This document describes Rx.
Permission to use, copy, modify, distribute, and sell this software and its documentation for any purpose is hereby granted without fee, provided that the above copyright notice appear in all copies and that both that copyright notice and this permission notice appear in supporting documentation.
BECAUSE THE PROGRAM IS LICENSED FREE OF CHARGE, THERE IS NO WARRANTY FOR THE PROGRAM, TO THE EXTENT PERMITTED BY APPLICABLE LAW. EXCEPT WHEN OTHERWISE STATED IN WRITING THE COPYRIGHT HOLDERS AND/OR OTHER PARTIES PROVIDE THE PROGRAM "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE PROGRAM IS WITH YOU. SHOULD THE PROGRAM PROVE DEFECTIVE, YOU ASSUME THE COST OF ALL NECESSARY SERVICING, REPAIR OR CORRECTION.
IN NO EVENT UNLESS REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING WILL ANY COPYRIGHT HOLDER, OR ANY OTHER PARTY WHO MAY MODIFY AND/OR REDISTRIBUTE THE PROGRAM AS PERMITTED ABOVE, BE LIABLE TO YOU FOR DAMAGES, INCLUDING ANY GENERAL, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THE USE OR INABILITY TO USE THE PROGRAM (INCLUDING BUT NOT LIMITED TO LOSS OF DATA OR DATA BEING RENDERED INACCURATE OR LOSSES SUSTAINED BY YOU OR THIRD PARTIES OR A FAILURE OF THE PROGRAM TO OPERATE WITH ANY OTHER PROGRAMS), EVEN IF SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.
Nothing to say here, yet.
The Posix Basic Regular Expression language is a notation for describing textual patterns. Regexps are typically used by comparing them to a string to see if that string matches the pattern, or by searching within a string for a substring that matches.
This chapter introduces the Posix regexp notation. This is not a formal or precise definition of Posix regexps – it is an intuitive and hopefully expository description of them.
In the simplest cases, a regexp is just a literal string that must match exactly. For example, the pattern:
regexp
matches the string "regexp" and no others.
Some characters have a special meaning when they occur in a regexp.
They aren't matched literally as in the previous example, but instead
denote a more general pattern. For example, the character *
is used to indicate that the preceeding element of a regexp may be
repeated 0, 1, or more times. In the pattern:
smooo*th
the * indicates that the preceeding o can be repeated 0 or
more times. So the pattern matches:
smooth
smoooth
smooooth
smoooooth
...
Suppose you want to write a pattern that literally matches a special
character like * – in other words, you don't want to * to
indicate a permissible repetition, but to match * literally. This
is accomplished by quoting the special character with a backslash.
The pattern:
smoo\*th
matches the string:
smoo*th
and no other strings.
In seven cases, the pattern is reversed – a backslash makes the
character special instead of making a special character normal. The
characters +, ?, |, (, and ) are
normal but the sequences \+, \?, \|, \(,
\), \{, and \} are special (their meaning is
described later).
The remaining sections of this chapter introduce and explain the various special characters that can occur in regexps.
A literal regexp is a string which contains no special characters. A literal regexp matches an identical string, but no other characters. For example:
literally
matches
literally
and nothing else.
Generally, whitespace characters, numbers, and letters are not special. Some punctuation characters are special and some are not (the syntax summary at the end of this chapter makes a convenient reference for which characters are special and which aren't).
This section introduces the special characters . and [.
. matches any character except the NULL character. For example:
p.ck
matches
pick
pack
puck
pbck
pcck
p.ck
...
[ begins a character set. A character set is similar to
. in that it matches not a single, literal character, but any
of a set of characters. [ is different from . in that
with [, you define the set of characters explicitly.
There are three basic forms a character set can take.
In the first form, the character set is spelled out:
[<cset-spec>] -- every character in <cset-spec> is in the set.
In the second form, the character set indicated is the negation of a character set is explicitly spelled out:
[^<cset-spec>] -- every character *not* in <cset-spec> is in the set.
A <cset-spec> is more or less an explicit enumeration of a set
of characters. It can be written as a string of individual characters:
[aeiou]
or as a range of characters:
[0-9]
These two forms can be mixed:
[A-za-z0-9_$]
Note that special regexp characters (such as *) are not
special within a character set. -, as illustrated above,
is special, except, as illustrated below, when it is the first
character mentioned.
This is a four-character set:
[-+*/]
The third form of a character set makes use of a pre-defined "character class":
[[:class-name:]] -- every character described by class-name is in the set.
The supported character classes are:
alnum - the set of alpha-numeric characters
alpha - the set of alphabetic characters
blank - tab and space
cntrl - the control characters
digit - decimal digits
graph - all printable characters except space
lower - lower case letters
print - the "printable" characters
punct - punctuation
space - whitespace characters
upper - upper case letters
xdigit - hexidecimal digits
Finally, character class sets can also be inverted:
[^[:space:]] - all non-whitespace characters
Character sets can be used in a regular expression anywhere a literal character can.
A subexpression is a regular expression enclosed in \( and
\). A subexpression can be used anywhere a single character or
character set can be used.
Subexpressions are useful for grouping regexp constructs. For example,
the repeat operator, *, usually applies to just the preceeding
character. Recall that:
smooo*th
matches
smooth
smoooth
...
Using a subexpression, we can apply * to a longer string:
banan\(an\)*a
matches
banana
bananana
banananana
...
Subexpressions also have a special meaning with regard to backreferences and substitutions (see See Backreferences).
* is the repeat operator. It applies to the preceeding
character, character set, subexpression or backreference. It indicates
that the preceeding element can be matched 0 or more times:
bana\(na\)*
matches
bana
banana
bananana
banananana
...
\+ is similar to * except that \+ requires the
preceeding element to be matched at least once. So while:
bana\(na\)*
matches
bana
bana(na\)\+
does not. Both match
banana
bananana
banananana
...
Thus, bana\(na\)+ is short-hand for banana\(na\)*.
\? indicates that the preceeding character, character set, or
subexpression is optional. It is permitted to match, or to be skipped:
CSNY\?
matches both
CSN
and
CSNY
An interval expression, \{m,n\} where m and n are
non-negative integers with n >= m, applies to the preceeding
character, character set, subexpression or backreference. It indicates
that the preceeding element must match at least m times and may
match as many as n times.
For example:
c\([ad]\)\{1,4\}
matches
car
cdr
caar
cdar
...
caaar
cdaar
...
cadddr
cddddr
An alternative is written:
regexp-1\|regexp-2\|regexp-3\|...
It matches anything matched by some regexp-n. For example:
Crosby, Stills, \(and Nash\|Nash, and Young\)
matches
Crosby, Stills, and Nash
and
Crosby, Stills, Nash, and Young
A backreference is written \n where n is some single digit
other than 0. To be a valid backreference, there must be at least
n parenthesized subexpressions in the pattern prior to the
backreference.
A backreference matches a literal copy of whatever was matched by the corresponding subexpression. For example,
\(.*\)-\1
matches:
go-go
ha-ha
wakka-wakka
...
In some applications, subexpressions are used to extract substrings.
For example, Emacs has the functions match-beginnning and
match-end which report the positions of strings matched by
subexpressions. These functions use the same numbering scheme for
subexpressions as backreferences, with the additional rule that
subexpression 0 is defined to be the whole regexp.
In some applications, subexpressions are used in string substitution. This again uses the backreference numbering scheme. For example, this sed command:
s/From:.*<\(.*\)>/To: \1/
first matches the line:
From: Joe Schmoe <schmoe@uspringfield.edu>
when it does, subexpression 1 matches "schmoe@uspringfield.edu". The command replaces the matched line with "To: \1" after doing subexpression substitution on it to get:
To: schmoe@uspringfield.edu
In summary, regexps can be:
abcd – matching a string literally
. – matching everything except NULL
[a-z_?], ^[a-z_?], [[:alpha:]] and
[^[:alpha:]] – matching character sets
\(subexp\) – grouping an expression into a subexpression.
\n – match a copy of whatever was matched by the nth subexpression.
The following special characters and sequences can be applied to a character, character set, subexpression, or backreference:
* – repeat the preceeding element 0 or more times.
\+ – repeat the preceeding element 1 or more times.
\? – match the preceeding element 0 or 1 time.
{m,n} – match the preceeding element at least m, and as
many as n times.
regexp-1\|regexp-2\|.. – match any regexp-n.
A special character, like . or * can be made into a literal
character by prefixing it with \.
A special sequence, like \+ or \? can be made into a
literal character by dropping the \.
Sometimes a regular expression appears to be ambiguous. For example, suppose we compare the pattern:
begin\|beginning
to the string
beginning
either just the first 5 characters will match, or the whole string will match.
In every case like this, the longer match is preferred. The whole string will match.
Sometimes there is ambiguity not about how many characters to match, but
where the subexpressions occur within the match. This can effect
extraction functions like Emacs' match-beginning or rewrite
functions like sed's s command. For example, consider matching
the pattern:
b\(\[^q]*\)\(ing\)?
against the string
beginning
One possibility is that the first subexpression matches "eginning" and the second is skipped. Another possibility is that the first subexpression matches "eginn" and the second matches "ing".
The rule is that consistant with matching as many characters as possible, the length of lower numbered subexpressions is maximized in preference to maximizing the length of later subexpressions.
In the case of the above example, the two possible matches are equal in overall length. Therefore, it comes down to maximizing the lower-numbered subexpression, \1. The correct answer is that \1 matches "eginning" and \2 is skipped.
This section is excerpted from The GNU C Library reference manual by Sandra Loosemore with Richard M. Stallman, Roland McGrath, and Andrew Oram.
The GNU C library supports the standard POSIX.2 interface. Programs using this interface should include the header file rxposix.h.
Before you can actually match a regular expression, you must compile it. This is not true compilation—it produces a special data structure, not machine instructions. But it is like ordinary compilation in that its purpose is to enable you to “execute” the pattern fast. (See Matching POSIX Regexps, for how to use the compiled regular expression for matching.)
There is a special data type for compiled regular expressions:
This type of object holds a compiled regular expression. It is actually a structure. It has just one field that your programs should look at:
re_nsub- This field holds the number of parenthetical subexpressions in the regular expression that was compiled.
There are several other fields, but we don't describe them here, because only the functions in the library should use them.
After you create a regex_t object, you can compile a regular
expression into it by calling regcomp.
The function
regcomp“compiles” a regular expression into a data structure that you can use withregexecto match against a string. The compiled regular expression format is designed for efficient matching.regcompstores it into*compiled.The parameter pattern points to the regular expression to be compiled. When using
regcomp, pattern must be 0-terminated. When usingregncomp, pattern must be len characters long.
regncompis not a standard function; strictly POSIX programs should avoid using it.It's up to you to allocate an object of type
regex_tand pass its address toregcomp.Before freeing the object of type
regex_tYou must pass it toregfree. Not doing so may cause subsequent calls to Rx functions to behave strangely.The argument cflags lets you specify various options that control the syntax and semantics of regular expressions. See Flags for POSIX Regexps.
If you use the flag
REG_NOSUB, thenregcompomits from the compiled regular expression the information necessary to record how subexpressions actually match. In this case, you might as well pass0for the matchptr and nmatch arguments when you callregexec.If you don't use
REG_NOSUB, then the compiled regular expression does have the capacity to record how subexpressions match. Also,regcomptells you how many subexpressions pattern has, by storing the number in compiled->re_nsub. You can use that value to decide how long an array to allocate to hold information about subexpression matches.
regcompreturns0if it succeeds in compiling the regular expression; otherwise, it returns a nonzero error code (see the table below). You can useregerrorto produce an error message string describing the reason for a nonzero value; see Regexp Cleanup.
Here are the possible nonzero values that regcomp can return:
REG_BADBRREG_BADPATREG_BADRPTREG_ECOLLATEREG_ECTYPEREG_EESCAPEREG_ESUBREGREG_EBRACKREG_EPARENREG_EBRACEREG_ERANGEREG_ESPACEregcomp ran out of memory.
These are the bit flags that you can use in the cflags operand when
compiling a regular expression with regcomp.
REG_EXTENDEDREG_ICASEREG_NOSUBREG_NEWLINEOtherwise, newline acts like any other ordinary character.
Once you have compiled a regular expression, as described in POSIX Regexp Compilation, you can match it against strings using
regexec. A match anywhere inside the string counts as success,
unless the regular expression contains anchor characters (`^' or
`$').
This function tries to match the compiled regular expression
*compiled against string.
regexecreturns0if the regular expression matches; otherwise, it returns a nonzero value. See the table below for what nonzero values mean. You can useregerrorto produce an error message string describing the reason for a nonzero value; see Regexp Cleanup.The parameter string points to the text to search. When using
regexec, string must be 0-terminated. When usingregnexec, string must be len characters long.
regnexecis not a standard function; strictly POSIX programs should avoid using it.The argument eflags is a word of bit flags that enable various options.
If you want to get information about what part of string actually matched the regular expression or its subexpressions, use the arguments matchptr and nmatch. Otherwise, pass
0for nmatch, andNULLfor matchptr. See Regexp Subexpressions.
You must match the regular expression with the same set of current locales that were in effect when you compiled the regular expression.
The function regexec accepts the following flags in the
eflags argument:
REG_NOTBOLREG_NOTEOLHere are the possible nonzero values that regexec can return:
REG_NOMATCHREG_ESPACEregexec ran out of memory.
When regexec matches parenthetical subexpressions of
pattern, it records which parts of string they match. It
returns that information by storing the offsets into an array whose
elements are structures of type regmatch_t. The first element of
the array (index 0) records the part of the string that matched
the entire regular expression. Each other element of the array records
the beginning and end of the part that matched a single parenthetical
subexpression.
This is the data type of the matcharray array that you pass to
regexec. It containes two structure fields, as follows:
rm_so- The offset in string of the beginning of a substring. Add this value to string to get the address of that part.
rm_eo- The offset in string of the end of the substring.
regoff_tis an alias for another signed integer type. The fields ofregmatch_thave typeregoff_t.
The regmatch_t elements correspond to subexpressions
positionally; the first element (index 1) records where the first
subexpression matched, the second element records the second
subexpression, and so on. The order of the subexpressions is the order
in which they begin.
When you call regexec, you specify how long the matchptr
array is, with the nmatch argument. This tells regexec how
many elements to store. If the actual regular expression has more than
nmatch subexpressions, then you won't get offset information about
the rest of them. But this doesn't alter whether the pattern matches a
particular string or not.
If you don't want regexec to return any information about where
the subexpressions matched, you can either supply 0 for
nmatch, or use the flag REG_NOSUB when you compile the
pattern with regcomp.
Sometimes a subexpression matches a substring of no characters. This
happens when `f\(o*\)' matches the string `fum'. (It really
matches just the `f'.) In this case, both of the offsets identify
the point in the string where the null substring was found. In this
example, the offsets are both 1.
Sometimes the entire regular expression can match without using some of
its subexpressions at all—for example, when `ba\(na\)*' matches the
string `ba', the parenthetical subexpression is not used. When
this happens, regexec stores -1 in both fields of the
element for that subexpression.
Sometimes matching the entire regular expression can match a particular
subexpression more than once—for example, when `ba\(na\)*'
matches the string `bananana', the parenthetical subexpression
matches three times. When this happens, regexec usually stores
the offsets of the last part of the string that matched the
subexpression. In the case of `bananana', these offsets are
6 and 8.
But the last match is not always the one that is chosen. It's more
accurate to say that the last opportunity to match is the one
that takes precedence. What this means is that when one subexpression
appears within another, then the results reported for the inner
subexpression reflect whatever happened on the last match of the outer
subexpression. For an example, consider `\(ba\(na\)*s \)*' matching
the string `bananas bas '. The last time the inner expression
actually matches is near the end of the first word. But it is
considered again in the second word, and fails to match there.
regexec reports nonuse of the “na” subexpression.
Another place where this rule applies is when the regular expression
`\(ba\(na\)*s \|nefer\(ti\)* \)*' matches `bananas nefertiti'.
The “na” subexpression does match in the first word, but it doesn't
match in the second word because the other alternative is used there.
Once again, the second repetition of the outer subexpression overrides
the first, and within that second repetition, the “na” subexpression
is not used. So regexec reports nonuse of the “na”
subexpression.
When you are finished using a compiled regular expression, you must
free the storage it uses by calling regfree.
Calling
regfreefrees all the storage that*compiled points to. This includes various internal fields of theregex_tstructure that aren't documented in this manual.
regfreedoes not free the object*compiled itself.
You should always free the space in a regex_t structure with
regfree before using the structure to compile another regular
expression.
When regcomp or regexec reports an error, you can use
the function regerror to turn it into an error message string.
This function produces an error message string for the error code errcode, and stores the string in length bytes of memory starting at buffer. For the compiled argument, supply the same compiled regular expression structure that
regcomporregexecwas working with when it got the error. Alternatively, you can supplyNULLfor compiled; you will still get a meaningful error message, but it might not be as detailed.If the error message can't fit in length bytes (including a terminating null character), then
regerrortruncates it. The string thatregerrorstores is always null-terminated even if it has been truncated.The return value of
regerroris the minimum length needed to store the entire error message. If this is less than length, then the error message was not truncated, and you can use it. Otherwise, you should callregerroragain with a larger buffer.Here is a function which uses
regerror, but always dynamically allocates a buffer for the error message:char *get_regerror (int errcode, regex_t *compiled) { size_t length = regerror (errcode, compiled, NULL, 0); char *buffer = xmalloc (length); (void) regerror (errcode, compiled, buffer, length); return buffer; }
This section is not finished documentation, but rather a collection of pointers towards some of the interesting, non-standard features of Rx.
Rx supports some unusual regexp syntax.
[[:cut N:]] sets pmatch[0].final_tag to N and causes the
matching to stop instantly. If N is 0, the overall match fails,
otherwise it succeeds.
[[:(:]] ... [[:):]] is just like \( ... \) except that in
the first case, no pmatch entries are changed, and the subexpression is
not counted in the numbering of parenthesized subexpressions.
[[:(:]] ... [[:):]] can be used when you do not need to know
where a subexpression matched but are only using parentheses to effect
the parsing of the regexp.
There are two reasons to use [[:(:]] ... [[:):]]:
1. regexec will run faster.
2. Currently, only 8 backreferencable subexpressions are supported:
\1 .. \9. Using [[:(:]] ... [[:):]] is a way to conserve
backreferencable subexpression names in an expression with many
parentheses.
regncomp and regnexec are non-standard generalizations of
regcomp and regexec.
Two mysterious parmaters can be used to trade-off performance and memory use.
At compile-time they are RX_DEFAULT_DFA_CACHE_SIZE and
RX_DEFAULT_NFA_DELAY.
If you want to mess with these (I generally don't advise it), I suggest experimenting for your particular application/memory situation; frob these by powers of two and try out the results on what you expect will be typical regexp workloads.
You can also set those parameters at run-time (before calling any regexp functions) by tweaking the corresponding variables:
rx_default_cache->bytes_allowed
and
rx_basic_unfaniverse_delay
rx_make_solutions, rx_next_solution, and
rx_free_solutions are a lower level alternative to the posix
functions. Using those functions, you can compare a compiled regexp to
a string that is not contiguous in memory or even a string that is not
entirely in memory at any one time.
The code in rxposix.c points out how those functions are used.
If you are only interested in pure regular expressions (no pmatch data,
no backreferences, and no counted subexpressions), you can parse a
regexp using rx_parse, convert it to an nfa using rx_unfa,
and run the dfa using rx_init_system, rx_advance_to_final,
and rx_terminate_system. The dfa Scheme primitives in
rgx.c may provide some guide.
There are two match algorithms. One is for truly regular regexps (those that can be reduced to a dfa). The other is for non-regular regexps.
The dfa algorithm implements the idea suggested in Compilers by Aho, Sethi and Ullman:
[One] approach [to pattern matching regular expressions] is to use a DFA, but avoid constructing all of the transition table by using a technique called "lazy transition evaluation". Here, transitions are computed at run time [when] actually needed. [T]ransitions are stored in a cache. [....] If the cache becomes full, we can erase some previously computed transition to make room for the new transition.
The implementation in Rx is generalized from that, but the above description covers what is used for Posix patterns.
The non-dfa algorithm implements a "recursive decomposition" technique described in email by Henry Spencer. For a given pattern, this algorithm first checks to see if a simpler, superset language, DFA-pattern matches. If it does, then this algorithm does the detail-work to see if the non-DFA pattern matches.
The detail work usually involves recursing on subpatterns. For example,
a concatentation of two subexpressions matches a string if the string
can be divided into two parts, each matching one subexpression, in the
right order. More than one solution is often possible for a given
pattern. This ambiguity is the subject of the "leftmost longest" rules
in the spec, and the back-tracking oriented stream-of-solution functions
rx_make_solutions, rx_next_solution and
rx_free_solutions.
rxspencer.[ch] -- The non-DFA algorithm
rxanal.[ch] rxsuper.[ch] rxnfa.[ch] -- The DFA algorithm