1REGEX(7) Linux Programmer's Manual REGEX(7)
2
6 regex - POSIX.2 regular expressions
7
9 Regular expressions ("RE"s), as defined in POSIX.2, come in two forms:
10 modern REs (roughly those of egrep; POSIX.2 calls these "extended" REs)
11 and obsolete REs (roughly those of ed(1); POSIX.2 "basic" REs). Obso‐
12 lete REs mostly exist for backward compatibility in some old programs;
13 they will be discussed at the end. POSIX.2 leaves some aspects of RE
14 syntax and semantics open; "(!)" marks decisions on these aspects that
15 may not be fully portable to other POSIX.2 implementations.
16 A (modern) RE is one(!) or more nonempty(!) branches, separated by '|'.
18 It matches anything that matches one of the branches.
19 A branch is one(!) or more pieces, concatenated. It matches a match
21 for the first, followed by a match for the second, and so on.
22 A piece is an atom possibly followed by a single(!) '*', '+', '?', or
24 bound. An atom followed by '*' matches a sequence of 0 or more matches
25 of the atom. An atom followed by '+' matches a sequence of 1 or more
26 matches of the atom. An atom followed by '?' matches a sequence of 0
27 or 1 matches of the atom.
28 A bound is '{' followed by an unsigned decimal integer, possibly fol‐
30 lowed by ',' possibly followed by another unsigned decimal integer,
31 always followed by '}'. The integers must lie between 0 and RE_DUP_MAX
32 (255(!)) inclusive, and if there are two of them, the first may not
33 exceed the second. An atom followed by a bound containing one integer
34 i and no comma matches a sequence of exactly i matches of the atom. An
35 atom followed by a bound containing one integer i and a comma matches a
36 sequence of i or more matches of the atom. An atom followed by a bound
37 containing two integers i and j matches a sequence of i through j
38 (inclusive) matches of the atom.
39 An atom is a regular expression enclosed in "()" (matching a match for
41 the regular expression), an empty set of "()" (matching the null
42 string)(!), a bracket expression (see below), '.' (matching any single
43 character), '^' (matching the null string at the beginning of a line),
44 '$' (matching the null string at the end of a line), a '\' followed by
45 one of the characters "^.[$()|*+?{\" (matching that character taken as
46 an ordinary character), a '\' followed by any other character(!)
47 (matching that character taken as an ordinary character, as if the '\'
48 had not been present(!)), or a single character with no other signifi‐
49 cance (matching that character). A '{' followed by a character other
50 than a digit is an ordinary character, not the beginning of a bound(!).
51 It is illegal to end an RE with '\'.
52 A bracket expression is a list of characters enclosed in "[]". It nor‐
54 mally matches any single character from the list (but see below). If
55 the list begins with '^', it matches any single character (but see
56 below) not from the rest of the list. If two characters in the list
57 are separated by '-', this is shorthand for the full range of charac‐
58 ters between those two (inclusive) in the collating sequence, for exam‐
59 ple, "[0-9]" in ASCII matches any decimal digit. It is illegal(!) for
60 two ranges to share an endpoint, for example, "a-c-e". Ranges are very
61 collating-sequence-dependent, and portable programs should avoid rely‐
62 ing on them.
63 To include a literal ']' in the list, make it the first character (fol‐
65 lowing a possible '^'). To include a literal '-', make it the first or
66 last character, or the second endpoint of a range. To use a literal
67 '-' as the first endpoint of a range, enclose it in "[." and ".]" to
68 make it a collating element (see below). With the exception of these
69 and some combinations using '[' (see next paragraphs), all other spe‐
70 cial characters, including '\', lose their special significance within
71 a bracket expression.
72 Within a bracket expression, a collating element (a character, a multi‐
74 character sequence that collates as if it were a single character, or a
75 collating-sequence name for either) enclosed in "[." and ".]" stands
76 for the sequence of characters of that collating element. The sequence
77 is a single element of the bracket expression's list. A bracket
78 expression containing a multicharacter collating element can thus match
79 more than one character, for example, if the collating sequence
80 includes a "ch" collating element, then the RE "[[.ch.]]*c" matches the
81 first five characters of "chchcc".
82 Within a bracket expression, a collating element enclosed in "[=" and
84 "=]" is an equivalence class, standing for the sequences of characters
85 of all collating elements equivalent to that one, including itself.
86 (If there are no other equivalent collating elements, the treatment is
87 as if the enclosing delimiters were "[." and ".]".) For example, if o
88 and o^ are the members of an equivalence class, then "[[=o=]]",
89 "[[=o^=]]", and "[oo^]" are all synonymous. An equivalence class may
90 not(!) be an endpoint of a range.
91 Within a bracket expression, the name of a character class enclosed in
93 "[:" and ":]" stands for the list of all characters belonging to that
94 class. Standard character class names are:
95 alnum digit punct
97 alpha graph space
98 blank lower upper
99 cntrl print xdigit
100 These stand for the character classes defined in wctype(3). A locale
102 may provide others. A character class may not be used as an endpoint
103 of a range.
104 In the event that an RE could match more than one substring of a given
106 string, the RE matches the one starting earliest in the string. If the
107 RE could match more than one substring starting at that point, it
108 matches the longest. Subexpressions also match the longest possible
109 substrings, subject to the constraint that the whole match be as long
110 as possible, with subexpressions starting earlier in the RE taking pri‐
111 ority over ones starting later. Note that higher-level subexpressions
112 thus take priority over their lower-level component subexpressions.
113 Match lengths are measured in characters, not collating elements. A
115 null string is considered longer than no match at all. For example,
116 "bb*" matches the three middle characters of "abbbc",
117 "(wee|week)(knights|nights)" matches all ten characters of "week‐
118 nights", when "(.*).*" is matched against "abc" the parenthesized sub‐
119 expression matches all three characters, and when "(a*)*" is matched
120 against "bc" both the whole RE and the parenthesized subexpression
121 match the null string.
122 If case-independent matching is specified, the effect is much as if all
124 case distinctions had vanished from the alphabet. When an alphabetic
125 that exists in multiple cases appears as an ordinary character outside
126 a bracket expression, it is effectively transformed into a bracket
127 expression containing both cases, for example, 'x' becomes "[xX]".
128 When it appears inside a bracket expression, all case counterparts of
129 it are added to the bracket expression, so that, for example, "[x]"
130 becomes "[xX]" and "[^x]" becomes "[^xX]".
131 No particular limit is imposed on the length of REs(!). Programs
133 intended to be portable should not employ REs longer than 256 bytes, as
134 an implementation can refuse to accept such REs and remain POSIX-com‐
135 pliant.
136 Obsolete ("basic") regular expressions differ in several respects.
138 '|', '+', and '?' are ordinary characters and there is no equivalent
139 for their functionality. The delimiters for bounds are "\{" and "\}",
140 with '{' and '}' by themselves ordinary characters. The parentheses
141 for nested subexpressions are "\(" and "\)", with '(' and ')' by them‐
142 selves ordinary characters. '^' is an ordinary character except at the
143 beginning of the RE or(!) the beginning of a parenthesized subexpres‐
144 sion, '$' is an ordinary character except at the end of the RE or(!)
145 the end of a parenthesized subexpression, and '*' is an ordinary char‐
146 acter if it appears at the beginning of the RE or the beginning of a
147 parenthesized subexpression (after a possible leading '^').
148 Finally, there is one new type of atom, a back reference: '\' followed
150 by a nonzero decimal digit d matches the same sequence of characters
151 matched by the dth parenthesized subexpression (numbering subexpres‐
152 sions by the positions of their opening parentheses, left to right), so
153 that, for example, "\([bc]\)\1" matches "bb" or "cc" but not "bc".
154
156 Having two kinds of REs is a botch.
157 The current POSIX.2 spec says that ')' is an ordinary character in the
159 absence of an unmatched '('; this was an unintentional result of a
160 wording error, and change is likely. Avoid relying on it.
161 Back references are a dreadful botch, posing major problems for effi‐
163 cient implementations. They are also somewhat vaguely defined (does
164 "a\(\(b\)*\2\)*d" match "abbbd"?). Avoid using them.
165 POSIX.2's specification of case-independent matching is vague. The
167 "one case implies all cases" definition given above is current consen‐
168 sus among implementors as to the right interpretation.
169
171 This page was taken from Henry Spencer's regex package.
172
174 grep(1), regex(3)
175 POSIX.2, section 2.8 (Regular Expression Notation).
177
179 This page is part of release 5.04 of the Linux man-pages project. A
180 description of the project, information about reporting bugs, and the
181 latest version of this page, can be found at
182 https://www.kernel.org/doc/man-pages/.
183 2009-01-12 REGEX(7)