         tried  from  left  to right, and options are not reset until the end of
4034         the  subpattern is reached, an option setting in one branch does affect         the subpattern is reached, an option setting in one branch does  affect
4035         subsequent branches, so the above patterns match "SUNDAY"  as  well  as         subsequent  branches,  so  the above patterns match "SUNDAY" as well as
4036         "Saturday".         "Saturday".
4037    
4038    
4039  DUPLICATE SUBPATTERN NUMBERS  DUPLICATE SUBPATTERN NUMBERS
4040    
4041         Perl 5.10 introduced a feature whereby each alternative in a subpattern         Perl 5.10 introduced a feature whereby each alternative in a subpattern
4042         uses the same numbers for its capturing parentheses. Such a  subpattern         uses  the same numbers for its capturing parentheses. Such a subpattern
4043         starts  with (?| and is itself a non-capturing subpattern. For example,         starts with (?| and is itself a non-capturing subpattern. For  example,
4044         consider this pattern:         consider this pattern:
4045    
4046           (?|(Sat)ur|(Sun))day           (?|(Sat)ur|(Sun))day
4047    
4048         Because the two alternatives are inside a (?| group, both sets of  cap-         Because  the two alternatives are inside a (?| group, both sets of cap-
4049         turing  parentheses  are  numbered one. Thus, when the pattern matches,         turing parentheses are numbered one. Thus, when  the  pattern  matches,
4050         you can look at captured substring number  one,  whichever  alternative         you  can  look  at captured substring number one, whichever alternative
4051         matched.  This  construct  is useful when you want to capture part, but         matched. This construct is useful when you want to  capture  part,  but
4052         not all, of one of a number of alternatives. Inside a (?| group, paren-         not all, of one of a number of alternatives. Inside a (?| group, paren-
4053         theses  are  numbered as usual, but the number is reset at the start of         theses are numbered as usual, but the number is reset at the  start  of
4054         each branch. The numbers of any capturing buffers that follow the  sub-         each  branch. The numbers of any capturing buffers that follow the sub-
4055         pattern  start after the highest number used in any branch. The follow-         pattern start after the highest number used in any branch. The  follow-
4056         ing example is taken from the Perl documentation.  The  numbers  under-         ing  example  is taken from the Perl documentation.  The numbers under-
4057         neath show in which buffer the captured content will be stored.         neath show in which buffer the captured content will be stored.
4058    
4059           # before  ---------------branch-reset----------- after           # before  ---------------branch-reset----------- after
4060           / ( a )  (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x           / ( a )  (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
4061           # 1            2         2  3        2     3     4           # 1            2         2  3        2     3     4
4062    
4063         A  back  reference  to a numbered subpattern uses the most recent value         A back reference to a numbered subpattern uses the  most  recent  value
4064         that is set for that number by any subpattern.  The  following  pattern         that  is  set  for that number by any subpattern. The following pattern
4065         matches "abcabc" or "defdef":         matches "abcabc" or "defdef":
4066    
4067           /(?|(abc)|(def))\1/           /(?|(abc)|(def))\1/
4068    
4069         In  contrast, a recursive or "subroutine" call to a numbered subpattern         In contrast, a recursive or "subroutine" call to a numbered  subpattern
4070         always refers to the first one in the pattern with  the  given  number.         always  refers  to  the first one in the pattern with the given number.
4071         The following pattern matches "abcabc" or "defabc":         The following pattern matches "abcabc" or "defabc":
4072    
4073           /(?|(abc)|(def))(?1)/           /(?|(abc)|(def))(?1)/
4074    
4075         If  a condition test for a subpattern's having matched refers to a non-         If a condition test for a subpattern's having matched refers to a  non-
4076         unique number, the test is true if any of the subpatterns of that  num-         unique  number, the test is true if any of the subpatterns of that num-
4077         ber have matched.         ber have matched.
4078    
4079         An  alternative approach to using this "branch reset" feature is to use         An alternative approach to using this "branch reset" feature is to  use
4080         duplicate named subpatterns, as described in the next section.         duplicate named subpatterns, as described in the next section.
4081    
4082    
4083  NAMED SUBPATTERNS  NAMED SUBPATTERNS
4084    
4085         Identifying capturing parentheses by number is simple, but  it  can  be         Identifying  capturing  parentheses  by number is simple, but it can be
4086         very  hard  to keep track of the numbers in complicated regular expres-         very hard to keep track of the numbers in complicated  regular  expres-
4087         sions. Furthermore, if an  expression  is  modified,  the  numbers  may         sions.  Furthermore,  if  an  expression  is  modified, the numbers may
4088         change.  To help with this difficulty, PCRE supports the naming of sub-         change. To help with this difficulty, PCRE supports the naming of  sub-
4089         patterns. This feature was not added to Perl until release 5.10. Python         patterns. This feature was not added to Perl until release 5.10. Python
4090         had  the  feature earlier, and PCRE introduced it at release 4.0, using         had the feature earlier, and PCRE introduced it at release  4.0,  using
4091         the Python syntax. PCRE now supports both the Perl and the Python  syn-         the  Python syntax. PCRE now supports both the Perl and the Python syn-
4092         tax.  Perl  allows  identically  numbered subpatterns to have different         tax. Perl allows identically numbered  subpatterns  to  have  different
4093         names, but PCRE does not.         names, but PCRE does not.
4094    
4095         In PCRE, a subpattern can be named in one of three  ways:  (?<name>...)         In  PCRE,  a subpattern can be named in one of three ways: (?<name>...)
4096         or  (?'name'...)  as in Perl, or (?P<name>...) as in Python. References         or (?'name'...) as in Perl, or (?P<name>...) as in  Python.  References
4097         to capturing parentheses from other parts of the pattern, such as  back         to  capturing parentheses from other parts of the pattern, such as back
4098         references,  recursion,  and conditions, can be made by name as well as         references, recursion, and conditions, can be made by name as  well  as
4099         by number.         by number.
4100    
4101         Names consist of up to  32  alphanumeric  characters  and  underscores.         Names  consist  of  up  to  32 alphanumeric characters and underscores.
4102         Named  capturing  parentheses  are  still  allocated numbers as well as         Named capturing parentheses are still  allocated  numbers  as  well  as
4103         names, exactly as if the names were not present. The PCRE API  provides         names,  exactly as if the names were not present. The PCRE API provides
4104         function calls for extracting the name-to-number translation table from         function calls for extracting the name-to-number translation table from
4105         a compiled pattern. There is also a convenience function for extracting         a compiled pattern. There is also a convenience function for extracting
4106         a captured substring by name.         a captured substring by name.
4107    
4108         By  default, a name must be unique within a pattern, but it is possible         By default, a name must be unique within a pattern, but it is  possible
4109         to relax this constraint by setting the PCRE_DUPNAMES option at compile         to relax this constraint by setting the PCRE_DUPNAMES option at compile
4110         time.  (Duplicate  names are also always permitted for subpatterns with         time. (Duplicate names are also always permitted for  subpatterns  with
4111         the same number, set up as described in the previous  section.)  Dupli-         the  same  number, set up as described in the previous section.) Dupli-
4112         cate  names  can  be useful for patterns where only one instance of the         cate names can be useful for patterns where only one  instance  of  the
4113         named parentheses can match. Suppose you want to match the  name  of  a         named  parentheses  can  match. Suppose you want to match the name of a
4114         weekday,  either as a 3-letter abbreviation or as the full name, and in         weekday, either as a 3-letter abbreviation or as the full name, and  in
4115         both cases you want to extract the abbreviation. This pattern (ignoring         both cases you want to extract the abbreviation. This pattern (ignoring
4116         the line breaks) does the job:         the line breaks) does the job:
4117    
# Line 4112  NAMED SUBPATTERNS Line 4121  NAMED SUBPATTERNS
4121           (?<DN>Thu)(?:rsday)?|           (?<DN>Thu)(?:rsday)?|
4122           (?<DN>Sat)(?:urday)?           (?<DN>Sat)(?:urday)?
4123    
4124         There  are  five capturing substrings, but only one is ever set after a         There are five capturing substrings, but only one is ever set  after  a
4125         match.  (An alternative way of solving this problem is to use a "branch         match.  (An alternative way of solving this problem is to use a "branch
4126         reset" subpattern, as described in the previous section.)         reset" subpattern, as described in the previous section.)
4127    
4128         The  convenience  function  for extracting the data by name returns the         The convenience function for extracting the data by  name  returns  the
4129         substring for the first (and in this example, the only)  subpattern  of         substring  for  the first (and in this example, the only) subpattern of
4130         that  name  that  matched.  This saves searching to find which numbered         that name that matched. This saves searching  to  find  which  numbered
4131         subpattern it was.         subpattern it was.
4132    
4133         If you make a back reference to  a  non-unique  named  subpattern  from         If  you  make  a  back  reference to a non-unique named subpattern from
4134         elsewhere  in the pattern, the one that corresponds to the first occur-         elsewhere in the pattern, the one that corresponds to the first  occur-
4135         rence of the name is used. In the absence of duplicate numbers (see the         rence of the name is used. In the absence of duplicate numbers (see the
4136         previous  section) this is the one with the lowest number. If you use a         previous section) this is the one with the lowest number. If you use  a
4137         named reference in a condition test (see the section  about  conditions         named  reference  in a condition test (see the section about conditions
4138         below),  either  to check whether a subpattern has matched, or to check         below), either to check whether a subpattern has matched, or  to  check
4139         for recursion, all subpatterns with the same name are  tested.  If  the         for  recursion,  all  subpatterns with the same name are tested. If the
4140         condition  is  true for any one of them, the overall condition is true.         condition is true for any one of them, the overall condition  is  true.
4141         This is the same behaviour as testing by number. For further details of         This is the same behaviour as testing by number. For further details of
4142         the interfaces for handling named subpatterns, see the pcreapi documen-         the interfaces for handling named subpatterns, see the pcreapi documen-
4143         tation.         tation.
4144    
4145         Warning: You cannot use different names to distinguish between two sub-         Warning: You cannot use different names to distinguish between two sub-
4146         patterns  with  the same number because PCRE uses only the numbers when         patterns with the same number because PCRE uses only the  numbers  when
4147         matching. For this reason, an error is given at compile time if differ-         matching. For this reason, an error is given at compile time if differ-
4148         ent  names  are given to subpatterns with the same number. However, you         ent names are given to subpatterns with the same number.  However,  you
4149         can give the same name to subpatterns with the same number,  even  when         can  give  the same name to subpatterns with the same number, even when
4150         PCRE_DUPNAMES is not set.         PCRE_DUPNAMES is not set.
4151    
4152    
4153  REPETITION  REPETITION
4154    
4155         Repetition  is  specified  by  quantifiers, which can follow any of the         Repetition is specified by quantifiers, which can  follow  any  of  the
4156         following items:         following items:
4157    
4158           a literal data character           a literal data character
# Line 4157  REPETITION Line 4166  REPETITION
4166           a parenthesized subpattern (unless it is an assertion)           a parenthesized subpattern (unless it is an assertion)
4167           a recursive or "subroutine" call to a subpattern           a recursive or "subroutine" call to a subpattern
4168    
4169         The general repetition quantifier specifies a minimum and maximum  num-         The  general repetition quantifier specifies a minimum and maximum num-
4170         ber  of  permitted matches, by giving the two numbers in curly brackets         ber of permitted matches, by giving the two numbers in  curly  brackets
4171         (braces), separated by a comma. The numbers must be  less  than  65536,         (braces),  separated  by  a comma. The numbers must be less than 65536,
4172         and the first must be less than or equal to the second. For example:         and the first must be less than or equal to the second. For example:
4173    
4174           z{2,4}           z{2,4}
4175    
4176         matches  "zz",  "zzz",  or  "zzzz". A closing brace on its own is not a         matches "zz", "zzz", or "zzzz". A closing brace on its  own  is  not  a
4177         special character. If the second number is omitted, but  the  comma  is         special  character.  If  the second number is omitted, but the comma is
4178         present,  there  is  no upper limit; if the second number and the comma         present, there is no upper limit; if the second number  and  the  comma
4179         are both omitted, the quantifier specifies an exact number of  required         are  both omitted, the quantifier specifies an exact number of required
4180         matches. Thus         matches. Thus
4181    
4182           [aeiou]{3,}           [aeiou]{3,}
# Line 4176  REPETITION Line 4185  REPETITION
4185    
4186           \d{8}           \d{8}
4187    
4188         matches  exactly  8  digits. An opening curly bracket that appears in a         matches exactly 8 digits. An opening curly bracket that  appears  in  a
4189         position where a quantifier is not allowed, or one that does not  match         position  where a quantifier is not allowed, or one that does not match
4190         the  syntax of a quantifier, is taken as a literal character. For exam-         the syntax of a quantifier, is taken as a literal character. For  exam-
4191         ple, {,6} is not a quantifier, but a literal string of four characters.         ple, {,6} is not a quantifier, but a literal string of four characters.
4192    
4193         In UTF-8 mode, quantifiers apply to UTF-8  characters  rather  than  to         In  UTF-8  mode,  quantifiers  apply to UTF-8 characters rather than to
4194         individual bytes. Thus, for example, \x{100}{2} matches two UTF-8 char-         individual bytes. Thus, for example, \x{100}{2} matches two UTF-8 char-
4195         acters, each of which is represented by a two-byte sequence. Similarly,         acters, each of which is represented by a two-byte sequence. Similarly,
4196         when Unicode property support is available, \X{3} matches three Unicode         when Unicode property support is available, \X{3} matches three Unicode
4197         extended sequences, each of which may be several bytes long  (and  they         extended  sequences,  each of which may be several bytes long (and they
4198         may be of different lengths).         may be of different lengths).
4199    
4200         The quantifier {0} is permitted, causing the expression to behave as if         The quantifier {0} is permitted, causing the expression to behave as if
4201         the previous item and the quantifier were not present. This may be use-         the previous item and the quantifier were not present. This may be use-
4202         ful  for  subpatterns that are referenced as subroutines from elsewhere         ful for subpatterns that are referenced as subroutines  from  elsewhere
4203         in the pattern. Items other than subpatterns that have a {0} quantifier         in the pattern. Items other than subpatterns that have a {0} quantifier
4204         are omitted from the compiled pattern.         are omitted from the compiled pattern.
4205    
4206         For  convenience, the three most common quantifiers have single-charac-         For convenience, the three most common quantifiers have  single-charac-
4207         ter abbreviations:         ter abbreviations:
4208    
4209           *    is equivalent to {0,}           *    is equivalent to {0,}
4210           +    is equivalent to {1,}           +    is equivalent to {1,}
4211           ?    is equivalent to {0,1}           ?    is equivalent to {0,1}
4212    
4213         It is possible to construct infinite loops by  following  a  subpattern         It  is  possible  to construct infinite loops by following a subpattern
4214         that can match no characters with a quantifier that has no upper limit,         that can match no characters with a quantifier that has no upper limit,
4215         for example:         for example:
4216    
4217           (a?)*           (a?)*
4218    
4219         Earlier versions of Perl and PCRE used to give an error at compile time         Earlier versions of Perl and PCRE used to give an error at compile time
4220         for  such  patterns. However, because there are cases where this can be         for such patterns. However, because there are cases where this  can  be
4221         useful, such patterns are now accepted, but if any  repetition  of  the         useful,  such  patterns  are now accepted, but if any repetition of the
4222         subpattern  does in fact match no characters, the loop is forcibly bro-         subpattern does in fact match no characters, the loop is forcibly  bro-
4223         ken.         ken.
4224    
4225         By default, the quantifiers are "greedy", that is, they match  as  much         By  default,  the quantifiers are "greedy", that is, they match as much
4226         as  possible  (up  to  the  maximum number of permitted times), without         as possible (up to the maximum  number  of  permitted  times),  without
4227         causing the rest of the pattern to fail. The classic example  of  where         causing  the  rest of the pattern to fail. The classic example of where
4228         this gives problems is in trying to match comments in C programs. These         this gives problems is in trying to match comments in C programs. These
4229         appear between /* and */ and within the comment,  individual  *  and  /         appear  between  /*  and  */ and within the comment, individual * and /
4230         characters  may  appear. An attempt to match C comments by applying the         characters may appear. An attempt to match C comments by  applying  the
4231         pattern         pattern
4232    
4233           /\*.*\*/           /\*.*\*/
# Line 4227  REPETITION Line 4236  REPETITION
4236    
4237           /* first comment */  not comment  /* second comment */           /* first comment */  not comment  /* second comment */
4238    
4239         fails, because it matches the entire string owing to the greediness  of         fails,  because it matches the entire string owing to the greediness of
4240         the .*  item.         the .*  item.
4241    
4242         However,  if  a quantifier is followed by a question mark, it ceases to         However, if a quantifier is followed by a question mark, it  ceases  to
4243         be greedy, and instead matches the minimum number of times possible, so         be greedy, and instead matches the minimum number of times possible, so
4244         the pattern         the pattern
4245    
4246           /\*.*?\*/           /\*.*?\*/
4247    
4248         does  the  right  thing with the C comments. The meaning of the various         does the right thing with the C comments. The meaning  of  the  various
4249         quantifiers is not otherwise changed,  just  the  preferred  number  of         quantifiers  is  not  otherwise  changed,  just the preferred number of
4250         matches.   Do  not  confuse this use of question mark with its use as a         matches.  Do not confuse this use of question mark with its  use  as  a
4251         quantifier in its own right. Because it has two uses, it can  sometimes         quantifier  in its own right. Because it has two uses, it can sometimes
4252         appear doubled, as in         appear doubled, as in
4253    
4254           \d??\d           \d??\d
# Line 4247  REPETITION Line 4256  REPETITION
4256         which matches one digit by preference, but can match two if that is the         which matches one digit by preference, but can match two if that is the
4257         only way the rest of the pattern matches.         only way the rest of the pattern matches.
4258    
4259         If the PCRE_UNGREEDY option is set (an option that is not available  in         If  the PCRE_UNGREEDY option is set (an option that is not available in
4260         Perl),  the  quantifiers are not greedy by default, but individual ones         Perl), the quantifiers are not greedy by default, but  individual  ones
4261         can be made greedy by following them with a  question  mark.  In  other         can  be  made  greedy  by following them with a question mark. In other
4262         words, it inverts the default behaviour.         words, it inverts the default behaviour.
4263    
4264         When  a  parenthesized  subpattern  is quantified with a minimum repeat         When a parenthesized subpattern is quantified  with  a  minimum  repeat
4265         count that is greater than 1 or with a limited maximum, more memory  is         count  that is greater than 1 or with a limited maximum, more memory is
4266         required  for  the  compiled  pattern, in proportion to the size of the         required for the compiled pattern, in proportion to  the  size  of  the
4267         minimum or maximum.         minimum or maximum.
4268    
4269         If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-         If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
4270         alent  to  Perl's  /s) is set, thus allowing the dot to match newlines,         alent to Perl's /s) is set, thus allowing the dot  to  match  newlines,
4271         the pattern is implicitly anchored, because whatever  follows  will  be         the  pattern  is  implicitly anchored, because whatever follows will be
4272         tried  against every character position in the subject string, so there         tried against every character position in the subject string, so  there
4273         is no point in retrying the overall match at  any  position  after  the         is  no  point  in  retrying the overall match at any position after the
4274         first.  PCRE  normally treats such a pattern as though it were preceded         first. PCRE normally treats such a pattern as though it  were  preceded
4275         by \A.         by \A.
4276    
4277         In cases where it is known that the subject  string  contains  no  new-         In  cases  where  it  is known that the subject string contains no new-
4278         lines,  it  is  worth setting PCRE_DOTALL in order to obtain this opti-         lines, it is worth setting PCRE_DOTALL in order to  obtain  this  opti-
4279         mization, or alternatively using ^ to indicate anchoring explicitly.         mization, or alternatively using ^ to indicate anchoring explicitly.
4280    
4281         However, there is one situation where the optimization cannot be  used.         However,  there is one situation where the optimization cannot be used.
4282         When .*  is inside capturing parentheses that are the subject of a back         When .*  is inside capturing parentheses that are the subject of a back
4283         reference elsewhere in the pattern, a match at the start may fail where         reference elsewhere in the pattern, a match at the start may fail where
4284         a later one succeeds. Consider, for example:         a later one succeeds. Consider, for example:
4285    
4286           (.*)abc\1           (.*)abc\1
4287    
4288         If  the subject is "xyz123abc123" the match point is the fourth charac-         If the subject is "xyz123abc123" the match point is the fourth  charac-
4289         ter. For this reason, such a pattern is not implicitly anchored.         ter. For this reason, such a pattern is not implicitly anchored.
4290    
4291         When a capturing subpattern is repeated, the value captured is the sub-         When a capturing subpattern is repeated, the value captured is the sub-
# Line 4285  REPETITION Line 4294  REPETITION
4294           (tweedle[dume]{3}\s*)+           (tweedle[dume]{3}\s*)+
4295    
4296         has matched "tweedledum tweedledee" the value of the captured substring         has matched "tweedledum tweedledee" the value of the captured substring
4297         is "tweedledee". However, if there are  nested  capturing  subpatterns,         is  "tweedledee".  However,  if there are nested capturing subpatterns,
4298         the  corresponding captured values may have been set in previous itera-         the corresponding captured values may have been set in previous  itera-
4299         tions. For example, after         tions. For example, after
4300    
4301           /(a|(b))+/           /(a|(b))+/
# Line 4296  REPETITION Line 4305  REPETITION
4305    
4306  ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS  ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS
4307    
4308         With both maximizing ("greedy") and minimizing ("ungreedy"  or  "lazy")         With  both  maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
4309         repetition,  failure  of what follows normally causes the repeated item         repetition, failure of what follows normally causes the  repeated  item
4310         to be re-evaluated to see if a different number of repeats  allows  the         to  be  re-evaluated to see if a different number of repeats allows the
4311         rest  of  the pattern to match. Sometimes it is useful to prevent this,         rest of the pattern to match. Sometimes it is useful to  prevent  this,
4312         either to change the nature of the match, or to cause it  fail  earlier         either  to  change the nature of the match, or to cause it fail earlier
4313         than  it otherwise might, when the author of the pattern knows there is         than it otherwise might, when the author of the pattern knows there  is
4314         no point in carrying on.         no point in carrying on.
4315    
4316         Consider, for example, the pattern \d+foo when applied to  the  subject         Consider,  for  example, the pattern \d+foo when applied to the subject
4317         line         line
4318    
4319           123456bar           123456bar
4320    
4321         After matching all 6 digits and then failing to match "foo", the normal         After matching all 6 digits and then failing to match "foo", the normal
4322         action of the matcher is to try again with only 5 digits  matching  the         action  of  the matcher is to try again with only 5 digits matching the
4323         \d+  item,  and  then  with  4,  and  so on, before ultimately failing.         \d+ item, and then with  4,  and  so  on,  before  ultimately  failing.
4324         "Atomic grouping" (a term taken from Jeffrey  Friedl's  book)  provides         "Atomic  grouping"  (a  term taken from Jeffrey Friedl's book) provides
4325         the  means for specifying that once a subpattern has matched, it is not         the means for specifying that once a subpattern has matched, it is  not
4326         to be re-evaluated in this way.         to be re-evaluated in this way.
4327    
4328         If we use atomic grouping for the previous example, the  matcher  gives         If  we  use atomic grouping for the previous example, the matcher gives
4329         up  immediately  on failing to match "foo" the first time. The notation         up immediately on failing to match "foo" the first time.  The  notation
4330         is a kind of special parenthesis, starting with (?> as in this example:         is a kind of special parenthesis, starting with (?> as in this example:
4331    
4332           (?>\d+)foo           (?>\d+)foo
4333    
4334         This kind of parenthesis "locks up" the  part of the  pattern  it  con-         This  kind  of  parenthesis "locks up" the  part of the pattern it con-
4335         tains  once  it  has matched, and a failure further into the pattern is         tains once it has matched, and a failure further into  the  pattern  is
4336         prevented from backtracking into it. Backtracking past it  to  previous         prevented  from  backtracking into it. Backtracking past it to previous
4337         items, however, works as normal.         items, however, works as normal.
4338    
4339         An  alternative  description  is that a subpattern of this type matches         An alternative description is that a subpattern of  this  type  matches
4340         the string of characters that an  identical  standalone  pattern  would         the  string  of  characters  that an identical standalone pattern would
4341         match, if anchored at the current point in the subject string.         match, if anchored at the current point in the subject string.
4342    
4343         Atomic grouping subpatterns are not capturing subpatterns. Simple cases         Atomic grouping subpatterns are not capturing subpatterns. Simple cases
4344         such as the above example can be thought of as a maximizing repeat that         such as the above example can be thought of as a maximizing repeat that
4345         must  swallow  everything  it can. So, while both \d+ and \d+? are pre-         must swallow everything it can. So, while both \d+ and  \d+?  are  pre-
4346         pared to adjust the number of digits they match in order  to  make  the         pared  to  adjust  the number of digits they match in order to make the
4347         rest of the pattern match, (?>\d+) can only match an entire sequence of         rest of the pattern match, (?>\d+) can only match an entire sequence of
4348         digits.         digits.
4349    
4350         Atomic groups in general can of course contain arbitrarily  complicated         Atomic  groups in general can of course contain arbitrarily complicated
4351         subpatterns,  and  can  be  nested. However, when the subpattern for an         subpatterns, and can be nested. However, when  the  subpattern  for  an
4352         atomic group is just a single repeated item, as in the example above, a         atomic group is just a single repeated item, as in the example above, a
4353         simpler  notation,  called  a "possessive quantifier" can be used. This         simpler notation, called a "possessive quantifier" can  be  used.  This
4354         consists of an additional + character  following  a  quantifier.  Using         consists  of  an  additional  + character following a quantifier. Using
4355         this notation, the previous example can be rewritten as         this notation, the previous example can be rewritten as
4356    
4357           \d++foo           \d++foo
# Line 4352  ATOMIC GROUPING AND POSSESSIVE QUANTIFIE Line 4361  ATOMIC GROUPING AND POSSESSIVE QUANTIFIE
4361    
4362           (abc|xyz){2,3}+           (abc|xyz){2,3}+
4363    
4364         Possessive  quantifiers  are  always  greedy;  the   setting   of   the         Possessive   quantifiers   are   always  greedy;  the  setting  of  the
4365         PCRE_UNGREEDY option is ignored. They are a convenient notation for the         PCRE_UNGREEDY option is ignored. They are a convenient notation for the
4366         simpler forms of atomic group. However, there is no difference  in  the         simpler  forms  of atomic group. However, there is no difference in the
4367         meaning  of  a  possessive  quantifier and the equivalent atomic group,         meaning of a possessive quantifier and  the  equivalent  atomic  group,
4368         though there may be a performance  difference;  possessive  quantifiers         though  there  may  be a performance difference; possessive quantifiers
4369         should be slightly faster.         should be slightly faster.
4370    
4371         The  possessive  quantifier syntax is an extension to the Perl 5.8 syn-         The possessive quantifier syntax is an extension to the Perl  5.8  syn-
4372         tax.  Jeffrey Friedl originated the idea (and the name)  in  the  first         tax.   Jeffrey  Friedl  originated the idea (and the name) in the first
4373         edition of his book. Mike McCloskey liked it, so implemented it when he         edition of his book. Mike McCloskey liked it, so implemented it when he
4374         built Sun's Java package, and PCRE copied it from there. It  ultimately         built  Sun's Java package, and PCRE copied it from there. It ultimately
4375         found its way into Perl at release 5.10.         found its way into Perl at release 5.10.
4376    
4377         PCRE has an optimization that automatically "possessifies" certain sim-         PCRE has an optimization that automatically "possessifies" certain sim-
4378         ple pattern constructs. For example, the sequence  A+B  is  treated  as         ple  pattern  constructs.  For  example, the sequence A+B is treated as
4379         A++B  because  there is no point in backtracking into a sequence of A's         A++B because there is no point in backtracking into a sequence  of  A's
4380         when B must follow.         when B must follow.
4381    
4382         When a pattern contains an unlimited repeat inside  a  subpattern  that         When  a  pattern  contains an unlimited repeat inside a subpattern that
4383         can  itself  be  repeated  an  unlimited number of times, the use of an         can itself be repeated an unlimited number of  times,  the  use  of  an
4384         atomic group is the only way to avoid some  failing  matches  taking  a         atomic  group  is  the  only way to avoid some failing matches taking a
4385         very long time indeed. The pattern         very long time indeed. The pattern
4386    
4387           (\D+|<\d+>)*[!?]           (\D+|<\d+>)*[!?]
4388    
4389         matches  an  unlimited number of substrings that either consist of non-         matches an unlimited number of substrings that either consist  of  non-
4390         digits, or digits enclosed in <>, followed by either ! or  ?.  When  it         digits,  or  digits  enclosed in <>, followed by either ! or ?. When it
4391         matches, it runs quickly. However, if it is applied to         matches, it runs quickly. However, if it is applied to
4392    
4393           aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa           aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
4394    
4395         it  takes  a  long  time  before reporting failure. This is because the         it takes a long time before reporting  failure.  This  is  because  the
4396         string can be divided between the internal \D+ repeat and the  external         string  can be divided between the internal \D+ repeat and the external
4397         *  repeat  in  a  large  number of ways, and all have to be tried. (The         * repeat in a large number of ways, and all  have  to  be  tried.  (The
4398         example uses [!?] rather than a single character at  the  end,  because         example  uses  [!?]  rather than a single character at the end, because
4399         both  PCRE  and  Perl have an optimization that allows for fast failure         both PCRE and Perl have an optimization that allows  for  fast  failure
4400         when a single character is used. They remember the last single  charac-         when  a single character is used. They remember the last single charac-
4401         ter  that  is required for a match, and fail early if it is not present         ter that is required for a match, and fail early if it is  not  present
4402         in the string.) If the pattern is changed so that  it  uses  an  atomic         in  the  string.)  If  the pattern is changed so that it uses an atomic
4403         group, like this:         group, like this:
4404    
4405           ((?>\D+)|<\d+>)*[!?]           ((?>\D+)|<\d+>)*[!?]
# Line 4402  BACK REFERENCES Line 4411  BACK REFERENCES
4411    
4412         Outside a character class, a backslash followed by a digit greater than         Outside a character class, a backslash followed by a digit greater than
4413         0 (and possibly further digits) is a back reference to a capturing sub-         0 (and possibly further digits) is a back reference to a capturing sub-
4414         pattern  earlier  (that is, to its left) in the pattern, provided there         pattern earlier (that is, to its left) in the pattern,  provided  there
4415         have been that many previous capturing left parentheses.         have been that many previous capturing left parentheses.
4416    
4417         However, if the decimal number following the backslash is less than 10,         However, if the decimal number following the backslash is less than 10,
4418         it  is  always  taken  as a back reference, and causes an error only if         it is always taken as a back reference, and causes  an  error  only  if
4419         there are not that many capturing left parentheses in the  entire  pat-         there  are  not that many capturing left parentheses in the entire pat-
4420         tern.  In  other words, the parentheses that are referenced need not be         tern. In other words, the parentheses that are referenced need  not  be
4421         to the left of the reference for numbers less than 10. A "forward  back         to  the left of the reference for numbers less than 10. A "forward back
4422         reference"  of  this  type can make sense when a repetition is involved         reference" of this type can make sense when a  repetition  is  involved
4423         and the subpattern to the right has participated in an  earlier  itera-         and  the  subpattern to the right has participated in an earlier itera-
4424         tion.         tion.
4425    
4426         It  is  not  possible to have a numerical "forward back reference" to a         It is not possible to have a numerical "forward back  reference"  to  a
4427         subpattern whose number is 10 or  more  using  this  syntax  because  a         subpattern  whose  number  is  10  or  more using this syntax because a
4428         sequence  such  as  \50 is interpreted as a character defined in octal.         sequence such as \50 is interpreted as a character  defined  in  octal.
4429         See the subsection entitled "Non-printing characters" above for further         See the subsection entitled "Non-printing characters" above for further
4430         details  of  the  handling of digits following a backslash. There is no         details of the handling of digits following a backslash.  There  is  no
4431         such problem when named parentheses are used. A back reference  to  any         such  problem  when named parentheses are used. A back reference to any
4432         subpattern is possible using named parentheses (see below).         subpattern is possible using named parentheses (see below).
4433    
4434         Another  way  of  avoiding  the ambiguity inherent in the use of digits         Another way of avoiding the ambiguity inherent in  the  use  of  digits
4435         following a backslash is to use the \g escape sequence, which is a fea-         following a backslash is to use the \g escape sequence, which is a fea-
4436         ture  introduced  in  Perl  5.10.  This  escape  must be followed by an         ture introduced in Perl 5.10.  This  escape  must  be  followed  by  an
4437         unsigned number or a negative number, optionally  enclosed  in  braces.         unsigned  number  or  a negative number, optionally enclosed in braces.
4438         These examples are all identical:         These examples are all identical:
4439    
4440           (ring), \1           (ring), \1
4441           (ring), \g1           (ring), \g1
4442           (ring), \g{1}           (ring), \g{1}
4443    
4444         An  unsigned number specifies an absolute reference without the ambigu-         An unsigned number specifies an absolute reference without the  ambigu-
4445         ity that is present in the older syntax. It is also useful when literal         ity that is present in the older syntax. It is also useful when literal
4446         digits follow the reference. A negative number is a relative reference.         digits follow the reference. A negative number is a relative reference.
4447         Consider this example:         Consider this example:
# Line 4440  BACK REFERENCES Line 4449  BACK REFERENCES
4449           (abc(def)ghi)\g{-1}           (abc(def)ghi)\g{-1}
4450    
4451         The sequence \g{-1} is a reference to the most recently started captur-         The sequence \g{-1} is a reference to the most recently started captur-
4452         ing  subpattern  before \g, that is, is it equivalent to \2. Similarly,         ing subpattern before \g, that is, is it equivalent to  \2.  Similarly,
4453         \g{-2} would be equivalent to \1. The use of relative references can be         \g{-2} would be equivalent to \1. The use of relative references can be
4454         helpful  in  long  patterns,  and  also in patterns that are created by         helpful in long patterns, and also in  patterns  that  are  created  by
4455         joining together fragments that contain references within themselves.         joining together fragments that contain references within themselves.
4456    
4457         A back reference matches whatever actually matched the  capturing  sub-         A  back  reference matches whatever actually matched the capturing sub-
4458         pattern  in  the  current subject string, rather than anything matching         pattern in the current subject string, rather  than  anything  matching
4459         the subpattern itself (see "Subpatterns as subroutines" below for a way         the subpattern itself (see "Subpatterns as subroutines" below for a way
4460         of doing that). So the pattern         of doing that). So the pattern
4461    
4462           (sens|respons)e and \1ibility           (sens|respons)e and \1ibility
4463    
4464         matches  "sense and sensibility" and "response and responsibility", but         matches "sense and sensibility" and "response and responsibility",  but
4465         not "sense and responsibility". If caseful matching is in force at  the         not  "sense and responsibility". If caseful matching is in force at the
4466         time  of the back reference, the case of letters is relevant. For exam-         time of the back reference, the case of letters is relevant. For  exam-
4467         ple,         ple,
4468    
4469           ((?i)rah)\s+\1           ((?i)rah)\s+\1
4470    
4471         matches "rah rah" and "RAH RAH", but not "RAH  rah",  even  though  the         matches  "rah  rah"  and  "RAH RAH", but not "RAH rah", even though the
4472         original capturing subpattern is matched caselessly.         original capturing subpattern is matched caselessly.
4473    
4474         There  are  several  different ways of writing back references to named         There are several different ways of writing back  references  to  named
4475         subpatterns. The .NET syntax \k{name} and the Perl syntax  \k<name>  or         subpatterns.  The  .NET syntax \k{name} and the Perl syntax \k<name> or
4476         \k'name'  are supported, as is the Python syntax (?P=name). Perl 5.10's         \k'name' are supported, as is the Python syntax (?P=name). Perl  5.10's
4477         unified back reference syntax, in which \g can be used for both numeric         unified back reference syntax, in which \g can be used for both numeric
4478         and  named  references,  is  also supported. We could rewrite the above         and named references, is also supported. We  could  rewrite  the  above
4479         example in any of the following ways:         example in any of the following ways:
4480    
4481           (?<p1>(?i)rah)\s+\k<p1>           (?<p1>(?i)rah)\s+\k<p1>
# Line 4474  BACK REFERENCES Line 4483  BACK REFERENCES
4483           (?P<p1>(?i)rah)\s+(?P=p1)           (?P<p1>(?i)rah)\s+(?P=p1)
4484           (?<p1>(?i)rah)\s+\g{p1}           (?<p1>(?i)rah)\s+\g{p1}
4485    
4486         A subpattern that is referenced by  name  may  appear  in  the  pattern         A  subpattern  that  is  referenced  by  name may appear in the pattern
4487         before or after the reference.         before or after the reference.
4488    
4489         There  may be more than one back reference to the same subpattern. If a         There may be more than one back reference to the same subpattern. If  a
4490         subpattern has not actually been used in a particular match,  any  back         subpattern  has  not actually been used in a particular match, any back
4491         references to it always fail by default. For example, the pattern         references to it always fail by default. For example, the pattern
4492    
4493           (a|(bc))\2           (a|(bc))\2
4494    
4495         always  fails  if  it starts to match "a" rather than "bc". However, if         always fails if it starts to match "a" rather than  "bc".  However,  if
4496         the PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back refer-         the PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back refer-
4497         ence to an unset value matches an empty string.         ence to an unset value matches an empty string.
4498    
4499         Because  there may be many capturing parentheses in a pattern, all dig-         Because there may be many capturing parentheses in a pattern, all  dig-
4500         its following a backslash are taken as part of a potential back  refer-         its  following a backslash are taken as part of a potential back refer-
4501         ence  number.   If  the  pattern continues with a digit character, some         ence number.  If the pattern continues with  a  digit  character,  some
4502         delimiter must  be  used  to  terminate  the  back  reference.  If  the         delimiter  must  be  used  to  terminate  the  back  reference.  If the
4503         PCRE_EXTENDED option is set, this can be whitespace. Otherwise, the \g{         PCRE_EXTENDED option is set, this can be whitespace. Otherwise, the \g{
4504         syntax or an empty comment (see "Comments" below) can be used.         syntax or an empty comment (see "Comments" below) can be used.
4505    
4506     Recursive back references     Recursive back references
4507    
4508         A back reference that occurs inside the parentheses to which it  refers         A  back reference that occurs inside the parentheses to which it refers
4509         fails  when  the subpattern is first used, so, for example, (a\1) never         fails when the subpattern is first used, so, for example,  (a\1)  never
4510         matches.  However, such references can be useful inside  repeated  sub-         matches.   However,  such references can be useful inside repeated sub-
4511         patterns. For example, the pattern         patterns. For example, the pattern
4512    
4513           (a|b\1)+           (a|b\1)+
4514    
4515         matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-         matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
4516         ation of the subpattern,  the  back  reference  matches  the  character         ation  of  the  subpattern,  the  back  reference matches the character
4517         string  corresponding  to  the previous iteration. In order for this to         string corresponding to the previous iteration. In order  for  this  to
4518         work, the pattern must be such that the first iteration does  not  need         work,  the  pattern must be such that the first iteration does not need
4519         to  match the back reference. This can be done using alternation, as in         to match the back reference. This can be done using alternation, as  in
4520         the example above, or by a quantifier with a minimum of zero.         the example above, or by a quantifier with a minimum of zero.
4521    
4522         Back references of this type cause the group that they reference to  be         Back  references of this type cause the group that they reference to be
4523         treated  as  an atomic group.  Once the whole group has been matched, a         treated as an atomic group.  Once the whole group has been  matched,  a
4524         subsequent matching failure cannot cause backtracking into  the  middle         subsequent  matching  failure cannot cause backtracking into the middle
4525         of the group.         of the group.
4526    
4527    
4528  ASSERTIONS  ASSERTIONS
4529    
4530         An  assertion  is  a  test on the characters following or preceding the         An assertion is a test on the characters  following  or  preceding  the
4531         current matching point that does not actually consume  any  characters.         current  matching  point that does not actually consume any characters.
4532         The  simple  assertions  coded  as  \b, \B, \A, \G, \Z, \z, ^ and $ are         The simple assertions coded as \b, \B, \A, \G, \Z,  \z,  ^  and  $  are
4533         described above.         described above.
4534    
4535         More complicated assertions are coded as  subpatterns.  There  are  two         More  complicated  assertions  are  coded as subpatterns. There are two
4536         kinds:  those  that  look  ahead of the current position in the subject         kinds: those that look ahead of the current  position  in  the  subject
4537         string, and those that look  behind  it.  An  assertion  subpattern  is         string,  and  those  that  look  behind  it. An assertion subpattern is
4538         matched  in  the  normal way, except that it does not cause the current         matched in the normal way, except that it does not  cause  the  current
4539         matching position to be changed.         matching position to be changed.
4540    
4541         Assertion subpatterns are not capturing subpatterns,  and  may  not  be         Assertion  subpatterns  are  not  capturing subpatterns, and may not be
4542         repeated,  because  it  makes no sense to assert the same thing several         repeated, because it makes no sense to assert the  same  thing  several
4543         times. If any kind of assertion contains capturing  subpatterns  within         times.  If  any kind of assertion contains capturing subpatterns within
4544         it,  these are counted for the purposes of numbering the capturing sub-         it, these are counted for the purposes of numbering the capturing  sub-
4545         patterns in the whole pattern.  However, substring capturing is carried         patterns in the whole pattern.  However, substring capturing is carried
4546         out  only  for  positive assertions, because it does not make sense for         out only for positive assertions, because it does not  make  sense  for
4547         negative assertions.         negative assertions.
4548    
4549     Lookahead assertions     Lookahead assertions
# Line 4544  ASSERTIONS Line 4553  ASSERTIONS
4553    
4554           \w+(?=;)           \w+(?=;)
4555    
4556         matches  a word followed by a semicolon, but does not include the semi-         matches a word followed by a semicolon, but does not include the  semi-
4557         colon in the match, and         colon in the match, and
4558    
4559           foo(?!bar)           foo(?!bar)
4560    
4561         matches any occurrence of "foo" that is not  followed  by  "bar".  Note         matches  any  occurrence  of  "foo" that is not followed by "bar". Note
4562         that the apparently similar pattern         that the apparently similar pattern
4563    
4564           (?!foo)bar           (?!foo)bar
4565    
4566         does  not  find  an  occurrence  of "bar" that is preceded by something         does not find an occurrence of "bar"  that  is  preceded  by  something
4567         other than "foo"; it finds any occurrence of "bar" whatsoever,  because         other  than "foo"; it finds any occurrence of "bar" whatsoever, because
4568         the assertion (?!foo) is always true when the next three characters are         the assertion (?!foo) is always true when the next three characters are
4569         "bar". A lookbehind assertion is needed to achieve the other effect.         "bar". A lookbehind assertion is needed to achieve the other effect.
4570    
4571         If you want to force a matching failure at some point in a pattern, the         If you want to force a matching failure at some point in a pattern, the
4572         most  convenient  way  to  do  it  is with (?!) because an empty string         most convenient way to do it is  with  (?!)  because  an  empty  string
4573         always matches, so an assertion that requires there not to be an  empty         always  matches, so an assertion that requires there not to be an empty
4574         string  must  always  fail.   The  Perl  5.10 backtracking control verb         string must always fail.   The  Perl  5.10  backtracking  control  verb
4575         (*FAIL) or (*F) is essentially a synonym for (?!).         (*FAIL) or (*F) is essentially a synonym for (?!).
4576    
4577     Lookbehind assertions     Lookbehind assertions
4578    
4579         Lookbehind assertions start with (?<= for positive assertions and  (?<!         Lookbehind  assertions start with (?<= for positive assertions and (?<!
4580         for negative assertions. For example,         for negative assertions. For example,
4581    
4582           (?<!foo)bar           (?<!foo)bar
4583    
4584         does  find  an  occurrence  of "bar" that is not preceded by "foo". The         does find an occurrence of "bar" that is not  preceded  by  "foo".  The
4585         contents of a lookbehind assertion are restricted  such  that  all  the         contents  of  a  lookbehind  assertion are restricted such that all the
4586         strings it matches must have a fixed length. However, if there are sev-         strings it matches must have a fixed length. However, if there are sev-
4587         eral top-level alternatives, they do not all  have  to  have  the  same         eral  top-level  alternatives,  they  do  not all have to have the same
4588         fixed length. Thus         fixed length. Thus
4589    
4590           (?<=bullock|donkey)           (?<=bullock|donkey)
# Line 4584  ASSERTIONS Line 4593  ASSERTIONS
4593    
4594           (?<!dogs?|cats?)           (?<!dogs?|cats?)
4595    
4596         causes  an  error at compile time. Branches that match different length         causes an error at compile time. Branches that match  different  length
4597         strings are permitted only at the top level of a lookbehind  assertion.         strings  are permitted only at the top level of a lookbehind assertion.
4598         This  is an extension compared with Perl (5.8 and 5.10), which requires         This is an extension compared with Perl (5.8 and 5.10), which  requires
4599         all branches to match the same length of string. An assertion such as         all branches to match the same length of string. An assertion such as
4600    
4601           (?<=ab(c|de))           (?<=ab(c|de))
4602    
4603         is not permitted, because its single top-level  branch  can  match  two         is  not  permitted,  because  its single top-level branch can match two
4604         different lengths, but it is acceptable to PCRE if rewritten to use two         different lengths, but it is acceptable to PCRE if rewritten to use two
4605         top-level branches:         top-level branches:
4606    
4607           (?<=abc|abde)           (?<=abc|abde)
4608    
4609         In some cases, the Perl 5.10 escape sequence \K (see above) can be used         In some cases, the Perl 5.10 escape sequence \K (see above) can be used
4610         instead  of  a  lookbehind  assertion  to  get  round  the fixed-length         instead of  a  lookbehind  assertion  to  get  round  the  fixed-length
4611         restriction.         restriction.
4612    
4613         The implementation of lookbehind assertions is, for  each  alternative,         The  implementation  of lookbehind assertions is, for each alternative,
4614         to  temporarily  move the current position back by the fixed length and         to temporarily move the current position back by the fixed  length  and
4615         then try to match. If there are insufficient characters before the cur-         then try to match. If there are insufficient characters before the cur-
4616         rent position, the assertion fails.         rent position, the assertion fails.
4617    
4618         PCRE does not allow the \C escape (which matches a single byte in UTF-8         PCRE does not allow the \C escape (which matches a single byte in UTF-8
4619         mode) to appear in lookbehind assertions, because it makes it  impossi-         mode)  to appear in lookbehind assertions, because it makes it impossi-
4620         ble  to  calculate the length of the lookbehind. The \X and \R escapes,         ble to calculate the length of the lookbehind. The \X and  \R  escapes,
4621         which can match different numbers of bytes, are also not permitted.         which can match different numbers of bytes, are also not permitted.
4622    
4623         "Subroutine" calls (see below) such as (?2) or (?&X) are  permitted  in         "Subroutine"  calls  (see below) such as (?2) or (?&X) are permitted in
4624         lookbehinds,  as  long as the subpattern matches a fixed-length string.         lookbehinds, as long as the subpattern matches a  fixed-length  string.
4625         Recursion, however, is not supported.         Recursion, however, is not supported.
4626    
4627         Possessive quantifiers can  be  used  in  conjunction  with  lookbehind         Possessive  quantifiers  can  be  used  in  conjunction with lookbehind
4628         assertions to specify efficient matching of fixed-length strings at the         assertions to specify efficient matching of fixed-length strings at the
4629         end of subject strings. Consider a simple pattern such as         end of subject strings. Consider a simple pattern such as
4630    
4631           abcd$           abcd$
4632    
4633         when applied to a long string that does  not  match.  Because  matching         when  applied  to  a  long string that does not match. Because matching
4634         proceeds from left to right, PCRE will look for each "a" in the subject         proceeds from left to right, PCRE will look for each "a" in the subject
4635         and then see if what follows matches the rest of the  pattern.  If  the         and  then  see  if what follows matches the rest of the pattern. If the
4636         pattern is specified as         pattern is specified as
4637    
4638           ^.*abcd$           ^.*abcd$
4639    
4640         the  initial .* matches the entire string at first, but when this fails         the initial .* matches the entire string at first, but when this  fails
4641         (because there is no following "a"), it backtracks to match all but the         (because there is no following "a"), it backtracks to match all but the
4642         last  character,  then all but the last two characters, and so on. Once         last character, then all but the last two characters, and so  on.  Once
4643         again the search for "a" covers the entire string, from right to  left,         again  the search for "a" covers the entire string, from right to left,
4644         so we are no better off. However, if the pattern is written as         so we are no better off. However, if the pattern is written as
4645    
4646           ^.*+(?<=abcd)           ^.*+(?<=abcd)
4647    
4648         there  can  be  no backtracking for the .*+ item; it can match only the         there can be no backtracking for the .*+ item; it can  match  only  the
4649         entire string. The subsequent lookbehind assertion does a  single  test         entire  string.  The subsequent lookbehind assertion does a single test
4650         on  the last four characters. If it fails, the match fails immediately.         on the last four characters. If it fails, the match fails  immediately.
4651         For long strings, this approach makes a significant difference  to  the         For  long  strings, this approach makes a significant difference to the
4652         processing time.         processing time.
4653    
4654     Using multiple assertions     Using multiple assertions
# Line 4648  ASSERTIONS Line 4657  ASSERTIONS
4657    
4658           (?<=\d{3})(?<!999)foo           (?<=\d{3})(?<!999)foo
4659    
4660         matches  "foo" preceded by three digits that are not "999". Notice that         matches "foo" preceded by three digits that are not "999". Notice  that
4661         each of the assertions is applied independently at the  same  point  in         each  of  the  assertions is applied independently at the same point in
4662         the  subject  string.  First  there  is a check that the previous three         the subject string. First there is a  check  that  the  previous  three
4663         characters are all digits, and then there is  a  check  that  the  same         characters  are  all  digits,  and  then there is a check that the same
4664         three characters are not "999".  This pattern does not match "foo" pre-         three characters are not "999".  This pattern does not match "foo" pre-
4665         ceded by six characters, the first of which are  digits  and  the  last         ceded  by  six  characters,  the first of which are digits and the last
4666         three  of  which  are not "999". For example, it doesn't match "123abc-         three of which are not "999". For example, it  doesn't  match  "123abc-
4667         foo". A pattern to do that is         foo". A pattern to do that is
4668    
4669           (?<=\d{3}...)(?<!999)foo           (?<=\d{3}...)(?<!999)foo
4670    
4671         This time the first assertion looks at the  preceding  six  characters,         This  time  the  first assertion looks at the preceding six characters,
4672         checking that the first three are digits, and then the second assertion         checking that the first three are digits, and then the second assertion
4673         checks that the preceding three characters are not "999".         checks that the preceding three characters are not "999".
4674    
# Line 4667  ASSERTIONS Line 4676  ASSERTIONS
4676    
4677           (?<=(?<!foo)bar)baz           (?<=(?<!foo)bar)baz
4678    
4679         matches an occurrence of "baz" that is preceded by "bar" which in  turn         matches  an occurrence of "baz" that is preceded by "bar" which in turn
4680         is not preceded by "foo", while         is not preceded by "foo", while
4681    
4682           (?<=\d{3}(?!999)...)foo           (?<=\d{3}(?!999)...)foo
4683    
4684         is  another pattern that matches "foo" preceded by three digits and any         is another pattern that matches "foo" preceded by three digits and  any
4685         three characters that are not "999".         three characters that are not "999".
4686    
4687    
4688  CONDITIONAL SUBPATTERNS  CONDITIONAL SUBPATTERNS
4689    
4690         It is possible to cause the matching process to obey a subpattern  con-         It  is possible to cause the matching process to obey a subpattern con-
4691         ditionally  or to choose between two alternative subpatterns, depending         ditionally or to choose between two alternative subpatterns,  depending
4692         on the result of an assertion, or whether a specific capturing  subpat-         on  the result of an assertion, or whether a specific capturing subpat-
4693         tern  has  already  been matched. The two possible forms of conditional         tern has already been matched. The two possible  forms  of  conditional
4694         subpattern are:         subpattern are:
4695    
4696           (?(condition)yes-pattern)           (?(condition)yes-pattern)
4697           (?(condition)yes-pattern|no-pattern)           (?(condition)yes-pattern|no-pattern)
4698    
4699         If the condition is satisfied, the yes-pattern is used;  otherwise  the         If  the  condition is satisfied, the yes-pattern is used; otherwise the
4700         no-pattern  (if  present)  is used. If there are more than two alterna-         no-pattern (if present) is used. If there are more  than  two  alterna-
4701         tives in the subpattern, a compile-time error occurs.         tives in the subpattern, a compile-time error occurs.
4702    
4703         There are four kinds of condition: references  to  subpatterns,  refer-         There  are  four  kinds of condition: references to subpatterns, refer-
4704         ences to recursion, a pseudo-condition called DEFINE, and assertions.         ences to recursion, a pseudo-condition called DEFINE, and assertions.
4705    
4706     Checking for a used subpattern by number     Checking for a used subpattern by number
4707    
4708         If  the  text between the parentheses consists of a sequence of digits,         If the text between the parentheses consists of a sequence  of  digits,
4709         the condition is true if a capturing subpattern of that number has pre-         the condition is true if a capturing subpattern of that number has pre-
4710         viously  matched.  If  there is more than one capturing subpattern with         viously matched. If there is more than one  capturing  subpattern  with
4711         the same number (see the earlier  section  about  duplicate  subpattern         the  same  number  (see  the earlier section about duplicate subpattern
4712         numbers), the condition is true if any of them have been set. An alter-         numbers), the condition is true if any of them have been set. An alter-
4713         native notation is to precede the digits with a plus or minus sign.  In         native  notation is to precede the digits with a plus or minus sign. In
4714         this  case, the subpattern number is relative rather than absolute. The         this case, the subpattern number is relative rather than absolute.  The
4715         most recently opened parentheses can be referenced by (?(-1), the  next         most  recently opened parentheses can be referenced by (?(-1), the next
4716         most  recent  by  (?(-2),  and so on. In looping constructs it can also         most recent by (?(-2), and so on. In looping  constructs  it  can  also
4717         make sense to refer  to  subsequent  groups  with  constructs  such  as         make  sense  to  refer  to  subsequent  groups  with constructs such as
4718         (?(+2).         (?(+2).
4719    
4720         Consider  the  following  pattern, which contains non-significant white         Consider the following pattern, which  contains  non-significant  white
4721         space to make it more readable (assume the PCRE_EXTENDED option) and to         space to make it more readable (assume the PCRE_EXTENDED option) and to
4722         divide it into three parts for ease of discussion:         divide it into three parts for ease of discussion:
4723    
4724           ( \( )?    [^()]+    (?(1) \) )           ( \( )?    [^()]+    (?(1) \) )
4725    
4726         The  first  part  matches  an optional opening parenthesis, and if that         The first part matches an optional opening  parenthesis,  and  if  that
4727         character is present, sets it as the first captured substring. The sec-         character is present, sets it as the first captured substring. The sec-
4728         ond  part  matches one or more characters that are not parentheses. The         ond part matches one or more characters that are not  parentheses.  The
4729         third part is a conditional subpattern that tests whether the first set         third part is a conditional subpattern that tests whether the first set
4730         of parentheses matched or not. If they did, that is, if subject started         of parentheses matched or not. If they did, that is, if subject started
4731         with an opening parenthesis, the condition is true, and so the yes-pat-         with an opening parenthesis, the condition is true, and so the yes-pat-
4732         tern  is  executed  and  a  closing parenthesis is required. Otherwise,         tern is executed and a  closing  parenthesis  is  required.  Otherwise,
4733         since no-pattern is not present, the  subpattern  matches  nothing.  In         since  no-pattern  is  not  present, the subpattern matches nothing. In
4734         other  words,  this  pattern  matches  a  sequence  of non-parentheses,         other words,  this  pattern  matches  a  sequence  of  non-parentheses,
4735         optionally enclosed in parentheses.         optionally enclosed in parentheses.
4736    
4737         If you were embedding this pattern in a larger one,  you  could  use  a         If  you  were  embedding  this pattern in a larger one, you could use a
4738         relative reference:         relative reference:
4739    
4740           ...other stuff... ( \( )?    [^()]+    (?(-1) \) ) ...           ...other stuff... ( \( )?    [^()]+    (?(-1) \) ) ...
4741    
4742         This  makes  the  fragment independent of the parentheses in the larger         This makes the fragment independent of the parentheses  in  the  larger
4743         pattern.         pattern.
4744    
4745     Checking for a used subpattern by name     Checking for a used subpattern by name
4746    
4747         Perl uses the syntax (?(<name>)...) or (?('name')...)  to  test  for  a         Perl  uses  the  syntax  (?(<name>)...) or (?('name')...) to test for a
4748         used  subpattern  by  name.  For compatibility with earlier versions of         used subpattern by name. For compatibility  with  earlier  versions  of
4749         PCRE, which had this facility before Perl, the syntax  (?(name)...)  is         PCRE,  which  had this facility before Perl, the syntax (?(name)...) is
4750         also  recognized. However, there is a possible ambiguity with this syn-         also recognized. However, there is a possible ambiguity with this  syn-
4751         tax, because subpattern names may  consist  entirely  of  digits.  PCRE         tax,  because  subpattern  names  may  consist entirely of digits. PCRE
4752         looks  first for a named subpattern; if it cannot find one and the name         looks first for a named subpattern; if it cannot find one and the  name
4753         consists entirely of digits, PCRE looks for a subpattern of  that  num-         consists  entirely  of digits, PCRE looks for a subpattern of that num-
4754         ber,  which must be greater than zero. Using subpattern names that con-         ber, which must be greater than zero. Using subpattern names that  con-
4755         sist entirely of digits is not recommended.         sist entirely of digits is not recommended.
4756    
4757         Rewriting the above example to use a named subpattern gives this:         Rewriting the above example to use a named subpattern gives this:
4758    
4759           (?<OPEN> \( )?    [^()]+    (?(<OPEN>) \) )           (?<OPEN> \( )?    [^()]+    (?(<OPEN>) \) )
4760    
4761         If the name used in a condition of this kind is a duplicate,  the  test         If  the  name used in a condition of this kind is a duplicate, the test
4762         is  applied to all subpatterns of the same name, and is true if any one         is applied to all subpatterns of the same name, and is true if any  one
4763         of them has matched.         of them has matched.
4764    
4765     Checking for pattern recursion     Checking for pattern recursion
4766    
4767         If the condition is the string (R), and there is no subpattern with the         If the condition is the string (R), and there is no subpattern with the
4768         name  R, the condition is true if a recursive call to the whole pattern         name R, the condition is true if a recursive call to the whole  pattern
4769         or any subpattern has been made. If digits or a name preceded by amper-         or any subpattern has been made. If digits or a name preceded by amper-
4770         sand follow the letter R, for example:         sand follow the letter R, for example:
4771    
# Line 4764  CONDITIONAL SUBPATTERNS Line 4773  CONDITIONAL SUBPATTERNS
4773    
4774         the condition is true if the most recent recursion is into a subpattern         the condition is true if the most recent recursion is into a subpattern
4775         whose number or name is given. This condition does not check the entire         whose number or name is given. This condition does not check the entire
4776         recursion  stack.  If  the  name  used in a condition of this kind is a         recursion stack. If the name used in a condition  of  this  kind  is  a
4777         duplicate, the test is applied to all subpatterns of the same name, and         duplicate, the test is applied to all subpatterns of the same name, and
4778         is true if any one of them is the most recent recursion.         is true if any one of them is the most recent recursion.
4779    
4780         At  "top  level",  all  these recursion test conditions are false.  The         At "top level", all these recursion test  conditions  are  false.   The
4781         syntax for recursive patterns is described below.         syntax for recursive patterns is described below.
4782    
4783     Defining subpatterns for use by reference only     Defining subpatterns for use by reference only
4784    
4785         If the condition is the string (DEFINE), and  there  is  no  subpattern         If  the  condition  is  the string (DEFINE), and there is no subpattern
4786         with  the  name  DEFINE,  the  condition is always false. In this case,         with the name DEFINE, the condition is  always  false.  In  this  case,
4787         there may be only one alternative  in  the  subpattern.  It  is  always         there  may  be  only  one  alternative  in the subpattern. It is always
4788         skipped  if  control  reaches  this  point  in the pattern; the idea of         skipped if control reaches this point  in  the  pattern;  the  idea  of
4789         DEFINE is that it can be used to define "subroutines" that can be  ref-         DEFINE  is that it can be used to define "subroutines" that can be ref-
4790         erenced  from elsewhere. (The use of "subroutines" is described below.)         erenced from elsewhere. (The use of "subroutines" is described  below.)
4791         For example, a pattern to match an IPv4 address could be  written  like         For  example,  a pattern to match an IPv4 address could be written like
4792         this (ignore whitespace and line breaks):         this (ignore whitespace and line breaks):
4793    
4794           (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )           (?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
4795           \b (?&byte) (\.(?&byte)){3} \b           \b (?&byte) (\.(?&byte)){3} \b
4796    
4797         The  first part of the pattern is a DEFINE group inside which a another         The first part of the pattern is a DEFINE group inside which a  another
4798         group named "byte" is defined. This matches an individual component  of         group  named "byte" is defined. This matches an individual component of
4799         an  IPv4  address  (a number less than 256). When matching takes place,         an IPv4 address (a number less than 256). When  matching  takes  place,
4800         this part of the pattern is skipped because DEFINE acts  like  a  false         this  part  of  the pattern is skipped because DEFINE acts like a false
4801         condition.  The  rest of the pattern uses references to the named group         condition. The rest of the pattern uses references to the  named  group
4802         to match the four dot-separated components of an IPv4 address,  insist-         to  match the four dot-separated components of an IPv4 address, insist-
4803         ing on a word boundary at each end.         ing on a word boundary at each end.
4804    
4805     Assertion conditions     Assertion conditions
4806    
4807         If  the  condition  is  not  in any of the above formats, it must be an         If the condition is not in any of the above  formats,  it  must  be  an
4808         assertion.  This may be a positive or negative lookahead or  lookbehind         assertion.   This may be a positive or negative lookahead or lookbehind
4809         assertion.  Consider  this  pattern,  again  containing non-significant         assertion. Consider  this  pattern,  again  containing  non-significant
4810         white space, and with the two alternatives on the second line:         white space, and with the two alternatives on the second line:
4811    
4812           (?(?=[^a-z]*[a-z])           (?(?=[^a-z]*[a-z])
4813           \d{2}-[a-z]{3}-\d{2}  |  \d{2}-\d{2}-\d{2} )           \d{2}-[a-z]{3}-\d{2}  |  \d{2}-\d{2}-\d{2} )
4814    
4815         The condition  is  a  positive  lookahead  assertion  that  matches  an         The  condition  is  a  positive  lookahead  assertion  that  matches an
4816         optional  sequence of non-letters followed by a letter. In other words,         optional sequence of non-letters followed by a letter. In other  words,
4817         it tests for the presence of at least one letter in the subject.  If  a         it  tests  for the presence of at least one letter in the subject. If a
4818         letter  is found, the subject is matched against the first alternative;         letter is found, the subject is matched against the first  alternative;
4819         otherwise it is  matched  against  the  second.  This  pattern  matches         otherwise  it  is  matched  against  the  second.  This pattern matches
4820         strings  in  one  of the two forms dd-aaa-dd or dd-dd-dd, where aaa are         strings in one of the two forms dd-aaa-dd or dd-dd-dd,  where  aaa  are
4821         letters and dd are digits.         letters and dd are digits.
4822    
4823    
4824  COMMENTS  COMMENTS
4825    
4826         The sequence (?# marks the start of a comment that continues up to  the         The  sequence (?# marks the start of a comment that continues up to the
4827         next  closing  parenthesis.  Nested  parentheses are not permitted. The         next closing parenthesis. Nested parentheses  are  not  permitted.  The
4828         characters that make up a comment play no part in the pattern  matching         characters  that make up a comment play no part in the pattern matching
4829         at all.         at all.
4830    
4831         If  the PCRE_EXTENDED option is set, an unescaped # character outside a         If the PCRE_EXTENDED option is set, an unescaped # character outside  a
4832         character class introduces a  comment  that  continues  to  immediately         character  class  introduces  a  comment  that continues to immediately
4833         after the next newline in the pattern.         after the next newline in the pattern.
4834    
4835    
4836  RECURSIVE PATTERNS  RECURSIVE PATTERNS
4837    
4838         Consider  the problem of matching a string in parentheses, allowing for         Consider the problem of matching a string in parentheses, allowing  for
4839         unlimited nested parentheses. Without the use of  recursion,  the  best         unlimited  nested  parentheses.  Without the use of recursion, the best
4840         that  can  be  done  is  to use a pattern that matches up to some fixed         that can be done is to use a pattern that  matches  up  to  some  fixed
4841         depth of nesting. It is not possible to  handle  an  arbitrary  nesting         depth  of  nesting.  It  is not possible to handle an arbitrary nesting
4842         depth.         depth.
4843    
4844         For some time, Perl has provided a facility that allows regular expres-         For some time, Perl has provided a facility that allows regular expres-
4845         sions to recurse (amongst other things). It does this by  interpolating         sions  to recurse (amongst other things). It does this by interpolating
4846         Perl  code in the expression at run time, and the code can refer to the         Perl code in the expression at run time, and the code can refer to  the
4847         expression itself. A Perl pattern using code interpolation to solve the         expression itself. A Perl pattern using code interpolation to solve the
4848         parentheses problem can be created like this:         parentheses problem can be created like this:
4849    
# Line 4844  RECURSIVE PATTERNS Line 4853  RECURSIVE PATTERNS
4853         refers recursively to the pattern in which it appears.         refers recursively to the pattern in which it appears.
4854    
4855         Obviously, PCRE cannot support the interpolation of Perl code. Instead,         Obviously, PCRE cannot support the interpolation of Perl code. Instead,
4856         it  supports  special  syntax  for recursion of the entire pattern, and         it supports special syntax for recursion of  the  entire  pattern,  and
4857         also for individual subpattern recursion.  After  its  introduction  in         also  for  individual  subpattern  recursion. After its introduction in
4858         PCRE  and  Python,  this  kind of recursion was subsequently introduced         PCRE and Python, this kind of  recursion  was  subsequently  introduced
4859         into Perl at release 5.10.         into Perl at release 5.10.
4860    
4861         A special item that consists of (? followed by a  number  greater  than         A  special  item  that consists of (? followed by a number greater than
4862         zero and a closing parenthesis is a recursive call of the subpattern of         zero and a closing parenthesis is a recursive call of the subpattern of
4863         the given number, provided that it occurs inside that  subpattern.  (If         the  given  number, provided that it occurs inside that subpattern. (If
4864         not,  it  is  a  "subroutine" call, which is described in the next sec-         not, it is a "subroutine" call, which is described  in  the  next  sec-
4865         tion.) The special item (?R) or (?0) is a recursive call of the  entire         tion.)  The special item (?R) or (?0) is a recursive call of the entire
4866         regular expression.         regular expression.
4867    
4868         This  PCRE  pattern  solves  the nested parentheses problem (assume the         This PCRE pattern solves the nested  parentheses  problem  (assume  the
4869         PCRE_EXTENDED option is set so that white space is ignored):         PCRE_EXTENDED option is set so that white space is ignored):
4870    
4871           \( ( [^()]++ | (?R) )* \)           \( ( [^()]++ | (?R) )* \)
4872    
4873         First it matches an opening parenthesis. Then it matches any number  of         First  it matches an opening parenthesis. Then it matches any number of
4874         substrings  which  can  either  be  a sequence of non-parentheses, or a         substrings which can either be a  sequence  of  non-parentheses,  or  a
4875         recursive match of the pattern itself (that is, a  correctly  parenthe-         recursive  match  of the pattern itself (that is, a correctly parenthe-
4876         sized substring).  Finally there is a closing parenthesis. Note the use         sized substring).  Finally there is a closing parenthesis. Note the use
4877         of a possessive quantifier to avoid backtracking into sequences of non-         of a possessive quantifier to avoid backtracking into sequences of non-
4878         parentheses.         parentheses.
4879    
4880         If  this  were  part of a larger pattern, you would not want to recurse         If this were part of a larger pattern, you would not  want  to  recurse
4881         the entire pattern, so instead you could use this:         the entire pattern, so instead you could use this:
4882    
4883           ( \( ( [^()]++ | (?1) )* \) )           ( \( ( [^()]++ | (?1) )* \) )
4884    
4885         We have put the pattern into parentheses, and caused the  recursion  to         We  have  put the pattern into parentheses, and caused the recursion to
4886         refer to them instead of the whole pattern.         refer to them instead of the whole pattern.
4887    
4888         In  a  larger  pattern,  keeping  track  of  parenthesis numbers can be         In a larger pattern,  keeping  track  of  parenthesis  numbers  can  be
4889         tricky. This is made easier by the use of relative references  (a  Perl         tricky.  This  is made easier by the use of relative references (a Perl
4890         5.10  feature).   Instead  of  (?1)  in the pattern above you can write         5.10 feature).  Instead of (?1) in the  pattern  above  you  can  write
4891         (?-2) to refer to the second most recently opened parentheses preceding         (?-2) to refer to the second most recently opened parentheses preceding
4892         the  recursion.  In  other  words,  a  negative number counts capturing         the recursion. In other  words,  a  negative  number  counts  capturing
4893         parentheses leftwards from the point at which it is encountered.         parentheses leftwards from the point at which it is encountered.
4894    
4895         It is also possible to refer to  subsequently  opened  parentheses,  by         It  is  also  possible  to refer to subsequently opened parentheses, by
4896         writing  references  such  as (?+2). However, these cannot be recursive         writing references such as (?+2). However, these  cannot  be  recursive
4897         because the reference is not inside the  parentheses  that  are  refer-         because  the  reference  is  not inside the parentheses that are refer-
4898         enced.  They  are  always  "subroutine" calls, as described in the next         enced. They are always "subroutine" calls, as  described  in  the  next
4899         section.         section.
4900    
4901         An alternative approach is to use named parentheses instead.  The  Perl         An  alternative  approach is to use named parentheses instead. The Perl
4902         syntax  for  this  is (?&name); PCRE's earlier syntax (?P>name) is also         syntax for this is (?&name); PCRE's earlier syntax  (?P>name)  is  also
4903         supported. We could rewrite the above example as follows:         supported. We could rewrite the above example as follows:
4904    
4905           (?<pn> \( ( [^()]++ | (?&pn) )* \) )           (?<pn> \( ( [^()]++ | (?&pn) )* \) )
4906    
4907         If there is more than one subpattern with the same name,  the  earliest         If  there  is more than one subpattern with the same name, the earliest
4908         one is used.         one is used.
4909    
4910         This  particular  example pattern that we have been looking at contains         This particular example pattern that we have been looking  at  contains
4911         nested unlimited repeats, and so the use of a possessive quantifier for         nested unlimited repeats, and so the use of a possessive quantifier for
4912         matching strings of non-parentheses is important when applying the pat-         matching strings of non-parentheses is important when applying the pat-
4913         tern to strings that do not match. For example, when  this  pattern  is         tern  to  strings  that do not match. For example, when this pattern is
4914         applied to         applied to
4915    
4916           (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()           (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
4917    
4918         it  yields  "no  match" quickly. However, if a possessive quantifier is         it yields "no match" quickly. However, if a  possessive  quantifier  is
4919         not used, the match runs for a very long time indeed because there  are         not  used, the match runs for a very long time indeed because there are
4920         so  many  different  ways the + and * repeats can carve up the subject,         so many different ways the + and * repeats can carve  up  the  subject,
4921         and all have to be tested before failure can be reported.         and all have to be tested before failure can be reported.
4922    
4923         At the end of a match, the values of capturing  parentheses  are  those         At  the  end  of a match, the values of capturing parentheses are those
4924         from  the outermost level. If you want to obtain intermediate values, a         from the outermost level. If you want to obtain intermediate values,  a
4925         callout function can be used (see below and the pcrecallout  documenta-         callout  function can be used (see below and the pcrecallout documenta-
4926         tion). If the pattern above is matched against         tion). If the pattern above is matched against
4927    
4928           (ab(cd)ef)           (ab(cd)ef)
4929    
4930         the  value  for  the  inner capturing parentheses (numbered 2) is "ef",         the value for the inner capturing parentheses  (numbered  2)  is  "ef",
4931         which is the last value taken on at the top level. If a capturing  sub-         which  is the last value taken on at the top level. If a capturing sub-
4932         pattern is not matched at the top level, its final value is unset, even         pattern is not matched at the top level, its final value is unset, even
4933         if it is (temporarily) set at a deeper level.         if it is (temporarily) set at a deeper level.
4934    
4935         If there are more than 15 capturing parentheses in a pattern, PCRE  has         If  there are more than 15 capturing parentheses in a pattern, PCRE has
4936         to  obtain extra memory to store data during a recursion, which it does         to obtain extra memory to store data during a recursion, which it  does
4937         by using pcre_malloc, freeing it via pcre_free afterwards. If no memory         by using pcre_malloc, freeing it via pcre_free afterwards. If no memory
4938         can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.         can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
4939    
4940         Do  not  confuse  the (?R) item with the condition (R), which tests for         Do not confuse the (?R) item with the condition (R),  which  tests  for
4941         recursion.  Consider this pattern, which matches text in  angle  brack-         recursion.   Consider  this pattern, which matches text in angle brack-
4942         ets,  allowing for arbitrary nesting. Only digits are allowed in nested         ets, allowing for arbitrary nesting. Only digits are allowed in  nested
4943         brackets (that is, when recursing), whereas any characters are  permit-         brackets  (that is, when recursing), whereas any characters are permit-
4944         ted at the outer level.         ted at the outer level.
4945    
4946           < (?: (?(R) \d++  | [^<>]*+) | (?R)) * >           < (?: (?(R) \d++  | [^<>]*+) | (?R)) * >
4947    
4948         In  this  pattern, (?(R) is the start of a conditional subpattern, with         In this pattern, (?(R) is the start of a conditional  subpattern,  with
4949         two different alternatives for the recursive and  non-recursive  cases.         two  different  alternatives for the recursive and non-recursive cases.
4950         The (?R) item is the actual recursive call.         The (?R) item is the actual recursive call.
4951    
4952     Recursion difference from Perl     Recursion difference from Perl
4953    
4954         In  PCRE (like Python, but unlike Perl), a recursive subpattern call is         In PCRE (like Python, but unlike Perl), a recursive subpattern call  is
4955         always treated as an atomic group. That is, once it has matched some of         always treated as an atomic group. That is, once it has matched some of
4956         the subject string, it is never re-entered, even if it contains untried         the subject string, it is never re-entered, even if it contains untried
4957         alternatives and there is a subsequent matching failure.  This  can  be         alternatives  and  there  is a subsequent matching failure. This can be
4958         illustrated  by the following pattern, which purports to match a palin-         illustrated by the following pattern, which purports to match a  palin-
4959         dromic string that contains an odd number of characters  (for  example,         dromic  string  that contains an odd number of characters (for example,
4960         "a", "aba", "abcba", "abcdcba"):         "a", "aba", "abcba", "abcdcba"):
4961    
4962           ^(.|(.)(?1)\2)$           ^(.|(.)(?1)\2)$
4963    
4964         The idea is that it either matches a single character, or two identical         The idea is that it either matches a single character, or two identical
4965         characters surrounding a sub-palindrome. In Perl, this  pattern  works;         characters  surrounding  a sub-palindrome. In Perl, this pattern works;
4966         in  PCRE  it  does  not if the pattern is longer than three characters.         in PCRE it does not if the pattern is  longer  than  three  characters.
4967         Consider the subject string "abcba":         Consider the subject string "abcba":
4968    
4969         At the top level, the first character is matched, but as it is  not  at         At  the  top level, the first character is matched, but as it is not at
4970         the end of the string, the first alternative fails; the second alterna-         the end of the string, the first alternative fails; the second alterna-
4971         tive is taken and the recursion kicks in. The recursive call to subpat-         tive is taken and the recursion kicks in. The recursive call to subpat-
4972         tern  1  successfully  matches the next character ("b"). (Note that the         tern 1 successfully matches the next character ("b").  (Note  that  the
4973         beginning and end of line tests are not part of the recursion).         beginning and end of line tests are not part of the recursion).
4974    
4975         Back at the top level, the next character ("c") is compared  with  what         Back  at  the top level, the next character ("c") is compared with what
4976         subpattern  2 matched, which was "a". This fails. Because the recursion         subpattern 2 matched, which was "a". This fails. Because the  recursion
4977         is treated as an atomic group, there are now  no  backtracking  points,         is  treated  as  an atomic group, there are now no backtracking points,
4978         and  so  the  entire  match fails. (Perl is able, at this point, to re-         and so the entire match fails. (Perl is able, at  this  point,  to  re-
4979         enter the recursion and try the second alternative.)  However,  if  the         enter  the  recursion  and try the second alternative.) However, if the
4980         pattern is written with the alternatives in the other order, things are         pattern is written with the alternatives in the other order, things are
4981         different:         different:
4982    
4983           ^((.)(?1)\2|.)$           ^((.)(?1)\2|.)$
4984    
4985         This time, the recursing alternative is tried first, and  continues  to         This  time,  the recursing alternative is tried first, and continues to
4986         recurse  until  it runs out of characters, at which point the recursion         recurse until it runs out of characters, at which point  the  recursion
4987         fails. But this time we do have  another  alternative  to  try  at  the         fails.  But  this  time  we  do  have another alternative to try at the
4988         higher  level.  That  is  the  big difference: in the previous case the         higher level. That is the big difference:  in  the  previous  case  the
4989         remaining alternative is at a deeper recursion level, which PCRE cannot         remaining alternative is at a deeper recursion level, which PCRE cannot
4990         use.         use.
4991    
4992         To change the pattern so that matches all palindromic strings, not just         To change the pattern so that matches all palindromic strings, not just
4993         those with an odd number of characters, it is tempting  to  change  the         those  with  an  odd number of characters, it is tempting to change the
4994         pattern to this:         pattern to this:
4995    
4996           ^((.)(?1)\2|.?)$           ^((.)(?1)\2|.?)$
4997    
4998         Again,  this  works  in Perl, but not in PCRE, and for the same reason.         Again, this works in Perl, but not in PCRE, and for  the  same  reason.
4999         When a deeper recursion has matched a single character,  it  cannot  be         When  a  deeper  recursion has matched a single character, it cannot be
5000         entered  again  in  order  to match an empty string. The solution is to         entered again in order to match an empty string.  The  solution  is  to
5001         separate the two cases, and write out the odd and even cases as  alter-         separate  the two cases, and write out the odd and even cases as alter-
5002         natives at the higher level:         natives at the higher level:
5003    
5004           ^(?:((.)(?1)\2|)|((.)(?3)\4|.))           ^(?:((.)(?1)\2|)|((.)(?3)\4|.))
5005    
5006         If  you  want  to match typical palindromic phrases, the pattern has to         If you want to match typical palindromic phrases, the  pattern  has  to
5007         ignore all non-word characters, which can be done like this:         ignore all non-word characters, which can be done like this:
5008    
5009           ^\W*+(?:((.)\W*+(?1)\W*+\2|)|((.)\W*+(?3)\W*+\4|\W*+.\W*+))\W*+$           ^\W*+(?:((.)\W*+(?1)\W*+\2|)|((.)\W*+(?3)\W*+\4|\W*+.\W*+))\W*+$
5010    
5011         If run with the PCRE_CASELESS option, this pattern matches phrases such         If run with the PCRE_CASELESS option, this pattern matches phrases such
5012         as "A man, a plan, a canal: Panama!" and it works well in both PCRE and         as "A man, a plan, a canal: Panama!" and it works well in both PCRE and
5013         Perl. Note the use of the possessive quantifier *+ to avoid  backtrack-         Perl.  Note the use of the possessive quantifier *+ to avoid backtrack-
5014         ing  into  sequences of non-word characters. Without this, PCRE takes a         ing into sequences of non-word characters. Without this, PCRE  takes  a
5015         great deal longer (ten times or more) to  match  typical  phrases,  and         great  deal  longer  (ten  times or more) to match typical phrases, and
5016         Perl takes so long that you think it has gone into a loop.         Perl takes so long that you think it has gone into a loop.
5017    
5018         WARNING:  The  palindrome-matching patterns above work only if the sub-         WARNING: The palindrome-matching patterns above work only if  the  sub-
5019         ject string does not start with a palindrome that is shorter  than  the         ject  string  does not start with a palindrome that is shorter than the
5020         entire  string.  For example, although "abcba" is correctly matched, if         entire string.  For example, although "abcba" is correctly matched,  if
5021         the subject is "ababa", PCRE finds the palindrome "aba" at  the  start,         the  subject  is "ababa", PCRE finds the palindrome "aba" at the start,
5022         then  fails at top level because the end of the string does not follow.         then fails at top level because the end of the string does not  follow.
5023         Once again, it cannot jump back into the recursion to try other  alter-         Once  again, it cannot jump back into the recursion to try other alter-
5024         natives, so the entire match fails.         natives, so the entire match fails.
5025    
5026    
5027  SUBPATTERNS AS SUBROUTINES  SUBPATTERNS AS SUBROUTINES
5028    
5029         If the syntax for a recursive subpattern reference (either by number or         If the syntax for a recursive subpattern reference (either by number or
5030         by name) is used outside the parentheses to which it refers,  it  oper-         by  name)  is used outside the parentheses to which it refers, it oper-
5031         ates  like a subroutine in a programming language. The "called" subpat-         ates like a subroutine in a programming language. The "called"  subpat-
5032         tern may be defined before or after the reference. A numbered reference         tern may be defined before or after the reference. A numbered reference
5033         can be absolute or relative, as in these examples:         can be absolute or relative, as in these examples:
5034    
# Line 5031  SUBPATTERNS AS SUBROUTINES Line 5040  SUBPATTERNS AS SUBROUTINES
5040    
5041           (sens|respons)e and \1ibility           (sens|respons)e and \1ibility
5042    
5043         matches  "sense and sensibility" and "response and responsibility", but         matches "sense and sensibility" and "response and responsibility",  but
5044         not "sense and responsibility". If instead the pattern         not "sense and responsibility". If instead the pattern
5045    
5046           (sens|respons)e and (?1)ibility           (sens|respons)e and (?1)ibility
5047    
5048         is used, it does match "sense and responsibility" as well as the  other         is  used, it does match "sense and responsibility" as well as the other
5049         two  strings.  Another  example  is  given  in the discussion of DEFINE         two strings. Another example is  given  in  the  discussion  of  DEFINE
5050         above.         above.
5051    
5052         Like recursive subpatterns, a subroutine call is always treated  as  an         Like  recursive  subpatterns, a subroutine call is always treated as an
5053         atomic  group. That is, once it has matched some of the subject string,         atomic group. That is, once it has matched some of the subject  string,
5054         it is never re-entered, even if it contains  untried  alternatives  and         it  is  never  re-entered, even if it contains untried alternatives and
5055         there  is a subsequent matching failure. Any capturing parentheses that         there is a subsequent matching failure. Any capturing parentheses  that
5056         are set during the subroutine call  revert  to  their  previous  values         are  set  during  the  subroutine  call revert to their previous values
5057         afterwards.         afterwards.
5058    
5059         When  a  subpattern is used as a subroutine, processing options such as         When a subpattern is used as a subroutine, processing options  such  as
5060         case-independence are fixed when the subpattern is defined. They cannot         case-independence are fixed when the subpattern is defined. They cannot
5061         be changed for different calls. For example, consider this pattern:         be changed for different calls. For example, consider this pattern:
5062    
5063           (abc)(?i:(?-1))           (abc)(?i:(?-1))
5064    
5065         It  matches  "abcabc". It does not match "abcABC" because the change of         It matches "abcabc". It does not match "abcABC" because the  change  of
5066         processing option does not affect the called subpattern.         processing option does not affect the called subpattern.
5067    
5068    
5069  ONIGURUMA SUBROUTINE SYNTAX  ONIGURUMA SUBROUTINE SYNTAX
5070    
5071         For compatibility with Oniguruma, the non-Perl syntax \g followed by  a         For  compatibility with Oniguruma, the non-Perl syntax \g followed by a
5072         name or a number enclosed either in angle brackets or single quotes, is         name or a number enclosed either in angle brackets or single quotes, is
5073         an alternative syntax for referencing a  subpattern  as  a  subroutine,         an  alternative  syntax  for  referencing a subpattern as a subroutine,
5074         possibly  recursively. Here are two of the examples used above, rewrit-         possibly recursively. Here are two of the examples used above,  rewrit-
5075         ten using this syntax:         ten using this syntax:
5076    
5077           (?<pn> \( ( (?>[^()]+) | \g<pn> )* \) )           (?<pn> \( ( (?>[^()]+) | \g<pn> )* \) )
5078           (sens|respons)e and \g'1'ibility           (sens|respons)e and \g'1'ibility
5079    
5080         PCRE supports an extension to Oniguruma: if a number is preceded  by  a         PCRE  supports  an extension to Oniguruma: if a number is preceded by a
5081         plus or a minus sign it is taken as a relative reference. For example:         plus or a minus sign it is taken as a relative reference. For example:
5082    
5083           (abc)(?i:\g<-1>)           (abc)(?i:\g<-1>)
5084    
5085         Note  that \g{...} (Perl syntax) and \g<...> (Oniguruma syntax) are not         Note that \g{...} (Perl syntax) and \g<...> (Oniguruma syntax) are  not
5086         synonymous. The former is a back reference; the latter is a  subroutine         synonymous.  The former is a back reference; the latter is a subroutine
5087         call.         call.
5088    
5089    
5090  CALLOUTS  CALLOUTS
5091    
5092         Perl has a feature whereby using the sequence (?{...}) causes arbitrary         Perl has a feature whereby using the sequence (?{...}) causes arbitrary
5093         Perl code to be obeyed in the middle of matching a regular  expression.         Perl  code to be obeyed in the middle of matching a regular expression.
5094         This makes it possible, amongst other things, to extract different sub-         This makes it possible, amongst other things, to extract different sub-
5095         strings that match the same pair of parentheses when there is a repeti-         strings that match the same pair of parentheses when there is a repeti-
5096         tion.         tion.
5097    
5098         PCRE provides a similar feature, but of course it cannot obey arbitrary         PCRE provides a similar feature, but of course it cannot obey arbitrary
5099         Perl code. The feature is called "callout". The caller of PCRE provides         Perl code. The feature is called "callout". The caller of PCRE provides
5100         an  external function by putting its entry point in the global variable         an external function by putting its entry point in the global  variable
5101         pcre_callout.  By default, this variable contains NULL, which  disables         pcre_callout.   By default, this variable contains NULL, which disables
5102         all calling out.         all calling out.
5103    
5104         Within  a  regular  expression,  (?C) indicates the points at which the         Within a regular expression, (?C) indicates the  points  at  which  the
5105         external function is to be called. If you want  to  identify  different         external  function  is  to be called. If you want to identify different
5106         callout  points, you can put a number less than 256 after the letter C.         callout points, you can put a number less than 256 after the letter  C.
5107         The default value is zero.  For example, this pattern has  two  callout         The  default  value is zero.  For example, this pattern has two callout
5108         points:         points:
5109    
5110           (?C1)abc(?C2)def           (?C1)abc(?C2)def
5111    
5112         If the PCRE_AUTO_CALLOUT flag is passed to pcre_compile(), callouts are         If the PCRE_AUTO_CALLOUT flag is passed to pcre_compile(), callouts are
5113         automatically installed before each item in the pattern. They  are  all         automatically  installed  before each item in the pattern. They are all
5114         numbered 255.         numbered 255.
5115    
5116         During matching, when PCRE reaches a callout point (and pcre_callout is         During matching, when PCRE reaches a callout point (and pcre_callout is
5117         set), the external function is called. It is provided with  the  number         set),  the  external function is called. It is provided with the number
5118         of  the callout, the position in the pattern, and, optionally, one item         of the callout, the position in the pattern, and, optionally, one  item
5119         of data originally supplied by the caller of pcre_exec().  The  callout         of  data  originally supplied by the caller of pcre_exec(). The callout
5120         function  may cause matching to proceed, to backtrack, or to fail alto-         function may cause matching to proceed, to backtrack, or to fail  alto-
5121         gether. A complete description of the interface to the callout function         gether. A complete description of the interface to the callout function
5122         is given in the pcrecallout documentation.         is given in the pcrecallout documentation.
5123    
5124    
5125  BACKTRACKING CONTROL  BACKTRACKING CONTROL
5126    
5127         Perl  5.10 introduced a number of "Special Backtracking Control Verbs",         Perl 5.10 introduced a number of "Special Backtracking Control  Verbs",
5128         which are described in the Perl documentation as "experimental and sub-         which are described in the Perl documentation as "experimental and sub-
5129         ject  to  change or removal in a future version of Perl". It goes on to         ject to change or removal in a future version of Perl". It goes  on  to
5130         say: "Their usage in production code should be noted to avoid  problems         say:  "Their usage in production code should be noted to avoid problems
5131         during upgrades." The same remarks apply to the PCRE features described         during upgrades." The same remarks apply to the PCRE features described
5132         in this section.         in this section.
5133    
5134         Since these verbs are specifically related  to  backtracking,  most  of         Since  these  verbs  are  specifically related to backtracking, most of
5135         them  can  be  used  only  when  the  pattern  is  to  be matched using         them can be  used  only  when  the  pattern  is  to  be  matched  using
5136         pcre_exec(), which uses a backtracking algorithm. With the exception of         pcre_exec(), which uses a backtracking algorithm. With the exception of
5137         (*FAIL), which behaves like a failing negative assertion, they cause an         (*FAIL), which behaves like a failing negative assertion, they cause an
5138         error if encountered by pcre_dfa_exec().         error if encountered by pcre_dfa_exec().
5139    
5140         If any of these verbs are used in an assertion or subroutine subpattern         If any of these verbs are used in an assertion or subroutine subpattern
5141         (including  recursive  subpatterns),  their  effect is confined to that         (including recursive subpatterns), their effect  is  confined  to  that
5142         subpattern; it does not extend to the surrounding  pattern.  Note  that         subpattern;  it  does  not extend to the surrounding pattern. Note that
5143         such  subpatterns are processed as anchored at the point where they are         such subpatterns are processed as anchored at the point where they  are
5144         tested.         tested.
5145    
5146         The new verbs make use of what was previously invalid syntax: an  open-         The  new verbs make use of what was previously invalid syntax: an open-
5147         ing parenthesis followed by an asterisk. In Perl, they are generally of         ing parenthesis followed by an asterisk. In Perl, they are generally of
5148         the form (*VERB:ARG) but PCRE does not support the use of arguments, so         the form (*VERB:ARG) but PCRE does not support the use of arguments, so
5149         its  general  form is just (*VERB). Any number of these verbs may occur         its general form is just (*VERB). Any number of these verbs  may  occur
5150         in a pattern. There are two kinds:         in a pattern. There are two kinds:
5151    
5152     Verbs that act immediately     Verbs that act immediately
# Line 5146  BACKTRACKING CONTROL Line 5155  BACKTRACKING CONTROL
5155    
5156            (*ACCEPT)            (*ACCEPT)
5157    
5158         This verb causes the match to end successfully, skipping the  remainder         This  verb causes the match to end successfully, skipping the remainder
5159         of  the pattern. When inside a recursion, only the innermost pattern is         of the pattern. When inside a recursion, only the innermost pattern  is
5160         ended immediately. If (*ACCEPT) is inside  capturing  parentheses,  the         ended  immediately.  If  (*ACCEPT) is inside capturing parentheses, the
5161         data  so  far  is  captured. (This feature was added to PCRE at release         data so far is captured. (This feature was added  to  PCRE  at  release
5162         8.00.) For example:         8.00.) For example:
5163    
5164           A((?:A|B(*ACCEPT)|C)D)           A((?:A|B(*ACCEPT)|C)D)
5165    
5166         This matches "AB", "AAD", or "ACD"; when it matches "AB", "B"  is  cap-         This  matches  "AB", "AAD", or "ACD"; when it matches "AB", "B" is cap-
5167         tured by the outer parentheses.         tured by the outer parentheses.
5168    
5169           (*FAIL) or (*F)           (*FAIL) or (*F)
5170    
5171         This  verb  causes the match to fail, forcing backtracking to occur. It         This verb causes the match to fail, forcing backtracking to  occur.  It
5172         is equivalent to (?!) but easier to read. The Perl documentation  notes         is  equivalent to (?!) but easier to read. The Perl documentation notes
5173         that  it  is  probably  useful only when combined with (?{}) or (??{}).         that it is probably useful only when combined  with  (?{})  or  (??{}).
5174         Those are, of course, Perl features that are not present in  PCRE.  The         Those  are,  of course, Perl features that are not present in PCRE. The
5175         nearest  equivalent is the callout feature, as for example in this pat-         nearest equivalent is the callout feature, as for example in this  pat-
5176         tern:         tern:
5177    
5178           a+(?C)(*FAIL)           a+(?C)(*FAIL)
5179    
5180         A match with the string "aaaa" always fails, but the callout  is  taken         A  match  with the string "aaaa" always fails, but the callout is taken
5181         before each backtrack happens (in this example, 10 times).         before each backtrack happens (in this example, 10 times).
5182    
5183     Verbs that act after backtracking     Verbs that act after backtracking
5184    
5185         The following verbs do nothing when they are encountered. Matching con-         The following verbs do nothing when they are encountered. Matching con-
5186         tinues with what follows, but if there is no subsequent match, a  fail-         tinues  with what follows, but if there is no subsequent match, a fail-
5187         ure  is  forced.   The  verbs  differ  in  exactly what kind of failure         ure is forced.  The verbs  differ  in  exactly  what  kind  of  failure
5188         occurs.         occurs.
5189    
5190           (*COMMIT)           (*COMMIT)
5191    
5192         This verb causes the whole match to fail outright if the  rest  of  the         This  verb  causes  the whole match to fail outright if the rest of the
5193         pattern  does  not match. Even if the pattern is unanchored, no further         pattern does not match. Even if the pattern is unanchored,  no  further
5194         attempts to find a match by advancing the starting  point  take  place.         attempts  to  find  a match by advancing the starting point take place.
5195         Once  (*COMMIT)  has been passed, pcre_exec() is committed to finding a         Once (*COMMIT) has been passed, pcre_exec() is committed to  finding  a
5196         match at the current starting point, or not at all. For example:         match at the current starting point, or not at all. For example:
5197    
5198           a+(*COMMIT)b           a+(*COMMIT)b
5199    
5200         This matches "xxaab" but not "aacaab". It can be thought of as  a  kind         This  matches  "xxaab" but not "aacaab". It can be thought of as a kind
5201         of dynamic anchor, or "I've started, so I must finish."         of dynamic anchor, or "I've started, so I must finish."
5202    
5203           (*PRUNE)           (*PRUNE)
5204    
5205         This  verb causes the match to fail at the current position if the rest         This verb causes the match to fail at the current position if the  rest
5206         of the pattern does not match. If the pattern is unanchored, the normal         of the pattern does not match. If the pattern is unanchored, the normal
5207         "bumpalong"  advance to the next starting character then happens. Back-         "bumpalong" advance to the next starting character then happens.  Back-
5208         tracking can occur as usual to the left of (*PRUNE), or  when  matching         tracking  can  occur as usual to the left of (*PRUNE), or when matching
5209         to  the right of (*PRUNE), but if there is no match to the right, back-         to the right of (*PRUNE), but if there is no match to the right,  back-
5210         tracking cannot cross (*PRUNE).  In simple cases, the use  of  (*PRUNE)         tracking  cannot  cross (*PRUNE).  In simple cases, the use of (*PRUNE)
5211         is just an alternative to an atomic group or possessive quantifier, but         is just an alternative to an atomic group or possessive quantifier, but
5212         there are some uses of (*PRUNE) that cannot be expressed in  any  other         there  are  some uses of (*PRUNE) that cannot be expressed in any other
5213         way.         way.
5214    
5215           (*SKIP)           (*SKIP)
5216    
5217         This  verb  is like (*PRUNE), except that if the pattern is unanchored,         This verb is like (*PRUNE), except that if the pattern  is  unanchored,
5218         the "bumpalong" advance is not to the next character, but to the  posi-         the  "bumpalong" advance is not to the next character, but to the posi-
5219         tion  in  the  subject where (*SKIP) was encountered. (*SKIP) signifies         tion in the subject where (*SKIP) was  encountered.  (*SKIP)  signifies
5220         that whatever text was matched leading up to it cannot  be  part  of  a         that  whatever  text  was  matched leading up to it cannot be part of a
5221         successful match. Consider:         successful match. Consider:
5222    
5223           a+(*SKIP)b           a+(*SKIP)b
5224    
5225         If  the  subject  is  "aaaac...",  after  the first match attempt fails         If the subject is "aaaac...",  after  the  first  match  attempt  fails
5226         (starting at the first character in the  string),  the  starting  point         (starting  at  the  first  character in the string), the starting point
5227         skips on to start the next attempt at "c". Note that a possessive quan-         skips on to start the next attempt at "c". Note that a possessive quan-
5228         tifer does not have the same effect as this example; although it  would         tifer  does not have the same effect as this example; although it would
5229         suppress  backtracking  during  the  first  match  attempt,  the second         suppress backtracking  during  the  first  match  attempt,  the  second
5230         attempt would start at the second character instead of skipping  on  to         attempt  would  start at the second character instead of skipping on to
5231         "c".         "c".
5232    
5233           (*THEN)           (*THEN)
5234    
5235         This verb causes a skip to the next alternation if the rest of the pat-         This verb causes a skip to the next alternation if the rest of the pat-
5236         tern does not match. That is, it cancels pending backtracking, but only         tern does not match. That is, it cancels pending backtracking, but only
5237         within  the  current  alternation.  Its name comes from the observation         within the current alternation. Its name  comes  from  the  observation
5238         that it can be used for a pattern-based if-then-else block:         that it can be used for a pattern-based if-then-else block:
5239    
5240           ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...           ( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
5241    
5242         If the COND1 pattern matches, FOO is tried (and possibly further  items         If  the COND1 pattern matches, FOO is tried (and possibly further items
5243         after  the  end  of  the group if FOO succeeds); on failure the matcher         after the end of the group if FOO succeeds);  on  failure  the  matcher
5244         skips to the second alternative and tries COND2,  without  backtracking         skips  to  the second alternative and tries COND2, without backtracking
5245         into  COND1.  If  (*THEN)  is  used outside of any alternation, it acts         into COND1. If (*THEN) is used outside  of  any  alternation,  it  acts
5246         exactly like (*PRUNE).         exactly like (*PRUNE).
5247    
5248    
# Line 5251  AUTHOR Line 5260  AUTHOR
5260    
5261  REVISION  REVISION
5262    
5263         Last updated: 11 January 2010         Last updated: 06 March 2010
5264         Copyright (c) 1997-2010 University of Cambridge.         Copyright (c) 1997-2010 University of Cambridge.
5265  ------------------------------------------------------------------------------  ------------------------------------------------------------------------------
5266    
# Line 5363  GENERAL CATEGORY PROPERTY CODES FOR \p a Line 5372  GENERAL CATEGORY PROPERTY CODES FOR \p a
5372    
5373  SCRIPT NAMES FOR \p AND \P  SCRIPT NAMES FOR \p AND \P
5374    
5375         Arabic,  Armenian,  Balinese,  Bengali,  Bopomofo,  Braille,  Buginese,         Arabic, Armenian, Avestan, Balinese, Bamum, Bengali, Bopomofo, Braille,
5376         Buhid, Canadian_Aboriginal, Carian, Cham, Cherokee, Common, Coptic, Cu-         Buginese, Buhid, Canadian_Aboriginal, Carian, Cham,  Cherokee,  Common,
5377         neiform,  Cypriot,  Cyrillic,  Deseret, Devanagari, Ethiopic, Georgian,         Coptic,   Cuneiform,  Cypriot,  Cyrillic,  Deseret,  Devanagari,  Egyp-
5378         Glagolitic, Gothic, Greek, Gujarati, Gurmukhi,  Han,  Hangul,  Hanunoo,         tian_Hieroglyphs,  Ethiopic,  Georgian,  Glagolitic,   Gothic,   Greek,
5379         Hebrew,  Hiragana,  Inherited, Kannada, Katakana, Kayah_Li, Kharoshthi,         Gujarati,  Gurmukhi,  Han,  Hangul,  Hanunoo,  Hebrew,  Hiragana, Impe-
5380         Khmer, Lao, Latin, Lepcha, Limbu, Linear_B, Lycian, Lydian,  Malayalam,         rial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscriptional_Parthian,
5381         Mongolian,  Myanmar,  New_Tai_Lue, Nko, Ogham, Old_Italic, Old_Persian,         Javanese,  Kaithi, Kannada, Katakana, Kayah_Li, Kharoshthi, Khmer, Lao,
5382         Ol_Chiki, Oriya, Osmanya, Phags_Pa, Phoenician, Rejang, Runic, Saurash-         Latin,  Lepcha,  Limbu,  Linear_B,  Lisu,  Lycian,  Lydian,  Malayalam,
5383         tra,  Shavian,  Sinhala,  Sudanese, Syloti_Nagri, Syriac, Tagalog, Tag-         Meetei_Mayek,  Mongolian, Myanmar, New_Tai_Lue, Nko, Ogham, Old_Italic,
5384         banwa,  Tai_Le,  Tamil,  Telugu,  Thaana,  Thai,   Tibetan,   Tifinagh,         Old_Persian, Old_South_Arabian, Old_Turkic, Ol_Chiki,  Oriya,  Osmanya,
5385           Phags_Pa,  Phoenician,  Rejang,  Runic, Samaritan, Saurashtra, Shavian,
5386           Sinhala, Sundanese, Syloti_Nagri, Syriac,  Tagalog,  Tagbanwa,  Tai_Le,
5387           Tai_Tham,  Tai_Viet,  Tamil,  Telugu,  Thaana, Thai, Tibetan, Tifinagh,
5388         Ugaritic, Vai, Yi.         Ugaritic, Vai, Yi.
5389    
5390    
# Line 5552  BACKTRACKING CONTROL Line 5564  BACKTRACKING CONTROL
5564           (*ACCEPT)       force successful match           (*ACCEPT)       force successful match
5565           (*FAIL)         force backtrack; synonym (*F)           (*FAIL)         force backtrack; synonym (*F)
5566    
5567         The  following  act only when a subsequent match failure causes a back-         The following act only when a subsequent match failure causes  a  back-
5568         track to reach them. They all force a match failure, but they differ in         track to reach them. They all force a match failure, but they differ in
5569         what happens afterwards. Those that advance the start-of-match point do         what happens afterwards. Those that advance the start-of-match point do
5570         so only if the pattern is not anchored.         so only if the pattern is not anchored.
# Line 5565  BACKTRACKING CONTROL Line 5577  BACKTRACKING CONTROL
5577    
5578  NEWLINE CONVENTIONS  NEWLINE CONVENTIONS
5579    
5580         These are recognized only at the very start of the pattern or  after  a         These  are  recognized only at the very start of the pattern or after a
5581         (*BSR_...) or (*UTF8) option.         (*BSR_...) or (*UTF8) option.
5582    
5583           (*CR)           carriage return only           (*CR)           carriage return only
# Line 5577  NEWLINE CONVENTIONS Line 5589  NEWLINE CONVENTIONS
5589    
5590  WHAT \R MATCHES  WHAT \R MATCHES
5591    
5592         These  are  recognized only at the very start of the pattern or after a         These are recognized only at the very start of the pattern or  after  a
5593         (*...) option that sets the newline convention or UTF-8 mode.         (*...) option that sets the newline convention or UTF-8 mode.
5594    
5595           (*BSR_ANYCRLF)  CR, LF, or CRLF           (*BSR_ANYCRLF)  CR, LF, or CRLF
# Line 5604  AUTHOR Line 5616  AUTHOR
5616    
5617  REVISION  REVISION
5618    
5619         Last updated: 11 April 2009         Last updated: 01 March 2010
5620         Copyright (c) 1997-2009 University of Cambridge.         Copyright (c) 1997-2010 University of Cambridge.
5621  ------------------------------------------------------------------------------  ------------------------------------------------------------------------------
5622    
5623    
# Line 6129  PCRE PERFORMANCE Line 6141  PCRE PERFORMANCE
6141         can affect both of them.         can affect both of them.
6142    
6143    
6144  MEMORY USAGE  COMPILED PATTERN MEMORY USAGE
6145    
6146         Patterns are compiled by PCRE into a reasonably efficient byte code, so         Patterns are compiled by PCRE into a reasonably efficient byte code, so
6147         that most simple patterns do not use much memory. However, there is one         that most simple patterns do not use much memory. However, there is one
6148         case where memory usage can be unexpectedly large. When a parenthesized         case  where  the memory usage of a compiled pattern can be unexpectedly
6149         subpattern has a quantifier with a minimum greater than 1 and/or a lim-         large. If a parenthesized subpattern has a quantifier  with  a  minimum
6150         ited  maximum,  the  whole subpattern is repeated in the compiled code.         greater  than  1  and/or  a  limited  maximum,  the whole subpattern is
6151         For example, the pattern         repeated in the compiled code. For example, the pattern
6152    
6153           (abc|def){2,4}           (abc|def){2,4}
6154    
# Line 6178  MEMORY USAGE Line 6190  MEMORY USAGE
6190         otherwise handle.         otherwise handle.
6191    
6192    
6193    STACK USAGE AT RUN TIME
6194    
6195           When  pcre_exec()  is  used  for matching, certain kinds of pattern can
6196           cause it to use large amounts of the process stack.  In  some  environ-
6197           ments  the default process stack is quite small, and if it runs out the
6198           result is often SIGSEGV.  This issue is probably  the  most  frequently
6199           raised  problem  with  PCRE. Rewriting your pattern can often help. The
6200           pcrestack documentation discusses this issue in detail.
6201    
6202    
6203  PROCESSING TIME  PROCESSING TIME
6204    
6205         Certain  items  in regular expression patterns are processed more effi-         Certain items in regular expression patterns are processed  more  effi-
6206         ciently than others. It is more efficient to use a character class like         ciently than others. It is more efficient to use a character class like
6207         [aeiou]   than   a   set   of  single-character  alternatives  such  as         [aeiou]  than  a  set  of   single-character   alternatives   such   as
6208         (a|e|i|o|u). In general, the simplest construction  that  provides  the         (a|e|i|o|u).  In  general,  the simplest construction that provides the
6209         required behaviour is usually the most efficient. Jeffrey Friedl's book         required behaviour is usually the most efficient. Jeffrey Friedl's book
6210         contains a lot of useful general discussion  about  optimizing  regular         contains  a  lot  of useful general discussion about optimizing regular
6211         expressions  for  efficient  performance.  This document contains a few         expressions for efficient performance. This  document  contains  a  few
6212         observations about PCRE.         observations about PCRE.
6213    
6214         Using Unicode character properties (the \p,  \P,  and  \X  escapes)  is         Using  Unicode  character  properties  (the  \p, \P, and \X escapes) is
6215         slow,  because PCRE has to scan a structure that contains data for over         slow, because PCRE has to scan a structure that contains data for  over
6216         fifteen thousand characters whenever it needs a  character's  property.         fifteen  thousand  characters whenever it needs a character's property.
6217         If  you  can  find  an  alternative pattern that does not use character         If you can find an alternative pattern  that  does  not  use  character
6218         properties, it will probably be faster.         properties, it will probably be faster.
6219    
6220         When a pattern begins with .* not in  parentheses,  or  in  parentheses         When  a  pattern  begins  with .* not in parentheses, or in parentheses
6221         that are not the subject of a backreference, and the PCRE_DOTALL option         that are not the subject of a backreference, and the PCRE_DOTALL option
6222         is set, the pattern is implicitly anchored by PCRE, since it can  match         is  set, the pattern is implicitly anchored by PCRE, since it can match
6223         only  at  the start of a subject string. However, if PCRE_DOTALL is not         only at the start of a subject string. However, if PCRE_DOTALL  is  not
6224         set, PCRE cannot make this optimization, because  the  .  metacharacter         set,  PCRE  cannot  make this optimization, because the . metacharacter
6225         does  not then match a newline, and if the subject string contains new-         does not then match a newline, and if the subject string contains  new-
6226         lines, the pattern may match from the character  immediately  following         lines,  the  pattern may match from the character immediately following
6227         one of them instead of from the very start. For example, the pattern         one of them instead of from the very start. For example, the pattern
6228    
6229           .*second           .*second
6230    
6231         matches  the subject "first\nand second" (where \n stands for a newline         matches the subject "first\nand second" (where \n stands for a  newline
6232         character), with the match starting at the seventh character. In  order         character),  with the match starting at the seventh character. In order
6233         to do this, PCRE has to retry the match starting after every newline in         to do this, PCRE has to retry the match starting after every newline in
6234         the subject.         the subject.
6235    
6236         If you are using such a pattern with subject strings that do  not  con-         If  you  are using such a pattern with subject strings that do not con-
6237         tain newlines, the best performance is obtained by setting PCRE_DOTALL,         tain newlines, the best performance is obtained by setting PCRE_DOTALL,
6238         or starting the pattern with ^.* or ^.*? to indicate  explicit  anchor-         or  starting  the pattern with ^.* or ^.*? to indicate explicit anchor-
6239         ing.  That saves PCRE from having to scan along the subject looking for         ing. That saves PCRE from having to scan along the subject looking  for
6240         a newline to restart at.         a newline to restart at.
6241    
6242         Beware of patterns that contain nested indefinite  repeats.  These  can         Beware  of  patterns  that contain nested indefinite repeats. These can
6243         take  a  long time to run when applied to a string that does not match.         take a long time to run when applied to a string that does  not  match.
6244         Consider the pattern fragment         Consider the pattern fragment
6245    
6246           ^(a+)*           ^(a+)*
6247    
6248         This can match "aaaa" in 16 different ways, and this  number  increases         This  can  match "aaaa" in 16 different ways, and this number increases
6249         very  rapidly  as the string gets longer. (The * repeat can match 0, 1,         very rapidly as the string gets longer. (The * repeat can match  0,  1,
6250         2, 3, or 4 times, and for each of those cases other than 0 or 4, the  +         2,  3, or 4 times, and for each of those cases other than 0 or 4, the +
6251         repeats  can  match  different numbers of times.) When the remainder of         repeats can match different numbers of times.) When  the  remainder  of
6252         the pattern is such that the entire match is going to fail, PCRE has in         the pattern is such that the entire match is going to fail, PCRE has in
6253         principle  to  try  every  possible  variation,  and  this  can take an         principle to try  every  possible  variation,  and  this  can  take  an
6254         extremely long time, even for relatively short strings.         extremely long time, even for relatively short strings.
6255    
6256         An optimization catches some of the more simple cases such as         An optimization catches some of the more simple cases such as
6257    
6258           (a+)*b           (a+)*b
6259    
6260         where a literal character follows. Before  embarking  on  the  standard         where  a  literal  character  follows. Before embarking on the standard
6261         matching  procedure,  PCRE checks that there is a "b" later in the sub-         matching procedure, PCRE checks that there is a "b" later in  the  sub-
6262         ject string, and if there is not, it fails the match immediately.  How-         ject  string, and if there is not, it fails the match immediately. How-
6263         ever,  when  there  is no following literal this optimization cannot be         ever, when there is no following literal this  optimization  cannot  be
6264         used. You can see the difference by comparing the behaviour of         used. You can see the difference by comparing the behaviour of
6265    
6266           (a+)*\d           (a+)*\d
6267    
6268         with the pattern above. The former gives  a  failure  almost  instantly         with  the  pattern  above.  The former gives a failure almost instantly
6269         when  applied  to  a  whole  line of "a" characters, whereas the latter         when applied to a whole line of  "a"  characters,  whereas  the  latter
6270         takes an appreciable time with strings longer than about 20 characters.         takes an appreciable time with strings longer than about 20 characters.
6271    
6272         In many cases, the solution to this kind of performance issue is to use         In many cases, the solution to this kind of performance issue is to use
# Line 6260  AUTHOR Line 6282  AUTHOR
6282    
6283  REVISION  REVISION
6284    
6285         Last updated: 06 March 2007         Last updated: 07 March 2010
6286         Copyright (c) 1997-2007 University of Cambridge.         Copyright (c) 1997-2010 University of Cambridge.
6287  ------------------------------------------------------------------------------  ------------------------------------------------------------------------------
6288    
6289    

Legend:
Removed from v.506  
changed lines
  Added in v.507

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12