Gitiles
Code Review Sign In
review.jami.net / jami-client-android / 92ef0f2f23fe89dbe290f8187624661c7deab966 / . / jni / libpcre / sources / doc / pcre.txt
blob: 4750c59d8d03e449bf745272d144a7808e71b5b7 [file] [log] [blame]
-----------------------------------------------------------------------------
This file contains a concatenation of the PCRE man pages, converted to plain
text format for ease of searching with a text editor, or for use on systems
that do not have a man page processor. The small individual files that give
synopses of each function in the library have not been included. Neither has
the pcredemo program. There are separate text files for the pcregrep and
pcretest commands.
-----------------------------------------------------------------------------
PCRE(3) PCRE(3)
NAME
PCRE - Perl-compatible regular expressions
INTRODUCTION
The PCRE library is a set of functions that implement regular expres-
sion pattern matching using the same syntax and semantics as Perl, with
just a few differences. Some features that appeared in Python and PCRE
before they appeared in Perl are also available using the Python syn-
tax, there is some support for one or two .NET and Oniguruma syntax
items, and there is an option for requesting some minor changes that
give better JavaScript compatibility.
The current implementation of PCRE corresponds approximately with Perl
5.12, including support for UTF-8 encoded strings and Unicode general
category properties. However, UTF-8 and Unicode support has to be
explicitly enabled; it is not the default. The Unicode tables corre-
spond to Unicode release 6.0.0.
In addition to the Perl-compatible matching function, PCRE contains an
alternative function that matches the same compiled patterns in a dif-
ferent way. In certain circumstances, the alternative function has some
advantages. For a discussion of the two matching algorithms, see the
pcrematching page.
PCRE is written in C and released as a C library. A number of people
have written wrappers and interfaces of various kinds. In particular,
Google Inc. have provided a comprehensive C++ wrapper. This is now
included as part of the PCRE distribution. The pcrecpp page has details
of this interface. Other people's contributions can be found in the
Contrib directory at the primary FTP site, which is:
ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre
Details of exactly which Perl regular expression features are and are
not supported by PCRE are given in separate documents. See the pcrepat-
tern and pcrecompat pages. There is a syntax summary in the pcresyntax
page.
Some features of PCRE can be included, excluded, or changed when the
library is built. The pcre_config() function makes it possible for a
client to discover which features are available. The features them-
selves are described in the pcrebuild page. Documentation about build-
ing PCRE for various operating systems can be found in the README and
NON-UNIX-USE files in the source distribution.
The library contains a number of undocumented internal functions and
data tables that are used by more than one of the exported external
functions, but which are not intended for use by external callers.
Their names all begin with "_pcre_", which hopefully will not provoke
any name clashes. In some environments, it is possible to control which
external symbols are exported when a shared library is built, and in
these cases the undocumented symbols are not exported.
USER DOCUMENTATION
The user documentation for PCRE comprises a number of different sec-
tions. In the "man" format, each of these is a separate "man page". In
the HTML format, each is a separate page, linked from the index page.
In the plain text format, all the sections, except the pcredemo sec-
tion, are concatenated, for ease of searching. The sections are as fol-
lows:
pcre this document
pcre-config show PCRE installation configuration information
pcreapi details of PCRE's native C API
pcrebuild options for building PCRE
pcrecallout details of the callout feature
pcrecompat discussion of Perl compatibility
pcrecpp details of the C++ wrapper
pcredemo a demonstration C program that uses PCRE
pcregrep description of the pcregrep command
pcrejit discussion of the just-in-time optimization support
pcrelimits details of size and other limits
pcrematching discussion of the two matching algorithms
pcrepartial details of the partial matching facility
pcrepattern syntax and semantics of supported
regular expressions
pcreperform discussion of performance issues
pcreposix the POSIX-compatible C API
pcreprecompile details of saving and re-using precompiled patterns
pcresample discussion of the pcredemo program
pcrestack discussion of stack usage
pcresyntax quick syntax reference
pcretest description of the pcretest testing command
pcreunicode discussion of Unicode and UTF-8 support
In addition, in the "man" and HTML formats, there is a short page for
each C library function, listing its arguments and results.
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
Putting an actual email address here seems to have been a spam magnet,
so I've taken it away. If you want to email me, use my two initials,
followed by the two digits 10, at the domain cam.ac.uk.
REVISION
Last updated: 24 August 2011
Copyright (c) 1997-2011 University of Cambridge.
------------------------------------------------------------------------------
PCREBUILD(3) PCREBUILD(3)
NAME
PCRE - Perl-compatible regular expressions
PCRE BUILD-TIME OPTIONS
This document describes the optional features of PCRE that can be
selected when the library is compiled. It assumes use of the configure
script, where the optional features are selected or deselected by pro-
viding options to configure before running the make command. However,
the same options can be selected in both Unix-like and non-Unix-like
environments using the GUI facility of cmake-gui if you are using CMake
instead of configure to build PCRE.
There is a lot more information about building PCRE in non-Unix-like
environments in the file called NON_UNIX_USE, which is part of the PCRE
distribution. You should consult this file as well as the README file
if you are building in a non-Unix-like environment.
The complete list of options for configure (which includes the standard
ones such as the selection of the installation directory) can be
obtained by running
./configure --help
The following sections include descriptions of options whose names
begin with --enable or --disable. These settings specify changes to the
defaults for the configure command. Because of the way that configure
works, --enable and --disable always come in pairs, so the complemen-
tary option always exists as well, but as it specifies the default, it
is not described.
BUILDING SHARED AND STATIC LIBRARIES
The PCRE building process uses libtool to build both shared and static
Unix libraries by default. You can suppress one of these by adding one
of
--disable-shared
--disable-static
to the configure command, as required.
C++ SUPPORT
By default, the configure script will search for a C++ compiler and C++
header files. If it finds them, it automatically builds the C++ wrapper
library for PCRE. You can disable this by adding
--disable-cpp
to the configure command.
UTF-8 SUPPORT
To build PCRE with support for UTF-8 Unicode character strings, add
--enable-utf8
to the configure command. Of itself, this does not make PCRE treat
strings as UTF-8. As well as compiling PCRE with this option, you also
have have to set the PCRE_UTF8 option when you call the pcre_compile()
or pcre_compile2() functions.
If you set --enable-utf8 when compiling in an EBCDIC environment, PCRE
expects its input to be either ASCII or UTF-8 (depending on the runtime
option). It is not possible to support both EBCDIC and UTF-8 codes in
the same version of the library. Consequently, --enable-utf8 and
--enable-ebcdic are mutually exclusive.
UNICODE CHARACTER PROPERTY SUPPORT
UTF-8 support allows PCRE to process character values greater than 255
in the strings that it handles. On its own, however, it does not pro-
vide any facilities for accessing the properties of such characters. If
you want to be able to use the pattern escapes \P, \p, and \X, which
refer to Unicode character properties, you must add
--enable-unicode-properties
to the configure command. This implies UTF-8 support, even if you have
not explicitly requested it.
Including Unicode property support adds around 30K of tables to the
PCRE library. Only the general category properties such as Lu and Nd
are supported. Details are given in the pcrepattern documentation.
JUST-IN-TIME COMPILER SUPPORT
Just-in-time compiler support is included in the build by specifying
--enable-jit
This support is available only for certain hardware architectures. If
this option is set for an unsupported architecture, a compile time
error occurs. See the pcrejit documentation for a discussion of JIT
usage. When JIT support is enabled, pcregrep automatically makes use of
it, unless you add
--disable-pcregrep-jit
to the "configure" command.
CODE VALUE OF NEWLINE
By default, PCRE interprets the linefeed (LF) character as indicating
the end of a line. This is the normal newline character on Unix-like
systems. You can compile PCRE to use carriage return (CR) instead, by
adding
--enable-newline-is-cr
to the configure command. There is also a --enable-newline-is-lf
option, which explicitly specifies linefeed as the newline character.
Alternatively, you can specify that line endings are to be indicated by
the two character sequence CRLF. If you want this, add
--enable-newline-is-crlf
to the configure command. There is a fourth option, specified by
--enable-newline-is-anycrlf
which causes PCRE to recognize any of the three sequences CR, LF, or
CRLF as indicating a line ending. Finally, a fifth option, specified by
--enable-newline-is-any
causes PCRE to recognize any Unicode newline sequence.
Whatever line ending convention is selected when PCRE is built can be
overridden when the library functions are called. At build time it is
conventional to use the standard for your operating system.
WHAT \R MATCHES
By default, the sequence \R in a pattern matches any Unicode newline
sequence, whatever has been selected as the line ending sequence. If
you specify
--enable-bsr-anycrlf
the default is changed so that \R matches only CR, LF, or CRLF. What-
ever is selected when PCRE is built can be overridden when the library
functions are called.
POSIX MALLOC USAGE
When PCRE is called through the POSIX interface (see the pcreposix doc-
umentation), additional working storage is required for holding the
pointers to capturing substrings, because PCRE requires three integers
per substring, whereas the POSIX interface provides only two. If the
number of expected substrings is small, the wrapper function uses space
on the stack, because this is faster than using malloc() for each call.
The default threshold above which the stack is no longer used is 10; it
can be changed by adding a setting such as
--with-posix-malloc-threshold=20
to the configure command.
HANDLING VERY LARGE PATTERNS
Within a compiled pattern, offset values are used to point from one
part to another (for example, from an opening parenthesis to an alter-
nation metacharacter). By default, two-byte values are used for these
offsets, leading to a maximum size for a compiled pattern of around
64K. This is sufficient to handle all but the most gigantic patterns.
Nevertheless, some people do want to process truyl enormous patterns,
so it is possible to compile PCRE to use three-byte or four-byte off-
sets by adding a setting such as
--with-link-size=3
to the configure command. The value given must be 2, 3, or 4. Using
longer offsets slows down the operation of PCRE because it has to load
additional bytes when handling them.
AVOIDING EXCESSIVE STACK USAGE
When matching with the pcre_exec() function, PCRE implements backtrack-
ing by making recursive calls to an internal function called match().
In environments where the size of the stack is limited, this can se-
verely limit PCRE's operation. (The Unix environment does not usually
suffer from this problem, but it may sometimes be necessary to increase
the maximum stack size. There is a discussion in the pcrestack docu-
mentation.) An alternative approach to recursion that uses memory from
the heap to remember data, instead of using recursive function calls,
has been implemented to work round the problem of limited stack size.
If you want to build a version of PCRE that works this way, add
--disable-stack-for-recursion
to the configure command. With this configuration, PCRE will use the
pcre_stack_malloc and pcre_stack_free variables to call memory manage-
ment functions. By default these point to malloc() and free(), but you
can replace the pointers so that your own functions are used instead.
Separate functions are provided rather than using pcre_malloc and
pcre_free because the usage is very predictable: the block sizes
requested are always the same, and the blocks are always freed in
reverse order. A calling program might be able to implement optimized
functions that perform better than malloc() and free(). PCRE runs
noticeably more slowly when built in this way. This option affects only
the pcre_exec() function; it is not relevant for pcre_dfa_exec().
LIMITING PCRE RESOURCE USAGE
Internally, PCRE has a function called match(), which it calls repeat-
edly (sometimes recursively) when matching a pattern with the
pcre_exec() function. By controlling the maximum number of times this
function may be called during a single matching operation, a limit can
be placed on the resources used by a single call to pcre_exec(). The
limit can be changed at run time, as described in the pcreapi documen-
tation. The default is 10 million, but this can be changed by adding a
setting such as
--with-match-limit=500000
to the configure command. This setting has no effect on the
pcre_dfa_exec() matching function.
In some environments it is desirable to limit the depth of recursive
calls of match() more strictly than the total number of calls, in order
to restrict the maximum amount of stack (or heap, if --disable-stack-
for-recursion is specified) that is used. A second limit controls this;
it defaults to the value that is set for --with-match-limit, which
imposes no additional constraints. However, you can set a lower limit
by adding, for example,
--with-match-limit-recursion=10000
to the configure command. This value can also be overridden at run
time.
CREATING CHARACTER TABLES AT BUILD TIME
PCRE uses fixed tables for processing characters whose code values are
less than 256. By default, PCRE is built with a set of tables that are
distributed in the file pcre_chartables.c.dist. These tables are for
ASCII codes only. If you add
--enable-rebuild-chartables
to the configure command, the distributed tables are no longer used.
Instead, a program called dftables is compiled and run. This outputs
the source for new set of tables, created in the default locale of your
C runtime system. (This method of replacing the tables does not work if
you are cross compiling, because dftables is run on the local host. If
you need to create alternative tables when cross compiling, you will
have to do so "by hand".)
USING EBCDIC CODE
PCRE assumes by default that it will run in an environment where the
character code is ASCII (or Unicode, which is a superset of ASCII).
This is the case for most computer operating systems. PCRE can, how-
ever, be compiled to run in an EBCDIC environment by adding
--enable-ebcdic
to the configure command. This setting implies --enable-rebuild-charta-
bles. You should only use it if you know that you are in an EBCDIC
environment (for example, an IBM mainframe operating system). The
--enable-ebcdic option is incompatible with --enable-utf8.
PCREGREP OPTIONS FOR COMPRESSED FILE SUPPORT
By default, pcregrep reads all files as plain text. You can build it so
that it recognizes files whose names end in .gz or .bz2, and reads them
with libz or libbz2, respectively, by adding one or both of
--enable-pcregrep-libz
--enable-pcregrep-libbz2
to the configure command. These options naturally require that the rel-
evant libraries are installed on your system. Configuration will fail
if they are not.
PCREGREP BUFFER SIZE
pcregrep uses an internal buffer to hold a "window" on the file it is
scanning, in order to be able to output "before" and "after" lines when
it finds a match. The size of the buffer is controlled by a parameter
whose default value is 20K. The buffer itself is three times this size,
but because of the way it is used for holding "before" lines, the long-
est line that is guaranteed to be processable is the parameter size.
You can change the default parameter value by adding, for example,
--with-pcregrep-bufsize=50K
to the configure command. The caller of pcregrep can, however, override
this value by specifying a run-time option.
PCRETEST OPTION FOR LIBREADLINE SUPPORT
If you add
--enable-pcretest-libreadline
to the configure command, pcretest is linked with the libreadline
library, and when its input is from a terminal, it reads it using the
readline() function. This provides line-editing and history facilities.
Note that libreadline is GPL-licensed, so if you distribute a binary of
pcretest linked in this way, there may be licensing issues.
Setting this option causes the -lreadline option to be added to the
pcretest build. In many operating environments with a sytem-installed
libreadline this is sufficient. However, in some environments (e.g. if
an unmodified distribution version of readline is in use), some extra
configuration may be necessary. The INSTALL file for libreadline says
this:
"Readline uses the termcap functions, but does not link with the
termcap or curses library itself, allowing applications which link
with readline the to choose an appropriate library."
If your environment has not been set up so that an appropriate library
is automatically included, you may need to add something like
LIBS="-ncurses"
immediately before the configure command.
SEE ALSO
pcreapi(3), pcre_config(3).
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 06 September 2011
Copyright (c) 1997-2011 University of Cambridge.
------------------------------------------------------------------------------
PCREMATCHING(3) PCREMATCHING(3)
NAME
PCRE - Perl-compatible regular expressions
PCRE MATCHING ALGORITHMS
This document describes the two different algorithms that are available
in PCRE for matching a compiled regular expression against a given sub-
ject string. The "standard" algorithm is the one provided by the
pcre_exec() function. This works in the same was as Perl's matching
function, and provides a Perl-compatible matching operation.
An alternative algorithm is provided by the pcre_dfa_exec() function;
this operates in a different way, and is not Perl-compatible. It has
advantages and disadvantages compared with the standard algorithm, and
these are described below.
When there is only one possible way in which a given subject string can
match a pattern, the two algorithms give the same answer. A difference
arises, however, when there are multiple possibilities. For example, if
the pattern
^<.*>
is matched against the string
<something> <something else> <something further>
there are three possible answers. The standard algorithm finds only one
of them, whereas the alternative algorithm finds all three.
REGULAR EXPRESSIONS AS TREES
The set of strings that are matched by a regular expression can be rep-
resented as a tree structure. An unlimited repetition in the pattern
makes the tree of infinite size, but it is still a tree. Matching the
pattern to a given subject string (from a given starting point) can be
thought of as a search of the tree. There are two ways to search a
tree: depth-first and breadth-first, and these correspond to the two
matching algorithms provided by PCRE.
THE STANDARD MATCHING ALGORITHM
In the terminology of Jeffrey Friedl's book "Mastering Regular Expres-
sions", the standard algorithm is an "NFA algorithm". It conducts a
depth-first search of the pattern tree. That is, it proceeds along a
single path through the tree, checking that the subject matches what is
required. When there is a mismatch, the algorithm tries any alterna-
tives at the current point, and if they all fail, it backs up to the
previous branch point in the tree, and tries the next alternative
branch at that level. This often involves backing up (moving to the
left) in the subject string as well. The order in which repetition
branches are tried is controlled by the greedy or ungreedy nature of
the quantifier.
If a leaf node is reached, a matching string has been found, and at
that point the algorithm stops. Thus, if there is more than one possi-
ble match, this algorithm returns the first one that it finds. Whether
this is the shortest, the longest, or some intermediate length depends
on the way the greedy and ungreedy repetition quantifiers are specified
in the pattern.
Because it ends up with a single path through the tree, it is rela-
tively straightforward for this algorithm to keep track of the sub-
strings that are matched by portions of the pattern in parentheses.
This provides support for capturing parentheses and back references.
THE ALTERNATIVE MATCHING ALGORITHM
This algorithm conducts a breadth-first search of the tree. Starting
from the first matching point in the subject, it scans the subject
string from left to right, once, character by character, and as it does
this, it remembers all the paths through the tree that represent valid
matches. In Friedl's terminology, this is a kind of "DFA algorithm",
though it is not implemented as a traditional finite state machine (it
keeps multiple states active simultaneously).
Although the general principle of this matching algorithm is that it
scans the subject string only once, without backtracking, there is one
exception: when a lookaround assertion is encountered, the characters
following or preceding the current point have to be independently
inspected.
The scan continues until either the end of the subject is reached, or
there are no more unterminated paths. At this point, terminated paths
represent the different matching possibilities (if there are none, the
match has failed). Thus, if there is more than one possible match,
this algorithm finds all of them, and in particular, it finds the long-
est. The matches are returned in decreasing order of length. There is
an option to stop the algorithm after the first match (which is neces-
sarily the shortest) is found.
Note that all the matches that are found start at the same point in the
subject. If the pattern
cat(er(pillar)?)?
is matched against the string "the caterpillar catchment", the result
will be the three strings "caterpillar", "cater", and "cat" that start
at the fifth character of the subject. The algorithm does not automati-
cally move on to find matches that start at later positions.
There are a number of features of PCRE regular expressions that are not
supported by the alternative matching algorithm. They are as follows:
1. Because the algorithm finds all possible matches, the greedy or
ungreedy nature of repetition quantifiers is not relevant. Greedy and
ungreedy quantifiers are treated in exactly the same way. However, pos-
sessive quantifiers can make a difference when what follows could also
match what is quantified, for example in a pattern like this:
^a++\w!
This pattern matches "aaab!" but not "aaa!", which would be matched by
a non-possessive quantifier. Similarly, if an atomic group is present,
it is matched as if it were a standalone pattern at the current point,
and the longest match is then "locked in" for the rest of the overall
pattern.
2. When dealing with multiple paths through the tree simultaneously, it
is not straightforward to keep track of captured substrings for the
different matching possibilities, and PCRE's implementation of this
algorithm does not attempt to do this. This means that no captured sub-
strings are available.
3. Because no substrings are captured, back references within the pat-
tern are not supported, and cause errors if encountered.
4. For the same reason, conditional expressions that use a backrefer-
ence as the condition or test for a specific group recursion are not
supported.
5. Because many paths through the tree may be active, the \K escape
sequence, which resets the start of the match when encountered (but may
be on some paths and not on others), is not supported. It causes an
error if encountered.
6. Callouts are supported, but the value of the capture_top field is
always 1, and the value of the capture_last field is always -1.
7. The \C escape sequence, which (in the standard algorithm) matches a
single byte, even in UTF-8 mode, is not supported in UTF-8 mode,
because the alternative algorithm moves through the subject string one
character at a time, for all active paths through the tree.
8. Except for (*FAIL), the backtracking control verbs such as (*PRUNE)
are not supported. (*FAIL) is supported, and behaves like a failing
negative assertion.
ADVANTAGES OF THE ALTERNATIVE ALGORITHM
Using the alternative matching algorithm provides the following advan-
tages:
1. All possible matches (at a single point in the subject) are automat-
ically found, and in particular, the longest match is found. To find
more than one match using the standard algorithm, you have to do kludgy
things with callouts.
2. Because the alternative algorithm scans the subject string just
once, and never needs to backtrack, it is possible to pass very long
subject strings to the matching function in several pieces, checking
for partial matching each time. Although it is possible to do multi-
segment matching using the standard algorithm (pcre_exec()), by retain-
ing partially matched substrings, it is more complicated. The pcrepar-
tial documentation gives details of partial matching and discusses
multi-segment matching.
DISADVANTAGES OF THE ALTERNATIVE ALGORITHM
The alternative algorithm suffers from a number of disadvantages:
1. It is substantially slower than the standard algorithm. This is
partly because it has to search for all possible matches, but is also
because it is less susceptible to optimization.
2. Capturing parentheses and back references are not supported.
3. Although atomic groups are supported, their use does not provide the
performance advantage that it does for the standard algorithm.
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 19 November 2011
Copyright (c) 1997-2010 University of Cambridge.
------------------------------------------------------------------------------
PCREAPI(3) PCREAPI(3)
NAME
PCRE - Perl-compatible regular expressions
PCRE NATIVE API BASIC FUNCTIONS
#include <pcre.h>
pcre *pcre_compile(const char *pattern, int options,
const char **errptr, int *erroffset,
const unsigned char *tableptr);
pcre *pcre_compile2(const char *pattern, int options,
int *errorcodeptr,
const char **errptr, int *erroffset,
const unsigned char *tableptr);
pcre_extra *pcre_study(const pcre *code, int options,
const char **errptr);
void pcre_free_study(pcre_extra *extra);
int pcre_exec(const pcre *code, const pcre_extra *extra,
const char *subject, int length, int startoffset,
int options, int *ovector, int ovecsize);
PCRE NATIVE API AUXILIARY FUNCTIONS
pcre_jit_stack *pcre_jit_stack_alloc(int startsize, int maxsize);
void pcre_jit_stack_free(pcre_jit_stack *stack);
void pcre_assign_jit_stack(pcre_extra *extra,
pcre_jit_callback callback, void *data);
int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
const char *subject, int length, int startoffset,
int options, int *ovector, int ovecsize,
int *workspace, int wscount);
int pcre_copy_named_substring(const pcre *code,
const char *subject, int *ovector,
int stringcount, const char *stringname,
char *buffer, int buffersize);
int pcre_copy_substring(const char *subject, int *ovector,
int stringcount, int stringnumber, char *buffer,
int buffersize);
int pcre_get_named_substring(const pcre *code,
const char *subject, int *ovector,
int stringcount, const char *stringname,
const char **stringptr);
int pcre_get_stringnumber(const pcre *code,
const char *name);
int pcre_get_stringtable_entries(const pcre *code,
const char *name, char **first, char **last);
int pcre_get_substring(const char *subject, int *ovector,
int stringcount, int stringnumber,
const char **stringptr);
int pcre_get_substring_list(const char *subject,
int *ovector, int stringcount, const char ***listptr);
void pcre_free_substring(const char *stringptr);
void pcre_free_substring_list(const char **stringptr);
const unsigned char *pcre_maketables(void);
int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
int what, void *where);
int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
int pcre_refcount(pcre *code, int adjust);
int pcre_config(int what, void *where);
char *pcre_version(void);
PCRE NATIVE API INDIRECTED FUNCTIONS
void *(*pcre_malloc)(size_t);
void (*pcre_free)(void *);
void *(*pcre_stack_malloc)(size_t);
void (*pcre_stack_free)(void *);
int (*pcre_callout)(pcre_callout_block *);
PCRE API OVERVIEW
PCRE has its own native API, which is described in this document. There
are also some wrapper functions that correspond to the POSIX regular
expression API, but they do not give access to all the functionality.
They are described in the pcreposix documentation. Both of these APIs
define a set of C function calls. A C++ wrapper is also distributed
with PCRE. It is documented in the pcrecpp page.
The native API C function prototypes are defined in the header file
pcre.h, and on Unix systems the library itself is called libpcre. It
can normally be accessed by adding -lpcre to the command for linking an
application that uses PCRE. The header file defines the macros
PCRE_MAJOR and PCRE_MINOR to contain the major and minor release num-
bers for the library. Applications can use these to include support
for different releases of PCRE.
In a Windows environment, if you want to statically link an application
program against a non-dll pcre.a file, you must define PCRE_STATIC
before including pcre.h or pcrecpp.h, because otherwise the pcre_mal-
loc() and pcre_free() exported functions will be declared
__declspec(dllimport), with unwanted results.
The functions pcre_compile(), pcre_compile2(), pcre_study(), and
pcre_exec() are used for compiling and matching regular expressions in
a Perl-compatible manner. A sample program that demonstrates the sim-
plest way of using them is provided in the file called pcredemo.c in
the PCRE source distribution. A listing of this program is given in the
pcredemo documentation, and the pcresample documentation describes how
to compile and run it.
Just-in-time compiler support is an optional feature of PCRE that can
be built in appropriate hardware environments. It greatly speeds up the
matching performance of many patterns. Simple programs can easily
request that it be used if available, by setting an option that is
ignored when it is not relevant. More complicated programs might need
to make use of the functions pcre_jit_stack_alloc(),
pcre_jit_stack_free(), and pcre_assign_jit_stack() in order to control
the JIT code's memory usage. These functions are discussed in the
pcrejit documentation.
A second matching function, pcre_dfa_exec(), which is not Perl-compati-
ble, is also provided. This uses a different algorithm for the match-
ing. The alternative algorithm finds all possible matches (at a given
point in the subject), and scans the subject just once (unless there
are lookbehind assertions). However, this algorithm does not return
captured substrings. A description of the two matching algorithms and
their advantages and disadvantages is given in the pcrematching docu-
mentation.
In addition to the main compiling and matching functions, there are
convenience functions for extracting captured substrings from a subject
string that is matched by pcre_exec(). They are:
pcre_copy_substring()
pcre_copy_named_substring()
pcre_get_substring()
pcre_get_named_substring()
pcre_get_substring_list()
pcre_get_stringnumber()
pcre_get_stringtable_entries()
pcre_free_substring() and pcre_free_substring_list() are also provided,
to free the memory used for extracted strings.
The function pcre_maketables() is used to build a set of character
tables in the current locale for passing to pcre_compile(),
pcre_exec(), or pcre_dfa_exec(). This is an optional facility that is
provided for specialist use. Most commonly, no special tables are
passed, in which case internal tables that are generated when PCRE is
built are used.
The function pcre_fullinfo() is used to find out information about a
compiled pattern; pcre_info() is an obsolete version that returns only
some of the available information, but is retained for backwards com-
patibility. The function pcre_version() returns a pointer to a string
containing the version of PCRE and its date of release.
The function pcre_refcount() maintains a reference count in a data
block containing a compiled pattern. This is provided for the benefit
of object-oriented applications.
The global variables pcre_malloc and pcre_free initially contain the
entry points of the standard malloc() and free() functions, respec-
tively. PCRE calls the memory management functions via these variables,
so a calling program can replace them if it wishes to intercept the
calls. This should be done before calling any PCRE functions.
The global variables pcre_stack_malloc and pcre_stack_free are also
indirections to memory management functions. These special functions
are used only when PCRE is compiled to use the heap for remembering
data, instead of recursive function calls, when running the pcre_exec()
function. See the pcrebuild documentation for details of how to do
this. It is a non-standard way of building PCRE, for use in environ-
ments that have limited stacks. Because of the greater use of memory
management, it runs more slowly. Separate functions are provided so
that special-purpose external code can be used for this case. When
used, these functions are always called in a stack-like manner (last
obtained, first freed), and always for memory blocks of the same size.
There is a discussion about PCRE's stack usage in the pcrestack docu-
mentation.
The global variable pcre_callout initially contains NULL. It can be set
by the caller to a "callout" function, which PCRE will then call at
specified points during a matching operation. Details are given in the
pcrecallout documentation.
NEWLINES
PCRE supports five different conventions for indicating line breaks in
strings: a single CR (carriage return) character, a single LF (line-
feed) character, the two-character sequence CRLF, any of the three pre-
ceding, or any Unicode newline sequence. The Unicode newline sequences
are the three just mentioned, plus the single characters VT (vertical
tab, U+000B), FF (formfeed, U+000C), NEL (next line, U+0085), LS (line
separator, U+2028), and PS (paragraph separator, U+2029).
Each of the first three conventions is used by at least one operating
system as its standard newline sequence. When PCRE is built, a default
can be specified. The default default is LF, which is the Unix stan-
dard. When PCRE is run, the default can be overridden, either when a
pattern is compiled, or when it is matched.
At compile time, the newline convention can be specified by the options
argument of pcre_compile(), or it can be specified by special text at
the start of the pattern itself; this overrides any other settings. See
the pcrepattern page for details of the special character sequences.
In the PCRE documentation the word "newline" is used to mean "the char-
acter or pair of characters that indicate a line break". The choice of
newline convention affects the handling of the dot, circumflex, and
dollar metacharacters, the handling of #-comments in /x mode, and, when
CRLF is a recognized line ending sequence, the match position advance-
ment for a non-anchored pattern. There is more detail about this in the
section on pcre_exec() options below.
The choice of newline convention does not affect the interpretation of
the \n or \r escape sequences, nor does it affect what \R matches,
which is controlled in a similar way, but by separate options.
MULTITHREADING
The PCRE functions can be used in multi-threading applications, with
the proviso that the memory management functions pointed to by
pcre_malloc, pcre_free, pcre_stack_malloc, and pcre_stack_free, and the
callout function pointed to by pcre_callout, are shared by all threads.
The compiled form of a regular expression is not altered during match-
ing, so the same compiled pattern can safely be used by several threads
at once.
If the just-in-time optimization feature is being used, it needs sepa-
rate memory stack areas for each thread. See the pcrejit documentation
for more details.
SAVING PRECOMPILED PATTERNS FOR LATER USE
The compiled form of a regular expression can be saved and re-used at a
later time, possibly by a different program, and even on a host other
than the one on which it was compiled. Details are given in the
pcreprecompile documentation. However, compiling a regular expression
with one version of PCRE for use with a different version is not guar-
anteed to work and may cause crashes.
CHECKING BUILD-TIME OPTIONS
int pcre_config(int what, void *where);
The function pcre_config() makes it possible for a PCRE client to dis-
cover which optional features have been compiled into the PCRE library.
The pcrebuild documentation has more details about these optional fea-
tures.
The first argument for pcre_config() is an integer, specifying which
information is required; the second argument is a pointer to a variable
into which the information is placed. The following information is
available:
PCRE_CONFIG_UTF8
The output is an integer that is set to one if UTF-8 support is avail-
able; otherwise it is set to zero.
PCRE_CONFIG_UNICODE_PROPERTIES
The output is an integer that is set to one if support for Unicode
character properties is available; otherwise it is set to zero.
PCRE_CONFIG_JIT
The output is an integer that is set to one if support for just-in-time
compiling is available; otherwise it is set to zero.
PCRE_CONFIG_NEWLINE
The output is an integer whose value specifies the default character
sequence that is recognized as meaning "newline". The four values that
are supported are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for ANYCRLF,
and -1 for ANY. Though they are derived from ASCII, the same values
are returned in EBCDIC environments. The default should normally corre-
spond to the standard sequence for your operating system.
PCRE_CONFIG_BSR
The output is an integer whose value indicates what character sequences
the \R escape sequence matches by default. A value of 0 means that \R
matches any Unicode line ending sequence; a value of 1 means that \R
matches only CR, LF, or CRLF. The default can be overridden when a pat-
tern is compiled or matched.
PCRE_CONFIG_LINK_SIZE
The output is an integer that contains the number of bytes used for
internal linkage in compiled regular expressions. The value is 2, 3, or
4. Larger values allow larger regular expressions to be compiled, at
the expense of slower matching. The default value of 2 is sufficient
for all but the most massive patterns, since it allows the compiled
pattern to be up to 64K in size.
PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
The output is an integer that contains the threshold above which the
POSIX interface uses malloc() for output vectors. Further details are
given in the pcreposix documentation.
PCRE_CONFIG_MATCH_LIMIT
The output is a long integer that gives the default limit for the num-
ber of internal matching function calls in a pcre_exec() execution.
Further details are given with pcre_exec() below.
PCRE_CONFIG_MATCH_LIMIT_RECURSION
The output is a long integer that gives the default limit for the depth
of recursion when calling the internal matching function in a
pcre_exec() execution. Further details are given with pcre_exec()
below.
PCRE_CONFIG_STACKRECURSE
The output is an integer that is set to one if internal recursion when
running pcre_exec() is implemented by recursive function calls that use
the stack to remember their state. This is the usual way that PCRE is
compiled. The output is zero if PCRE was compiled to use blocks of data
on the heap instead of recursive function calls. In this case,
pcre_stack_malloc and pcre_stack_free are called to manage memory
blocks on the heap, thus avoiding the use of the stack.
COMPILING A PATTERN
pcre *pcre_compile(const char *pattern, int options,
const char **errptr, int *erroffset,
const unsigned char *tableptr);
pcre *pcre_compile2(const char *pattern, int options,
int *errorcodeptr,
const char **errptr, int *erroffset,
const unsigned char *tableptr);
Either of the functions pcre_compile() or pcre_compile2() can be called
to compile a pattern into an internal form. The only difference between
the two interfaces is that pcre_compile2() has an additional argument,
errorcodeptr, via which a numerical error code can be returned. To
avoid too much repetition, we refer just to pcre_compile() below, but
the information applies equally to pcre_compile2().
The pattern is a C string terminated by a binary zero, and is passed in
the pattern argument. A pointer to a single block of memory that is
obtained via pcre_malloc is returned. This contains the compiled code
and related data. The pcre type is defined for the returned block; this
is a typedef for a structure whose contents are not externally defined.
It is up to the caller to free the memory (via pcre_free) when it is no
longer required.
Although the compiled code of a PCRE regex is relocatable, that is, it
does not depend on memory location, the complete pcre data block is not
fully relocatable, because it may contain a copy of the tableptr argu-
ment, which is an address (see below).
The options argument contains various bit settings that affect the com-
pilation. It should be zero if no options are required. The available
options are described below. Some of them (in particular, those that
are compatible with Perl, but some others as well) can also be set and
unset from within the pattern (see the detailed description in the
pcrepattern documentation). For those options that can be different in
different parts of the pattern, the contents of the options argument
specifies their settings at the start of compilation and execution. The
PCRE_ANCHORED, PCRE_BSR_xxx, PCRE_NEWLINE_xxx, PCRE_NO_UTF8_CHECK, and
PCRE_NO_START_OPT options can be set at the time of matching as well as
at compile time.
If errptr is NULL, pcre_compile() returns NULL immediately. Otherwise,
if compilation of a pattern fails, pcre_compile() returns NULL, and
sets the variable pointed to by errptr to point to a textual error mes-
sage. This is a static string that is part of the library. You must not
try to free it. Normally, the offset from the start of the pattern to
the byte that was being processed when the error was discovered is
placed in the variable pointed to by erroffset, which must not be NULL
(if it is, an immediate error is given). However, for an invalid UTF-8
string, the offset is that of the first byte of the failing character.
Also, some errors are not detected until checks are carried out when
the whole pattern has been scanned; in these cases the offset passed
back is the length of the pattern.
Note that the offset is in bytes, not characters, even in UTF-8 mode.
It may sometimes point into the middle of a UTF-8 character.
If pcre_compile2() is used instead of pcre_compile(), and the error-
codeptr argument is not NULL, a non-zero error code number is returned
via this argument in the event of an error. This is in addition to the
textual error message. Error codes and messages are listed below.
If the final argument, tableptr, is NULL, PCRE uses a default set of
character tables that are built when PCRE is compiled, using the
default C locale. Otherwise, tableptr must be an address that is the
result of a call to pcre_maketables(). This value is stored with the
compiled pattern, and used again by pcre_exec(), unless another table
pointer is passed to it. For more discussion, see the section on locale
support below.
This code fragment shows a typical straightforward call to pcre_com-
pile():
pcre *re;
const char *error;
int erroffset;
re = pcre_compile(
"^A.*Z", /* the pattern */
0, /* default options */
&error, /* for error message */
&erroffset, /* for error offset */
NULL); /* use default character tables */
The following names for option bits are defined in the pcre.h header
file:
PCRE_ANCHORED
If this bit is set, the pattern is forced to be "anchored", that is, it
is constrained to match only at the first matching point in the string
that is being searched (the "subject string"). This effect can also be
achieved by appropriate constructs in the pattern itself, which is the
only way to do it in Perl.
PCRE_AUTO_CALLOUT
If this bit is set, pcre_compile() automatically inserts callout items,
all with number 255, before each pattern item. For discussion of the
callout facility, see the pcrecallout documentation.
PCRE_BSR_ANYCRLF
PCRE_BSR_UNICODE
These options (which are mutually exclusive) control what the \R escape
sequence matches. The choice is either to match only CR, LF, or CRLF,
or to match any Unicode newline sequence. The default is specified when
PCRE is built. It can be overridden from within the pattern, or by set-
ting an option when a compiled pattern is matched.
PCRE_CASELESS
If this bit is set, letters in the pattern match both upper and lower
case letters. It is equivalent to Perl's /i option, and it can be
changed within a pattern by a (?i) option setting. In UTF-8 mode, PCRE
always understands the concept of case for characters whose values are
less than 128, so caseless matching is always possible. For characters
with higher values, the concept of case is supported if PCRE is com-
piled with Unicode property support, but not otherwise. If you want to
use caseless matching for characters 128 and above, you must ensure
that PCRE is compiled with Unicode property support as well as with
UTF-8 support.
PCRE_DOLLAR_ENDONLY
If this bit is set, a dollar metacharacter in the pattern matches only
at the end of the subject string. Without this option, a dollar also
matches immediately before a newline at the end of the string (but not
before any other newlines). The PCRE_DOLLAR_ENDONLY option is ignored
if PCRE_MULTILINE is set. There is no equivalent to this option in
Perl, and no way to set it within a pattern.
PCRE_DOTALL
If this bit is set, a dot metacharacter in the pattern matches a char-
acter of any value, including one that indicates a newline. However, it
only ever matches one character, even if newlines are coded as CRLF.
Without this option, a dot does not match when the current position is
at a newline. This option is equivalent to Perl's /s option, and it can
be changed within a pattern by a (?s) option setting. A negative class
such as [^a] always matches newline characters, independent of the set-
ting of this option.
PCRE_DUPNAMES
If this bit is set, names used to identify capturing subpatterns need
not be unique. This can be helpful for certain types of pattern when it
is known that only one instance of the named subpattern can ever be
matched. There are more details of named subpatterns below; see also
the pcrepattern documentation.
PCRE_EXTENDED
If this bit is set, whitespace data characters in the pattern are
totally ignored except when escaped or inside a character class. White-
space does not include the VT character (code 11). In addition, charac-
ters between an unescaped # outside a character class and the next new-
line, inclusive, are also ignored. This is equivalent to Perl's /x
option, and it can be changed within a pattern by a (?x) option set-
ting.
Which characters are interpreted as newlines is controlled by the
options passed to pcre_compile() or by a special sequence at the start
of the pattern, as described in the section entitled "Newline conven-
tions" in the pcrepattern documentation. Note that the end of this type
of comment is a literal newline sequence in the pattern; escape
sequences that happen to represent a newline do not count.
This option makes it possible to include comments inside complicated
patterns. Note, however, that this applies only to data characters.
Whitespace characters may never appear within special character
sequences in a pattern, for example within the sequence (?( that intro-
duces a conditional subpattern.
PCRE_EXTRA
This option was invented in order to turn on additional functionality
of PCRE that is incompatible with Perl, but it is currently of very
little use. When set, any backslash in a pattern that is followed by a
letter that has no special meaning causes an error, thus reserving
these combinations for future expansion. By default, as in Perl, a
backslash followed by a letter with no special meaning is treated as a
literal. (Perl can, however, be persuaded to give an error for this, by
running it with the -w option.) There are at present no other features
controlled by this option. It can also be set by a (?X) option setting
within a pattern.
PCRE_FIRSTLINE
If this option is set, an unanchored pattern is required to match
before or at the first newline in the subject string, though the
matched text may continue over the newline.
PCRE_JAVASCRIPT_COMPAT
If this option is set, PCRE's behaviour is changed in some ways so that
it is compatible with JavaScript rather than Perl. The changes are as
follows:
(1) A lone closing square bracket in a pattern causes a compile-time
error, because this is illegal in JavaScript (by default it is treated
as a data character). Thus, the pattern AB]CD becomes illegal when this
option is set.
(2) At run time, a back reference to an unset subpattern group matches
an empty string (by default this causes the current matching alterna-
tive to fail). A pattern such as (\1)(a) succeeds when this option is
set (assuming it can find an "a" in the subject), whereas it fails by
default, for Perl compatibility.
(3) \U matches an upper case "U" character; by default \U causes a com-
pile time error (Perl uses \U to upper case subsequent characters).
(4) \u matches a lower case "u" character unless it is followed by four
hexadecimal digits, in which case the hexadecimal number defines the
code point to match. By default, \u causes a compile time error (Perl
uses it to upper case the following character).
(5) \x matches a lower case "x" character unless it is followed by two
hexadecimal digits, in which case the hexadecimal number defines the
code point to match. By default, as in Perl, a hexadecimal number is
always expected after \x, but it may have zero, one, or two digits (so,
for example, \xz matches a binary zero character followed by z).
PCRE_MULTILINE
By default, PCRE treats the subject string as consisting of a single
line of characters (even if it actually contains newlines). The "start
of line" metacharacter (^) matches only at the start of the string,
while the "end of line" metacharacter ($) matches only at the end of
the string, or before a terminating newline (unless PCRE_DOLLAR_ENDONLY
is set). This is the same as Perl.
When PCRE_MULTILINE it is set, the "start of line" and "end of line"
constructs match immediately following or immediately before internal
newlines in the subject string, respectively, as well as at the very
start and end. This is equivalent to Perl's /m option, and it can be
changed within a pattern by a (?m) option setting. If there are no new-
lines in a subject string, or no occurrences of ^ or $ in a pattern,
setting PCRE_MULTILINE has no effect.
PCRE_NEWLINE_CR
PCRE_NEWLINE_LF
PCRE_NEWLINE_CRLF
PCRE_NEWLINE_ANYCRLF
PCRE_NEWLINE_ANY
These options override the default newline definition that was chosen
when PCRE was built. Setting the first or the second specifies that a
newline is indicated by a single character (CR or LF, respectively).
Setting PCRE_NEWLINE_CRLF specifies that a newline is indicated by the
two-character CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies
that any of the three preceding sequences should be recognized. Setting
PCRE_NEWLINE_ANY specifies that any Unicode newline sequence should be
recognized. The Unicode newline sequences are the three just mentioned,
plus the single characters VT (vertical tab, U+000B), FF (formfeed,
U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
(paragraph separator, U+2029). The last two are recognized only in
UTF-8 mode.
The newline setting in the options word uses three bits that are
treated as a number, giving eight possibilities. Currently only six are
used (default plus the five values above). This means that if you set
more than one newline option, the combination may or may not be sensi-
ble. For example, PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to
PCRE_NEWLINE_CRLF, but other combinations may yield unused numbers and
cause an error.
The only time that a line break in a pattern is specially recognized
when compiling is when PCRE_EXTENDED is set. CR and LF are whitespace
characters, and so are ignored in this mode. Also, an unescaped # out-
side a character class indicates a comment that lasts until after the
next line break sequence. In other circumstances, line break sequences
in patterns are treated as literal data.
The newline option that is set at compile time becomes the default that
is used for pcre_exec() and pcre_dfa_exec(), but it can be overridden.
PCRE_NO_AUTO_CAPTURE
If this option is set, it disables the use of numbered capturing paren-
theses in the pattern. Any opening parenthesis that is not followed by
? behaves as if it were followed by ?: but named parentheses can still
be used for capturing (and they acquire numbers in the usual way).
There is no equivalent of this option in Perl.
NO_START_OPTIMIZE
This is an option that acts at matching time; that is, it is really an
option for pcre_exec() or pcre_dfa_exec(). If it is set at compile
time, it is remembered with the compiled pattern and assumed at match-
ing time. For details see the discussion of PCRE_NO_START_OPTIMIZE
below.
PCRE_UCP
This option changes the way PCRE processes \B, \b, \D, \d, \S, \s, \W,
\w, and some of the POSIX character classes. By default, only ASCII
characters are recognized, but if PCRE_UCP is set, Unicode properties
are used instead to classify characters. More details are given in the
section on generic character types in the pcrepattern page. If you set
PCRE_UCP, matching one of the items it affects takes much longer. The
option is available only if PCRE has been compiled with Unicode prop-
erty support.
PCRE_UNGREEDY
This option inverts the "greediness" of the quantifiers so that they
are not greedy by default, but become greedy if followed by "?". It is
not compatible with Perl. It can also be set by a (?U) option setting
within the pattern.
PCRE_UTF8
This option causes PCRE to regard both the pattern and the subject as
strings of UTF-8 characters instead of single-byte character strings.
However, it is available only when PCRE is built to include UTF-8 sup-
port. If not, the use of this option provokes an error. Details of how
this option changes the behaviour of PCRE are given in the pcreunicode
page.
PCRE_NO_UTF8_CHECK
When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
automatically checked. There is a discussion about the validity of
UTF-8 strings in the main pcre page. If an invalid UTF-8 sequence of
bytes is found, pcre_compile() returns an error. If you already know
that your pattern is valid, and you want to skip this check for perfor-
mance reasons, you can set the PCRE_NO_UTF8_CHECK option. When it is
set, the effect of passing an invalid UTF-8 string as a pattern is
undefined. It may cause your program to crash. Note that this option
can also be passed to pcre_exec() and pcre_dfa_exec(), to suppress the
UTF-8 validity checking of subject strings.
COMPILATION ERROR CODES
The following table lists the error codes than may be returned by
pcre_compile2(), along with the error messages that may be returned by
both compiling functions. As PCRE has developed, some error codes have
fallen out of use. To avoid confusion, they have not been re-used.
0 no error
1 \ at end of pattern
2 \c at end of pattern
3 unrecognized character follows \
4 numbers out of order in {} quantifier
5 number too big in {} quantifier
6 missing terminating ] for character class
7 invalid escape sequence in character class
8 range out of order in character class
9 nothing to repeat
10 [this code is not in use]
11 internal error: unexpected repeat
12 unrecognized character after (? or (?-
13 POSIX named classes are supported only within a class
14 missing )
15 reference to non-existent subpattern
16 erroffset passed as NULL
17 unknown option bit(s) set
18 missing ) after comment
19 [this code is not in use]
20 regular expression is too large
21 failed to get memory
22 unmatched parentheses
23 internal error: code overflow
24 unrecognized character after (?<
25 lookbehind assertion is not fixed length
26 malformed number or name after (?(
27 conditional group contains more than two branches
28 assertion expected after (?(
29 (?R or (?[+-]digits must be followed by )
30 unknown POSIX class name
31 POSIX collating elements are not supported
32 this version of PCRE is not compiled with PCRE_UTF8 support
33 [this code is not in use]
34 character value in \x{...} sequence is too large
35 invalid condition (?(0)
36 \C not allowed in lookbehind assertion
37 PCRE does not support \L, \l, \N{name}, \U, or \u
38 number after (?C is > 255
39 closing ) for (?C expected
40 recursive call could loop indefinitely
41 unrecognized character after (?P
42 syntax error in subpattern name (missing terminator)
43 two named subpatterns have the same name
44 invalid UTF-8 string
45 support for \P, \p, and \X has not been compiled
46 malformed \P or \p sequence
47 unknown property name after \P or \p
48 subpattern name is too long (maximum 32 characters)
49 too many named subpatterns (maximum 10000)
50 [this code is not in use]
51 octal value is greater than \377 (not in UTF-8 mode)
52 internal error: overran compiling workspace
53 internal error: previously-checked referenced subpattern
not found
54 DEFINE group contains more than one branch
55 repeating a DEFINE group is not allowed
56 inconsistent NEWLINE options
57 \g is not followed by a braced, angle-bracketed, or quoted
name/number or by a plain number
58 a numbered reference must not be zero
59 an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)
60 (*VERB) not recognized
61 number is too big
62 subpattern name expected
63 digit expected after (?+
64 ] is an invalid data character in JavaScript compatibility mode
65 different names for subpatterns of the same number are
not allowed
66 (*MARK) must have an argument
67 this version of PCRE is not compiled with PCRE_UCP support
68 \c must be followed by an ASCII character
69 \k is not followed by a braced, angle-bracketed, or quoted name
The numbers 32 and 10000 in errors 48 and 49 are defaults; different
values may be used if the limits were changed when PCRE was built.
STUDYING A PATTERN
pcre_extra *pcre_study(const pcre *code, int options
const char **errptr);
If a compiled pattern is going to be used several times, it is worth
spending more time analyzing it in order to speed up the time taken for
matching. The function pcre_study() takes a pointer to a compiled pat-
tern as its first argument. If studying the pattern produces additional
information that will help speed up matching, pcre_study() returns a
pointer to a pcre_extra block, in which the study_data field points to
the results of the study.
The returned value from pcre_study() can be passed directly to
pcre_exec() or pcre_dfa_exec(). However, a pcre_extra block also con-
tains other fields that can be set by the caller before the block is
passed; these are described below in the section on matching a pattern.
If studying the pattern does not produce any useful information,
pcre_study() returns NULL. In that circumstance, if the calling program
wants to pass any of the other fields to pcre_exec() or
pcre_dfa_exec(), it must set up its own pcre_extra block.
The second argument of pcre_study() contains option bits. There is only
one option: PCRE_STUDY_JIT_COMPILE. If this is set, and the just-in-
time compiler is available, the pattern is further compiled into
machine code that executes much faster than the pcre_exec() matching
function. If the just-in-time compiler is not available, this option is
ignored. All other bits in the options argument must be zero.
JIT compilation is a heavyweight optimization. It can take some time
for patterns to be analyzed, and for one-off matches and simple pat-
terns the benefit of faster execution might be offset by a much slower
study time. Not all patterns can be optimized by the JIT compiler. For
those that cannot be handled, matching automatically falls back to the
pcre_exec() interpreter. For more details, see the pcrejit documenta-
tion.
The third argument for pcre_study() is a pointer for an error message.
If studying succeeds (even if no data is returned), the variable it
points to is set to NULL. Otherwise it is set to point to a textual
error message. This is a static string that is part of the library. You
must not try to free it. You should test the error pointer for NULL
after calling pcre_study(), to be sure that it has run successfully.
When you are finished with a pattern, you can free the memory used for
the study data by calling pcre_free_study(). This function was added to
the API for release 8.20. For earlier versions, the memory could be
freed with pcre_free(), just like the pattern itself. This will still
work in cases where PCRE_STUDY_JIT_COMPILE is not used, but it is
advisable to change to the new function when convenient.
This is a typical way in which pcre_study() is used (except that in a
real application there should be tests for errors):
int rc;
pcre *re;
pcre_extra *sd;
re = pcre_compile("pattern", 0, &error, &erroroffset, NULL);
sd = pcre_study(
re, /* result of pcre_compile() */
0, /* no options */
&error); /* set to NULL or points to a message */
rc = pcre_exec( /* see below for details of pcre_exec() options */
re, sd, "subject", 7, 0, 0, ovector, 30);
...
pcre_free_study(sd);
pcre_free(re);
Studying a pattern does two things: first, a lower bound for the length
of subject string that is needed to match the pattern is computed. This
does not mean that there are any strings of that length that match, but
it does guarantee that no shorter strings match. The value is used by
pcre_exec() and pcre_dfa_exec() to avoid wasting time by trying to
match strings that are shorter than the lower bound. You can find out
the value in a calling program via the pcre_fullinfo() function.
Studying a pattern is also useful for non-anchored patterns that do not
have a single fixed starting character. A bitmap of possible starting
bytes is created. This speeds up finding a position in the subject at
which to start matching.
These two optimizations apply to both pcre_exec() and pcre_dfa_exec().
However, they are not used by pcre_exec() if pcre_study() is called
with the PCRE_STUDY_JIT_COMPILE option, and just-in-time compiling is
successful. The optimizations can be disabled by setting the
PCRE_NO_START_OPTIMIZE option when calling pcre_exec() or
pcre_dfa_exec(). You might want to do this if your pattern contains
callouts or (*MARK) (which cannot be handled by the JIT compiler), and
you want to make use of these facilities in cases where matching fails.
See the discussion of PCRE_NO_START_OPTIMIZE below.
LOCALE SUPPORT
PCRE handles caseless matching, and determines whether characters are
letters, digits, or whatever, by reference to a set of tables, indexed
by character value. When running in UTF-8 mode, this applies only to
characters with codes less than 128. By default, higher-valued codes
never match escapes such as \w or \d, but they can be tested with \p if
PCRE is built with Unicode character property support. Alternatively,
the PCRE_UCP option can be set at compile time; this causes \w and
friends to use Unicode property support instead of built-in tables. The
use of locales with Unicode is discouraged. If you are handling charac-
ters with codes greater than 128, you should either use UTF-8 and Uni-
code, or use locales, but not try to mix the two.
PCRE contains an internal set of tables that are used when the final
argument of pcre_compile() is NULL. These are sufficient for many
applications. Normally, the internal tables recognize only ASCII char-
acters. However, when PCRE is built, it is possible to cause the inter-
nal tables to be rebuilt in the default "C" locale of the local system,
which may cause them to be different.
The internal tables can always be overridden by tables supplied by the
application that calls PCRE. These may be created in a different locale
from the default. As more and more applications change to using Uni-
code, the need for this locale support is expected to die away.
External tables are built by calling the pcre_maketables() function,
which has no arguments, in the relevant locale. The result can then be
passed to pcre_compile() or pcre_exec() as often as necessary. For
example, to build and use tables that are appropriate for the French
locale (where accented characters with values greater than 128 are
treated as letters), the following code could be used:
setlocale(LC_CTYPE, "fr_FR");
tables = pcre_maketables();
re = pcre_compile(..., tables);
The locale name "fr_FR" is used on Linux and other Unix-like systems;
if you are using Windows, the name for the French locale is "french".
When pcre_maketables() runs, the tables are built in memory that is
obtained via pcre_malloc. It is the caller's responsibility to ensure
that the memory containing the tables remains available for as long as
it is needed.
The pointer that is passed to pcre_compile() is saved with the compiled
pattern, and the same tables are used via this pointer by pcre_study()
and normally also by pcre_exec(). Thus, by default, for any single pat-
tern, compilation, studying and matching all happen in the same locale,
but different patterns can be compiled in different locales.
It is possible to pass a table pointer or NULL (indicating the use of
the internal tables) to pcre_exec(). Although not intended for this
purpose, this facility could be used to match a pattern in a different
locale from the one in which it was compiled. Passing table pointers at
run time is discussed below in the section on matching a pattern.
INFORMATION ABOUT A PATTERN
int pcre_fullinfo(const pcre *code, const pcre_extra *extra,
int what, void *where);
The pcre_fullinfo() function returns information about a compiled pat-
tern. It replaces the obsolete pcre_info() function, which is neverthe-
less retained for backwards compability (and is documented below).
The first argument for pcre_fullinfo() is a pointer to the compiled
pattern. The second argument is the result of pcre_study(), or NULL if
the pattern was not studied. The third argument specifies which piece
of information is required, and the fourth argument is a pointer to a
variable to receive the data. The yield of the function is zero for
success, or one of the following negative numbers:
PCRE_ERROR_NULL the argument code was NULL
the argument where was NULL
PCRE_ERROR_BADMAGIC the "magic number" was not found
PCRE_ERROR_BADOPTION the value of what was invalid
The "magic number" is placed at the start of each compiled pattern as
an simple check against passing an arbitrary memory pointer. Here is a
typical call of pcre_fullinfo(), to obtain the length of the compiled
pattern:
int rc;
size_t length;
rc = pcre_fullinfo(
re, /* result of pcre_compile() */
sd, /* result of pcre_study(), or NULL */
PCRE_INFO_SIZE, /* what is required */
&length); /* where to put the data */
The possible values for the third argument are defined in pcre.h, and
are as follows:
PCRE_INFO_BACKREFMAX
Return the number of the highest back reference in the pattern. The
fourth argument should point to an int variable. Zero is returned if
there are no back references.
PCRE_INFO_CAPTURECOUNT
Return the number of capturing subpatterns in the pattern. The fourth
argument should point to an int variable.
PCRE_INFO_DEFAULT_TABLES
Return a pointer to the internal default character tables within PCRE.
The fourth argument should point to an unsigned char * variable. This
information call is provided for internal use by the pcre_study() func-
tion. External callers can cause PCRE to use its internal tables by
passing a NULL table pointer.
PCRE_INFO_FIRSTBYTE
Return information about the first byte of any matched string, for a
non-anchored pattern. The fourth argument should point to an int vari-
able. (This option used to be called PCRE_INFO_FIRSTCHAR; the old name
is still recognized for backwards compatibility.)
If there is a fixed first byte, for example, from a pattern such as
(cat|cow|coyote), its value is returned. Otherwise, if either
(a) the pattern was compiled with the PCRE_MULTILINE option, and every
branch starts with "^", or
(b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not
set (if it were set, the pattern would be anchored),
-1 is returned, indicating that the pattern matches only at the start
of a subject string or after any newline within the string. Otherwise
-2 is returned. For anchored patterns, -2 is returned.
PCRE_INFO_FIRSTTABLE
If the pattern was studied, and this resulted in the construction of a
256-bit table indicating a fixed set of bytes for the first byte in any
matching string, a pointer to the table is returned. Otherwise NULL is
returned. The fourth argument should point to an unsigned char * vari-
able.
PCRE_INFO_HASCRORLF
Return 1 if the pattern contains any explicit matches for CR or LF
characters, otherwise 0. The fourth argument should point to an int
variable. An explicit match is either a literal CR or LF character, or
\r or \n.
PCRE_INFO_JCHANGED
Return 1 if the (?J) or (?-J) option setting is used in the pattern,
otherwise 0. The fourth argument should point to an int variable. (?J)
and (?-J) set and unset the local PCRE_DUPNAMES option, respectively.
PCRE_INFO_JIT
Return 1 if the pattern was studied with the PCRE_STUDY_JIT_COMPILE
option, and just-in-time compiling was successful. The fourth argument
should point to an int variable. A return value of 0 means that JIT
support is not available in this version of PCRE, or that the pattern
was not studied with the PCRE_STUDY_JIT_COMPILE option, or that the JIT
compiler could not handle this particular pattern. See the pcrejit doc-
umentation for details of what can and cannot be handled.
PCRE_INFO_JITSIZE
If the pattern was successfully studied with the PCRE_STUDY_JIT_COMPILE
option, return the size of the JIT compiled code, otherwise return
zero. The fourth argument should point to a size_t variable.
PCRE_INFO_LASTLITERAL
Return the value of the rightmost literal byte that must exist in any
matched string, other than at its start, if such a byte has been
recorded. The fourth argument should point to an int variable. If there
is no such byte, -1 is returned. For anchored patterns, a last literal
byte is recorded only if it follows something of variable length. For
example, for the pattern /^a\d+z\d+/ the returned value is "z", but for
/^a\dz\d/ the returned value is -1.
PCRE_INFO_MINLENGTH
If the pattern was studied and a minimum length for matching subject
strings was computed, its value is returned. Otherwise the returned
value is -1. The value is a number of characters, not bytes (this may
be relevant in UTF-8 mode). The fourth argument should point to an int
variable. A non-negative value is a lower bound to the length of any
matching string. There may not be any strings of that length that do
actually match, but every string that does match is at least that long.
PCRE_INFO_NAMECOUNT
PCRE_INFO_NAMEENTRYSIZE
PCRE_INFO_NAMETABLE
PCRE supports the use of named as well as numbered capturing parenthe-
ses. The names are just an additional way of identifying the parenthe-
ses, which still acquire numbers. Several convenience functions such as
pcre_get_named_substring() are provided for extracting captured sub-
strings by name. It is also possible to extract the data directly, by
first converting the name to a number in order to access the correct
pointers in the output vector (described with pcre_exec() below). To do
the conversion, you need to use the name-to-number map, which is
described by these three values.
The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT
gives the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size
of each entry; both of these return an int value. The entry size
depends on the length of the longest name. PCRE_INFO_NAMETABLE returns
a pointer to the first entry of the table (a pointer to char). The
first two bytes of each entry are the number of the capturing parenthe-
sis, most significant byte first. The rest of the entry is the corre-
sponding name, zero terminated.
The names are in alphabetical order. Duplicate names may appear if (?|
is used to create multiple groups with the same number, as described in
the section on duplicate subpattern numbers in the pcrepattern page.
Duplicate names for subpatterns with different numbers are permitted
only if PCRE_DUPNAMES is set. In all cases of duplicate names, they
appear in the table in the order in which they were found in the pat-
tern. In the absence of (?| this is the order of increasing number;
when (?| is used this is not necessarily the case because later subpat-
terns may have lower numbers.
As a simple example of the name/number table, consider the following
pattern (assume PCRE_EXTENDED is set, so white space - including new-
lines - is ignored):
(?<date> (?<year>(\d\d)?\d\d) -
(?<month>\d\d) - (?<day>\d\d) )
There are four named subpatterns, so the table has four entries, and
each entry in the table is eight bytes long. The table is as follows,
with non-printing bytes shows in hexadecimal, and undefined bytes shown
as ??:
00 01 d a t e 00 ??
00 05 d a y 00 ?? ??
00 04 m o n t h 00
00 02 y e a r 00 ??
When writing code to extract data from named subpatterns using the
name-to-number map, remember that the length of the entries is likely
to be different for each compiled pattern.
PCRE_INFO_OKPARTIAL
Return 1 if the pattern can be used for partial matching with
pcre_exec(), otherwise 0. The fourth argument should point to an int
variable. From release 8.00, this always returns 1, because the
restrictions that previously applied to partial matching have been
lifted. The pcrepartial documentation gives details of partial match-
ing.
PCRE_INFO_OPTIONS
Return a copy of the options with which the pattern was compiled. The
fourth argument should point to an unsigned long int variable. These
option bits are those specified in the call to pcre_compile(), modified
by any top-level option settings at the start of the pattern itself. In
other words, they are the options that will be in force when matching
starts. For example, if the pattern /(?im)abc(?-i)d/ is compiled with
the PCRE_EXTENDED option, the result is PCRE_CASELESS, PCRE_MULTILINE,
and PCRE_EXTENDED.
A pattern is automatically anchored by PCRE if all of its top-level
alternatives begin with one of the following:
^ unless PCRE_MULTILINE is set
\A always
\G always
.* if PCRE_DOTALL is set and there are no back
references to the subpattern in which .* appears
For such patterns, the PCRE_ANCHORED bit is set in the options returned
by pcre_fullinfo().
PCRE_INFO_SIZE
Return the size of the compiled pattern. The fourth argument should
point to a size_t variable. This value does not include the size of the
pcre structure that is returned by pcre_compile(). The value that is
passed as the argument to pcre_malloc() when pcre_compile() is getting
memory in which to place the compiled data is the value returned by
this option plus the size of the pcre structure. Studying a compiled
pattern, with or without JIT, does not alter the value returned by this
option.
PCRE_INFO_STUDYSIZE
Return the size of the data block pointed to by the study_data field in
a pcre_extra block. If pcre_extra is NULL, or there is no study data,
zero is returned. The fourth argument should point to a size_t vari-
able. The study_data field is set by pcre_study() to record informa-
tion that will speed up matching (see the section entitled "Studying a
pattern" above). The format of the study_data block is private, but its
length is made available via this option so that it can be saved and
restored (see the pcreprecompile documentation for details).
OBSOLETE INFO FUNCTION
int pcre_info(const pcre *code, int *optptr, int *firstcharptr);
The pcre_info() function is now obsolete because its interface is too
restrictive to return all the available data about a compiled pattern.
New programs should use pcre_fullinfo() instead. The yield of
pcre_info() is the number of capturing subpatterns, or one of the fol-
lowing negative numbers:
PCRE_ERROR_NULL the argument code was NULL
PCRE_ERROR_BADMAGIC the "magic number" was not found
If the optptr argument is not NULL, a copy of the options with which
the pattern was compiled is placed in the integer it points to (see
PCRE_INFO_OPTIONS above).
If the pattern is not anchored and the firstcharptr argument is not
NULL, it is used to pass back information about the first character of
any matched string (see PCRE_INFO_FIRSTBYTE above).
REFERENCE COUNTS
int pcre_refcount(pcre *code, int adjust);
The pcre_refcount() function is used to maintain a reference count in
the data block that contains a compiled pattern. It is provided for the
benefit of applications that operate in an object-oriented manner,
where different parts of the application may be using the same compiled
pattern, but you want to free the block when they are all done.
When a pattern is compiled, the reference count field is initialized to
zero. It is changed only by calling this function, whose action is to
add the adjust value (which may be positive or negative) to it. The
yield of the function is the new value. However, the value of the count
is constrained to lie between 0 and 65535, inclusive. If the new value
is outside these limits, it is forced to the appropriate limit value.
Except when it is zero, the reference count is not correctly preserved
if a pattern is compiled on one host and then transferred to a host
whose byte-order is different. (This seems a highly unlikely scenario.)
MATCHING A PATTERN: THE TRADITIONAL FUNCTION
int pcre_exec(const pcre *code, const pcre_extra *extra,
const char *subject, int length, int startoffset,
int options, int *ovector, int ovecsize);
The function pcre_exec() is called to match a subject string against a
compiled pattern, which is passed in the code argument. If the pattern
was studied, the result of the study should be passed in the extra
argument. You can call pcre_exec() with the same code and extra argu-
ments as many times as you like, in order to match different subject
strings with the same pattern.
This function is the main matching facility of the library, and it
operates in a Perl-like manner. For specialist use there is also an
alternative matching function, which is described below in the section
about the pcre_dfa_exec() function.
In most applications, the pattern will have been compiled (and option-
ally studied) in the same process that calls pcre_exec(). However, it
is possible to save compiled patterns and study data, and then use them
later in different processes, possibly even on different hosts. For a
discussion about this, see the pcreprecompile documentation.
Here is an example of a simple call to pcre_exec():
int rc;
int ovector[30];
rc = pcre_exec(
re, /* result of pcre_compile() */
NULL, /* we didn't study the pattern */
"some string", /* the subject string */
11, /* the length of the subject string */
0, /* start at offset 0 in the subject */
0, /* default options */
ovector, /* vector of integers for substring information */
30); /* number of elements (NOT size in bytes) */
Extra data for pcre_exec()
If the extra argument is not NULL, it must point to a pcre_extra data
block. The pcre_study() function returns such a block (when it doesn't
return NULL), but you can also create one for yourself, and pass addi-
tional information in it. The pcre_extra block contains the following
fields (not necessarily in this order):
unsigned long int flags;
void *study_data;
void *executable_jit;
unsigned long int match_limit;
unsigned long int match_limit_recursion;
void *callout_data;
const unsigned char *tables;
unsigned char **mark;
The flags field is a bitmap that specifies which of the other fields
are set. The flag bits are:
PCRE_EXTRA_STUDY_DATA
PCRE_EXTRA_EXECUTABLE_JIT
PCRE_EXTRA_MATCH_LIMIT
PCRE_EXTRA_MATCH_LIMIT_RECURSION
PCRE_EXTRA_CALLOUT_DATA
PCRE_EXTRA_TABLES
PCRE_EXTRA_MARK
Other flag bits should be set to zero. The study_data field and some-
times the executable_jit field are set in the pcre_extra block that is
returned by pcre_study(), together with the appropriate flag bits. You
should not set these yourself, but you may add to the block by setting
the other fields and their corresponding flag bits.
The match_limit field provides a means of preventing PCRE from using up
a vast amount of resources when running patterns that are not going to
match, but which have a very large number of possibilities in their
search trees. The classic example is a pattern that uses nested unlim-
ited repeats.
Internally, pcre_exec() uses a function called match(), which it calls
repeatedly (sometimes recursively). The limit set by match_limit is
imposed on the number of times this function is called during a match,
which has the effect of limiting the amount of backtracking that can
take place. For patterns that are not anchored, the count restarts from
zero for each position in the subject string.
When pcre_exec() is called with a pattern that was successfully studied
with the PCRE_STUDY_JIT_COMPILE option, the way that the matching is
executed is entirely different. However, there is still the possibility
of runaway matching that goes on for a very long time, and so the
match_limit value is also used in this case (but in a different way) to
limit how long the matching can continue.
The default value for the limit can be set when PCRE is built; the
default default is 10 million, which handles all but the most extreme
cases. You can override the default by suppling pcre_exec() with a
pcre_extra block in which match_limit is set, and
PCRE_EXTRA_MATCH_LIMIT is set in the flags field. If the limit is
exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
The match_limit_recursion field is similar to match_limit, but instead
of limiting the total number of times that match() is called, it limits
the depth of recursion. The recursion depth is a smaller number than
the total number of calls, because not all calls to match() are recur-
sive. This limit is of use only if it is set smaller than match_limit.
Limiting the recursion depth limits the amount of machine stack that
can be used, or, when PCRE has been compiled to use memory on the heap
instead of the stack, the amount of heap memory that can be used. This
limit is not relevant, and is ignored, if the pattern was successfully
studied with PCRE_STUDY_JIT_COMPILE.
The default value for match_limit_recursion can be set when PCRE is
built; the default default is the same value as the default for
match_limit. You can override the default by suppling pcre_exec() with
a pcre_extra block in which match_limit_recursion is set, and
PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the flags field. If the
limit is exceeded, pcre_exec() returns PCRE_ERROR_RECURSIONLIMIT.
The callout_data field is used in conjunction with the "callout" fea-
ture, and is described in the pcrecallout documentation.
The tables field is used to pass a character tables pointer to
pcre_exec(); this overrides the value that is stored with the compiled
pattern. A non-NULL value is stored with the compiled pattern only if
custom tables were supplied to pcre_compile() via its tableptr argu-
ment. If NULL is passed to pcre_exec() using this mechanism, it forces
PCRE's internal tables to be used. This facility is helpful when re-
using patterns that have been saved after compiling with an external
set of tables, because the external tables might be at a different
address when pcre_exec() is called. See the pcreprecompile documenta-
tion for a discussion of saving compiled patterns for later use.
If PCRE_EXTRA_MARK is set in the flags field, the mark field must be
set to point to a char * variable. If the pattern contains any back-
tracking control verbs such as (*MARK:NAME), and the execution ends up
with a name to pass back, a pointer to the name string (zero termi-
nated) is placed in the variable pointed to by the mark field. The
names are within the compiled pattern; if you wish to retain such a
name you must copy it before freeing the memory of a compiled pattern.
If there is no name to pass back, the variable pointed to by the mark
field set to NULL. For details of the backtracking control verbs, see
the section entitled "Backtracking control" in the pcrepattern documen-
tation.
Option bits for pcre_exec()
The unused bits of the options argument for pcre_exec() must be zero.
The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_xxx,
PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
PCRE_NO_START_OPTIMIZE, PCRE_NO_UTF8_CHECK, PCRE_PARTIAL_SOFT, and
PCRE_PARTIAL_HARD.
If the pattern was successfully studied with the PCRE_STUDY_JIT_COMPILE
option, the only supported options for JIT execution are
PCRE_NO_UTF8_CHECK, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, and
PCRE_NOTEMPTY_ATSTART. Note in particular that partial matching is not
supported. If an unsupported option is used, JIT execution is disabled
and the normal interpretive code in pcre_exec() is run.
PCRE_ANCHORED
The PCRE_ANCHORED option limits pcre_exec() to matching at the first
matching position. If a pattern was compiled with PCRE_ANCHORED, or
turned out to be anchored by virtue of its contents, it cannot be made
unachored at matching time.
PCRE_BSR_ANYCRLF
PCRE_BSR_UNICODE
These options (which are mutually exclusive) control what the \R escape
sequence matches. The choice is either to match only CR, LF, or CRLF,
or to match any Unicode newline sequence. These options override the
choice that was made or defaulted when the pattern was compiled.
PCRE_NEWLINE_CR
PCRE_NEWLINE_LF
PCRE_NEWLINE_CRLF
PCRE_NEWLINE_ANYCRLF
PCRE_NEWLINE_ANY
These options override the newline definition that was chosen or
defaulted when the pattern was compiled. For details, see the descrip-
tion of pcre_compile() above. During matching, the newline choice
affects the behaviour of the dot, circumflex, and dollar metacharac-
ters. It may also alter the way the match position is advanced after a
match failure for an unanchored pattern.
When PCRE_NEWLINE_CRLF, PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is
set, and a match attempt for an unanchored pattern fails when the cur-
rent position is at a CRLF sequence, and the pattern contains no
explicit matches for CR or LF characters, the match position is
advanced by two characters instead of one, in other words, to after the
CRLF.
The above rule is a compromise that makes the most common cases work as
expected. For example, if the pattern is .+A (and the PCRE_DOTALL
option is not set), it does not match the string "\r\nA" because, after
failing at the start, it skips both the CR and the LF before retrying.
However, the pattern [\r\n]A does match that string, because it con-
tains an explicit CR or LF reference, and so advances only by one char-
acter after the first failure.
An explicit match for CR of LF is either a literal appearance of one of
those characters, or one of the \r or \n escape sequences. Implicit
matches such as [^X] do not count, nor does \s (which includes CR and
LF in the characters that it matches).
Notwithstanding the above, anomalous effects may still occur when CRLF
is a valid newline sequence and explicit \r or \n escapes appear in the
pattern.
PCRE_NOTBOL
This option specifies that first character of the subject string is not
the beginning of a line, so the circumflex metacharacter should not
match before it. Setting this without PCRE_MULTILINE (at compile time)
causes circumflex never to match. This option affects only the behav-
iour of the circumflex metacharacter. It does not affect \A.
PCRE_NOTEOL
This option specifies that the end of the subject string is not the end
of a line, so the dollar metacharacter should not match it nor (except
in multiline mode) a newline immediately before it. Setting this with-
out PCRE_MULTILINE (at compile time) causes dollar never to match. This
option affects only the behaviour of the dollar metacharacter. It does
not affect \Z or \z.
PCRE_NOTEMPTY
An empty string is not considered to be a valid match if this option is
set. If there are alternatives in the pattern, they are tried. If all
the alternatives match the empty string, the entire match fails. For
example, if the pattern
a?b?
is applied to a string not beginning with "a" or "b", it matches an
empty string at the start of the subject. With PCRE_NOTEMPTY set, this
match is not valid, so PCRE searches further into the string for occur-
rences of "a" or "b".
PCRE_NOTEMPTY_ATSTART
This is like PCRE_NOTEMPTY, except that an empty string match that is
not at the start of the subject is permitted. If the pattern is
anchored, such a match can occur only if the pattern contains \K.
Perl has no direct equivalent of PCRE_NOTEMPTY or
PCRE_NOTEMPTY_ATSTART, but it does make a special case of a pattern
match of the empty string within its split() function, and when using
the /g modifier. It is possible to emulate Perl's behaviour after
matching a null string by first trying the match again at the same off-
set with PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED, and then if that
fails, by advancing the starting offset (see below) and trying an ordi-
nary match again. There is some code that demonstrates how to do this
in the pcredemo sample program. In the most general case, you have to
check to see if the newline convention recognizes CRLF as a newline,
and if so, and the current character is CR followed by LF, advance the
starting offset by two characters instead of one.
PCRE_NO_START_OPTIMIZE
There are a number of optimizations that pcre_exec() uses at the start
of a match, in order to speed up the process. For example, if it is
known that an unanchored match must start with a specific character, it
searches the subject for that character, and fails immediately if it
cannot find it, without actually running the main matching function.
This means that a special item such as (*COMMIT) at the start of a pat-
tern is not considered until after a suitable starting point for the
match has been found. When callouts or (*MARK) items are in use, these
"start-up" optimizations can cause them to be skipped if the pattern is
never actually used. The start-up optimizations are in effect a pre-
scan of the subject that takes place before the pattern is run.
The PCRE_NO_START_OPTIMIZE option disables the start-up optimizations,
possibly causing performance to suffer, but ensuring that in cases
where the result is "no match", the callouts do occur, and that items
such as (*COMMIT) and (*MARK) are considered at every possible starting
position in the subject string. If PCRE_NO_START_OPTIMIZE is set at
compile time, it cannot be unset at matching time.
Setting PCRE_NO_START_OPTIMIZE can change the outcome of a matching
operation. Consider the pattern
(*COMMIT)ABC
When this is compiled, PCRE records the fact that a match must start
with the character "A". Suppose the subject string is "DEFABC". The
start-up optimization scans along the subject, finds "A" and runs the
first match attempt from there. The (*COMMIT) item means that the pat-
tern must match the current starting position, which in this case, it
does. However, if the same match is run with PCRE_NO_START_OPTIMIZE
set, the initial scan along the subject string does not happen. The
first match attempt is run starting from "D" and when this fails,
(*COMMIT) prevents any further matches being tried, so the overall
result is "no match". If the pattern is studied, more start-up opti-
mizations may be used. For example, a minimum length for the subject
may be recorded. Consider the pattern
(*MARK:A)(X|Y)
The minimum length for a match is one character. If the subject is
"ABC", there will be attempts to match "ABC", "BC", "C", and then
finally an empty string. If the pattern is studied, the final attempt
does not take place, because PCRE knows that the subject is too short,
and so the (*MARK) is never encountered. In this case, studying the
pattern does not affect the overall match result, which is still "no
match", but it does affect the auxiliary information that is returned.
PCRE_NO_UTF8_CHECK
When PCRE_UTF8 is set at compile time, the validity of the subject as a
UTF-8 string is automatically checked when pcre_exec() is subsequently
called. The value of startoffset is also checked to ensure that it
points to the start of a UTF-8 character. There is a discussion about
the validity of UTF-8 strings in the section on UTF-8 support in the
main pcre page. If an invalid UTF-8 sequence of bytes is found,
pcre_exec() returns the error PCRE_ERROR_BADUTF8 or, if PCRE_PAR-
TIAL_HARD is set and the problem is a truncated UTF-8 character at the
end of the subject, PCRE_ERROR_SHORTUTF8. In both cases, information
about the precise nature of the error may also be returned (see the
descriptions of these errors in the section entitled Error return val-
ues from pcre_exec() below). If startoffset contains a value that does
not point to the start of a UTF-8 character (or to the end of the sub-
ject), PCRE_ERROR_BADUTF8_OFFSET is returned.
If you already know that your subject is valid, and you want to skip
these checks for performance reasons, you can set the
PCRE_NO_UTF8_CHECK option when calling pcre_exec(). You might want to
do this for the second and subsequent calls to pcre_exec() if you are
making repeated calls to find all the matches in a single subject
string. However, you should be sure that the value of startoffset
points to the start of a UTF-8 character (or the end of the subject).
When PCRE_NO_UTF8_CHECK is set, the effect of passing an invalid UTF-8
string as a subject or an invalid value of startoffset is undefined.
Your program may crash.
PCRE_PARTIAL_HARD
PCRE_PARTIAL_SOFT
These options turn on the partial matching feature. For backwards com-
patibility, PCRE_PARTIAL is a synonym for PCRE_PARTIAL_SOFT. A partial
match occurs if the end of the subject string is reached successfully,
but there are not enough subject characters to complete the match. If
this happens when PCRE_PARTIAL_SOFT (but not PCRE_PARTIAL_HARD) is set,
matching continues by testing any remaining alternatives. Only if no
complete match can be found is PCRE_ERROR_PARTIAL returned instead of
PCRE_ERROR_NOMATCH. In other words, PCRE_PARTIAL_SOFT says that the
caller is prepared to handle a partial match, but only if no complete
match can be found.
If PCRE_PARTIAL_HARD is set, it overrides PCRE_PARTIAL_SOFT. In this
case, if a partial match is found, pcre_exec() immediately returns
PCRE_ERROR_PARTIAL, without considering any other alternatives. In
other words, when PCRE_PARTIAL_HARD is set, a partial match is consid-
ered to be more important that an alternative complete match.
In both cases, the portion of the string that was inspected when the
partial match was found is set as the first matching string. There is a
more detailed discussion of partial and multi-segment matching, with
examples, in the pcrepartial documentation.
The string to be matched by pcre_exec()
The subject string is passed to pcre_exec() as a pointer in subject, a
length (in bytes) in length, and a starting byte offset in startoffset.
If this is negative or greater than the length of the subject,
pcre_exec() returns PCRE_ERROR_BADOFFSET. When the starting offset is
zero, the search for a match starts at the beginning of the subject,
and this is by far the most common case. In UTF-8 mode, the byte offset
must point to the start of a UTF-8 character (or the end of the sub-
ject). Unlike the pattern string, the subject may contain binary zero
bytes.
A non-zero starting offset is useful when searching for another match
in the same subject by calling pcre_exec() again after a previous suc-
cess. Setting startoffset differs from just passing over a shortened
string and setting PCRE_NOTBOL in the case of a pattern that begins
with any kind of lookbehind. For example, consider the pattern
\Biss\B
which finds occurrences of "iss" in the middle of words. (\B matches
only if the current position in the subject is not a word boundary.)
When applied to the string "Mississipi" the first call to pcre_exec()
finds the first occurrence. If pcre_exec() is called again with just
the remainder of the subject, namely "issipi", it does not match,
because \B is always false at the start of the subject, which is deemed
to be a word boundary. However, if pcre_exec() is passed the entire
string again, but with startoffset set to 4, it finds the second occur-
rence of "iss" because it is able to look behind the starting point to
discover that it is preceded by a letter.
Finding all the matches in a subject is tricky when the pattern can
match an empty string. It is possible to emulate Perl's /g behaviour by
first trying the match again at the same offset, with the
PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED options, and then if that
fails, advancing the starting offset and trying an ordinary match
again. There is some code that demonstrates how to do this in the pcre-
demo sample program. In the most general case, you have to check to see
if the newline convention recognizes CRLF as a newline, and if so, and
the current character is CR followed by LF, advance the starting offset
by two characters instead of one.
If a non-zero starting offset is passed when the pattern is anchored,
one attempt to match at the given offset is made. This can only succeed
if the pattern does not require the match to be at the start of the
subject.
How pcre_exec() returns captured substrings
In general, a pattern matches a certain portion of the subject, and in
addition, further substrings from the subject may be picked out by
parts of the pattern. Following the usage in Jeffrey Friedl's book,
this is called "capturing" in what follows, and the phrase "capturing
subpattern" is used for a fragment of a pattern that picks out a sub-
string. PCRE supports several other kinds of parenthesized subpattern
that do not cause substrings to be captured.
Captured substrings are returned to the caller via a vector of integers
whose address is passed in ovector. The number of elements in the vec-
tor is passed in ovecsize, which must be a non-negative number. Note:
this argument is NOT the size of ovector in bytes.
The first two-thirds of the vector is used to pass back captured sub-
strings, each substring using a pair of integers. The remaining third
of the vector is used as workspace by pcre_exec() while matching cap-
turing subpatterns, and is not available for passing back information.
The number passed in ovecsize should always be a multiple of three. If
it is not, it is rounded down.
When a match is successful, information about captured substrings is
returned in pairs of integers, starting at the beginning of ovector,
and continuing up to two-thirds of its length at the most. The first
element of each pair is set to the byte offset of the first character
in a substring, and the second is set to the byte offset of the first
character after the end of a substring. Note: these values are always
byte offsets, even in UTF-8 mode. They are not character counts.
The first pair of integers, ovector[0] and ovector[1], identify the
portion of the subject string matched by the entire pattern. The next
pair is used for the first capturing subpattern, and so on. The value
returned by pcre_exec() is one more than the highest numbered pair that
has been set. For example, if two substrings have been captured, the
returned value is 3. If there are no capturing subpatterns, the return
value from a successful match is 1, indicating that just the first pair
of offsets has been set.
If a capturing subpattern is matched repeatedly, it is the last portion
of the string that it matched that is returned.
If the vector is too small to hold all the captured substring offsets,
it is used as far as possible (up to two-thirds of its length), and the
function returns a value of zero. If neither the actual string matched
not any captured substrings are of interest, pcre_exec() may be called
with ovector passed as NULL and ovecsize as zero. However, if the pat-
tern contains back references and the ovector is not big enough to
remember the related substrings, PCRE has to get additional memory for
use during matching. Thus it is usually advisable to supply an ovector
of reasonable size.
There are some cases where zero is returned (indicating vector over-
flow) when in fact the vector is exactly the right size for the final
match. For example, consider the pattern
(a)(?:(b)c|bd)
If a vector of 6 elements (allowing for only 1 captured substring) is
given with subject string "abd", pcre_exec() will try to set the second
captured string, thereby recording a vector overflow, before failing to
match "c" and backing up to try the second alternative. The zero
return, however, does correctly indicate that the maximum number of
slots (namely 2) have been filled. In similar cases where there is tem-
porary overflow, but the final number of used slots is actually less
than the maximum, a non-zero value is returned.
The pcre_fullinfo() function can be used to find out how many capturing
subpatterns there are in a compiled pattern. The smallest size for
ovector that will allow for n captured substrings, in addition to the
offsets of the substring matched by the whole pattern, is (n+1)*3.
It is possible for capturing subpattern number n+1 to match some part
of the subject when subpattern n has not been used at all. For example,
if the string "abc" is matched against the pattern (a|(z))(bc) the
return from the function is 4, and subpatterns 1 and 3 are matched, but
2 is not. When this happens, both values in the offset pairs corre-
sponding to unused subpatterns are set to -1.
Offset values that correspond to unused subpatterns at the end of the
expression are also set to -1. For example, if the string "abc" is
matched against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not
matched. The return from the function is 2, because the highest used
capturing subpattern number is 1, and the offsets for for the second
and third capturing subpatterns (assuming the vector is large enough,
of course) are set to -1.
Note: Elements in the first two-thirds of ovector that do not corre-
spond to capturing parentheses in the pattern are never changed. That
is, if a pattern contains n capturing parentheses, no more than ovec-
tor[0] to ovector[2n+1] are set by pcre_exec(). The other elements (in
the first two-thirds) retain whatever values they previously had.
Some convenience functions are provided for extracting the captured
substrings as separate strings. These are described below.
Error return values from pcre_exec()
If pcre_exec() fails, it returns a negative number. The following are
defined in the header file:
PCRE_ERROR_NOMATCH (-1)
The subject string did not match the pattern.
PCRE_ERROR_NULL (-2)
Either code or subject was passed as NULL, or ovector was NULL and
ovecsize was not zero.
PCRE_ERROR_BADOPTION (-3)
An unrecognized bit was set in the options argument.
PCRE_ERROR_BADMAGIC (-4)
PCRE stores a 4-byte "magic number" at the start of the compiled code,
to catch the case when it is passed a junk pointer and to detect when a
pattern that was compiled in an environment of one endianness is run in
an environment with the other endianness. This is the error that PCRE
gives when the magic number is not present.
PCRE_ERROR_UNKNOWN_OPCODE (-5)
While running the pattern match, an unknown item was encountered in the
compiled pattern. This error could be caused by a bug in PCRE or by
overwriting of the compiled pattern.
PCRE_ERROR_NOMEMORY (-6)
If a pattern contains back references, but the ovector that is passed
to pcre_exec() is not big enough to remember the referenced substrings,
PCRE gets a block of memory at the start of matching to use for this
purpose. If the call via pcre_malloc() fails, this error is given. The
memory is automatically freed at the end of matching.
This error is also given if pcre_stack_malloc() fails in pcre_exec().
This can happen only when PCRE has been compiled with --disable-stack-
for-recursion.
PCRE_ERROR_NOSUBSTRING (-7)
This error is used by the pcre_copy_substring(), pcre_get_substring(),
and pcre_get_substring_list() functions (see below). It is never
returned by pcre_exec().
PCRE_ERROR_MATCHLIMIT (-8)
The backtracking limit, as specified by the match_limit field in a
pcre_extra structure (or defaulted) was reached. See the description
above.
PCRE_ERROR_CALLOUT (-9)
This error is never generated by pcre_exec() itself. It is provided for
use by callout functions that want to yield a distinctive error code.
See the pcrecallout documentation for details.
PCRE_ERROR_BADUTF8 (-10)
A string that contains an invalid UTF-8 byte sequence was passed as a
subject, and the PCRE_NO_UTF8_CHECK option was not set. If the size of
the output vector (ovecsize) is at least 2, the byte offset to the
start of the the invalid UTF-8 character is placed in the first ele-
ment, and a reason code is placed in the second element. The reason
codes are listed in the following section. For backward compatibility,
if PCRE_PARTIAL_HARD is set and the problem is a truncated UTF-8 char-
acter at the end of the subject (reason codes 1 to 5),
PCRE_ERROR_SHORTUTF8 is returned instead of PCRE_ERROR_BADUTF8.
PCRE_ERROR_BADUTF8_OFFSET (-11)
The UTF-8 byte sequence that was passed as a subject was checked and
found to be valid (the PCRE_NO_UTF8_CHECK option was not set), but the
value of startoffset did not point to the beginning of a UTF-8 charac-
ter or the end of the subject.
PCRE_ERROR_PARTIAL (-12)
The subject string did not match, but it did match partially. See the
pcrepartial documentation for details of partial matching.
PCRE_ERROR_BADPARTIAL (-13)
This code is no longer in use. It was formerly returned when the
PCRE_PARTIAL option was used with a compiled pattern containing items
that were not supported for partial matching. From release 8.00
onwards, there are no restrictions on partial matching.
PCRE_ERROR_INTERNAL (-14)
An unexpected internal error has occurred. This error could be caused
by a bug in PCRE or by overwriting of the compiled pattern.
PCRE_ERROR_BADCOUNT (-15)
This error is given if the value of the ovecsize argument is negative.
PCRE_ERROR_RECURSIONLIMIT (-21)
The internal recursion limit, as specified by the match_limit_recursion
field in a pcre_extra structure (or defaulted) was reached. See the
description above.
PCRE_ERROR_BADNEWLINE (-23)
An invalid combination of PCRE_NEWLINE_xxx options was given.
PCRE_ERROR_BADOFFSET (-24)
The value of startoffset was negative or greater than the length of the
subject, that is, the value in length.
PCRE_ERROR_SHORTUTF8 (-25)
This error is returned instead of PCRE_ERROR_BADUTF8 when the subject
string ends with a truncated UTF-8 character and the PCRE_PARTIAL_HARD
option is set. Information about the failure is returned as for
PCRE_ERROR_BADUTF8. It is in fact sufficient to detect this case, but
this special error code for PCRE_PARTIAL_HARD precedes the implementa-
tion of returned information; it is retained for backwards compatibil-
ity.
PCRE_ERROR_RECURSELOOP (-26)
This error is returned when pcre_exec() detects a recursion loop within
the pattern. Specifically, it means that either the whole pattern or a
subpattern has been called recursively for the second time at the same
position in the subject string. Some simple patterns that might do this
are detected and faulted at compile time, but more complicated cases,
in particular mutual recursions between two different subpatterns, can-
not be detected until run time.
PCRE_ERROR_JIT_STACKLIMIT (-27)
This error is returned when a pattern that was successfully studied
using the PCRE_STUDY_JIT_COMPILE option is being matched, but the mem-
ory available for the just-in-time processing stack is not large
enough. See the pcrejit documentation for more details.
Error numbers -16 to -20 and -22 are not used by pcre_exec().
Reason codes for invalid UTF-8 strings
When pcre_exec() returns either PCRE_ERROR_BADUTF8 or PCRE_ERROR_SHORT-
UTF8, and the size of the output vector (ovecsize) is at least 2, the
offset of the start of the invalid UTF-8 character is placed in the
first output vector element (ovector[0]) and a reason code is placed in
the second element (ovector[1]). The reason codes are given names in
the pcre.h header file:
PCRE_UTF8_ERR1
PCRE_UTF8_ERR2
PCRE_UTF8_ERR3
PCRE_UTF8_ERR4
PCRE_UTF8_ERR5
The string ends with a truncated UTF-8 character; the code specifies
how many bytes are missing (1 to 5). Although RFC 3629 restricts UTF-8
characters to be no longer than 4 bytes, the encoding scheme (origi-
nally defined by RFC 2279) allows for up to 6 bytes, and this is
checked first; hence the possibility of 4 or 5 missing bytes.
PCRE_UTF8_ERR6
PCRE_UTF8_ERR7
PCRE_UTF8_ERR8
PCRE_UTF8_ERR9
PCRE_UTF8_ERR10
The two most significant bits of the 2nd, 3rd, 4th, 5th, or 6th byte of
the character do not have the binary value 0b10 (that is, either the
most significant bit is 0, or the next bit is 1).
PCRE_UTF8_ERR11
PCRE_UTF8_ERR12
A character that is valid by the RFC 2279 rules is either 5 or 6 bytes
long; these code points are excluded by RFC 3629.
PCRE_UTF8_ERR13
A 4-byte character has a value greater than 0x10fff; these code points
are excluded by RFC 3629.
PCRE_UTF8_ERR14
A 3-byte character has a value in the range 0xd800 to 0xdfff; this
range of code points are reserved by RFC 3629 for use with UTF-16, and
so are excluded from UTF-8.
PCRE_UTF8_ERR15
PCRE_UTF8_ERR16
PCRE_UTF8_ERR17
PCRE_UTF8_ERR18
PCRE_UTF8_ERR19
A 2-, 3-, 4-, 5-, or 6-byte character is "overlong", that is, it codes
for a value that can be represented by fewer bytes, which is invalid.
For example, the two bytes 0xc0, 0xae give the value 0x2e, whose cor-
rect coding uses just one byte.
PCRE_UTF8_ERR20
The two most significant bits of the first byte of a character have the
binary value 0b10 (that is, the most significant bit is 1 and the sec-
ond is 0). Such a byte can only validly occur as the second or subse-
quent byte of a multi-byte character.
PCRE_UTF8_ERR21
The first byte of a character has the value 0xfe or 0xff. These values
can never occur in a valid UTF-8 string.
EXTRACTING CAPTURED SUBSTRINGS BY NUMBER
int pcre_copy_substring(const char *subject, int *ovector,
int stringcount, int stringnumber, char *buffer,
int buffersize);
int pcre_get_substring(const char *subject, int *ovector,
int stringcount, int stringnumber,
const char **stringptr);
int pcre_get_substring_list(const char *subject,
int *ovector, int stringcount, const char ***listptr);
Captured substrings can be accessed directly by using the offsets
returned by pcre_exec() in ovector. For convenience, the functions
pcre_copy_substring(), pcre_get_substring(), and pcre_get_sub-
string_list() are provided for extracting captured substrings as new,
separate, zero-terminated strings. These functions identify substrings
by number. The next section describes functions for extracting named
substrings.
A substring that contains a binary zero is correctly extracted and has
a further zero added on the end, but the result is not, of course, a C
string. However, you can process such a string by referring to the
length that is returned by pcre_copy_substring() and pcre_get_sub-
string(). Unfortunately, the interface to pcre_get_substring_list() is
not adequate for handling strings containing binary zeros, because the
end of the final string is not independently indicated.
The first three arguments are the same for all three of these func-
tions: subject is the subject string that has just been successfully
matched, ovector is a pointer to the vector of integer offsets that was
passed to pcre_exec(), and stringcount is the number of substrings that
were captured by the match, including the substring that matched the
entire regular expression. This is the value returned by pcre_exec() if
it is greater than zero. If pcre_exec() returned zero, indicating that
it ran out of space in ovector, the value passed as stringcount should
be the number of elements in the vector divided by three.
The functions pcre_copy_substring() and pcre_get_substring() extract a
single substring, whose number is given as stringnumber. A value of
zero extracts the substring that matched the entire pattern, whereas
higher values extract the captured substrings. For pcre_copy_sub-
string(), the string is placed in buffer, whose length is given by
buffersize, while for pcre_get_substring() a new block of memory is
obtained via pcre_malloc, and its address is returned via stringptr.
The yield of the function is the length of the string, not including
the terminating zero, or one of these error codes:
PCRE_ERROR_NOMEMORY (-6)
The buffer was too small for pcre_copy_substring(), or the attempt to
get memory failed for pcre_get_substring().
PCRE_ERROR_NOSUBSTRING (-7)
There is no substring whose number is stringnumber.
The pcre_get_substring_list() function extracts all available sub-
strings and builds a list of pointers to them. All this is done in a
single block of memory that is obtained via pcre_malloc. The address of
the memory block is returned via listptr, which is also the start of
the list of string pointers. The end of the list is marked by a NULL
pointer. The yield of the function is zero if all went well, or the
error code
PCRE_ERROR_NOMEMORY (-6)
if the attempt to get the memory block failed.
When any of these functions encounter a substring that is unset, which
can happen when capturing subpattern number n+1 matches some part of
the subject, but subpattern n has not been used at all, they return an
empty string. This can be distinguished from a genuine zero-length sub-
string by inspecting the appropriate offset in ovector, which is nega-
tive for unset substrings.
The two convenience functions pcre_free_substring() and pcre_free_sub-
string_list() can be used to free the memory returned by a previous
call of pcre_get_substring() or pcre_get_substring_list(), respec-
tively. They do nothing more than call the function pointed to by
pcre_free, which of course could be called directly from a C program.
However, PCRE is used in some situations where it is linked via a spe-
cial interface to another programming language that cannot use
pcre_free directly; it is for these cases that the functions are pro-
vided.
EXTRACTING CAPTURED SUBSTRINGS BY NAME
int pcre_get_stringnumber(const pcre *code,
const char *name);
int pcre_copy_named_substring(const pcre *code,
const char *subject, int *ovector,
int stringcount, const char *stringname,
char *buffer, int buffersize);
int pcre_get_named_substring(const pcre *code,
const char *subject, int *ovector,
int stringcount, const char *stringname,
const char **stringptr);
To extract a substring by name, you first have to find associated num-
ber. For example, for this pattern
(a+)b(?<xxx>\d+)...
the number of the subpattern called "xxx" is 2. If the name is known to
be unique (PCRE_DUPNAMES was not set), you can find the number from the
name by calling pcre_get_stringnumber(). The first argument is the com-
piled pattern, and the second is the name. The yield of the function is
the subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no
subpattern of that name.
Given the number, you can extract the substring directly, or use one of
the functions described in the previous section. For convenience, there
are also two functions that do the whole job.
Most of the arguments of pcre_copy_named_substring() and
pcre_get_named_substring() are the same as those for the similarly
named functions that extract by number. As these are described in the
previous section, they are not re-described here. There are just two
differences:
First, instead of a substring number, a substring name is given. Sec-
ond, there is an extra argument, given at the start, which is a pointer
to the compiled pattern. This is needed in order to gain access to the
name-to-number translation table.
These functions call pcre_get_stringnumber(), and if it succeeds, they
then call pcre_copy_substring() or pcre_get_substring(), as appropri-
ate. NOTE: If PCRE_DUPNAMES is set and there are duplicate names, the
behaviour may not be what you want (see the next section).
Warning: If the pattern uses the (?| feature to set up multiple subpat-
terns with the same number, as described in the section on duplicate
subpattern numbers in the pcrepattern page, you cannot use names to
distinguish the different subpatterns, because names are not included
in the compiled code. The matching process uses only numbers. For this
reason, the use of different names for subpatterns of the same number
causes an error at compile time.
DUPLICATE SUBPATTERN NAMES
int pcre_get_stringtable_entries(const pcre *code,
const char *name, char **first, char **last);
When a pattern is compiled with the PCRE_DUPNAMES option, names for
subpatterns are not required to be unique. (Duplicate names are always
allowed for subpatterns with the same number, created by using the (?|
feature. Indeed, if such subpatterns are named, they are required to
use the same names.)
Normally, patterns with duplicate names are such that in any one match,
only one of the named subpatterns participates. An example is shown in
the pcrepattern documentation.
When duplicates are present, pcre_copy_named_substring() and
pcre_get_named_substring() return the first substring corresponding to
the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING
(-7) is returned; no data is returned. The pcre_get_stringnumber()
function returns one of the numbers that are associated with the name,
but it is not defined which it is.
If you want to get full details of all captured substrings for a given
name, you must use the pcre_get_stringtable_entries() function. The
first argument is the compiled pattern, and the second is the name. The
third and fourth are pointers to variables which are updated by the
function. After it has run, they point to the first and last entries in
the name-to-number table for the given name. The function itself
returns the length of each entry, or PCRE_ERROR_NOSUBSTRING (-7) if
there are none. The format of the table is described above in the sec-
tion entitled Information about a pattern above. Given all the rele-
vant entries for the name, you can extract each of their numbers, and
hence the captured data, if any.
FINDING ALL POSSIBLE MATCHES
The traditional matching function uses a similar algorithm to Perl,
which stops when it finds the first match, starting at a given point in
the subject. If you want to find all possible matches, or the longest
possible match, consider using the alternative matching function (see
below) instead. If you cannot use the alternative function, but still
need to find all possible matches, you can kludge it up by making use
of the callout facility, which is described in the pcrecallout documen-
tation.
What you have to do is to insert a callout right at the end of the pat-
tern. When your callout function is called, extract and save the cur-
rent matched substring. Then return 1, which forces pcre_exec() to
backtrack and try other alternatives. Ultimately, when it runs out of
matches, pcre_exec() will yield PCRE_ERROR_NOMATCH.
MATCHING A PATTERN: THE ALTERNATIVE FUNCTION
int pcre_dfa_exec(const pcre *code, const pcre_extra *extra,
const char *subject, int length, int startoffset,
int options, int *ovector, int ovecsize,
int *workspace, int wscount);
The function pcre_dfa_exec() is called to match a subject string
against a compiled pattern, using a matching algorithm that scans the
subject string just once, and does not backtrack. This has different
characteristics to the normal algorithm, and is not compatible with
Perl. Some of the features of PCRE patterns are not supported. Never-
theless, there are times when this kind of matching can be useful. For
a discussion of the two matching algorithms, and a list of features
that pcre_dfa_exec() does not support, see the pcrematching documenta-
tion.
The arguments for the pcre_dfa_exec() function are the same as for
pcre_exec(), plus two extras. The ovector argument is used in a differ-
ent way, and this is described below. The other common arguments are
used in the same way as for pcre_exec(), so their description is not
repeated here.
The two additional arguments provide workspace for the function. The
workspace vector should contain at least 20 elements. It is used for
keeping track of multiple paths through the pattern tree. More
workspace will be needed for patterns and subjects where there are a
lot of potential matches.
Here is an example of a simple call to pcre_dfa_exec():
int rc;
int ovector[10];
int wspace[20];
rc = pcre_dfa_exec(
re, /* result of pcre_compile() */
NULL, /* we didn't study the pattern */
"some string", /* the subject string */
11, /* the length of the subject string */
0, /* start at offset 0 in the subject */
0, /* default options */
ovector, /* vector of integers for substring information */
10, /* number of elements (NOT size in bytes) */
wspace, /* working space vector */
20); /* number of elements (NOT size in bytes) */
Option bits for pcre_dfa_exec()
The unused bits of the options argument for pcre_dfa_exec() must be
zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEW-
LINE_xxx, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY,
PCRE_NOTEMPTY_ATSTART, PCRE_NO_UTF8_CHECK, PCRE_BSR_ANYCRLF,
PCRE_BSR_UNICODE, PCRE_NO_START_OPTIMIZE, PCRE_PARTIAL_HARD, PCRE_PAR-
TIAL_SOFT, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART. All but the last
four of these are exactly the same as for pcre_exec(), so their
description is not repeated here.
PCRE_PARTIAL_HARD
PCRE_PARTIAL_SOFT
These have the same general effect as they do for pcre_exec(), but the
details are slightly different. When PCRE_PARTIAL_HARD is set for
pcre_dfa_exec(), it returns PCRE_ERROR_PARTIAL if the end of the sub-
ject is reached and there is still at least one matching possibility
that requires additional characters. This happens even if some complete
matches have also been found. When PCRE_PARTIAL_SOFT is set, the return
code PCRE_ERROR_NOMATCH is converted into PCRE_ERROR_PARTIAL if the end
of the subject is reached, there have been no complete matches, but
there is still at least one matching possibility. The portion of the
string that was inspected when the longest partial match was found is
set as the first matching string in both cases. There is a more
detailed discussion of partial and multi-segment matching, with exam-
ples, in the pcrepartial documentation.
PCRE_DFA_SHORTEST
Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to
stop as soon as it has found one match. Because of the way the alterna-
tive algorithm works, this is necessarily the shortest possible match
at the first possible matching point in the subject string.
PCRE_DFA_RESTART
When pcre_dfa_exec() returns a partial match, it is possible to call it
again, with additional subject characters, and have it continue with
the same match. The PCRE_DFA_RESTART option requests this action; when
it is set, the workspace and wscount options must reference the same
vector as before because data about the match so far is left in them
after a partial match. There is more discussion of this facility in the
pcrepartial documentation.
Successful returns from pcre_dfa_exec()
When pcre_dfa_exec() succeeds, it may have matched more than one sub-
string in the subject. Note, however, that all the matches from one run
of the function start at the same point in the subject. The shorter
matches are all initial substrings of the longer matches. For example,
if the pattern
<.*>
is matched against the string
This is <something> <something else> <something further> no more
the three matched strings are
<something>
<something> <something else>
<something> <something else> <something further>
On success, the yield of the function is a number greater than zero,
which is the number of matched substrings. The substrings themselves
are returned in ovector. Each string uses two elements; the first is
the offset to the start, and the second is the offset to the end. In
fact, all the strings have the same start offset. (Space could have
been saved by giving this only once, but it was decided to retain some
compatibility with the way pcre_exec() returns data, even though the
meaning of the strings is different.)
The strings are returned in reverse order of length; that is, the long-
est matching string is given first. If there were too many matches to
fit into ovector, the yield of the function is zero, and the vector is
filled with the longest matches. Unlike pcre_exec(), pcre_dfa_exec()
can use the entire ovector for returning matched strings.
Error returns from pcre_dfa_exec()
The pcre_dfa_exec() function returns a negative number when it fails.
Many of the errors are the same as for pcre_exec(), and these are
described above. There are in addition the following errors that are
specific to pcre_dfa_exec():
PCRE_ERROR_DFA_UITEM (-16)
This return is given if pcre_dfa_exec() encounters an item in the pat-
tern that it does not support, for instance, the use of \C or a back
reference.
PCRE_ERROR_DFA_UCOND (-17)
This return is given if pcre_dfa_exec() encounters a condition item
that uses a back reference for the condition, or a test for recursion
in a specific group. These are not supported.
PCRE_ERROR_DFA_UMLIMIT (-18)
This return is given if pcre_dfa_exec() is called with an extra block
that contains a setting of the match_limit or match_limit_recursion
fields. This is not supported (these fields are meaningless for DFA
matching).
PCRE_ERROR_DFA_WSSIZE (-19)
This return is given if pcre_dfa_exec() runs out of space in the
workspace vector.
PCRE_ERROR_DFA_RECURSE (-20)
When a recursive subpattern is processed, the matching function calls
itself recursively, using private vectors for ovector and workspace.
This error is given if the output vector is not large enough. This
should be extremely rare, as a vector of size 1000 is used.
SEE ALSO
pcrebuild(3), pcrecallout(3), pcrecpp(3)(3), pcrematching(3), pcrepar-
tial(3), pcreposix(3), pcreprecompile(3), pcresample(3), pcrestack(3).
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 02 December 2011
Copyright (c) 1997-2011 University of Cambridge.
------------------------------------------------------------------------------
PCRECALLOUT(3) PCRECALLOUT(3)
NAME
PCRE - Perl-compatible regular expressions
PCRE CALLOUTS
int (*pcre_callout)(pcre_callout_block *);
PCRE provides a feature called "callout", which is a means of temporar-
ily passing control to the caller of PCRE in the middle of pattern
matching. The caller of PCRE provides an external function by putting
its entry point in the global variable pcre_callout. By default, this
variable contains NULL, which disables all calling out.
Within a regular expression, (?C) indicates the points at which the
external function is to be called. Different callout points can be
identified by putting a number less than 256 after the letter C. The
default value is zero. For example, this pattern has two callout
points:
(?C1)abc(?C2)def
If the PCRE_AUTO_CALLOUT option bit is set when pcre_compile() or
pcre_compile2() is called, PCRE automatically inserts callouts, all
with number 255, before each item in the pattern. For example, if
PCRE_AUTO_CALLOUT is used with the pattern
A(\d{2}|--)
it is processed as if it were
(?C255)A(?C255)((?C255)\d{2}(?C255)|(?C255)-(?C255)-(?C255))(?C255)
Notice that there is a callout before and after each parenthesis and
alternation bar. Automatic callouts can be used for tracking the
progress of pattern matching. The pcretest command has an option that
sets automatic callouts; when it is used, the output indicates how the
pattern is matched. This is useful information when you are trying to
optimize the performance of a particular pattern.
The use of callouts in a pattern makes it ineligible for optimization
by the just-in-time compiler. Studying such a pattern with the
PCRE_STUDY_JIT_COMPILE option always fails.
MISSING CALLOUTS
You should be aware that, because of optimizations in the way PCRE
matches patterns by default, callouts sometimes do not happen. For
example, if the pattern is
ab(?C4)cd
PCRE knows that any matching string must contain the letter "d". If the
subject string is "abyz", the lack of "d" means that matching doesn't
ever start, and the callout is never reached. However, with "abyd",
though the result is still no match, the callout is obeyed.
If the pattern is studied, PCRE knows the minimum length of a matching
string, and will immediately give a "no match" return without actually
running a match if the subject is not long enough, or, for unanchored
patterns, if it has been scanned far enough.
You can disable these optimizations by passing the PCRE_NO_START_OPTI-
MIZE option to pcre_compile(), pcre_exec(), or pcre_dfa_exec(), or by
starting the pattern with (*NO_START_OPT). This slows down the matching
process, but does ensure that callouts such as the example above are
obeyed.
THE CALLOUT INTERFACE
During matching, when PCRE reaches a callout point, the external func-
tion defined by pcre_callout is called (if it is set). This applies to
both the pcre_exec() and the pcre_dfa_exec() matching functions. The
only argument to the callout function is a pointer to a pcre_callout
block. This structure contains the following fields:
int version;
int callout_number;
int *offset_vector;
const char *subject;
int subject_length;
int start_match;
int current_position;
int capture_top;
int capture_last;
void *callout_data;
int pattern_position;
int next_item_length;
const unsigned char *mark;
The version field is an integer containing the version number of the
block format. The initial version was 0; the current version is 2. The
version number will change again in future if additional fields are
added, but the intention is never to remove any of the existing fields.
The callout_number field contains the number of the callout, as com-
piled into the pattern (that is, the number after ?C for manual call-
outs, and 255 for automatically generated callouts).
The offset_vector field is a pointer to the vector of offsets that was
passed by the caller to pcre_exec() or pcre_dfa_exec(). When
pcre_exec() is used, the contents can be inspected in order to extract
substrings that have been matched so far, in the same way as for
extracting substrings after a match has completed. For pcre_dfa_exec()
this field is not useful.
The subject and subject_length fields contain copies of the values that
were passed to pcre_exec().
The start_match field normally contains the offset within the subject
at which the current match attempt started. However, if the escape
sequence \K has been encountered, this value is changed to reflect the
modified starting point. If the pattern is not anchored, the callout
function may be called several times from the same point in the pattern
for different starting points in the subject.
The current_position field contains the offset within the subject of
the current match pointer.
When the pcre_exec() function is used, the capture_top field contains
one more than the number of the highest numbered captured substring so
far. If no substrings have been captured, the value of capture_top is
one. This is always the case when pcre_dfa_exec() is used, because it
does not support captured substrings.
The capture_last field contains the number of the most recently cap-
tured substring. If no substrings have been captured, its value is -1.
This is always the case when pcre_dfa_exec() is used.
The callout_data field contains a value that is passed to pcre_exec()
or pcre_dfa_exec() specifically so that it can be passed back in call-
outs. It is passed in the pcre_callout field of the pcre_extra data
structure. If no such data was passed, the value of callout_data in a
pcre_callout block is NULL. There is a description of the pcre_extra
structure in the pcreapi documentation.
The pattern_position field is present from version 1 of the pcre_call-
out structure. It contains the offset to the next item to be matched in
the pattern string.
The next_item_length field is present from version 1 of the pcre_call-
out structure. It contains the length of the next item to be matched in
the pattern string. When the callout immediately precedes an alterna-
tion bar, a closing parenthesis, or the end of the pattern, the length
is zero. When the callout precedes an opening parenthesis, the length
is that of the entire subpattern.
The pattern_position and next_item_length fields are intended to help
in distinguishing between different automatic callouts, which all have
the same callout number. However, they are set for all callouts.
The mark field is present from version 2 of the pcre_callout structure.
In callouts from pcre_exec() it contains a pointer to the zero-termi-
nated name of the most recently passed (*MARK), (*PRUNE), or (*THEN)
item in the match, or NULL if no such items have been passed. Instances
of (*PRUNE) or (*THEN) without a name do not obliterate a previous
(*MARK). In callouts from pcre_dfa_exec() this field always contains
NULL.
RETURN VALUES
The external callout function returns an integer to PCRE. If the value
is zero, matching proceeds as normal. If the value is greater than
zero, matching fails at the current point, but the testing of other
matching possibilities goes ahead, just as if a lookahead assertion had
failed. If the value is less than zero, the match is abandoned, and
pcre_exec() or pcre_dfa_exec() returns the negative value.
Negative values should normally be chosen from the set of
PCRE_ERROR_xxx values. In particular, PCRE_ERROR_NOMATCH forces a stan-
dard "no match" failure. The error number PCRE_ERROR_CALLOUT is
reserved for use by callout functions; it will never be used by PCRE
itself.
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 30 November 2011
Copyright (c) 1997-2011 University of Cambridge.
------------------------------------------------------------------------------
PCRECOMPAT(3) PCRECOMPAT(3)
NAME
PCRE - Perl-compatible regular expressions
DIFFERENCES BETWEEN PCRE AND PERL
This document describes the differences in the ways that PCRE and Perl
handle regular expressions. The differences described here are with
respect to Perl versions 5.10 and above.
1. PCRE has only a subset of Perl's UTF-8 and Unicode support. Details
of what it does have are given in the pcreunicode page.
2. PCRE allows repeat quantifiers only on parenthesized assertions, but
they do not mean what you might think. For example, (?!a){3} does not
assert that the next three characters are not "a". It just asserts that
the next character is not "a" three times (in principle: PCRE optimizes
this to run the assertion just once). Perl allows repeat quantifiers on
other assertions such as \b, but these do not seem to have any use.
3. Capturing subpatterns that occur inside negative lookahead asser-
tions are counted, but their entries in the offsets vector are never
set. Perl sets its numerical variables from any such patterns that are
matched before the assertion fails to match something (thereby succeed-
ing), but only if the negative lookahead assertion contains just one
branch.
4. Though binary zero characters are supported in the subject string,
they are not allowed in a pattern string because it is passed as a nor-
mal C string, terminated by zero. The escape sequence \0 can be used in
the pattern to represent a binary zero.
5. The following Perl escape sequences are not supported: \l, \u, \L,
\U, and \N when followed by a character name or Unicode value. (\N on
its own, matching a non-newline character, is supported.) In fact these
are implemented by Perl's general string-handling and are not part of
its pattern matching engine. If any of these are encountered by PCRE,
an error is generated by default. However, if the PCRE_JAVASCRIPT_COM-
PAT option is set, \U and \u are interpreted as JavaScript interprets
them.
6. The Perl escape sequences \p, \P, and \X are supported only if PCRE
is built with Unicode character property support. The properties that
can be tested with \p and \P are limited to the general category prop-
erties such as Lu and Nd, script names such as Greek or Han, and the
derived properties Any and L&. PCRE does support the Cs (surrogate)
property, which Perl does not; the Perl documentation says "Because
Perl hides the need for the user to understand the internal representa-
tion of Unicode characters, there is no need to implement the somewhat
messy concept of surrogates."
7. PCRE implements a simpler version of \X than Perl, which changed to
make \X match what Unicode calls an "extended grapheme cluster". This
is more complicated than an extended Unicode sequence, which is what
PCRE matches.
8. PCRE does support the \Q...\E escape for quoting substrings. Charac-
ters in between are treated as literals. This is slightly different
from Perl in that $ and @ are also handled as literals inside the
quotes. In Perl, they cause variable interpolation (but of course PCRE
does not have variables). Note the following examples:
Pattern PCRE matches Perl matches
\Qabc$xyz\E abc$xyz abc followed by the
contents of $xyz
\Qabc\$xyz\E abc\$xyz abc\$xyz
\Qabc\E\$\Qxyz\E abc$xyz abc$xyz
The \Q...\E sequence is recognized both inside and outside character
classes.
9. Fairly obviously, PCRE does not support the (?{code}) and (??{code})
constructions. However, there is support for recursive patterns. This
is not available in Perl 5.8, but it is in Perl 5.10. Also, the PCRE
"callout" feature allows an external function to be called during pat-
tern matching. See the pcrecallout documentation for details.
10. Subpatterns that are called as subroutines (whether or not recur-
sively) are always treated as atomic groups in PCRE. This is like
Python, but unlike Perl. Captured values that are set outside a sub-
routine call can be reference from inside in PCRE, but not in Perl.
There is a discussion that explains these differences in more detail in
the section on recursion differences from Perl in the pcrepattern page.
11. If (*THEN) is present in a group that is called as a subroutine,
its action is limited to that group, even if the group does not contain
any | characters.
12. There are some differences that are concerned with the settings of
captured strings when part of a pattern is repeated. For example,
matching "aba" against the pattern /^(a(b)?)+$/ in Perl leaves $2
unset, but in PCRE it is set to "b".
13. PCRE's handling of duplicate subpattern numbers and duplicate sub-
pattern names is not as general as Perl's. This is a consequence of the
fact the PCRE works internally just with numbers, using an external ta-
ble to translate between numbers and names. In particular, a pattern
such as (?|(?<a>A)|(?<b)B), where the two capturing parentheses have
the same number but different names, is not supported, and causes an
error at compile time. If it were allowed, it would not be possible to
distinguish which parentheses matched, because both names map to cap-
turing subpattern number 1. To avoid this confusing situation, an error
is given at compile time.
14. Perl recognizes comments in some places that PCRE does not, for
example, between the ( and ? at the start of a subpattern. If the /x
modifier is set, Perl allows whitespace between ( and ? but PCRE never
does, even if the PCRE_EXTENDED option is set.
15. PCRE provides some extensions to the Perl regular expression facil-
ities. Perl 5.10 includes new features that are not in earlier ver-
sions of Perl, some of which (such as named parentheses) have been in
PCRE for some time. This list is with respect to Perl 5.10:
(a) Although lookbehind assertions in PCRE must match fixed length
strings, each alternative branch of a lookbehind assertion can match a
different length of string. Perl requires them all to have the same
length.
(b) If PCRE_DOLLAR_ENDONLY is set and PCRE_MULTILINE is not set, the $
meta-character matches only at the very end of the string.
(c) If PCRE_EXTRA is set, a backslash followed by a letter with no spe-
cial meaning is faulted. Otherwise, like Perl, the backslash is quietly
ignored. (Perl can be made to issue a warning.)
(d) If PCRE_UNGREEDY is set, the greediness of the repetition quanti-
fiers is inverted, that is, by default they are not greedy, but if fol-
lowed by a question mark they are.
(e) PCRE_ANCHORED can be used at matching time to force a pattern to be
tried only at the first matching position in the subject string.
(f) The PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
and PCRE_NO_AUTO_CAPTURE options for pcre_exec() have no Perl equiva-
lents.
(g) The \R escape sequence can be restricted to match only CR, LF, or
CRLF by the PCRE_BSR_ANYCRLF option.
(h) The callout facility is PCRE-specific.
(i) The partial matching facility is PCRE-specific.
(j) Patterns compiled by PCRE can be saved and re-used at a later time,
even on different hosts that have the other endianness. However, this
does not apply to optimized data created by the just-in-time compiler.
(k) The alternative matching function (pcre_dfa_exec()) matches in a
different way and is not Perl-compatible.
(l) PCRE recognizes some special sequences such as (*CR) at the start
of a pattern that set overall options that cannot be changed within the
pattern.
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 14 November 2011
Copyright (c) 1997-2011 University of Cambridge.
------------------------------------------------------------------------------
PCREPATTERN(3) PCREPATTERN(3)
NAME
PCRE - Perl-compatible regular expressions
PCRE REGULAR EXPRESSION DETAILS
The syntax and semantics of the regular expressions that are supported
by PCRE are described in detail below. There is a quick-reference syn-
tax summary in the pcresyntax page. PCRE tries to match Perl syntax and
semantics as closely as it can. PCRE also supports some alternative
regular expression syntax (which does not conflict with the Perl syn-
tax) in order to provide some compatibility with regular expressions in
Python, .NET, and Oniguruma.
Perl's regular expressions are described in its own documentation, and
regular expressions in general are covered in a number of books, some
of which have copious examples. Jeffrey Friedl's "Mastering Regular
Expressions", published by O'Reilly, covers regular expressions in
great detail. This description of PCRE's regular expressions is
intended as reference material.
The original operation of PCRE was on strings of one-byte characters.
However, there is now also support for UTF-8 character strings. To use
this, PCRE must be built to include UTF-8 support, and you must call
pcre_compile() or pcre_compile2() with the PCRE_UTF8 option. There is
also a special sequence that can be given at the start of a pattern:
(*UTF8)
Starting a pattern with this sequence is equivalent to setting the
PCRE_UTF8 option. This feature is not Perl-compatible. How setting
UTF-8 mode affects pattern matching is mentioned in several places
below. There is also a summary of UTF-8 features in the pcreunicode
page.
Another special sequence that may appear at the start of a pattern or
in combination with (*UTF8) is:
(*UCP)
This has the same effect as setting the PCRE_UCP option: it causes
sequences such as \d and \w to use Unicode properties to determine
character types, instead of recognizing only characters with codes less
than 128 via a lookup table.
If a pattern starts with (*NO_START_OPT), it has the same effect as
setting the PCRE_NO_START_OPTIMIZE option either at compile or matching
time. There are also some more of these special sequences that are con-
cerned with the handling of newlines; they are described below.
The remainder of this document discusses the patterns that are sup-
ported by PCRE when its main matching function, pcre_exec(), is used.
From release 6.0, PCRE offers a second matching function,
pcre_dfa_exec(), which matches using a different algorithm that is not
Perl-compatible. Some of the features discussed below are not available
when pcre_dfa_exec() is used. The advantages and disadvantages of the
alternative function, and how it differs from the normal function, are
discussed in the pcrematching page.
NEWLINE CONVENTIONS
PCRE supports five different conventions for indicating line breaks in
strings: a single CR (carriage return) character, a single LF (line-
feed) character, the two-character sequence CRLF, any of the three pre-
ceding, or any Unicode newline sequence. The pcreapi page has further
discussion about newlines, and shows how to set the newline convention
in the options arguments for the compiling and matching functions.
It is also possible to specify a newline convention by starting a pat-
tern string with one of the following five sequences:
(*CR) carriage return
(*LF) linefeed
(*CRLF) carriage return, followed by linefeed
(*ANYCRLF) any of the three above
(*ANY) all Unicode newline sequences
These override the default and the options given to pcre_compile() or
pcre_compile2(). For example, on a Unix system where LF is the default
newline sequence, the pattern
(*CR)a.b
changes the convention to CR. That pattern matches "a\nb" because LF is
no longer a newline. Note that these special settings, which are not
Perl-compatible, are recognized only at the very start of a pattern,
and that they must be in upper case. If more than one of them is
present, the last one is used.
The newline convention affects the interpretation of the dot metachar-
acter when PCRE_DOTALL is not set, and also the behaviour of \N. How-
ever, it does not affect what the \R escape sequence matches. By
default, this is any Unicode newline sequence, for Perl compatibility.
However, this can be changed; see the description of \R in the section
entitled "Newline sequences" below. A change of \R setting can be com-
bined with a change of newline convention.
CHARACTERS AND METACHARACTERS
A regular expression is a pattern that is matched against a subject
string from left to right. Most characters stand for themselves in a
pattern, and match the corresponding characters in the subject. As a
trivial example, the pattern
The quick brown fox
matches a portion of a subject string that is identical to itself. When
caseless matching is specified (the PCRE_CASELESS option), letters are
matched independently of case. In UTF-8 mode, PCRE always understands
the concept of case for characters whose values are less than 128, so
caseless matching is always possible. For characters with higher val-
ues, the concept of case is supported if PCRE is compiled with Unicode
property support, but not otherwise. If you want to use caseless
matching for characters 128 and above, you must ensure that PCRE is
compiled with Unicode property support as well as with UTF-8 support.
The power of regular expressions comes from the ability to include
alternatives and repetitions in the pattern. These are encoded in the
pattern by the use of metacharacters, which do not stand for themselves
but instead are interpreted in some special way.
There are two different sets of metacharacters: those that are recog-
nized anywhere in the pattern except within square brackets, and those
that are recognized within square brackets. Outside square brackets,
the metacharacters are as follows:
\ general escape character with several uses
^ assert start of string (or line, in multiline mode)
$ assert end of string (or line, in multiline mode)
. match any character except newline (by default)
[ start character class definition
| start of alternative branch
( start subpattern
) end subpattern
? extends the meaning of (
also 0 or 1 quantifier
also quantifier minimizer
* 0 or more quantifier
+ 1 or more quantifier
also "possessive quantifier"
{ start min/max quantifier
Part of a pattern that is in square brackets is called a "character
class". In a character class the only metacharacters are:
\ general escape character
^ negate the class, but only if the first character
- indicates character range
[ POSIX character class (only if followed by POSIX
syntax)
] terminates the character class
The following sections describe the use of each of the metacharacters.
BACKSLASH
The backslash character has several uses. Firstly, if it is followed by
a character that is not a number or a letter, it takes away any special
meaning that character may have. This use of backslash as an escape
character applies both inside and outside character classes.
For example, if you want to match a * character, you write \* in the
pattern. This escaping action applies whether or not the following
character would otherwise be interpreted as a metacharacter, so it is
always safe to precede a non-alphanumeric with backslash to specify
that it stands for itself. In particular, if you want to match a back-
slash, you write \\.
In UTF-8 mode, only ASCII numbers and letters have any special meaning
after a backslash. All other characters (in particular, those whose
codepoints are greater than 127) are treated as literals.
If a pattern is compiled with the PCRE_EXTENDED option, whitespace in
the pattern (other than in a character class) and characters between a
# outside a character class and the next newline are ignored. An escap-
ing backslash can be used to include a whitespace or # character as
part of the pattern.
If you want to remove the special meaning from a sequence of charac-
ters, you can do so by putting them between \Q and \E. This is differ-
ent from Perl in that $ and @ are handled as literals in \Q...\E
sequences in PCRE, whereas in Perl, $ and @ cause variable interpola-
tion. Note the following examples:
Pattern PCRE matches Perl matches
\Qabc$xyz\E abc$xyz abc followed by the
contents of $xyz
\Qabc\$xyz\E abc\$xyz abc\$xyz
\Qabc\E\$\Qxyz\E abc$xyz abc$xyz
The \Q...\E sequence is recognized both inside and outside character
classes. An isolated \E that is not preceded by \Q is ignored. If \Q
is not followed by \E later in the pattern, the literal interpretation
continues to the end of the pattern (that is, \E is assumed at the
end). If the isolated \Q is inside a character class, this causes an
error, because the character class is not terminated.
Non-printing characters
A second use of backslash provides a way of encoding non-printing char-
acters in patterns in a visible manner. There is no restriction on the
appearance of non-printing characters, apart from the binary zero that
terminates a pattern, but when a pattern is being prepared by text
editing, it is often easier to use one of the following escape
sequences than the binary character it represents:
\a alarm, that is, the BEL character (hex 07)
\cx "control-x", where x is any ASCII character
\e escape (hex 1B)
\f formfeed (hex 0C)
\n linefeed (hex 0A)
\r carriage return (hex 0D)
\t tab (hex 09)
\ddd character with octal code ddd, or back reference
\xhh character with hex code hh
\x{hhh..} character with hex code hhh.. (non-JavaScript mode)
\uhhhh character with hex code hhhh (JavaScript mode only)
The precise effect of \cx is as follows: if x is a lower case letter,
it is converted to upper case. Then bit 6 of the character (hex 40) is
inverted. Thus \cz becomes hex 1A (z is 7A), but \c{ becomes hex 3B ({
is 7B), while \c; becomes hex 7B (; is 3B). If the byte following \c
has a value greater than 127, a compile-time error occurs. This locks
out non-ASCII characters in both byte mode and UTF-8 mode. (When PCRE
is compiled in EBCDIC mode, all byte values are valid. A lower case
letter is converted to upper case, and then the 0xc0 bits are flipped.)
By default, after \x, from zero to two hexadecimal digits are read
(letters can be in upper or lower case). Any number of hexadecimal dig-
its may appear between \x{ and }, but the value of the character code
must be less than 256 in non-UTF-8 mode, and less than 2**31 in UTF-8
mode. That is, the maximum value in hexadecimal is 7FFFFFFF. Note that
this is bigger than the largest Unicode code point, which is 10FFFF.
If characters other than hexadecimal digits appear between \x{ and },
or if there is no terminating }, this form of escape is not recognized.
Instead, the initial \x will be interpreted as a basic hexadecimal
escape, with no following digits, giving a character whose value is
zero.
If the PCRE_JAVASCRIPT_COMPAT option is set, the interpretation of \x
is as just described only when it is followed by two hexadecimal dig-
its. Otherwise, it matches a literal "x" character. In JavaScript
mode, support for code points greater than 256 is provided by \u, which
must be followed by four hexadecimal digits; otherwise it matches a
literal "u" character.
Characters whose value is less than 256 can be defined by either of the
two syntaxes for \x (or by \u in JavaScript mode). There is no differ-
ence in the way they are handled. For example, \xdc is exactly the same
as \x{dc} (or \u00dc in JavaScript mode).
After \0 up to two further octal digits are read. If there are fewer
than two digits, just those that are present are used. Thus the
sequence \0\x\07 specifies two binary zeros followed by a BEL character
(code value 7). Make sure you supply two digits after the initial zero
if the pattern character that follows is itself an octal digit.
The handling of a backslash followed by a digit other than 0 is compli-
cated. Outside a character class, PCRE reads it and any following dig-
its as a decimal number. If the number is less than 10, or if there
have been at least that many previous capturing left parentheses in the
expression, the entire sequence is taken as a back reference. A
description of how this works is given later, following the discussion
of parenthesized subpatterns.
Inside a character class, or if the decimal number is greater than 9
and there have not been that many capturing subpatterns, PCRE re-reads
up to three octal digits following the backslash, and uses them to gen-
erate a data character. Any subsequent digits stand for themselves. In
non-UTF-8 mode, the value of a character specified in octal must be
less than \400. In UTF-8 mode, values up to \777 are permitted. For
example:
\040 is another way of writing a space
\40 is the same, provided there are fewer than 40
previous capturing subpatterns
\7 is always a back reference
\11 might be a back reference, or another way of
writing a tab
\011 is always a tab
\0113 is a tab followed by the character "3"
\113 might be a back reference, otherwise the
character with octal code 113
\377 might be a back reference, otherwise
the byte consisting entirely of 1 bits
\81 is either a back reference, or a binary zero
followed by the two characters "8" and "1"
Note that octal values of 100 or greater must not be introduced by a
leading zero, because no more than three octal digits are ever read.
All the sequences that define a single character value can be used both
inside and outside character classes. In addition, inside a character
class, \b is interpreted as the backspace character (hex 08).
\N is not allowed in a character class. \B, \R, and \X are not special
inside a character class. Like other unrecognized escape sequences,
they are treated as the literal characters "B", "R", and "X" by
default, but cause an error if the PCRE_EXTRA option is set. Outside a
character class, these sequences have different meanings.
Unsupported escape sequences
In Perl, the sequences \l, \L, \u, and \U are recognized by its string
handler and used to modify the case of following characters. By
default, PCRE does not support these escape sequences. However, if the
PCRE_JAVASCRIPT_COMPAT option is set, \U matches a "U" character, and
\u can be used to define a character by code point, as described in the
previous section.
Absolute and relative back references
The sequence \g followed by an unsigned or a negative number, option-
ally enclosed in braces, is an absolute or relative back reference. A
named back reference can be coded as \g{name}. Back references are dis-
cussed later, following the discussion of parenthesized subpatterns.
Absolute and relative subroutine calls
For compatibility with Oniguruma, the non-Perl syntax \g followed by a
name or a number enclosed either in angle brackets or single quotes, is
an alternative syntax for referencing a subpattern as a "subroutine".
Details are discussed later. Note that \g{...} (Perl syntax) and
\g<...> (Oniguruma syntax) are not synonymous. The former is a back
reference; the latter is a subroutine call.
Generic character types
Another use of backslash is for specifying generic character types:
\d any decimal digit
\D any character that is not a decimal digit
\h any horizontal whitespace character
\H any character that is not a horizontal whitespace character
\s any whitespace character
\S any character that is not a whitespace character
\v any vertical whitespace character
\V any character that is not a vertical whitespace character
\w any "word" character
\W any "non-word" character
There is also the single sequence \N, which matches a non-newline char-
acter. This is the same as the "." metacharacter when PCRE_DOTALL is
not set. Perl also uses \N to match characters by name; PCRE does not
support this.
Each pair of lower and upper case escape sequences partitions the com-
plete set of characters into two disjoint sets. Any given character
matches one, and only one, of each pair. The sequences can appear both
inside and outside character classes. They each match one character of
the appropriate type. If the current matching point is at the end of
the subject string, all of them fail, because there is no character to
match.
For compatibility with Perl, \s does not match the VT character (code
11). This makes it different from the the POSIX "space" class. The \s
characters are HT (9), LF (10), FF (12), CR (13), and space (32). If
"use locale;" is included in a Perl script, \s may match the VT charac-
ter. In PCRE, it never does.
A "word" character is an underscore or any character that is a letter
or digit. By default, the definition of letters and digits is con-
trolled by PCRE's low-valued character tables, and may vary if locale-
specific matching is taking place (see "Locale support" in the pcreapi
page). For example, in a French locale such as "fr_FR" in Unix-like
systems, or "french" in Windows, some character codes greater than 128
are used for accented letters, and these are then matched by \w. The
use of locales with Unicode is discouraged.
By default, in UTF-8 mode, characters with values greater than 128
never match \d, \s, or \w, and always match \D, \S, and \W. These
sequences retain their original meanings from before UTF-8 support was
available, mainly for efficiency reasons. However, if PCRE is compiled
with Unicode property support, and the PCRE_UCP option is set, the be-
haviour is changed so that Unicode properties are used to determine
character types, as follows:
\d any character that \p{Nd} matches (decimal digit)
\s any character that \p{Z} matches, plus HT, LF, FF, CR
\w any character that \p{L} or \p{N} matches, plus underscore
The upper case escapes match the inverse sets of characters. Note that
\d matches only decimal digits, whereas \w matches any Unicode digit,
as well as any Unicode letter, and underscore. Note also that PCRE_UCP
affects \b, and \B because they are defined in terms of \w and \W.
Matching these sequences is noticeably slower when PCRE_UCP is set.
The sequences \h, \H, \v, and \V are features that were added to Perl
at release 5.10. In contrast to the other sequences, which match only
ASCII characters by default, these always match certain high-valued
codepoints in UTF-8 mode, whether or not PCRE_UCP is set. The horizon-
tal space characters are:
U+0009 Horizontal tab
U+0020 Space
U+00A0 Non-break space
U+1680 Ogham space mark
U+180E Mongolian vowel separator
U+2000 En quad
U+2001 Em quad
U+2002 En space
U+2003 Em space
U+2004 Three-per-em space
U+2005 Four-per-em space
U+2006 Six-per-em space
U+2007 Figure space
U+2008 Punctuation space
U+2009 Thin space
U+200A Hair space
U+202F Narrow no-break space
U+205F Medium mathematical space
U+3000 Ideographic space
The vertical space characters are:
U+000A Linefeed
U+000B Vertical tab
U+000C Formfeed
U+000D Carriage return
U+0085 Next line
U+2028 Line separator
U+2029 Paragraph separator
Newline sequences
Outside a character class, by default, the escape sequence \R matches
any Unicode newline sequence. In non-UTF-8 mode \R is equivalent to the
following:
(?>\r\n|\n|\x0b|\f|\r|\x85)
This is an example of an "atomic group", details of which are given
below. This particular group matches either the two-character sequence
CR followed by LF, or one of the single characters LF (linefeed,
U+000A), VT (vertical tab, U+000B), FF (formfeed, U+000C), CR (carriage
return, U+000D), or NEL (next line, U+0085). The two-character sequence
is treated as a single unit that cannot be split.
In UTF-8 mode, two additional characters whose codepoints are greater
than 255 are added: LS (line separator, U+2028) and PS (paragraph sepa-
rator, U+2029). Unicode character property support is not needed for
these characters to be recognized.
It is possible to restrict \R to match only CR, LF, or CRLF (instead of
the complete set of Unicode line endings) by setting the option
PCRE_BSR_ANYCRLF either at compile time or when the pattern is matched.
(BSR is an abbrevation for "backslash R".) This can be made the default
when PCRE is built; if this is the case, the other behaviour can be
requested via the PCRE_BSR_UNICODE option. It is also possible to
specify these settings by starting a pattern string with one of the
following sequences:
(*BSR_ANYCRLF) CR, LF, or CRLF only
(*BSR_UNICODE) any Unicode newline sequence
These override the default and the options given to pcre_compile() or
pcre_compile2(), but they can be overridden by options given to
pcre_exec() or pcre_dfa_exec(). Note that these special settings, which
are not Perl-compatible, are recognized only at the very start of a
pattern, and that they must be in upper case. If more than one of them
is present, the last one is used. They can be combined with a change of
newline convention; for example, a pattern can start with:
(*ANY)(*BSR_ANYCRLF)
They can also be combined with the (*UTF8) or (*UCP) special sequences.
Inside a character class, \R is treated as an unrecognized escape
sequence, and so matches the letter "R" by default, but causes an error
if PCRE_EXTRA is set.
Unicode character properties
When PCRE is built with Unicode character property support, three addi-
tional escape sequences that match characters with specific properties
are available. When not in UTF-8 mode, these sequences are of course
limited to testing characters whose codepoints are less than 256, but
they do work in this mode. The extra escape sequences are:
\p{xx} a character with the xx property
\P{xx} a character without the xx property
\X an extended Unicode sequence
The property names represented by xx above are limited to the Unicode
script names, the general category properties, "Any", which matches any
character (including newline), and some special PCRE properties
(described in the next section). Other Perl properties such as "InMu-
sicalSymbols" are not currently supported by PCRE. Note that \P{Any}
does not match any characters, so always causes a match failure.
Sets of Unicode characters are defined as belonging to certain scripts.
A character from one of these sets can be matched using a script name.
For example:
\p{Greek}
\P{Han}
Those that are not part of an identified script are lumped together as
"Common". The current list of scripts is:
Arabic, Armenian, Avestan, Balinese, Bamum, Bengali, Bopomofo, Braille,
Buginese, Buhid, Canadian_Aboriginal, Carian, Cham, Cherokee, Common,
Coptic, Cuneiform, Cypriot, Cyrillic, Deseret, Devanagari, Egyp-
tian_Hieroglyphs, Ethiopic, Georgian, Glagolitic, Gothic, Greek,
Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hiragana, Impe-
rial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscriptional_Parthian,
Javanese, Kaithi, Kannada, Katakana, Kayah_Li, Kharoshthi, Khmer, Lao,
Latin, Lepcha, Limbu, Linear_B, Lisu, Lycian, Lydian, Malayalam,
Meetei_Mayek, Mongolian, Myanmar, New_Tai_Lue, Nko, Ogham, Old_Italic,
Old_Persian, Old_South_Arabian, Old_Turkic, Ol_Chiki, Oriya, Osmanya,
Phags_Pa, Phoenician, Rejang, Runic, Samaritan, Saurashtra, Shavian,
Sinhala, Sundanese, Syloti_Nagri, Syriac, Tagalog, Tagbanwa, Tai_Le,
Tai_Tham, Tai_Viet, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh,
Ugaritic, Vai, Yi.
Each character has exactly one Unicode general category property, spec-
ified by a two-letter abbreviation. For compatibility with Perl, nega-
tion can be specified by including a circumflex between the opening
brace and the property name. For example, \p{^Lu} is the same as
\P{Lu}.
If only one letter is specified with \p or \P, it includes all the gen-
eral category properties that start with that letter. In this case, in
the absence of negation, the curly brackets in the escape sequence are
optional; these two examples have the same effect:
\p{L}
\pL
The following general category property codes are supported:
C Other
Cc Control
Cf Format
Cn Unassigned
Co Private use
Cs Surrogate
L Letter
Ll Lower case letter
Lm Modifier letter
Lo Other letter
Lt Title case letter
Lu Upper case letter
M Mark
Mc Spacing mark
Me Enclosing mark
Mn Non-spacing mark
N Number
Nd Decimal number
Nl Letter number
No Other number
P Punctuation
Pc Connector punctuation
Pd Dash punctuation
Pe Close punctuation
Pf Final punctuation
Pi Initial punctuation
Po Other punctuation
Ps Open punctuation
S Symbol
Sc Currency symbol
Sk Modifier symbol
Sm Mathematical symbol
So Other symbol
Z Separator
Zl Line separator
Zp Paragraph separator
Zs Space separator
The special property L& is also supported: it matches a character that
has the Lu, Ll, or Lt property, in other words, a letter that is not
classified as a modifier or "other".
The Cs (Surrogate) property applies only to characters in the range
U+D800 to U+DFFF. Such characters are not valid in UTF-8 strings (see
RFC 3629) and so cannot be tested by PCRE, unless UTF-8 validity check-
ing has been turned off (see the discussion of PCRE_NO_UTF8_CHECK in
the pcreapi page). Perl does not support the Cs property.
The long synonyms for property names that Perl supports (such as
\p{Letter}) are not supported by PCRE, nor is it permitted to prefix
any of these properties with "Is".
No character that is in the Unicode table has the Cn (unassigned) prop-
erty. Instead, this property is assumed for any code point that is not
in the Unicode table.
Specifying caseless matching does not affect these escape sequences.
For example, \p{Lu} always matches only upper case letters.
The \X escape matches any number of Unicode characters that form an
extended Unicode sequence. \X is equivalent to
(?>\PM\pM*)
That is, it matches a character without the "mark" property, followed
by zero or more characters with the "mark" property, and treats the
sequence as an atomic group (see below). Characters with the "mark"
property are typically accents that affect the preceding character.
None of them have codepoints less than 256, so in non-UTF-8 mode \X
matches any one character.
Note that recent versions of Perl have changed \X to match what Unicode
calls an "extended grapheme cluster", which has a more complicated def-
inition.
Matching characters by Unicode property is not fast, because PCRE has
to search a structure that contains data for over fifteen thousand
characters. That is why the traditional escape sequences such as \d and
\w do not use Unicode properties in PCRE by default, though you can
make them do so by setting the PCRE_UCP option for pcre_compile() or by
starting the pattern with (*UCP).
PCRE's additional properties
As well as the standard Unicode properties described in the previous
section, PCRE supports four more that make it possible to convert tra-
ditional escape sequences such as \w and \s and POSIX character classes
to use Unicode properties. PCRE uses these non-standard, non-Perl prop-
erties internally when PCRE_UCP is set. They are:
Xan Any alphanumeric character
Xps Any POSIX space character
Xsp Any Perl space character
Xwd Any Perl "word" character
Xan matches characters that have either the L (letter) or the N (num-
ber) property. Xps matches the characters tab, linefeed, vertical tab,
formfeed, or carriage return, and any other character that has the Z
(separator) property. Xsp is the same as Xps, except that vertical tab
is excluded. Xwd matches the same characters as Xan, plus underscore.
Resetting the match start
The escape sequence \K causes any previously matched characters not to
be included in the final matched sequence. For example, the pattern:
foo\Kbar
matches "foobar", but reports that it has matched "bar". This feature
is similar to a lookbehind assertion (described below). However, in
this case, the part of the subject before the real match does not have
to be of fixed length, as lookbehind assertions do. The use of \K does
not interfere with the setting of captured substrings. For example,
when the pattern
(foo)\Kbar
matches "foobar", the first substring is still set to "foo".
Perl documents that the use of \K within assertions is "not well
defined". In PCRE, \K is acted upon when it occurs inside positive
assertions, but is ignored in negative assertions.
Simple assertions
The final use of backslash is for certain simple assertions. An asser-
tion specifies a condition that has to be met at a particular point in
a match, without consuming any characters from the subject string. The
use of subpatterns for more complicated assertions is described below.
The backslashed assertions are:
\b matches at a word boundary
\B matches when not at a word boundary
\A matches at the start of the subject
\Z matches at the end of the subject
also matches before a newline at the end of the subject
\z matches only at the end of the subject
\G matches at the first matching position in the subject
Inside a character class, \b has a different meaning; it matches the
backspace character. If any other of these assertions appears in a
character class, by default it matches the corresponding literal char-
acter (for example, \B matches the letter B). However, if the
PCRE_EXTRA option is set, an "invalid escape sequence" error is gener-
ated instead.
A word boundary is a position in the subject string where the current
character and the previous character do not both match \w or \W (i.e.
one matches \w and the other matches \W), or the start or end of the
string if the first or last character matches \w, respectively. In
UTF-8 mode, the meanings of \w and \W can be changed by setting the
PCRE_UCP option. When this is done, it also affects \b and \B. Neither
PCRE nor Perl has a separate "start of word" or "end of word" metase-
quence. However, whatever follows \b normally determines which it is.
For example, the fragment \ba matches "a" at the start of a word.
The \A, \Z, and \z assertions differ from the traditional circumflex
and dollar (described in the next section) in that they only ever match
at the very start and end of the subject string, whatever options are
set. Thus, they are independent of multiline mode. These three asser-
tions are not affected by the PCRE_NOTBOL or PCRE_NOTEOL options, which
affect only the behaviour of the circumflex and dollar metacharacters.
However, if the startoffset argument of pcre_exec() is non-zero, indi-
cating that matching is to start at a point other than the beginning of
the subject, \A can never match. The difference between \Z and \z is
that \Z matches before a newline at the end of the string as well as at
the very end, whereas \z matches only at the end.
The \G assertion is true only when the current matching position is at
the start point of the match, as specified by the startoffset argument
of pcre_exec(). It differs from \A when the value of startoffset is
non-zero. By calling pcre_exec() multiple times with appropriate argu-
ments, you can mimic Perl's /g option, and it is in this kind of imple-
mentation where \G can be useful.
Note, however, that PCRE's interpretation of \G, as the start of the
current match, is subtly different from Perl's, which defines it as the
end of the previous match. In Perl, these can be different when the
previously matched string was empty. Because PCRE does just one match
at a time, it cannot reproduce this behaviour.
If all the alternatives of a pattern begin with \G, the expression is
anchored to the starting match position, and the "anchored" flag is set
in the compiled regular expression.
CIRCUMFLEX AND DOLLAR
Outside a character class, in the default matching mode, the circumflex
character is an assertion that is true only if the current matching
point is at the start of the subject string. If the startoffset argu-
ment of pcre_exec() is non-zero, circumflex can never match if the
PCRE_MULTILINE option is unset. Inside a character class, circumflex
has an entirely different meaning (see below).
Circumflex need not be the first character of the pattern if a number
of alternatives are involved, but it should be the first thing in each
alternative in which it appears if the pattern is ever to match that
branch. If all possible alternatives start with a circumflex, that is,
if the pattern is constrained to match only at the start of the sub-
ject, it is said to be an "anchored" pattern. (There are also other
constructs that can cause a pattern to be anchored.)
A dollar character is an assertion that is true only if the current
matching point is at the end of the subject string, or immediately
before a newline at the end of the string (by default). Dollar need not
be the last character of the pattern if a number of alternatives are
involved, but it should be the last item in any branch in which it
appears. Dollar has no special meaning in a character class.
The meaning of dollar can be changed so that it matches only at the
very end of the string, by setting the PCRE_DOLLAR_ENDONLY option at
compile time. This does not affect the \Z assertion.
The meanings of the circumflex and dollar characters are changed if the
PCRE_MULTILINE option is set. When this is the case, a circumflex
matches immediately after internal newlines as well as at the start of
the subject string. It does not match after a newline that ends the
string. A dollar matches before any newlines in the string, as well as
at the very end, when PCRE_MULTILINE is set. When newline is specified
as the two-character sequence CRLF, isolated CR and LF characters do
not indicate newlines.
For example, the pattern /^abc$/ matches the subject string "def\nabc"
(where \n represents a newline) in multiline mode, but not otherwise.
Consequently, patterns that are anchored in single line mode because
all branches start with ^ are not anchored in multiline mode, and a
match for circumflex is possible when the startoffset argument of
pcre_exec() is non-zero. The PCRE_DOLLAR_ENDONLY option is ignored if
PCRE_MULTILINE is set.
Note that the sequences \A, \Z, and \z can be used to match the start
and end of the subject in both modes, and if all branches of a pattern
start with \A it is always anchored, whether or not PCRE_MULTILINE is
set.
FULL STOP (PERIOD, DOT) AND \N
Outside a character class, a dot in the pattern matches any one charac-
ter in the subject string except (by default) a character that signi-
fies the end of a line. In UTF-8 mode, the matched character may be
more than one byte long.
When a line ending is defined as a single character, dot never matches
that character; when the two-character sequence CRLF is used, dot does
not match CR if it is immediately followed by LF, but otherwise it
matches all characters (including isolated CRs and LFs). When any Uni-
code line endings are being recognized, dot does not match CR or LF or
any of the other line ending characters.
The behaviour of dot with regard to newlines can be changed. If the
PCRE_DOTALL option is set, a dot matches any one character, without
exception. If the two-character sequence CRLF is present in the subject
string, it takes two dots to match it.
The handling of dot is entirely independent of the handling of circum-
flex and dollar, the only relationship being that they both involve
newlines. Dot has no special meaning in a character class.
The escape sequence \N behaves like a dot, except that it is not
affected by the PCRE_DOTALL option. In other words, it matches any
character except one that signifies the end of a line. Perl also uses
\N to match characters by name; PCRE does not support this.
MATCHING A SINGLE BYTE
Outside a character class, the escape sequence \C matches any one byte,
both in and out of UTF-8 mode. Unlike a dot, it always matches line-
ending characters. The feature is provided in Perl in order to match
individual bytes in UTF-8 mode, but it is unclear how it can usefully
be used. Because \C breaks up characters into individual bytes, match-
ing one byte with \C in UTF-8 mode means that the rest of the string
may start with a malformed UTF-8 character. This has undefined results,
because PCRE assumes that it is dealing with valid UTF-8 strings (and
by default it checks this at the start of processing unless the
PCRE_NO_UTF8_CHECK option is used).
PCRE does not allow \C to appear in lookbehind assertions (described
below) in UTF-8 mode, because this would make it impossible to calcu-
late the length of the lookbehind.
In general, the \C escape sequence is best avoided in UTF-8 mode. How-
ever, one way of using it that avoids the problem of malformed UTF-8
characters is to use a lookahead to check the length of the next char-
acter, as in this pattern (ignore white space and line breaks):
(?| (?=[\x00-\x7f])(\C) |
(?=[\x80-\x{7ff}])(\C)(\C) |
(?=[\x{800}-\x{ffff}])(\C)(\C)(\C) |
(?=[\x{10000}-\x{1fffff}])(\C)(\C)(\C)(\C))
A group that starts with (?| resets the capturing parentheses numbers
in each alternative (see "Duplicate Subpattern Numbers" below). The
assertions at the start of each branch check the next UTF-8 character
for values whose encoding uses 1, 2, 3, or 4 bytes, respectively. The
character's individual bytes are then captured by the appropriate num-
ber of groups.
SQUARE BRACKETS AND CHARACTER CLASSES
An opening square bracket introduces a character class, terminated by a
closing square bracket. A closing square bracket on its own is not spe-
cial by default. However, if the PCRE_JAVASCRIPT_COMPAT option is set,
a lone closing square bracket causes a compile-time error. If a closing
square bracket is required as a member of the class, it should be the
first data character in the class (after an initial circumflex, if
present) or escaped with a backslash.
A character class matches a single character in the subject. In UTF-8
mode, the character may be more than one byte long. A matched character
must be in the set of characters defined by the class, unless the first
character in the class definition is a circumflex, in which case the
subject character must not be in the set defined by the class. If a
circumflex is actually required as a member of the class, ensure it is
not the first character, or escape it with a backslash.
For example, the character class [aeiou] matches any lower case vowel,
while [^aeiou] matches any character that is not a lower case vowel.
Note that a circumflex is just a convenient notation for specifying the
characters that are in the class by enumerating those that are not. A
class that starts with a circumflex is not an assertion; it still con-
sumes a character from the subject string, and therefore it fails if
the current pointer is at the end of the string.
In UTF-8 mode, characters with values greater than 255 can be included
in a class as a literal string of bytes, or by using the \x{ escaping
mechanism.
When caseless matching is set, any letters in a class represent both
their upper case and lower case versions, so for example, a caseless
[aeiou] matches "A" as well as "a", and a caseless [^aeiou] does not
match "A", whereas a caseful version would. In UTF-8 mode, PCRE always
understands the concept of case for characters whose values are less
than 128, so caseless matching is always possible. For characters with
higher values, the concept of case is supported if PCRE is compiled
with Unicode property support, but not otherwise. If you want to use
caseless matching in UTF8-mode for characters 128 and above, you must
ensure that PCRE is compiled with Unicode property support as well as
with UTF-8 support.
Characters that might indicate line breaks are never treated in any
special way when matching character classes, whatever line-ending
sequence is in use, and whatever setting of the PCRE_DOTALL and
PCRE_MULTILINE options is used. A class such as [^a] always matches one
of these characters.
The minus (hyphen) character can be used to specify a range of charac-
ters in a character class. For example, [d-m] matches any letter
between d and m, inclusive. If a minus character is required in a
class, it must be escaped with a backslash or appear in a position
where it cannot be interpreted as indicating a range, typically as the
first or last character in the class.
It is not possible to have the literal character "]" as the end charac-
ter of a range. A pattern such as [W-]46] is interpreted as a class of
two characters ("W" and "-") followed by a literal string "46]", so it
would match "W46]" or "-46]". However, if the "]" is escaped with a
backslash it is interpreted as the end of range, so [W-\]46] is inter-
preted as a class containing a range followed by two other characters.
The octal or hexadecimal representation of "]" can also be used to end
a range.
Ranges operate in the collating sequence of character values. They can
also be used for characters specified numerically, for example
[\000-\037]. In UTF-8 mode, ranges can include characters whose values
are greater than 255, for example [\x{100}-\x{2ff}].
If a range that includes letters is used when caseless matching is set,
it matches the letters in either case. For example, [W-c] is equivalent
to [][\\^_`wxyzabc], matched caselessly, and in non-UTF-8 mode, if
character tables for a French locale are in use, [\xc8-\xcb] matches
accented E characters in both cases. In UTF-8 mode, PCRE supports the
concept of case for characters with values greater than 128 only when
it is compiled with Unicode property support.
The character escape sequences \d, \D, \h, \H, \p, \P, \s, \S, \v, \V,
\w, and \W may appear in a character class, and add the characters that
they match to the class. For example, [\dABCDEF] matches any hexadeci-
mal digit. In UTF-8 mode, the PCRE_UCP option affects the meanings of
\d, \s, \w and their upper case partners, just as it does when they
appear outside a character class, as described in the section entitled
"Generic character types" above. The escape sequence \b has a different
meaning inside a character class; it matches the backspace character.
The sequences \B, \N, \R, and \X are not special inside a character
class. Like any other unrecognized escape sequences, they are treated
as the literal characters "B", "N", "R", and "X" by default, but cause
an error if the PCRE_EXTRA option is set.
A circumflex can conveniently be used with the upper case character
types to specify a more restricted set of characters than the matching
lower case type. For example, the class [^\W_] matches any letter or
digit, but not underscore, whereas [\w] includes underscore. A positive
character class should be read as "something OR something OR ..." and a
negative class as "NOT something AND NOT something AND NOT ...".
The only metacharacters that are recognized in character classes are
backslash, hyphen (only where it can be interpreted as specifying a
range), circumflex (only at the start), opening square bracket (only
when it can be interpreted as introducing a POSIX class name - see the
next section), and the terminating closing square bracket. However,
escaping other non-alphanumeric characters does no harm.
POSIX CHARACTER CLASSES
Perl supports the POSIX notation for character classes. This uses names
enclosed by [: and :] within the enclosing square brackets. PCRE also
supports this notation. For example,
[01[:alpha:]%]
matches "0", "1", any alphabetic character, or "%". The supported class
names are:
alnum letters and digits
alpha letters
ascii character codes 0 - 127
blank space or tab only
cntrl control characters
digit decimal digits (same as \d)
graph printing characters, excluding space
lower lower case letters
print printing characters, including space
punct printing characters, excluding letters and digits and space
space white space (not quite the same as \s)
upper upper case letters
word "word" characters (same as \w)
xdigit hexadecimal digits
The "space" characters are HT (9), LF (10), VT (11), FF (12), CR (13),
and space (32). Notice that this list includes the VT character (code
11). This makes "space" different to \s, which does not include VT (for
Perl compatibility).
The name "word" is a Perl extension, and "blank" is a GNU extension
from Perl 5.8. Another Perl extension is negation, which is indicated
by a ^ character after the colon. For example,
[12[:^digit:]]
matches "1", "2", or any non-digit. PCRE (and Perl) also recognize the
POSIX syntax [.ch.] and [=ch=] where "ch" is a "collating element", but
these are not supported, and an error is given if they are encountered.
By default, in UTF-8 mode, characters with values greater than 128 do
not match any of the POSIX character classes. However, if the PCRE_UCP
option is passed to pcre_compile(), some of the classes are changed so
that Unicode character properties are used. This is achieved by replac-
ing the POSIX classes by other sequences, as follows:
[:alnum:] becomes \p{Xan}
[:alpha:] becomes \p{L}
[:blank:] becomes \h
[:digit:] becomes \p{Nd}
[:lower:] becomes \p{Ll}
[:space:] becomes \p{Xps}
[:upper:] becomes \p{Lu}
[:word:] becomes \p{Xwd}
Negated versions, such as [:^alpha:] use \P instead of \p. The other
POSIX classes are unchanged, and match only characters with code points
less than 128.
VERTICAL BAR
Vertical bar characters are used to separate alternative patterns. For
example, the pattern
gilbert|sullivan
matches either "gilbert" or "sullivan". Any number of alternatives may
appear, and an empty alternative is permitted (matching the empty
string). The matching process tries each alternative in turn, from left
to right, and the first one that succeeds is used. If the alternatives
are within a subpattern (defined below), "succeeds" means matching the
rest of the main pattern as well as the alternative in the subpattern.
INTERNAL OPTION SETTING
The settings of the PCRE_CASELESS, PCRE_MULTILINE, PCRE_DOTALL, and
PCRE_EXTENDED options (which are Perl-compatible) can be changed from
within the pattern by a sequence of Perl option letters enclosed
between "(?" and ")". The option letters are
i for PCRE_CASELESS
m for PCRE_MULTILINE
s for PCRE_DOTALL
x for PCRE_EXTENDED
For example, (?im) sets caseless, multiline matching. It is also possi-
ble to unset these options by preceding the letter with a hyphen, and a
combined setting and unsetting such as (?im-sx), which sets PCRE_CASE-
LESS and PCRE_MULTILINE while unsetting PCRE_DOTALL and PCRE_EXTENDED,
is also permitted. If a letter appears both before and after the
hyphen, the option is unset.
The PCRE-specific options PCRE_DUPNAMES, PCRE_UNGREEDY, and PCRE_EXTRA
can be changed in the same way as the Perl-compatible options by using
the characters J, U and X respectively.
When one of these option changes occurs at top level (that is, not
inside subpattern parentheses), the change applies to the remainder of
the pattern that follows. If the change is placed right at the start of
a pattern, PCRE extracts it into the global options (and it will there-
fore show up in data extracted by the pcre_fullinfo() function).
An option change within a subpattern (see below for a description of
subpatterns) affects only that part of the subpattern that follows it,
so
(a(?i)b)c
matches abc and aBc and no other strings (assuming PCRE_CASELESS is not
used). By this means, options can be made to have different settings
in different parts of the pattern. Any changes made in one alternative
do carry on into subsequent branches within the same subpattern. For
example,
(a(?i)b|c)
matches "ab", "aB", "c", and "C", even though when matching "C" the
first branch is abandoned before the option setting. This is because
the effects of option settings happen at compile time. There would be
some very weird behaviour otherwise.
Note: There are other PCRE-specific options that can be set by the
application when the compile or match functions are called. In some
cases the pattern can contain special leading sequences such as (*CRLF)
to override what the application has set or what has been defaulted.
Details are given in the section entitled "Newline sequences" above.
There are also the (*UTF8) and (*UCP) leading sequences that can be
used to set UTF-8 and Unicode property modes; they are equivalent to
setting the PCRE_UTF8 and the PCRE_UCP options, respectively.
SUBPATTERNS
Subpatterns are delimited by parentheses (round brackets), which can be
nested. Turning part of a pattern into a subpattern does two things:
1. It localizes a set of alternatives. For example, the pattern
cat(aract|erpillar|)
matches "cataract", "caterpillar", or "cat". Without the parentheses,
it would match "cataract", "erpillar" or an empty string.
2. It sets up the subpattern as a capturing subpattern. This means
that, when the whole pattern matches, that portion of the subject
string that matched the subpattern is passed back to the caller via the
ovector argument of pcre_exec(). Opening parentheses are counted from
left to right (starting from 1) to obtain numbers for the capturing
subpatterns. For example, if the string "the red king" is matched
against the pattern
the ((red|white) (king|queen))
the captured substrings are "red king", "red", and "king", and are num-
bered 1, 2, and 3, respectively.
The fact that plain parentheses fulfil two functions is not always
helpful. There are often times when a grouping subpattern is required
without a capturing requirement. If an opening parenthesis is followed
by a question mark and a colon, the subpattern does not do any captur-
ing, and is not counted when computing the number of any subsequent
capturing subpatterns. For example, if the string "the white queen" is
matched against the pattern
the ((?:red|white) (king|queen))
the captured substrings are "white queen" and "queen", and are numbered
1 and 2. The maximum number of capturing subpatterns is 65535.
As a convenient shorthand, if any option settings are required at the
start of a non-capturing subpattern, the option letters may appear
between the "?" and the ":". Thus the two patterns
(?i:saturday|sunday)
(?:(?i)saturday|sunday)
match exactly the same set of strings. Because alternative branches are
tried from left to right, and options are not reset until the end of
the subpattern is reached, an option setting in one branch does affect
subsequent branches, so the above patterns match "SUNDAY" as well as
"Saturday".
DUPLICATE SUBPATTERN NUMBERS
Perl 5.10 introduced a feature whereby each alternative in a subpattern
uses the same numbers for its capturing parentheses. Such a subpattern
starts with (?| and is itself a non-capturing subpattern. For example,
consider this pattern:
(?|(Sat)ur|(Sun))day
Because the two alternatives are inside a (?| group, both sets of cap-
turing parentheses are numbered one. Thus, when the pattern matches,
you can look at captured substring number one, whichever alternative
matched. This construct is useful when you want to capture part, but
not all, of one of a number of alternatives. Inside a (?| group, paren-
theses are numbered as usual, but the number is reset at the start of
each branch. The numbers of any capturing parentheses that follow the
subpattern start after the highest number used in any branch. The fol-
lowing example is taken from the Perl documentation. The numbers under-
neath show in which buffer the captured content will be stored.
# before ---------------branch-reset----------- after
/ ( a ) (?| x ( y ) z | (p (q) r) | (t) u (v) ) ( z ) /x
# 1 2 2 3 2 3 4
A back reference to a numbered subpattern uses the most recent value
that is set for that number by any subpattern. The following pattern
matches "abcabc" or "defdef":
/(?|(abc)|(def))\1/
In contrast, a subroutine call to a numbered subpattern always refers
to the first one in the pattern with the given number. The following
pattern matches "abcabc" or "defabc":
/(?|(abc)|(def))(?1)/
If a condition test for a subpattern's having matched refers to a non-
unique number, the test is true if any of the subpatterns of that num-
ber have matched.
An alternative approach to using this "branch reset" feature is to use
duplicate named subpatterns, as described in the next section.
NAMED SUBPATTERNS
Identifying capturing parentheses by number is simple, but it can be
very hard to keep track of the numbers in complicated regular expres-
sions. Furthermore, if an expression is modified, the numbers may
change. To help with this difficulty, PCRE supports the naming of sub-
patterns. This feature was not added to Perl until release 5.10. Python
had the feature earlier, and PCRE introduced it at release 4.0, using
the Python syntax. PCRE now supports both the Perl and the Python syn-
tax. Perl allows identically numbered subpatterns to have different
names, but PCRE does not.
In PCRE, a subpattern can be named in one of three ways: (?<name>...)
or (?'name'...) as in Perl, or (?P<name>...) as in Python. References
to capturing parentheses from other parts of the pattern, such as back
references, recursion, and conditions, can be made by name as well as
by number.
Names consist of up to 32 alphanumeric characters and underscores.
Named capturing parentheses are still allocated numbers as well as
names, exactly as if the names were not present. The PCRE API provides
function calls for extracting the name-to-number translation table from
a compiled pattern. There is also a convenience function for extracting
a captured substring by name.
By default, a name must be unique within a pattern, but it is possible
to relax this constraint by setting the PCRE_DUPNAMES option at compile
time. (Duplicate names are also always permitted for subpatterns with
the same number, set up as described in the previous section.) Dupli-
cate names can be useful for patterns where only one instance of the
named parentheses can match. Suppose you want to match the name of a
weekday, either as a 3-letter abbreviation or as the full name, and in
both cases you want to extract the abbreviation. This pattern (ignoring
the line breaks) does the job:
(?<DN>Mon|Fri|Sun)(?:day)?|
(?<DN>Tue)(?:sday)?|
(?<DN>Wed)(?:nesday)?|
(?<DN>Thu)(?:rsday)?|
(?<DN>Sat)(?:urday)?
There are five capturing substrings, but only one is ever set after a
match. (An alternative way of solving this problem is to use a "branch
reset" subpattern, as described in the previous section.)
The convenience function for extracting the data by name returns the
substring for the first (and in this example, the only) subpattern of
that name that matched. This saves searching to find which numbered
subpattern it was.
If you make a back reference to a non-unique named subpattern from
elsewhere in the pattern, the one that corresponds to the first occur-
rence of the name is used. In the absence of duplicate numbers (see the
previous section) this is the one with the lowest number. If you use a
named reference in a condition test (see the section about conditions
below), either to check whether a subpattern has matched, or to check
for recursion, all subpatterns with the same name are tested. If the
condition is true for any one of them, the overall condition is true.
This is the same behaviour as testing by number. For further details of
the interfaces for handling named subpatterns, see the pcreapi documen-
tation.
Warning: You cannot use different names to distinguish between two sub-
patterns with the same number because PCRE uses only the numbers when
matching. For this reason, an error is given at compile time if differ-
ent names are given to subpatterns with the same number. However, you
can give the same name to subpatterns with the same number, even when
PCRE_DUPNAMES is not set.
REPETITION
Repetition is specified by quantifiers, which can follow any of the
following items:
a literal data character
the dot metacharacter
the \C escape sequence
the \X escape sequence (in UTF-8 mode with Unicode properties)
the \R escape sequence
an escape such as \d or \pL that matches a single character
a character class
a back reference (see next section)
a parenthesized subpattern (including assertions)
a subroutine call to a subpattern (recursive or otherwise)
The general repetition quantifier specifies a minimum and maximum num-
ber of permitted matches, by giving the two numbers in curly brackets
(braces), separated by a comma. The numbers must be less than 65536,
and the first must be less than or equal to the second. For example:
z{2,4}
matches "zz", "zzz", or "zzzz". A closing brace on its own is not a
special character. If the second number is omitted, but the comma is
present, there is no upper limit; if the second number and the comma
are both omitted, the quantifier specifies an exact number of required
matches. Thus
[aeiou]{3,}
matches at least 3 successive vowels, but may match many more, while
\d{8}
matches exactly 8 digits. An opening curly bracket that appears in a
position where a quantifier is not allowed, or one that does not match
the syntax of a quantifier, is taken as a literal character. For exam-
ple, {,6} is not a quantifier, but a literal string of four characters.
In UTF-8 mode, quantifiers apply to UTF-8 characters rather than to
individual bytes. Thus, for example, \x{100}{2} matches two UTF-8 char-
acters, each of which is represented by a two-byte sequence. Similarly,
when Unicode property support is available, \X{3} matches three Unicode
extended sequences, each of which may be several bytes long (and they
may be of different lengths).
The quantifier {0} is permitted, causing the expression to behave as if
the previous item and the quantifier were not present. This may be use-
ful for subpatterns that are referenced as subroutines from elsewhere
in the pattern (but see also the section entitled "Defining subpatterns
for use by reference only" below). Items other than subpatterns that
have a {0} quantifier are omitted from the compiled pattern.
For convenience, the three most common quantifiers have single-charac-
ter abbreviations:
* is equivalent to {0,}
+ is equivalent to {1,}
? is equivalent to {0,1}
It is possible to construct infinite loops by following a subpattern
that can match no characters with a quantifier that has no upper limit,
for example:
(a?)*
Earlier versions of Perl and PCRE used to give an error at compile time
for such patterns. However, because there are cases where this can be
useful, such patterns are now accepted, but if any repetition of the
subpattern does in fact match no characters, the loop is forcibly bro-
ken.
By default, the quantifiers are "greedy", that is, they match as much
as possible (up to the maximum number of permitted times), without
causing the rest of the pattern to fail. The classic example of where
this gives problems is in trying to match comments in C programs. These
appear between /* and */ and within the comment, individual * and /
characters may appear. An attempt to match C comments by applying the
pattern
/\*.*\*/
to the string
/* first comment */ not comment /* second comment */
fails, because it matches the entire string owing to the greediness of
the .* item.
However, if a quantifier is followed by a question mark, it ceases to
be greedy, and instead matches the minimum number of times possible, so
the pattern
/\*.*?\*/
does the right thing with the C comments. The meaning of the various
quantifiers is not otherwise changed, just the preferred number of
matches. Do not confuse this use of question mark with its use as a
quantifier in its own right. Because it has two uses, it can sometimes
appear doubled, as in
\d??\d
which matches one digit by preference, but can match two if that is the
only way the rest of the pattern matches.
If the PCRE_UNGREEDY option is set (an option that is not available in
Perl), the quantifiers are not greedy by default, but individual ones
can be made greedy by following them with a question mark. In other
words, it inverts the default behaviour.
When a parenthesized subpattern is quantified with a minimum repeat
count that is greater than 1 or with a limited maximum, more memory is
required for the compiled pattern, in proportion to the size of the
minimum or maximum.
If a pattern starts with .* or .{0,} and the PCRE_DOTALL option (equiv-
alent to Perl's /s) is set, thus allowing the dot to match newlines,
the pattern is implicitly anchored, because whatever follows will be
tried against every character position in the subject string, so there
is no point in retrying the overall match at any position after the
first. PCRE normally treats such a pattern as though it were preceded
by \A.
In cases where it is known that the subject string contains no new-
lines, it is worth setting PCRE_DOTALL in order to obtain this opti-
mization, or alternatively using ^ to indicate anchoring explicitly.
However, there is one situation where the optimization cannot be used.
When .* is inside capturing parentheses that are the subject of a back
reference elsewhere in the pattern, a match at the start may fail where
a later one succeeds. Consider, for example:
(.*)abc\1
If the subject is "xyz123abc123" the match point is the fourth charac-
ter. For this reason, such a pattern is not implicitly anchored.
When a capturing subpattern is repeated, the value captured is the sub-
string that matched the final iteration. For example, after
(tweedle[dume]{3}\s*)+
has matched "tweedledum tweedledee" the value of the captured substring
is "tweedledee". However, if there are nested capturing subpatterns,
the corresponding captured values may have been set in previous itera-
tions. For example, after
/(a|(b))+/
matches "aba" the value of the second captured substring is "b".
ATOMIC GROUPING AND POSSESSIVE QUANTIFIERS
With both maximizing ("greedy") and minimizing ("ungreedy" or "lazy")
repetition, failure of what follows normally causes the repeated item
to be re-evaluated to see if a different number of repeats allows the
rest of the pattern to match. Sometimes it is useful to prevent this,
either to change the nature of the match, or to cause it fail earlier
than it otherwise might, when the author of the pattern knows there is
no point in carrying on.
Consider, for example, the pattern \d+foo when applied to the subject
line
123456bar
After matching all 6 digits and then failing to match "foo", the normal
action of the matcher is to try again with only 5 digits matching the
\d+ item, and then with 4, and so on, before ultimately failing.
"Atomic grouping" (a term taken from Jeffrey Friedl's book) provides
the means for specifying that once a subpattern has matched, it is not
to be re-evaluated in this way.
If we use atomic grouping for the previous example, the matcher gives
up immediately on failing to match "foo" the first time. The notation
is a kind of special parenthesis, starting with (?> as in this example:
(?>\d+)foo
This kind of parenthesis "locks up" the part of the pattern it con-
tains once it has matched, and a failure further into the pattern is
prevented from backtracking into it. Backtracking past it to previous
items, however, works as normal.
An alternative description is that a subpattern of this type matches
the string of characters that an identical standalone pattern would
match, if anchored at the current point in the subject string.
Atomic grouping subpatterns are not capturing subpatterns. Simple cases
such as the above example can be thought of as a maximizing repeat that
must swallow everything it can. So, while both \d+ and \d+? are pre-
pared to adjust the number of digits they match in order to make the
rest of the pattern match, (?>\d+) can only match an entire sequence of
digits.
Atomic groups in general can of course contain arbitrarily complicated
subpatterns, and can be nested. However, when the subpattern for an
atomic group is just a single repeated item, as in the example above, a
simpler notation, called a "possessive quantifier" can be used. This
consists of an additional + character following a quantifier. Using
this notation, the previous example can be rewritten as
\d++foo
Note that a possessive quantifier can be used with an entire group, for
example:
(abc|xyz){2,3}+
Possessive quantifiers are always greedy; the setting of the
PCRE_UNGREEDY option is ignored. They are a convenient notation for the
simpler forms of atomic group. However, there is no difference in the
meaning of a possessive quantifier and the equivalent atomic group,
though there may be a performance difference; possessive quantifiers
should be slightly faster.
The possessive quantifier syntax is an extension to the Perl 5.8 syn-
tax. Jeffrey Friedl originated the idea (and the name) in the first
edition of his book. Mike McCloskey liked it, so implemented it when he
built Sun's Java package, and PCRE copied it from there. It ultimately
found its way into Perl at release 5.10.
PCRE has an optimization that automatically "possessifies" certain sim-
ple pattern constructs. For example, the sequence A+B is treated as
A++B because there is no point in backtracking into a sequence of A's
when B must follow.
When a pattern contains an unlimited repeat inside a subpattern that
can itself be repeated an unlimited number of times, the use of an
atomic group is the only way to avoid some failing matches taking a
very long time indeed. The pattern
(\D+|<\d+>)*[!?]
matches an unlimited number of substrings that either consist of non-
digits, or digits enclosed in <>, followed by either ! or ?. When it
matches, it runs quickly. However, if it is applied to
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
it takes a long time before reporting failure. This is because the
string can be divided between the internal \D+ repeat and the external
* repeat in a large number of ways, and all have to be tried. (The
example uses [!?] rather than a single character at the end, because
both PCRE and Perl have an optimization that allows for fast failure
when a single character is used. They remember the last single charac-
ter that is required for a match, and fail early if it is not present
in the string.) If the pattern is changed so that it uses an atomic
group, like this:
((?>\D+)|<\d+>)*[!?]
sequences of non-digits cannot be broken, and failure happens quickly.
BACK REFERENCES
Outside a character class, a backslash followed by a digit greater than
0 (and possibly further digits) is a back reference to a capturing sub-
pattern earlier (that is, to its left) in the pattern, provided there
have been that many previous capturing left parentheses.
However, if the decimal number following the backslash is less than 10,
it is always taken as a back reference, and causes an error only if
there are not that many capturing left parentheses in the entire pat-
tern. In other words, the parentheses that are referenced need not be
to the left of the reference for numbers less than 10. A "forward back
reference" of this type can make sense when a repetition is involved
and the subpattern to the right has participated in an earlier itera-
tion.
It is not possible to have a numerical "forward back reference" to a
subpattern whose number is 10 or more using this syntax because a
sequence such as \50 is interpreted as a character defined in octal.
See the subsection entitled "Non-printing characters" above for further
details of the handling of digits following a backslash. There is no
such problem when named parentheses are used. A back reference to any
subpattern is possible using named parentheses (see below).
Another way of avoiding the ambiguity inherent in the use of digits
following a backslash is to use the \g escape sequence. This escape
must be followed by an unsigned number or a negative number, optionally
enclosed in braces. These examples are all identical:
(ring), \1
(ring), \g1
(ring), \g{1}
An unsigned number specifies an absolute reference without the ambigu-
ity that is present in the older syntax. It is also useful when literal
digits follow the reference. A negative number is a relative reference.
Consider this example:
(abc(def)ghi)\g{-1}
The sequence \g{-1} is a reference to the most recently started captur-
ing subpattern before \g, that is, is it equivalent to \2 in this exam-
ple. Similarly, \g{-2} would be equivalent to \1. The use of relative
references can be helpful in long patterns, and also in patterns that
are created by joining together fragments that contain references
within themselves.
A back reference matches whatever actually matched the capturing sub-
pattern in the current subject string, rather than anything matching
the subpattern itself (see "Subpatterns as subroutines" below for a way
of doing that). So the pattern
(sens|respons)e and \1ibility
matches "sense and sensibility" and "response and responsibility", but
not "sense and responsibility". If caseful matching is in force at the
time of the back reference, the case of letters is relevant. For exam-
ple,
((?i)rah)\s+\1
matches "rah rah" and "RAH RAH", but not "RAH rah", even though the
original capturing subpattern is matched caselessly.
There are several different ways of writing back references to named
subpatterns. The .NET syntax \k{name} and the Perl syntax \k<name> or
\k'name' are supported, as is the Python syntax (?P=name). Perl 5.10's
unified back reference syntax, in which \g can be used for both numeric
and named references, is also supported. We could rewrite the above
example in any of the following ways:
(?<p1>(?i)rah)\s+\k<p1>
(?'p1'(?i)rah)\s+\k{p1}
(?P<p1>(?i)rah)\s+(?P=p1)
(?<p1>(?i)rah)\s+\g{p1}
A subpattern that is referenced by name may appear in the pattern
before or after the reference.
There may be more than one back reference to the same subpattern. If a
subpattern has not actually been used in a particular match, any back
references to it always fail by default. For example, the pattern
(a|(bc))\2
always fails if it starts to match "a" rather than "bc". However, if
the PCRE_JAVASCRIPT_COMPAT option is set at compile time, a back refer-
ence to an unset value matches an empty string.
Because there may be many capturing parentheses in a pattern, all dig-
its following a backslash are taken as part of a potential back refer-
ence number. If the pattern continues with a digit character, some
delimiter must be used to terminate the back reference. If the
PCRE_EXTENDED option is set, this can be whitespace. Otherwise, the \g{
syntax or an empty comment (see "Comments" below) can be used.
Recursive back references
A back reference that occurs inside the parentheses to which it refers
fails when the subpattern is first used, so, for example, (a\1) never
matches. However, such references can be useful inside repeated sub-
patterns. For example, the pattern
(a|b\1)+
matches any number of "a"s and also "aba", "ababbaa" etc. At each iter-
ation of the subpattern, the back reference matches the character
string corresponding to the previous iteration. In order for this to
work, the pattern must be such that the first iteration does not need
to match the back reference. This can be done using alternation, as in
the example above, or by a quantifier with a minimum of zero.
Back references of this type cause the group that they reference to be
treated as an atomic group. Once the whole group has been matched, a
subsequent matching failure cannot cause backtracking into the middle
of the group.
ASSERTIONS
An assertion is a test on the characters following or preceding the
current matching point that does not actually consume any characters.
The simple assertions coded as \b, \B, \A, \G, \Z, \z, ^ and $ are
described above.
More complicated assertions are coded as subpatterns. There are two
kinds: those that look ahead of the current position in the subject
string, and those that look behind it. An assertion subpattern is
matched in the normal way, except that it does not cause the current
matching position to be changed.
Assertion subpatterns are not capturing subpatterns. If such an asser-
tion contains capturing subpatterns within it, these are counted for
the purposes of numbering the capturing subpatterns in the whole pat-
tern. However, substring capturing is carried out only for positive
assertions, because it does not make sense for negative assertions.
For compatibility with Perl, assertion subpatterns may be repeated;
though it makes no sense to assert the same thing several times, the
side effect of capturing parentheses may occasionally be useful. In
practice, there only three cases:
(1) If the quantifier is {0}, the assertion is never obeyed during
matching. However, it may contain internal capturing parenthesized
groups that are called from elsewhere via the subroutine mechanism.
(2) If quantifier is {0,n} where n is greater than zero, it is treated
as if it were {0,1}. At run time, the rest of the pattern match is
tried with and without the assertion, the order depending on the greed-
iness of the quantifier.
(3) If the minimum repetition is greater than zero, the quantifier is
ignored. The assertion is obeyed just once when encountered during
matching.
Lookahead assertions
Lookahead assertions start with (?= for positive assertions and (?! for
negative assertions. For example,
\w+(?=;)
matches a word followed by a semicolon, but does not include the semi-
colon in the match, and
foo(?!bar)
matches any occurrence of "foo" that is not followed by "bar". Note
that the apparently similar pattern
(?!foo)bar
does not find an occurrence of "bar" that is preceded by something
other than "foo"; it finds any occurrence of "bar" whatsoever, because
the assertion (?!foo) is always true when the next three characters are
"bar". A lookbehind assertion is needed to achieve the other effect.
If you want to force a matching failure at some point in a pattern, the
most convenient way to do it is with (?!) because an empty string
always matches, so an assertion that requires there not to be an empty
string must always fail. The backtracking control verb (*FAIL) or (*F)
is a synonym for (?!).
Lookbehind assertions
Lookbehind assertions start with (?<= for positive assertions and (?<!
for negative assertions. For example,
(?<!foo)bar
does find an occurrence of "bar" that is not preceded by "foo". The
contents of a lookbehind assertion are restricted such that all the
strings it matches must have a fixed length. However, if there are sev-
eral top-level alternatives, they do not all have to have the same
fixed length. Thus
(?<=bullock|donkey)
is permitted, but
(?<!dogs?|cats?)
causes an error at compile time. Branches that match different length
strings are permitted only at the top level of a lookbehind assertion.
This is an extension compared with Perl, which requires all branches to
match the same length of string. An assertion such as
(?<=ab(c|de))
is not permitted, because its single top-level branch can match two
different lengths, but it is acceptable to PCRE if rewritten to use two
top-level branches:
(?<=abc|abde)
In some cases, the escape sequence \K (see above) can be used instead
of a lookbehind assertion to get round the fixed-length restriction.
The implementation of lookbehind assertions is, for each alternative,
to temporarily move the current position back by the fixed length and
then try to match. If there are insufficient characters before the cur-
rent position, the assertion fails.
In UTF-8 mode, PCRE does not allow the \C escape (which matches a sin-
gle byte, even in UTF-8 mode) to appear in lookbehind assertions,
because it makes it impossible to calculate the length of the lookbe-
hind. The \X and \R escapes, which can match different numbers of
bytes, are also not permitted.
"Subroutine" calls (see below) such as (?2) or (?&X) are permitted in
lookbehinds, as long as the subpattern matches a fixed-length string.
Recursion, however, is not supported.
Possessive quantifiers can be used in conjunction with lookbehind
assertions to specify efficient matching of fixed-length strings at the
end of subject strings. Consider a simple pattern such as
abcd$
when applied to a long string that does not match. Because matching
proceeds from left to right, PCRE will look for each "a" in the subject
and then see if what follows matches the rest of the pattern. If the
pattern is specified as
^.*abcd$
the initial .* matches the entire string at first, but when this fails
(because there is no following "a"), it backtracks to match all but the
last character, then all but the last two characters, and so on. Once
again the search for "a" covers the entire string, from right to left,
so we are no better off. However, if the pattern is written as
^.*+(?<=abcd)
there can be no backtracking for the .*+ item; it can match only the
entire string. The subsequent lookbehind assertion does a single test
on the last four characters. If it fails, the match fails immediately.
For long strings, this approach makes a significant difference to the
processing time.
Using multiple assertions
Several assertions (of any sort) may occur in succession. For example,
(?<=\d{3})(?<!999)foo
matches "foo" preceded by three digits that are not "999". Notice that
each of the assertions is applied independently at the same point in
the subject string. First there is a check that the previous three
characters are all digits, and then there is a check that the same
three characters are not "999". This pattern does not match "foo" pre-
ceded by six characters, the first of which are digits and the last
three of which are not "999". For example, it doesn't match "123abc-
foo". A pattern to do that is
(?<=\d{3}...)(?<!999)foo
This time the first assertion looks at the preceding six characters,
checking that the first three are digits, and then the second assertion
checks that the preceding three characters are not "999".
Assertions can be nested in any combination. For example,
(?<=(?<!foo)bar)baz
matches an occurrence of "baz" that is preceded by "bar" which in turn
is not preceded by "foo", while
(?<=\d{3}(?!999)...)foo
is another pattern that matches "foo" preceded by three digits and any
three characters that are not "999".
CONDITIONAL SUBPATTERNS
It is possible to cause the matching process to obey a subpattern con-
ditionally or to choose between two alternative subpatterns, depending
on the result of an assertion, or whether a specific capturing subpat-
tern has already been matched. The two possible forms of conditional
subpattern are:
(?(condition)yes-pattern)
(?(condition)yes-pattern|no-pattern)
If the condition is satisfied, the yes-pattern is used; otherwise the
no-pattern (if present) is used. If there are more than two alterna-
tives in the subpattern, a compile-time error occurs. Each of the two
alternatives may itself contain nested subpatterns of any form, includ-
ing conditional subpatterns; the restriction to two alternatives
applies only at the level of the condition. This pattern fragment is an
example where the alternatives are complex:
(?(1) (A|B|C) | (D | (?(2)E|F) | E) )
There are four kinds of condition: references to subpatterns, refer-
ences to recursion, a pseudo-condition called DEFINE, and assertions.
Checking for a used subpattern by number
If the text between the parentheses consists of a sequence of digits,
the condition is true if a capturing subpattern of that number has pre-
viously matched. If there is more than one capturing subpattern with
the same number (see the earlier section about duplicate subpattern
numbers), the condition is true if any of them have matched. An alter-
native notation is to precede the digits with a plus or minus sign. In
this case, the subpattern number is relative rather than absolute. The
most recently opened parentheses can be referenced by (?(-1), the next
most recent by (?(-2), and so on. Inside loops it can also make sense
to refer to subsequent groups. The next parentheses to be opened can be
referenced as (?(+1), and so on. (The value zero in any of these forms
is not used; it provokes a compile-time error.)
Consider the following pattern, which contains non-significant white
space to make it more readable (assume the PCRE_EXTENDED option) and to
divide it into three parts for ease of discussion:
( \( )? [^()]+ (?(1) \) )
The first part matches an optional opening parenthesis, and if that
character is present, sets it as the first captured substring. The sec-
ond part matches one or more characters that are not parentheses. The
third part is a conditional subpattern that tests whether or not the
first set of parentheses matched. If they did, that is, if subject
started with an opening parenthesis, the condition is true, and so the
yes-pattern is executed and a closing parenthesis is required. Other-
wise, since no-pattern is not present, the subpattern matches nothing.
In other words, this pattern matches a sequence of non-parentheses,
optionally enclosed in parentheses.
If you were embedding this pattern in a larger one, you could use a
relative reference:
...other stuff... ( \( )? [^()]+ (?(-1) \) ) ...
This makes the fragment independent of the parentheses in the larger
pattern.
Checking for a used subpattern by name
Perl uses the syntax (?(<name>)...) or (?('name')...) to test for a
used subpattern by name. For compatibility with earlier versions of
PCRE, which had this facility before Perl, the syntax (?(name)...) is
also recognized. However, there is a possible ambiguity with this syn-
tax, because subpattern names may consist entirely of digits. PCRE
looks first for a named subpattern; if it cannot find one and the name
consists entirely of digits, PCRE looks for a subpattern of that num-
ber, which must be greater than zero. Using subpattern names that con-
sist entirely of digits is not recommended.
Rewriting the above example to use a named subpattern gives this:
(?<OPEN> \( )? [^()]+ (?(<OPEN>) \) )
If the name used in a condition of this kind is a duplicate, the test
is applied to all subpatterns of the same name, and is true if any one
of them has matched.
Checking for pattern recursion
If the condition is the string (R), and there is no subpattern with the
name R, the condition is true if a recursive call to the whole pattern
or any subpattern has been made. If digits or a name preceded by amper-
sand follow the letter R, for example:
(?(R3)...) or (?(R&name)...)
the condition is true if the most recent recursion is into a subpattern
whose number or name is given. This condition does not check the entire
recursion stack. If the name used in a condition of this kind is a
duplicate, the test is applied to all subpatterns of the same name, and
is true if any one of them is the most recent recursion.
At "top level", all these recursion test conditions are false. The
syntax for recursive patterns is described below.
Defining subpatterns for use by reference only
If the condition is the string (DEFINE), and there is no subpattern
with the name DEFINE, the condition is always false. In this case,
there may be only one alternative in the subpattern. It is always
skipped if control reaches this point in the pattern; the idea of
DEFINE is that it can be used to define subroutines that can be refer-
enced from elsewhere. (The use of subroutines is described below.) For
example, a pattern to match an IPv4 address such as "192.168.23.245"
could be written like this (ignore whitespace and line breaks):
(?(DEFINE) (?<byte> 2[0-4]\d | 25[0-5] | 1\d\d | [1-9]?\d) )
\b (?&byte) (\.(?&byte)){3} \b
The first part of the pattern is a DEFINE group inside which a another
group named "byte" is defined. This matches an individual component of
an IPv4 address (a number less than 256). When matching takes place,
this part of the pattern is skipped because DEFINE acts like a false
condition. The rest of the pattern uses references to the named group
to match the four dot-separated components of an IPv4 address, insist-
ing on a word boundary at each end.
Assertion conditions
If the condition is not in any of the above formats, it must be an
assertion. This may be a positive or negative lookahead or lookbehind
assertion. Consider this pattern, again containing non-significant
white space, and with the two alternatives on the second line:
(?(?=[^a-z]*[a-z])
\d{2}-[a-z]{3}-\d{2} | \d{2}-\d{2}-\d{2} )
The condition is a positive lookahead assertion that matches an
optional sequence of non-letters followed by a letter. In other words,
it tests for the presence of at least one letter in the subject. If a
letter is found, the subject is matched against the first alternative;
otherwise it is matched against the second. This pattern matches
strings in one of the two forms dd-aaa-dd or dd-dd-dd, where aaa are
letters and dd are digits.
COMMENTS
There are two ways of including comments in patterns that are processed
by PCRE. In both cases, the start of the comment must not be in a char-
acter class, nor in the middle of any other sequence of related charac-
ters such as (?: or a subpattern name or number. The characters that
make up a comment play no part in the pattern matching.
The sequence (?# marks the start of a comment that continues up to the
next closing parenthesis. Nested parentheses are not permitted. If the
PCRE_EXTENDED option is set, an unescaped # character also introduces a
comment, which in this case continues to immediately after the next
newline character or character sequence in the pattern. Which charac-
ters are interpreted as newlines is controlled by the options passed to
pcre_compile() or by a special sequence at the start of the pattern, as
described in the section entitled "Newline conventions" above. Note
that the end of this type of comment is a literal newline sequence in
the pattern; escape sequences that happen to represent a newline do not
count. For example, consider this pattern when PCRE_EXTENDED is set,
and the default newline convention is in force:
abc #comment \n still comment
On encountering the # character, pcre_compile() skips along, looking
for a newline in the pattern. The sequence \n is still literal at this
stage, so it does not terminate the comment. Only an actual character
with the code value 0x0a (the default newline) does so.
RECURSIVE PATTERNS
Consider the problem of matching a string in parentheses, allowing for
unlimited nested parentheses. Without the use of recursion, the best
that can be done is to use a pattern that matches up to some fixed
depth of nesting. It is not possible to handle an arbitrary nesting
depth.
For some time, Perl has provided a facility that allows regular expres-
sions to recurse (amongst other things). It does this by interpolating
Perl code in the expression at run time, and the code can refer to the
expression itself. A Perl pattern using code interpolation to solve the
parentheses problem can be created like this:
$re = qr{\( (?: (?>[^()]+) | (?p{$re}) )* \)}x;
The (?p{...}) item interpolates Perl code at run time, and in this case
refers recursively to the pattern in which it appears.
Obviously, PCRE cannot support the interpolation of Perl code. Instead,
it supports special syntax for recursion of the entire pattern, and
also for individual subpattern recursion. After its introduction in
PCRE and Python, this kind of recursion was subsequently introduced
into Perl at release 5.10.
A special item that consists of (? followed by a number greater than
zero and a closing parenthesis is a recursive subroutine call of the
subpattern of the given number, provided that it occurs inside that
subpattern. (If not, it is a non-recursive subroutine call, which is
described in the next section.) The special item (?R) or (?0) is a
recursive call of the entire regular expression.
This PCRE pattern solves the nested parentheses problem (assume the
PCRE_EXTENDED option is set so that white space is ignored):
\( ( [^()]++ | (?R) )* \)
First it matches an opening parenthesis. Then it matches any number of
substrings which can either be a sequence of non-parentheses, or a
recursive match of the pattern itself (that is, a correctly parenthe-
sized substring). Finally there is a closing parenthesis. Note the use
of a possessive quantifier to avoid backtracking into sequences of non-
parentheses.
If this were part of a larger pattern, you would not want to recurse
the entire pattern, so instead you could use this:
( \( ( [^()]++ | (?1) )* \) )
We have put the pattern into parentheses, and caused the recursion to
refer to them instead of the whole pattern.
In a larger pattern, keeping track of parenthesis numbers can be
tricky. This is made easier by the use of relative references. Instead
of (?1) in the pattern above you can write (?-2) to refer to the second
most recently opened parentheses preceding the recursion. In other
words, a negative number counts capturing parentheses leftwards from
the point at which it is encountered.
It is also possible to refer to subsequently opened parentheses, by
writing references such as (?+2). However, these cannot be recursive
because the reference is not inside the parentheses that are refer-
enced. They are always non-recursive subroutine calls, as described in
the next section.
An alternative approach is to use named parentheses instead. The Perl
syntax for this is (?&name); PCRE's earlier syntax (?P>name) is also
supported. We could rewrite the above example as follows:
(?<pn> \( ( [^()]++ | (?&pn) )* \) )
If there is more than one subpattern with the same name, the earliest
one is used.
This particular example pattern that we have been looking at contains
nested unlimited repeats, and so the use of a possessive quantifier for
matching strings of non-parentheses is important when applying the pat-
tern to strings that do not match. For example, when this pattern is
applied to
(aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa()
it yields "no match" quickly. However, if a possessive quantifier is
not used, the match runs for a very long time indeed because there are
so many different ways the + and * repeats can carve up the subject,
and all have to be tested before failure can be reported.
At the end of a match, the values of capturing parentheses are those
from the outermost level. If you want to obtain intermediate values, a
callout function can be used (see below and the pcrecallout documenta-
tion). If the pattern above is matched against
(ab(cd)ef)
the value for the inner capturing parentheses (numbered 2) is "ef",
which is the last value taken on at the top level. If a capturing sub-
pattern is not matched at the top level, its final captured value is
unset, even if it was (temporarily) set at a deeper level during the
matching process.
If there are more than 15 capturing parentheses in a pattern, PCRE has
to obtain extra memory to store data during a recursion, which it does
by using pcre_malloc, freeing it via pcre_free afterwards. If no memory
can be obtained, the match fails with the PCRE_ERROR_NOMEMORY error.
Do not confuse the (?R) item with the condition (R), which tests for
recursion. Consider this pattern, which matches text in angle brack-
ets, allowing for arbitrary nesting. Only digits are allowed in nested
brackets (that is, when recursing), whereas any characters are permit-
ted at the outer level.
< (?: (?(R) \d++ | [^<>]*+) | (?R)) * >
In this pattern, (?(R) is the start of a conditional subpattern, with
two different alternatives for the recursive and non-recursive cases.
The (?R) item is the actual recursive call.
Differences in recursion processing between PCRE and Perl
Recursion processing in PCRE differs from Perl in two important ways.
In PCRE (like Python, but unlike Perl), a recursive subpattern call is
always treated as an atomic group. That is, once it has matched some of
the subject string, it is never re-entered, even if it contains untried
alternatives and there is a subsequent matching failure. This can be
illustrated by the following pattern, which purports to match a palin-
dromic string that contains an odd number of characters (for example,
"a", "aba", "abcba", "abcdcba"):
^(.|(.)(?1)\2)$
The idea is that it either matches a single character, or two identical
characters surrounding a sub-palindrome. In Perl, this pattern works;
in PCRE it does not if the pattern is longer than three characters.
Consider the subject string "abcba":
At the top level, the first character is matched, but as it is not at
the end of the string, the first alternative fails; the second alterna-
tive is taken and the recursion kicks in. The recursive call to subpat-
tern 1 successfully matches the next character ("b"). (Note that the
beginning and end of line tests are not part of the recursion).
Back at the top level, the next character ("c") is compared with what
subpattern 2 matched, which was "a". This fails. Because the recursion
is treated as an atomic group, there are now no backtracking points,
and so the entire match fails. (Perl is able, at this point, to re-
enter the recursion and try the second alternative.) However, if the
pattern is written with the alternatives in the other order, things are
different:
^((.)(?1)\2|.)$
This time, the recursing alternative is tried first, and continues to
recurse until it runs out of characters, at which point the recursion
fails. But this time we do have another alternative to try at the
higher level. That is the big difference: in the previous case the
remaining alternative is at a deeper recursion level, which PCRE cannot
use.
To change the pattern so that it matches all palindromic strings, not
just those with an odd number of characters, it is tempting to change
the pattern to this:
^((.)(?1)\2|.?)$
Again, this works in Perl, but not in PCRE, and for the same reason.
When a deeper recursion has matched a single character, it cannot be
entered again in order to match an empty string. The solution is to
separate the two cases, and write out the odd and even cases as alter-
natives at the higher level:
^(?:((.)(?1)\2|)|((.)(?3)\4|.))
If you want to match typical palindromic phrases, the pattern has to
ignore all non-word characters, which can be done like this:
^\W*+(?:((.)\W*+(?1)\W*+\2|)|((.)\W*+(?3)\W*+\4|\W*+.\W*+))\W*+$
If run with the PCRE_CASELESS option, this pattern matches phrases such
as "A man, a plan, a canal: Panama!" and it works well in both PCRE and
Perl. Note the use of the possessive quantifier *+ to avoid backtrack-
ing into sequences of non-word characters. Without this, PCRE takes a
great deal longer (ten times or more) to match typical phrases, and
Perl takes so long that you think it has gone into a loop.
WARNING: The palindrome-matching patterns above work only if the sub-
ject string does not start with a palindrome that is shorter than the
entire string. For example, although "abcba" is correctly matched, if
the subject is "ababa", PCRE finds the palindrome "aba" at the start,
then fails at top level because the end of the string does not follow.
Once again, it cannot jump back into the recursion to try other alter-
natives, so the entire match fails.
The second way in which PCRE and Perl differ in their recursion pro-
cessing is in the handling of captured values. In Perl, when a subpat-
tern is called recursively or as a subpattern (see the next section),
it has no access to any values that were captured outside the recur-
sion, whereas in PCRE these values can be referenced. Consider this
pattern:
^(.)(\1|a(?2))
In PCRE, this pattern matches "bab". The first capturing parentheses
match "b", then in the second group, when the back reference \1 fails
to match "b", the second alternative matches "a" and then recurses. In
the recursion, \1 does now match "b" and so the whole match succeeds.
In Perl, the pattern fails to match because inside the recursive call
\1 cannot access the externally set value.
SUBPATTERNS AS SUBROUTINES
If the syntax for a recursive subpattern call (either by number or by
name) is used outside the parentheses to which it refers, it operates
like a subroutine in a programming language. The called subpattern may
be defined before or after the reference. A numbered reference can be
absolute or relative, as in these examples:
(...(absolute)...)...(?2)...
(...(relative)...)...(?-1)...
(...(?+1)...(relative)...
An earlier example pointed out that the pattern
(sens|respons)e and \1ibility
matches "sense and sensibility" and "response and responsibility", but
not "sense and responsibility". If instead the pattern
(sens|respons)e and (?1)ibility
is used, it does match "sense and responsibility" as well as the other
two strings. Another example is given in the discussion of DEFINE
above.
All subroutine calls, whether recursive or not, are always treated as
atomic groups. That is, once a subroutine has matched some of the sub-
ject string, it is never re-entered, even if it contains untried alter-
natives and there is a subsequent matching failure. Any capturing
parentheses that are set during the subroutine call revert to their
previous values afterwards.
Processing options such as case-independence are fixed when a subpat-
tern is defined, so if it is used as a subroutine, such options cannot
be changed for different calls. For example, consider this pattern:
(abc)(?i:(?-1))
It matches "abcabc". It does not match "abcABC" because the change of
processing option does not affect the called subpattern.
ONIGURUMA SUBROUTINE SYNTAX
For compatibility with Oniguruma, the non-Perl syntax \g followed by a
name or a number enclosed either in angle brackets or single quotes, is
an alternative syntax for referencing a subpattern as a subroutine,
possibly recursively. Here are two of the examples used above, rewrit-
ten using this syntax:
(?<pn> \( ( (?>[^()]+) | \g<pn> )* \) )
(sens|respons)e and \g'1'ibility
PCRE supports an extension to Oniguruma: if a number is preceded by a
plus or a minus sign it is taken as a relative reference. For example:
(abc)(?i:\g<-1>)
Note that \g{...} (Perl syntax) and \g<...> (Oniguruma syntax) are not
synonymous. The former is a back reference; the latter is a subroutine
call.
CALLOUTS
Perl has a feature whereby using the sequence (?{...}) causes arbitrary
Perl code to be obeyed in the middle of matching a regular expression.
This makes it possible, amongst other things, to extract different sub-
strings that match the same pair of parentheses when there is a repeti-
tion.
PCRE provides a similar feature, but of course it cannot obey arbitrary
Perl code. The feature is called "callout". The caller of PCRE provides
an external function by putting its entry point in the global variable
pcre_callout. By default, this variable contains NULL, which disables
all calling out.
Within a regular expression, (?C) indicates the points at which the
external function is to be called. If you want to identify different
callout points, you can put a number less than 256 after the letter C.
The default value is zero. For example, this pattern has two callout
points:
(?C1)abc(?C2)def
If the PCRE_AUTO_CALLOUT flag is passed to pcre_compile(), callouts are
automatically installed before each item in the pattern. They are all
numbered 255.
During matching, when PCRE reaches a callout point (and pcre_callout is
set), the external function is called. It is provided with the number
of the callout, the position in the pattern, and, optionally, one item
of data originally supplied by the caller of pcre_exec(). The callout
function may cause matching to proceed, to backtrack, or to fail alto-
gether. A complete description of the interface to the callout function
is given in the pcrecallout documentation.
BACKTRACKING CONTROL
Perl 5.10 introduced a number of "Special Backtracking Control Verbs",
which are described in the Perl documentation as "experimental and sub-
ject to change or removal in a future version of Perl". It goes on to
say: "Their usage in production code should be noted to avoid problems
during upgrades." The same remarks apply to the PCRE features described
in this section.
Since these verbs are specifically related to backtracking, most of
them can be used only when the pattern is to be matched using
pcre_exec(), which uses a backtracking algorithm. With the exception of
(*FAIL), which behaves like a failing negative assertion, they cause an
error if encountered by pcre_dfa_exec().
If any of these verbs are used in an assertion or in a subpattern that
is called as a subroutine (whether or not recursively), their effect is
confined to that subpattern; it does not extend to the surrounding pat-
tern, with one exception: the name from a *(MARK), (*PRUNE), or (*THEN)
that is encountered in a successful positive assertion is passed back
when a match succeeds (compare capturing parentheses in assertions).
Note that such subpatterns are processed as anchored at the point where
they are tested. Note also that Perl's treatment of subroutines is dif-
ferent in some cases.
The new verbs make use of what was previously invalid syntax: an open-
ing parenthesis followed by an asterisk. They are generally of the form
(*VERB) or (*VERB:NAME). Some may take either form, with differing be-
haviour, depending on whether or not an argument is present. A name is
any sequence of characters that does not include a closing parenthesis.
If the name is empty, that is, if the closing parenthesis immediately
follows the colon, the effect is as if the colon were not there. Any
number of these verbs may occur in a pattern.
PCRE contains some optimizations that are used to speed up matching by
running some checks at the start of each match attempt. For example, it
may know the minimum length of matching subject, or that a particular
character must be present. When one of these optimizations suppresses
the running of a match, any included backtracking verbs will not, of
course, be processed. You can suppress the start-of-match optimizations
by setting the PCRE_NO_START_OPTIMIZE option when calling pcre_com-
pile() or pcre_exec(), or by starting the pattern with (*NO_START_OPT).
Experiments with Perl suggest that it too has similar optimizations,
sometimes leading to anomalous results.
Verbs that act immediately
The following verbs act as soon as they are encountered. They may not
be followed by a name.
(*ACCEPT)
This verb causes the match to end successfully, skipping the remainder
of the pattern. However, when it is inside a subpattern that is called
as a subroutine, only that subpattern is ended successfully. Matching
then continues at the outer level. If (*ACCEPT) is inside capturing
parentheses, the data so far is captured. For example:
A((?:A|B(*ACCEPT)|C)D)
This matches "AB", "AAD", or "ACD"; when it matches "AB", "B" is cap-
tured by the outer parentheses.
(*FAIL) or (*F)
This verb causes a matching failure, forcing backtracking to occur. It
is equivalent to (?!) but easier to read. The Perl documentation notes
that it is probably useful only when combined with (?{}) or (??{}).
Those are, of course, Perl features that are not present in PCRE. The
nearest equivalent is the callout feature, as for example in this pat-
tern:
a+(?C)(*FAIL)
A match with the string "aaaa" always fails, but the callout is taken
before each backtrack happens (in this example, 10 times).
Recording which path was taken
There is one verb whose main purpose is to track how a match was
arrived at, though it also has a secondary use in conjunction with
advancing the match starting point (see (*SKIP) below).
(*MARK:NAME) or (*:NAME)
A name is always required with this verb. There may be as many
instances of (*MARK) as you like in a pattern, and their names do not
have to be unique.
When a match succeeds, the name of the last-encountered (*MARK) on the
matching path is passed back to the caller via the pcre_extra data
structure, as described in the section on pcre_extra in the pcreapi
documentation. Here is an example of pcretest output, where the /K mod-
ifier requests the retrieval and outputting of (*MARK) data:
re> /X(*MARK:A)Y|X(*MARK:B)Z/K
data> XY
0: XY
MK: A
XZ
0: XZ
MK: B
The (*MARK) name is tagged with "MK:" in this output, and in this exam-
ple it indicates which of the two alternatives matched. This is a more
efficient way of obtaining this information than putting each alterna-
tive in its own capturing parentheses.
If (*MARK) is encountered in a positive assertion, its name is recorded
and passed back if it is the last-encountered. This does not happen for
negative assertions.
After a partial match or a failed match, the name of the last encoun-
tered (*MARK) in the entire match process is returned. For example:
re> /X(*MARK:A)Y|X(*MARK:B)Z/K
data> XP
No match, mark = B
Note that in this unanchored example the mark is retained from the
match attempt that started at the letter "X". Subsequent match attempts
starting at "P" and then with an empty string do not get as far as the
(*MARK) item, but nevertheless do not reset it.
Verbs that act after backtracking
The following verbs do nothing when they are encountered. Matching con-
tinues with what follows, but if there is no subsequent match, causing
a backtrack to the verb, a failure is forced. That is, backtracking
cannot pass to the left of the verb. However, when one of these verbs
appears inside an atomic group, its effect is confined to that group,
because once the group has been matched, there is never any backtrack-
ing into it. In this situation, backtracking can "jump back" to the
left of the entire atomic group. (Remember also, as stated above, that
this localization also applies in subroutine calls and assertions.)
These verbs differ in exactly what kind of failure occurs when back-
tracking reaches them.
(*COMMIT)
This verb, which may not be followed by a name, causes the whole match
to fail outright if the rest of the pattern does not match. Even if the
pattern is unanchored, no further attempts to find a match by advancing
the starting point take place. Once (*COMMIT) has been passed,
pcre_exec() is committed to finding a match at the current starting
point, or not at all. For example:
a+(*COMMIT)b
This matches "xxaab" but not "aacaab". It can be thought of as a kind
of dynamic anchor, or "I've started, so I must finish." The name of the
most recently passed (*MARK) in the path is passed back when (*COMMIT)
forces a match failure.
Note that (*COMMIT) at the start of a pattern is not the same as an
anchor, unless PCRE's start-of-match optimizations are turned off, as
shown in this pcretest example:
re> /(*COMMIT)abc/
data> xyzabc
0: abc
xyzabc\Y
No match
PCRE knows that any match must start with "a", so the optimization
skips along the subject to "a" before running the first match attempt,
which succeeds. When the optimization is disabled by the \Y escape in
the second subject, the match starts at "x" and so the (*COMMIT) causes
it to fail without trying any other starting points.
(*PRUNE) or (*PRUNE:NAME)
This verb causes the match to fail at the current starting position in
the subject if the rest of the pattern does not match. If the pattern
is unanchored, the normal "bumpalong" advance to the next starting
character then happens. Backtracking can occur as usual to the left of
(*PRUNE), before it is reached, or when matching to the right of
(*PRUNE), but if there is no match to the right, backtracking cannot
cross (*PRUNE). In simple cases, the use of (*PRUNE) is just an alter-
native to an atomic group or possessive quantifier, but there are some
uses of (*PRUNE) that cannot be expressed in any other way. The behav-
iour of (*PRUNE:NAME) is the same as (*MARK:NAME)(*PRUNE). In an
anchored pattern (*PRUNE) has the same effect as (*COMMIT).
(*SKIP)
This verb, when given without a name, is like (*PRUNE), except that if
the pattern is unanchored, the "bumpalong" advance is not to the next
character, but to the position in the subject where (*SKIP) was encoun-
tered. (*SKIP) signifies that whatever text was matched leading up to
it cannot be part of a successful match. Consider:
a+(*SKIP)b
If the subject is "aaaac...", after the first match attempt fails
(starting at the first character in the string), the starting point
skips on to start the next attempt at "c". Note that a possessive quan-
tifer does not have the same effect as this example; although it would
suppress backtracking during the first match attempt, the second
attempt would start at the second character instead of skipping on to
"c".
(*SKIP:NAME)
When (*SKIP) has an associated name, its behaviour is modified. If the
following pattern fails to match, the previous path through the pattern
is searched for the most recent (*MARK) that has the same name. If one
is found, the "bumpalong" advance is to the subject position that cor-
responds to that (*MARK) instead of to where (*SKIP) was encountered.
If no (*MARK) with a matching name is found, the (*SKIP) is ignored.
(*THEN) or (*THEN:NAME)
This verb causes a skip to the next innermost alternative if the rest
of the pattern does not match. That is, it cancels pending backtrack-
ing, but only within the current alternative. Its name comes from the
observation that it can be used for a pattern-based if-then-else block:
( COND1 (*THEN) FOO | COND2 (*THEN) BAR | COND3 (*THEN) BAZ ) ...
If the COND1 pattern matches, FOO is tried (and possibly further items
after the end of the group if FOO succeeds); on failure, the matcher
skips to the second alternative and tries COND2, without backtracking
into COND1. The behaviour of (*THEN:NAME) is exactly the same as
(*MARK:NAME)(*THEN). If (*THEN) is not inside an alternation, it acts
like (*PRUNE).
Note that a subpattern that does not contain a | character is just a
part of the enclosing alternative; it is not a nested alternation with
only one alternative. The effect of (*THEN) extends beyond such a sub-
pattern to the enclosing alternative. Consider this pattern, where A,
B, etc. are complex pattern fragments that do not contain any | charac-
ters at this level:
A (B(*THEN)C) | D
If A and B are matched, but there is a failure in C, matching does not
backtrack into A; instead it moves to the next alternative, that is, D.
However, if the subpattern containing (*THEN) is given an alternative,
it behaves differently:
A (B(*THEN)C | (*FAIL)) | D
The effect of (*THEN) is now confined to the inner subpattern. After a
failure in C, matching moves to (*FAIL), which causes the whole subpat-
tern to fail because there are no more alternatives to try. In this
case, matching does now backtrack into A.
Note also that a conditional subpattern is not considered as having two
alternatives, because only one is ever used. In other words, the |
character in a conditional subpattern has a different meaning. Ignoring
white space, consider:
^.*? (?(?=a) a | b(*THEN)c )
If the subject is "ba", this pattern does not match. Because .*? is
ungreedy, it initially matches zero characters. The condition (?=a)
then fails, the character "b" is matched, but "c" is not. At this
point, matching does not backtrack to .*? as might perhaps be expected
from the presence of the | character. The conditional subpattern is
part of the single alternative that comprises the whole pattern, and so
the match fails. (If there was a backtrack into .*?, allowing it to
match "b", the match would succeed.)
The verbs just described provide four different "strengths" of control
when subsequent matching fails. (*THEN) is the weakest, carrying on the
match at the next alternative. (*PRUNE) comes next, failing the match
at the current starting position, but allowing an advance to the next
character (for an unanchored pattern). (*SKIP) is similar, except that
the advance may be more than one character. (*COMMIT) is the strongest,
causing the entire match to fail.
If more than one such verb is present in a pattern, the "strongest" one
wins. For example, consider this pattern, where A, B, etc. are complex
pattern fragments:
(A(*COMMIT)B(*THEN)C|D)
Once A has matched, PCRE is committed to this match, at the current
starting position. If subsequently B matches, but C does not, the nor-
mal (*THEN) action of trying the next alternative (that is, D) does not
happen because (*COMMIT) overrides.
SEE ALSO
pcreapi(3), pcrecallout(3), pcrematching(3), pcresyntax(3), pcre(3).
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 29 November 2011
Copyright (c) 1997-2011 University of Cambridge.
------------------------------------------------------------------------------
PCRESYNTAX(3) PCRESYNTAX(3)
NAME
PCRE - Perl-compatible regular expressions
PCRE REGULAR EXPRESSION SYNTAX SUMMARY
The full syntax and semantics of the regular expressions that are sup-
ported by PCRE are described in the pcrepattern documentation. This
document contains just a quick-reference summary of the syntax.
QUOTING
\x where x is non-alphanumeric is a literal x
\Q...\E treat enclosed characters as literal
CHARACTERS
\a alarm, that is, the BEL character (hex 07)
\cx "control-x", where x is any ASCII character
\e escape (hex 1B)
\f formfeed (hex 0C)
\n newline (hex 0A)
\r carriage return (hex 0D)
\t tab (hex 09)
\ddd character with octal code ddd, or backreference
\xhh character with hex code hh
\x{hhh..} character with hex code hhh..
CHARACTER TYPES
. any character except newline;
in dotall mode, any character whatsoever
\C one byte, even in UTF-8 mode (best avoided)
\d a decimal digit
\D a character that is not a decimal digit
\h a horizontal whitespace character
\H a character that is not a horizontal whitespace character
\N a character that is not a newline
\p{xx} a character with the xx property
\P{xx} a character without the xx property
\R a newline sequence
\s a whitespace character
\S a character that is not a whitespace character
\v a vertical whitespace character
\V a character that is not a vertical whitespace character
\w a "word" character
\W a "non-word" character
\X an extended Unicode sequence
In PCRE, by default, \d, \D, \s, \S, \w, and \W recognize only ASCII
characters, even in UTF-8 mode. However, this can be changed by setting
the PCRE_UCP option.
GENERAL CATEGORY PROPERTIES FOR \p and \P
C Other
Cc Control
Cf Format
Cn Unassigned
Co Private use
Cs Surrogate
L Letter
Ll Lower case letter
Lm Modifier letter
Lo Other letter
Lt Title case letter
Lu Upper case letter
L& Ll, Lu, or Lt
M Mark
Mc Spacing mark
Me Enclosing mark
Mn Non-spacing mark
N Number
Nd Decimal number
Nl Letter number
No Other number
P Punctuation
Pc Connector punctuation
Pd Dash punctuation
Pe Close punctuation
Pf Final punctuation
Pi Initial punctuation
Po Other punctuation
Ps Open punctuation
S Symbol
Sc Currency symbol
Sk Modifier symbol
Sm Mathematical symbol
So Other symbol
Z Separator
Zl Line separator
Zp Paragraph separator
Zs Space separator
PCRE SPECIAL CATEGORY PROPERTIES FOR \p and \P
Xan Alphanumeric: union of properties L and N
Xps POSIX space: property Z or tab, NL, VT, FF, CR
Xsp Perl space: property Z or tab, NL, FF, CR
Xwd Perl word: property Xan or underscore
SCRIPT NAMES FOR \p AND \P
Arabic, Armenian, Avestan, Balinese, Bamum, Bengali, Bopomofo, Braille,
Buginese, Buhid, Canadian_Aboriginal, Carian, Cham, Cherokee, Common,
Coptic, Cuneiform, Cypriot, Cyrillic, Deseret, Devanagari, Egyp-
tian_Hieroglyphs, Ethiopic, Georgian, Glagolitic, Gothic, Greek,
Gujarati, Gurmukhi, Han, Hangul, Hanunoo, Hebrew, Hiragana, Impe-
rial_Aramaic, Inherited, Inscriptional_Pahlavi, Inscriptional_Parthian,
Javanese, Kaithi, Kannada, Katakana, Kayah_Li, Kharoshthi, Khmer, Lao,
Latin, Lepcha, Limbu, Linear_B, Lisu, Lycian, Lydian, Malayalam,
Meetei_Mayek, Mongolian, Myanmar, New_Tai_Lue, Nko, Ogham, Old_Italic,
Old_Persian, Old_South_Arabian, Old_Turkic, Ol_Chiki, Oriya, Osmanya,
Phags_Pa, Phoenician, Rejang, Runic, Samaritan, Saurashtra, Shavian,
Sinhala, Sundanese, Syloti_Nagri, Syriac, Tagalog, Tagbanwa, Tai_Le,
Tai_Tham, Tai_Viet, Tamil, Telugu, Thaana, Thai, Tibetan, Tifinagh,
Ugaritic, Vai, Yi.
CHARACTER CLASSES
[...] positive character class
[^...] negative character class
[x-y] range (can be used for hex characters)
[[:xxx:]] positive POSIX named set
[[:^xxx:]] negative POSIX named set
alnum alphanumeric
alpha alphabetic
ascii 0-127
blank space or tab
cntrl control character
digit decimal digit
graph printing, excluding space
lower lower case letter
print printing, including space
punct printing, excluding alphanumeric
space whitespace
upper upper case letter
word same as \w
xdigit hexadecimal digit
In PCRE, POSIX character set names recognize only ASCII characters by
default, but some of them use Unicode properties if PCRE_UCP is set.
You can use \Q...\E inside a character class.
QUANTIFIERS
? 0 or 1, greedy
?+ 0 or 1, possessive
?? 0 or 1, lazy
* 0 or more, greedy
*+ 0 or more, possessive
*? 0 or more, lazy
+ 1 or more, greedy
++ 1 or more, possessive
+? 1 or more, lazy
{n} exactly n
{n,m} at least n, no more than m, greedy
{n,m}+ at least n, no more than m, possessive
{n,m}? at least n, no more than m, lazy
{n,} n or more, greedy
{n,}+ n or more, possessive
{n,}? n or more, lazy
ANCHORS AND SIMPLE ASSERTIONS
\b word boundary
\B not a word boundary
^ start of subject
also after internal newline in multiline mode
\A start of subject
$ end of subject
also before newline at end of subject
also before internal newline in multiline mode
\Z end of subject
also before newline at end of subject
\z end of subject
\G first matching position in subject
MATCH POINT RESET
\K reset start of match
ALTERNATION
expr|expr|expr...
CAPTURING
(...) capturing group
(?<name>...) named capturing group (Perl)
(?'name'...) named capturing group (Perl)
(?P<name>...) named capturing group (Python)
(?:...) non-capturing group
(?|...) non-capturing group; reset group numbers for
capturing groups in each alternative
ATOMIC GROUPS
(?>...) atomic, non-capturing group
COMMENT
(?#....) comment (not nestable)
OPTION SETTING
(?i) caseless
(?J) allow duplicate names
(?m) multiline
(?s) single line (dotall)
(?U) default ungreedy (lazy)
(?x) extended (ignore white space)
(?-...) unset option(s)
The following are recognized only at the start of a pattern or after
one of the newline-setting options with similar syntax:
(*NO_START_OPT) no start-match optimization (PCRE_NO_START_OPTIMIZE)
(*UTF8) set UTF-8 mode (PCRE_UTF8)
(*UCP) set PCRE_UCP (use Unicode properties for \d etc)
LOOKAHEAD AND LOOKBEHIND ASSERTIONS
(?=...) positive look ahead
(?!...) negative look ahead
(?<=...) positive look behind
(?<!...) negative look behind
Each top-level branch of a look behind must be of a fixed length.
BACKREFERENCES
\n reference by number (can be ambiguous)
\gn reference by number
\g{n} reference by number
\g{-n} relative reference by number
\k<name> reference by name (Perl)
\k'name' reference by name (Perl)
\g{name} reference by name (Perl)
\k{name} reference by name (.NET)
(?P=name) reference by name (Python)
SUBROUTINE REFERENCES (POSSIBLY RECURSIVE)
(?R) recurse whole pattern
(?n) call subpattern by absolute number
(?+n) call subpattern by relative number
(?-n) call subpattern by relative number
(?&name) call subpattern by name (Perl)
(?P>name) call subpattern by name (Python)
\g<name> call subpattern by name (Oniguruma)
\g'name' call subpattern by name (Oniguruma)
\g<n> call subpattern by absolute number (Oniguruma)
\g'n' call subpattern by absolute number (Oniguruma)
\g<+n> call subpattern by relative number (PCRE extension)
\g'+n' call subpattern by relative number (PCRE extension)
\g<-n> call subpattern by relative number (PCRE extension)
\g'-n' call subpattern by relative number (PCRE extension)
CONDITIONAL PATTERNS
(?(condition)yes-pattern)
(?(condition)yes-pattern|no-pattern)
(?(n)... absolute reference condition
(?(+n)... relative reference condition
(?(-n)... relative reference condition
(?(<name>)... named reference condition (Perl)
(?('name')... named reference condition (Perl)
(?(name)... named reference condition (PCRE)
(?(R)... overall recursion condition
(?(Rn)... specific group recursion condition
(?(R&name)... specific recursion condition
(?(DEFINE)... define subpattern for reference
(?(assert)... assertion condition
BACKTRACKING CONTROL
The following act immediately they are reached:
(*ACCEPT) force successful match
(*FAIL) force backtrack; synonym (*F)
The following act only when a subsequent match failure causes a back-
track to reach them. They all force a match failure, but they differ in
what happens afterwards. Those that advance the start-of-match point do
so only if the pattern is not anchored.
(*COMMIT) overall failure, no advance of starting point
(*PRUNE) advance to next starting character
(*SKIP) advance start to current matching position
(*THEN) local failure, backtrack to next alternation
NEWLINE CONVENTIONS
These are recognized only at the very start of the pattern or after a
(*BSR_...) or (*UTF8) or (*UCP) option.
(*CR) carriage return only
(*LF) linefeed only
(*CRLF) carriage return followed by linefeed
(*ANYCRLF) all three of the above
(*ANY) any Unicode newline sequence
WHAT \R MATCHES
These are recognized only at the very start of the pattern or after a
(*...) option that sets the newline convention or UTF-8 or UCP mode.
(*BSR_ANYCRLF) CR, LF, or CRLF
(*BSR_UNICODE) any Unicode newline sequence
CALLOUTS
(?C) callout
(?Cn) callout with data n
SEE ALSO
pcrepattern(3), pcreapi(3), pcrecallout(3), pcrematching(3), pcre(3).
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 21 November 2010
Copyright (c) 1997-2010 University of Cambridge.
------------------------------------------------------------------------------
PCREUNICODE(3) PCREUNICODE(3)
NAME
PCRE - Perl-compatible regular expressions
UTF-8 AND UNICODE PROPERTY SUPPORT
In order process UTF-8 strings, you must build PCRE to include UTF-8
support in the code, and, in addition, you must call pcre_compile()
with the PCRE_UTF8 option flag, or the pattern must start with the
sequence (*UTF8). When either of these is the case, both the pattern
and any subject strings that are matched against it are treated as
UTF-8 strings instead of strings of 1-byte characters. PCRE does not
support any other formats (in particular, it does not support UTF-16).
If you compile PCRE with UTF-8 support, but do not use it at run time,
the library will be a bit bigger, but the additional run time overhead
is limited to testing the PCRE_UTF8 flag occasionally, so should not be
very big.
If PCRE is built with Unicode character property support (which implies
UTF-8 support), the escape sequences \p{..}, \P{..}, and \X are sup-
ported. The available properties that can be tested are limited to the
general category properties such as Lu for an upper case letter or Nd
for a decimal number, the Unicode script names such as Arabic or Han,
and the derived properties Any and L&. A full list is given in the
pcrepattern documentation. Only the short names for properties are sup-
ported. For example, \p{L} matches a letter. Its Perl synonym, \p{Let-
ter}, is not supported. Furthermore, in Perl, many properties may
optionally be prefixed by "Is", for compatibility with Perl 5.6. PCRE
does not support this.
Validity of UTF-8 strings
When you set the PCRE_UTF8 flag, the strings passed as patterns and
subjects are (by default) checked for validity on entry to the relevant
functions. From release 7.3 of PCRE, the check is according the rules
of RFC 3629, which are themselves derived from the Unicode specifica-
tion. Earlier releases of PCRE followed the rules of RFC 2279, which
allows the full range of 31-bit values (0 to 0x7FFFFFFF). The current
check allows only values in the range U+0 to U+10FFFF, excluding U+D800
to U+DFFF.
The excluded code points are the "Low Surrogate Area" of Unicode, of
which the Unicode Standard says this: "The Low Surrogate Area does not
contain any character assignments, consequently no character code
charts or namelists are provided for this area. Surrogates are reserved
for use with UTF-16 and then must be used in pairs." The code points
that are encoded by UTF-16 pairs are available as independent code
points in the UTF-8 encoding. (In other words, the whole surrogate
thing is a fudge for UTF-16 which unfortunately messes up UTF-8.)
If an invalid UTF-8 string is passed to PCRE, an error return is given.
At compile time, the only additional information is the offset to the
first byte of the failing character. The runtime functions pcre_exec()
and pcre_dfa_exec() also pass back this information, as well as a more
detailed reason code if the caller has provided memory in which to do
this.
In some situations, you may already know that your strings are valid,
and therefore want to skip these checks in order to improve perfor-
mance. If you set the PCRE_NO_UTF8_CHECK flag at compile time or at run
time, PCRE assumes that the pattern or subject it is given (respec-
tively) contains only valid UTF-8 codes. In this case, it does not
diagnose an invalid UTF-8 string.
If you pass an invalid UTF-8 string when PCRE_NO_UTF8_CHECK is set,
what happens depends on why the string is invalid. If the string con-
forms to the "old" definition of UTF-8 (RFC 2279), it is processed as a
string of characters in the range 0 to 0x7FFFFFFF by pcre_dfa_exec()
and the interpreted version of pcre_exec(). In other words, apart from
the initial validity test, these functions (when in UTF-8 mode) handle
strings according to the more liberal rules of RFC 2279. However, the
just-in-time (JIT) optimization for pcre_exec() supports only RFC 3629.
If you are using JIT optimization, or if the string does not even con-
form to RFC 2279, the result is undefined. Your program may crash.
If you want to process strings of values in the full range 0 to
0x7FFFFFFF, encoded in a UTF-8-like manner as per the old RFC, you can
set PCRE_NO_UTF8_CHECK to bypass the more restrictive test. However, in
this situation, you will have to apply your own validity check, and
avoid the use of JIT optimization.
General comments about UTF-8 mode
1. An unbraced hexadecimal escape sequence (such as \xb3) matches a
two-byte UTF-8 character if the value is greater than 127.
2. Octal numbers up to \777 are recognized, and match two-byte UTF-8
characters for values greater than \177.
3. Repeat quantifiers apply to complete UTF-8 characters, not to indi-
vidual bytes, for example: \x{100}{3}.
4. The dot metacharacter matches one UTF-8 character instead of a sin-
gle byte.
5. The escape sequence \C can be used to match a single byte in UTF-8
mode, but its use can lead to some strange effects because it breaks up
multibyte characters (see the description of \C in the pcrepattern doc-
umentation). The use of \C is not supported in the alternative matching
function pcre_dfa_exec(), nor is it supported in UTF-8 mode by the JIT
optimization of pcre_exec(). If JIT optimization is requested for a
UTF-8 pattern that contains \C, it will not succeed, and so the match-
ing will be carried out by the normal interpretive function.
6. The character escapes \b, \B, \d, \D, \s, \S, \w, and \W correctly
test characters of any code value, but, by default, the characters that
PCRE recognizes as digits, spaces, or word characters remain the same
set as before, all with values less than 256. This remains true even
when PCRE is built to include Unicode property support, because to do
otherwise would slow down PCRE in many common cases. Note in particular
that this applies to \b and \B, because they are defined in terms of \w
and \W. If you really want to test for a wider sense of, say, "digit",
you can use explicit Unicode property tests such as \p{Nd}. Alterna-
tively, if you set the PCRE_UCP option, the way that the character
escapes work is changed so that Unicode properties are used to deter-
mine which characters match. There are more details in the section on
generic character types in the pcrepattern documentation.
7. Similarly, characters that match the POSIX named character classes
are all low-valued characters, unless the PCRE_UCP option is set.
8. However, the horizontal and vertical whitespace matching escapes
(\h, \H, \v, and \V) do match all the appropriate Unicode characters,
whether or not PCRE_UCP is set.
9. Case-insensitive matching applies only to characters whose values
are less than 128, unless PCRE is built with Unicode property support.
Even when Unicode property support is available, PCRE still uses its
own character tables when checking the case of low-valued characters,
so as not to degrade performance. The Unicode property information is
used only for characters with higher values. Furthermore, PCRE supports
case-insensitive matching only when there is a one-to-one mapping
between a letter's cases. There are a small number of many-to-one map-
pings in Unicode; these are not supported by PCRE.
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 19 October 2011
Copyright (c) 1997-2011 University of Cambridge.
------------------------------------------------------------------------------
PCREJIT(3) PCREJIT(3)
NAME
PCRE - Perl-compatible regular expressions
PCRE JUST-IN-TIME COMPILER SUPPORT
Just-in-time compiling is a heavyweight optimization that can greatly
speed up pattern matching. However, it comes at the cost of extra pro-
cessing before the match is performed. Therefore, it is of most benefit
when the same pattern is going to be matched many times. This does not
necessarily mean many calls of pcre_exec(); if the pattern is not
anchored, matching attempts may take place many times at various posi-
tions in the subject, even for a single call to pcre_exec(). If the
subject string is very long, it may still pay to use JIT for one-off
matches.
JIT support applies only to the traditional matching function,
pcre_exec(). It does not apply when pcre_dfa_exec() is being used. The
code for this support was written by Zoltan Herczeg.
AVAILABILITY OF JIT SUPPORT
JIT support is an optional feature of PCRE. The "configure" option
--enable-jit (or equivalent CMake option) must be set when PCRE is
built if you want to use JIT. The support is limited to the following
hardware platforms:
ARM v5, v7, and Thumb2
Intel x86 32-bit and 64-bit
MIPS 32-bit
Power PC 32-bit and 64-bit (experimental)
The Power PC support is designated as experimental because it has not
been fully tested. If --enable-jit is set on an unsupported platform,
compilation fails.
A program that is linked with PCRE 8.20 or later can tell if JIT sup-
port is available by calling pcre_config() with the PCRE_CONFIG_JIT
option. The result is 1 when JIT is available, and 0 otherwise. How-
ever, a simple program does not need to check this in order to use JIT.
The API is implemented in a way that falls back to the ordinary PCRE
code if JIT is not available.
If your program may sometimes be linked with versions of PCRE that are
older than 8.20, but you want to use JIT when it is available, you can
test the values of PCRE_MAJOR and PCRE_MINOR, or the existence of a JIT
macro such as PCRE_CONFIG_JIT, for compile-time control of your code.
SIMPLE USE OF JIT
You have to do two things to make use of the JIT support in the sim-
plest way:
(1) Call pcre_study() with the PCRE_STUDY_JIT_COMPILE option for
each compiled pattern, and pass the resulting pcre_extra block to
pcre_exec().
(2) Use pcre_free_study() to free the pcre_extra block when it is
no longer needed instead of just freeing it yourself. This
ensures that any JIT data is also freed.
For a program that may be linked with pre-8.20 versions of PCRE, you
can insert
#ifndef PCRE_STUDY_JIT_COMPILE
#define PCRE_STUDY_JIT_COMPILE 0
#endif
so that no option is passed to pcre_study(), and then use something
like this to free the study data:
#ifdef PCRE_CONFIG_JIT
pcre_free_study(study_ptr);
#else
pcre_free(study_ptr);
#endif
In some circumstances you may need to call additional functions. These
are described in the section entitled "Controlling the JIT stack"
below.
If JIT support is not available, PCRE_STUDY_JIT_COMPILE is ignored, and
no JIT data is set up. Otherwise, the compiled pattern is passed to the
JIT compiler, which turns it into machine code that executes much
faster than the normal interpretive code. When pcre_exec() is passed a
pcre_extra block containing a pointer to JIT code, it obeys that
instead of the normal code. The result is identical, but the code runs
much faster.
There are some pcre_exec() options that are not supported for JIT exe-
cution. There are also some pattern items that JIT cannot handle.
Details are given below. In both cases, execution automatically falls
back to the interpretive code.
If the JIT compiler finds an unsupported item, no JIT data is gener-
ated. You can find out if JIT execution is available after studying a
pattern by calling pcre_fullinfo() with the PCRE_INFO_JIT option. A
result of 1 means that JIT compilation was successful. A result of 0
means that JIT support is not available, or the pattern was not studied
with PCRE_STUDY_JIT_COMPILE, or the JIT compiler was not able to handle
the pattern.
Once a pattern has been studied, with or without JIT, it can be used as
many times as you like for matching different subject strings.
UNSUPPORTED OPTIONS AND PATTERN ITEMS
The only pcre_exec() options that are supported for JIT execution are
PCRE_NO_UTF8_CHECK, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, and
PCRE_NOTEMPTY_ATSTART. Note in particular that partial matching is not
supported.
The unsupported pattern items are:
\C match a single byte; not supported in UTF-8 mode
(?Cn) callouts
(*COMMIT) )
(*MARK) )
(*PRUNE) ) the backtracking control verbs
(*SKIP) )
(*THEN) )
Support for some of these may be added in future.
RETURN VALUES FROM JIT EXECUTION
When a pattern is matched using JIT execution, the return values are
the same as those given by the interpretive pcre_exec() code, with the
addition of one new error code: PCRE_ERROR_JIT_STACKLIMIT. This means
that the memory used for the JIT stack was insufficient. See "Control-
ling the JIT stack" below for a discussion of JIT stack usage. For com-
patibility with the interpretive pcre_exec() code, no more than two-
thirds of the ovector argument is used for passing back captured sub-
strings.
The error code PCRE_ERROR_MATCHLIMIT is returned by the JIT code if
searching a very large pattern tree goes on for too long, as it is in
the same circumstance when JIT is not used, but the details of exactly
what is counted are not the same. The PCRE_ERROR_RECURSIONLIMIT error
code is never returned by JIT execution.
SAVING AND RESTORING COMPILED PATTERNS
The code that is generated by the JIT compiler is architecture-spe-
cific, and is also position dependent. For those reasons it cannot be
saved (in a file or database) and restored later like the bytecode and
other data of a compiled pattern. Saving and restoring compiled pat-
terns is not something many people do. More detail about this facility
is given in the pcreprecompile documentation. It should be possible to
run pcre_study() on a saved and restored pattern, and thereby recreate
the JIT data, but because JIT compilation uses significant resources,
it is probably not worth doing this; you might as well recompile the
original pattern.
CONTROLLING THE JIT STACK
When the compiled JIT code runs, it needs a block of memory to use as a
stack. By default, it uses 32K on the machine stack. However, some
large or complicated patterns need more than this. The error
PCRE_ERROR_JIT_STACKLIMIT is given when there is not enough stack.
Three functions are provided for managing blocks of memory for use as
JIT stacks. There is further discussion about the use of JIT stacks in
the section entitled "JIT stack FAQ" below.
The pcre_jit_stack_alloc() function creates a JIT stack. Its arguments
are a starting size and a maximum size, and it returns a pointer to an
opaque structure of type pcre_jit_stack, or NULL if there is an error.
The pcre_jit_stack_free() function can be used to free a stack that is
no longer needed. (For the technically minded: the address space is
allocated by mmap or VirtualAlloc.)
JIT uses far less memory for recursion than the interpretive code, and
a maximum stack size of 512K to 1M should be more than enough for any
pattern.
The pcre_assign_jit_stack() function specifies which stack JIT code
should use. Its arguments are as follows:
pcre_extra *extra
pcre_jit_callback callback
void *data
The extra argument must be the result of studying a pattern with
PCRE_STUDY_JIT_COMPILE. There are three cases for the values of the
other two options:
(1) If callback is NULL and data is NULL, an internal 32K block
on the machine stack is used.
(2) If callback is NULL and data is not NULL, data must be
a valid JIT stack, the result of calling pcre_jit_stack_alloc().
(3) If callback not NULL, it must point to a function that is called
with data as an argument at the start of matching, in order to
set up a JIT stack. If the result is NULL, the internal 32K stack
is used; otherwise the return value must be a valid JIT stack,
the result of calling pcre_jit_stack_alloc().
You may safely assign the same JIT stack to more than one pattern, as
long as they are all matched sequentially in the same thread. In a mul-
tithread application, each thread must use its own JIT stack.
Strictly speaking, even more is allowed. You can assign the same stack
to any number of patterns as long as they are not used for matching by
multiple threads at the same time. For example, you can assign the same
stack to all compiled patterns, and use a global mutex in the callback
to wait until the stack is available for use. However, this is an inef-
ficient solution, and not recommended.
This is a suggestion for how a typical multithreaded program might
operate:
During thread initalization
thread_local_var = pcre_jit_stack_alloc(...)
During thread exit
pcre_jit_stack_free(thread_local_var)
Use a one-line callback function
return thread_local_var
All the functions described in this section do nothing if JIT is not
available, and pcre_assign_jit_stack() does nothing unless the extra
argument is non-NULL and points to a pcre_extra block that is the
result of a successful study with PCRE_STUDY_JIT_COMPILE.
JIT STACK FAQ
(1) Why do we need JIT stacks?
PCRE (and JIT) is a recursive, depth-first engine, so it needs a stack
where the local data of the current node is pushed before checking its
child nodes. Allocating real machine stack on some platforms is diffi-
cult. For example, the stack chain needs to be updated every time if we
extend the stack on PowerPC. Although it is possible, its updating
time overhead decreases performance. So we do the recursion in memory.
(2) Why don't we simply allocate blocks of memory with malloc()?
Modern operating systems have a nice feature: they can reserve an
address space instead of allocating memory. We can safely allocate mem-
ory pages inside this address space, so the stack could grow without
moving memory data (this is important because of pointers). Thus we can
allocate 1M address space, and use only a single memory page (usually
4K) if that is enough. However, we can still grow up to 1M anytime if
needed.
(3) Who "owns" a JIT stack?
The owner of the stack is the user program, not the JIT studied pattern
or anything else. The user program must ensure that if a stack is used
by pcre_exec(), (that is, it is assigned to the pattern currently run-
ning), that stack must not be used by any other threads (to avoid over-
writing the same memory area). The best practice for multithreaded pro-
grams is to allocate a stack for each thread, and return this stack
through the JIT callback function.
(4) When should a JIT stack be freed?
You can free a JIT stack at any time, as long as it will not be used by
pcre_exec() again. When you assign the stack to a pattern, only a
pointer is set. There is no reference counting or any other magic. You
can free the patterns and stacks in any order, anytime. Just do not
call pcre_exec() with a pattern pointing to an already freed stack, as
that will cause SEGFAULT. (Also, do not free a stack currently used by
pcre_exec() in another thread). You can also replace the stack for a
pattern at any time. You can even free the previous stack before
assigning a replacement.
(5) Should I allocate/free a stack every time before/after calling
pcre_exec()?
No, because this is too costly in terms of resources. However, you
could implement some clever idea which release the stack if it is not
used in let's say two minutes. The JIT callback can help to achive this
without keeping a list of the currently JIT studied patterns.
(6) OK, the stack is for long term memory allocation. But what happens
if a pattern causes stack overflow with a stack of 1M? Is that 1M kept
until the stack is freed?
Especially on embedded sytems, it might be a good idea to release mem-
ory sometimes without freeing the stack. There is no API for this at
the moment. Probably a function call which returns with the currently
allocated memory for any stack and another which allows releasing mem-
ory (shrinking the stack) would be a good idea if someone needs this.
(7) This is too much of a headache. Isn't there any better solution for
JIT stack handling?
No, thanks to Windows. If POSIX threads were used everywhere, we could
throw out this complicated API.
EXAMPLE CODE
This is a single-threaded example that specifies a JIT stack without
using a callback.
int rc;
int ovector[30];
pcre *re;
pcre_extra *extra;
pcre_jit_stack *jit_stack;
re = pcre_compile(pattern, 0, &error, &erroffset, NULL);
/* Check for errors */
extra = pcre_study(re, PCRE_STUDY_JIT_COMPILE, &error);
jit_stack = pcre_jit_stack_alloc(32*1024, 512*1024);
/* Check for error (NULL) */
pcre_assign_jit_stack(extra, NULL, jit_stack);
rc = pcre_exec(re, extra, subject, length, 0, 0, ovector, 30);
/* Check results */
pcre_free(re);
pcre_free_study(extra);
pcre_jit_stack_free(jit_stack);
SEE ALSO
pcreapi(3)
AUTHOR
Philip Hazel (FAQ by Zoltan Herczeg)
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 26 November 2011
Copyright (c) 1997-2011 University of Cambridge.
------------------------------------------------------------------------------
PCREPARTIAL(3) PCREPARTIAL(3)
NAME
PCRE - Perl-compatible regular expressions
PARTIAL MATCHING IN PCRE
In normal use of PCRE, if the subject string that is passed to
pcre_exec() or pcre_dfa_exec() matches as far as it goes, but is too
short to match the entire pattern, PCRE_ERROR_NOMATCH is returned.
There are circumstances where it might be helpful to distinguish this
case from other cases in which there is no match.
Consider, for example, an application where a human is required to type
in data for a field with specific formatting requirements. An example
might be a date in the form ddmmmyy, defined by this pattern:
^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$
If the application sees the user's keystrokes one by one, and can check
that what has been typed so far is potentially valid, it is able to
raise an error as soon as a mistake is made, by beeping and not
reflecting the character that has been typed, for example. This immedi-
ate feedback is likely to be a better user interface than a check that
is delayed until the entire string has been entered. Partial matching
can also be useful when the subject string is very long and is not all
available at once.
PCRE supports partial matching by means of the PCRE_PARTIAL_SOFT and
PCRE_PARTIAL_HARD options, which can be set when calling pcre_exec() or
pcre_dfa_exec(). For backwards compatibility, PCRE_PARTIAL is a synonym
for PCRE_PARTIAL_SOFT. The essential difference between the two options
is whether or not a partial match is preferred to an alternative com-
plete match, though the details differ between the two matching func-
tions. If both options are set, PCRE_PARTIAL_HARD takes precedence.
Setting a partial matching option for pcre_exec() disables the use of
any just-in-time code that was set up by calling pcre_study() with the
PCRE_STUDY_JIT_COMPILE option. It also disables two of PCRE's standard
optimizations. PCRE remembers the last literal byte in a pattern, and
abandons matching immediately if such a byte is not present in the sub-
ject string. This optimization cannot be used for a subject string that
might match only partially. If the pattern was studied, PCRE knows the
minimum length of a matching string, and does not bother to run the
matching function on shorter strings. This optimization is also dis-
abled for partial matching.
PARTIAL MATCHING USING pcre_exec()
A partial match occurs during a call to pcre_exec() when the end of the
subject string is reached successfully, but matching cannot continue
because more characters are needed. However, at least one character in
the subject must have been inspected. This character need not form part
of the final matched string; lookbehind assertions and the \K escape
sequence provide ways of inspecting characters before the start of a
matched substring. The requirement for inspecting at least one charac-
ter exists because an empty string can always be matched; without such
a restriction there would always be a partial match of an empty string
at the end of the subject.
If there are at least two slots in the offsets vector when pcre_exec()
returns with a partial match, the first slot is set to the offset of
the earliest character that was inspected when the partial match was
found. For convenience, the second offset points to the end of the sub-
ject so that a substring can easily be identified.
For the majority of patterns, the first offset identifies the start of
the partially matched string. However, for patterns that contain look-
behind assertions, or \K, or begin with \b or \B, earlier characters
have been inspected while carrying out the match. For example:
/(?<=abc)123/
This pattern matches "123", but only if it is preceded by "abc". If the
subject string is "xyzabc12", the offsets after a partial match are for
the substring "abc12", because all these characters are needed if
another match is tried with extra characters added to the subject.
What happens when a partial match is identified depends on which of the
two partial matching options are set.
PCRE_PARTIAL_SOFT with pcre_exec()
If PCRE_PARTIAL_SOFT is set when pcre_exec() identifies a partial
match, the partial match is remembered, but matching continues as nor-
mal, and other alternatives in the pattern are tried. If no complete
match can be found, pcre_exec() returns PCRE_ERROR_PARTIAL instead of
PCRE_ERROR_NOMATCH.
This option is "soft" because it prefers a complete match over a par-
tial match. All the various matching items in a pattern behave as if
the subject string is potentially complete. For example, \z, \Z, and $
match at the end of the subject, as normal, and for \b and \B the end
of the subject is treated as a non-alphanumeric.
If there is more than one partial match, the first one that was found
provides the data that is returned. Consider this pattern:
/123\w+X|dogY/
If this is matched against the subject string "abc123dog", both alter-
natives fail to match, but the end of the subject is reached during
matching, so PCRE_ERROR_PARTIAL is returned. The offsets are set to 3
and 9, identifying "123dog" as the first partial match that was found.
(In this example, there are two partial matches, because "dog" on its
own partially matches the second alternative.)
PCRE_PARTIAL_HARD with pcre_exec()
If PCRE_PARTIAL_HARD is set for pcre_exec(), it returns PCRE_ERROR_PAR-
TIAL as soon as a partial match is found, without continuing to search
for possible complete matches. This option is "hard" because it prefers
an earlier partial match over a later complete match. For this reason,
the assumption is made that the end of the supplied subject string may
not be the true end of the available data, and so, if \z, \Z, \b, \B,
or $ are encountered at the end of the subject, the result is
PCRE_ERROR_PARTIAL.
Setting PCRE_PARTIAL_HARD also affects the way pcre_exec() checks UTF-8
subject strings for validity. Normally, an invalid UTF-8 sequence
causes the error PCRE_ERROR_BADUTF8. However, in the special case of a
truncated UTF-8 character at the end of the subject, PCRE_ERROR_SHORT-
UTF8 is returned when PCRE_PARTIAL_HARD is set.
Comparing hard and soft partial matching
The difference between the two partial matching options can be illus-
trated by a pattern such as:
/dog(sbody)?/
This matches either "dog" or "dogsbody", greedily (that is, it prefers
the longer string if possible). If it is matched against the string
"dog" with PCRE_PARTIAL_SOFT, it yields a complete match for "dog".
However, if PCRE_PARTIAL_HARD is set, the result is PCRE_ERROR_PARTIAL.
On the other hand, if the pattern is made ungreedy the result is dif-
ferent:
/dog(sbody)??/
In this case the result is always a complete match because pcre_exec()
finds that first, and it never continues after finding a match. It
might be easier to follow this explanation by thinking of the two pat-
terns like this:
/dog(sbody)?/ is the same as /dogsbody|dog/
/dog(sbody)??/ is the same as /dog|dogsbody/
The second pattern will never match "dogsbody" when pcre_exec() is
used, because it will always find the shorter match first.
PARTIAL MATCHING USING pcre_dfa_exec()
The pcre_dfa_exec() function moves along the subject string character
by character, without backtracking, searching for all possible matches
simultaneously. If the end of the subject is reached before the end of
the pattern, there is the possibility of a partial match, again pro-
vided that at least one character has been inspected.
When PCRE_PARTIAL_SOFT is set, PCRE_ERROR_PARTIAL is returned only if
there have been no complete matches. Otherwise, the complete matches
are returned. However, if PCRE_PARTIAL_HARD is set, a partial match
takes precedence over any complete matches. The portion of the string
that was inspected when the longest partial match was found is set as
the first matching string, provided there are at least two slots in the
offsets vector.
Because pcre_dfa_exec() always searches for all possible matches, and
there is no difference between greedy and ungreedy repetition, its be-
haviour is different from pcre_exec when PCRE_PARTIAL_HARD is set. Con-
sider the string "dog" matched against the ungreedy pattern shown
above:
/dog(sbody)??/
Whereas pcre_exec() stops as soon as it finds the complete match for
"dog", pcre_dfa_exec() also finds the partial match for "dogsbody", and
so returns that when PCRE_PARTIAL_HARD is set.
PARTIAL MATCHING AND WORD BOUNDARIES
If a pattern ends with one of sequences \b or \B, which test for word
boundaries, partial matching with PCRE_PARTIAL_SOFT can give counter-
intuitive results. Consider this pattern:
/\bcat\b/
This matches "cat", provided there is a word boundary at either end. If
the subject string is "the cat", the comparison of the final "t" with a
following character cannot take place, so a partial match is found.
However, pcre_exec() carries on with normal matching, which matches \b
at the end of the subject when the last character is a letter, thus
finding a complete match. The result, therefore, is not PCRE_ERROR_PAR-
TIAL. The same thing happens with pcre_dfa_exec(), because it also
finds the complete match.
Using PCRE_PARTIAL_HARD in this case does yield PCRE_ERROR_PARTIAL,
because then the partial match takes precedence.
FORMERLY RESTRICTED PATTERNS
For releases of PCRE prior to 8.00, because of the way certain internal
optimizations were implemented in the pcre_exec() function, the
PCRE_PARTIAL option (predecessor of PCRE_PARTIAL_SOFT) could not be
used with all patterns. From release 8.00 onwards, the restrictions no
longer apply, and partial matching with pcre_exec() can be requested
for any pattern.
Items that were formerly restricted were repeated single characters and
repeated metasequences. If PCRE_PARTIAL was set for a pattern that did
not conform to the restrictions, pcre_exec() returned the error code
PCRE_ERROR_BADPARTIAL (-13). This error code is no longer in use. The
PCRE_INFO_OKPARTIAL call to pcre_fullinfo() to find out if a compiled
pattern can be used for partial matching now always returns 1.
EXAMPLE OF PARTIAL MATCHING USING PCRETEST
If the escape sequence \P is present in a pcretest data line, the
PCRE_PARTIAL_SOFT option is used for the match. Here is a run of
pcretest that uses the date example quoted above:
re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
data> 25jun04\P
0: 25jun04
1: jun
data> 25dec3\P
Partial match: 23dec3
data> 3ju\P
Partial match: 3ju
data> 3juj\P
No match
data> j\P
No match
The first data string is matched completely, so pcretest shows the
matched substrings. The remaining four strings do not match the com-
plete pattern, but the first two are partial matches. Similar output is
obtained when pcre_dfa_exec() is used.
If the escape sequence \P is present more than once in a pcretest data
line, the PCRE_PARTIAL_HARD option is set for the match.
MULTI-SEGMENT MATCHING WITH pcre_dfa_exec()
When a partial match has been found using pcre_dfa_exec(), it is possi-
ble to continue the match by providing additional subject data and
calling pcre_dfa_exec() again with the same compiled regular expres-
sion, this time setting the PCRE_DFA_RESTART option. You must pass the
same working space as before, because this is where details of the pre-
vious partial match are stored. Here is an example using pcretest,
using the \R escape sequence to set the PCRE_DFA_RESTART option (\D
specifies the use of pcre_dfa_exec()):
re> /^\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d$/
data> 23ja\P\D
Partial match: 23ja
data> n05\R\D
0: n05
The first call has "23ja" as the subject, and requests partial match-
ing; the second call has "n05" as the subject for the continued
(restarted) match. Notice that when the match is complete, only the
last part is shown; PCRE does not retain the previously partially-
matched string. It is up to the calling program to do that if it needs
to.
You can set the PCRE_PARTIAL_SOFT or PCRE_PARTIAL_HARD options with
PCRE_DFA_RESTART to continue partial matching over multiple segments.
This facility can be used to pass very long subject strings to
pcre_dfa_exec().
MULTI-SEGMENT MATCHING WITH pcre_exec()
From release 8.00, pcre_exec() can also be used to do multi-segment
matching. Unlike pcre_dfa_exec(), it is not possible to restart the
previous match with a new segment of data. Instead, new data must be
added to the previous subject string, and the entire match re-run,
starting from the point where the partial match occurred. Earlier data
can be discarded. It is best to use PCRE_PARTIAL_HARD in this situa-
tion, because it does not treat the end of a segment as the end of the
subject when matching \z, \Z, \b, \B, and $. Consider an unanchored
pattern that matches dates:
re> /\d?\d(jan|feb|mar|apr|may|jun|jul|aug|sep|oct|nov|dec)\d\d/
data> The date is 23ja\P\P
Partial match: 23ja
At this stage, an application could discard the text preceding "23ja",
add on text from the next segment, and call pcre_exec() again. Unlike
pcre_dfa_exec(), the entire matching string must always be available,
and the complete matching process occurs for each call, so more memory
and more processing time is needed.
Note: If the pattern contains lookbehind assertions, or \K, or starts
with \b or \B, the string that is returned for a partial match will
include characters that precede the partially matched string itself,
because these must be retained when adding on more characters for a
subsequent matching attempt.
ISSUES WITH MULTI-SEGMENT MATCHING
Certain types of pattern may give problems with multi-segment matching,
whichever matching function is used.
1. If the pattern contains a test for the beginning of a line, you need
to pass the PCRE_NOTBOL option when the subject string for any call
does start at the beginning of a line. There is also a PCRE_NOTEOL
option, but in practice when doing multi-segment matching you should be
using PCRE_PARTIAL_HARD, which includes the effect of PCRE_NOTEOL.
2. Lookbehind assertions at the start of a pattern are catered for in
the offsets that are returned for a partial match. However, in theory,
a lookbehind assertion later in the pattern could require even earlier
characters to be inspected, and it might not have been reached when a
partial match occurs. This is probably an extremely unlikely case; you
could guard against it to a certain extent by always including extra
characters at the start.
3. Matching a subject string that is split into multiple segments may
not always produce exactly the same result as matching over one single
long string, especially when PCRE_PARTIAL_SOFT is used. The section
"Partial Matching and Word Boundaries" above describes an issue that
arises if the pattern ends with \b or \B. Another kind of difference
may occur when there are multiple matching possibilities, because (for
PCRE_PARTIAL_SOFT) a partial match result is given only when there are
no completed matches. This means that as soon as the shortest match has
been found, continuation to a new subject segment is no longer possi-
ble. Consider again this pcretest example:
re> /dog(sbody)?/
data> dogsb\P
0: dog
data> do\P\D
Partial match: do
data> gsb\R\P\D
0: g
data> dogsbody\D
0: dogsbody
1: dog
The first data line passes the string "dogsb" to pcre_exec(), setting
the PCRE_PARTIAL_SOFT option. Although the string is a partial match
for "dogsbody", the result is not PCRE_ERROR_PARTIAL, because the
shorter string "dog" is a complete match. Similarly, when the subject
is presented to pcre_dfa_exec() in several parts ("do" and "gsb" being
the first two) the match stops when "dog" has been found, and it is not
possible to continue. On the other hand, if "dogsbody" is presented as
a single string, pcre_dfa_exec() finds both matches.
Because of these problems, it is best to use PCRE_PARTIAL_HARD when
matching multi-segment data. The example above then behaves differ-
ently:
re> /dog(sbody)?/
data> dogsb\P\P
Partial match: dogsb
data> do\P\D
Partial match: do
data> gsb\R\P\P\D
Partial match: gsb
4. Patterns that contain alternatives at the top level which do not all
start with the same pattern item may not work as expected when
PCRE_DFA_RESTART is used with pcre_dfa_exec(). For example, consider
this pattern:
1234|3789
If the first part of the subject is "ABC123", a partial match of the
first alternative is found at offset 3. There is no partial match for
the second alternative, because such a match does not start at the same
point in the subject string. Attempting to continue with the string
"7890" does not yield a match because only those alternatives that
match at one point in the subject are remembered. The problem arises
because the start of the second alternative matches within the first
alternative. There is no problem with anchored patterns or patterns
such as:
1234|ABCD
where no string can be a partial match for both alternatives. This is
not a problem if pcre_exec() is used, because the entire match has to
be rerun each time:
re> /1234|3789/
data> ABC123\P\P
Partial match: 123
data> 1237890
0: 3789
Of course, instead of using PCRE_DFA_RESTART, the same technique of re-
running the entire match can also be used with pcre_dfa_exec(). Another
possibility is to work with two buffers. If a partial match at offset n
in the first buffer is followed by "no match" when PCRE_DFA_RESTART is
used on the second buffer, you can then try a new match starting at
offset n+1 in the first buffer.
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 26 August 2011
Copyright (c) 1997-2011 University of Cambridge.
------------------------------------------------------------------------------
PCREPRECOMPILE(3) PCREPRECOMPILE(3)
NAME
PCRE - Perl-compatible regular expressions
SAVING AND RE-USING PRECOMPILED PCRE PATTERNS
If you are running an application that uses a large number of regular
expression patterns, it may be useful to store them in a precompiled
form instead of having to compile them every time the application is
run. If you are not using any private character tables (see the
pcre_maketables() documentation), this is relatively straightforward.
If you are using private tables, it is a little bit more complicated.
However, if you are using the just-in-time optimization feature of
pcre_study(), it is not possible to save and reload the JIT data.
If you save compiled patterns to a file, you can copy them to a differ-
ent host and run them there. This works even if the new host has the
opposite endianness to the one on which the patterns were compiled.
There may be a small performance penalty, but it should be insignifi-
cant. However, compiling regular expressions with one version of PCRE
for use with a different version is not guaranteed to work and may
cause crashes, and saving and restoring a compiled pattern loses any
JIT optimization data.
SAVING A COMPILED PATTERN
The value returned by pcre_compile() points to a single block of memory
that holds the compiled pattern and associated data. You can find the
length of this block in bytes by calling pcre_fullinfo() with an argu-
ment of PCRE_INFO_SIZE. You can then save the data in any appropriate
manner. Here is sample code that compiles a pattern and writes it to a
file. It assumes that the variable fd refers to a file that is open for
output:
int erroroffset, rc, size;
char *error;
pcre *re;
re = pcre_compile("my pattern", 0, &error, &erroroffset, NULL);
if (re == NULL) { ... handle errors ... }
rc = pcre_fullinfo(re, NULL, PCRE_INFO_SIZE, &size);
if (rc < 0) { ... handle errors ... }
rc = fwrite(re, 1, size, fd);
if (rc != size) { ... handle errors ... }
In this example, the bytes that comprise the compiled pattern are
copied exactly. Note that this is binary data that may contain any of
the 256 possible byte values. On systems that make a distinction
between binary and non-binary data, be sure that the file is opened for
binary output.
If you want to write more than one pattern to a file, you will have to
devise a way of separating them. For binary data, preceding each pat-
tern with its length is probably the most straightforward approach.
Another possibility is to write out the data in hexadecimal instead of
binary, one pattern to a line.
Saving compiled patterns in a file is only one possible way of storing
them for later use. They could equally well be saved in a database, or
in the memory of some daemon process that passes them via sockets to
the processes that want them.
If the pattern has been studied, it is also possible to save the normal
study data in a similar way to the compiled pattern itself. However, if
the PCRE_STUDY_JIT_COMPILE was used, the just-in-time data that is cre-
ated cannot be saved because it is too dependent on the current envi-
ronment. When studying generates additional information, pcre_study()
returns a pointer to a pcre_extra data block. Its format is defined in
the section on matching a pattern in the pcreapi documentation. The
study_data field points to the binary study data, and this is what you
must save (not the pcre_extra block itself). The length of the study
data can be obtained by calling pcre_fullinfo() with an argument of
PCRE_INFO_STUDYSIZE. Remember to check that pcre_study() did return a
non-NULL value before trying to save the study data.
RE-USING A PRECOMPILED PATTERN
Re-using a precompiled pattern is straightforward. Having reloaded it
into main memory, you pass its pointer to pcre_exec() or
pcre_dfa_exec() in the usual way. This should work even on another
host, and even if that host has the opposite endianness to the one
where the pattern was compiled.
However, if you passed a pointer to custom character tables when the
pattern was compiled (the tableptr argument of pcre_compile()), you
must now pass a similar pointer to pcre_exec() or pcre_dfa_exec(),
because the value saved with the compiled pattern will obviously be
nonsense. A field in a pcre_extra() block is used to pass this data, as
described in the section on matching a pattern in the pcreapi documen-
tation.
If you did not provide custom character tables when the pattern was
compiled, the pointer in the compiled pattern is NULL, which causes
pcre_exec() to use PCRE's internal tables. Thus, you do not need to
take any special action at run time in this case.
If you saved study data with the compiled pattern, you need to create
your own pcre_extra data block and set the study_data field to point to
the reloaded study data. You must also set the PCRE_EXTRA_STUDY_DATA
bit in the flags field to indicate that study data is present. Then
pass the pcre_extra block to pcre_exec() or pcre_dfa_exec() in the
usual way. If the pattern was studied for just-in-time optimization,
that data cannot be saved, and so is lost by a save/restore cycle.
COMPATIBILITY WITH DIFFERENT PCRE RELEASES
In general, it is safest to recompile all saved patterns when you
update to a new PCRE release, though not all updates actually require
this.
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 26 August 2011
Copyright (c) 1997-2011 University of Cambridge.
------------------------------------------------------------------------------
PCREPERFORM(3) PCREPERFORM(3)
NAME
PCRE - Perl-compatible regular expressions
PCRE PERFORMANCE
Two aspects of performance are discussed below: memory usage and pro-
cessing time. The way you express your pattern as a regular expression
can affect both of them.
COMPILED PATTERN MEMORY USAGE
Patterns are compiled by PCRE into a reasonably efficient byte code, so
that most simple patterns do not use much memory. However, there is one
case where the memory usage of a compiled pattern can be unexpectedly
large. If a parenthesized subpattern has a quantifier with a minimum
greater than 1 and/or a limited maximum, the whole subpattern is
repeated in the compiled code. For example, the pattern
(abc|def){2,4}
is compiled as if it were
(abc|def)(abc|def)((abc|def)(abc|def)?)?
(Technical aside: It is done this way so that backtrack points within
each of the repetitions can be independently maintained.)
For regular expressions whose quantifiers use only small numbers, this
is not usually a problem. However, if the numbers are large, and par-
ticularly if such repetitions are nested, the memory usage can become
an embarrassment. For example, the very simple pattern
((ab){1,1000}c){1,3}
uses 51K bytes when compiled. When PCRE is compiled with its default
internal pointer size of two bytes, the size limit on a compiled pat-
tern is 64K, and this is reached with the above pattern if the outer
repetition is increased from 3 to 4. PCRE can be compiled to use larger
internal pointers and thus handle larger compiled patterns, but it is
better to try to rewrite your pattern to use less memory if you can.
One way of reducing the memory usage for such patterns is to make use
of PCRE's "subroutine" facility. Re-writing the above pattern as
((ab)(?2){0,999}c)(?1){0,2}
reduces the memory requirements to 18K, and indeed it remains under 20K
even with the outer repetition increased to 100. However, this pattern
is not exactly equivalent, because the "subroutine" calls are treated
as atomic groups into which there can be no backtracking if there is a
subsequent matching failure. Therefore, PCRE cannot do this kind of
rewriting automatically. Furthermore, there is a noticeable loss of
speed when executing the modified pattern. Nevertheless, if the atomic
grouping is not a problem and the loss of speed is acceptable, this
kind of rewriting will allow you to process patterns that PCRE cannot
otherwise handle.
STACK USAGE AT RUN TIME
When pcre_exec() is used for matching, certain kinds of pattern can
cause it to use large amounts of the process stack. In some environ-
ments the default process stack is quite small, and if it runs out the
result is often SIGSEGV. This issue is probably the most frequently
raised problem with PCRE. Rewriting your pattern can often help. The
pcrestack documentation discusses this issue in detail.
PROCESSING TIME
Certain items in regular expression patterns are processed more effi-
ciently than others. It is more efficient to use a character class like
[aeiou] than a set of single-character alternatives such as
(a|e|i|o|u). In general, the simplest construction that provides the
required behaviour is usually the most efficient. Jeffrey Friedl's book
contains a lot of useful general discussion about optimizing regular
expressions for efficient performance. This document contains a few
observations about PCRE.
Using Unicode character properties (the \p, \P, and \X escapes) is
slow, because PCRE has to scan a structure that contains data for over
fifteen thousand characters whenever it needs a character's property.
If you can find an alternative pattern that does not use character
properties, it will probably be faster.
By default, the escape sequences \b, \d, \s, and \w, and the POSIX
character classes such as [:alpha:] do not use Unicode properties,
partly for backwards compatibility, and partly for performance reasons.
However, you can set PCRE_UCP if you want Unicode character properties
to be used. This can double the matching time for items such as \d,
when matched with pcre_exec(); the performance loss is less with
pcre_dfa_exec(), and in both cases there is not much difference for \b.
When a pattern begins with .* not in parentheses, or in parentheses
that are not the subject of a backreference, and the PCRE_DOTALL option
is set, the pattern is implicitly anchored by PCRE, since it can match
only at the start of a subject string. However, if PCRE_DOTALL is not
set, PCRE cannot make this optimization, because the . metacharacter
does not then match a newline, and if the subject string contains new-
lines, the pattern may match from the character immediately following
one of them instead of from the very start. For example, the pattern
.*second
matches the subject "first\nand second" (where \n stands for a newline
character), with the match starting at the seventh character. In order
to do this, PCRE has to retry the match starting after every newline in
the subject.
If you are using such a pattern with subject strings that do not con-
tain newlines, the best performance is obtained by setting PCRE_DOTALL,
or starting the pattern with ^.* or ^.*? to indicate explicit anchor-
ing. That saves PCRE from having to scan along the subject looking for
a newline to restart at.
Beware of patterns that contain nested indefinite repeats. These can
take a long time to run when applied to a string that does not match.
Consider the pattern fragment
^(a+)*
This can match "aaaa" in 16 different ways, and this number increases
very rapidly as the string gets longer. (The * repeat can match 0, 1,
2, 3, or 4 times, and for each of those cases other than 0 or 4, the +
repeats can match different numbers of times.) When the remainder of
the pattern is such that the entire match is going to fail, PCRE has in
principle to try every possible variation, and this can take an
extremely long time, even for relatively short strings.
An optimization catches some of the more simple cases such as
(a+)*b
where a literal character follows. Before embarking on the standard
matching procedure, PCRE checks that there is a "b" later in the sub-
ject string, and if there is not, it fails the match immediately. How-
ever, when there is no following literal this optimization cannot be
used. You can see the difference by comparing the behaviour of
(a+)*\d
with the pattern above. The former gives a failure almost instantly
when applied to a whole line of "a" characters, whereas the latter
takes an appreciable time with strings longer than about 20 characters.
In many cases, the solution to this kind of performance issue is to use
an atomic group or a possessive quantifier.
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 16 May 2010
Copyright (c) 1997-2010 University of Cambridge.
------------------------------------------------------------------------------
PCREPOSIX(3) PCREPOSIX(3)
NAME
PCRE - Perl-compatible regular expressions.
SYNOPSIS OF POSIX API
#include <pcreposix.h>
int regcomp(regex_t *preg, const char *pattern,
int cflags);
int regexec(regex_t *preg, const char *string,
size_t nmatch, regmatch_t pmatch[], int eflags);
size_t regerror(int errcode, const regex_t *preg,
char *errbuf, size_t errbuf_size);
void regfree(regex_t *preg);
DESCRIPTION
This set of functions provides a POSIX-style API to the PCRE regular
expression package. See the pcreapi documentation for a description of
PCRE's native API, which contains much additional functionality.
The functions described here are just wrapper functions that ultimately
call the PCRE native API. Their prototypes are defined in the
pcreposix.h header file, and on Unix systems the library itself is
called pcreposix.a, so can be accessed by adding -lpcreposix to the
command for linking an application that uses them. Because the POSIX
functions call the native ones, it is also necessary to add -lpcre.
I have implemented only those POSIX option bits that can be reasonably
mapped to PCRE native options. In addition, the option REG_EXTENDED is
defined with the value zero. This has no effect, but since programs
that are written to the POSIX interface often use it, this makes it
easier to slot in PCRE as a replacement library. Other POSIX options
are not even defined.
There are also some other options that are not defined by POSIX. These
have been added at the request of users who want to make use of certain
PCRE-specific features via the POSIX calling interface.
When PCRE is called via these functions, it is only the API that is
POSIX-like in style. The syntax and semantics of the regular expres-
sions themselves are still those of Perl, subject to the setting of
various PCRE options, as described below. "POSIX-like in style" means
that the API approximates to the POSIX definition; it is not fully
POSIX-compatible, and in multi-byte encoding domains it is probably
even less compatible.
The header for these functions is supplied as pcreposix.h to avoid any
potential clash with other POSIX libraries. It can, of course, be
renamed or aliased as regex.h, which is the "correct" name. It provides
two structure types, regex_t for compiled internal forms, and reg-
match_t for returning captured substrings. It also defines some con-
stants whose names start with "REG_"; these are used for setting
options and identifying error codes.
COMPILING A PATTERN
The function regcomp() is called to compile a pattern into an internal
form. The pattern is a C string terminated by a binary zero, and is
passed in the argument pattern. The preg argument is a pointer to a
regex_t structure that is used as a base for storing information about
the compiled regular expression.
The argument cflags is either zero, or contains one or more of the bits
defined by the following macros:
REG_DOTALL
The PCRE_DOTALL option is set when the regular expression is passed for
compilation to the native function. Note that REG_DOTALL is not part of
the POSIX standard.
REG_ICASE
The PCRE_CASELESS option is set when the regular expression is passed
for compilation to the native function.
REG_NEWLINE
The PCRE_MULTILINE option is set when the regular expression is passed
for compilation to the native function. Note that this does not mimic
the defined POSIX behaviour for REG_NEWLINE (see the following sec-
tion).
REG_NOSUB
The PCRE_NO_AUTO_CAPTURE option is set when the regular expression is
passed for compilation to the native function. In addition, when a pat-
tern that is compiled with this flag is passed to regexec() for match-
ing, the nmatch and pmatch arguments are ignored, and no captured
strings are returned.
REG_UCP
The PCRE_UCP option is set when the regular expression is passed for
compilation to the native function. This causes PCRE to use Unicode
properties when matchine \d, \w, etc., instead of just recognizing
ASCII values. Note that REG_UTF8 is not part of the POSIX standard.
REG_UNGREEDY
The PCRE_UNGREEDY option is set when the regular expression is passed
for compilation to the native function. Note that REG_UNGREEDY is not
part of the POSIX standard.
REG_UTF8
The PCRE_UTF8 option is set when the regular expression is passed for
compilation to the native function. This causes the pattern itself and
all data strings used for matching it to be treated as UTF-8 strings.
Note that REG_UTF8 is not part of the POSIX standard.
In the absence of these flags, no options are passed to the native
function. This means the the regex is compiled with PCRE default
semantics. In particular, the way it handles newline characters in the
subject string is the Perl way, not the POSIX way. Note that setting
PCRE_MULTILINE has only some of the effects specified for REG_NEWLINE.
It does not affect the way newlines are matched by . (they are not) or
by a negative class such as [^a] (they are).
The yield of regcomp() is zero on success, and non-zero otherwise. The
preg structure is filled in on success, and one member of the structure
is public: re_nsub contains the number of capturing subpatterns in the
regular expression. Various error codes are defined in the header file.
NOTE: If the yield of regcomp() is non-zero, you must not attempt to
use the contents of the preg structure. If, for example, you pass it to
regexec(), the result is undefined and your program is likely to crash.
MATCHING NEWLINE CHARACTERS
This area is not simple, because POSIX and Perl take different views of
things. It is not possible to get PCRE to obey POSIX semantics, but
then PCRE was never intended to be a POSIX engine. The following table
lists the different possibilities for matching newline characters in
PCRE:
Default Change with
. matches newline no PCRE_DOTALL
newline matches [^a] yes not changeable
$ matches \n at end yes PCRE_DOLLARENDONLY
$ matches \n in middle no PCRE_MULTILINE
^ matches \n in middle no PCRE_MULTILINE
This is the equivalent table for POSIX:
Default Change with
. matches newline yes REG_NEWLINE
newline matches [^a] yes REG_NEWLINE
$ matches \n at end no REG_NEWLINE
$ matches \n in middle no REG_NEWLINE
^ matches \n in middle no REG_NEWLINE
PCRE's behaviour is the same as Perl's, except that there is no equiva-
lent for PCRE_DOLLAR_ENDONLY in Perl. In both PCRE and Perl, there is
no way to stop newline from matching [^a].
The default POSIX newline handling can be obtained by setting
PCRE_DOTALL and PCRE_DOLLAR_ENDONLY, but there is no way to make PCRE
behave exactly as for the REG_NEWLINE action.
MATCHING A PATTERN
The function regexec() is called to match a compiled pattern preg
against a given string, which is by default terminated by a zero byte
(but see REG_STARTEND below), subject to the options in eflags. These
can be:
REG_NOTBOL
The PCRE_NOTBOL option is set when calling the underlying PCRE matching
function.
REG_NOTEMPTY
The PCRE_NOTEMPTY option is set when calling the underlying PCRE match-
ing function. Note that REG_NOTEMPTY is not part of the POSIX standard.
However, setting this option can give more POSIX-like behaviour in some
situations.
REG_NOTEOL
The PCRE_NOTEOL option is set when calling the underlying PCRE matching
function.
REG_STARTEND
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. This is a BSD extension, compatible with but not specified by
IEEE Standard 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.
If the pattern was compiled with the REG_NOSUB flag, no data about any
matched strings is returned. The nmatch and pmatch arguments of
regexec() are ignored.
If the value of nmatch is zero, or if the value pmatch is NULL, no data
about any matched strings is returned.
Otherwise,the portion of the string that was matched, and also any cap-
tured substrings, are returned via the pmatch argument, which points to
an array of nmatch structures of type regmatch_t, containing the mem-
bers rm_so and rm_eo. These contain the offset to the first character
of each substring and the offset to the first character after the end
of each substring, respectively. The 0th element of the vector relates
to the entire portion of string that was matched; subsequent elements
relate to the capturing subpatterns of the regular expression. Unused
entries in the array have both structure members set to -1.
A successful match yields a zero return; various error codes are
defined in the header file, of which REG_NOMATCH is the "expected"
failure code.
ERROR MESSAGES
The regerror() function maps a non-zero errorcode from either regcomp()
or regexec() to a printable message. If preg is not NULL, the error
should have arisen from the use of that structure. A message terminated
by a binary zero is placed in errbuf. The length of the message,
including the zero, is limited to errbuf_size. The yield of the func-
tion is the size of buffer needed to hold the whole message.
MEMORY USAGE
Compiling a regular expression causes memory to be allocated and asso-
ciated with the preg structure. The function regfree() frees all such
memory, after which preg may no longer be used as a compiled expres-
sion.
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 16 May 2010
Copyright (c) 1997-2010 University of Cambridge.
------------------------------------------------------------------------------
PCRECPP(3) PCRECPP(3)
NAME
PCRE - Perl-compatible regular expressions.
SYNOPSIS OF C++ WRAPPER
#include <pcrecpp.h>
DESCRIPTION
The C++ wrapper for PCRE was provided by Google Inc. Some additional
functionality was added by Giuseppe Maxia. This brief man page was con-
structed from the notes in the pcrecpp.h file, which should be con-
sulted for further details.
MATCHING INTERFACE
The "FullMatch" operation checks that supplied text matches a supplied
pattern exactly. If pointer arguments are supplied, it copies matched
sub-strings that match sub-patterns into them.
Example: successful match
pcrecpp::RE re("h.*o");
re.FullMatch("hello");
Example: unsuccessful match (requires full match):
pcrecpp::RE re("e");
!re.FullMatch("hello");
Example: creating a temporary RE object:
pcrecpp::RE("h.*o").FullMatch("hello");
You can pass in a "const char*" or a "string" for "text". The examples
below tend to use a const char*. You can, as in the different examples
above, store the RE object explicitly in a variable or use a temporary
RE object. The examples below use one mode or the other arbitrarily.
Either could correctly be used for any of these examples.
You must supply extra pointer arguments to extract matched subpieces.
Example: extracts "ruby" into "s" and 1234 into "i"
int i;
string s;
pcrecpp::RE re("(\\w+):(\\d+)");
re.FullMatch("ruby:1234", &s, &i);
Example: does not try to extract any extra sub-patterns
re.FullMatch("ruby:1234", &s);
Example: does not try to extract into NULL
re.FullMatch("ruby:1234", NULL, &i);
Example: integer overflow causes failure
!re.FullMatch("ruby:1234567891234", NULL, &i);
Example: fails because there aren't enough sub-patterns:
!pcrecpp::RE("\\w+:\\d+").FullMatch("ruby:1234", &s);
Example: fails because string cannot be stored in integer
!pcrecpp::RE("(.*)").FullMatch("ruby", &i);
The provided pointer arguments can be pointers to any scalar numeric
type, or one of:
string (matched piece is copied to string)
StringPiece (StringPiece is mutated to point to matched piece)
T (where "bool T::ParseFrom(const char*, int)" exists)
NULL (the corresponding matched sub-pattern is not copied)
The function returns true iff all of the following conditions are sat-
isfied:
a. "text" matches "pattern" exactly;
b. The number of matched sub-patterns is >= number of supplied
pointers;
c. The "i"th argument has a suitable type for holding the
string captured as the "i"th sub-pattern. If you pass in
void * NULL for the "i"th argument, or a non-void * NULL
of the correct type, or pass fewer arguments than the
number of sub-patterns, "i"th captured sub-pattern is
ignored.
CAVEAT: An optional sub-pattern that does not exist in the matched
string is assigned the empty string. Therefore, the following will
return false (because the empty string is not a valid number):
int number;
pcrecpp::RE::FullMatch("abc", "[a-z]+(\\d+)?", &number);
The matching interface supports at most 16 arguments per call. If you
need more, consider using the more general interface
pcrecpp::RE::DoMatch. See pcrecpp.h for the signature for DoMatch.
NOTE: Do not use no_arg, which is used internally to mark the end of a
list of optional arguments, as a placeholder for missing arguments, as
this can lead to segfaults.
QUOTING METACHARACTERS
You can use the "QuoteMeta" operation to insert backslashes before all
potentially meaningful characters in a string. The returned string,
used as a regular expression, will exactly match the original string.
Example:
string quoted = RE::QuoteMeta(unquoted);
Note that it's legal to escape a character even if it has no special
meaning in a regular expression -- so this function does that. (This
also makes it identical to the perl function of the same name; see
"perldoc -f quotemeta".) For example, "1.5-2.0?" becomes
"1\.5\-2\.0\?".
PARTIAL MATCHES
You can use the "PartialMatch" operation when you want the pattern to
match any substring of the text.
Example: simple search for a string:
pcrecpp::RE("ell").PartialMatch("hello");
Example: find first number in a string:
int number;
pcrecpp::RE re("(\\d+)");
re.PartialMatch("x*100 + 20", &number);
assert(number == 100);
UTF-8 AND THE MATCHING INTERFACE
By default, pattern and text are plain text, one byte per character.
The UTF8 flag, passed to the constructor, causes both pattern and
string to be treated as UTF-8 text, still a byte stream but potentially
multiple bytes per character. In practice, the text is likelier to be
UTF-8 than the pattern, but the match returned may depend on the UTF8
flag, so always use it when matching UTF8 text. For example, "." will
match one byte normally but with UTF8 set may match up to three bytes
of a multi-byte character.
Example:
pcrecpp::RE_Options options;
options.set_utf8();
pcrecpp::RE re(utf8_pattern, options);
re.FullMatch(utf8_string);
Example: using the convenience function UTF8():
pcrecpp::RE re(utf8_pattern, pcrecpp::UTF8());
re.FullMatch(utf8_string);
NOTE: The UTF8 flag is ignored if pcre was not configured with the
--enable-utf8 flag.
PASSING MODIFIERS TO THE REGULAR EXPRESSION ENGINE
PCRE defines some modifiers to change the behavior of the regular
expression engine. The C++ wrapper defines an auxiliary class,
RE_Options, as a vehicle to pass such modifiers to a RE class. Cur-
rently, the following modifiers are supported:
modifier description Perl corresponding
PCRE_CASELESS case insensitive match /i
PCRE_MULTILINE multiple lines match /m
PCRE_DOTALL dot matches newlines /s
PCRE_DOLLAR_ENDONLY $ matches only at end N/A
PCRE_EXTRA strict escape parsing N/A
PCRE_EXTENDED ignore whitespaces /x
PCRE_UTF8 handles UTF8 chars built-in
PCRE_UNGREEDY reverses * and *? N/A
PCRE_NO_AUTO_CAPTURE disables capturing parens N/A (*)
(*) Both Perl and PCRE allow non capturing parentheses by means of the
"?:" modifier within the pattern itself. e.g. (?:ab|cd) does not cap-
ture, while (ab|cd) does.
For a full account on how each modifier works, please check the PCRE
API reference page.
For each modifier, there are two member functions whose name is made
out of the modifier in lowercase, without the "PCRE_" prefix. For
instance, PCRE_CASELESS is handled by
bool caseless()
which returns true if the modifier is set, and
RE_Options & set_caseless(bool)
which sets or unsets the modifier. Moreover, PCRE_EXTRA_MATCH_LIMIT can
be accessed through the set_match_limit() and match_limit() member
functions. Setting match_limit to a non-zero value will limit the exe-
cution of pcre to keep it from doing bad things like blowing the stack
or taking an eternity to return a result. A value of 5000 is good
enough to stop stack blowup in a 2MB thread stack. Setting match_limit
to zero disables match limiting. Alternatively, you can call
match_limit_recursion() which uses PCRE_EXTRA_MATCH_LIMIT_RECURSION to
limit how much PCRE recurses. match_limit() limits the number of
matches PCRE does; match_limit_recursion() limits the depth of internal
recursion, and therefore the amount of stack that is used.
Normally, to pass one or more modifiers to a RE class, you declare a
RE_Options object, set the appropriate options, and pass this object to
a RE constructor. Example:
RE_Options opt;
opt.set_caseless(true);
if (RE("HELLO", opt).PartialMatch("hello world")) ...
RE_options has two constructors. The default constructor takes no argu-
ments and creates a set of flags that are off by default. The optional
parameter option_flags is to facilitate transfer of legacy code from C
programs. This lets you do
RE(pattern,
RE_Options(PCRE_CASELESS|PCRE_MULTILINE)).PartialMatch(str);
However, new code is better off doing
RE(pattern,
RE_Options().set_caseless(true).set_multiline(true))
.PartialMatch(str);
If you are going to pass one of the most used modifiers, there are some
convenience functions that return a RE_Options class with the appropri-
ate modifier already set: CASELESS(), UTF8(), MULTILINE(), DOTALL(),
and EXTENDED().
If you need to set several options at once, and you don't want to go
through the pains of declaring a RE_Options object and setting several
options, there is a parallel method that give you such ability on the
fly. You can concatenate several set_xxxxx() member functions, since
each of them returns a reference to its class object. For example, to
pass PCRE_CASELESS, PCRE_EXTENDED, and PCRE_MULTILINE to a RE with one
statement, you may write:
RE(" ^ xyz \\s+ .* blah$",
RE_Options()
.set_caseless(true)
.set_extended(true)
.set_multiline(true)).PartialMatch(sometext);
SCANNING TEXT INCREMENTALLY
The "Consume" operation may be useful if you want to repeatedly match
regular expressions at the front of a string and skip over them as they
match. This requires use of the "StringPiece" type, which represents a
sub-range of a real string. Like RE, StringPiece is defined in the
pcrecpp namespace.
Example: read lines of the form "var = value" from a string.
string contents = ...; // Fill string somehow
pcrecpp::StringPiece input(contents); // Wrap in a StringPiece
string var;
int value;
pcrecpp::RE re("(\\w+) = (\\d+)\n");
while (re.Consume(&input, &var, &value)) {
...;
}
Each successful call to "Consume" will set "var/value", and also
advance "input" so it points past the matched text.
The "FindAndConsume" operation is similar to "Consume" but does not
anchor your match at the beginning of the string. For example, you
could extract all words from a string by repeatedly calling
pcrecpp::RE("(\\w+)").FindAndConsume(&input, &word)
PARSING HEX/OCTAL/C-RADIX NUMBERS
By default, if you pass a pointer to a numeric value, the corresponding
text is interpreted as a base-10 number. You can instead wrap the
pointer with a call to one of the operators Hex(), Octal(), or CRadix()
to interpret the text in another base. The CRadix operator interprets
C-style "0" (base-8) and "0x" (base-16) prefixes, but defaults to
base-10.
Example:
int a, b, c, d;
pcrecpp::RE re("(.*) (.*) (.*) (.*)");
re.FullMatch("100 40 0100 0x40",
pcrecpp::Octal(&a), pcrecpp::Hex(&b),
pcrecpp::CRadix(&c), pcrecpp::CRadix(&d));
will leave 64 in a, b, c, and d.
REPLACING PARTS OF STRINGS
You can replace the first match of "pattern" in "str" with "rewrite".
Within "rewrite", backslash-escaped digits (\1 to \9) can be used to
insert text matching corresponding parenthesized group from the pat-
tern. \0 in "rewrite" refers to the entire matching text. For example:
string s = "yabba dabba doo";
pcrecpp::RE("b+").Replace("d", &s);
will leave "s" containing "yada dabba doo". The result is true if the
pattern matches and a replacement occurs, false otherwise.
GlobalReplace is like Replace except that it replaces all occurrences
of the pattern in the string with the rewrite. Replacements are not
subject to re-matching. For example:
string s = "yabba dabba doo";
pcrecpp::RE("b+").GlobalReplace("d", &s);
will leave "s" containing "yada dada doo". It returns the number of
replacements made.
Extract is like Replace, except that if the pattern matches, "rewrite"
is copied into "out" (an additional argument) with substitutions. The
non-matching portions of "text" are ignored. Returns true iff a match
occurred and the extraction happened successfully; if no match occurs,
the string is left unaffected.
AUTHOR
The C++ wrapper was contributed by Google Inc.
Copyright (c) 2007 Google Inc.
REVISION
Last updated: 17 March 2009
Minor typo fixed: 25 July 2011
------------------------------------------------------------------------------
PCRESAMPLE(3) PCRESAMPLE(3)
NAME
PCRE - Perl-compatible regular expressions
PCRE SAMPLE PROGRAM
A simple, complete demonstration program, to get you started with using
PCRE, is supplied in the file pcredemo.c in the PCRE distribution. A
listing of this program is given in the pcredemo documentation. If you
do not have a copy of the PCRE distribution, you can save this listing
to re-create pcredemo.c.
The program compiles the regular expression that is its first argument,
and matches it against the subject string in its second argument. No
PCRE options are set, and default character tables are used. If match-
ing succeeds, the program outputs the portion of the subject that
matched, together with the contents of any captured substrings.
If the -g option is given on the command line, the program then goes on
to check for further matches of the same regular expression in the same
subject string. The logic is a little bit tricky because of the possi-
bility of matching an empty string. Comments in the code explain what
is going on.
If PCRE is installed in the standard include and library directories
for your operating system, you should be able to compile the demonstra-
tion program using this command:
gcc -o pcredemo pcredemo.c -lpcre
If PCRE is installed elsewhere, you may need to add additional options
to the command line. For example, on a Unix-like system that has PCRE
installed in /usr/local, you can compile the demonstration program
using a command like this:
gcc -o pcredemo -I/usr/local/include pcredemo.c \
-L/usr/local/lib -lpcre
In a Windows environment, if you want to statically link the program
against a non-dll pcre.a file, you must uncomment the line that defines
PCRE_STATIC before including pcre.h, because otherwise the pcre_mal-
loc() and pcre_free() exported functions will be declared
__declspec(dllimport), with unwanted results.
Once you have compiled and linked the demonstration program, you can
run simple tests like this:
./pcredemo 'cat|dog' 'the cat sat on the mat'
./pcredemo -g 'cat|dog' 'the dog sat on the cat'
Note that there is a much more comprehensive test program, called
pcretest, which supports many more facilities for testing regular
expressions and the PCRE library. The pcredemo program is provided as a
simple coding example.
If you try to run pcredemo when PCRE is not installed in the standard
library directory, you may get an error like this on some operating
systems (e.g. Solaris):
ld.so.1: a.out: fatal: libpcre.so.0: open failed: No such file or
directory
This is caused by the way shared library support works on those sys-
tems. You need to add
-R/usr/local/lib
(for example) to the compile command to get round this problem.
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 17 November 2010
Copyright (c) 1997-2010 University of Cambridge.
------------------------------------------------------------------------------
PCRELIMITS(3) PCRELIMITS(3)
NAME
PCRE - Perl-compatible regular expressions
SIZE AND OTHER LIMITATIONS
There are some size limitations in PCRE but it is hoped that they will
never in practice be relevant.
The maximum length of a compiled pattern is 65539 (sic) bytes if PCRE
is compiled with the default internal linkage size of 2. If you want to
process regular expressions that are truly enormous, you can compile
PCRE with an internal linkage size of 3 or 4 (see the README file in
the source distribution and the pcrebuild documentation for details).
In these cases the limit is substantially larger. However, the speed
of execution is slower.
All values in repeating quantifiers must be less than 65536.
There is no limit to the number of parenthesized subpatterns, but there
can be no more than 65535 capturing subpatterns.
There is a limit to the number of forward references to subsequent sub-
patterns of around 200,000. Repeated forward references with fixed
upper limits, for example, (?2){0,100} when subpattern number 2 is to
the right, are included in the count. There is no limit to the number
of backward references.
The maximum length of name for a named subpattern is 32 characters, and
the maximum number of named subpatterns is 10000.
The maximum length of a subject string is the largest positive number
that an integer variable can hold. However, when using the traditional
matching function, PCRE uses recursion to handle subpatterns and indef-
inite repetition. This means that the available stack space may limit
the size of a subject string that can be processed by certain patterns.
For a discussion of stack issues, see the pcrestack documentation.
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 30 November 2011
Copyright (c) 1997-2011 University of Cambridge.
------------------------------------------------------------------------------
PCRESTACK(3) PCRESTACK(3)
NAME
PCRE - Perl-compatible regular expressions
PCRE DISCUSSION OF STACK USAGE
When you call pcre_exec(), it makes use of an internal function called
match(). This calls itself recursively at branch points in the pattern,
in order to remember the state of the match so that it can back up and
try a different alternative if the first one fails. As matching pro-
ceeds deeper and deeper into the tree of possibilities, the recursion
depth increases. The match() function is also called in other circum-
stances, for example, whenever a parenthesized sub-pattern is entered,
and in certain cases of repetition.
Not all calls of match() increase the recursion depth; for an item such
as a* it may be called several times at the same level, after matching
different numbers of a's. Furthermore, in a number of cases where the
result of the recursive call would immediately be passed back as the
result of the current call (a "tail recursion"), the function is just
restarted instead.
The above comments apply when pcre_exec() is run in its normal inter-
pretive manner. If the pattern was studied with the PCRE_STUDY_JIT_COM-
PILE option, and just-in-time compiling was successful, and the options
passed to pcre_exec() were not incompatible, the matching process uses
the JIT-compiled code instead of the match() function. In this case,
the memory requirements are handled entirely differently. See the pcre-
jit documentation for details.
The pcre_dfa_exec() function operates in an entirely different way, and
uses recursion only when there is a regular expression recursion or
subroutine call in the pattern. This includes the processing of asser-
tion and "once-only" subpatterns, which are handled like subroutine
calls. Normally, these are never very deep, and the limit on the com-
plexity of pcre_dfa_exec() is controlled by the amount of workspace it
is given. However, it is possible to write patterns with runaway infi-
nite recursions; such patterns will cause pcre_dfa_exec() to run out of
stack. At present, there is no protection against this.
The comments that follow do NOT apply to pcre_dfa_exec(); they are rel-
evant only for pcre_exec() without the JIT optimization.
Reducing pcre_exec()'s stack usage
Each time that match() is actually called recursively, it uses memory
from the process stack. For certain kinds of pattern and data, very
large amounts of stack may be needed, despite the recognition of "tail
recursion". You can often reduce the amount of recursion, and there-
fore the amount of stack used, by modifying the pattern that is being
matched. Consider, for example, this pattern:
([^<]|<(?!inet))+
It matches from wherever it starts until it encounters "<inet" or the
end of the data, and is the kind of pattern that might be used when
processing an XML file. Each iteration of the outer parentheses matches
either one character that is not "<" or a "<" that is not followed by
"inet". However, each time a parenthesis is processed, a recursion
occurs, so this formulation uses a stack frame for each matched charac-
ter. For a long string, a lot of stack is required. Consider now this
rewritten pattern, which matches exactly the same strings:
([^<]++|<(?!inet))+
This uses very much less stack, because runs of characters that do not
contain "<" are "swallowed" in one item inside the parentheses. Recur-
sion happens only when a "<" character that is not followed by "inet"
is encountered (and we assume this is relatively rare). A possessive
quantifier is used to stop any backtracking into the runs of non-"<"
characters, but that is not related to stack usage.
This example shows that one way of avoiding stack problems when match-
ing long subject strings is to write repeated parenthesized subpatterns
to match more than one character whenever possible.
Compiling PCRE to use heap instead of stack for pcre_exec()
In environments where stack memory is constrained, you might want to
compile PCRE to use heap memory instead of stack for remembering back-
up points when pcre_exec() is running. This makes it run a lot more
slowly, however. Details of how to do this are given in the pcrebuild
documentation. When built in this way, instead of using the stack, PCRE
obtains and frees memory by calling the functions that are pointed to
by the pcre_stack_malloc and pcre_stack_free variables. By default,
these point to malloc() and free(), but you can replace the pointers to
cause PCRE to use your own functions. Since the block sizes are always
the same, and are always freed in reverse order, it may be possible to
implement customized memory handlers that are more efficient than the
standard functions.
Limiting pcre_exec()'s stack usage
You can set limits on the number of times that match() is called, both
in total and recursively. If a limit is exceeded, pcre_exec() returns
an error code. Setting suitable limits should prevent it from running
out of stack. The default values of the limits are very large, and
unlikely ever to operate. They can be changed when PCRE is built, and
they can also be set when pcre_exec() is called. For details of these
interfaces, see the pcrebuild documentation and the section on extra
data for pcre_exec() in the pcreapi documentation.
As a very rough rule of thumb, you should reckon on about 500 bytes per
recursion. Thus, if you want to limit your stack usage to 8Mb, you
should set the limit at 16000 recursions. A 64Mb stack, on the other
hand, can support around 128000 recursions.
In Unix-like environments, the pcretest test program has a command line
option (-S) that can be used to increase the size of its stack. As long
as the stack is large enough, another option (-M) can be used to find
the smallest limits that allow a particular pattern to match a given
subject string. This is done by calling pcre_exec() repeatedly with
different limits.
Changing stack size in Unix-like systems
In Unix-like environments, there is not often a problem with the stack
unless very long strings are involved, though the default limit on
stack size varies from system to system. Values from 8Mb to 64Mb are
common. You can find your default limit by running the command:
ulimit -s
Unfortunately, the effect of running out of stack is often SIGSEGV,
though sometimes a more explicit error message is given. You can nor-
mally increase the limit on stack size by code such as this:
struct rlimit rlim;
getrlimit(RLIMIT_STACK, &rlim);
rlim.rlim_cur = 100*1024*1024;
setrlimit(RLIMIT_STACK, &rlim);
This reads the current limits (soft and hard) using getrlimit(), then
attempts to increase the soft limit to 100Mb using setrlimit(). You
must do this before calling pcre_exec().
Changing stack size in Mac OS X
Using setrlimit(), as described above, should also work on Mac OS X. It
is also possible to set a stack size when linking a program. There is a
discussion about stack sizes in Mac OS X at this web site:
http://developer.apple.com/qa/qa2005/qa1419.html.
AUTHOR
Philip Hazel
University Computing Service
Cambridge CB2 3QH, England.
REVISION
Last updated: 26 August 2011
Copyright (c) 1997-2011 University of Cambridge.
------------------------------------------------------------------------------