xref: /freebsd/lib/libc/regex/re_format.7 (revision 3e5645b78f476816ca3b5acc28b29bbafbb9c444)
1.\" Copyright (c) 1992, 1993, 1994 Henry Spencer.
2.\" Copyright (c) 1992, 1993, 1994
3.\"	The Regents of the University of California.  All rights reserved.
4.\"
5.\" This code is derived from software contributed to Berkeley by
6.\" Henry Spencer.
7.\"
8.\" Redistribution and use in source and binary forms, with or without
9.\" modification, are permitted provided that the following conditions
10.\" are met:
11.\" 1. Redistributions of source code must retain the above copyright
12.\"    notice, this list of conditions and the following disclaimer.
13.\" 2. Redistributions in binary form must reproduce the above copyright
14.\"    notice, this list of conditions and the following disclaimer in the
15.\"    documentation and/or other materials provided with the distribution.
16.\" 3. All advertising materials mentioning features or use of this software
17.\"    must display the following acknowledgement:
18.\"	This product includes software developed by the University of
19.\"	California, Berkeley and its contributors.
20.\" 4. Neither the name of the University nor the names of its contributors
21.\"    may be used to endorse or promote products derived from this software
22.\"    without specific prior written permission.
23.\"
24.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34.\" SUCH DAMAGE.
35.\"
36.\"	@(#)re_format.7	8.3 (Berkeley) 3/20/94
37.\" $FreeBSD$
38.\"
39.Dd June 30, 2014
40.Dt RE_FORMAT 7
41.Os
42.Sh NAME
43.Nm re_format
44.Nd POSIX 1003.2 regular expressions
45.Sh DESCRIPTION
46Regular expressions
47.Pq Dq RE Ns s ,
48as defined in
49.St -p1003.2 ,
50come in two forms:
51modern REs (roughly those of
52.Xr egrep 1 ;
531003.2 calls these
54.Dq extended
55REs)
56and obsolete REs (roughly those of
57.Xr ed 1 ;
581003.2
59.Dq basic
60REs).
61Obsolete REs mostly exist for backward compatibility in some old programs;
62they will be discussed at the end.
63.St -p1003.2
64leaves some aspects of RE syntax and semantics open;
65`\(dd' marks decisions on these aspects that
66may not be fully portable to other
67.St -p1003.2
68implementations.
69.Pp
70A (modern) RE is one\(dd or more non-empty\(dd
71.Em branches ,
72separated by
73.Ql \&| .
74It matches anything that matches one of the branches.
75.Pp
76A branch is one\(dd or more
77.Em pieces ,
78concatenated.
79It matches a match for the first, followed by a match for the second, etc.
80.Pp
81A piece is an
82.Em atom
83possibly followed
84by a single\(dd
85.Ql \&* ,
86.Ql \&+ ,
87.Ql \&? ,
88or
89.Em bound .
90An atom followed by
91.Ql \&*
92matches a sequence of 0 or more matches of the atom.
93An atom followed by
94.Ql \&+
95matches a sequence of 1 or more matches of the atom.
96An atom followed by
97.Ql ?\&
98matches a sequence of 0 or 1 matches of the atom.
99.Pp
100A
101.Em bound
102is
103.Ql \&{
104followed by an unsigned decimal integer,
105possibly followed by
106.Ql \&,
107possibly followed by another unsigned decimal integer,
108always followed by
109.Ql \&} .
110The integers must lie between 0 and
111.Dv RE_DUP_MAX
112(255\(dd) inclusive,
113and if there are two of them, the first may not exceed the second.
114An atom followed by a bound containing one integer
115.Em i
116and no comma matches
117a sequence of exactly
118.Em i
119matches of the atom.
120An atom followed by a bound
121containing one integer
122.Em i
123and a comma matches
124a sequence of
125.Em i
126or more matches of the atom.
127An atom followed by a bound
128containing two integers
129.Em i
130and
131.Em j
132matches
133a sequence of
134.Em i
135through
136.Em j
137(inclusive) matches of the atom.
138.Pp
139An atom is a regular expression enclosed in
140.Ql ()
141(matching a match for the
142regular expression),
143an empty set of
144.Ql ()
145(matching the null string)\(dd,
146a
147.Em bracket expression
148(see below),
149.Ql .\&
150(matching any single character),
151.Ql \&^
152(matching the null string at the beginning of a line),
153.Ql \&$
154(matching the null string at the end of a line), a
155.Ql \e
156followed by one of the characters
157.Ql ^.[$()|*+?{\e
158(matching that character taken as an ordinary character),
159a
160.Ql \e
161followed by any other character\(dd
162(matching that character taken as an ordinary character,
163as if the
164.Ql \e
165had not been present\(dd),
166or a single character with no other significance (matching that character).
167A
168.Ql \&{
169followed by a character other than a digit is an ordinary
170character, not the beginning of a bound\(dd.
171It is illegal to end an RE with
172.Ql \e .
173.Pp
174A
175.Em bracket expression
176is a list of characters enclosed in
177.Ql [] .
178It normally matches any single character from the list (but see below).
179If the list begins with
180.Ql \&^ ,
181it matches any single character
182(but see below)
183.Em not
184from the rest of the list.
185If two characters in the list are separated by
186.Ql \&- ,
187this is shorthand
188for the full
189.Em range
190of characters between those two (inclusive) in the
191collating sequence,
192.No e.g. Ql [0-9]
193in ASCII matches any decimal digit.
194It is illegal\(dd for two ranges to share an
195endpoint,
196.No e.g. Ql a-c-e .
197Ranges are very collating-sequence-dependent,
198and portable programs should avoid relying on them.
199.Pp
200To include a literal
201.Ql \&]
202in the list, make it the first character
203(following a possible
204.Ql \&^ ) .
205To include a literal
206.Ql \&- ,
207make it the first or last character,
208or the second endpoint of a range.
209To use a literal
210.Ql \&-
211as the first endpoint of a range,
212enclose it in
213.Ql [.\&
214and
215.Ql .]\&
216to make it a collating element (see below).
217With the exception of these and some combinations using
218.Ql \&[
219(see next paragraphs), all other special characters, including
220.Ql \e ,
221lose their special significance within a bracket expression.
222.Pp
223Within a bracket expression, a collating element (a character,
224a multi-character sequence that collates as if it were a single character,
225or a collating-sequence name for either)
226enclosed in
227.Ql [.\&
228and
229.Ql .]\&
230stands for the
231sequence of characters of that collating element.
232The sequence is a single element of the bracket expression's list.
233A bracket expression containing a multi-character collating element
234can thus match more than one character,
235e.g.\& if the collating sequence includes a
236.Ql ch
237collating element,
238then the RE
239.Ql [[.ch.]]*c
240matches the first five characters
241of
242.Ql chchcc .
243.Pp
244Within a bracket expression, a collating element enclosed in
245.Ql [=
246and
247.Ql =]
248is an equivalence class, standing for the sequences of characters
249of all collating elements equivalent to that one, including itself.
250(If there are no other equivalent collating elements,
251the treatment is as if the enclosing delimiters were
252.Ql [.\&
253and
254.Ql .] . )
255For example, if
256.Ql x
257and
258.Ql y
259are the members of an equivalence class,
260then
261.Ql [[=x=]] ,
262.Ql [[=y=]] ,
263and
264.Ql [xy]
265are all synonymous.
266An equivalence class may not\(dd be an endpoint
267of a range.
268.Pp
269Within a bracket expression, the name of a
270.Em character class
271enclosed in
272.Ql [:
273and
274.Ql :]
275stands for the list of all characters belonging to that
276class.
277Standard character class names are:
278.Bl -column "alnum" "digit" "xdigit" -offset indent
279.It Em "alnum	digit	punct"
280.It Em "alpha	graph	space"
281.It Em "blank	lower	upper"
282.It Em "cntrl	print	xdigit"
283.El
284.Pp
285These stand for the character classes defined in
286.Xr ctype 3 .
287A locale may provide others.
288A character class may not be used as an endpoint of a range.
289.Pp
290A bracketed expression like
291.Ql [[:class:]]
292can be used to match a single character that belongs to a character
293class.
294The reverse, matching any character that does not belong to a specific
295class, the negation operator of bracket expressions may be used:
296.Ql [^[:class:]] .
297.Pp
298There are two special cases\(dd of bracket expressions:
299the bracket expressions
300.Ql [[:<:]]
301and
302.Ql [[:>:]]
303match the null string at the beginning and end of a word respectively.
304A word is defined as a sequence of word characters
305which is neither preceded nor followed by
306word characters.
307A word character is an
308.Em alnum
309character (as defined by
310.Xr ctype 3 )
311or an underscore.
312This is an extension,
313compatible with but not specified by
314.St -p1003.2 ,
315and should be used with
316caution in software intended to be portable to other systems.
317The additional word delimiters
318.Ql \e<
319and
320.Ql \e>
321are provided to ease compatibility with traditional
322.Xr svr4 4
323systems but are not portable and should be avoided.
324.Pp
325In the event that an RE could match more than one substring of a given
326string,
327the RE matches the one starting earliest in the string.
328If the RE could match more than one substring starting at that point,
329it matches the longest.
330Subexpressions also match the longest possible substrings, subject to
331the constraint that the whole match be as long as possible,
332with subexpressions starting earlier in the RE taking priority over
333ones starting later.
334Note that higher-level subexpressions thus take priority over
335their lower-level component subexpressions.
336.Pp
337Match lengths are measured in characters, not collating elements.
338A null string is considered longer than no match at all.
339For example,
340.Ql bb*
341matches the three middle characters of
342.Ql abbbc ,
343.Ql (wee|week)(knights|nights)
344matches all ten characters of
345.Ql weeknights ,
346when
347.Ql (.*).*\&
348is matched against
349.Ql abc
350the parenthesized subexpression
351matches all three characters, and
352when
353.Ql (a*)*
354is matched against
355.Ql bc
356both the whole RE and the parenthesized
357subexpression match the null string.
358.Pp
359If case-independent matching is specified,
360the effect is much as if all case distinctions had vanished from the
361alphabet.
362When an alphabetic that exists in multiple cases appears as an
363ordinary character outside a bracket expression, it is effectively
364transformed into a bracket expression containing both cases,
365.No e.g. Ql x
366becomes
367.Ql [xX] .
368When it appears inside a bracket expression, all case counterparts
369of it are added to the bracket expression, so that (e.g.)
370.Ql [x]
371becomes
372.Ql [xX]
373and
374.Ql [^x]
375becomes
376.Ql [^xX] .
377.Pp
378No particular limit is imposed on the length of REs\(dd.
379Programs intended to be portable should not employ REs longer
380than 256 bytes,
381as an implementation can refuse to accept such REs and remain
382POSIX-compliant.
383.Pp
384Obsolete
385.Pq Dq basic
386regular expressions differ in several respects.
387.Ql \&|
388is an ordinary character and there is no equivalent
389for its functionality.
390.Ql \&+
391and
392.Ql ?\&
393are ordinary characters, and their functionality
394can be expressed using bounds
395.Po
396.Ql {1,}
397or
398.Ql {0,1}
399respectively
400.Pc .
401Also note that
402.Ql x+
403in modern REs is equivalent to
404.Ql xx* .
405The delimiters for bounds are
406.Ql \e{
407and
408.Ql \e} ,
409with
410.Ql \&{
411and
412.Ql \&}
413by themselves ordinary characters.
414The parentheses for nested subexpressions are
415.Ql \e(
416and
417.Ql \e) ,
418with
419.Ql \&(
420and
421.Ql \&)
422by themselves ordinary characters.
423.Ql \&^
424is an ordinary character except at the beginning of the
425RE or\(dd the beginning of a parenthesized subexpression,
426.Ql \&$
427is an ordinary character except at the end of the
428RE or\(dd the end of a parenthesized subexpression,
429and
430.Ql \&*
431is an ordinary character if it appears at the beginning of the
432RE or the beginning of a parenthesized subexpression
433(after a possible leading
434.Ql \&^ ) .
435Finally, there is one new type of atom, a
436.Em back reference :
437.Ql \e
438followed by a non-zero decimal digit
439.Em d
440matches the same sequence of characters
441matched by the
442.Em d Ns th
443parenthesized subexpression
444(numbering subexpressions by the positions of their opening parentheses,
445left to right),
446so that (e.g.)
447.Ql \e([bc]\e)\e1
448matches
449.Ql bb
450or
451.Ql cc
452but not
453.Ql bc .
454.Sh SEE ALSO
455.Xr regex 3
456.Rs
457.%T Regular Expression Notation
458.%R IEEE Std
459.%N 1003.2
460.%P section 2.8
461.Re
462.Sh BUGS
463Having two kinds of REs is a botch.
464.Pp
465The current
466.St -p1003.2
467spec says that
468.Ql \&)
469is an ordinary character in
470the absence of an unmatched
471.Ql \&( ;
472this was an unintentional result of a wording error,
473and change is likely.
474Avoid relying on it.
475.Pp
476Back references are a dreadful botch,
477posing major problems for efficient implementations.
478They are also somewhat vaguely defined
479(does
480.Ql a\e(\e(b\e)*\e2\e)*d
481match
482.Ql abbbd ? ) .
483Avoid using them.
484.Pp
485.St -p1003.2
486specification of case-independent matching is vague.
487The
488.Dq one case implies all cases
489definition given above
490is current consensus among implementors as to the right interpretation.
491.Pp
492The syntax for word boundaries is incredibly ugly.
493