/[pcre]/code/trunk/doc/pcreapi.3
ViewVC logotype

Contents of /code/trunk/doc/pcreapi.3

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1392 - (show annotations)
Wed Nov 6 18:00:09 2013 UTC (7 years ago) by ph10
File size: 127087 byte(s)
Give errors for [A-\d] and [a-[:digit:]] etc.
1 .TH PCREAPI 3 "05 November 2013" "PCRE 8.34"
2 .SH NAME
3 PCRE - Perl-compatible regular expressions
4 .sp
5 .B #include <pcre.h>
6 .
7 .
8 .SH "PCRE NATIVE API BASIC FUNCTIONS"
9 .rs
10 .sp
11 .nf
12 .B pcre *pcre_compile(const char *\fIpattern\fP, int \fIoptions\fP,
13 .B " const char **\fIerrptr\fP, int *\fIerroffset\fP,"
14 .B " const unsigned char *\fItableptr\fP);"
15 .sp
16 .B pcre *pcre_compile2(const char *\fIpattern\fP, int \fIoptions\fP,
17 .B " int *\fIerrorcodeptr\fP,"
18 .B " const char **\fIerrptr\fP, int *\fIerroffset\fP,"
19 .B " const unsigned char *\fItableptr\fP);"
20 .sp
21 .B pcre_extra *pcre_study(const pcre *\fIcode\fP, int \fIoptions\fP,
22 .B " const char **\fIerrptr\fP);"
23 .sp
24 .B void pcre_free_study(pcre_extra *\fIextra\fP);
25 .sp
26 .B int pcre_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
27 .B " const char *\fIsubject\fP, int \fIlength\fP, int \fIstartoffset\fP,"
28 .B " int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP);"
29 .sp
30 .B int pcre_dfa_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
31 .B " const char *\fIsubject\fP, int \fIlength\fP, int \fIstartoffset\fP,"
32 .B " int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP,"
33 .B " int *\fIworkspace\fP, int \fIwscount\fP);"
34 .fi
35 .
36 .
37 .SH "PCRE NATIVE API STRING EXTRACTION FUNCTIONS"
38 .rs
39 .sp
40 .nf
41 .B int pcre_copy_named_substring(const pcre *\fIcode\fP,
42 .B " const char *\fIsubject\fP, int *\fIovector\fP,"
43 .B " int \fIstringcount\fP, const char *\fIstringname\fP,"
44 .B " char *\fIbuffer\fP, int \fIbuffersize\fP);"
45 .sp
46 .B int pcre_copy_substring(const char *\fIsubject\fP, int *\fIovector\fP,
47 .B " int \fIstringcount\fP, int \fIstringnumber\fP, char *\fIbuffer\fP,"
48 .B " int \fIbuffersize\fP);"
49 .sp
50 .B int pcre_get_named_substring(const pcre *\fIcode\fP,
51 .B " const char *\fIsubject\fP, int *\fIovector\fP,"
52 .B " int \fIstringcount\fP, const char *\fIstringname\fP,"
53 .B " const char **\fIstringptr\fP);"
54 .sp
55 .B int pcre_get_stringnumber(const pcre *\fIcode\fP,
56 .B " const char *\fIname\fP);"
57 .sp
58 .B int pcre_get_stringtable_entries(const pcre *\fIcode\fP,
59 .B " const char *\fIname\fP, char **\fIfirst\fP, char **\fIlast\fP);"
60 .sp
61 .B int pcre_get_substring(const char *\fIsubject\fP, int *\fIovector\fP,
62 .B " int \fIstringcount\fP, int \fIstringnumber\fP,"
63 .B " const char **\fIstringptr\fP);"
64 .sp
65 .B int pcre_get_substring_list(const char *\fIsubject\fP,
66 .B " int *\fIovector\fP, int \fIstringcount\fP, const char ***\fIlistptr\fP);"
67 .sp
68 .B void pcre_free_substring(const char *\fIstringptr\fP);
69 .sp
70 .B void pcre_free_substring_list(const char **\fIstringptr\fP);
71 .fi
72 .
73 .
74 .SH "PCRE NATIVE API AUXILIARY FUNCTIONS"
75 .rs
76 .sp
77 .nf
78 .B int pcre_jit_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
79 .B " const char *\fIsubject\fP, int \fIlength\fP, int \fIstartoffset\fP,"
80 .B " int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP,"
81 .B " pcre_jit_stack *\fIjstack\fP);"
82 .sp
83 .B pcre_jit_stack *pcre_jit_stack_alloc(int \fIstartsize\fP, int \fImaxsize\fP);
84 .sp
85 .B void pcre_jit_stack_free(pcre_jit_stack *\fIstack\fP);
86 .sp
87 .B void pcre_assign_jit_stack(pcre_extra *\fIextra\fP,
88 .B " pcre_jit_callback \fIcallback\fP, void *\fIdata\fP);"
89 .sp
90 .B const unsigned char *pcre_maketables(void);
91 .sp
92 .B int pcre_fullinfo(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
93 .B " int \fIwhat\fP, void *\fIwhere\fP);"
94 .sp
95 .B int pcre_refcount(pcre *\fIcode\fP, int \fIadjust\fP);
96 .sp
97 .B int pcre_config(int \fIwhat\fP, void *\fIwhere\fP);
98 .sp
99 .B const char *pcre_version(void);
100 .sp
101 .B int pcre_pattern_to_host_byte_order(pcre *\fIcode\fP,
102 .B " pcre_extra *\fIextra\fP, const unsigned char *\fItables\fP);"
103 .fi
104 .
105 .
106 .SH "PCRE NATIVE API INDIRECTED FUNCTIONS"
107 .rs
108 .sp
109 .nf
110 .B void *(*pcre_malloc)(size_t);
111 .sp
112 .B void (*pcre_free)(void *);
113 .sp
114 .B void *(*pcre_stack_malloc)(size_t);
115 .sp
116 .B void (*pcre_stack_free)(void *);
117 .sp
118 .B int (*pcre_callout)(pcre_callout_block *);
119 .fi
120 .
121 .
122 .SH "PCRE 8-BIT, 16-BIT, AND 32-BIT LIBRARIES"
123 .rs
124 .sp
125 As well as support for 8-bit character strings, PCRE also supports 16-bit
126 strings (from release 8.30) and 32-bit strings (from release 8.32), by means of
127 two additional libraries. They can be built as well as, or instead of, the
128 8-bit library. To avoid too much complication, this document describes the
129 8-bit versions of the functions, with only occasional references to the 16-bit
130 and 32-bit libraries.
131 .P
132 The 16-bit and 32-bit functions operate in the same way as their 8-bit
133 counterparts; they just use different data types for their arguments and
134 results, and their names start with \fBpcre16_\fP or \fBpcre32_\fP instead of
135 \fBpcre_\fP. For every option that has UTF8 in its name (for example,
136 PCRE_UTF8), there are corresponding 16-bit and 32-bit names with UTF8 replaced
137 by UTF16 or UTF32, respectively. This facility is in fact just cosmetic; the
138 16-bit and 32-bit option names define the same bit values.
139 .P
140 References to bytes and UTF-8 in this document should be read as references to
141 16-bit data units and UTF-16 when using the 16-bit library, or 32-bit data
142 units and UTF-32 when using the 32-bit library, unless specified otherwise.
143 More details of the specific differences for the 16-bit and 32-bit libraries
144 are given in the
145 .\" HREF
146 \fBpcre16\fP
147 .\"
148 and
149 .\" HREF
150 \fBpcre32\fP
151 .\"
152 pages.
153 .
154 .
155 .SH "PCRE API OVERVIEW"
156 .rs
157 .sp
158 PCRE has its own native API, which is described in this document. There are
159 also some wrapper functions (for the 8-bit library only) that correspond to the
160 POSIX regular expression API, but they do not give access to all the
161 functionality. They are described in the
162 .\" HREF
163 \fBpcreposix\fP
164 .\"
165 documentation. Both of these APIs define a set of C function calls. A C++
166 wrapper (again for the 8-bit library only) is also distributed with PCRE. It is
167 documented in the
168 .\" HREF
169 \fBpcrecpp\fP
170 .\"
171 page.
172 .P
173 The native API C function prototypes are defined in the header file
174 \fBpcre.h\fP, and on Unix-like systems the (8-bit) library itself is called
175 \fBlibpcre\fP. It can normally be accessed by adding \fB-lpcre\fP to the
176 command for linking an application that uses PCRE. The header file defines the
177 macros PCRE_MAJOR and PCRE_MINOR to contain the major and minor release numbers
178 for the library. Applications can use these to include support for different
179 releases of PCRE.
180 .P
181 In a Windows environment, if you want to statically link an application program
182 against a non-dll \fBpcre.a\fP file, you must define PCRE_STATIC before
183 including \fBpcre.h\fP or \fBpcrecpp.h\fP, because otherwise the
184 \fBpcre_malloc()\fP and \fBpcre_free()\fP exported functions will be declared
185 \fB__declspec(dllimport)\fP, with unwanted results.
186 .P
187 The functions \fBpcre_compile()\fP, \fBpcre_compile2()\fP, \fBpcre_study()\fP,
188 and \fBpcre_exec()\fP are used for compiling and matching regular expressions
189 in a Perl-compatible manner. A sample program that demonstrates the simplest
190 way of using them is provided in the file called \fIpcredemo.c\fP in the PCRE
191 source distribution. A listing of this program is given in the
192 .\" HREF
193 \fBpcredemo\fP
194 .\"
195 documentation, and the
196 .\" HREF
197 \fBpcresample\fP
198 .\"
199 documentation describes how to compile and run it.
200 .P
201 Just-in-time compiler support is an optional feature of PCRE that can be built
202 in appropriate hardware environments. It greatly speeds up the matching
203 performance of many patterns. Simple programs can easily request that it be
204 used if available, by setting an option that is ignored when it is not
205 relevant. More complicated programs might need to make use of the functions
206 \fBpcre_jit_stack_alloc()\fP, \fBpcre_jit_stack_free()\fP, and
207 \fBpcre_assign_jit_stack()\fP in order to control the JIT code's memory usage.
208 .P
209 From release 8.32 there is also a direct interface for JIT execution, which
210 gives improved performance. The JIT-specific functions are discussed in the
211 .\" HREF
212 \fBpcrejit\fP
213 .\"
214 documentation.
215 .P
216 A second matching function, \fBpcre_dfa_exec()\fP, which is not
217 Perl-compatible, is also provided. This uses a different algorithm for the
218 matching. The alternative algorithm finds all possible matches (at a given
219 point in the subject), and scans the subject just once (unless there are
220 lookbehind assertions). However, this algorithm does not return captured
221 substrings. A description of the two matching algorithms and their advantages
222 and disadvantages is given in the
223 .\" HREF
224 \fBpcrematching\fP
225 .\"
226 documentation.
227 .P
228 In addition to the main compiling and matching functions, there are convenience
229 functions for extracting captured substrings from a subject string that is
230 matched by \fBpcre_exec()\fP. They are:
231 .sp
232 \fBpcre_copy_substring()\fP
233 \fBpcre_copy_named_substring()\fP
234 \fBpcre_get_substring()\fP
235 \fBpcre_get_named_substring()\fP
236 \fBpcre_get_substring_list()\fP
237 \fBpcre_get_stringnumber()\fP
238 \fBpcre_get_stringtable_entries()\fP
239 .sp
240 \fBpcre_free_substring()\fP and \fBpcre_free_substring_list()\fP are also
241 provided, to free the memory used for extracted strings.
242 .P
243 The function \fBpcre_maketables()\fP is used to build a set of character tables
244 in the current locale for passing to \fBpcre_compile()\fP, \fBpcre_exec()\fP,
245 or \fBpcre_dfa_exec()\fP. This is an optional facility that is provided for
246 specialist use. Most commonly, no special tables are passed, in which case
247 internal tables that are generated when PCRE is built are used.
248 .P
249 The function \fBpcre_fullinfo()\fP is used to find out information about a
250 compiled pattern. The function \fBpcre_version()\fP returns a pointer to a
251 string containing the version of PCRE and its date of release.
252 .P
253 The function \fBpcre_refcount()\fP maintains a reference count in a data block
254 containing a compiled pattern. This is provided for the benefit of
255 object-oriented applications.
256 .P
257 The global variables \fBpcre_malloc\fP and \fBpcre_free\fP initially contain
258 the entry points of the standard \fBmalloc()\fP and \fBfree()\fP functions,
259 respectively. PCRE calls the memory management functions via these variables,
260 so a calling program can replace them if it wishes to intercept the calls. This
261 should be done before calling any PCRE functions.
262 .P
263 The global variables \fBpcre_stack_malloc\fP and \fBpcre_stack_free\fP are also
264 indirections to memory management functions. These special functions are used
265 only when PCRE is compiled to use the heap for remembering data, instead of
266 recursive function calls, when running the \fBpcre_exec()\fP function. See the
267 .\" HREF
268 \fBpcrebuild\fP
269 .\"
270 documentation for details of how to do this. It is a non-standard way of
271 building PCRE, for use in environments that have limited stacks. Because of the
272 greater use of memory management, it runs more slowly. Separate functions are
273 provided so that special-purpose external code can be used for this case. When
274 used, these functions are always called in a stack-like manner (last obtained,
275 first freed), and always for memory blocks of the same size. There is a
276 discussion about PCRE's stack usage in the
277 .\" HREF
278 \fBpcrestack\fP
279 .\"
280 documentation.
281 .P
282 The global variable \fBpcre_callout\fP initially contains NULL. It can be set
283 by the caller to a "callout" function, which PCRE will then call at specified
284 points during a matching operation. Details are given in the
285 .\" HREF
286 \fBpcrecallout\fP
287 .\"
288 documentation.
289 .
290 .
291 .\" HTML <a name="newlines"></a>
292 .SH NEWLINES
293 .rs
294 .sp
295 PCRE supports five different conventions for indicating line breaks in
296 strings: a single CR (carriage return) character, a single LF (linefeed)
297 character, the two-character sequence CRLF, any of the three preceding, or any
298 Unicode newline sequence. The Unicode newline sequences are the three just
299 mentioned, plus the single characters VT (vertical tab, U+000B), FF (form feed,
300 U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
301 (paragraph separator, U+2029).
302 .P
303 Each of the first three conventions is used by at least one operating system as
304 its standard newline sequence. When PCRE is built, a default can be specified.
305 The default default is LF, which is the Unix standard. When PCRE is run, the
306 default can be overridden, either when a pattern is compiled, or when it is
307 matched.
308 .P
309 At compile time, the newline convention can be specified by the \fIoptions\fP
310 argument of \fBpcre_compile()\fP, or it can be specified by special text at the
311 start of the pattern itself; this overrides any other settings. See the
312 .\" HREF
313 \fBpcrepattern\fP
314 .\"
315 page for details of the special character sequences.
316 .P
317 In the PCRE documentation the word "newline" is used to mean "the character or
318 pair of characters that indicate a line break". The choice of newline
319 convention affects the handling of the dot, circumflex, and dollar
320 metacharacters, the handling of #-comments in /x mode, and, when CRLF is a
321 recognized line ending sequence, the match position advancement for a
322 non-anchored pattern. There is more detail about this in the
323 .\" HTML <a href="#execoptions">
324 .\" </a>
325 section on \fBpcre_exec()\fP options
326 .\"
327 below.
328 .P
329 The choice of newline convention does not affect the interpretation of
330 the \en or \er escape sequences, nor does it affect what \eR matches, which is
331 controlled in a similar way, but by separate options.
332 .
333 .
334 .SH MULTITHREADING
335 .rs
336 .sp
337 The PCRE functions can be used in multi-threading applications, with the
338 proviso that the memory management functions pointed to by \fBpcre_malloc\fP,
339 \fBpcre_free\fP, \fBpcre_stack_malloc\fP, and \fBpcre_stack_free\fP, and the
340 callout function pointed to by \fBpcre_callout\fP, are shared by all threads.
341 .P
342 The compiled form of a regular expression is not altered during matching, so
343 the same compiled pattern can safely be used by several threads at once.
344 .P
345 If the just-in-time optimization feature is being used, it needs separate
346 memory stack areas for each thread. See the
347 .\" HREF
348 \fBpcrejit\fP
349 .\"
350 documentation for more details.
351 .
352 .
353 .SH "SAVING PRECOMPILED PATTERNS FOR LATER USE"
354 .rs
355 .sp
356 The compiled form of a regular expression can be saved and re-used at a later
357 time, possibly by a different program, and even on a host other than the one on
358 which it was compiled. Details are given in the
359 .\" HREF
360 \fBpcreprecompile\fP
361 .\"
362 documentation, which includes a description of the
363 \fBpcre_pattern_to_host_byte_order()\fP function. However, compiling a regular
364 expression with one version of PCRE for use with a different version is not
365 guaranteed to work and may cause crashes.
366 .
367 .
368 .SH "CHECKING BUILD-TIME OPTIONS"
369 .rs
370 .sp
371 .B int pcre_config(int \fIwhat\fP, void *\fIwhere\fP);
372 .PP
373 The function \fBpcre_config()\fP makes it possible for a PCRE client to
374 discover which optional features have been compiled into the PCRE library. The
375 .\" HREF
376 \fBpcrebuild\fP
377 .\"
378 documentation has more details about these optional features.
379 .P
380 The first argument for \fBpcre_config()\fP is an integer, specifying which
381 information is required; the second argument is a pointer to a variable into
382 which the information is placed. The returned value is zero on success, or the
383 negative error code PCRE_ERROR_BADOPTION if the value in the first argument is
384 not recognized. The following information is available:
385 .sp
386 PCRE_CONFIG_UTF8
387 .sp
388 The output is an integer that is set to one if UTF-8 support is available;
389 otherwise it is set to zero. This value should normally be given to the 8-bit
390 version of this function, \fBpcre_config()\fP. If it is given to the 16-bit
391 or 32-bit version of this function, the result is PCRE_ERROR_BADOPTION.
392 .sp
393 PCRE_CONFIG_UTF16
394 .sp
395 The output is an integer that is set to one if UTF-16 support is available;
396 otherwise it is set to zero. This value should normally be given to the 16-bit
397 version of this function, \fBpcre16_config()\fP. If it is given to the 8-bit
398 or 32-bit version of this function, the result is PCRE_ERROR_BADOPTION.
399 .sp
400 PCRE_CONFIG_UTF32
401 .sp
402 The output is an integer that is set to one if UTF-32 support is available;
403 otherwise it is set to zero. This value should normally be given to the 32-bit
404 version of this function, \fBpcre32_config()\fP. If it is given to the 8-bit
405 or 16-bit version of this function, the result is PCRE_ERROR_BADOPTION.
406 .sp
407 PCRE_CONFIG_UNICODE_PROPERTIES
408 .sp
409 The output is an integer that is set to one if support for Unicode character
410 properties is available; otherwise it is set to zero.
411 .sp
412 PCRE_CONFIG_JIT
413 .sp
414 The output is an integer that is set to one if support for just-in-time
415 compiling is available; otherwise it is set to zero.
416 .sp
417 PCRE_CONFIG_JITTARGET
418 .sp
419 The output is a pointer to a zero-terminated "const char *" string. If JIT
420 support is available, the string contains the name of the architecture for
421 which the JIT compiler is configured, for example "x86 32bit (little endian +
422 unaligned)". If JIT support is not available, the result is NULL.
423 .sp
424 PCRE_CONFIG_NEWLINE
425 .sp
426 The output is an integer whose value specifies the default character sequence
427 that is recognized as meaning "newline". The values that are supported in
428 ASCII/Unicode environments are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for
429 ANYCRLF, and -1 for ANY. In EBCDIC environments, CR, ANYCRLF, and ANY yield the
430 same values. However, the value for LF is normally 21, though some EBCDIC
431 environments use 37. The corresponding values for CRLF are 3349 and 3365. The
432 default should normally correspond to the standard sequence for your operating
433 system.
434 .sp
435 PCRE_CONFIG_BSR
436 .sp
437 The output is an integer whose value indicates what character sequences the \eR
438 escape sequence matches by default. A value of 0 means that \eR matches any
439 Unicode line ending sequence; a value of 1 means that \eR matches only CR, LF,
440 or CRLF. The default can be overridden when a pattern is compiled or matched.
441 .sp
442 PCRE_CONFIG_LINK_SIZE
443 .sp
444 The output is an integer that contains the number of bytes used for internal
445 linkage in compiled regular expressions. For the 8-bit library, the value can
446 be 2, 3, or 4. For the 16-bit library, the value is either 2 or 4 and is still
447 a number of bytes. For the 32-bit library, the value is either 2 or 4 and is
448 still a number of bytes. The default value of 2 is sufficient for all but the
449 most massive patterns, since it allows the compiled pattern to be up to 64K in
450 size. Larger values allow larger regular expressions to be compiled, at the
451 expense of slower matching.
452 .sp
453 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
454 .sp
455 The output is an integer that contains the threshold above which the POSIX
456 interface uses \fBmalloc()\fP for output vectors. Further details are given in
457 the
458 .\" HREF
459 \fBpcreposix\fP
460 .\"
461 documentation.
462 .sp
463 PCRE_CONFIG_PARENS_LIMIT
464 .sp
465 The output is a long integer that gives the maximum depth of nesting of
466 parentheses (of any kind) in a pattern. This limit is imposed to cap the amount
467 of system stack used when a pattern is compiled. It is specified when PCRE is
468 built; the default is 250.
469 .sp
470 PCRE_CONFIG_MATCH_LIMIT
471 .sp
472 The output is a long integer that gives the default limit for the number of
473 internal matching function calls in a \fBpcre_exec()\fP execution. Further
474 details are given with \fBpcre_exec()\fP below.
475 .sp
476 PCRE_CONFIG_MATCH_LIMIT_RECURSION
477 .sp
478 The output is a long integer that gives the default limit for the depth of
479 recursion when calling the internal matching function in a \fBpcre_exec()\fP
480 execution. Further details are given with \fBpcre_exec()\fP below.
481 .sp
482 PCRE_CONFIG_STACKRECURSE
483 .sp
484 The output is an integer that is set to one if internal recursion when running
485 \fBpcre_exec()\fP is implemented by recursive function calls that use the stack
486 to remember their state. This is the usual way that PCRE is compiled. The
487 output is zero if PCRE was compiled to use blocks of data on the heap instead
488 of recursive function calls. In this case, \fBpcre_stack_malloc\fP and
489 \fBpcre_stack_free\fP are called to manage memory blocks on the heap, thus
490 avoiding the use of the stack.
491 .
492 .
493 .SH "COMPILING A PATTERN"
494 .rs
495 .sp
496 .nf
497 .B pcre *pcre_compile(const char *\fIpattern\fP, int \fIoptions\fP,
498 .B " const char **\fIerrptr\fP, int *\fIerroffset\fP,"
499 .B " const unsigned char *\fItableptr\fP);"
500 .sp
501 .B pcre *pcre_compile2(const char *\fIpattern\fP, int \fIoptions\fP,
502 .B " int *\fIerrorcodeptr\fP,"
503 .B " const char **\fIerrptr\fP, int *\fIerroffset\fP,"
504 .B " const unsigned char *\fItableptr\fP);"
505 .fi
506 .P
507 Either of the functions \fBpcre_compile()\fP or \fBpcre_compile2()\fP can be
508 called to compile a pattern into an internal form. The only difference between
509 the two interfaces is that \fBpcre_compile2()\fP has an additional argument,
510 \fIerrorcodeptr\fP, via which a numerical error code can be returned. To avoid
511 too much repetition, we refer just to \fBpcre_compile()\fP below, but the
512 information applies equally to \fBpcre_compile2()\fP.
513 .P
514 The pattern is a C string terminated by a binary zero, and is passed in the
515 \fIpattern\fP argument. A pointer to a single block of memory that is obtained
516 via \fBpcre_malloc\fP is returned. This contains the compiled code and related
517 data. The \fBpcre\fP type is defined for the returned block; this is a typedef
518 for a structure whose contents are not externally defined. It is up to the
519 caller to free the memory (via \fBpcre_free\fP) when it is no longer required.
520 .P
521 Although the compiled code of a PCRE regex is relocatable, that is, it does not
522 depend on memory location, the complete \fBpcre\fP data block is not
523 fully relocatable, because it may contain a copy of the \fItableptr\fP
524 argument, which is an address (see below).
525 .P
526 The \fIoptions\fP argument contains various bit settings that affect the
527 compilation. It should be zero if no options are required. The available
528 options are described below. Some of them (in particular, those that are
529 compatible with Perl, but some others as well) can also be set and unset from
530 within the pattern (see the detailed description in the
531 .\" HREF
532 \fBpcrepattern\fP
533 .\"
534 documentation). For those options that can be different in different parts of
535 the pattern, the contents of the \fIoptions\fP argument specifies their
536 settings at the start of compilation and execution. The PCRE_ANCHORED,
537 PCRE_BSR_\fIxxx\fP, PCRE_NEWLINE_\fIxxx\fP, PCRE_NO_UTF8_CHECK, and
538 PCRE_NO_START_OPTIMIZE options can be set at the time of matching as well as at
539 compile time.
540 .P
541 If \fIerrptr\fP is NULL, \fBpcre_compile()\fP returns NULL immediately.
542 Otherwise, if compilation of a pattern fails, \fBpcre_compile()\fP returns
543 NULL, and sets the variable pointed to by \fIerrptr\fP to point to a textual
544 error message. This is a static string that is part of the library. You must
545 not try to free it. Normally, the offset from the start of the pattern to the
546 data unit that was being processed when the error was discovered is placed in
547 the variable pointed to by \fIerroffset\fP, which must not be NULL (if it is,
548 an immediate error is given). However, for an invalid UTF-8 or UTF-16 string,
549 the offset is that of the first data unit of the failing character.
550 .P
551 Some errors are not detected until the whole pattern has been scanned; in these
552 cases, the offset passed back is the length of the pattern. Note that the
553 offset is in data units, not characters, even in a UTF mode. It may sometimes
554 point into the middle of a UTF-8 or UTF-16 character.
555 .P
556 If \fBpcre_compile2()\fP is used instead of \fBpcre_compile()\fP, and the
557 \fIerrorcodeptr\fP argument is not NULL, a non-zero error code number is
558 returned via this argument in the event of an error. This is in addition to the
559 textual error message. Error codes and messages are listed below.
560 .P
561 If the final argument, \fItableptr\fP, is NULL, PCRE uses a default set of
562 character tables that are built when PCRE is compiled, using the default C
563 locale. Otherwise, \fItableptr\fP must be an address that is the result of a
564 call to \fBpcre_maketables()\fP. This value is stored with the compiled
565 pattern, and used again by \fBpcre_exec()\fP, unless another table pointer is
566 passed to it. For more discussion, see the section on locale support below.
567 .P
568 This code fragment shows a typical straightforward call to \fBpcre_compile()\fP:
569 .sp
570 pcre *re;
571 const char *error;
572 int erroffset;
573 re = pcre_compile(
574 "^A.*Z", /* the pattern */
575 0, /* default options */
576 &error, /* for error message */
577 &erroffset, /* for error offset */
578 NULL); /* use default character tables */
579 .sp
580 The following names for option bits are defined in the \fBpcre.h\fP header
581 file:
582 .sp
583 PCRE_ANCHORED
584 .sp
585 If this bit is set, the pattern is forced to be "anchored", that is, it is
586 constrained to match only at the first matching point in the string that is
587 being searched (the "subject string"). This effect can also be achieved by
588 appropriate constructs in the pattern itself, which is the only way to do it in
589 Perl.
590 .sp
591 PCRE_AUTO_CALLOUT
592 .sp
593 If this bit is set, \fBpcre_compile()\fP automatically inserts callout items,
594 all with number 255, before each pattern item. For discussion of the callout
595 facility, see the
596 .\" HREF
597 \fBpcrecallout\fP
598 .\"
599 documentation.
600 .sp
601 PCRE_BSR_ANYCRLF
602 PCRE_BSR_UNICODE
603 .sp
604 These options (which are mutually exclusive) control what the \eR escape
605 sequence matches. The choice is either to match only CR, LF, or CRLF, or to
606 match any Unicode newline sequence. The default is specified when PCRE is
607 built. It can be overridden from within the pattern, or by setting an option
608 when a compiled pattern is matched.
609 .sp
610 PCRE_CASELESS
611 .sp
612 If this bit is set, letters in the pattern match both upper and lower case
613 letters. It is equivalent to Perl's /i option, and it can be changed within a
614 pattern by a (?i) option setting. In UTF-8 mode, PCRE always understands the
615 concept of case for characters whose values are less than 128, so caseless
616 matching is always possible. For characters with higher values, the concept of
617 case is supported if PCRE is compiled with Unicode property support, but not
618 otherwise. If you want to use caseless matching for characters 128 and above,
619 you must ensure that PCRE is compiled with Unicode property support as well as
620 with UTF-8 support.
621 .sp
622 PCRE_DOLLAR_ENDONLY
623 .sp
624 If this bit is set, a dollar metacharacter in the pattern matches only at the
625 end of the subject string. Without this option, a dollar also matches
626 immediately before a newline at the end of the string (but not before any other
627 newlines). The PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
628 There is no equivalent to this option in Perl, and no way to set it within a
629 pattern.
630 .sp
631 PCRE_DOTALL
632 .sp
633 If this bit is set, a dot metacharacter in the pattern matches a character of
634 any value, including one that indicates a newline. However, it only ever
635 matches one character, even if newlines are coded as CRLF. Without this option,
636 a dot does not match when the current position is at a newline. This option is
637 equivalent to Perl's /s option, and it can be changed within a pattern by a
638 (?s) option setting. A negative class such as [^a] always matches newline
639 characters, independent of the setting of this option.
640 .sp
641 PCRE_DUPNAMES
642 .sp
643 If this bit is set, names used to identify capturing subpatterns need not be
644 unique. This can be helpful for certain types of pattern when it is known that
645 only one instance of the named subpattern can ever be matched. There are more
646 details of named subpatterns below; see also the
647 .\" HREF
648 \fBpcrepattern\fP
649 .\"
650 documentation.
651 .sp
652 PCRE_EXTENDED
653 .sp
654 If this bit is set, white space data characters in the pattern are totally
655 ignored except when escaped or inside a character class. White space did not
656 used to include the VT character (code 11), because Perl did not treat this
657 character as white space. However, Perl changed at release 5.18, so PCRE
658 followed at release 8.34, and VT is now treated as white space. PCRE_EXTENDED
659 also causes characters between an unescaped # outside a character class and the
660 next newline, inclusive, to be ignored. PCRE_EXTENDED is equivalent to
661 Perl's /x option, and it can be changed within a pattern by a (?x) option
662 setting.
663 .P
664 Which characters are interpreted as newlines is controlled by the options
665 passed to \fBpcre_compile()\fP or by a special sequence at the start of the
666 pattern, as described in the section entitled
667 .\" HTML <a href="pcrepattern.html#newlines">
668 .\" </a>
669 "Newline conventions"
670 .\"
671 in the \fBpcrepattern\fP documentation. Note that the end of this type of
672 comment is a literal newline sequence in the pattern; escape sequences that
673 happen to represent a newline do not count.
674 .P
675 This option makes it possible to include comments inside complicated patterns.
676 Note, however, that this applies only to data characters. White space characters
677 may never appear within special character sequences in a pattern, for example
678 within the sequence (?( that introduces a conditional subpattern.
679 .sp
680 PCRE_EXTRA
681 .sp
682 This option was invented in order to turn on additional functionality of PCRE
683 that is incompatible with Perl, but it is currently of very little use. When
684 set, any backslash in a pattern that is followed by a letter that has no
685 special meaning causes an error, thus reserving these combinations for future
686 expansion. By default, as in Perl, a backslash followed by a letter with no
687 special meaning is treated as a literal. (Perl can, however, be persuaded to
688 give an error for this, by running it with the -w option.) There are at present
689 no other features controlled by this option. It can also be set by a (?X)
690 option setting within a pattern.
691 .sp
692 PCRE_FIRSTLINE
693 .sp
694 If this option is set, an unanchored pattern is required to match before or at
695 the first newline in the subject string, though the matched text may continue
696 over the newline.
697 .sp
698 PCRE_JAVASCRIPT_COMPAT
699 .sp
700 If this option is set, PCRE's behaviour is changed in some ways so that it is
701 compatible with JavaScript rather than Perl. The changes are as follows:
702 .P
703 (1) A lone closing square bracket in a pattern causes a compile-time error,
704 because this is illegal in JavaScript (by default it is treated as a data
705 character). Thus, the pattern AB]CD becomes illegal when this option is set.
706 .P
707 (2) At run time, a back reference to an unset subpattern group matches an empty
708 string (by default this causes the current matching alternative to fail). A
709 pattern such as (\e1)(a) succeeds when this option is set (assuming it can find
710 an "a" in the subject), whereas it fails by default, for Perl compatibility.
711 .P
712 (3) \eU matches an upper case "U" character; by default \eU causes a compile
713 time error (Perl uses \eU to upper case subsequent characters).
714 .P
715 (4) \eu matches a lower case "u" character unless it is followed by four
716 hexadecimal digits, in which case the hexadecimal number defines the code point
717 to match. By default, \eu causes a compile time error (Perl uses it to upper
718 case the following character).
719 .P
720 (5) \ex matches a lower case "x" character unless it is followed by two
721 hexadecimal digits, in which case the hexadecimal number defines the code point
722 to match. By default, as in Perl, a hexadecimal number is always expected after
723 \ex, but it may have zero, one, or two digits (so, for example, \exz matches a
724 binary zero character followed by z).
725 .sp
726 PCRE_MULTILINE
727 .sp
728 By default, for the purposes of matching "start of line" and "end of line",
729 PCRE treats the subject string as consisting of a single line of characters,
730 even if it actually contains newlines. The "start of line" metacharacter (^)
731 matches only at the start of the string, and the "end of line" metacharacter
732 ($) matches only at the end of the string, or before a terminating newline
733 (except when PCRE_DOLLAR_ENDONLY is set). Note, however, that unless
734 PCRE_DOTALL is set, the "any character" metacharacter (.) does not match at a
735 newline. This behaviour (for ^, $, and dot) is the same as Perl.
736 .P
737 When PCRE_MULTILINE it is set, the "start of line" and "end of line" constructs
738 match immediately following or immediately before internal newlines in the
739 subject string, respectively, as well as at the very start and end. This is
740 equivalent to Perl's /m option, and it can be changed within a pattern by a
741 (?m) option setting. If there are no newlines in a subject string, or no
742 occurrences of ^ or $ in a pattern, setting PCRE_MULTILINE has no effect.
743 .sp
744 PCRE_NEVER_UTF
745 .sp
746 This option locks out interpretation of the pattern as UTF-8 (or UTF-16 or
747 UTF-32 in the 16-bit and 32-bit libraries). In particular, it prevents the
748 creator of the pattern from switching to UTF interpretation by starting the
749 pattern with (*UTF). This may be useful in applications that process patterns
750 from external sources. The combination of PCRE_UTF8 and PCRE_NEVER_UTF also
751 causes an error.
752 .sp
753 PCRE_NEWLINE_CR
754 PCRE_NEWLINE_LF
755 PCRE_NEWLINE_CRLF
756 PCRE_NEWLINE_ANYCRLF
757 PCRE_NEWLINE_ANY
758 .sp
759 These options override the default newline definition that was chosen when PCRE
760 was built. Setting the first or the second specifies that a newline is
761 indicated by a single character (CR or LF, respectively). Setting
762 PCRE_NEWLINE_CRLF specifies that a newline is indicated by the two-character
763 CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies that any of the three
764 preceding sequences should be recognized. Setting PCRE_NEWLINE_ANY specifies
765 that any Unicode newline sequence should be recognized.
766 .P
767 In an ASCII/Unicode environment, the Unicode newline sequences are the three
768 just mentioned, plus the single characters VT (vertical tab, U+000B), FF (form
769 feed, U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
770 (paragraph separator, U+2029). For the 8-bit library, the last two are
771 recognized only in UTF-8 mode.
772 .P
773 When PCRE is compiled to run in an EBCDIC (mainframe) environment, the code for
774 CR is 0x0d, the same as ASCII. However, the character code for LF is normally
775 0x15, though in some EBCDIC environments 0x25 is used. Whichever of these is
776 not LF is made to correspond to Unicode's NEL character. EBCDIC codes are all
777 less than 256. For more details, see the
778 .\" HREF
779 \fBpcrebuild\fP
780 .\"
781 documentation.
782 .P
783 The newline setting in the options word uses three bits that are treated
784 as a number, giving eight possibilities. Currently only six are used (default
785 plus the five values above). This means that if you set more than one newline
786 option, the combination may or may not be sensible. For example,
787 PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to PCRE_NEWLINE_CRLF, but
788 other combinations may yield unused numbers and cause an error.
789 .P
790 The only time that a line break in a pattern is specially recognized when
791 compiling is when PCRE_EXTENDED is set. CR and LF are white space characters,
792 and so are ignored in this mode. Also, an unescaped # outside a character class
793 indicates a comment that lasts until after the next line break sequence. In
794 other circumstances, line break sequences in patterns are treated as literal
795 data.
796 .P
797 The newline option that is set at compile time becomes the default that is used
798 for \fBpcre_exec()\fP and \fBpcre_dfa_exec()\fP, but it can be overridden.
799 .sp
800 PCRE_NO_AUTO_CAPTURE
801 .sp
802 If this option is set, it disables the use of numbered capturing parentheses in
803 the pattern. Any opening parenthesis that is not followed by ? behaves as if it
804 were followed by ?: but named parentheses can still be used for capturing (and
805 they acquire numbers in the usual way). There is no equivalent of this option
806 in Perl.
807 .sp
808 PCRE_NO_AUTO_POSSESSIFY
809 .sp
810 If this option is set, it disables "auto-possessification". This is an
811 optimization that, for example, turns a+b into a++b in order to avoid
812 backtracks into a+ that can never be successful.
813 .sp
814 PCRE_NO_START_OPTIMIZE
815 .sp
816 This is an option that acts at matching time; that is, it is really an option
817 for \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP. If it is set at compile time,
818 it is remembered with the compiled pattern and assumed at matching time. This
819 is necessary if you want to use JIT execution, because the JIT compiler needs
820 to know whether or not this option is set. For details see the discussion of
821 PCRE_NO_START_OPTIMIZE
822 .\" HTML <a href="#execoptions">
823 .\" </a>
824 below.
825 .\"
826 .sp
827 PCRE_UCP
828 .sp
829 This option changes the way PCRE processes \eB, \eb, \eD, \ed, \eS, \es, \eW,
830 \ew, and some of the POSIX character classes. By default, only ASCII characters
831 are recognized, but if PCRE_UCP is set, Unicode properties are used instead to
832 classify characters. More details are given in the section on
833 .\" HTML <a href="pcre.html#genericchartypes">
834 .\" </a>
835 generic character types
836 .\"
837 in the
838 .\" HREF
839 \fBpcrepattern\fP
840 .\"
841 page. If you set PCRE_UCP, matching one of the items it affects takes much
842 longer. The option is available only if PCRE has been compiled with Unicode
843 property support.
844 .sp
845 PCRE_UNGREEDY
846 .sp
847 This option inverts the "greediness" of the quantifiers so that they are not
848 greedy by default, but become greedy if followed by "?". It is not compatible
849 with Perl. It can also be set by a (?U) option setting within the pattern.
850 .sp
851 PCRE_UTF8
852 .sp
853 This option causes PCRE to regard both the pattern and the subject as strings
854 of UTF-8 characters instead of single-byte strings. However, it is available
855 only when PCRE is built to include UTF support. If not, the use of this option
856 provokes an error. Details of how this option changes the behaviour of PCRE are
857 given in the
858 .\" HREF
859 \fBpcreunicode\fP
860 .\"
861 page.
862 .sp
863 PCRE_NO_UTF8_CHECK
864 .sp
865 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
866 automatically checked. There is a discussion about the
867 .\" HTML <a href="pcreunicode.html#utf8strings">
868 .\" </a>
869 validity of UTF-8 strings
870 .\"
871 in the
872 .\" HREF
873 \fBpcreunicode\fP
874 .\"
875 page. If an invalid UTF-8 sequence is found, \fBpcre_compile()\fP returns an
876 error. If you already know that your pattern is valid, and you want to skip
877 this check for performance reasons, you can set the PCRE_NO_UTF8_CHECK option.
878 When it is set, the effect of passing an invalid UTF-8 string as a pattern is
879 undefined. It may cause your program to crash or loop. Note that this option
880 can also be passed to \fBpcre_exec()\fP and \fBpcre_dfa_exec()\fP, to suppress
881 the validity checking of subject strings only. If the same string is being
882 matched many times, the option can be safely set for the second and subsequent
883 matchings to improve performance.
884 .
885 .
886 .SH "COMPILATION ERROR CODES"
887 .rs
888 .sp
889 The following table lists the error codes than may be returned by
890 \fBpcre_compile2()\fP, along with the error messages that may be returned by
891 both compiling functions. Note that error messages are always 8-bit ASCII
892 strings, even in 16-bit or 32-bit mode. As PCRE has developed, some error codes
893 have fallen out of use. To avoid confusion, they have not been re-used.
894 .sp
895 0 no error
896 1 \e at end of pattern
897 2 \ec at end of pattern
898 3 unrecognized character follows \e
899 4 numbers out of order in {} quantifier
900 5 number too big in {} quantifier
901 6 missing terminating ] for character class
902 7 invalid escape sequence in character class
903 8 range out of order in character class
904 9 nothing to repeat
905 10 [this code is not in use]
906 11 internal error: unexpected repeat
907 12 unrecognized character after (? or (?-
908 13 POSIX named classes are supported only within a class
909 14 missing )
910 15 reference to non-existent subpattern
911 16 erroffset passed as NULL
912 17 unknown option bit(s) set
913 18 missing ) after comment
914 19 [this code is not in use]
915 20 regular expression is too large
916 21 failed to get memory
917 22 unmatched parentheses
918 23 internal error: code overflow
919 24 unrecognized character after (?<
920 25 lookbehind assertion is not fixed length
921 26 malformed number or name after (?(
922 27 conditional group contains more than two branches
923 28 assertion expected after (?(
924 29 (?R or (?[+-]digits must be followed by )
925 30 unknown POSIX class name
926 31 POSIX collating elements are not supported
927 32 this version of PCRE is compiled without UTF support
928 33 [this code is not in use]
929 34 character value in \ex{} or \eo{} is too large
930 35 invalid condition (?(0)
931 36 \eC not allowed in lookbehind assertion
932 37 PCRE does not support \eL, \el, \eN{name}, \eU, or \eu
933 38 number after (?C is > 255
934 39 closing ) for (?C expected
935 40 recursive call could loop indefinitely
936 41 unrecognized character after (?P
937 42 syntax error in subpattern name (missing terminator)
938 43 two named subpatterns have the same name
939 44 invalid UTF-8 string (specifically UTF-8)
940 45 support for \eP, \ep, and \eX has not been compiled
941 46 malformed \eP or \ep sequence
942 47 unknown property name after \eP or \ep
943 48 subpattern name is too long (maximum 32 characters)
944 49 too many named subpatterns (maximum 10000)
945 50 [this code is not in use]
946 51 octal value is greater than \e377 in 8-bit non-UTF-8 mode
947 52 internal error: overran compiling workspace
948 53 internal error: previously-checked referenced subpattern
949 not found
950 54 DEFINE group contains more than one branch
951 55 repeating a DEFINE group is not allowed
952 56 inconsistent NEWLINE options
953 57 \eg is not followed by a braced, angle-bracketed, or quoted
954 name/number or by a plain number
955 58 a numbered reference must not be zero
956 59 an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)
957 60 (*VERB) not recognized or malformed
958 61 number is too big
959 62 subpattern name expected
960 63 digit expected after (?+
961 64 ] is an invalid data character in JavaScript compatibility mode
962 65 different names for subpatterns of the same number are
963 not allowed
964 66 (*MARK) must have an argument
965 67 this version of PCRE is not compiled with Unicode property
966 support
967 68 \ec must be followed by an ASCII character
968 69 \ek is not followed by a braced, angle-bracketed, or quoted name
969 70 internal error: unknown opcode in find_fixedlength()
970 71 \eN is not supported in a class
971 72 too many forward references
972 73 disallowed Unicode code point (>= 0xd800 && <= 0xdfff)
973 74 invalid UTF-16 string (specifically UTF-16)
974 75 name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)
975 76 character value in \eu.... sequence is too large
976 77 invalid UTF-32 string (specifically UTF-32)
977 78 setting UTF is disabled by the application
978 79 non-hex character in \ex{} (closing brace missing?)
979 80 non-octal character in \eo{} (closing brace missing?)
980 81 missing opening brace after \eo
981 82 parentheses are too deeply nested
982 83 invalid range in character class
983 .sp
984 The numbers 32 and 10000 in errors 48 and 49 are defaults; different values may
985 be used if the limits were changed when PCRE was built.
986 .
987 .
988 .\" HTML <a name="studyingapattern"></a>
989 .SH "STUDYING A PATTERN"
990 .rs
991 .sp
992 .nf
993 .B pcre_extra *pcre_study(const pcre *\fIcode\fP, int \fIoptions\fP,
994 .B " const char **\fIerrptr\fP);"
995 .fi
996 .PP
997 If a compiled pattern is going to be used several times, it is worth spending
998 more time analyzing it in order to speed up the time taken for matching. The
999 function \fBpcre_study()\fP takes a pointer to a compiled pattern as its first
1000 argument. If studying the pattern produces additional information that will
1001 help speed up matching, \fBpcre_study()\fP returns a pointer to a
1002 \fBpcre_extra\fP block, in which the \fIstudy_data\fP field points to the
1003 results of the study.
1004 .P
1005 The returned value from \fBpcre_study()\fP can be passed directly to
1006 \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP. However, a \fBpcre_extra\fP block
1007 also contains other fields that can be set by the caller before the block is
1008 passed; these are described
1009 .\" HTML <a href="#extradata">
1010 .\" </a>
1011 below
1012 .\"
1013 in the section on matching a pattern.
1014 .P
1015 If studying the pattern does not produce any useful information,
1016 \fBpcre_study()\fP returns NULL by default. In that circumstance, if the
1017 calling program wants to pass any of the other fields to \fBpcre_exec()\fP or
1018 \fBpcre_dfa_exec()\fP, it must set up its own \fBpcre_extra\fP block. However,
1019 if \fBpcre_study()\fP is called with the PCRE_STUDY_EXTRA_NEEDED option, it
1020 returns a \fBpcre_extra\fP block even if studying did not find any additional
1021 information. It may still return NULL, however, if an error occurs in
1022 \fBpcre_study()\fP.
1023 .P
1024 The second argument of \fBpcre_study()\fP contains option bits. There are three
1025 further options in addition to PCRE_STUDY_EXTRA_NEEDED:
1026 .sp
1027 PCRE_STUDY_JIT_COMPILE
1028 PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE
1029 PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE
1030 .sp
1031 If any of these are set, and the just-in-time compiler is available, the
1032 pattern is further compiled into machine code that executes much faster than
1033 the \fBpcre_exec()\fP interpretive matching function. If the just-in-time
1034 compiler is not available, these options are ignored. All undefined bits in the
1035 \fIoptions\fP argument must be zero.
1036 .P
1037 JIT compilation is a heavyweight optimization. It can take some time for
1038 patterns to be analyzed, and for one-off matches and simple patterns the
1039 benefit of faster execution might be offset by a much slower study time.
1040 Not all patterns can be optimized by the JIT compiler. For those that cannot be
1041 handled, matching automatically falls back to the \fBpcre_exec()\fP
1042 interpreter. For more details, see the
1043 .\" HREF
1044 \fBpcrejit\fP
1045 .\"
1046 documentation.
1047 .P
1048 The third argument for \fBpcre_study()\fP is a pointer for an error message. If
1049 studying succeeds (even if no data is returned), the variable it points to is
1050 set to NULL. Otherwise it is set to point to a textual error message. This is a
1051 static string that is part of the library. You must not try to free it. You
1052 should test the error pointer for NULL after calling \fBpcre_study()\fP, to be
1053 sure that it has run successfully.
1054 .P
1055 When you are finished with a pattern, you can free the memory used for the
1056 study data by calling \fBpcre_free_study()\fP. This function was added to the
1057 API for release 8.20. For earlier versions, the memory could be freed with
1058 \fBpcre_free()\fP, just like the pattern itself. This will still work in cases
1059 where JIT optimization is not used, but it is advisable to change to the new
1060 function when convenient.
1061 .P
1062 This is a typical way in which \fBpcre_study\fP() is used (except that in a
1063 real application there should be tests for errors):
1064 .sp
1065 int rc;
1066 pcre *re;
1067 pcre_extra *sd;
1068 re = pcre_compile("pattern", 0, &error, &erroroffset, NULL);
1069 sd = pcre_study(
1070 re, /* result of pcre_compile() */
1071 0, /* no options */
1072 &error); /* set to NULL or points to a message */
1073 rc = pcre_exec( /* see below for details of pcre_exec() options */
1074 re, sd, "subject", 7, 0, 0, ovector, 30);
1075 ...
1076 pcre_free_study(sd);
1077 pcre_free(re);
1078 .sp
1079 Studying a pattern does two things: first, a lower bound for the length of
1080 subject string that is needed to match the pattern is computed. This does not
1081 mean that there are any strings of that length that match, but it does
1082 guarantee that no shorter strings match. The value is used to avoid wasting
1083 time by trying to match strings that are shorter than the lower bound. You can
1084 find out the value in a calling program via the \fBpcre_fullinfo()\fP function.
1085 .P
1086 Studying a pattern is also useful for non-anchored patterns that do not have a
1087 single fixed starting character. A bitmap of possible starting bytes is
1088 created. This speeds up finding a position in the subject at which to start
1089 matching. (In 16-bit mode, the bitmap is used for 16-bit values less than 256.
1090 In 32-bit mode, the bitmap is used for 32-bit values less than 256.)
1091 .P
1092 These two optimizations apply to both \fBpcre_exec()\fP and
1093 \fBpcre_dfa_exec()\fP, and the information is also used by the JIT compiler.
1094 The optimizations can be disabled by setting the PCRE_NO_START_OPTIMIZE option.
1095 You might want to do this if your pattern contains callouts or (*MARK) and you
1096 want to make use of these facilities in cases where matching fails.
1097 .P
1098 PCRE_NO_START_OPTIMIZE can be specified at either compile time or execution
1099 time. However, if PCRE_NO_START_OPTIMIZE is passed to \fBpcre_exec()\fP, (that
1100 is, after any JIT compilation has happened) JIT execution is disabled. For JIT
1101 execution to work with PCRE_NO_START_OPTIMIZE, the option must be set at
1102 compile time.
1103 .P
1104 There is a longer discussion of PCRE_NO_START_OPTIMIZE
1105 .\" HTML <a href="#execoptions">
1106 .\" </a>
1107 below.
1108 .\"
1109 .
1110 .
1111 .\" HTML <a name="localesupport"></a>
1112 .SH "LOCALE SUPPORT"
1113 .rs
1114 .sp
1115 PCRE handles caseless matching, and determines whether characters are letters,
1116 digits, or whatever, by reference to a set of tables, indexed by character
1117 value. When running in UTF-8 mode, this applies only to characters
1118 with codes less than 128. By default, higher-valued codes never match escapes
1119 such as \ew or \ed, but they can be tested with \ep if PCRE is built with
1120 Unicode character property support. Alternatively, the PCRE_UCP option can be
1121 set at compile time; this causes \ew and friends to use Unicode property
1122 support instead of built-in tables. The use of locales with Unicode is
1123 discouraged. If you are handling characters with codes greater than 128, you
1124 should either use UTF-8 and Unicode, or use locales, but not try to mix the
1125 two.
1126 .P
1127 PCRE contains an internal set of tables that are used when the final argument
1128 of \fBpcre_compile()\fP is NULL. These are sufficient for many applications.
1129 Normally, the internal tables recognize only ASCII characters. However, when
1130 PCRE is built, it is possible to cause the internal tables to be rebuilt in the
1131 default "C" locale of the local system, which may cause them to be different.
1132 .P
1133 The internal tables can always be overridden by tables supplied by the
1134 application that calls PCRE. These may be created in a different locale from
1135 the default. As more and more applications change to using Unicode, the need
1136 for this locale support is expected to die away.
1137 .P
1138 External tables are built by calling the \fBpcre_maketables()\fP function,
1139 which has no arguments, in the relevant locale. The result can then be passed
1140 to \fBpcre_compile()\fP or \fBpcre_exec()\fP as often as necessary. For
1141 example, to build and use tables that are appropriate for the French locale
1142 (where accented characters with values greater than 128 are treated as letters),
1143 the following code could be used:
1144 .sp
1145 setlocale(LC_CTYPE, "fr_FR");
1146 tables = pcre_maketables();
1147 re = pcre_compile(..., tables);
1148 .sp
1149 The locale name "fr_FR" is used on Linux and other Unix-like systems; if you
1150 are using Windows, the name for the French locale is "french".
1151 .P
1152 When \fBpcre_maketables()\fP runs, the tables are built in memory that is
1153 obtained via \fBpcre_malloc\fP. It is the caller's responsibility to ensure
1154 that the memory containing the tables remains available for as long as it is
1155 needed.
1156 .P
1157 The pointer that is passed to \fBpcre_compile()\fP is saved with the compiled
1158 pattern, and the same tables are used via this pointer by \fBpcre_study()\fP
1159 and normally also by \fBpcre_exec()\fP. Thus, by default, for any single
1160 pattern, compilation, studying and matching all happen in the same locale, but
1161 different patterns can be compiled in different locales.
1162 .P
1163 It is possible to pass a table pointer or NULL (indicating the use of the
1164 internal tables) to \fBpcre_exec()\fP. Although not intended for this purpose,
1165 this facility could be used to match a pattern in a different locale from the
1166 one in which it was compiled. Passing table pointers at run time is discussed
1167 below in the section on matching a pattern.
1168 .
1169 .
1170 .\" HTML <a name="infoaboutpattern"></a>
1171 .SH "INFORMATION ABOUT A PATTERN"
1172 .rs
1173 .sp
1174 .nf
1175 .B int pcre_fullinfo(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
1176 .B " int \fIwhat\fP, void *\fIwhere\fP);"
1177 .fi
1178 .PP
1179 The \fBpcre_fullinfo()\fP function returns information about a compiled
1180 pattern. It replaces the \fBpcre_info()\fP function, which was removed from the
1181 library at version 8.30, after more than 10 years of obsolescence.
1182 .P
1183 The first argument for \fBpcre_fullinfo()\fP is a pointer to the compiled
1184 pattern. The second argument is the result of \fBpcre_study()\fP, or NULL if
1185 the pattern was not studied. The third argument specifies which piece of
1186 information is required, and the fourth argument is a pointer to a variable
1187 to receive the data. The yield of the function is zero for success, or one of
1188 the following negative numbers:
1189 .sp
1190 PCRE_ERROR_NULL the argument \fIcode\fP was NULL
1191 the argument \fIwhere\fP was NULL
1192 PCRE_ERROR_BADMAGIC the "magic number" was not found
1193 PCRE_ERROR_BADENDIANNESS the pattern was compiled with different
1194 endianness
1195 PCRE_ERROR_BADOPTION the value of \fIwhat\fP was invalid
1196 PCRE_ERROR_UNSET the requested field is not set
1197 .sp
1198 The "magic number" is placed at the start of each compiled pattern as an simple
1199 check against passing an arbitrary memory pointer. The endianness error can
1200 occur if a compiled pattern is saved and reloaded on a different host. Here is
1201 a typical call of \fBpcre_fullinfo()\fP, to obtain the length of the compiled
1202 pattern:
1203 .sp
1204 int rc;
1205 size_t length;
1206 rc = pcre_fullinfo(
1207 re, /* result of pcre_compile() */
1208 sd, /* result of pcre_study(), or NULL */
1209 PCRE_INFO_SIZE, /* what is required */
1210 &length); /* where to put the data */
1211 .sp
1212 The possible values for the third argument are defined in \fBpcre.h\fP, and are
1213 as follows:
1214 .sp
1215 PCRE_INFO_BACKREFMAX
1216 .sp
1217 Return the number of the highest back reference in the pattern. The fourth
1218 argument should point to an \fBint\fP variable. Zero is returned if there are
1219 no back references.
1220 .sp
1221 PCRE_INFO_CAPTURECOUNT
1222 .sp
1223 Return the number of capturing subpatterns in the pattern. The fourth argument
1224 should point to an \fBint\fP variable.
1225 .sp
1226 PCRE_INFO_DEFAULT_TABLES
1227 .sp
1228 Return a pointer to the internal default character tables within PCRE. The
1229 fourth argument should point to an \fBunsigned char *\fP variable. This
1230 information call is provided for internal use by the \fBpcre_study()\fP
1231 function. External callers can cause PCRE to use its internal tables by passing
1232 a NULL table pointer.
1233 .sp
1234 PCRE_INFO_FIRSTBYTE
1235 .sp
1236 Return information about the first data unit of any matched string, for a
1237 non-anchored pattern. (The name of this option refers to the 8-bit library,
1238 where data units are bytes.) The fourth argument should point to an \fBint\fP
1239 variable.
1240 .P
1241 If there is a fixed first value, for example, the letter "c" from a pattern
1242 such as (cat|cow|coyote), its value is returned. In the 8-bit library, the
1243 value is always less than 256. In the 16-bit library the value can be up to
1244 0xffff. In the 32-bit library the value can be up to 0x10ffff.
1245 .P
1246 If there is no fixed first value, and if either
1247 .sp
1248 (a) the pattern was compiled with the PCRE_MULTILINE option, and every branch
1249 starts with "^", or
1250 .sp
1251 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not set
1252 (if it were set, the pattern would be anchored),
1253 .sp
1254 -1 is returned, indicating that the pattern matches only at the start of a
1255 subject string or after any newline within the string. Otherwise -2 is
1256 returned. For anchored patterns, -2 is returned.
1257 .P
1258 Since for the 32-bit library using the non-UTF-32 mode, this function is unable
1259 to return the full 32-bit range of the character, this value is deprecated;
1260 instead the PCRE_INFO_FIRSTCHARACTERFLAGS and PCRE_INFO_FIRSTCHARACTER values
1261 should be used.
1262 .sp
1263 PCRE_INFO_FIRSTTABLE
1264 .sp
1265 If the pattern was studied, and this resulted in the construction of a 256-bit
1266 table indicating a fixed set of values for the first data unit in any matching
1267 string, a pointer to the table is returned. Otherwise NULL is returned. The
1268 fourth argument should point to an \fBunsigned char *\fP variable.
1269 .sp
1270 PCRE_INFO_HASCRORLF
1271 .sp
1272 Return 1 if the pattern contains any explicit matches for CR or LF characters,
1273 otherwise 0. The fourth argument should point to an \fBint\fP variable. An
1274 explicit match is either a literal CR or LF character, or \er or \en.
1275 .sp
1276 PCRE_INFO_JCHANGED
1277 .sp
1278 Return 1 if the (?J) or (?-J) option setting is used in the pattern, otherwise
1279 0. The fourth argument should point to an \fBint\fP variable. (?J) and
1280 (?-J) set and unset the local PCRE_DUPNAMES option, respectively.
1281 .sp
1282 PCRE_INFO_JIT
1283 .sp
1284 Return 1 if the pattern was studied with one of the JIT options, and
1285 just-in-time compiling was successful. The fourth argument should point to an
1286 \fBint\fP variable. A return value of 0 means that JIT support is not available
1287 in this version of PCRE, or that the pattern was not studied with a JIT option,
1288 or that the JIT compiler could not handle this particular pattern. See the
1289 .\" HREF
1290 \fBpcrejit\fP
1291 .\"
1292 documentation for details of what can and cannot be handled.
1293 .sp
1294 PCRE_INFO_JITSIZE
1295 .sp
1296 If the pattern was successfully studied with a JIT option, return the size of
1297 the JIT compiled code, otherwise return zero. The fourth argument should point
1298 to a \fBsize_t\fP variable.
1299 .sp
1300 PCRE_INFO_LASTLITERAL
1301 .sp
1302 Return the value of the rightmost literal data unit that must exist in any
1303 matched string, other than at its start, if such a value has been recorded. The
1304 fourth argument should point to an \fBint\fP variable. If there is no such
1305 value, -1 is returned. For anchored patterns, a last literal value is recorded
1306 only if it follows something of variable length. For example, for the pattern
1307 /^a\ed+z\ed+/ the returned value is "z", but for /^a\edz\ed/ the returned value
1308 is -1.
1309 .P
1310 Since for the 32-bit library using the non-UTF-32 mode, this function is unable
1311 to return the full 32-bit range of characters, this value is deprecated;
1312 instead the PCRE_INFO_REQUIREDCHARFLAGS and PCRE_INFO_REQUIREDCHAR values should
1313 be used.
1314 .sp
1315 PCRE_INFO_MATCH_EMPTY
1316 .sp
1317 Return 1 if the pattern can match an empty string, otherwise 0. The fourth
1318 argument should point to an \fBint\fP variable.
1319 .sp
1320 PCRE_INFO_MATCHLIMIT
1321 .sp
1322 If the pattern set a match limit by including an item of the form
1323 (*LIMIT_MATCH=nnnn) at the start, the value is returned. The fourth argument
1324 should point to an unsigned 32-bit integer. If no such value has been set, the
1325 call to \fBpcre_fullinfo()\fP returns the error PCRE_ERROR_UNSET.
1326 .sp
1327 PCRE_INFO_MAXLOOKBEHIND
1328 .sp
1329 Return the number of characters (NB not data units) in the longest lookbehind
1330 assertion in the pattern. This information is useful when doing multi-segment
1331 matching using the partial matching facilities. Note that the simple assertions
1332 \eb and \eB require a one-character lookbehind. \eA also registers a
1333 one-character lookbehind, though it does not actually inspect the previous
1334 character. This is to ensure that at least one character from the old segment
1335 is retained when a new segment is processed. Otherwise, if there are no
1336 lookbehinds in the pattern, \eA might match incorrectly at the start of a new
1337 segment.
1338 .sp
1339 PCRE_INFO_MINLENGTH
1340 .sp
1341 If the pattern was studied and a minimum length for matching subject strings
1342 was computed, its value is returned. Otherwise the returned value is -1. The
1343 value is a number of characters, which in UTF mode may be different from the
1344 number of data units. The fourth argument should point to an \fBint\fP
1345 variable. A non-negative value is a lower bound to the length of any matching
1346 string. There may not be any strings of that length that do actually match, but
1347 every string that does match is at least that long.
1348 .sp
1349 PCRE_INFO_NAMECOUNT
1350 PCRE_INFO_NAMEENTRYSIZE
1351 PCRE_INFO_NAMETABLE
1352 .sp
1353 PCRE supports the use of named as well as numbered capturing parentheses. The
1354 names are just an additional way of identifying the parentheses, which still
1355 acquire numbers. Several convenience functions such as
1356 \fBpcre_get_named_substring()\fP are provided for extracting captured
1357 substrings by name. It is also possible to extract the data directly, by first
1358 converting the name to a number in order to access the correct pointers in the
1359 output vector (described with \fBpcre_exec()\fP below). To do the conversion,
1360 you need to use the name-to-number map, which is described by these three
1361 values.
1362 .P
1363 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT gives
1364 the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size of each
1365 entry; both of these return an \fBint\fP value. The entry size depends on the
1366 length of the longest name. PCRE_INFO_NAMETABLE returns a pointer to the first
1367 entry of the table. This is a pointer to \fBchar\fP in the 8-bit library, where
1368 the first two bytes of each entry are the number of the capturing parenthesis,
1369 most significant byte first. In the 16-bit library, the pointer points to
1370 16-bit data units, the first of which contains the parenthesis number. In the
1371 32-bit library, the pointer points to 32-bit data units, the first of which
1372 contains the parenthesis number. The rest of the entry is the corresponding
1373 name, zero terminated.
1374 .P
1375 The names are in alphabetical order. If (?| is used to create multiple groups
1376 with the same number, as described in the
1377 .\" HTML <a href="pcrepattern.html#dupsubpatternnumber">
1378 .\" </a>
1379 section on duplicate subpattern numbers
1380 .\"
1381 in the
1382 .\" HREF
1383 \fBpcrepattern\fP
1384 .\"
1385 page, the groups may be given the same name, but there is only one entry in the
1386 table. Different names for groups of the same number are not permitted.
1387 Duplicate names for subpatterns with different numbers are permitted,
1388 but only if PCRE_DUPNAMES is set. They appear in the table in the order in
1389 which they were found in the pattern. In the absence of (?| this is the order
1390 of increasing number; when (?| is used this is not necessarily the case because
1391 later subpatterns may have lower numbers.
1392 .P
1393 As a simple example of the name/number table, consider the following pattern
1394 after compilation by the 8-bit library (assume PCRE_EXTENDED is set, so white
1395 space - including newlines - is ignored):
1396 .sp
1397 .\" JOIN
1398 (?<date> (?<year>(\ed\ed)?\ed\ed) -
1399 (?<month>\ed\ed) - (?<day>\ed\ed) )
1400 .sp
1401 There are four named subpatterns, so the table has four entries, and each entry
1402 in the table is eight bytes long. The table is as follows, with non-printing
1403 bytes shows in hexadecimal, and undefined bytes shown as ??:
1404 .sp
1405 00 01 d a t e 00 ??
1406 00 05 d a y 00 ?? ??
1407 00 04 m o n t h 00
1408 00 02 y e a r 00 ??
1409 .sp
1410 When writing code to extract data from named subpatterns using the
1411 name-to-number map, remember that the length of the entries is likely to be
1412 different for each compiled pattern.
1413 .sp
1414 PCRE_INFO_OKPARTIAL
1415 .sp
1416 Return 1 if the pattern can be used for partial matching with
1417 \fBpcre_exec()\fP, otherwise 0. The fourth argument should point to an
1418 \fBint\fP variable. From release 8.00, this always returns 1, because the
1419 restrictions that previously applied to partial matching have been lifted. The
1420 .\" HREF
1421 \fBpcrepartial\fP
1422 .\"
1423 documentation gives details of partial matching.
1424 .sp
1425 PCRE_INFO_OPTIONS
1426 .sp
1427 Return a copy of the options with which the pattern was compiled. The fourth
1428 argument should point to an \fBunsigned long int\fP variable. These option bits
1429 are those specified in the call to \fBpcre_compile()\fP, modified by any
1430 top-level option settings at the start of the pattern itself. In other words,
1431 they are the options that will be in force when matching starts. For example,
1432 if the pattern /(?im)abc(?-i)d/ is compiled with the PCRE_EXTENDED option, the
1433 result is PCRE_CASELESS, PCRE_MULTILINE, and PCRE_EXTENDED.
1434 .P
1435 A pattern is automatically anchored by PCRE if all of its top-level
1436 alternatives begin with one of the following:
1437 .sp
1438 ^ unless PCRE_MULTILINE is set
1439 \eA always
1440 \eG always
1441 .\" JOIN
1442 .* if PCRE_DOTALL is set and there are no back
1443 references to the subpattern in which .* appears
1444 .sp
1445 For such patterns, the PCRE_ANCHORED bit is set in the options returned by
1446 \fBpcre_fullinfo()\fP.
1447 .sp
1448 PCRE_INFO_RECURSIONLIMIT
1449 .sp
1450 If the pattern set a recursion limit by including an item of the form
1451 (*LIMIT_RECURSION=nnnn) at the start, the value is returned. The fourth
1452 argument should point to an unsigned 32-bit integer. If no such value has been
1453 set, the call to \fBpcre_fullinfo()\fP returns the error PCRE_ERROR_UNSET.
1454 .sp
1455 PCRE_INFO_SIZE
1456 .sp
1457 Return the size of the compiled pattern in bytes (for all three libraries). The
1458 fourth argument should point to a \fBsize_t\fP variable. This value does not
1459 include the size of the \fBpcre\fP structure that is returned by
1460 \fBpcre_compile()\fP. The value that is passed as the argument to
1461 \fBpcre_malloc()\fP when \fBpcre_compile()\fP is getting memory in which to
1462 place the compiled data is the value returned by this option plus the size of
1463 the \fBpcre\fP structure. Studying a compiled pattern, with or without JIT,
1464 does not alter the value returned by this option.
1465 .sp
1466 PCRE_INFO_STUDYSIZE
1467 .sp
1468 Return the size in bytes (for all three libraries) of the data block pointed to
1469 by the \fIstudy_data\fP field in a \fBpcre_extra\fP block. If \fBpcre_extra\fP
1470 is NULL, or there is no study data, zero is returned. The fourth argument
1471 should point to a \fBsize_t\fP variable. The \fIstudy_data\fP field is set by
1472 \fBpcre_study()\fP to record information that will speed up matching (see the
1473 section entitled
1474 .\" HTML <a href="#studyingapattern">
1475 .\" </a>
1476 "Studying a pattern"
1477 .\"
1478 above). The format of the \fIstudy_data\fP block is private, but its length
1479 is made available via this option so that it can be saved and restored (see the
1480 .\" HREF
1481 \fBpcreprecompile\fP
1482 .\"
1483 documentation for details).
1484 .sp
1485 PCRE_INFO_FIRSTCHARACTERFLAGS
1486 .sp
1487 Return information about the first data unit of any matched string, for a
1488 non-anchored pattern. The fourth argument should point to an \fBint\fP
1489 variable.
1490 .P
1491 If there is a fixed first value, for example, the letter "c" from a pattern
1492 such as (cat|cow|coyote), 1 is returned, and the character value can be
1493 retrieved using PCRE_INFO_FIRSTCHARACTER.
1494 .P
1495 If there is no fixed first value, and if either
1496 .sp
1497 (a) the pattern was compiled with the PCRE_MULTILINE option, and every branch
1498 starts with "^", or
1499 .sp
1500 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not set
1501 (if it were set, the pattern would be anchored),
1502 .sp
1503 2 is returned, indicating that the pattern matches only at the start of a
1504 subject string or after any newline within the string. Otherwise 0 is
1505 returned. For anchored patterns, 0 is returned.
1506 .sp
1507 PCRE_INFO_FIRSTCHARACTER
1508 .sp
1509 Return the fixed first character value in the situation where
1510 PCRE_INFO_FIRSTCHARACTERFLAGS returns 1; otherwise return 0. The fourth
1511 argument should point to an \fBuint_t\fP variable.
1512 .P
1513 In the 8-bit library, the value is always less than 256. In the 16-bit library
1514 the value can be up to 0xffff. In the 32-bit library in UTF-32 mode the value
1515 can be up to 0x10ffff, and up to 0xffffffff when not using UTF-32 mode.
1516 .sp
1517 PCRE_INFO_REQUIREDCHARFLAGS
1518 .sp
1519 Returns 1 if there is a rightmost literal data unit that must exist in any
1520 matched string, other than at its start. The fourth argument should point to
1521 an \fBint\fP variable. If there is no such value, 0 is returned. If returning
1522 1, the character value itself can be retrieved using PCRE_INFO_REQUIREDCHAR.
1523 .P
1524 For anchored patterns, a last literal value is recorded only if it follows
1525 something of variable length. For example, for the pattern /^a\ed+z\ed+/ the
1526 returned value 1 (with "z" returned from PCRE_INFO_REQUIREDCHAR), but for
1527 /^a\edz\ed/ the returned value is 0.
1528 .sp
1529 PCRE_INFO_REQUIREDCHAR
1530 .sp
1531 Return the value of the rightmost literal data unit that must exist in any
1532 matched string, other than at its start, if such a value has been recorded. The
1533 fourth argument should point to an \fBuint32_t\fP variable. If there is no such
1534 value, 0 is returned.
1535 .
1536 .
1537 .SH "REFERENCE COUNTS"
1538 .rs
1539 .sp
1540 .B int pcre_refcount(pcre *\fIcode\fP, int \fIadjust\fP);
1541 .PP
1542 The \fBpcre_refcount()\fP function is used to maintain a reference count in the
1543 data block that contains a compiled pattern. It is provided for the benefit of
1544 applications that operate in an object-oriented manner, where different parts
1545 of the application may be using the same compiled pattern, but you want to free
1546 the block when they are all done.
1547 .P
1548 When a pattern is compiled, the reference count field is initialized to zero.
1549 It is changed only by calling this function, whose action is to add the
1550 \fIadjust\fP value (which may be positive or negative) to it. The yield of the
1551 function is the new value. However, the value of the count is constrained to
1552 lie between 0 and 65535, inclusive. If the new value is outside these limits,
1553 it is forced to the appropriate limit value.
1554 .P
1555 Except when it is zero, the reference count is not correctly preserved if a
1556 pattern is compiled on one host and then transferred to a host whose byte-order
1557 is different. (This seems a highly unlikely scenario.)
1558 .
1559 .
1560 .SH "MATCHING A PATTERN: THE TRADITIONAL FUNCTION"
1561 .rs
1562 .sp
1563 .nf
1564 .B int pcre_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
1565 .B " const char *\fIsubject\fP," int \fIlength\fP, int \fIstartoffset\fP,
1566 .B " int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP);"
1567 .fi
1568 .P
1569 The function \fBpcre_exec()\fP is called to match a subject string against a
1570 compiled pattern, which is passed in the \fIcode\fP argument. If the
1571 pattern was studied, the result of the study should be passed in the
1572 \fIextra\fP argument. You can call \fBpcre_exec()\fP with the same \fIcode\fP
1573 and \fIextra\fP arguments as many times as you like, in order to match
1574 different subject strings with the same pattern.
1575 .P
1576 This function is the main matching facility of the library, and it operates in
1577 a Perl-like manner. For specialist use there is also an alternative matching
1578 function, which is described
1579 .\" HTML <a href="#dfamatch">
1580 .\" </a>
1581 below
1582 .\"
1583 in the section about the \fBpcre_dfa_exec()\fP function.
1584 .P
1585 In most applications, the pattern will have been compiled (and optionally
1586 studied) in the same process that calls \fBpcre_exec()\fP. However, it is
1587 possible to save compiled patterns and study data, and then use them later
1588 in different processes, possibly even on different hosts. For a discussion
1589 about this, see the
1590 .\" HREF
1591 \fBpcreprecompile\fP
1592 .\"
1593 documentation.
1594 .P
1595 Here is an example of a simple call to \fBpcre_exec()\fP:
1596 .sp
1597 int rc;
1598 int ovector[30];
1599 rc = pcre_exec(
1600 re, /* result of pcre_compile() */
1601 NULL, /* we didn't study the pattern */
1602 "some string", /* the subject string */
1603 11, /* the length of the subject string */
1604 0, /* start at offset 0 in the subject */
1605 0, /* default options */
1606 ovector, /* vector of integers for substring information */
1607 30); /* number of elements (NOT size in bytes) */
1608 .
1609 .
1610 .\" HTML <a name="extradata"></a>
1611 .SS "Extra data for \fBpcre_exec()\fR"
1612 .rs
1613 .sp
1614 If the \fIextra\fP argument is not NULL, it must point to a \fBpcre_extra\fP
1615 data block. The \fBpcre_study()\fP function returns such a block (when it
1616 doesn't return NULL), but you can also create one for yourself, and pass
1617 additional information in it. The \fBpcre_extra\fP block contains the following
1618 fields (not necessarily in this order):
1619 .sp
1620 unsigned long int \fIflags\fP;
1621 void *\fIstudy_data\fP;
1622 void *\fIexecutable_jit\fP;
1623 unsigned long int \fImatch_limit\fP;
1624 unsigned long int \fImatch_limit_recursion\fP;
1625 void *\fIcallout_data\fP;
1626 const unsigned char *\fItables\fP;
1627 unsigned char **\fImark\fP;
1628 .sp
1629 In the 16-bit version of this structure, the \fImark\fP field has type
1630 "PCRE_UCHAR16 **".
1631 .sp
1632 In the 32-bit version of this structure, the \fImark\fP field has type
1633 "PCRE_UCHAR32 **".
1634 .P
1635 The \fIflags\fP field is used to specify which of the other fields are set. The
1636 flag bits are:
1637 .sp
1638 PCRE_EXTRA_CALLOUT_DATA
1639 PCRE_EXTRA_EXECUTABLE_JIT
1640 PCRE_EXTRA_MARK
1641 PCRE_EXTRA_MATCH_LIMIT
1642 PCRE_EXTRA_MATCH_LIMIT_RECURSION
1643 PCRE_EXTRA_STUDY_DATA
1644 PCRE_EXTRA_TABLES
1645 .sp
1646 Other flag bits should be set to zero. The \fIstudy_data\fP field and sometimes
1647 the \fIexecutable_jit\fP field are set in the \fBpcre_extra\fP block that is
1648 returned by \fBpcre_study()\fP, together with the appropriate flag bits. You
1649 should not set these yourself, but you may add to the block by setting other
1650 fields and their corresponding flag bits.
1651 .P
1652 The \fImatch_limit\fP field provides a means of preventing PCRE from using up a
1653 vast amount of resources when running patterns that are not going to match,
1654 but which have a very large number of possibilities in their search trees. The
1655 classic example is a pattern that uses nested unlimited repeats.
1656 .P
1657 Internally, \fBpcre_exec()\fP uses a function called \fBmatch()\fP, which it
1658 calls repeatedly (sometimes recursively). The limit set by \fImatch_limit\fP is
1659 imposed on the number of times this function is called during a match, which
1660 has the effect of limiting the amount of backtracking that can take place. For
1661 patterns that are not anchored, the count restarts from zero for each position
1662 in the subject string.
1663 .P
1664 When \fBpcre_exec()\fP is called with a pattern that was successfully studied
1665 with a JIT option, the way that the matching is executed is entirely different.
1666 However, there is still the possibility of runaway matching that goes on for a
1667 very long time, and so the \fImatch_limit\fP value is also used in this case
1668 (but in a different way) to limit how long the matching can continue.
1669 .P
1670 The default value for the limit can be set when PCRE is built; the default
1671 default is 10 million, which handles all but the most extreme cases. You can
1672 override the default by suppling \fBpcre_exec()\fP with a \fBpcre_extra\fP
1673 block in which \fImatch_limit\fP is set, and PCRE_EXTRA_MATCH_LIMIT is set in
1674 the \fIflags\fP field. If the limit is exceeded, \fBpcre_exec()\fP returns
1675 PCRE_ERROR_MATCHLIMIT.
1676 .P
1677 A value for the match limit may also be supplied by an item at the start of a
1678 pattern of the form
1679 .sp
1680 (*LIMIT_MATCH=d)
1681 .sp
1682 where d is a decimal number. However, such a setting is ignored unless d is
1683 less than the limit set by the caller of \fBpcre_exec()\fP or, if no such limit
1684 is set, less than the default.
1685 .P
1686 The \fImatch_limit_recursion\fP field is similar to \fImatch_limit\fP, but
1687 instead of limiting the total number of times that \fBmatch()\fP is called, it
1688 limits the depth of recursion. The recursion depth is a smaller number than the
1689 total number of calls, because not all calls to \fBmatch()\fP are recursive.
1690 This limit is of use only if it is set smaller than \fImatch_limit\fP.
1691 .P
1692 Limiting the recursion depth limits the amount of machine stack that can be
1693 used, or, when PCRE has been compiled to use memory on the heap instead of the
1694 stack, the amount of heap memory that can be used. This limit is not relevant,
1695 and is ignored, when matching is done using JIT compiled code.
1696 .P
1697 The default value for \fImatch_limit_recursion\fP can be set when PCRE is
1698 built; the default default is the same value as the default for
1699 \fImatch_limit\fP. You can override the default by suppling \fBpcre_exec()\fP
1700 with a \fBpcre_extra\fP block in which \fImatch_limit_recursion\fP is set, and
1701 PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the \fIflags\fP field. If the limit
1702 is exceeded, \fBpcre_exec()\fP returns PCRE_ERROR_RECURSIONLIMIT.
1703 .P
1704 A value for the recursion limit may also be supplied by an item at the start of
1705 a pattern of the form
1706 .sp
1707 (*LIMIT_RECURSION=d)
1708 .sp
1709 where d is a decimal number. However, such a setting is ignored unless d is
1710 less than the limit set by the caller of \fBpcre_exec()\fP or, if no such limit
1711 is set, less than the default.
1712 .P
1713 The \fIcallout_data\fP field is used in conjunction with the "callout" feature,
1714 and is described in the
1715 .\" HREF
1716 \fBpcrecallout\fP
1717 .\"
1718 documentation.
1719 .P
1720 The \fItables\fP field is used to pass a character tables pointer to
1721 \fBpcre_exec()\fP; this overrides the value that is stored with the compiled
1722 pattern. A non-NULL value is stored with the compiled pattern only if custom
1723 tables were supplied to \fBpcre_compile()\fP via its \fItableptr\fP argument.
1724 If NULL is passed to \fBpcre_exec()\fP using this mechanism, it forces PCRE's
1725 internal tables to be used. This facility is helpful when re-using patterns
1726 that have been saved after compiling with an external set of tables, because
1727 the external tables might be at a different address when \fBpcre_exec()\fP is
1728 called. See the
1729 .\" HREF
1730 \fBpcreprecompile\fP
1731 .\"
1732 documentation for a discussion of saving compiled patterns for later use.
1733 .P
1734 If PCRE_EXTRA_MARK is set in the \fIflags\fP field, the \fImark\fP field must
1735 be set to point to a suitable variable. If the pattern contains any
1736 backtracking control verbs such as (*MARK:NAME), and the execution ends up with
1737 a name to pass back, a pointer to the name string (zero terminated) is placed
1738 in the variable pointed to by the \fImark\fP field. The names are within the
1739 compiled pattern; if you wish to retain such a name you must copy it before
1740 freeing the memory of a compiled pattern. If there is no name to pass back, the
1741 variable pointed to by the \fImark\fP field is set to NULL. For details of the
1742 backtracking control verbs, see the section entitled
1743 .\" HTML <a href="pcrepattern#backtrackcontrol">
1744 .\" </a>
1745 "Backtracking control"
1746 .\"
1747 in the
1748 .\" HREF
1749 \fBpcrepattern\fP
1750 .\"
1751 documentation.
1752 .
1753 .
1754 .\" HTML <a name="execoptions"></a>
1755 .SS "Option bits for \fBpcre_exec()\fP"
1756 .rs
1757 .sp
1758 The unused bits of the \fIoptions\fP argument for \fBpcre_exec()\fP must be
1759 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_\fIxxx\fP,
1760 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
1761 PCRE_NO_START_OPTIMIZE, PCRE_NO_UTF8_CHECK, PCRE_PARTIAL_HARD, and
1762 PCRE_PARTIAL_SOFT.
1763 .P
1764 If the pattern was successfully studied with one of the just-in-time (JIT)
1765 compile options, the only supported options for JIT execution are
1766 PCRE_NO_UTF8_CHECK, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY,
1767 PCRE_NOTEMPTY_ATSTART, PCRE_PARTIAL_HARD, and PCRE_PARTIAL_SOFT. If an
1768 unsupported option is used, JIT execution is disabled and the normal
1769 interpretive code in \fBpcre_exec()\fP is run.
1770 .sp
1771 PCRE_ANCHORED
1772 .sp
1773 The PCRE_ANCHORED option limits \fBpcre_exec()\fP to matching at the first
1774 matching position. If a pattern was compiled with PCRE_ANCHORED, or turned out
1775 to be anchored by virtue of its contents, it cannot be made unachored at
1776 matching time.
1777 .sp
1778 PCRE_BSR_ANYCRLF
1779 PCRE_BSR_UNICODE
1780 .sp
1781 These options (which are mutually exclusive) control what the \eR escape
1782 sequence matches. The choice is either to match only CR, LF, or CRLF, or to
1783 match any Unicode newline sequence. These options override the choice that was
1784 made or defaulted when the pattern was compiled.
1785 .sp
1786 PCRE_NEWLINE_CR
1787 PCRE_NEWLINE_LF
1788 PCRE_NEWLINE_CRLF
1789 PCRE_NEWLINE_ANYCRLF
1790 PCRE_NEWLINE_ANY
1791 .sp
1792 These options override the newline definition that was chosen or defaulted when
1793 the pattern was compiled. For details, see the description of
1794 \fBpcre_compile()\fP above. During matching, the newline choice affects the
1795 behaviour of the dot, circumflex, and dollar metacharacters. It may also alter
1796 the way the match position is advanced after a match failure for an unanchored
1797 pattern.
1798 .P
1799 When PCRE_NEWLINE_CRLF, PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is set, and a
1800 match attempt for an unanchored pattern fails when the current position is at a
1801 CRLF sequence, and the pattern contains no explicit matches for CR or LF
1802 characters, the match position is advanced by two characters instead of one, in
1803 other words, to after the CRLF.
1804 .P
1805 The above rule is a compromise that makes the most common cases work as
1806 expected. For example, if the pattern is .+A (and the PCRE_DOTALL option is not
1807 set), it does not match the string "\er\enA" because, after failing at the
1808 start, it skips both the CR and the LF before retrying. However, the pattern
1809 [\er\en]A does match that string, because it contains an explicit CR or LF
1810 reference, and so advances only by one character after the first failure.
1811 .P
1812 An explicit match for CR of LF is either a literal appearance of one of those
1813 characters, or one of the \er or \en escape sequences. Implicit matches such as
1814 [^X] do not count, nor does \es (which includes CR and LF in the characters
1815 that it matches).
1816 .P
1817 Notwithstanding the above, anomalous effects may still occur when CRLF is a
1818 valid newline sequence and explicit \er or \en escapes appear in the pattern.
1819 .sp
1820 PCRE_NOTBOL
1821 .sp
1822 This option specifies that first character of the subject string is not the
1823 beginning of a line, so the circumflex metacharacter should not match before
1824 it. Setting this without PCRE_MULTILINE (at compile time) causes circumflex
1825 never to match. This option affects only the behaviour of the circumflex
1826 metacharacter. It does not affect \eA.
1827 .sp
1828 PCRE_NOTEOL
1829 .sp
1830 This option specifies that the end of the subject string is not the end of a
1831 line, so the dollar metacharacter should not match it nor (except in multiline
1832 mode) a newline immediately before it. Setting this without PCRE_MULTILINE (at
1833 compile time) causes dollar never to match. This option affects only the
1834 behaviour of the dollar metacharacter. It does not affect \eZ or \ez.
1835 .sp
1836 PCRE_NOTEMPTY
1837 .sp
1838 An empty string is not considered to be a valid match if this option is set. If
1839 there are alternatives in the pattern, they are tried. If all the alternatives
1840 match the empty string, the entire match fails. For example, if the pattern
1841 .sp
1842 a?b?
1843 .sp
1844 is applied to a string not beginning with "a" or "b", it matches an empty
1845 string at the start of the subject. With PCRE_NOTEMPTY set, this match is not
1846 valid, so PCRE searches further into the string for occurrences of "a" or "b".
1847 .sp
1848 PCRE_NOTEMPTY_ATSTART
1849 .sp
1850 This is like PCRE_NOTEMPTY, except that an empty string match that is not at
1851 the start of the subject is permitted. If the pattern is anchored, such a match
1852 can occur only if the pattern contains \eK.
1853 .P
1854 Perl has no direct equivalent of PCRE_NOTEMPTY or PCRE_NOTEMPTY_ATSTART, but it
1855 does make a special case of a pattern match of the empty string within its
1856 \fBsplit()\fP function, and when using the /g modifier. It is possible to
1857 emulate Perl's behaviour after matching a null string by first trying the match
1858 again at the same offset with PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED, and then
1859 if that fails, by advancing the starting offset (see below) and trying an
1860 ordinary match again. There is some code that demonstrates how to do this in
1861 the
1862 .\" HREF
1863 \fBpcredemo\fP
1864 .\"
1865 sample program. In the most general case, you have to check to see if the
1866 newline convention recognizes CRLF as a newline, and if so, and the current
1867 character is CR followed by LF, advance the starting offset by two characters
1868 instead of one.
1869 .sp
1870 PCRE_NO_START_OPTIMIZE
1871 .sp
1872 There are a number of optimizations that \fBpcre_exec()\fP uses at the start of
1873 a match, in order to speed up the process. For example, if it is known that an
1874 unanchored match must start with a specific character, it searches the subject
1875 for that character, and fails immediately if it cannot find it, without
1876 actually running the main matching function. This means that a special item
1877 such as (*COMMIT) at the start of a pattern is not considered until after a
1878 suitable starting point for the match has been found. Also, when callouts or
1879 (*MARK) items are in use, these "start-up" optimizations can cause them to be
1880 skipped if the pattern is never actually used. The start-up optimizations are
1881 in effect a pre-scan of the subject that takes place before the pattern is run.
1882 .P
1883 The PCRE_NO_START_OPTIMIZE option disables the start-up optimizations, possibly
1884 causing performance to suffer, but ensuring that in cases where the result is
1885 "no match", the callouts do occur, and that items such as (*COMMIT) and (*MARK)
1886 are considered at every possible starting position in the subject string. If
1887 PCRE_NO_START_OPTIMIZE is set at compile time, it cannot be unset at matching
1888 time. The use of PCRE_NO_START_OPTIMIZE at matching time (that is, passing it
1889 to \fBpcre_exec()\fP) disables JIT execution; in this situation, matching is
1890 always done using interpretively.
1891 .P
1892 Setting PCRE_NO_START_OPTIMIZE can change the outcome of a matching operation.
1893 Consider the pattern
1894 .sp
1895 (*COMMIT)ABC
1896 .sp
1897 When this is compiled, PCRE records the fact that a match must start with the
1898 character "A". Suppose the subject string is "DEFABC". The start-up
1899 optimization scans along the subject, finds "A" and runs the first match
1900 attempt from there. The (*COMMIT) item means that the pattern must match the
1901 current starting position, which in this case, it does. However, if the same
1902 match is run with PCRE_NO_START_OPTIMIZE set, the initial scan along the
1903 subject string does not happen. The first match attempt is run starting from
1904 "D" and when this fails, (*COMMIT) prevents any further matches being tried, so
1905 the overall result is "no match". If the pattern is studied, more start-up
1906 optimizations may be used. For example, a minimum length for the subject may be
1907 recorded. Consider the pattern
1908 .sp
1909 (*MARK:A)(X|Y)
1910 .sp
1911 The minimum length for a match is one character. If the subject is "ABC", there
1912 will be attempts to match "ABC", "BC", "C", and then finally an empty string.
1913 If the pattern is studied, the final attempt does not take place, because PCRE
1914 knows that the subject is too short, and so the (*MARK) is never encountered.
1915 In this case, studying the pattern does not affect the overall match result,
1916 which is still "no match", but it does affect the auxiliary information that is
1917 returned.
1918 .sp
1919 PCRE_NO_UTF8_CHECK
1920 .sp
1921 When PCRE_UTF8 is set at compile time, the validity of the subject as a UTF-8
1922 string is automatically checked when \fBpcre_exec()\fP is subsequently called.
1923 The entire string is checked before any other processing takes place. The value
1924 of \fIstartoffset\fP is also checked to ensure that it points to the start of a
1925 UTF-8 character. There is a discussion about the
1926 .\" HTML <a href="pcreunicode.html#utf8strings">
1927 .\" </a>
1928 validity of UTF-8 strings
1929 .\"
1930 in the
1931 .\" HREF
1932 \fBpcreunicode\fP
1933 .\"
1934 page. If an invalid sequence of bytes is found, \fBpcre_exec()\fP returns the
1935 error PCRE_ERROR_BADUTF8 or, if PCRE_PARTIAL_HARD is set and the problem is a
1936 truncated character at the end of the subject, PCRE_ERROR_SHORTUTF8. In both
1937 cases, information about the precise nature of the error may also be returned
1938 (see the descriptions of these errors in the section entitled \fIError return
1939 values from\fP \fBpcre_exec()\fP
1940 .\" HTML <a href="#errorlist">
1941 .\" </a>
1942 below).
1943 .\"
1944 If \fIstartoffset\fP contains a value that does not point to the start of a
1945 UTF-8 character (or to the end of the subject), PCRE_ERROR_BADUTF8_OFFSET is
1946 returned.
1947 .P
1948 If you already know that your subject is valid, and you want to skip these
1949 checks for performance reasons, you can set the PCRE_NO_UTF8_CHECK option when
1950 calling \fBpcre_exec()\fP. You might want to do this for the second and
1951 subsequent calls to \fBpcre_exec()\fP if you are making repeated calls to find
1952 all the matches in a single subject string. However, you should be sure that
1953 the value of \fIstartoffset\fP points to the start of a character (or the end
1954 of the subject). When PCRE_NO_UTF8_CHECK is set, the effect of passing an
1955 invalid string as a subject or an invalid value of \fIstartoffset\fP is
1956 undefined. Your program may crash or loop.
1957 .sp
1958 PCRE_PARTIAL_HARD
1959 PCRE_PARTIAL_SOFT
1960 .sp
1961 These options turn on the partial matching feature. For backwards
1962 compatibility, PCRE_PARTIAL is a synonym for PCRE_PARTIAL_SOFT. A partial match
1963 occurs if the end of the subject string is reached successfully, but there are
1964 not enough subject characters to complete the match. If this happens when
1965 PCRE_PARTIAL_SOFT (but not PCRE_PARTIAL_HARD) is set, matching continues by
1966 testing any remaining alternatives. Only if no complete match can be found is
1967 PCRE_ERROR_PARTIAL returned instead of PCRE_ERROR_NOMATCH. In other words,
1968 PCRE_PARTIAL_SOFT says that the caller is prepared to handle a partial match,
1969 but only if no complete match can be found.
1970 .P
1971 If PCRE_PARTIAL_HARD is set, it overrides PCRE_PARTIAL_SOFT. In this case, if a
1972 partial match is found, \fBpcre_exec()\fP immediately returns
1973 PCRE_ERROR_PARTIAL, without considering any other alternatives. In other words,
1974 when PCRE_PARTIAL_HARD is set, a partial match is considered to be more
1975 important that an alternative complete match.
1976 .P
1977 In both cases, the portion of the string that was inspected when the partial
1978 match was found is set as the first matching string. There is a more detailed
1979 discussion of partial and multi-segment matching, with examples, in the
1980 .\" HREF
1981 \fBpcrepartial\fP
1982 .\"
1983 documentation.
1984 .
1985 .
1986 .SS "The string to be matched by \fBpcre_exec()\fP"
1987 .rs
1988 .sp
1989 The subject string is passed to \fBpcre_exec()\fP as a pointer in
1990 \fIsubject\fP, a length in \fIlength\fP, and a starting offset in
1991 \fIstartoffset\fP. The units for \fIlength\fP and \fIstartoffset\fP are bytes
1992 for the 8-bit library, 16-bit data items for the 16-bit library, and 32-bit
1993 data items for the 32-bit library.
1994 .P
1995 If \fIstartoffset\fP is negative or greater than the length of the subject,
1996 \fBpcre_exec()\fP returns PCRE_ERROR_BADOFFSET. When the starting offset is
1997 zero, the search for a match starts at the beginning of the subject, and this
1998 is by far the most common case. In UTF-8 or UTF-16 mode, the offset must point
1999 to the start of a character, or the end of the subject (in UTF-32 mode, one
2000 data unit equals one character, so all offsets are valid). Unlike the pattern
2001 string, the subject may contain binary zeroes.
2002 .P
2003 A non-zero starting offset is useful when searching for another match in the
2004 same subject by calling \fBpcre_exec()\fP again after a previous success.
2005 Setting \fIstartoffset\fP differs from just passing over a shortened string and
2006 setting PCRE_NOTBOL in the case of a pattern that begins with any kind of
2007 lookbehind. For example, consider the pattern
2008 .sp
2009 \eBiss\eB
2010 .sp
2011 which finds occurrences of "iss" in the middle of words. (\eB matches only if
2012 the current position in the subject is not a word boundary.) When applied to
2013 the string "Mississipi" the first call to \fBpcre_exec()\fP finds the first
2014 occurrence. If \fBpcre_exec()\fP is called again with just the remainder of the
2015 subject, namely "issipi", it does not match, because \eB is always false at the
2016 start of the subject, which is deemed to be a word boundary. However, if
2017 \fBpcre_exec()\fP is passed the entire string again, but with \fIstartoffset\fP
2018 set to 4, it finds the second occurrence of "iss" because it is able to look
2019 behind the starting point to discover that it is preceded by a letter.
2020 .P
2021 Finding all the matches in a subject is tricky when the pattern can match an
2022 empty string. It is possible to emulate Perl's /g behaviour by first trying the
2023 match again at the same offset, with the PCRE_NOTEMPTY_ATSTART and
2024 PCRE_ANCHORED options, and then if that fails, advancing the starting offset
2025 and trying an ordinary match again. There is some code that demonstrates how to
2026 do this in the
2027 .\" HREF
2028 \fBpcredemo\fP
2029 .\"
2030 sample program. In the most general case, you have to check to see if the
2031 newline convention recognizes CRLF as a newline, and if so, and the current
2032 character is CR followed by LF, advance the starting offset by two characters
2033 instead of one.
2034 .P
2035 If a non-zero starting offset is passed when the pattern is anchored, one
2036 attempt to match at the given offset is made. This can only succeed if the
2037 pattern does not require the match to be at the start of the subject.
2038 .
2039 .
2040 .SS "How \fBpcre_exec()\fP returns captured substrings"
2041 .rs
2042 .sp
2043 In general, a pattern matches a certain portion of the subject, and in
2044 addition, further substrings from the subject may be picked out by parts of the
2045 pattern. Following the usage in Jeffrey Friedl's book, this is called
2046 "capturing" in what follows, and the phrase "capturing subpattern" is used for
2047 a fragment of a pattern that picks out a substring. PCRE supports several other
2048 kinds of parenthesized subpattern that do not cause substrings to be captured.
2049 .P
2050 Captured substrings are returned to the caller via a vector of integers whose
2051 address is passed in \fIovector\fP. The number of elements in the vector is
2052 passed in \fIovecsize\fP, which must be a non-negative number. \fBNote\fP: this
2053 argument is NOT the size of \fIovector\fP in bytes.
2054 .P
2055 The first two-thirds of the vector is used to pass back captured substrings,
2056 each substring using a pair of integers. The remaining third of the vector is
2057 used as workspace by \fBpcre_exec()\fP while matching capturing subpatterns,
2058 and is not available for passing back information. The number passed in
2059 \fIovecsize\fP should always be a multiple of three. If it is not, it is
2060 rounded down.
2061 .P
2062 When a match is successful, information about captured substrings is returned
2063 in pairs of integers, starting at the beginning of \fIovector\fP, and
2064 continuing up to two-thirds of its length at the most. The first element of
2065 each pair is set to the offset of the first character in a substring, and the
2066 second is set to the offset of the first character after the end of a
2067 substring. These values are always data unit offsets, even in UTF mode. They
2068 are byte offsets in the 8-bit library, 16-bit data item offsets in the 16-bit
2069 library, and 32-bit data item offsets in the 32-bit library. \fBNote\fP: they
2070 are not character counts.
2071 .P
2072 The first pair of integers, \fIovector[0]\fP and \fIovector[1]\fP, identify the
2073 portion of the subject string matched by the entire pattern. The next pair is
2074 used for the first capturing subpattern, and so on. The value returned by
2075 \fBpcre_exec()\fP is one more than the highest numbered pair that has been set.
2076 For example, if two substrings have been captured, the returned value is 3. If
2077 there are no capturing subpatterns, the return value from a successful match is
2078 1, indicating that just the first pair of offsets has been set.
2079 .P
2080 If a capturing subpattern is matched repeatedly, it is the last portion of the
2081 string that it matched that is returned.
2082 .P
2083 If the vector is too small to hold all the captured substring offsets, it is
2084 used as far as possible (up to two-thirds of its length), and the function
2085 returns a value of zero. If neither the actual string matched nor any captured
2086 substrings are of interest, \fBpcre_exec()\fP may be called with \fIovector\fP
2087 passed as NULL and \fIovecsize\fP as zero. However, if the pattern contains
2088 back references and the \fIovector\fP is not big enough to remember the related
2089 substrings, PCRE has to get additional memory for use during matching. Thus it
2090 is usually advisable to supply an \fIovector\fP of reasonable size.
2091 .P
2092 There are some cases where zero is returned (indicating vector overflow) when
2093 in fact the vector is exactly the right size for the final match. For example,
2094 consider the pattern
2095 .sp
2096 (a)(?:(b)c|bd)
2097 .sp
2098 If a vector of 6 elements (allowing for only 1 captured substring) is given
2099 with subject string "abd", \fBpcre_exec()\fP will try to set the second
2100 captured string, thereby recording a vector overflow, before failing to match
2101 "c" and backing up to try the second alternative. The zero return, however,
2102 does correctly indicate that the maximum number of slots (namely 2) have been
2103 filled. In similar cases where there is temporary overflow, but the final
2104 number of used slots is actually less than the maximum, a non-zero value is
2105 returned.
2106 .P
2107 The \fBpcre_fullinfo()\fP function can be used to find out how many capturing
2108 subpatterns there are in a compiled pattern. The smallest size for
2109 \fIovector\fP that will allow for \fIn\fP captured substrings, in addition to
2110 the offsets of the substring matched by the whole pattern, is (\fIn\fP+1)*3.
2111 .P
2112 It is possible for capturing subpattern number \fIn+1\fP to match some part of
2113 the subject when subpattern \fIn\fP has not been used at all. For example, if
2114 the string "abc" is matched against the pattern (a|(z))(bc) the return from the
2115 function is 4, and subpatterns 1 and 3 are matched, but 2 is not. When this
2116 happens, both values in the offset pairs corresponding to unused subpatterns
2117 are set to -1.
2118 .P
2119 Offset values that correspond to unused subpatterns at the end of the
2120 expression are also set to -1. For example, if the string "abc" is matched
2121 against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not matched. The
2122 return from the function is 2, because the highest used capturing subpattern
2123 number is 1, and the offsets for for the second and third capturing subpatterns
2124 (assuming the vector is large enough, of course) are set to -1.
2125 .P
2126 \fBNote\fP: Elements in the first two-thirds of \fIovector\fP that do not
2127 correspond to capturing parentheses in the pattern are never changed. That is,
2128 if a pattern contains \fIn\fP capturing parentheses, no more than
2129 \fIovector[0]\fP to \fIovector[2n+1]\fP are set by \fBpcre_exec()\fP. The other
2130 elements (in the first two-thirds) retain whatever values they previously had.
2131 .P
2132 Some convenience functions are provided for extracting the captured substrings
2133 as separate strings. These are described below.
2134 .
2135 .
2136 .\" HTML <a name="errorlist"></a>
2137 .SS "Error return values from \fBpcre_exec()\fP"
2138 .rs
2139 .sp
2140 If \fBpcre_exec()\fP fails, it returns a negative number. The following are
2141 defined in the header file:
2142 .sp
2143 PCRE_ERROR_NOMATCH (-1)
2144 .sp
2145 The subject string did not match the pattern.
2146 .sp
2147 PCRE_ERROR_NULL (-2)
2148 .sp
2149 Either \fIcode\fP or \fIsubject\fP was passed as NULL, or \fIovector\fP was
2150 NULL and \fIovecsize\fP was not zero.
2151 .sp
2152 PCRE_ERROR_BADOPTION (-3)
2153 .sp
2154 An unrecognized bit was set in the \fIoptions\fP argument.
2155 .sp
2156 PCRE_ERROR_BADMAGIC (-4)
2157 .sp
2158 PCRE stores a 4-byte "magic number" at the start of the compiled code, to catch
2159 the case when it is passed a junk pointer and to detect when a pattern that was
2160 compiled in an environment of one endianness is run in an environment with the
2161 other endianness. This is the error that PCRE gives when the magic number is
2162 not present.
2163 .sp
2164 PCRE_ERROR_UNKNOWN_OPCODE (-5)
2165 .sp
2166 While running the pattern match, an unknown item was encountered in the
2167 compiled pattern. This error could be caused by a bug in PCRE or by overwriting
2168 of the compiled pattern.
2169 .sp
2170 PCRE_ERROR_NOMEMORY (-6)
2171 .sp
2172 If a pattern contains back references, but the \fIovector\fP that is passed to
2173 \fBpcre_exec()\fP is not big enough to remember the referenced substrings, PCRE
2174 gets a block of memory at the start of matching to use for this purpose. If the
2175 call via \fBpcre_malloc()\fP fails, this error is given. The memory is
2176 automatically freed at the end of matching.
2177 .P
2178 This error is also given if \fBpcre_stack_malloc()\fP fails in
2179 \fBpcre_exec()\fP. This can happen only when PCRE has been compiled with
2180 \fB--disable-stack-for-recursion\fP.
2181 .sp
2182 PCRE_ERROR_NOSUBSTRING (-7)
2183 .sp
2184 This error is used by the \fBpcre_copy_substring()\fP,
2185 \fBpcre_get_substring()\fP, and \fBpcre_get_substring_list()\fP functions (see
2186 below). It is never returned by \fBpcre_exec()\fP.
2187 .sp
2188 PCRE_ERROR_MATCHLIMIT (-8)
2189 .sp
2190 The backtracking limit, as specified by the \fImatch_limit\fP field in a
2191 \fBpcre_extra\fP structure (or defaulted) was reached. See the description
2192 above.
2193 .sp
2194 PCRE_ERROR_CALLOUT (-9)
2195 .sp
2196 This error is never generated by \fBpcre_exec()\fP itself. It is provided for
2197 use by callout functions that want to yield a distinctive error code. See the
2198 .\" HREF
2199 \fBpcrecallout\fP
2200 .\"
2201 documentation for details.
2202 .sp
2203 PCRE_ERROR_BADUTF8 (-10)
2204 .sp
2205 A string that contains an invalid UTF-8 byte sequence was passed as a subject,
2206 and the PCRE_NO_UTF8_CHECK option was not set. If the size of the output vector
2207 (\fIovecsize\fP) is at least 2, the byte offset to the start of the the invalid
2208 UTF-8 character is placed in the first element, and a reason code is placed in
2209 the second element. The reason codes are listed in the
2210 .\" HTML <a href="#badutf8reasons">
2211 .\" </a>
2212 following section.
2213 .\"
2214 For backward compatibility, if PCRE_PARTIAL_HARD is set and the problem is a
2215 truncated UTF-8 character at the end of the subject (reason codes 1 to 5),
2216 PCRE_ERROR_SHORTUTF8 is returned instead of PCRE_ERROR_BADUTF8.
2217 .sp
2218 PCRE_ERROR_BADUTF8_OFFSET (-11)
2219 .sp
2220 The UTF-8 byte sequence that was passed as a subject was checked and found to
2221 be valid (the PCRE_NO_UTF8_CHECK option was not set), but the value of
2222 \fIstartoffset\fP did not point to the beginning of a UTF-8 character or the
2223 end of the subject.
2224 .sp
2225 PCRE_ERROR_PARTIAL (-12)
2226 .sp
2227 The subject string did not match, but it did match partially. See the
2228 .\" HREF
2229 \fBpcrepartial\fP
2230 .\"
2231 documentation for details of partial matching.
2232 .sp
2233 PCRE_ERROR_BADPARTIAL (-13)
2234 .sp
2235 This code is no longer in use. It was formerly returned when the PCRE_PARTIAL
2236 option was used with a compiled pattern containing items that were not
2237 supported for partial matching. From release 8.00 onwards, there are no
2238 restrictions on partial matching.
2239 .sp
2240 PCRE_ERROR_INTERNAL (-14)
2241 .sp
2242 An unexpected internal error has occurred. This error could be caused by a bug
2243 in PCRE or by overwriting of the compiled pattern.
2244 .sp
2245 PCRE_ERROR_BADCOUNT (-15)
2246 .sp
2247 This error is given if the value of the \fIovecsize\fP argument is negative.
2248 .sp
2249 PCRE_ERROR_RECURSIONLIMIT (-21)
2250 .sp
2251 The internal recursion limit, as specified by the \fImatch_limit_recursion\fP
2252 field in a \fBpcre_extra\fP structure (or defaulted) was reached. See the
2253 description above.
2254 .sp
2255 PCRE_ERROR_BADNEWLINE (-23)
2256 .sp
2257 An invalid combination of PCRE_NEWLINE_\fIxxx\fP options was given.
2258 .sp
2259 PCRE_ERROR_BADOFFSET (-24)
2260 .sp
2261 The value of \fIstartoffset\fP was negative or greater than the length of the
2262 subject, that is, the value in \fIlength\fP.
2263 .sp
2264 PCRE_ERROR_SHORTUTF8 (-25)
2265 .sp
2266 This error is returned instead of PCRE_ERROR_BADUTF8 when the subject string
2267 ends with a truncated UTF-8 character and the PCRE_PARTIAL_HARD option is set.
2268 Information about the failure is returned as for PCRE_ERROR_BADUTF8. It is in
2269 fact sufficient to detect this case, but this special error code for
2270 PCRE_PARTIAL_HARD precedes the implementation of returned information; it is
2271 retained for backwards compatibility.
2272 .sp
2273 PCRE_ERROR_RECURSELOOP (-26)
2274 .sp
2275 This error is returned when \fBpcre_exec()\fP detects a recursion loop within
2276 the pattern. Specifically, it means that either the whole pattern or a
2277 subpattern has been called recursively for the second time at the same position
2278 in the subject string. Some simple patterns that might do this are detected and
2279 faulted at compile time, but more complicated cases, in particular mutual
2280 recursions between two different subpatterns, cannot be detected until run
2281 time.
2282 .sp
2283 PCRE_ERROR_JIT_STACKLIMIT (-27)
2284 .sp
2285 This error is returned when a pattern that was successfully studied using a
2286 JIT compile option is being matched, but the memory available for the
2287 just-in-time processing stack is not large enough. See the
2288 .\" HREF
2289 \fBpcrejit\fP
2290 .\"
2291 documentation for more details.
2292 .sp
2293 PCRE_ERROR_BADMODE (-28)
2294 .sp
2295 This error is given if a pattern that was compiled by the 8-bit library is
2296 passed to a 16-bit or 32-bit library function, or vice versa.
2297 .sp
2298 PCRE_ERROR_BADENDIANNESS (-29)
2299 .sp
2300 This error is given if a pattern that was compiled and saved is reloaded on a
2301 host with different endianness. The utility function
2302 \fBpcre_pattern_to_host_byte_order()\fP can be used to convert such a pattern
2303 so that it runs on the new host.
2304 .sp
2305 PCRE_ERROR_JIT_BADOPTION
2306 .sp
2307 This error is returned when a pattern that was successfully studied using a JIT
2308 compile option is being matched, but the matching mode (partial or complete
2309 match) does not correspond to any JIT compilation mode. When the JIT fast path
2310 function is used, this error may be also given for invalid options. See the
2311 .\" HREF
2312 \fBpcrejit\fP
2313 .\"
2314 documentation for more details.
2315 .sp
2316 PCRE_ERROR_BADLENGTH (-32)
2317 .sp
2318 This error is given if \fBpcre_exec()\fP is called with a negative value for
2319 the \fIlength\fP argument.
2320 .P
2321 Error numbers -16 to -20, -22, and 30 are not used by \fBpcre_exec()\fP.
2322 .
2323 .
2324 .\" HTML <a name="badutf8reasons"></a>
2325 .SS "Reason codes for invalid UTF-8 strings"
2326 .rs
2327 .sp
2328 This section applies only to the 8-bit library. The corresponding information
2329 for the 16-bit and 32-bit libraries is given in the
2330 .\" HREF
2331 \fBpcre16\fP
2332 .\"
2333 and
2334 .\" HREF
2335 \fBpcre32\fP
2336 .\"
2337 pages.
2338 .P
2339 When \fBpcre_exec()\fP returns either PCRE_ERROR_BADUTF8 or
2340 PCRE_ERROR_SHORTUTF8, and the size of the output vector (\fIovecsize\fP) is at
2341 least 2, the offset of the start of the invalid UTF-8 character is placed in
2342 the first output vector element (\fIovector[0]\fP) and a reason code is placed
2343 in the second element (\fIovector[1]\fP). The reason codes are given names in
2344 the \fBpcre.h\fP header file:
2345 .sp
2346 PCRE_UTF8_ERR1
2347 PCRE_UTF8_ERR2
2348 PCRE_UTF8_ERR3
2349 PCRE_UTF8_ERR4
2350 PCRE_UTF8_ERR5
2351 .sp
2352 The string ends with a truncated UTF-8 character; the code specifies how many
2353 bytes are missing (1 to 5). Although RFC 3629 restricts UTF-8 characters to be
2354 no longer than 4 bytes, the encoding scheme (originally defined by RFC 2279)
2355 allows for up to 6 bytes, and this is checked first; hence the possibility of
2356 4 or 5 missing bytes.
2357 .sp
2358 PCRE_UTF8_ERR6
2359 PCRE_UTF8_ERR7
2360 PCRE_UTF8_ERR8
2361 PCRE_UTF8_ERR9
2362 PCRE_UTF8_ERR10
2363 .sp
2364 The two most significant bits of the 2nd, 3rd, 4th, 5th, or 6th byte of the
2365 character do not have the binary value 0b10 (that is, either the most
2366 significant bit is 0, or the next bit is 1).
2367 .sp
2368 PCRE_UTF8_ERR11
2369 PCRE_UTF8_ERR12
2370 .sp
2371 A character that is valid by the RFC 2279 rules is either 5 or 6 bytes long;
2372 these code points are excluded by RFC 3629.
2373 .sp
2374 PCRE_UTF8_ERR13
2375 .sp
2376 A 4-byte character has a value greater than 0x10fff; these code points are
2377 excluded by RFC 3629.
2378 .sp
2379 PCRE_UTF8_ERR14
2380 .sp
2381 A 3-byte character has a value in the range 0xd800 to 0xdfff; this range of
2382 code points are reserved by RFC 3629 for use with UTF-16, and so are excluded
2383 from UTF-8.
2384 .sp
2385 PCRE_UTF8_ERR15
2386 PCRE_UTF8_ERR16
2387 PCRE_UTF8_ERR17
2388 PCRE_UTF8_ERR18
2389 PCRE_UTF8_ERR19
2390 .sp
2391 A 2-, 3-, 4-, 5-, or 6-byte character is "overlong", that is, it codes for a
2392 value that can be represented by fewer bytes, which is invalid. For example,
2393 the two bytes 0xc0, 0xae give the value 0x2e, whose correct coding uses just
2394 one byte.
2395 .sp
2396 PCRE_UTF8_ERR20
2397 .sp
2398 The two most significant bits of the first byte of a character have the binary
2399 value 0b10 (that is, the most significant bit is 1 and the second is 0). Such a
2400 byte can only validly occur as the second or subsequent byte of a multi-byte
2401 character.
2402 .sp
2403 PCRE_UTF8_ERR21
2404 .sp
2405 The first byte of a character has the value 0xfe or 0xff. These values can
2406 never occur in a valid UTF-8 string.
2407 .sp
2408 PCRE_UTF8_ERR22
2409 .sp
2410 This error code was formerly used when the presence of a so-called
2411 "non-character" caused an error. Unicode corrigendum #9 makes it clear that
2412 such characters should not cause a string to be rejected, and so this code is
2413 no longer in use and is never returned.
2414 .
2415 .
2416 .SH "EXTRACTING CAPTURED SUBSTRINGS BY NUMBER"
2417 .rs
2418 .sp
2419 .nf
2420 .B int pcre_copy_substring(const char *\fIsubject\fP, int *\fIovector\fP,
2421 .B " int \fIstringcount\fP, int \fIstringnumber\fP, char *\fIbuffer\fP,"
2422 .B " int \fIbuffersize\fP);"
2423 .sp
2424 .B int pcre_get_substring(const char *\fIsubject\fP, int *\fIovector\fP,
2425 .B " int \fIstringcount\fP, int \fIstringnumber\fP,"
2426 .B " const char **\fIstringptr\fP);"
2427 .sp
2428 .B int pcre_get_substring_list(const char *\fIsubject\fP,
2429 .B " int *\fIovector\fP, int \fIstringcount\fP, const char ***\fIlistptr\fP);"
2430 .fi
2431 .PP
2432 Captured substrings can be accessed directly by using the offsets returned by
2433 \fBpcre_exec()\fP in \fIovector\fP. For convenience, the functions
2434 \fBpcre_copy_substring()\fP, \fBpcre_get_substring()\fP, and
2435 \fBpcre_get_substring_list()\fP are provided for extracting captured substrings
2436 as new, separate, zero-terminated strings. These functions identify substrings
2437 by number. The next section describes functions for extracting named
2438 substrings.
2439 .P
2440 A substring that contains a binary zero is correctly extracted and has a
2441 further zero added on the end, but the result is not, of course, a C string.
2442 However, you can process such a string by referring to the length that is
2443 returned by \fBpcre_copy_substring()\fP and \fBpcre_get_substring()\fP.
2444 Unfortunately, the interface to \fBpcre_get_substring_list()\fP is not adequate
2445 for handling strings containing binary zeros, because the end of the final
2446 string is not independently indicated.
2447 .P
2448 The first three arguments are the same for all three of these functions:
2449 \fIsubject\fP is the subject string that has just been successfully matched,
2450 \fIovector\fP is a pointer to the vector of integer offsets that was passed to
2451 \fBpcre_exec()\fP, and \fIstringcount\fP is the number of substrings that were
2452 captured by the match, including the substring that matched the entire regular
2453 expression. This is the value returned by \fBpcre_exec()\fP if it is greater
2454 than zero. If \fBpcre_exec()\fP returned zero, indicating that it ran out of
2455 space in \fIovector\fP, the value passed as \fIstringcount\fP should be the
2456 number of elements in the vector divided by three.
2457 .P
2458 The functions \fBpcre_copy_substring()\fP and \fBpcre_get_substring()\fP
2459 extract a single substring, whose number is given as \fIstringnumber\fP. A
2460 value of zero extracts the substring that matched the entire pattern, whereas
2461 higher values extract the captured substrings. For \fBpcre_copy_substring()\fP,
2462 the string is placed in \fIbuffer\fP, whose length is given by
2463 \fIbuffersize\fP, while for \fBpcre_get_substring()\fP a new block of memory is
2464 obtained via \fBpcre_malloc\fP, and its address is returned via
2465 \fIstringptr\fP. The yield of the function is the length of the string, not
2466 including the terminating zero, or one of these error codes:
2467 .sp
2468 PCRE_ERROR_NOMEMORY (-6)
2469 .sp
2470 The buffer was too small for \fBpcre_copy_substring()\fP, or the attempt to get
2471 memory failed for \fBpcre_get_substring()\fP.
2472 .sp
2473 PCRE_ERROR_NOSUBSTRING (-7)
2474 .sp
2475 There is no substring whose number is \fIstringnumber\fP.
2476 .P
2477 The \fBpcre_get_substring_list()\fP function extracts all available substrings
2478 and builds a list of pointers to them. All this is done in a single block of
2479 memory that is obtained via \fBpcre_malloc\fP. The address of the memory block
2480 is returned via \fIlistptr\fP, which is also the start of the list of string
2481 pointers. The end of the list is marked by a NULL pointer. The yield of the
2482 function is zero if all went well, or the error code
2483 .sp
2484 PCRE_ERROR_NOMEMORY (-6)
2485 .sp
2486 if the attempt to get the memory block failed.
2487 .P
2488 When any of these functions encounter a substring that is unset, which can
2489 happen when capturing subpattern number \fIn+1\fP matches some part of the
2490 subject, but subpattern \fIn\fP has not been used at all, they return an empty
2491 string. This can be distinguished from a genuine zero-length substring by
2492 inspecting the appropriate offset in \fIovector\fP, which is negative for unset
2493 substrings.
2494 .P
2495 The two convenience functions \fBpcre_free_substring()\fP and
2496 \fBpcre_free_substring_list()\fP can be used to free the memory returned by
2497 a previous call of \fBpcre_get_substring()\fP or
2498 \fBpcre_get_substring_list()\fP, respectively. They do nothing more than call
2499 the function pointed to by \fBpcre_free\fP, which of course could be called
2500 directly from a C program. However, PCRE is used in some situations where it is
2501 linked via a special interface to another programming language that cannot use
2502 \fBpcre_free\fP directly; it is for these cases that the functions are
2503 provided.
2504 .
2505 .
2506 .SH "EXTRACTING CAPTURED SUBSTRINGS BY NAME"
2507 .rs
2508 .sp
2509 .nf
2510 .B int pcre_get_stringnumber(const pcre *\fIcode\fP,
2511 .B " const char *\fIname\fP);"
2512 .sp
2513 .B int pcre_copy_named_substring(const pcre *\fIcode\fP,
2514 .B " const char *\fIsubject\fP, int *\fIovector\fP,"
2515 .B " int \fIstringcount\fP, const char *\fIstringname\fP,"
2516 .B " char *\fIbuffer\fP, int \fIbuffersize\fP);"
2517 .sp
2518 .B int pcre_get_named_substring(const pcre *\fIcode\fP,
2519 .B " const char *\fIsubject\fP, int *\fIovector\fP,"
2520 .B " int \fIstringcount\fP, const char *\fIstringname\fP,"
2521 .B " const char **\fIstringptr\fP);"
2522 .fi
2523 .PP
2524 To extract a substring by name, you first have to find associated number.
2525 For example, for this pattern
2526 .sp
2527 (a+)b(?<xxx>\ed+)...
2528 .sp
2529 the number of the subpattern called "xxx" is 2. If the name is known to be
2530 unique (PCRE_DUPNAMES was not set), you can find the number from the name by
2531 calling \fBpcre_get_stringnumber()\fP. The first argument is the compiled
2532 pattern, and the second is the name. The yield of the function is the
2533 subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no subpattern of
2534 that name.
2535 .P
2536 Given the number, you can extract the substring directly, or use one of the
2537 functions described in the previous section. For convenience, there are also
2538 two functions that do the whole job.
2539 .P
2540 Most of the arguments of \fBpcre_copy_named_substring()\fP and
2541 \fBpcre_get_named_substring()\fP are the same as those for the similarly named
2542 functions that extract by number. As these are described in the previous
2543 section, they are not re-described here. There are just two differences:
2544 .P
2545 First, instead of a substring number, a substring name is given. Second, there
2546 is an extra argument, given at the start, which is a pointer to the compiled
2547 pattern. This is needed in order to gain access to the name-to-number
2548 translation table.
2549 .P
2550 These functions call \fBpcre_get_stringnumber()\fP, and if it succeeds, they
2551 then call \fBpcre_copy_substring()\fP or \fBpcre_get_substring()\fP, as
2552 appropriate. \fBNOTE:\fP If PCRE_DUPNAMES is set and there are duplicate names,
2553 the behaviour may not be what you want (see the next section).
2554 .P
2555 \fBWarning:\fP If the pattern uses the (?| feature to set up multiple
2556 subpatterns with the same number, as described in the
2557 .\" HTML <a href="pcrepattern.html#dupsubpatternnumber">
2558 .\" </a>
2559 section on duplicate subpattern numbers
2560 .\"
2561 in the
2562 .\" HREF
2563 \fBpcrepattern\fP
2564 .\"
2565 page, you cannot use names to distinguish the different subpatterns, because
2566 names are not included in the compiled code. The matching process uses only
2567 numbers. For this reason, the use of different names for subpatterns of the
2568 same number causes an error at compile time.
2569 .
2570 .
2571 .SH "DUPLICATE SUBPATTERN NAMES"
2572 .rs
2573 .sp
2574 .nf
2575 .B int pcre_get_stringtable_entries(const pcre *\fIcode\fP,
2576 .B " const char *\fIname\fP, char **\fIfirst\fP, char **\fIlast\fP);"
2577 .fi
2578 .PP
2579 When a pattern is compiled with the PCRE_DUPNAMES option, names for subpatterns
2580 are not required to be unique. (Duplicate names are always allowed for
2581 subpatterns with the same number, created by using the (?| feature. Indeed, if
2582 such subpatterns are named, they are required to use the same names.)
2583 .P
2584 Normally, patterns with duplicate names are such that in any one match, only
2585 one of the named subpatterns participates. An example is shown in the
2586 .\" HREF
2587 \fBpcrepattern\fP
2588 .\"
2589 documentation.
2590 .P
2591 When duplicates are present, \fBpcre_copy_named_substring()\fP and
2592 \fBpcre_get_named_substring()\fP return the first substring corresponding to
2593 the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING (-7) is
2594 returned; no data is returned. The \fBpcre_get_stringnumber()\fP function
2595 returns one of the numbers that are associated with the name, but it is not
2596 defined which it is.
2597 .P
2598 If you want to get full details of all captured substrings for a given name,
2599 you must use the \fBpcre_get_stringtable_entries()\fP function. The first
2600 argument is the compiled pattern, and the second is the name. The third and
2601 fourth are pointers to variables which are updated by the function. After it
2602 has run, they point to the first and last entries in the name-to-number table
2603 for the given name. The function itself returns the length of each entry, or
2604 PCRE_ERROR_NOSUBSTRING (-7) if there are none. The format of the table is
2605 described above in the section entitled \fIInformation about a pattern\fP
2606 .\" HTML <a href="#infoaboutpattern">
2607 .\" </a>
2608 above.
2609 .\"
2610 Given all the relevant entries for the name, you can extract each of their
2611 numbers, and hence the captured data, if any.
2612 .
2613 .
2614 .SH "FINDING ALL POSSIBLE MATCHES"
2615 .rs
2616 .sp
2617 The traditional matching function uses a similar algorithm to Perl, which stops
2618 when it finds the first match, starting at a given point in the subject. If you
2619 want to find all possible matches, or the longest possible match, consider
2620 using the alternative matching function (see below) instead. If you cannot use
2621 the alternative function, but still need to find all possible matches, you
2622 can kludge it up by making use of the callout facility, which is described in
2623 the
2624 .\" HREF
2625 \fBpcrecallout\fP
2626 .\"
2627 documentation.
2628 .P
2629 What you have to do is to insert a callout right at the end of the pattern.
2630 When your callout function is called, extract and save the current matched
2631 substring. Then return 1, which forces \fBpcre_exec()\fP to backtrack and try
2632 other alternatives. Ultimately, when it runs out of matches, \fBpcre_exec()\fP
2633 will yield PCRE_ERROR_NOMATCH.
2634 .
2635 .
2636 .SH "OBTAINING AN ESTIMATE OF STACK USAGE"
2637 .rs
2638 .sp
2639 Matching certain patterns using \fBpcre_exec()\fP can use a lot of process
2640 stack, which in certain environments can be rather limited in size. Some users
2641 find it helpful to have an estimate of the amount of stack that is used by
2642 \fBpcre_exec()\fP, to help them set recursion limits, as described in the
2643 .\" HREF
2644 \fBpcrestack\fP
2645 .\"
2646 documentation. The estimate that is output by \fBpcretest\fP when called with
2647 the \fB-m\fP and \fB-C\fP options is obtained by calling \fBpcre_exec\fP with
2648 the values NULL, NULL, NULL, -999, and -999 for its first five arguments.
2649 .P
2650 Normally, if its first argument is NULL, \fBpcre_exec()\fP immediately returns
2651 the negative error code PCRE_ERROR_NULL, but with this special combination of
2652 arguments, it returns instead a negative number whose absolute value is the
2653 approximate stack frame size in bytes. (A negative number is used so that it is
2654 clear that no match has happened.) The value is approximate because in some
2655 cases, recursive calls to \fBpcre_exec()\fP occur when there are one or two
2656 additional variables on the stack.
2657 .P
2658 If PCRE has been compiled to use the heap instead of the stack for recursion,
2659 the value returned is the size of each block that is obtained from the heap.
2660 .
2661 .
2662 .\" HTML <a name="dfamatch"></a>
2663 .SH "MATCHING A PATTERN: THE ALTERNATIVE FUNCTION"
2664 .rs
2665 .sp
2666 .nf
2667 .B int pcre_dfa_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
2668 .B " const char *\fIsubject\fP, int \fIlength\fP, int \fIstartoffset\fP,"
2669 .B " int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP,"
2670 .B " int *\fIworkspace\fP, int \fIwscount\fP);"
2671 .fi
2672 .P
2673 The function \fBpcre_dfa_exec()\fP is called to match a subject string against
2674 a compiled pattern, using a matching algorithm that scans the subject string
2675 just once, and does not backtrack. This has different characteristics to the
2676 normal algorithm, and is not compatible with Perl. Some of the features of PCRE
2677 patterns are not supported. Nevertheless, there are times when this kind of
2678 matching can be useful. For a discussion of the two matching algorithms, and a
2679 list of features that \fBpcre_dfa_exec()\fP does not support, see the
2680 .\" HREF
2681 \fBpcrematching\fP
2682 .\"
2683 documentation.
2684 .P
2685 The arguments for the \fBpcre_dfa_exec()\fP function are the same as for
2686 \fBpcre_exec()\fP, plus two extras. The \fIovector\fP argument is used in a
2687 different way, and this is described below. The other common arguments are used
2688 in the same way as for \fBpcre_exec()\fP, so their description is not repeated
2689 here.
2690 .P
2691 The two additional arguments provide workspace for the function. The workspace
2692 vector should contain at least 20 elements. It is used for keeping track of
2693 multiple paths through the pattern tree. More workspace will be needed for
2694 patterns and subjects where there are a lot of potential matches.
2695 .P
2696 Here is an example of a simple call to \fBpcre_dfa_exec()\fP:
2697 .sp
2698 int rc;
2699 int ovector[10];
2700 int wspace[20];
2701 rc = pcre_dfa_exec(
2702 re, /* result of pcre_compile() */
2703 NULL, /* we didn't study the pattern */
2704 "some string", /* the subject string */
2705 11, /* the length of the subject string */
2706 0, /* start at offset 0 in the subject */
2707 0, /* default options */
2708 ovector, /* vector of integers for substring information */
2709 10, /* number of elements (NOT size in bytes) */
2710 wspace, /* working space vector */
2711 20); /* number of elements (NOT size in bytes) */
2712 .
2713 .SS "Option bits for \fBpcre_dfa_exec()\fP"
2714 .rs
2715 .sp
2716 The unused bits of the \fIoptions\fP argument for \fBpcre_dfa_exec()\fP must be
2717 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_\fIxxx\fP,
2718 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
2719 PCRE_NO_UTF8_CHECK, PCRE_BSR_ANYCRLF, PCRE_BSR_UNICODE, PCRE_NO_START_OPTIMIZE,
2720 PCRE_PARTIAL_HARD, PCRE_PARTIAL_SOFT, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART.
2721 All but the last four of these are exactly the same as for \fBpcre_exec()\fP,
2722 so their description is not repeated here.
2723 .sp
2724 PCRE_PARTIAL_HARD
2725 PCRE_PARTIAL_SOFT
2726 .sp
2727 These have the same general effect as they do for \fBpcre_exec()\fP, but the
2728 details are slightly different. When PCRE_PARTIAL_HARD is set for
2729 \fBpcre_dfa_exec()\fP, it returns PCRE_ERROR_PARTIAL if the end of the subject
2730 is reached and there is still at least one matching possibility that requires
2731 additional characters. This happens even if some complete matches have also
2732 been found. When PCRE_PARTIAL_SOFT is set, the return code PCRE_ERROR_NOMATCH
2733 is converted into PCRE_ERROR_PARTIAL if the end of the subject is reached,
2734 there have been no complete matches, but there is still at least one matching
2735 possibility. The portion of the string that was inspected when the longest
2736 partial match was found is set as the first matching string in both cases.
2737 There is a more detailed discussion of partial and multi-segment matching, with
2738 examples, in the
2739 .\" HREF
2740 \fBpcrepartial\fP
2741 .\"
2742 documentation.
2743 .sp
2744 PCRE_DFA_SHORTEST
2745 .sp
2746 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to stop as
2747 soon as it has found one match. Because of the way the alternative algorithm
2748 works, this is necessarily the shortest possible match at the first possible
2749 matching point in the subject string.
2750 .sp
2751 PCRE_DFA_RESTART
2752 .sp
2753 When \fBpcre_dfa_exec()\fP returns a partial match, it is possible to call it
2754 again, with additional subject characters, and have it continue with the same
2755 match. The PCRE_DFA_RESTART option requests this action; when it is set, the
2756 \fIworkspace\fP and \fIwscount\fP options must reference the same vector as
2757 before because data about the match so far is left in them after a partial
2758 match. There is more discussion of this facility in the
2759 .\" HREF
2760 \fBpcrepartial\fP
2761 .\"
2762 documentation.
2763 .
2764 .
2765 .SS "Successful returns from \fBpcre_dfa_exec()\fP"
2766 .rs
2767 .sp
2768 When \fBpcre_dfa_exec()\fP succeeds, it may have matched more than one
2769 substring in the subject. Note, however, that all the matches from one run of
2770 the function start at the same point in the subject. The shorter matches are
2771 all initial substrings of the longer matches. For example, if the pattern
2772 .sp
2773 <.*>
2774 .sp
2775 is matched against the string
2776 .sp
2777 This is <something> <something else> <something further> no more
2778 .sp
2779 the three matched strings are
2780 .sp
2781 <something>
2782 <something> <something else>
2783 <something> <something else> <something further>
2784 .sp
2785 On success, the yield of the function is a number greater than zero, which is
2786 the number of matched substrings. The substrings themselves are returned in
2787 \fIovector\fP. Each string uses two elements; the first is the offset to the
2788 start, and the second is the offset to the end. In fact, all the strings have
2789 the same start offset. (Space could have been saved by giving this only once,
2790 but it was decided to retain some compatibility with the way \fBpcre_exec()\fP
2791 returns data, even though the meaning of the strings is different.)
2792 .P
2793 The strings are returned in reverse order of length; that is, the longest
2794 matching string is given first. If there were too many matches to fit into
2795 \fIovector\fP, the yield of the function is zero, and the vector is filled with
2796 the longest matches. Unlike \fBpcre_exec()\fP, \fBpcre_dfa_exec()\fP can use
2797 the entire \fIovector\fP for returning matched strings.
2798
2799 NOTE: PCRE's "auto-possessification" optimization usually applies to character
2800 repeats at the end of a pattern (as well as internally). For example, the
2801 pattern "a\ed+" is compiled as if it were "a\ed++" because there is no point
2802 even considering the possibility of backtracking into the repeated digits. For
2803 DFA matching, this means that only one possible match is found. If you really
2804 do want multiple matches in such cases, either use an ungreedy repeat
2805 ("a\ed+?") or set the PCRE_NO_AUTO_POSSESSIFY option when compiling.
2806 .
2807 .
2808 .SS "Error returns from \fBpcre_dfa_exec()\fP"
2809 .rs
2810 .sp
2811 The \fBpcre_dfa_exec()\fP function returns a negative number when it fails.
2812 Many of the errors are the same as for \fBpcre_exec()\fP, and these are
2813 described
2814 .\" HTML <a href="#errorlist">
2815 .\" </a>
2816 above.
2817 .\"
2818 There are in addition the following errors that are specific to
2819 \fBpcre_dfa_exec()\fP:
2820 .sp
2821 PCRE_ERROR_DFA_UITEM (-16)
2822 .sp
2823 This return is given if \fBpcre_dfa_exec()\fP encounters an item in the pattern
2824 that it does not support, for instance, the use of \eC or a back reference.
2825 .sp
2826 PCRE_ERROR_DFA_UCOND (-17)
2827 .sp
2828 This return is given if \fBpcre_dfa_exec()\fP encounters a condition item that
2829 uses a back reference for the condition, or a test for recursion in a specific
2830 group. These are not supported.
2831 .sp
2832 PCRE_ERROR_DFA_UMLIMIT (-18)
2833 .sp
2834 This return is given if \fBpcre_dfa_exec()\fP is called with an \fIextra\fP
2835 block that contains a setting of the \fImatch_limit\fP or
2836 \fImatch_limit_recursion\fP fields. This is not supported (these fields are
2837 meaningless for DFA matching).
2838 .sp
2839 PCRE_ERROR_DFA_WSSIZE (-19)
2840 .sp
2841 This return is given if \fBpcre_dfa_exec()\fP runs out of space in the
2842 \fIworkspace\fP vector.
2843 .sp
2844 PCRE_ERROR_DFA_RECURSE (-20)
2845 .sp
2846 When a recursive subpattern is processed, the matching function calls itself
2847 recursively, using private vectors for \fIovector\fP and \fIworkspace\fP. This
2848 error is given if the output vector is not large enough. This should be
2849 extremely rare, as a vector of size 1000 is used.
2850 .sp
2851 PCRE_ERROR_DFA_BADRESTART (-30)
2852 .sp
2853 When \fBpcre_dfa_exec()\fP is called with the \fBPCRE_DFA_RESTART\fP option,
2854 some plausibility checks are made on the contents of the workspace, which
2855 should contain data about the previous partial match. If any of these checks
2856 fail, this error is given.
2857 .
2858 .
2859 .SH "SEE ALSO"
2860 .rs
2861 .sp
2862 \fBpcre16\fP(3), \fBpcre32\fP(3), \fBpcrebuild\fP(3), \fBpcrecallout\fP(3),
2863 \fBpcrecpp(3)\fP(3), \fBpcrematching\fP(3), \fBpcrepartial\fP(3),
2864 \fBpcreposix\fP(3), \fBpcreprecompile\fP(3), \fBpcresample\fP(3),
2865 \fBpcrestack\fP(3).
2866 .
2867 .
2868 .SH AUTHOR
2869 .rs
2870 .sp
2871 .nf
2872 Philip Hazel
2873 University Computing Service
2874 Cambridge CB2 3QH, England.
2875 .fi
2876 .
2877 .
2878 .SH REVISION
2879 .rs
2880 .sp
2881 .nf
2882 Last updated: 05 November 2013
2883 Copyright (c) 1997-2013 University of Cambridge.
2884 .fi

Properties

Name Value
svn:eol-style native
svn:keywords "Author Date Id Revision Url"

  ViewVC Help
Powered by ViewVC 1.1.5