REGEX(3) Library Functions Manual REGEX(3)

regcomp, regcomp_l, regerror, regexec, regfree, regncomp, regncomp_l, regnexec, regnwcomp, regnwcomp_l, regnwexec, regwcomp, regwcomp_l, regwexec — regular-expression library


#include <regex.h>

int
regcomp(regex_t *restrict preg, const char *restrict pattern, int cflags);

size_t
regerror(int errcode, const regex_t *restrict preg, char *restrict errbuf, size_t errbuf_size);

int
regexec(const regex_t *restrict preg, const char *restrict string, size_t nmatch, regmatch_t pmatch[restrict], int eflags);

void
regfree(regex_t *preg);


int
regncomp(regex_t *restrict preg, const char *restrict pattern, size_t len, int cflags);

int
regnexec(const regex_t *restrict preg, const char *restrict string, size_t len, size_t nmatch, regmatch_t pmatch[restrict], int eflags);

int
regwcomp(regex_t *restrict preg, const wchar_t *restrict widepat, int cflags);

int
regwexec(const regex_t *restrict preg, const wchar_t *restrict widestr, size_t nmatch, regmatch_t pmatch[restrict], int eflags);

int
regwncomp(regex_t *restrict preg, const wchar_t *restrict widepat, size_t len, int cflags);

int
regwnexec(const regex_t *restrict preg, const wchar_t *restrict widestr, size_t len, size_t nmatch, regmatch_t pmatch[restrict], int eflags);

#include <regex.h>
#include <xlocale.h>

int
regcomp_l(regex_t *restrict preg, const char *restrict pattern, int cflags, locale_t restrict);

int
regncomp_l(regex_t *restrict preg, const char *restrict pattern, size_t len, int cflags, locale_t restrict);

int
regwcomp_l(regex_t *restrict preg, const wchar_t *restrict widepat, int cflags, locale_t restrict);

int
regwncomp_l(regex_t *restrict preg, const wchar_t *restrict widepat, size_t len, int cflags, locale_t restrict);

These routines implement IEEE Std 1003.2 (“POSIX.2”) regular expressions (“RE”s); see re_format(7). The () function compiles an RE, written as a string, into an internal form. regexec() matches that internal form against a string and reports results. regerror() transforms error codes from either into human-readable messages. regfree() frees any dynamically-allocated storage used by the internal form of an RE.

The header <regex.h> declares two structure types, regex_t and regmatch_t, the former for compiled internal forms and the latter for match reporting. It also declares the four functions, a type regoff_t, and a number of constants with names starting with “REG_”.

The () function compiles the regular expression contained in the pattern string, subject to the flags in cflags, and places the results in the regex_t structure pointed to by preg. The cflags argument is the bitwise OR of zero or more of the following flags:

Compile modern (“extended”) REs, rather than the obsolete (“basic”) REs that are the default.
This is a synonym for 0, provided as a counterpart to REG_EXTENDED to improve readability.
Compile with recognition of all special characters turned off. All characters are thus considered ordinary, so the “RE” is a literal string. This is an extension, compatible with but not specified by IEEE Std 1003.2 (“POSIX.2”), and should be used with caution in software intended to be portable to other systems. REG_EXTENDED and REG_NOSPEC may not be used in the same call to regcomp().
An alias of REG_NOSPEC.
Compile for matching that ignores upper/lower case distinctions. See re_format(7).
Compile for matching that need only report success or failure, not what was matched.
Compile for newline-sensitive matching. By default, newline is a completely ordinary character with no special meaning in either REs or strings. With this flag, ‘[^’ bracket expressions and ‘.’ never match newline, a ‘^’ anchor matches the null string after any newline in the string in addition to its normal function, and the ‘$’ anchor matches the null string before any newline in the string in addition to its normal function.
(Note that REG_PEND is not recognized by any of the wide character or “n” variants. Besides, the “n” variants can be used instead of REG_PEND; see EXTENDED APIS below.) The regular expression ends, not at the first NUL, but just before the character pointed to by the re_endp member of the structure pointed to by preg. The re_endp member is of type const char *. This flag permits inclusion of NULs in the RE; they are considered ordinary characters. This is an extension, compatible with but not specified by IEEE Std 1003.2 (“POSIX.2”), and should be used with caution in software intended to be portable to other systems.
Recognized enhanced regular expression features; see re_format(7) for details. This is an extension not specified by IEEE Std 1003.2 (“POSIX.2”), and should be used with caution in software intended to be portable to other systems.
Use minimal (non-greedy) repetitions instead of the normal greedy ones; see re_format(7) for details. (This only applies when both REG_ENHANCED and REG_EXTENDED are also set.) This is an extension not specified by IEEE Std 1003.2 (“POSIX.2”), and should be used with caution in software intended to be portable to other systems.
Alias of REG_MINIMAL.

When successful, () returns 0 and fills in the structure pointed to by preg. One member of that structure (other than re_endp) is publicized: re_nsub, of type size_t, contains the number of parenthesized subexpressions within the RE (except that the value of this member is undefined if the REG_NOSUB flag was used). If regcomp() fails, it returns a non-zero error code; see DIAGNOSTICS.

The () function matches the compiled RE pointed to by preg against the string, subject to the flags in eflags, and reports results using nmatch, pmatch, and the returned value. The RE must have been compiled by a previous invocation of regcomp(). The compiled form is not altered during execution of regexec(), so a single compiled RE can be used simultaneously by multiple threads.

By default, the NUL-terminated string pointed to by string is considered to be the text of an entire line, minus any terminating newline. The eflags argument is the bitwise OR of zero or more of the following flags:

The first character of the string is not the beginning of a line, so the ‘^’ anchor should not match before it. This does not affect the behavior of newlines under REG_NEWLINE.
The NUL terminating the string does not end a line, so the ‘$’ anchor should not match before it. This does not affect the behavior of newlines under REG_NEWLINE.
The string is considered to start at string + pmatch[0].rm_so and to have a terminating NUL located at string + pmatch[0].rm_eo (there need not actually be a NUL at that location), regardless of the value of nmatch. See below for the definition of pmatch and nmatch. This is an extension, compatible with but not specified by IEEE Std 1003.2 (“POSIX.2”), and should be used with caution in software intended to be portable to other systems. Note that a non-zero rm_so does not imply REG_NOTBOL; REG_STARTEND affects only the location of the string, not how it is matched.

See re_format(7) for a discussion of what is matched in situations where an RE or a portion thereof could match any of several substrings of string.

Normally, () returns 0 for success and the non-zero code REG_NOMATCH for failure. Other non-zero error codes may be returned in exceptional situations; see DIAGNOSTICS.

If REG_NOSUB was specified in the compilation of the RE, or if nmatch is 0, () ignores the pmatch argument (but see below for the case where REG_STARTEND is specified). Otherwise, pmatch points to an array of nmatch structures of type regmatch_t. Such a structure has at least the members rm_so and rm_eo, both of type regoff_t (a signed arithmetic type at least as large as an off_t and a ssize_t), containing respectively the offset of the first character of a substring and the offset of the first character after the end of the substring. Offsets are measured from the beginning of the string argument given to regexec(). An empty substring is denoted by equal offsets, both indicating the character following the empty substring.

The 0th member of the pmatch array is filled in to indicate what substring of string was matched by the entire RE. Remaining members report what substring was matched by parenthesized subexpressions within the RE; member i reports subexpression i, with subexpressions counted (starting at 1) by the order of their opening parentheses in the RE, left to right. Unused entries in the array (corresponding either to subexpressions that did not participate in the match at all, or to subexpressions that do not exist in the RE (that is, i > preg->re_nsub)) have both rm_so and rm_eo set to -1. If a subexpression participated in the match several times, the reported substring is the last one it matched. (Note, as an example in particular, that when the RE ‘(b*)+’ matches ‘bbb’, the parenthesized subexpression matches each of the three ‘b’s and then an infinite number of empty strings following the last ‘b’, so the reported substring is one of the empties.)

If REG_STARTEND is specified, pmatch must point to at least one regmatch_t (even if nmatch is 0 or REG_NOSUB was specified), to hold the input offsets for REG_STARTEND. Use for output is still entirely controlled by nmatch; if nmatch is 0 or REG_NOSUB was specified, the value of pmatch[0] will not be changed by a successful ().

The () function maps a non-zero errcode from either regcomp() or regexec() to a human-readable, printable message. If preg is non-NULL, the error code should have arisen from use of the regex_t pointed to by preg, and if the error code came from regcomp(), it should have been the result from the most recent regcomp() using that regex_t. The (regerror() may be able to supply a more detailed message using information from the regex_t.) The regerror() function places the NUL-terminated message into the buffer pointed to by errbuf, limiting the length (including the NUL) to at most errbuf_size bytes. If the whole message will not fit, as much of it as will fit before the terminating NUL is supplied. In any case, the returned value is the size of buffer needed to hold the whole message (including terminating NUL). If errbuf_size is 0, errbuf is ignored but the return value is still correct.

If the errcode given to () is first ORed with REG_ITOA, the “message” that results is the printable name of the error code, e.g. “REG_NOMATCH”, rather than an explanation thereof. If errcode is REG_ATOI, then preg shall be non-NULL and the re_endp member of the structure it points to must point to the printable name of an error code; in this case, the result in errbuf is the decimal digits of the numeric value of the error code (0 if the name is not recognized). REG_ITOA and REG_ATOI are intended primarily as debugging facilities; they are extensions, compatible with but not specified by IEEE Std 1003.2 (“POSIX.2”), and should be used with caution in software intended to be portable to other systems. Be warned also that they are considered experimental and changes are possible.

The () function frees any dynamically-allocated storage associated with the compiled RE pointed to by preg. The remaining regex_t is no longer a valid compiled RE and the effect of supplying it to regexec() or regerror() is undefined.

None of these functions references global variables except for tables of constants; all are safe for use from multiple threads if the arguments are safe.

These extended APIs are available in Mac OS X 10.8 and beyond, when the deployment target is 10.8 or later. It should also be noted that any of the () variants may be used to initialize a regex_t structure, that can then be passed to any of the regexec() variants. So it is quite legal to compile a wide character RE and use it to match a multibyte character string, or vice versa.

The () routine compiles regular expressions like regcomp(), but the length of the regular expression string is specified, allowing a string that is not NUL terminated and/or contains NUL characters. This is a modern replacement for using regcomp() with the REG_PEND option.

Similarly, the () routine is like regexec(), but the length of the string to match is specified, allowing a string that is not NUL terminated and/or contains NUL characters.

The () and () variants take a wide-character (wchar_t) string for the regular expression and string to match. And () and () are variants that allow specifying the wide character string length, and so allows wide character strings that are not NUL terminated and/or contains NUL characters.

When regcomp() or one of its variants is run, the regular expression is compiled into an internal form, which may include specific information about the locale currently in effect, such as equivalence classes or multi-character collation symbols. So a reference to the current locale is also stored with the internal form, so that when regexec() is run, it can use the same locale (even if the locale is changed in-between the calls to regcomp() and regexec()).

To provide more direct control over which locale is used, routines with “_l” appended to their names are provided that work just like the variants without the “_l”, except that a locale (via a locale_t variable type) is specified directly. Note that only variants of () have “_l” variants, since the regexec() variants just use the reference to the locale stored in the internal form.

The regex implementation in Mac OS X 10.8 and later is based on a heavily modified subset of TRE (http://laurikari.net/tre/). This provides improved performance, better conformance and additional features. However, both API and binary compatibility have been maintained with previous releases, so binaries built on previous releases should work on 10.8 and later, and binaries built on 10.8 and later should be able to run on previous releases (as long as none of the new variants or new features are used.

There are a number of decisions that IEEE Std 1003.2 (“POSIX.2”) leaves up to the implementor, either by explicitly saying “undefined” or by virtue of them being forbidden by the RE grammar. This implementation treats them as follows.

See re_format(7) for a discussion of the definition of case-independent matching.

There is no particular limit on the length of REs, except insofar as memory is limited. Memory usage is approximately linear in RE size, and largely insensitive to RE complexity, except for bounded repetitions. See BUGS for one short RE using them that will run almost any system out of memory.

A backslashed character other than one specifically given a magic meaning by IEEE Std 1003.2 (“POSIX.2”) (such magic meanings occur only in obsolete [“basic”] REs) is taken as an ordinary character.

Any unmatched ‘[’ is a REG_EBRACK error.

Equivalence classes cannot begin or end bracket-expression ranges. The endpoint of one range cannot begin another.

RE_DUP_MAX, the limit on repetition counts in bounded repetitions, is 255.

A repetition operator (‘?’, ‘*’, ‘+’, or bounds) cannot follow another repetition operator, except for the use of ‘?’ for minimal repetition (for enhanced extended REs; see re_format(7) for details). A repetition operator cannot begin an expression or subexpression or follow ‘^’ or ‘|’.

‘|’ cannot appear first or last in a (sub)expression or after another ‘|’, i.e., an operand of ‘|’ cannot be an empty subexpression. An empty parenthesized subexpression, ‘()’, is legal and matches an empty (sub)string. An empty string is not a legal RE.

A ‘{’ followed by a digit is considered the beginning of bounds for a bounded repetition, which must then follow the syntax for bounds. A ‘{’ followed by a digit is considered an ordinary character.

‘^’ and ‘$’ beginning and ending subexpressions in obsolete (“basic”) REs are anchors, not ordinary characters.

Non-zero error codes from regcomp() and regexec() include the following:

The regexec() function failed to match
invalid regular expression
invalid collating element
invalid character class
‘\’ applied to unescapable character
invalid backreference number
brackets ‘[ ]’ not balanced
parentheses ‘( )’ not balanced
braces ‘{ }’ not balanced
invalid repetition count(s) in ‘{ }’
invalid character range in ‘[ ]’
ran out of memory
‘?’, ‘*’, or ‘+’ operand invalid
empty (sub)expression
cannot happen - you found a bug
invalid argument, e.g. negative-length string
illegal byte sequence (bad multibyte character)

grep(1), re_format(7)

IEEE Std 1003.2 (“POSIX.2”), sections 2.8 (Regular Expression Notation) and B.5 (C Binding for Regular Expression Matching).

The regex implementation is based on a heavily modified subset of TRE (http://laurikari.net/tre/), originally written by Ville Laurikari. Previous releases used an implementation originally written by Henry Spencer, and altered for inclusion in the 4.4BSD distribution.

The beginning-of-line and end-of-line anchors ( “^” and “$”) are currently implemented so that repetitions can not be applied to them. The standards are unclear about whether this is legal, but other regex packages do support this case. It is best to avoid this non-portable (and not really very useful) case.

The back-reference code is subtle and doubts linger about its correctness in complex cases.

The regexec() variants use one of two internal matching engines. The normal one is linear worst-case time in the length of the text being searched, and quadratic worst-case time in the length of the used regular expression. When back-references are used, a slower, backtracking engine is used. While all backtracking matching engines suffer from extreme slowness for certain pathological cases, the normal engines doesn't suffer from these cases. It is advised to avoid back-references whenever possible.

The regcomp() variants implements bounded repetitions by macro expansion, which is costly in time and space if counts are large or bounded repetitions are nested. An RE like, say, ‘((((a{1,100}){1,100}){1,100}){1,100}){1,100}’ will (eventually) run almost any existing machine out of swap space.

Due to a mistake in IEEE Std 1003.2 (“POSIX.2”), things like ‘a)b’ are legal REs because ‘)’ is a special character only in the presence of a previous unmatched ‘(’. This cannot be fixed until the spec is fixed.

The standard's definition of back references is vague. For example, does ‘a\(\(b\)*\2\)*d’ match ‘abbbd’? Until the standard is clarified, behavior in such cases should not be relied on.

Sept 29, 2011 Mac OS X 12