1<html> 2<head> 3<title>pcrecompat specification</title> 4</head> 5<body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB"> 6<h1>pcrecompat man page</h1> 7<p> 8Return to the <a href="index.html">PCRE index page</a>. 9</p> 10<p> 11This page is part of the PCRE HTML documentation. It was generated automatically 12from the original man page. If there is any nonsense in it, please consult the 13man page, in case the conversion went wrong. 14<br> 15<br><b> 16DIFFERENCES BETWEEN PCRE AND PERL 17</b><br> 18<P> 19This document describes the differences in the ways that PCRE and Perl handle 20regular expressions. The differences described here are with respect to Perl 21versions 5.10 and above. 22</P> 23<P> 241. PCRE has only a subset of Perl's Unicode support. Details of what it does 25have are given in the 26<a href="pcreunicode.html"><b>pcreunicode</b></a> 27page. 28</P> 29<P> 302. PCRE allows repeat quantifiers only on parenthesized assertions, but they do 31not mean what you might think. For example, (?!a){3} does not assert that the 32next three characters are not "a". It just asserts that the next character is 33not "a" three times (in principle: PCRE optimizes this to run the assertion 34just once). Perl allows repeat quantifiers on other assertions such as \b, but 35these do not seem to have any use. 36</P> 37<P> 383. Capturing subpatterns that occur inside negative lookahead assertions are 39counted, but their entries in the offsets vector are never set. Perl sets its 40numerical variables from any such patterns that are matched before the 41assertion fails to match something (thereby succeeding), but only if the 42negative lookahead assertion contains just one branch. 43</P> 44<P> 454. Though binary zero characters are supported in the subject string, they are 46not allowed in a pattern string because it is passed as a normal C string, 47terminated by zero. The escape sequence \0 can be used in the pattern to 48represent a binary zero. 49</P> 50<P> 515. The following Perl escape sequences are not supported: \l, \u, \L, 52\U, and \N when followed by a character name or Unicode value. (\N on its 53own, matching a non-newline character, is supported.) In fact these are 54implemented by Perl's general string-handling and are not part of its pattern 55matching engine. If any of these are encountered by PCRE, an error is 56generated by default. However, if the PCRE_JAVASCRIPT_COMPAT option is set, 57\U and \u are interpreted as JavaScript interprets them. 58</P> 59<P> 606. The Perl escape sequences \p, \P, and \X are supported only if PCRE is 61built with Unicode character property support. The properties that can be 62tested with \p and \P are limited to the general category properties such as 63Lu and Nd, script names such as Greek or Han, and the derived properties Any 64and L&. PCRE does support the Cs (surrogate) property, which Perl does not; the 65Perl documentation says "Because Perl hides the need for the user to understand 66the internal representation of Unicode characters, there is no need to 67implement the somewhat messy concept of surrogates." 68</P> 69<P> 707. PCRE implements a simpler version of \X than Perl, which changed to make 71\X match what Unicode calls an "extended grapheme cluster". This is more 72complicated than an extended Unicode sequence, which is what PCRE matches. 73</P> 74<P> 758. PCRE does support the \Q...\E escape for quoting substrings. Characters in 76between are treated as literals. This is slightly different from Perl in that $ 77and @ are also handled as literals inside the quotes. In Perl, they cause 78variable interpolation (but of course PCRE does not have variables). Note the 79following examples: 80<pre> 81 Pattern PCRE matches Perl matches 82 83 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz 84 \Qabc\$xyz\E abc\$xyz abc\$xyz 85 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz 86</pre> 87The \Q...\E sequence is recognized both inside and outside character classes. 88</P> 89<P> 909. Fairly obviously, PCRE does not support the (?{code}) and (??{code}) 91constructions. However, there is support for recursive patterns. This is not 92available in Perl 5.8, but it is in Perl 5.10. Also, the PCRE "callout" 93feature allows an external function to be called during pattern matching. See 94the 95<a href="pcrecallout.html"><b>pcrecallout</b></a> 96documentation for details. 97</P> 98<P> 9910. Subpatterns that are called as subroutines (whether or not recursively) are 100always treated as atomic groups in PCRE. This is like Python, but unlike Perl. 101Captured values that are set outside a subroutine call can be reference from 102inside in PCRE, but not in Perl. There is a discussion that explains these 103differences in more detail in the 104<a href="pcrepattern.html#recursiondifference">section on recursion differences from Perl</a> 105in the 106<a href="pcrepattern.html"><b>pcrepattern</b></a> 107page. 108</P> 109<P> 11011. If any of the backtracking control verbs are used in an assertion or in a 111subpattern that is called as a subroutine (whether or not recursively), their 112effect is confined to that subpattern; it does not extend to the surrounding 113pattern. This is not always the case in Perl. In particular, if (*THEN) is 114present in a group that is called as a subroutine, its action is limited to 115that group, even if the group does not contain any | characters. There is one 116exception to this: the name from a *(MARK), (*PRUNE), or (*THEN) that is 117encountered in a successful positive assertion <i>is</i> passed back when a 118match succeeds (compare capturing parentheses in assertions). Note that such 119subpatterns are processed as anchored at the point where they are tested. 120</P> 121<P> 12212. There are some differences that are concerned with the settings of captured 123strings when part of a pattern is repeated. For example, matching "aba" against 124the pattern /^(a(b)?)+$/ in Perl leaves $2 unset, but in PCRE it is set to "b". 125</P> 126<P> 12713. PCRE's handling of duplicate subpattern numbers and duplicate subpattern 128names is not as general as Perl's. This is a consequence of the fact the PCRE 129works internally just with numbers, using an external table to translate 130between numbers and names. In particular, a pattern such as (?|(?<a>A)|(?<b)B), 131where the two capturing parentheses have the same number but different names, 132is not supported, and causes an error at compile time. If it were allowed, it 133would not be possible to distinguish which parentheses matched, because both 134names map to capturing subpattern number 1. To avoid this confusing situation, 135an error is given at compile time. 136</P> 137<P> 13814. Perl recognizes comments in some places that PCRE does not, for example, 139between the ( and ? at the start of a subpattern. If the /x modifier is set, 140Perl allows white space between ( and ? but PCRE never does, even if the 141PCRE_EXTENDED option is set. 142</P> 143<P> 14415. PCRE provides some extensions to the Perl regular expression facilities. 145Perl 5.10 includes new features that are not in earlier versions of Perl, some 146of which (such as named parentheses) have been in PCRE for some time. This list 147is with respect to Perl 5.10: 148<br> 149<br> 150(a) Although lookbehind assertions in PCRE must match fixed length strings, 151each alternative branch of a lookbehind assertion can match a different length 152of string. Perl requires them all to have the same length. 153<br> 154<br> 155(b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $ 156meta-character matches only at the very end of the string. 157<br> 158<br> 159(c) If PCRE_EXTRA is set, a backslash followed by a letter with no special 160meaning is faulted. Otherwise, like Perl, the backslash is quietly ignored. 161(Perl can be made to issue a warning.) 162<br> 163<br> 164(d) If PCRE_UNGREEDY is set, the greediness of the repetition quantifiers is 165inverted, that is, by default they are not greedy, but if followed by a 166question mark they are. 167<br> 168<br> 169(e) PCRE_ANCHORED can be used at matching time to force a pattern to be tried 170only at the first matching position in the subject string. 171<br> 172<br> 173(f) The PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART, and 174PCRE_NO_AUTO_CAPTURE options for <b>pcre_exec()</b> have no Perl equivalents. 175<br> 176<br> 177(g) The \R escape sequence can be restricted to match only CR, LF, or CRLF 178by the PCRE_BSR_ANYCRLF option. 179<br> 180<br> 181(h) The callout facility is PCRE-specific. 182<br> 183<br> 184(i) The partial matching facility is PCRE-specific. 185<br> 186<br> 187(j) Patterns compiled by PCRE can be saved and re-used at a later time, even on 188different hosts that have the other endianness. However, this does not apply to 189optimized data created by the just-in-time compiler. 190<br> 191<br> 192(k) The alternative matching functions (<b>pcre_dfa_exec()</b> and 193<b>pcre16_dfa_exec()</b>) match in a different way and are not Perl-compatible. 194<br> 195<br> 196(l) PCRE recognizes some special sequences such as (*CR) at the start of 197a pattern that set overall options that cannot be changed within the pattern. 198</P> 199<br><b> 200AUTHOR 201</b><br> 202<P> 203Philip Hazel 204<br> 205University Computing Service 206<br> 207Cambridge CB2 3QH, England. 208<br> 209</P> 210<br><b> 211REVISION 212</b><br> 213<P> 214Last updated: 01 June 2012 215<br> 216Copyright © 1997-2012 University of Cambridge. 217<br> 218<p> 219Return to the <a href="index.html">PCRE index page</a>. 220</p> 221