ViewVC logotype

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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1766 - (show annotations)
Tue Jun 15 16:15:50 2021 UTC (3 months ago) by ph10
File MIME type: text/html
File size: 135355 byte(s)
Error occurred while calculating annotation data.
Final source tidies for 8.45 release.
1 <html>
2 <head>
3 <title>pcreapi specification</title>
4 </head>
5 <body bgcolor="#FFFFFF" text="#00005A" link="#0066FF" alink="#3399FF" vlink="#2222BB">
6 <h1>pcreapi man page</h1>
7 <p>
8 Return to the <a href="index.html">PCRE index page</a>.
9 </p>
10 <p>
11 This page is part of the PCRE HTML documentation. It was generated automatically
12 from the original man page. If there is any nonsense in it, please consult the
13 man page, in case the conversion went wrong.
14 <br>
15 <ul>
16 <li><a name="TOC1" href="#SEC1">PCRE NATIVE API BASIC FUNCTIONS</a>
18 <li><a name="TOC3" href="#SEC3">PCRE NATIVE API AUXILIARY FUNCTIONS</a>
20 <li><a name="TOC5" href="#SEC5">PCRE 8-BIT, 16-BIT, AND 32-BIT LIBRARIES</a>
21 <li><a name="TOC6" href="#SEC6">PCRE API OVERVIEW</a>
22 <li><a name="TOC7" href="#SEC7">NEWLINES</a>
23 <li><a name="TOC8" href="#SEC8">MULTITHREADING</a>
25 <li><a name="TOC10" href="#SEC10">CHECKING BUILD-TIME OPTIONS</a>
26 <li><a name="TOC11" href="#SEC11">COMPILING A PATTERN</a>
27 <li><a name="TOC12" href="#SEC12">COMPILATION ERROR CODES</a>
28 <li><a name="TOC13" href="#SEC13">STUDYING A PATTERN</a>
29 <li><a name="TOC14" href="#SEC14">LOCALE SUPPORT</a>
30 <li><a name="TOC15" href="#SEC15">INFORMATION ABOUT A PATTERN</a>
31 <li><a name="TOC16" href="#SEC16">REFERENCE COUNTS</a>
35 <li><a name="TOC20" href="#SEC20">DUPLICATE SUBPATTERN NAMES</a>
36 <li><a name="TOC21" href="#SEC21">FINDING ALL POSSIBLE MATCHES</a>
37 <li><a name="TOC22" href="#SEC22">OBTAINING AN ESTIMATE OF STACK USAGE</a>
39 <li><a name="TOC24" href="#SEC24">SEE ALSO</a>
40 <li><a name="TOC25" href="#SEC25">AUTHOR</a>
41 <li><a name="TOC26" href="#SEC26">REVISION</a>
42 </ul>
43 <P>
44 <b>#include &#60;pcre.h&#62;</b>
45 </P>
46 <br><a name="SEC1" href="#TOC1">PCRE NATIVE API BASIC FUNCTIONS</a><br>
47 <P>
48 <b>pcre *pcre_compile(const char *<i>pattern</i>, int <i>options</i>,</b>
49 <b> const char **<i>errptr</i>, int *<i>erroffset</i>,</b>
50 <b> const unsigned char *<i>tableptr</i>);</b>
51 <br>
52 <br>
53 <b>pcre *pcre_compile2(const char *<i>pattern</i>, int <i>options</i>,</b>
54 <b> int *<i>errorcodeptr</i>,</b>
55 <b> const char **<i>errptr</i>, int *<i>erroffset</i>,</b>
56 <b> const unsigned char *<i>tableptr</i>);</b>
57 <br>
58 <br>
59 <b>pcre_extra *pcre_study(const pcre *<i>code</i>, int <i>options</i>,</b>
60 <b> const char **<i>errptr</i>);</b>
61 <br>
62 <br>
63 <b>void pcre_free_study(pcre_extra *<i>extra</i>);</b>
64 <br>
65 <br>
66 <b>int pcre_exec(const pcre *<i>code</i>, const pcre_extra *<i>extra</i>,</b>
67 <b> const char *<i>subject</i>, int <i>length</i>, int <i>startoffset</i>,</b>
68 <b> int <i>options</i>, int *<i>ovector</i>, int <i>ovecsize</i>);</b>
69 <br>
70 <br>
71 <b>int pcre_dfa_exec(const pcre *<i>code</i>, const pcre_extra *<i>extra</i>,</b>
72 <b> const char *<i>subject</i>, int <i>length</i>, int <i>startoffset</i>,</b>
73 <b> int <i>options</i>, int *<i>ovector</i>, int <i>ovecsize</i>,</b>
74 <b> int *<i>workspace</i>, int <i>wscount</i>);</b>
75 </P>
77 <P>
78 <b>int pcre_copy_named_substring(const pcre *<i>code</i>,</b>
79 <b> const char *<i>subject</i>, int *<i>ovector</i>,</b>
80 <b> int <i>stringcount</i>, const char *<i>stringname</i>,</b>
81 <b> char *<i>buffer</i>, int <i>buffersize</i>);</b>
82 <br>
83 <br>
84 <b>int pcre_copy_substring(const char *<i>subject</i>, int *<i>ovector</i>,</b>
85 <b> int <i>stringcount</i>, int <i>stringnumber</i>, char *<i>buffer</i>,</b>
86 <b> int <i>buffersize</i>);</b>
87 <br>
88 <br>
89 <b>int pcre_get_named_substring(const pcre *<i>code</i>,</b>
90 <b> const char *<i>subject</i>, int *<i>ovector</i>,</b>
91 <b> int <i>stringcount</i>, const char *<i>stringname</i>,</b>
92 <b> const char **<i>stringptr</i>);</b>
93 <br>
94 <br>
95 <b>int pcre_get_stringnumber(const pcre *<i>code</i>,</b>
96 <b> const char *<i>name</i>);</b>
97 <br>
98 <br>
99 <b>int pcre_get_stringtable_entries(const pcre *<i>code</i>,</b>
100 <b> const char *<i>name</i>, char **<i>first</i>, char **<i>last</i>);</b>
101 <br>
102 <br>
103 <b>int pcre_get_substring(const char *<i>subject</i>, int *<i>ovector</i>,</b>
104 <b> int <i>stringcount</i>, int <i>stringnumber</i>,</b>
105 <b> const char **<i>stringptr</i>);</b>
106 <br>
107 <br>
108 <b>int pcre_get_substring_list(const char *<i>subject</i>,</b>
109 <b> int *<i>ovector</i>, int <i>stringcount</i>, const char ***<i>listptr</i>);</b>
110 <br>
111 <br>
112 <b>void pcre_free_substring(const char *<i>stringptr</i>);</b>
113 <br>
114 <br>
115 <b>void pcre_free_substring_list(const char **<i>stringptr</i>);</b>
116 </P>
117 <br><a name="SEC3" href="#TOC1">PCRE NATIVE API AUXILIARY FUNCTIONS</a><br>
118 <P>
119 <b>int pcre_jit_exec(const pcre *<i>code</i>, const pcre_extra *<i>extra</i>,</b>
120 <b> const char *<i>subject</i>, int <i>length</i>, int <i>startoffset</i>,</b>
121 <b> int <i>options</i>, int *<i>ovector</i>, int <i>ovecsize</i>,</b>
122 <b> pcre_jit_stack *<i>jstack</i>);</b>
123 <br>
124 <br>
125 <b>pcre_jit_stack *pcre_jit_stack_alloc(int <i>startsize</i>, int <i>maxsize</i>);</b>
126 <br>
127 <br>
128 <b>void pcre_jit_stack_free(pcre_jit_stack *<i>stack</i>);</b>
129 <br>
130 <br>
131 <b>void pcre_assign_jit_stack(pcre_extra *<i>extra</i>,</b>
132 <b> pcre_jit_callback <i>callback</i>, void *<i>data</i>);</b>
133 <br>
134 <br>
135 <b>const unsigned char *pcre_maketables(void);</b>
136 <br>
137 <br>
138 <b>int pcre_fullinfo(const pcre *<i>code</i>, const pcre_extra *<i>extra</i>,</b>
139 <b> int <i>what</i>, void *<i>where</i>);</b>
140 <br>
141 <br>
142 <b>int pcre_refcount(pcre *<i>code</i>, int <i>adjust</i>);</b>
143 <br>
144 <br>
145 <b>int pcre_config(int <i>what</i>, void *<i>where</i>);</b>
146 <br>
147 <br>
148 <b>const char *pcre_version(void);</b>
149 <br>
150 <br>
151 <b>int pcre_pattern_to_host_byte_order(pcre *<i>code</i>,</b>
152 <b> pcre_extra *<i>extra</i>, const unsigned char *<i>tables</i>);</b>
153 </P>
154 <br><a name="SEC4" href="#TOC1">PCRE NATIVE API INDIRECTED FUNCTIONS</a><br>
155 <P>
156 <b>void *(*pcre_malloc)(size_t);</b>
157 <br>
158 <br>
159 <b>void (*pcre_free)(void *);</b>
160 <br>
161 <br>
162 <b>void *(*pcre_stack_malloc)(size_t);</b>
163 <br>
164 <br>
165 <b>void (*pcre_stack_free)(void *);</b>
166 <br>
167 <br>
168 <b>int (*pcre_callout)(pcre_callout_block *);</b>
169 <br>
170 <br>
171 <b>int (*pcre_stack_guard)(void);</b>
172 </P>
173 <br><a name="SEC5" href="#TOC1">PCRE 8-BIT, 16-BIT, AND 32-BIT LIBRARIES</a><br>
174 <P>
175 As well as support for 8-bit character strings, PCRE also supports 16-bit
176 strings (from release 8.30) and 32-bit strings (from release 8.32), by means of
177 two additional libraries. They can be built as well as, or instead of, the
178 8-bit library. To avoid too much complication, this document describes the
179 8-bit versions of the functions, with only occasional references to the 16-bit
180 and 32-bit libraries.
181 </P>
182 <P>
183 The 16-bit and 32-bit functions operate in the same way as their 8-bit
184 counterparts; they just use different data types for their arguments and
185 results, and their names start with <b>pcre16_</b> or <b>pcre32_</b> instead of
186 <b>pcre_</b>. For every option that has UTF8 in its name (for example,
187 PCRE_UTF8), there are corresponding 16-bit and 32-bit names with UTF8 replaced
188 by UTF16 or UTF32, respectively. This facility is in fact just cosmetic; the
189 16-bit and 32-bit option names define the same bit values.
190 </P>
191 <P>
192 References to bytes and UTF-8 in this document should be read as references to
193 16-bit data units and UTF-16 when using the 16-bit library, or 32-bit data
194 units and UTF-32 when using the 32-bit library, unless specified otherwise.
195 More details of the specific differences for the 16-bit and 32-bit libraries
196 are given in the
197 <a href="pcre16.html"><b>pcre16</b></a>
198 and
199 <a href="pcre32.html"><b>pcre32</b></a>
200 pages.
201 </P>
202 <br><a name="SEC6" href="#TOC1">PCRE API OVERVIEW</a><br>
203 <P>
204 PCRE has its own native API, which is described in this document. There are
205 also some wrapper functions (for the 8-bit library only) that correspond to the
206 POSIX regular expression API, but they do not give access to all the
207 functionality. They are described in the
208 <a href="pcreposix.html"><b>pcreposix</b></a>
209 documentation. Both of these APIs define a set of C function calls. A C++
210 wrapper (again for the 8-bit library only) is also distributed with PCRE. It is
211 documented in the
212 <a href="pcrecpp.html"><b>pcrecpp</b></a>
213 page.
214 </P>
215 <P>
216 The native API C function prototypes are defined in the header file
217 <b>pcre.h</b>, and on Unix-like systems the (8-bit) library itself is called
218 <b>libpcre</b>. It can normally be accessed by adding <b>-lpcre</b> to the
219 command for linking an application that uses PCRE. The header file defines the
220 macros PCRE_MAJOR and PCRE_MINOR to contain the major and minor release numbers
221 for the library. Applications can use these to include support for different
222 releases of PCRE.
223 </P>
224 <P>
225 In a Windows environment, if you want to statically link an application program
226 against a non-dll <b>pcre.a</b> file, you must define PCRE_STATIC before
227 including <b>pcre.h</b> or <b>pcrecpp.h</b>, because otherwise the
228 <b>pcre_malloc()</b> and <b>pcre_free()</b> exported functions will be declared
229 <b>__declspec(dllimport)</b>, with unwanted results.
230 </P>
231 <P>
232 The functions <b>pcre_compile()</b>, <b>pcre_compile2()</b>, <b>pcre_study()</b>,
233 and <b>pcre_exec()</b> are used for compiling and matching regular expressions
234 in a Perl-compatible manner. A sample program that demonstrates the simplest
235 way of using them is provided in the file called <i>pcredemo.c</i> in the PCRE
236 source distribution. A listing of this program is given in the
237 <a href="pcredemo.html"><b>pcredemo</b></a>
238 documentation, and the
239 <a href="pcresample.html"><b>pcresample</b></a>
240 documentation describes how to compile and run it.
241 </P>
242 <P>
243 Just-in-time compiler support is an optional feature of PCRE that can be built
244 in appropriate hardware environments. It greatly speeds up the matching
245 performance of many patterns. Simple programs can easily request that it be
246 used if available, by setting an option that is ignored when it is not
247 relevant. More complicated programs might need to make use of the functions
248 <b>pcre_jit_stack_alloc()</b>, <b>pcre_jit_stack_free()</b>, and
249 <b>pcre_assign_jit_stack()</b> in order to control the JIT code's memory usage.
250 </P>
251 <P>
252 From release 8.32 there is also a direct interface for JIT execution, which
253 gives improved performance. The JIT-specific functions are discussed in the
254 <a href="pcrejit.html"><b>pcrejit</b></a>
255 documentation.
256 </P>
257 <P>
258 A second matching function, <b>pcre_dfa_exec()</b>, which is not
259 Perl-compatible, is also provided. This uses a different algorithm for the
260 matching. The alternative algorithm finds all possible matches (at a given
261 point in the subject), and scans the subject just once (unless there are
262 lookbehind assertions). However, this algorithm does not return captured
263 substrings. A description of the two matching algorithms and their advantages
264 and disadvantages is given in the
265 <a href="pcrematching.html"><b>pcrematching</b></a>
266 documentation.
267 </P>
268 <P>
269 In addition to the main compiling and matching functions, there are convenience
270 functions for extracting captured substrings from a subject string that is
271 matched by <b>pcre_exec()</b>. They are:
272 <pre>
273 <b>pcre_copy_substring()</b>
274 <b>pcre_copy_named_substring()</b>
275 <b>pcre_get_substring()</b>
276 <b>pcre_get_named_substring()</b>
277 <b>pcre_get_substring_list()</b>
278 <b>pcre_get_stringnumber()</b>
279 <b>pcre_get_stringtable_entries()</b>
280 </pre>
281 <b>pcre_free_substring()</b> and <b>pcre_free_substring_list()</b> are also
282 provided, to free the memory used for extracted strings.
283 </P>
284 <P>
285 The function <b>pcre_maketables()</b> is used to build a set of character tables
286 in the current locale for passing to <b>pcre_compile()</b>, <b>pcre_exec()</b>,
287 or <b>pcre_dfa_exec()</b>. This is an optional facility that is provided for
288 specialist use. Most commonly, no special tables are passed, in which case
289 internal tables that are generated when PCRE is built are used.
290 </P>
291 <P>
292 The function <b>pcre_fullinfo()</b> is used to find out information about a
293 compiled pattern. The function <b>pcre_version()</b> returns a pointer to a
294 string containing the version of PCRE and its date of release.
295 </P>
296 <P>
297 The function <b>pcre_refcount()</b> maintains a reference count in a data block
298 containing a compiled pattern. This is provided for the benefit of
299 object-oriented applications.
300 </P>
301 <P>
302 The global variables <b>pcre_malloc</b> and <b>pcre_free</b> initially contain
303 the entry points of the standard <b>malloc()</b> and <b>free()</b> functions,
304 respectively. PCRE calls the memory management functions via these variables,
305 so a calling program can replace them if it wishes to intercept the calls. This
306 should be done before calling any PCRE functions.
307 </P>
308 <P>
309 The global variables <b>pcre_stack_malloc</b> and <b>pcre_stack_free</b> are also
310 indirections to memory management functions. These special functions are used
311 only when PCRE is compiled to use the heap for remembering data, instead of
312 recursive function calls, when running the <b>pcre_exec()</b> function. See the
313 <a href="pcrebuild.html"><b>pcrebuild</b></a>
314 documentation for details of how to do this. It is a non-standard way of
315 building PCRE, for use in environments that have limited stacks. Because of the
316 greater use of memory management, it runs more slowly. Separate functions are
317 provided so that special-purpose external code can be used for this case. When
318 used, these functions always allocate memory blocks of the same size. There is
319 a discussion about PCRE's stack usage in the
320 <a href="pcrestack.html"><b>pcrestack</b></a>
321 documentation.
322 </P>
323 <P>
324 The global variable <b>pcre_callout</b> initially contains NULL. It can be set
325 by the caller to a "callout" function, which PCRE will then call at specified
326 points during a matching operation. Details are given in the
327 <a href="pcrecallout.html"><b>pcrecallout</b></a>
328 documentation.
329 </P>
330 <P>
331 The global variable <b>pcre_stack_guard</b> initially contains NULL. It can be
332 set by the caller to a function that is called by PCRE whenever it starts
333 to compile a parenthesized part of a pattern. When parentheses are nested, PCRE
334 uses recursive function calls, which use up the system stack. This function is
335 provided so that applications with restricted stacks can force a compilation
336 error if the stack runs out. The function should return zero if all is well, or
337 non-zero to force an error.
338 <a name="newlines"></a></P>
339 <br><a name="SEC7" href="#TOC1">NEWLINES</a><br>
340 <P>
341 PCRE supports five different conventions for indicating line breaks in
342 strings: a single CR (carriage return) character, a single LF (linefeed)
343 character, the two-character sequence CRLF, any of the three preceding, or any
344 Unicode newline sequence. The Unicode newline sequences are the three just
345 mentioned, plus the single characters VT (vertical tab, U+000B), FF (form feed,
346 U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
347 (paragraph separator, U+2029).
348 </P>
349 <P>
350 Each of the first three conventions is used by at least one operating system as
351 its standard newline sequence. When PCRE is built, a default can be specified.
352 The default default is LF, which is the Unix standard. When PCRE is run, the
353 default can be overridden, either when a pattern is compiled, or when it is
354 matched.
355 </P>
356 <P>
357 At compile time, the newline convention can be specified by the <i>options</i>
358 argument of <b>pcre_compile()</b>, or it can be specified by special text at the
359 start of the pattern itself; this overrides any other settings. See the
360 <a href="pcrepattern.html"><b>pcrepattern</b></a>
361 page for details of the special character sequences.
362 </P>
363 <P>
364 In the PCRE documentation the word "newline" is used to mean "the character or
365 pair of characters that indicate a line break". The choice of newline
366 convention affects the handling of the dot, circumflex, and dollar
367 metacharacters, the handling of #-comments in /x mode, and, when CRLF is a
368 recognized line ending sequence, the match position advancement for a
369 non-anchored pattern. There is more detail about this in the
370 <a href="#execoptions">section on <b>pcre_exec()</b> options</a>
371 below.
372 </P>
373 <P>
374 The choice of newline convention does not affect the interpretation of
375 the \n or \r escape sequences, nor does it affect what \R matches, which is
376 controlled in a similar way, but by separate options.
377 </P>
378 <br><a name="SEC8" href="#TOC1">MULTITHREADING</a><br>
379 <P>
380 The PCRE functions can be used in multi-threading applications, with the
381 proviso that the memory management functions pointed to by <b>pcre_malloc</b>,
382 <b>pcre_free</b>, <b>pcre_stack_malloc</b>, and <b>pcre_stack_free</b>, and the
383 callout and stack-checking functions pointed to by <b>pcre_callout</b> and
384 <b>pcre_stack_guard</b>, are shared by all threads.
385 </P>
386 <P>
387 The compiled form of a regular expression is not altered during matching, so
388 the same compiled pattern can safely be used by several threads at once.
389 </P>
390 <P>
391 If the just-in-time optimization feature is being used, it needs separate
392 memory stack areas for each thread. See the
393 <a href="pcrejit.html"><b>pcrejit</b></a>
394 documentation for more details.
395 </P>
396 <br><a name="SEC9" href="#TOC1">SAVING PRECOMPILED PATTERNS FOR LATER USE</a><br>
397 <P>
398 The compiled form of a regular expression can be saved and re-used at a later
399 time, possibly by a different program, and even on a host other than the one on
400 which it was compiled. Details are given in the
401 <a href="pcreprecompile.html"><b>pcreprecompile</b></a>
402 documentation, which includes a description of the
403 <b>pcre_pattern_to_host_byte_order()</b> function. However, compiling a regular
404 expression with one version of PCRE for use with a different version is not
405 guaranteed to work and may cause crashes.
406 </P>
407 <br><a name="SEC10" href="#TOC1">CHECKING BUILD-TIME OPTIONS</a><br>
408 <P>
409 <b>int pcre_config(int <i>what</i>, void *<i>where</i>);</b>
410 </P>
411 <P>
412 The function <b>pcre_config()</b> makes it possible for a PCRE client to
413 discover which optional features have been compiled into the PCRE library. The
414 <a href="pcrebuild.html"><b>pcrebuild</b></a>
415 documentation has more details about these optional features.
416 </P>
417 <P>
418 The first argument for <b>pcre_config()</b> is an integer, specifying which
419 information is required; the second argument is a pointer to a variable into
420 which the information is placed. The returned value is zero on success, or the
421 negative error code PCRE_ERROR_BADOPTION if the value in the first argument is
422 not recognized. The following information is available:
423 <pre>
425 </pre>
426 The output is an integer that is set to one if UTF-8 support is available;
427 otherwise it is set to zero. This value should normally be given to the 8-bit
428 version of this function, <b>pcre_config()</b>. If it is given to the 16-bit
429 or 32-bit version of this function, the result is PCRE_ERROR_BADOPTION.
430 <pre>
432 </pre>
433 The output is an integer that is set to one if UTF-16 support is available;
434 otherwise it is set to zero. This value should normally be given to the 16-bit
435 version of this function, <b>pcre16_config()</b>. If it is given to the 8-bit
436 or 32-bit version of this function, the result is PCRE_ERROR_BADOPTION.
437 <pre>
439 </pre>
440 The output is an integer that is set to one if UTF-32 support is available;
441 otherwise it is set to zero. This value should normally be given to the 32-bit
442 version of this function, <b>pcre32_config()</b>. If it is given to the 8-bit
443 or 16-bit version of this function, the result is PCRE_ERROR_BADOPTION.
444 <pre>
446 </pre>
447 The output is an integer that is set to one if support for Unicode character
448 properties is available; otherwise it is set to zero.
449 <pre>
451 </pre>
452 The output is an integer that is set to one if support for just-in-time
453 compiling is available; otherwise it is set to zero.
454 <pre>
456 </pre>
457 The output is a pointer to a zero-terminated "const char *" string. If JIT
458 support is available, the string contains the name of the architecture for
459 which the JIT compiler is configured, for example "x86 32bit (little endian +
460 unaligned)". If JIT support is not available, the result is NULL.
461 <pre>
463 </pre>
464 The output is an integer whose value specifies the default character sequence
465 that is recognized as meaning "newline". The values that are supported in
466 ASCII/Unicode environments are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for
467 ANYCRLF, and -1 for ANY. In EBCDIC environments, CR, ANYCRLF, and ANY yield the
468 same values. However, the value for LF is normally 21, though some EBCDIC
469 environments use 37. The corresponding values for CRLF are 3349 and 3365. The
470 default should normally correspond to the standard sequence for your operating
471 system.
472 <pre>
474 </pre>
475 The output is an integer whose value indicates what character sequences the \R
476 escape sequence matches by default. A value of 0 means that \R matches any
477 Unicode line ending sequence; a value of 1 means that \R matches only CR, LF,
478 or CRLF. The default can be overridden when a pattern is compiled or matched.
479 <pre>
481 </pre>
482 The output is an integer that contains the number of bytes used for internal
483 linkage in compiled regular expressions. For the 8-bit library, the value can
484 be 2, 3, or 4. For the 16-bit library, the value is either 2 or 4 and is still
485 a number of bytes. For the 32-bit library, the value is either 2 or 4 and is
486 still a number of bytes. The default value of 2 is sufficient for all but the
487 most massive patterns, since it allows the compiled pattern to be up to 64K in
488 size. Larger values allow larger regular expressions to be compiled, at the
489 expense of slower matching.
490 <pre>
492 </pre>
493 The output is an integer that contains the threshold above which the POSIX
494 interface uses <b>malloc()</b> for output vectors. Further details are given in
495 the
496 <a href="pcreposix.html"><b>pcreposix</b></a>
497 documentation.
498 <pre>
500 </pre>
501 The output is a long integer that gives the maximum depth of nesting of
502 parentheses (of any kind) in a pattern. This limit is imposed to cap the amount
503 of system stack used when a pattern is compiled. It is specified when PCRE is
504 built; the default is 250. This limit does not take into account the stack that
505 may already be used by the calling application. For finer control over
506 compilation stack usage, you can set a pointer to an external checking function
507 in <b>pcre_stack_guard</b>.
508 <pre>
510 </pre>
511 The output is a long integer that gives the default limit for the number of
512 internal matching function calls in a <b>pcre_exec()</b> execution. Further
513 details are given with <b>pcre_exec()</b> below.
514 <pre>
516 </pre>
517 The output is a long integer that gives the default limit for the depth of
518 recursion when calling the internal matching function in a <b>pcre_exec()</b>
519 execution. Further details are given with <b>pcre_exec()</b> below.
520 <pre>
522 </pre>
523 The output is an integer that is set to one if internal recursion when running
524 <b>pcre_exec()</b> is implemented by recursive function calls that use the stack
525 to remember their state. This is the usual way that PCRE is compiled. The
526 output is zero if PCRE was compiled to use blocks of data on the heap instead
527 of recursive function calls. In this case, <b>pcre_stack_malloc</b> and
528 <b>pcre_stack_free</b> are called to manage memory blocks on the heap, thus
529 avoiding the use of the stack.
530 </P>
531 <br><a name="SEC11" href="#TOC1">COMPILING A PATTERN</a><br>
532 <P>
533 <b>pcre *pcre_compile(const char *<i>pattern</i>, int <i>options</i>,</b>
534 <b> const char **<i>errptr</i>, int *<i>erroffset</i>,</b>
535 <b> const unsigned char *<i>tableptr</i>);</b>
536 <br>
537 <br>
538 <b>pcre *pcre_compile2(const char *<i>pattern</i>, int <i>options</i>,</b>
539 <b> int *<i>errorcodeptr</i>,</b>
540 <b> const char **<i>errptr</i>, int *<i>erroffset</i>,</b>
541 <b> const unsigned char *<i>tableptr</i>);</b>
542 </P>
543 <P>
544 Either of the functions <b>pcre_compile()</b> or <b>pcre_compile2()</b> can be
545 called to compile a pattern into an internal form. The only difference between
546 the two interfaces is that <b>pcre_compile2()</b> has an additional argument,
547 <i>errorcodeptr</i>, via which a numerical error code can be returned. To avoid
548 too much repetition, we refer just to <b>pcre_compile()</b> below, but the
549 information applies equally to <b>pcre_compile2()</b>.
550 </P>
551 <P>
552 The pattern is a C string terminated by a binary zero, and is passed in the
553 <i>pattern</i> argument. A pointer to a single block of memory that is obtained
554 via <b>pcre_malloc</b> is returned. This contains the compiled code and related
555 data. The <b>pcre</b> type is defined for the returned block; this is a typedef
556 for a structure whose contents are not externally defined. It is up to the
557 caller to free the memory (via <b>pcre_free</b>) when it is no longer required.
558 </P>
559 <P>
560 Although the compiled code of a PCRE regex is relocatable, that is, it does not
561 depend on memory location, the complete <b>pcre</b> data block is not
562 fully relocatable, because it may contain a copy of the <i>tableptr</i>
563 argument, which is an address (see below).
564 </P>
565 <P>
566 The <i>options</i> argument contains various bit settings that affect the
567 compilation. It should be zero if no options are required. The available
568 options are described below. Some of them (in particular, those that are
569 compatible with Perl, but some others as well) can also be set and unset from
570 within the pattern (see the detailed description in the
571 <a href="pcrepattern.html"><b>pcrepattern</b></a>
572 documentation). For those options that can be different in different parts of
573 the pattern, the contents of the <i>options</i> argument specifies their
574 settings at the start of compilation and execution. The PCRE_ANCHORED,
575 PCRE_BSR_<i>xxx</i>, PCRE_NEWLINE_<i>xxx</i>, PCRE_NO_UTF8_CHECK, and
576 PCRE_NO_START_OPTIMIZE options can be set at the time of matching as well as at
577 compile time.
578 </P>
579 <P>
580 If <i>errptr</i> is NULL, <b>pcre_compile()</b> returns NULL immediately.
581 Otherwise, if compilation of a pattern fails, <b>pcre_compile()</b> returns
582 NULL, and sets the variable pointed to by <i>errptr</i> to point to a textual
583 error message. This is a static string that is part of the library. You must
584 not try to free it. Normally, the offset from the start of the pattern to the
585 data unit that was being processed when the error was discovered is placed in
586 the variable pointed to by <i>erroffset</i>, which must not be NULL (if it is,
587 an immediate error is given). However, for an invalid UTF-8 or UTF-16 string,
588 the offset is that of the first data unit of the failing character.
589 </P>
590 <P>
591 Some errors are not detected until the whole pattern has been scanned; in these
592 cases, the offset passed back is the length of the pattern. Note that the
593 offset is in data units, not characters, even in a UTF mode. It may sometimes
594 point into the middle of a UTF-8 or UTF-16 character.
595 </P>
596 <P>
597 If <b>pcre_compile2()</b> is used instead of <b>pcre_compile()</b>, and the
598 <i>errorcodeptr</i> argument is not NULL, a non-zero error code number is
599 returned via this argument in the event of an error. This is in addition to the
600 textual error message. Error codes and messages are listed below.
601 </P>
602 <P>
603 If the final argument, <i>tableptr</i>, is NULL, PCRE uses a default set of
604 character tables that are built when PCRE is compiled, using the default C
605 locale. Otherwise, <i>tableptr</i> must be an address that is the result of a
606 call to <b>pcre_maketables()</b>. This value is stored with the compiled
607 pattern, and used again by <b>pcre_exec()</b> and <b>pcre_dfa_exec()</b> when the
608 pattern is matched. For more discussion, see the section on locale support
609 below.
610 </P>
611 <P>
612 This code fragment shows a typical straightforward call to <b>pcre_compile()</b>:
613 <pre>
614 pcre *re;
615 const char *error;
616 int erroffset;
617 re = pcre_compile(
618 "^A.*Z", /* the pattern */
619 0, /* default options */
620 &error, /* for error message */
621 &erroffset, /* for error offset */
622 NULL); /* use default character tables */
623 </pre>
624 The following names for option bits are defined in the <b>pcre.h</b> header
625 file:
626 <pre>
628 </pre>
629 If this bit is set, the pattern is forced to be "anchored", that is, it is
630 constrained to match only at the first matching point in the string that is
631 being searched (the "subject string"). This effect can also be achieved by
632 appropriate constructs in the pattern itself, which is the only way to do it in
633 Perl.
634 <pre>
636 </pre>
637 If this bit is set, <b>pcre_compile()</b> automatically inserts callout items,
638 all with number 255, before each pattern item. For discussion of the callout
639 facility, see the
640 <a href="pcrecallout.html"><b>pcrecallout</b></a>
641 documentation.
642 <pre>
645 </pre>
646 These options (which are mutually exclusive) control what the \R escape
647 sequence matches. The choice is either to match only CR, LF, or CRLF, or to
648 match any Unicode newline sequence. The default is specified when PCRE is
649 built. It can be overridden from within the pattern, or by setting an option
650 when a compiled pattern is matched.
651 <pre>
653 </pre>
654 If this bit is set, letters in the pattern match both upper and lower case
655 letters. It is equivalent to Perl's /i option, and it can be changed within a
656 pattern by a (?i) option setting. In UTF-8 mode, PCRE always understands the
657 concept of case for characters whose values are less than 128, so caseless
658 matching is always possible. For characters with higher values, the concept of
659 case is supported if PCRE is compiled with Unicode property support, but not
660 otherwise. If you want to use caseless matching for characters 128 and above,
661 you must ensure that PCRE is compiled with Unicode property support as well as
662 with UTF-8 support.
663 <pre>
665 </pre>
666 If this bit is set, a dollar metacharacter in the pattern matches only at the
667 end of the subject string. Without this option, a dollar also matches
668 immediately before a newline at the end of the string (but not before any other
669 newlines). The PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
670 There is no equivalent to this option in Perl, and no way to set it within a
671 pattern.
672 <pre>
674 </pre>
675 If this bit is set, a dot metacharacter in the pattern matches a character of
676 any value, including one that indicates a newline. However, it only ever
677 matches one character, even if newlines are coded as CRLF. Without this option,
678 a dot does not match when the current position is at a newline. This option is
679 equivalent to Perl's /s option, and it can be changed within a pattern by a
680 (?s) option setting. A negative class such as [^a] always matches newline
681 characters, independent of the setting of this option.
682 <pre>
684 </pre>
685 If this bit is set, names used to identify capturing subpatterns need not be
686 unique. This can be helpful for certain types of pattern when it is known that
687 only one instance of the named subpattern can ever be matched. There are more
688 details of named subpatterns below; see also the
689 <a href="pcrepattern.html"><b>pcrepattern</b></a>
690 documentation.
691 <pre>
693 </pre>
694 If this bit is set, most white space characters in the pattern are totally
695 ignored except when escaped or inside a character class. However, white space
696 is not allowed within sequences such as (?&#62; that introduce various
697 parenthesized subpatterns, nor within a numerical quantifier such as {1,3}.
698 However, ignorable white space is permitted between an item and a following
699 quantifier and between a quantifier and a following + that indicates
700 possessiveness.
701 </P>
702 <P>
703 White space did not used to include the VT character (code 11), because Perl
704 did not treat this character as white space. However, Perl changed at release
705 5.18, so PCRE followed at release 8.34, and VT is now treated as white space.
706 </P>
707 <P>
708 PCRE_EXTENDED also causes characters between an unescaped # outside a character
709 class and the next newline, inclusive, to be ignored. PCRE_EXTENDED is
710 equivalent to Perl's /x option, and it can be changed within a pattern by a
711 (?x) option setting.
712 </P>
713 <P>
714 Which characters are interpreted as newlines is controlled by the options
715 passed to <b>pcre_compile()</b> or by a special sequence at the start of the
716 pattern, as described in the section entitled
717 <a href="pcrepattern.html#newlines">"Newline conventions"</a>
718 in the <b>pcrepattern</b> documentation. Note that the end of this type of
719 comment is a literal newline sequence in the pattern; escape sequences that
720 happen to represent a newline do not count.
721 </P>
722 <P>
723 This option makes it possible to include comments inside complicated patterns.
724 Note, however, that this applies only to data characters. White space characters
725 may never appear within special character sequences in a pattern, for example
726 within the sequence (?( that introduces a conditional subpattern.
727 <pre>
729 </pre>
730 This option was invented in order to turn on additional functionality of PCRE
731 that is incompatible with Perl, but it is currently of very little use. When
732 set, any backslash in a pattern that is followed by a letter that has no
733 special meaning causes an error, thus reserving these combinations for future
734 expansion. By default, as in Perl, a backslash followed by a letter with no
735 special meaning is treated as a literal. (Perl can, however, be persuaded to
736 give an error for this, by running it with the -w option.) There are at present
737 no other features controlled by this option. It can also be set by a (?X)
738 option setting within a pattern.
739 <pre>
741 </pre>
742 If this option is set, an unanchored pattern is required to match before or at
743 the first newline in the subject string, though the matched text may continue
744 over the newline.
745 <pre>
747 </pre>
748 If this option is set, PCRE's behaviour is changed in some ways so that it is
749 compatible with JavaScript rather than Perl. The changes are as follows:
750 </P>
751 <P>
752 (1) A lone closing square bracket in a pattern causes a compile-time error,
753 because this is illegal in JavaScript (by default it is treated as a data
754 character). Thus, the pattern AB]CD becomes illegal when this option is set.
755 </P>
756 <P>
757 (2) At run time, a back reference to an unset subpattern group matches an empty
758 string (by default this causes the current matching alternative to fail). A
759 pattern such as (\1)(a) succeeds when this option is set (assuming it can find
760 an "a" in the subject), whereas it fails by default, for Perl compatibility.
761 </P>
762 <P>
763 (3) \U matches an upper case "U" character; by default \U causes a compile
764 time error (Perl uses \U to upper case subsequent characters).
765 </P>
766 <P>
767 (4) \u matches a lower case "u" character unless it is followed by four
768 hexadecimal digits, in which case the hexadecimal number defines the code point
769 to match. By default, \u causes a compile time error (Perl uses it to upper
770 case the following character).
771 </P>
772 <P>
773 (5) \x matches a lower case "x" character unless it is followed by two
774 hexadecimal digits, in which case the hexadecimal number defines the code point
775 to match. By default, as in Perl, a hexadecimal number is always expected after
776 \x, but it may have zero, one, or two digits (so, for example, \xz matches a
777 binary zero character followed by z).
778 <pre>
780 </pre>
781 By default, for the purposes of matching "start of line" and "end of line",
782 PCRE treats the subject string as consisting of a single line of characters,
783 even if it actually contains newlines. The "start of line" metacharacter (^)
784 matches only at the start of the string, and the "end of line" metacharacter
785 ($) matches only at the end of the string, or before a terminating newline
786 (except when PCRE_DOLLAR_ENDONLY is set). Note, however, that unless
787 PCRE_DOTALL is set, the "any character" metacharacter (.) does not match at a
788 newline. This behaviour (for ^, $, and dot) is the same as Perl.
789 </P>
790 <P>
791 When PCRE_MULTILINE it is set, the "start of line" and "end of line" constructs
792 match immediately following or immediately before internal newlines in the
793 subject string, respectively, as well as at the very start and end. This is
794 equivalent to Perl's /m option, and it can be changed within a pattern by a
795 (?m) option setting. If there are no newlines in a subject string, or no
796 occurrences of ^ or $ in a pattern, setting PCRE_MULTILINE has no effect.
797 <pre>
799 </pre>
800 This option locks out interpretation of the pattern as UTF-8 (or UTF-16 or
801 UTF-32 in the 16-bit and 32-bit libraries). In particular, it prevents the
802 creator of the pattern from switching to UTF interpretation by starting the
803 pattern with (*UTF). This may be useful in applications that process patterns
804 from external sources. The combination of PCRE_UTF8 and PCRE_NEVER_UTF also
805 causes an error.
806 <pre>
812 </pre>
813 These options override the default newline definition that was chosen when PCRE
814 was built. Setting the first or the second specifies that a newline is
815 indicated by a single character (CR or LF, respectively). Setting
816 PCRE_NEWLINE_CRLF specifies that a newline is indicated by the two-character
817 CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies that any of the three
818 preceding sequences should be recognized. Setting PCRE_NEWLINE_ANY specifies
819 that any Unicode newline sequence should be recognized.
820 </P>
821 <P>
822 In an ASCII/Unicode environment, the Unicode newline sequences are the three
823 just mentioned, plus the single characters VT (vertical tab, U+000B), FF (form
824 feed, U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
825 (paragraph separator, U+2029). For the 8-bit library, the last two are
826 recognized only in UTF-8 mode.
827 </P>
828 <P>
829 When PCRE is compiled to run in an EBCDIC (mainframe) environment, the code for
830 CR is 0x0d, the same as ASCII. However, the character code for LF is normally
831 0x15, though in some EBCDIC environments 0x25 is used. Whichever of these is
832 not LF is made to correspond to Unicode's NEL character. EBCDIC codes are all
833 less than 256. For more details, see the
834 <a href="pcrebuild.html"><b>pcrebuild</b></a>
835 documentation.
836 </P>
837 <P>
838 The newline setting in the options word uses three bits that are treated
839 as a number, giving eight possibilities. Currently only six are used (default
840 plus the five values above). This means that if you set more than one newline
841 option, the combination may or may not be sensible. For example,
843 other combinations may yield unused numbers and cause an error.
844 </P>
845 <P>
846 The only time that a line break in a pattern is specially recognized when
847 compiling is when PCRE_EXTENDED is set. CR and LF are white space characters,
848 and so are ignored in this mode. Also, an unescaped # outside a character class
849 indicates a comment that lasts until after the next line break sequence. In
850 other circumstances, line break sequences in patterns are treated as literal
851 data.
852 </P>
853 <P>
854 The newline option that is set at compile time becomes the default that is used
855 for <b>pcre_exec()</b> and <b>pcre_dfa_exec()</b>, but it can be overridden.
856 <pre>
858 </pre>
859 If this option is set, it disables the use of numbered capturing parentheses in
860 the pattern. Any opening parenthesis that is not followed by ? behaves as if it
861 were followed by ?: but named parentheses can still be used for capturing (and
862 they acquire numbers in the usual way). There is no equivalent of this option
863 in Perl.
864 <pre>
866 </pre>
867 If this option is set, it disables "auto-possessification". This is an
868 optimization that, for example, turns a+b into a++b in order to avoid
869 backtracks into a+ that can never be successful. However, if callouts are in
870 use, auto-possessification means that some of them are never taken. You can set
871 this option if you want the matching functions to do a full unoptimized search
872 and run all the callouts, but it is mainly provided for testing purposes.
873 <pre>
875 </pre>
876 This is an option that acts at matching time; that is, it is really an option
877 for <b>pcre_exec()</b> or <b>pcre_dfa_exec()</b>. If it is set at compile time,
878 it is remembered with the compiled pattern and assumed at matching time. This
879 is necessary if you want to use JIT execution, because the JIT compiler needs
880 to know whether or not this option is set. For details see the discussion of
882 <a href="#execoptions">below.</a>
883 <pre>
885 </pre>
886 This option changes the way PCRE processes \B, \b, \D, \d, \S, \s, \W,
887 \w, and some of the POSIX character classes. By default, only ASCII characters
888 are recognized, but if PCRE_UCP is set, Unicode properties are used instead to
889 classify characters. More details are given in the section on
890 <a href="pcre.html#genericchartypes">generic character types</a>
891 in the
892 <a href="pcrepattern.html"><b>pcrepattern</b></a>
893 page. If you set PCRE_UCP, matching one of the items it affects takes much
894 longer. The option is available only if PCRE has been compiled with Unicode
895 property support.
896 <pre>
898 </pre>
899 This option inverts the "greediness" of the quantifiers so that they are not
900 greedy by default, but become greedy if followed by "?". It is not compatible
901 with Perl. It can also be set by a (?U) option setting within the pattern.
902 <pre>
904 </pre>
905 This option causes PCRE to regard both the pattern and the subject as strings
906 of UTF-8 characters instead of single-byte strings. However, it is available
907 only when PCRE is built to include UTF support. If not, the use of this option
908 provokes an error. Details of how this option changes the behaviour of PCRE are
909 given in the
910 <a href="pcreunicode.html"><b>pcreunicode</b></a>
911 page.
912 <pre>
914 </pre>
915 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
916 automatically checked. There is a discussion about the
917 <a href="pcreunicode.html#utf8strings">validity of UTF-8 strings</a>
918 in the
919 <a href="pcreunicode.html"><b>pcreunicode</b></a>
920 page. If an invalid UTF-8 sequence is found, <b>pcre_compile()</b> returns an
921 error. If you already know that your pattern is valid, and you want to skip
922 this check for performance reasons, you can set the PCRE_NO_UTF8_CHECK option.
923 When it is set, the effect of passing an invalid UTF-8 string as a pattern is
924 undefined. It may cause your program to crash or loop. Note that this option
925 can also be passed to <b>pcre_exec()</b> and <b>pcre_dfa_exec()</b>, to suppress
926 the validity checking of subject strings only. If the same string is being
927 matched many times, the option can be safely set for the second and subsequent
928 matchings to improve performance.
929 </P>
930 <br><a name="SEC12" href="#TOC1">COMPILATION ERROR CODES</a><br>
931 <P>
932 The following table lists the error codes than may be returned by
933 <b>pcre_compile2()</b>, along with the error messages that may be returned by
934 both compiling functions. Note that error messages are always 8-bit ASCII
935 strings, even in 16-bit or 32-bit mode. As PCRE has developed, some error codes
936 have fallen out of use. To avoid confusion, they have not been re-used.
937 <pre>
938 0 no error
939 1 \ at end of pattern
940 2 \c at end of pattern
941 3 unrecognized character follows \
942 4 numbers out of order in {} quantifier
943 5 number too big in {} quantifier
944 6 missing terminating ] for character class
945 7 invalid escape sequence in character class
946 8 range out of order in character class
947 9 nothing to repeat
948 10 [this code is not in use]
949 11 internal error: unexpected repeat
950 12 unrecognized character after (? or (?-
951 13 POSIX named classes are supported only within a class
952 14 missing )
953 15 reference to non-existent subpattern
954 16 erroffset passed as NULL
955 17 unknown option bit(s) set
956 18 missing ) after comment
957 19 [this code is not in use]
958 20 regular expression is too large
959 21 failed to get memory
960 22 unmatched parentheses
961 23 internal error: code overflow
962 24 unrecognized character after (?&#60;
963 25 lookbehind assertion is not fixed length
964 26 malformed number or name after (?(
965 27 conditional group contains more than two branches
966 28 assertion expected after (?(
967 29 (?R or (?[+-]digits must be followed by )
968 30 unknown POSIX class name
969 31 POSIX collating elements are not supported
970 32 this version of PCRE is compiled without UTF support
971 33 [this code is not in use]
972 34 character value in \x{} or \o{} is too large
973 35 invalid condition (?(0)
974 36 \C not allowed in lookbehind assertion
975 37 PCRE does not support \L, \l, \N{name}, \U, or \u
976 38 number after (?C is &#62; 255
977 39 closing ) for (?C expected
978 40 recursive call could loop indefinitely
979 41 unrecognized character after (?P
980 42 syntax error in subpattern name (missing terminator)
981 43 two named subpatterns have the same name
982 44 invalid UTF-8 string (specifically UTF-8)
983 45 support for \P, \p, and \X has not been compiled
984 46 malformed \P or \p sequence
985 47 unknown property name after \P or \p
986 48 subpattern name is too long (maximum 32 characters)
987 49 too many named subpatterns (maximum 10000)
988 50 [this code is not in use]
989 51 octal value is greater than \377 in 8-bit non-UTF-8 mode
990 52 internal error: overran compiling workspace
991 53 internal error: previously-checked referenced subpattern
992 not found
993 54 DEFINE group contains more than one branch
994 55 repeating a DEFINE group is not allowed
995 56 inconsistent NEWLINE options
996 57 \g is not followed by a braced, angle-bracketed, or quoted
997 name/number or by a plain number
998 58 a numbered reference must not be zero
999 59 an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)
1000 60 (*VERB) not recognized or malformed
1001 61 number is too big
1002 62 subpattern name expected
1003 63 digit expected after (?+
1004 64 ] is an invalid data character in JavaScript compatibility mode
1005 65 different names for subpatterns of the same number are
1006 not allowed
1007 66 (*MARK) must have an argument
1008 67 this version of PCRE is not compiled with Unicode property
1009 support
1010 68 \c must be followed by an ASCII character
1011 69 \k is not followed by a braced, angle-bracketed, or quoted name
1012 70 internal error: unknown opcode in find_fixedlength()
1013 71 \N is not supported in a class
1014 72 too many forward references
1015 73 disallowed Unicode code point (&#62;= 0xd800 && &#60;= 0xdfff)
1016 74 invalid UTF-16 string (specifically UTF-16)
1017 75 name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)
1018 76 character value in \u.... sequence is too large
1019 77 invalid UTF-32 string (specifically UTF-32)
1020 78 setting UTF is disabled by the application
1021 79 non-hex character in \x{} (closing brace missing?)
1022 80 non-octal character in \o{} (closing brace missing?)
1023 81 missing opening brace after \o
1024 82 parentheses are too deeply nested
1025 83 invalid range in character class
1026 84 group name must start with a non-digit
1027 85 parentheses are too deeply nested (stack check)
1028 </pre>
1029 The numbers 32 and 10000 in errors 48 and 49 are defaults; different values may
1030 be used if the limits were changed when PCRE was built.
1031 <a name="studyingapattern"></a></P>
1032 <br><a name="SEC13" href="#TOC1">STUDYING A PATTERN</a><br>
1033 <P>
1034 <b>pcre_extra *pcre_study(const pcre *<i>code</i>, int <i>options</i>,</b>
1035 <b> const char **<i>errptr</i>);</b>
1036 </P>
1037 <P>
1038 If a compiled pattern is going to be used several times, it is worth spending
1039 more time analyzing it in order to speed up the time taken for matching. The
1040 function <b>pcre_study()</b> takes a pointer to a compiled pattern as its first
1041 argument. If studying the pattern produces additional information that will
1042 help speed up matching, <b>pcre_study()</b> returns a pointer to a
1043 <b>pcre_extra</b> block, in which the <i>study_data</i> field points to the
1044 results of the study.
1045 </P>
1046 <P>
1047 The returned value from <b>pcre_study()</b> can be passed directly to
1048 <b>pcre_exec()</b> or <b>pcre_dfa_exec()</b>. However, a <b>pcre_extra</b> block
1049 also contains other fields that can be set by the caller before the block is
1050 passed; these are described
1051 <a href="#extradata">below</a>
1052 in the section on matching a pattern.
1053 </P>
1054 <P>
1055 If studying the pattern does not produce any useful information,
1056 <b>pcre_study()</b> returns NULL by default. In that circumstance, if the
1057 calling program wants to pass any of the other fields to <b>pcre_exec()</b> or
1058 <b>pcre_dfa_exec()</b>, it must set up its own <b>pcre_extra</b> block. However,
1059 if <b>pcre_study()</b> is called with the PCRE_STUDY_EXTRA_NEEDED option, it
1060 returns a <b>pcre_extra</b> block even if studying did not find any additional
1061 information. It may still return NULL, however, if an error occurs in
1062 <b>pcre_study()</b>.
1063 </P>
1064 <P>
1065 The second argument of <b>pcre_study()</b> contains option bits. There are three
1066 further options in addition to PCRE_STUDY_EXTRA_NEEDED:
1067 <pre>
1071 </pre>
1072 If any of these are set, and the just-in-time compiler is available, the
1073 pattern is further compiled into machine code that executes much faster than
1074 the <b>pcre_exec()</b> interpretive matching function. If the just-in-time
1075 compiler is not available, these options are ignored. All undefined bits in the
1076 <i>options</i> argument must be zero.
1077 </P>
1078 <P>
1079 JIT compilation is a heavyweight optimization. It can take some time for
1080 patterns to be analyzed, and for one-off matches and simple patterns the
1081 benefit of faster execution might be offset by a much slower study time.
1082 Not all patterns can be optimized by the JIT compiler. For those that cannot be
1083 handled, matching automatically falls back to the <b>pcre_exec()</b>
1084 interpreter. For more details, see the
1085 <a href="pcrejit.html"><b>pcrejit</b></a>
1086 documentation.
1087 </P>
1088 <P>
1089 The third argument for <b>pcre_study()</b> is a pointer for an error message. If
1090 studying succeeds (even if no data is returned), the variable it points to is
1091 set to NULL. Otherwise it is set to point to a textual error message. This is a
1092 static string that is part of the library. You must not try to free it. You
1093 should test the error pointer for NULL after calling <b>pcre_study()</b>, to be
1094 sure that it has run successfully.
1095 </P>
1096 <P>
1097 When you are finished with a pattern, you can free the memory used for the
1098 study data by calling <b>pcre_free_study()</b>. This function was added to the
1099 API for release 8.20. For earlier versions, the memory could be freed with
1100 <b>pcre_free()</b>, just like the pattern itself. This will still work in cases
1101 where JIT optimization is not used, but it is advisable to change to the new
1102 function when convenient.
1103 </P>
1104 <P>
1105 This is a typical way in which <b>pcre_study</b>() is used (except that in a
1106 real application there should be tests for errors):
1107 <pre>
1108 int rc;
1109 pcre *re;
1110 pcre_extra *sd;
1111 re = pcre_compile("pattern", 0, &error, &erroroffset, NULL);
1112 sd = pcre_study(
1113 re, /* result of pcre_compile() */
1114 0, /* no options */
1115 &error); /* set to NULL or points to a message */
1116 rc = pcre_exec( /* see below for details of pcre_exec() options */
1117 re, sd, "subject", 7, 0, 0, ovector, 30);
1118 ...
1119 pcre_free_study(sd);
1120 pcre_free(re);
1121 </pre>
1122 Studying a pattern does two things: first, a lower bound for the length of
1123 subject string that is needed to match the pattern is computed. This does not
1124 mean that there are any strings of that length that match, but it does
1125 guarantee that no shorter strings match. The value is used to avoid wasting
1126 time by trying to match strings that are shorter than the lower bound. You can
1127 find out the value in a calling program via the <b>pcre_fullinfo()</b> function.
1128 </P>
1129 <P>
1130 Studying a pattern is also useful for non-anchored patterns that do not have a
1131 single fixed starting character. A bitmap of possible starting bytes is
1132 created. This speeds up finding a position in the subject at which to start
1133 matching. (In 16-bit mode, the bitmap is used for 16-bit values less than 256.
1134 In 32-bit mode, the bitmap is used for 32-bit values less than 256.)
1135 </P>
1136 <P>
1137 These two optimizations apply to both <b>pcre_exec()</b> and
1138 <b>pcre_dfa_exec()</b>, and the information is also used by the JIT compiler.
1139 The optimizations can be disabled by setting the PCRE_NO_START_OPTIMIZE option.
1140 You might want to do this if your pattern contains callouts or (*MARK) and you
1141 want to make use of these facilities in cases where matching fails.
1142 </P>
1143 <P>
1144 PCRE_NO_START_OPTIMIZE can be specified at either compile time or execution
1145 time. However, if PCRE_NO_START_OPTIMIZE is passed to <b>pcre_exec()</b>, (that
1146 is, after any JIT compilation has happened) JIT execution is disabled. For JIT
1147 execution to work with PCRE_NO_START_OPTIMIZE, the option must be set at
1148 compile time.
1149 </P>
1150 <P>
1151 There is a longer discussion of PCRE_NO_START_OPTIMIZE
1152 <a href="#execoptions">below.</a>
1153 <a name="localesupport"></a></P>
1154 <br><a name="SEC14" href="#TOC1">LOCALE SUPPORT</a><br>
1155 <P>
1156 PCRE handles caseless matching, and determines whether characters are letters,
1157 digits, or whatever, by reference to a set of tables, indexed by character
1158 code point. When running in UTF-8 mode, or in the 16- or 32-bit libraries, this
1159 applies only to characters with code points less than 256. By default,
1160 higher-valued code points never match escapes such as \w or \d. However, if
1161 PCRE is built with Unicode property support, all characters can be tested with
1162 \p and \P, or, alternatively, the PCRE_UCP option can be set when a pattern
1163 is compiled; this causes \w and friends to use Unicode property support
1164 instead of the built-in tables.
1165 </P>
1166 <P>
1167 The use of locales with Unicode is discouraged. If you are handling characters
1168 with code points greater than 128, you should either use Unicode support, or
1169 use locales, but not try to mix the two.
1170 </P>
1171 <P>
1172 PCRE contains an internal set of tables that are used when the final argument
1173 of <b>pcre_compile()</b> is NULL. These are sufficient for many applications.
1174 Normally, the internal tables recognize only ASCII characters. However, when
1175 PCRE is built, it is possible to cause the internal tables to be rebuilt in the
1176 default "C" locale of the local system, which may cause them to be different.
1177 </P>
1178 <P>
1179 The internal tables can always be overridden by tables supplied by the
1180 application that calls PCRE. These may be created in a different locale from
1181 the default. As more and more applications change to using Unicode, the need
1182 for this locale support is expected to die away.
1183 </P>
1184 <P>
1185 External tables are built by calling the <b>pcre_maketables()</b> function,
1186 which has no arguments, in the relevant locale. The result can then be passed
1187 to <b>pcre_compile()</b> as often as necessary. For example, to build and use
1188 tables that are appropriate for the French locale (where accented characters
1189 with values greater than 128 are treated as letters), the following code could
1190 be used:
1191 <pre>
1192 setlocale(LC_CTYPE, "fr_FR");
1193 tables = pcre_maketables();
1194 re = pcre_compile(..., tables);
1195 </pre>
1196 The locale name "fr_FR" is used on Linux and other Unix-like systems; if you
1197 are using Windows, the name for the French locale is "french".
1198 </P>
1199 <P>
1200 When <b>pcre_maketables()</b> runs, the tables are built in memory that is
1201 obtained via <b>pcre_malloc</b>. It is the caller's responsibility to ensure
1202 that the memory containing the tables remains available for as long as it is
1203 needed.
1204 </P>
1205 <P>
1206 The pointer that is passed to <b>pcre_compile()</b> is saved with the compiled
1207 pattern, and the same tables are used via this pointer by <b>pcre_study()</b>
1208 and also by <b>pcre_exec()</b> and <b>pcre_dfa_exec()</b>. Thus, for any single
1209 pattern, compilation, studying and matching all happen in the same locale, but
1210 different patterns can be processed in different locales.
1211 </P>
1212 <P>
1213 It is possible to pass a table pointer or NULL (indicating the use of the
1214 internal tables) to <b>pcre_exec()</b> or <b>pcre_dfa_exec()</b> (see the
1215 discussion below in the section on matching a pattern). This facility is
1216 provided for use with pre-compiled patterns that have been saved and reloaded.
1217 Character tables are not saved with patterns, so if a non-standard table was
1218 used at compile time, it must be provided again when the reloaded pattern is
1219 matched. Attempting to use this facility to match a pattern in a different
1220 locale from the one in which it was compiled is likely to lead to anomalous
1221 (usually incorrect) results.
1222 <a name="infoaboutpattern"></a></P>
1223 <br><a name="SEC15" href="#TOC1">INFORMATION ABOUT A PATTERN</a><br>
1224 <P>
1225 <b>int pcre_fullinfo(const pcre *<i>code</i>, const pcre_extra *<i>extra</i>,</b>
1226 <b> int <i>what</i>, void *<i>where</i>);</b>
1227 </P>
1228 <P>
1229 The <b>pcre_fullinfo()</b> function returns information about a compiled
1230 pattern. It replaces the <b>pcre_info()</b> function, which was removed from the
1231 library at version 8.30, after more than 10 years of obsolescence.
1232 </P>
1233 <P>
1234 The first argument for <b>pcre_fullinfo()</b> is a pointer to the compiled
1235 pattern. The second argument is the result of <b>pcre_study()</b>, or NULL if
1236 the pattern was not studied. The third argument specifies which piece of
1237 information is required, and the fourth argument is a pointer to a variable
1238 to receive the data. The yield of the function is zero for success, or one of
1239 the following negative numbers:
1240 <pre>
1241 PCRE_ERROR_NULL the argument <i>code</i> was NULL
1242 the argument <i>where</i> was NULL
1243 PCRE_ERROR_BADMAGIC the "magic number" was not found
1244 PCRE_ERROR_BADENDIANNESS the pattern was compiled with different
1245 endianness
1246 PCRE_ERROR_BADOPTION the value of <i>what</i> was invalid
1247 PCRE_ERROR_UNSET the requested field is not set
1248 </pre>
1249 The "magic number" is placed at the start of each compiled pattern as a simple
1250 check against passing an arbitrary memory pointer. The endianness error can
1251 occur if a compiled pattern is saved and reloaded on a different host. Here is
1252 a typical call of <b>pcre_fullinfo()</b>, to obtain the length of the compiled
1253 pattern:
1254 <pre>
1255 int rc;
1256 size_t length;
1257 rc = pcre_fullinfo(
1258 re, /* result of pcre_compile() */
1259 sd, /* result of pcre_study(), or NULL */
1260 PCRE_INFO_SIZE, /* what is required */
1261 &length); /* where to put the data */
1262 </pre>
1263 The possible values for the third argument are defined in <b>pcre.h</b>, and are
1264 as follows:
1265 <pre>
1267 </pre>
1268 Return the number of the highest back reference in the pattern. The fourth
1269 argument should point to an <b>int</b> variable. Zero is returned if there are
1270 no back references.
1271 <pre>
1273 </pre>
1274 Return the number of capturing subpatterns in the pattern. The fourth argument
1275 should point to an <b>int</b> variable.
1276 <pre>
1278 </pre>
1279 Return a pointer to the internal default character tables within PCRE. The
1280 fourth argument should point to an <b>unsigned char *</b> variable. This
1281 information call is provided for internal use by the <b>pcre_study()</b>
1282 function. External callers can cause PCRE to use its internal tables by passing
1283 a NULL table pointer.
1284 <pre>
1285 PCRE_INFO_FIRSTBYTE (deprecated)
1286 </pre>
1287 Return information about the first data unit of any matched string, for a
1288 non-anchored pattern. The name of this option refers to the 8-bit library,
1289 where data units are bytes. The fourth argument should point to an <b>int</b>
1290 variable. Negative values are used for special cases. However, this means that
1291 when the 32-bit library is in non-UTF-32 mode, the full 32-bit range of
1292 characters cannot be returned. For this reason, this value is deprecated; use
1294 </P>
1295 <P>
1296 If there is a fixed first value, for example, the letter "c" from a pattern
1297 such as (cat|cow|coyote), its value is returned. In the 8-bit library, the
1298 value is always less than 256. In the 16-bit library the value can be up to
1299 0xffff. In the 32-bit library the value can be up to 0x10ffff.
1300 </P>
1301 <P>
1302 If there is no fixed first value, and if either
1303 <br>
1304 <br>
1305 (a) the pattern was compiled with the PCRE_MULTILINE option, and every branch
1306 starts with "^", or
1307 <br>
1308 <br>
1309 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not set
1310 (if it were set, the pattern would be anchored),
1311 <br>
1312 <br>
1313 -1 is returned, indicating that the pattern matches only at the start of a
1314 subject string or after any newline within the string. Otherwise -2 is
1315 returned. For anchored patterns, -2 is returned.
1316 <pre>
1318 </pre>
1319 Return the value of the first data unit (non-UTF character) of any matched
1320 string in the situation where PCRE_INFO_FIRSTCHARACTERFLAGS returns 1;
1321 otherwise return 0. The fourth argument should point to a <b>uint_t</b>
1322 variable.
1323 </P>
1324 <P>
1325 In the 8-bit library, the value is always less than 256. In the 16-bit library
1326 the value can be up to 0xffff. In the 32-bit library in UTF-32 mode the value
1327 can be up to 0x10ffff, and up to 0xffffffff when not using UTF-32 mode.
1328 <pre>
1330 </pre>
1331 Return information about the first data unit of any matched string, for a
1332 non-anchored pattern. The fourth argument should point to an <b>int</b>
1333 variable.
1334 </P>
1335 <P>
1336 If there is a fixed first value, for example, the letter "c" from a pattern
1337 such as (cat|cow|coyote), 1 is returned, and the character value can be
1338 retrieved using PCRE_INFO_FIRSTCHARACTER. If there is no fixed first value, and
1339 if either
1340 <br>
1341 <br>
1342 (a) the pattern was compiled with the PCRE_MULTILINE option, and every branch
1343 starts with "^", or
1344 <br>
1345 <br>
1346 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not set
1347 (if it were set, the pattern would be anchored),
1348 <br>
1349 <br>
1350 2 is returned, indicating that the pattern matches only at the start of a
1351 subject string or after any newline within the string. Otherwise 0 is
1352 returned. For anchored patterns, 0 is returned.
1353 <pre>
1355 </pre>
1356 If the pattern was studied, and this resulted in the construction of a 256-bit
1357 table indicating a fixed set of values for the first data unit in any matching
1358 string, a pointer to the table is returned. Otherwise NULL is returned. The
1359 fourth argument should point to an <b>unsigned char *</b> variable.
1360 <pre>
1362 </pre>
1363 Return 1 if the pattern contains any explicit matches for CR or LF characters,
1364 otherwise 0. The fourth argument should point to an <b>int</b> variable. An
1365 explicit match is either a literal CR or LF character, or \r or \n.
1366 <pre>
1368 </pre>
1369 Return 1 if the (?J) or (?-J) option setting is used in the pattern, otherwise
1370 0. The fourth argument should point to an <b>int</b> variable. (?J) and
1371 (?-J) set and unset the local PCRE_DUPNAMES option, respectively.
1372 <pre>
1374 </pre>
1375 Return 1 if the pattern was studied with one of the JIT options, and
1376 just-in-time compiling was successful. The fourth argument should point to an
1377 <b>int</b> variable. A return value of 0 means that JIT support is not available
1378 in this version of PCRE, or that the pattern was not studied with a JIT option,
1379 or that the JIT compiler could not handle this particular pattern. See the
1380 <a href="pcrejit.html"><b>pcrejit</b></a>
1381 documentation for details of what can and cannot be handled.
1382 <pre>
1384 </pre>
1385 If the pattern was successfully studied with a JIT option, return the size of
1386 the JIT compiled code, otherwise return zero. The fourth argument should point
1387 to a <b>size_t</b> variable.
1388 <pre>
1390 </pre>
1391 Return the value of the rightmost literal data unit that must exist in any
1392 matched string, other than at its start, if such a value has been recorded. The
1393 fourth argument should point to an <b>int</b> variable. If there is no such
1394 value, -1 is returned. For anchored patterns, a last literal value is recorded
1395 only if it follows something of variable length. For example, for the pattern
1396 /^a\d+z\d+/ the returned value is "z", but for /^a\dz\d/ the returned value
1397 is -1.
1398 </P>
1399 <P>
1400 Since for the 32-bit library using the non-UTF-32 mode, this function is unable
1401 to return the full 32-bit range of characters, this value is deprecated;
1403 be used.
1404 <pre>
1406 </pre>
1407 Return 1 if the pattern can match an empty string, otherwise 0. The fourth
1408 argument should point to an <b>int</b> variable.
1409 <pre>
1411 </pre>
1412 If the pattern set a match limit by including an item of the form
1413 (*LIMIT_MATCH=nnnn) at the start, the value is returned. The fourth argument
1414 should point to an unsigned 32-bit integer. If no such value has been set, the
1415 call to <b>pcre_fullinfo()</b> returns the error PCRE_ERROR_UNSET.
1416 <pre>
1418 </pre>
1419 Return the number of characters (NB not data units) in the longest lookbehind
1420 assertion in the pattern. This information is useful when doing multi-segment
1421 matching using the partial matching facilities. Note that the simple assertions
1422 \b and \B require a one-character lookbehind. \A also registers a
1423 one-character lookbehind, though it does not actually inspect the previous
1424 character. This is to ensure that at least one character from the old segment
1425 is retained when a new segment is processed. Otherwise, if there are no
1426 lookbehinds in the pattern, \A might match incorrectly at the start of a new
1427 segment.
1428 <pre>
1430 </pre>
1431 If the pattern was studied and a minimum length for matching subject strings
1432 was computed, its value is returned. Otherwise the returned value is -1. The
1433 value is a number of characters, which in UTF mode may be different from the
1434 number of data units. The fourth argument should point to an <b>int</b>
1435 variable. A non-negative value is a lower bound to the length of any matching
1436 string. There may not be any strings of that length that do actually match, but
1437 every string that does match is at least that long.
1438 <pre>
1442 </pre>
1443 PCRE supports the use of named as well as numbered capturing parentheses. The
1444 names are just an additional way of identifying the parentheses, which still
1445 acquire numbers. Several convenience functions such as
1446 <b>pcre_get_named_substring()</b> are provided for extracting captured
1447 substrings by name. It is also possible to extract the data directly, by first
1448 converting the name to a number in order to access the correct pointers in the
1449 output vector (described with <b>pcre_exec()</b> below). To do the conversion,
1450 you need to use the name-to-number map, which is described by these three
1451 values.
1452 </P>
1453 <P>
1454 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT gives
1455 the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size of each
1456 entry; both of these return an <b>int</b> value. The entry size depends on the
1457 length of the longest name. PCRE_INFO_NAMETABLE returns a pointer to the first
1458 entry of the table. This is a pointer to <b>char</b> in the 8-bit library, where
1459 the first two bytes of each entry are the number of the capturing parenthesis,
1460 most significant byte first. In the 16-bit library, the pointer points to
1461 16-bit data units, the first of which contains the parenthesis number. In the
1462 32-bit library, the pointer points to 32-bit data units, the first of which
1463 contains the parenthesis number. The rest of the entry is the corresponding
1464 name, zero terminated.
1465 </P>
1466 <P>
1467 The names are in alphabetical order. If (?| is used to create multiple groups
1468 with the same number, as described in the
1469 <a href="pcrepattern.html#dupsubpatternnumber">section on duplicate subpattern numbers</a>
1470 in the
1471 <a href="pcrepattern.html"><b>pcrepattern</b></a>
1472 page, the groups may be given the same name, but there is only one entry in the
1473 table. Different names for groups of the same number are not permitted.
1474 Duplicate names for subpatterns with different numbers are permitted,
1475 but only if PCRE_DUPNAMES is set. They appear in the table in the order in
1476 which they were found in the pattern. In the absence of (?| this is the order
1477 of increasing number; when (?| is used this is not necessarily the case because
1478 later subpatterns may have lower numbers.
1479 </P>
1480 <P>
1481 As a simple example of the name/number table, consider the following pattern
1482 after compilation by the 8-bit library (assume PCRE_EXTENDED is set, so white
1483 space - including newlines - is ignored):
1484 <pre>
1485 (?&#60;date&#62; (?&#60;year&#62;(\d\d)?\d\d) - (?&#60;month&#62;\d\d) - (?&#60;day&#62;\d\d) )
1486 </pre>
1487 There are four named subpatterns, so the table has four entries, and each entry
1488 in the table is eight bytes long. The table is as follows, with non-printing
1489 bytes shows in hexadecimal, and undefined bytes shown as ??:
1490 <pre>
1491 00 01 d a t e 00 ??
1492 00 05 d a y 00 ?? ??
1493 00 04 m o n t h 00
1494 00 02 y e a r 00 ??
1495 </pre>
1496 When writing code to extract data from named subpatterns using the
1497 name-to-number map, remember that the length of the entries is likely to be
1498 different for each compiled pattern.
1499 <pre>
1501 </pre>
1502 Return 1 if the pattern can be used for partial matching with
1503 <b>pcre_exec()</b>, otherwise 0. The fourth argument should point to an
1504 <b>int</b> variable. From release 8.00, this always returns 1, because the
1505 restrictions that previously applied to partial matching have been lifted. The
1506 <a href="pcrepartial.html"><b>pcrepartial</b></a>
1507 documentation gives details of partial matching.
1508 <pre>
1510 </pre>
1511 Return a copy of the options with which the pattern was compiled. The fourth
1512 argument should point to an <b>unsigned long int</b> variable. These option bits
1513 are those specified in the call to <b>pcre_compile()</b>, modified by any
1514 top-level option settings at the start of the pattern itself. In other words,
1515 they are the options that will be in force when matching starts. For example,
1516 if the pattern /(?im)abc(?-i)d/ is compiled with the PCRE_EXTENDED option, the
1518 </P>
1519 <P>
1520 A pattern is automatically anchored by PCRE if all of its top-level
1521 alternatives begin with one of the following:
1522 <pre>
1523 ^ unless PCRE_MULTILINE is set
1524 \A always
1525 \G always
1526 .* if PCRE_DOTALL is set and there are no back references to the subpattern in which .* appears
1527 </pre>
1528 For such patterns, the PCRE_ANCHORED bit is set in the options returned by
1529 <b>pcre_fullinfo()</b>.
1530 <pre>
1532 </pre>
1533 If the pattern set a recursion limit by including an item of the form
1534 (*LIMIT_RECURSION=nnnn) at the start, the value is returned. The fourth
1535 argument should point to an unsigned 32-bit integer. If no such value has been
1536 set, the call to <b>pcre_fullinfo()</b> returns the error PCRE_ERROR_UNSET.
1537 <pre>
1539 </pre>
1540 Return the size of the compiled pattern in bytes (for all three libraries). The
1541 fourth argument should point to a <b>size_t</b> variable. This value does not
1542 include the size of the <b>pcre</b> structure that is returned by
1543 <b>pcre_compile()</b>. The value that is passed as the argument to
1544 <b>pcre_malloc()</b> when <b>pcre_compile()</b> is getting memory in which to
1545 place the compiled data is the value returned by this option plus the size of
1546 the <b>pcre</b> structure. Studying a compiled pattern, with or without JIT,
1547 does not alter the value returned by this option.
1548 <pre>
1550 </pre>
1551 Return the size in bytes (for all three libraries) of the data block pointed to
1552 by the <i>study_data</i> field in a <b>pcre_extra</b> block. If <b>pcre_extra</b>
1553 is NULL, or there is no study data, zero is returned. The fourth argument
1554 should point to a <b>size_t</b> variable. The <i>study_data</i> field is set by
1555 <b>pcre_study()</b> to record information that will speed up matching (see the
1556 section entitled
1557 <a href="#studyingapattern">"Studying a pattern"</a>
1558 above). The format of the <i>study_data</i> block is private, but its length
1559 is made available via this option so that it can be saved and restored (see the
1560 <a href="pcreprecompile.html"><b>pcreprecompile</b></a>
1561 documentation for details).
1562 <pre>
1564 </pre>
1565 Returns 1 if there is a rightmost literal data unit that must exist in any
1566 matched string, other than at its start. The fourth argument should point to
1567 an <b>int</b> variable. If there is no such value, 0 is returned. If returning
1568 1, the character value itself can be retrieved using PCRE_INFO_REQUIREDCHAR.
1569 </P>
1570 <P>
1571 For anchored patterns, a last literal value is recorded only if it follows
1572 something of variable length. For example, for the pattern /^a\d+z\d+/ the
1573 returned value 1 (with "z" returned from PCRE_INFO_REQUIREDCHAR), but for
1574 /^a\dz\d/ the returned value is 0.
1575 <pre>
1577 </pre>
1578 Return the value of the rightmost literal data unit that must exist in any
1579 matched string, other than at its start, if such a value has been recorded. The
1580 fourth argument should point to a <b>uint32_t</b> variable. If there is no such
1581 value, 0 is returned.
1582 </P>
1583 <br><a name="SEC16" href="#TOC1">REFERENCE COUNTS</a><br>
1584 <P>
1585 <b>int pcre_refcount(pcre *<i>code</i>, int <i>adjust</i>);</b>
1586 </P>
1587 <P>
1588 The <b>pcre_refcount()</b> function is used to maintain a reference count in the
1589 data block that contains a compiled pattern. It is provided for the benefit of
1590 applications that operate in an object-oriented manner, where different parts
1591 of the application may be using the same compiled pattern, but you want to free
1592 the block when they are all done.
1593 </P>
1594 <P>
1595 When a pattern is compiled, the reference count field is initialized to zero.
1596 It is changed only by calling this function, whose action is to add the
1597 <i>adjust</i> value (which may be positive or negative) to it. The yield of the
1598 function is the new value. However, the value of the count is constrained to
1599 lie between 0 and 65535, inclusive. If the new value is outside these limits,
1600 it is forced to the appropriate limit value.
1601 </P>
1602 <P>
1603 Except when it is zero, the reference count is not correctly preserved if a
1604 pattern is compiled on one host and then transferred to a host whose byte-order
1605 is different. (This seems a highly unlikely scenario.)
1606 </P>
1607 <br><a name="SEC17" href="#TOC1">MATCHING A PATTERN: THE TRADITIONAL FUNCTION</a><br>
1608 <P>
1609 <b>int pcre_exec(const pcre *<i>code</i>, const pcre_extra *<i>extra</i>,</b>
1610 <b> const char *<i>subject</i>, int <i>length</i>, int <i>startoffset</i>,</b>
1611 <b> int <i>options</i>, int *<i>ovector</i>, int <i>ovecsize</i>);</b>
1612 </P>
1613 <P>
1614 The function <b>pcre_exec()</b> is called to match a subject string against a
1615 compiled pattern, which is passed in the <i>code</i> argument. If the
1616 pattern was studied, the result of the study should be passed in the
1617 <i>extra</i> argument. You can call <b>pcre_exec()</b> with the same <i>code</i>
1618 and <i>extra</i> arguments as many times as you like, in order to match
1619 different subject strings with the same pattern.
1620 </P>
1621 <P>
1622 This function is the main matching facility of the library, and it operates in
1623 a Perl-like manner. For specialist use there is also an alternative matching
1624 function, which is described
1625 <a href="#dfamatch">below</a>
1626 in the section about the <b>pcre_dfa_exec()</b> function.
1627 </P>
1628 <P>
1629 In most applications, the pattern will have been compiled (and optionally
1630 studied) in the same process that calls <b>pcre_exec()</b>. However, it is
1631 possible to save compiled patterns and study data, and then use them later
1632 in different processes, possibly even on different hosts. For a discussion
1633 about this, see the
1634 <a href="pcreprecompile.html"><b>pcreprecompile</b></a>
1635 documentation.
1636 </P>
1637 <P>
1638 Here is an example of a simple call to <b>pcre_exec()</b>:
1639 <pre>
1640 int rc;
1641 int ovector[30];
1642 rc = pcre_exec(
1643 re, /* result of pcre_compile() */
1644 NULL, /* we didn't study the pattern */
1645 "some string", /* the subject string */
1646 11, /* the length of the subject string */
1647 0, /* start at offset 0 in the subject */
1648 0, /* default options */
1649 ovector, /* vector of integers for substring information */
1650 30); /* number of elements (NOT size in bytes) */
1651 <a name="extradata"></a></PRE>
1652 </P>
1653 <br><b>
1654 Extra data for <b>pcre_exec()</b>
1655 </b><br>
1656 <P>
1657 If the <i>extra</i> argument is not NULL, it must point to a <b>pcre_extra</b>
1658 data block. The <b>pcre_study()</b> function returns such a block (when it
1659 doesn't return NULL), but you can also create one for yourself, and pass
1660 additional information in it. The <b>pcre_extra</b> block contains the following
1661 fields (not necessarily in this order):
1662 <pre>
1663 unsigned long int <i>flags</i>;
1664 void *<i>study_data</i>;
1665 void *<i>executable_jit</i>;
1666 unsigned long int <i>match_limit</i>;
1667 unsigned long int <i>match_limit_recursion</i>;
1668 void *<i>callout_data</i>;
1669 const unsigned char *<i>tables</i>;
1670 unsigned char **<i>mark</i>;
1671 </pre>
1672 In the 16-bit version of this structure, the <i>mark</i> field has type
1673 "PCRE_UCHAR16 **".
1674 <br>
1675 <br>
1676 In the 32-bit version of this structure, the <i>mark</i> field has type
1677 "PCRE_UCHAR32 **".
1678 </P>
1679 <P>
1680 The <i>flags</i> field is used to specify which of the other fields are set. The
1681 flag bits are:
1682 <pre>
1690 </pre>
1691 Other flag bits should be set to zero. The <i>study_data</i> field and sometimes
1692 the <i>executable_jit</i> field are set in the <b>pcre_extra</b> block that is
1693 returned by <b>pcre_study()</b>, together with the appropriate flag bits. You
1694 should not set these yourself, but you may add to the block by setting other
1695 fields and their corresponding flag bits.
1696 </P>
1697 <P>
1698 The <i>match_limit</i> field provides a means of preventing PCRE from using up a
1699 vast amount of resources when running patterns that are not going to match,
1700 but which have a very large number of possibilities in their search trees. The
1701 classic example is a pattern that uses nested unlimited repeats.
1702 </P>
1703 <P>
1704 Internally, <b>pcre_exec()</b> uses a function called <b>match()</b>, which it
1705 calls repeatedly (sometimes recursively). The limit set by <i>match_limit</i> is
1706 imposed on the number of times this function is called during a match, which
1707 has the effect of limiting the amount of backtracking that can take place. For
1708 patterns that are not anchored, the count restarts from zero for each position
1709 in the subject string.
1710 </P>
1711 <P>
1712 When <b>pcre_exec()</b> is called with a pattern that was successfully studied
1713 with a JIT option, the way that the matching is executed is entirely different.
1714 However, there is still the possibility of runaway matching that goes on for a
1715 very long time, and so the <i>match_limit</i> value is also used in this case
1716 (but in a different way) to limit how long the matching can continue.
1717 </P>
1718 <P>
1719 The default value for the limit can be set when PCRE is built; the default
1720 default is 10 million, which handles all but the most extreme cases. You can
1721 override the default by supplying <b>pcre_exec()</b> with a <b>pcre_extra</b>
1722 block in which <i>match_limit</i> is set, and PCRE_EXTRA_MATCH_LIMIT is set in
1723 the <i>flags</i> field. If the limit is exceeded, <b>pcre_exec()</b> returns
1725 </P>
1726 <P>
1727 A value for the match limit may also be supplied by an item at the start of a
1728 pattern of the form
1729 <pre>
1730 (*LIMIT_MATCH=d)
1731 </pre>
1732 where d is a decimal number. However, such a setting is ignored unless d is
1733 less than the limit set by the caller of <b>pcre_exec()</b> or, if no such limit
1734 is set, less than the default.
1735 </P>
1736 <P>
1737 The <i>match_limit_recursion</i> field is similar to <i>match_limit</i>, but
1738 instead of limiting the total number of times that <b>match()</b> is called, it
1739 limits the depth of recursion. The recursion depth is a smaller number than the
1740 total number of calls, because not all calls to <b>match()</b> are recursive.
1741 This limit is of use only if it is set smaller than <i>match_limit</i>.
1742 </P>
1743 <P>
1744 Limiting the recursion depth limits the amount of machine stack that can be
1745 used, or, when PCRE has been compiled to use memory on the heap instead of the
1746 stack, the amount of heap memory that can be used. This limit is not relevant,
1747 and is ignored, when matching is done using JIT compiled code.
1748 </P>
1749 <P>
1750 The default value for <i>match_limit_recursion</i> can be set when PCRE is
1751 built; the default default is the same value as the default for
1752 <i>match_limit</i>. You can override the default by supplying <b>pcre_exec()</b>
1753 with a <b>pcre_extra</b> block in which <i>match_limit_recursion</i> is set, and
1754 PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the <i>flags</i> field. If the limit
1755 is exceeded, <b>pcre_exec()</b> returns PCRE_ERROR_RECURSIONLIMIT.
1756 </P>
1757 <P>
1758 A value for the recursion limit may also be supplied by an item at the start of
1759 a pattern of the form
1760 <pre>
1762 </pre>
1763 where d is a decimal number. However, such a setting is ignored unless d is
1764 less than the limit set by the caller of <b>pcre_exec()</b> or, if no such limit
1765 is set, less than the default.
1766 </P>
1767 <P>
1768 The <i>callout_data</i> field is used in conjunction with the "callout" feature,
1769 and is described in the
1770 <a href="pcrecallout.html"><b>pcrecallout</b></a>
1771 documentation.
1772 </P>
1773 <P>
1774 The <i>tables</i> field is provided for use with patterns that have been
1775 pre-compiled using custom character tables, saved to disc or elsewhere, and
1776 then reloaded, because the tables that were used to compile a pattern are not
1777 saved with it. See the
1778 <a href="pcreprecompile.html"><b>pcreprecompile</b></a>
1779 documentation for a discussion of saving compiled patterns for later use. If
1780 NULL is passed using this mechanism, it forces PCRE's internal tables to be
1781 used.
1782 </P>
1783 <P>
1784 <b>Warning:</b> The tables that <b>pcre_exec()</b> uses must be the same as those
1785 that were used when the pattern was compiled. If this is not the case, the
1786 behaviour of <b>pcre_exec()</b> is undefined. Therefore, when a pattern is
1787 compiled and matched in the same process, this field should never be set. In
1788 this (the most common) case, the correct table pointer is automatically passed
1789 with the compiled pattern from <b>pcre_compile()</b> to <b>pcre_exec()</b>.
1790 </P>
1791 <P>
1792 If PCRE_EXTRA_MARK is set in the <i>flags</i> field, the <i>mark</i> field must
1793 be set to point to a suitable variable. If the pattern contains any
1794 backtracking control verbs such as (*MARK:NAME), and the execution ends up with
1795 a name to pass back, a pointer to the name string (zero terminated) is placed
1796 in the variable pointed to by the <i>mark</i> field. The names are within the
1797 compiled pattern; if you wish to retain such a name you must copy it before
1798 freeing the memory of a compiled pattern. If there is no name to pass back, the
1799 variable pointed to by the <i>mark</i> field is set to NULL. For details of the
1800 backtracking control verbs, see the section entitled
1801 <a href="pcrepattern#backtrackcontrol">"Backtracking control"</a>
1802 in the
1803 <a href="pcrepattern.html"><b>pcrepattern</b></a>
1804 documentation.
1805 <a name="execoptions"></a></P>
1806 <br><b>
1807 Option bits for <b>pcre_exec()</b>
1808 </b><br>
1809 <P>
1810 The unused bits of the <i>options</i> argument for <b>pcre_exec()</b> must be
1811 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_<i>xxx</i>,
1815 </P>
1816 <P>
1817 If the pattern was successfully studied with one of the just-in-time (JIT)
1818 compile options, the only supported options for JIT execution are
1821 unsupported option is used, JIT execution is disabled and the normal
1822 interpretive code in <b>pcre_exec()</b> is run.
1823 <pre>
1825 </pre>
1826 The PCRE_ANCHORED option limits <b>pcre_exec()</b> to matching at the first
1827 matching position. If a pattern was compiled with PCRE_ANCHORED, or turned out
1828 to be anchored by virtue of its contents, it cannot be made unachored at
1829 matching time.
1830 <pre>
1833 </pre>
1834 These options (which are mutually exclusive) control what the \R escape
1835 sequence matches. The choice is either to match only CR, LF, or CRLF, or to
1836 match any Unicode newline sequence. These options override the choice that was
1837 made or defaulted when the pattern was compiled.
1838 <pre>
1844 </pre>
1845 These options override the newline definition that was chosen or defaulted when
1846 the pattern was compiled. For details, see the description of
1847 <b>pcre_compile()</b> above. During matching, the newline choice affects the
1848 behaviour of the dot, circumflex, and dollar metacharacters. It may also alter
1849 the way the match position is advanced after a match failure for an unanchored
1850 pattern.
1851 </P>
1852 <P>
1854 match attempt for an unanchored pattern fails when the current position is at a
1855 CRLF sequence, and the pattern contains no explicit matches for CR or LF
1856 characters, the match position is advanced by two characters instead of one, in
1857 other words, to after the CRLF.
1858 </P>
1859 <P>
1860 The above rule is a compromise that makes the most common cases work as
1861 expected. For example, if the pattern is .+A (and the PCRE_DOTALL option is not
1862 set), it does not match the string "\r\nA" because, after failing at the
1863 start, it skips both the CR and the LF before retrying. However, the pattern
1864 [\r\n]A does match that string, because it contains an explicit CR or LF
1865 reference, and so advances only by one character after the first failure.
1866 </P>
1867 <P>
1868 An explicit match for CR of LF is either a literal appearance of one of those
1869 characters, or one of the \r or \n escape sequences. Implicit matches such as
1870 [^X] do not count, nor does \s (which includes CR and LF in the characters
1871 that it matches).
1872 </P>
1873 <P>
1874 Notwithstanding the above, anomalous effects may still occur when CRLF is a
1875 valid newline sequence and explicit \r or \n escapes appear in the pattern.
1876 <pre>
1878 </pre>
1879 This option specifies that first character of the subject string is not the
1880 beginning of a line, so the circumflex metacharacter should not match before
1881 it. Setting this without PCRE_MULTILINE (at compile time) causes circumflex
1882 never to match. This option affects only the behaviour of the circumflex
1883 metacharacter. It does not affect \A.
1884 <pre>
1886 </pre>
1887 This option specifies that the end of the subject string is not the end of a
1888 line, so the dollar metacharacter should not match it nor (except in multiline
1889 mode) a newline immediately before it. Setting this without PCRE_MULTILINE (at
1890 compile time) causes dollar never to match. This option affects only the
1891 behaviour of the dollar metacharacter. It does not affect \Z or \z.
1892 <pre>
1894 </pre>
1895 An empty string is not considered to be a valid match if this option is set. If
1896 there are alternatives in the pattern, they are tried. If all the alternatives
1897 match the empty string, the entire match fails. For example, if the pattern
1898 <pre>
1899 a?b?
1900 </pre>
1901 is applied to a string not beginning with "a" or "b", it matches an empty
1902 string at the start of the subject. With PCRE_NOTEMPTY set, this match is not
1903 valid, so PCRE searches further into the string for occurrences of "a" or "b".
1904 <pre>
1906 </pre>
1907 This is like PCRE_NOTEMPTY, except that an empty string match that is not at
1908 the start of the subject is permitted. If the pattern is anchored, such a match
1909 can occur only if the pattern contains \K.
1910 </P>
1911 <P>
1912 Perl has no direct equivalent of PCRE_NOTEMPTY or PCRE_NOTEMPTY_ATSTART, but it
1913 does make a special case of a pattern match of the empty string within its
1914 <b>split()</b> function, and when using the /g modifier. It is possible to
1915 emulate Perl's behaviour after matching a null string by first trying the match
1916 again at the same offset with PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED, and then
1917 if that fails, by advancing the starting offset (see below) and trying an
1918 ordinary match again. There is some code that demonstrates how to do this in
1919 the
1920 <a href="pcredemo.html"><b>pcredemo</b></a>
1921 sample program. In the most general case, you have to check to see if the
1922 newline convention recognizes CRLF as a newline, and if so, and the current
1923 character is CR followed by LF, advance the starting offset by two characters
1924 instead of one.
1925 <pre>
1927 </pre>
1928 There are a number of optimizations that <b>pcre_exec()</b> uses at the start of
1929 a match, in order to speed up the process. For example, if it is known that an
1930 unanchored match must start with a specific character, it searches the subject
1931 for that character, and fails immediately if it cannot find it, without
1932 actually running the main matching function. This means that a special item
1933 such as (*COMMIT) at the start of a pattern is not considered until after a
1934 suitable starting point for the match has been found. Also, when callouts or
1935 (*MARK) items are in use, these "start-up" optimizations can cause them to be
1936 skipped if the pattern is never actually used. The start-up optimizations are
1937 in effect a pre-scan of the subject that takes place before the pattern is run.
1938 </P>
1939 <P>
1940 The PCRE_NO_START_OPTIMIZE option disables the start-up optimizations, possibly
1941 causing performance to suffer, but ensuring that in cases where the result is
1942 "no match", the callouts do occur, and that items such as (*COMMIT) and (*MARK)
1943 are considered at every possible starting position in the subject string. If
1944 PCRE_NO_START_OPTIMIZE is set at compile time, it cannot be unset at matching
1945 time. The use of PCRE_NO_START_OPTIMIZE at matching time (that is, passing it
1946 to <b>pcre_exec()</b>) disables JIT execution; in this situation, matching is
1947 always done using interpretively.
1948 </P>
1949 <P>
1950 Setting PCRE_NO_START_OPTIMIZE can change the outcome of a matching operation.
1951 Consider the pattern
1952 <pre>
1954 </pre>
1955 When this is compiled, PCRE records the fact that a match must start with the
1956 character "A". Suppose the subject string is "DEFABC". The start-up
1957 optimization scans along the subject, finds "A" and runs the first match
1958 attempt from there. The (*COMMIT) item means that the pattern must match the
1959 current starting position, which in this case, it does. However, if the same
1960 match is run with PCRE_NO_START_OPTIMIZE set, the initial scan along the
1961 subject string does not happen. The first match attempt is run starting from
1962 "D" and when this fails, (*COMMIT) prevents any further matches being tried, so
1963 the overall result is "no match". If the pattern is studied, more start-up
1964 optimizations may be used. For example, a minimum length for the subject may be
1965 recorded. Consider the pattern
1966 <pre>
1967 (*MARK:A)(X|Y)
1968 </pre>
1969 The minimum length for a match is one character. If the subject is "ABC", there
1970 will be attempts to match "ABC", "BC", "C", and then finally an empty string.
1971 If the pattern is studied, the final attempt does not take place, because PCRE
1972 knows that the subject is too short, and so the (*MARK) is never encountered.
1973 In this case, studying the pattern does not affect the overall match result,
1974 which is still "no match", but it does affect the auxiliary information that is
1975 returned.
1976 <pre>
1978 </pre>
1979 When PCRE_UTF8 is set at compile time, the validity of the subject as a UTF-8
1980 string is automatically checked when <b>pcre_exec()</b> is subsequently called.
1981 The entire string is checked before any other processing takes place. The value
1982 of <i>startoffset</i> is also checked to ensure that it points to the start of a
1983 UTF-8 character. There is a discussion about the
1984 <a href="pcreunicode.html#utf8strings">validity of UTF-8 strings</a>
1985 in the
1986 <a href="pcreunicode.html"><b>pcreunicode</b></a>
1987 page. If an invalid sequence of bytes is found, <b>pcre_exec()</b> returns the
1988 error PCRE_ERROR_BADUTF8 or, if PCRE_PARTIAL_HARD is set and the problem is a
1989 truncated character at the end of the subject, PCRE_ERROR_SHORTUTF8. In both
1990 cases, information about the precise nature of the error may also be returned
1991 (see the descriptions of these errors in the section entitled \fIError return
1992 values from\fP <b>pcre_exec()</b>
1993 <a href="#errorlist">below).</a>
1994 If <i>startoffset</i> contains a value that does not point to the start of a
1995 UTF-8 character (or to the end of the subject), PCRE_ERROR_BADUTF8_OFFSET is
1996 returned.
1997 </P>
1998 <P>
1999 If you already know that your subject is valid, and you want to skip these
2000 checks for performance reasons, you can set the PCRE_NO_UTF8_CHECK option when
2001 calling <b>pcre_exec()</b>. You might want to do this for the second and
2002 subsequent calls to <b>pcre_exec()</b> if you are making repeated calls to find
2003 all the matches in a single subject string. However, you should be sure that
2004 the value of <i>startoffset</i> points to the start of a character (or the end
2005 of the subject). When PCRE_NO_UTF8_CHECK is set, the effect of passing an
2006 invalid string as a subject or an invalid value of <i>startoffset</i> is
2007 undefined. Your program may crash or loop.
2008 <pre>
2011 </pre>
2012 These options turn on the partial matching feature. For backwards
2013 compatibility, PCRE_PARTIAL is a synonym for PCRE_PARTIAL_SOFT. A partial match
2014 occurs if the end of the subject string is reached successfully, but there are
2015 not enough subject characters to complete the match. If this happens when
2016 PCRE_PARTIAL_SOFT (but not PCRE_PARTIAL_HARD) is set, matching continues by
2017 testing any remaining alternatives. Only if no complete match can be found is
2018 PCRE_ERROR_PARTIAL returned instead of PCRE_ERROR_NOMATCH. In other words,
2019 PCRE_PARTIAL_SOFT says that the caller is prepared to handle a partial match,
2020 but only if no complete match can be found.
2021 </P>
2022 <P>
2023 If PCRE_PARTIAL_HARD is set, it overrides PCRE_PARTIAL_SOFT. In this case, if a
2024 partial match is found, <b>pcre_exec()</b> immediately returns
2025 PCRE_ERROR_PARTIAL, without considering any other alternatives. In other words,
2026 when PCRE_PARTIAL_HARD is set, a partial match is considered to be more
2027 important that an alternative complete match.
2028 </P>
2029 <P>
2030 In both cases, the portion of the string that was inspected when the partial
2031 match was found is set as the first matching string. There is a more detailed
2032 discussion of partial and multi-segment matching, with examples, in the
2033 <a href="pcrepartial.html"><b>pcrepartial</b></a>
2034 documentation.
2035 </P>
2036 <br><b>
2037 The string to be matched by <b>pcre_exec()</b>
2038 </b><br>
2039 <P>
2040 The subject string is passed to <b>pcre_exec()</b> as a pointer in
2041 <i>subject</i>, a length in <i>length</i>, and a starting offset in
2042 <i>startoffset</i>. The units for <i>length</i> and <i>startoffset</i> are bytes
2043 for the 8-bit library, 16-bit data items for the 16-bit library, and 32-bit
2044 data items for the 32-bit library.
2045 </P>
2046 <P>
2047 If <i>startoffset</i> is negative or greater than the length of the subject,
2048 <b>pcre_exec()</b> returns PCRE_ERROR_BADOFFSET. When the starting offset is
2049 zero, the search for a match starts at the beginning of the subject, and this
2050 is by far the most common case. In UTF-8 or UTF-16 mode, the offset must point
2051 to the start of a character, or the end of the subject (in UTF-32 mode, one
2052 data unit equals one character, so all offsets are valid). Unlike the pattern
2053 string, the subject may contain binary zeroes.
2054 </P>
2055 <P>
2056 A non-zero starting offset is useful when searching for another match in the
2057 same subject by calling <b>pcre_exec()</b> again after a previous success.
2058 Setting <i>startoffset</i> differs from just passing over a shortened string and
2059 setting PCRE_NOTBOL in the case of a pattern that begins with any kind of
2060 lookbehind. For example, consider the pattern
2061 <pre>
2062 \Biss\B
2063 </pre>
2064 which finds occurrences of "iss" in the middle of words. (\B matches only if
2065 the current position in the subject is not a word boundary.) When applied to
2066 the string "Mississippi" the first call to <b>pcre_exec()</b> finds the first
2067 occurrence. If <b>pcre_exec()</b> is called again with just the remainder of the
2068 subject, namely "issippi", it does not match, because \B is always false at
2069 the start of the subject, which is deemed to be a word boundary. However, if
2070 <b>pcre_exec()</b> is passed the entire string again, but with <i>startoffset</i>
2071 set to 4, it finds the second occurrence of "iss" because it is able to look
2072 behind the starting point to discover that it is preceded by a letter.
2073 </P>
2074 <P>
2075 Finding all the matches in a subject is tricky when the pattern can match an
2076 empty string. It is possible to emulate Perl's /g behaviour by first trying the
2077 match again at the same offset, with the PCRE_NOTEMPTY_ATSTART and
2078 PCRE_ANCHORED options, and then if that fails, advancing the starting offset
2079 and trying an ordinary match again. There is some code that demonstrates how to
2080 do this in the
2081 <a href="pcredemo.html"><b>pcredemo</b></a>
2082 sample program. In the most general case, you have to check to see if the
2083 newline convention recognizes CRLF as a newline, and if so, and the current
2084 character is CR followed by LF, advance the starting offset by two characters
2085 instead of one.
2086 </P>
2087 <P>
2088 If a non-zero starting offset is passed when the pattern is anchored, one
2089 attempt to match at the given offset is made. This can only succeed if the
2090 pattern does not require the match to be at the start of the subject.
2091 </P>
2092 <br><b>
2093 How <b>pcre_exec()</b> returns captured substrings
2094 </b><br>
2095 <P>
2096 In general, a pattern matches a certain portion of the subject, and in
2097 addition, further substrings from the subject may be picked out by parts of the
2098 pattern. Following the usage in Jeffrey Friedl's book, this is called
2099 "capturing" in what follows, and the phrase "capturing subpattern" is used for
2100 a fragment of a pattern that picks out a substring. PCRE supports several other
2101 kinds of parenthesized subpattern that do not cause substrings to be captured.
2102 </P>
2103 <P>
2104 Captured substrings are returned to the caller via a vector of integers whose
2105 address is passed in <i>ovector</i>. The number of elements in the vector is
2106 passed in <i>ovecsize</i>, which must be a non-negative number. <b>Note</b>: this
2107 argument is NOT the size of <i>ovector</i> in bytes.
2108 </P>
2109 <P>
2110 The first two-thirds of the vector is used to pass back captured substrings,
2111 each substring using a pair of integers. The remaining third of the vector is
2112 used as workspace by <b>pcre_exec()</b> while matching capturing subpatterns,
2113 and is not available for passing back information. The number passed in
2114 <i>ovecsize</i> should always be a multiple of three. If it is not, it is
2115 rounded down.
2116 </P>
2117 <P>
2118 When a match is successful, information about captured substrings is returned
2119 in pairs of integers, starting at the beginning of <i>ovector</i>, and
2120 continuing up to two-thirds of its length at the most. The first element of
2121 each pair is set to the offset of the first character in a substring, and the
2122 second is set to the offset of the first character after the end of a
2123 substring. These values are always data unit offsets, even in UTF mode. They
2124 are byte offsets in the 8-bit library, 16-bit data item offsets in the 16-bit
2125 library, and 32-bit data item offsets in the 32-bit library. <b>Note</b>: they
2126 are not character counts.
2127 </P>
2128 <P>
2129 The first pair of integers, <i>ovector[0]</i> and <i>ovector[1]</i>, identify the
2130 portion of the subject string matched by the entire pattern. The next pair is
2131 used for the first capturing subpattern, and so on. The value returned by
2132 <b>pcre_exec()</b> is one more than the highest numbered pair that has been set.
2133 For example, if two substrings have been captured, the returned value is 3. If
2134 there are no capturing subpatterns, the return value from a successful match is
2135 1, indicating that just the first pair of offsets has been set.
2136 </P>
2137 <P>
2138 If a capturing subpattern is matched repeatedly, it is the last portion of the
2139 string that it matched that is returned.
2140 </P>
2141 <P>
2142 If the vector is too small to hold all the captured substring offsets, it is
2143 used as far as possible (up to two-thirds of its length), and the function
2144 returns a value of zero. If neither the actual string matched nor any captured
2145 substrings are of interest, <b>pcre_exec()</b> may be called with <i>ovector</i>
2146 passed as NULL and <i>ovecsize</i> as zero. However, if the pattern contains
2147 back references and the <i>ovector</i> is not big enough to remember the related
2148 substrings, PCRE has to get additional memory for use during matching. Thus it
2149 is usually advisable to supply an <i>ovector</i> of reasonable size.
2150 </P>
2151 <P>
2152 There are some cases where zero is returned (indicating vector overflow) when
2153 in fact the vector is exactly the right size for the final match. For example,
2154 consider the pattern
2155 <pre>
2156 (a)(?:(b)c|bd)
2157 </pre>
2158 If a vector of 6 elements (allowing for only 1 captured substring) is given
2159 with subject string "abd", <b>pcre_exec()</b> will try to set the second
2160 captured string, thereby recording a vector overflow, before failing to match
2161 "c" and backing up to try the second alternative. The zero return, however,
2162 does correctly indicate that the maximum number of slots (namely 2) have been
2163 filled. In similar cases where there is temporary overflow, but the final
2164 number of used slots is actually less than the maximum, a non-zero value is
2165 returned.
2166 </P>
2167 <P>
2168 The <b>pcre_fullinfo()</b> function can be used to find out how many capturing
2169 subpatterns there are in a compiled pattern. The smallest size for
2170 <i>ovector</i> that will allow for <i>n</i> captured substrings, in addition to
2171 the offsets of the substring matched by the whole pattern, is (<i>n</i>+1)*3.
2172 </P>
2173 <P>
2174 It is possible for capturing subpattern number <i>n+1</i> to match some part of
2175 the subject when subpattern <i>n</i> has not been used at all. For example, if
2176 the string "abc" is matched against the pattern (a|(z))(bc) the return from the
2177 function is 4, and subpatterns 1 and 3 are matched, but 2 is not. When this
2178 happens, both values in the offset pairs corresponding to unused subpatterns
2179 are set to -1.
2180 </P>
2181 <P>
2182 Offset values that correspond to unused subpatterns at the end of the
2183 expression are also set to -1. For example, if the string "abc" is matched
2184 against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not matched. The
2185 return from the function is 2, because the highest used capturing subpattern
2186 number is 1, and the offsets for for the second and third capturing subpatterns
2187 (assuming the vector is large enough, of course) are set to -1.
2188 </P>
2189 <P>
2190 <b>Note</b>: Elements in the first two-thirds of <i>ovector</i> that do not
2191 correspond to capturing parentheses in the pattern are never changed. That is,
2192 if a pattern contains <i>n</i> capturing parentheses, no more than
2193 <i>ovector[0]</i> to <i>ovector[2n+1]</i> are set by <b>pcre_exec()</b>. The other
2194 elements (in the first two-thirds) retain whatever values they previously had.
2195 </P>
2196 <P>
2197 Some convenience functions are provided for extracting the captured substrings
2198 as separate strings. These are described below.
2199 <a name="errorlist"></a></P>
2200 <br><b>
2201 Error return values from <b>pcre_exec()</b>
2202 </b><br>
2203 <P>
2204 If <b>pcre_exec()</b> fails, it returns a negative number. The following are
2205 defined in the header file:
2206 <pre>
2208 </pre>
2209 The subject string did not match the pattern.
2210 <pre>
2212 </pre>
2213 Either <i>code</i> or <i>subject</i> was passed as NULL, or <i>ovector</i> was
2214 NULL and <i>ovecsize</i> was not zero.
2215 <pre>
2217 </pre>
2218 An unrecognized bit was set in the <i>options</i> argument.
2219 <pre>
2221 </pre>
2222 PCRE stores a 4-byte "magic number" at the start of the compiled code, to catch
2223 the case when it is passed a junk pointer and to detect when a pattern that was
2224 compiled in an environment of one endianness is run in an environment with the
2225 other endianness. This is the error that PCRE gives when the magic number is
2226 not present.
2227 <pre>
2229 </pre>
2230 While running the pattern match, an unknown item was encountered in the
2231 compiled pattern. This error could be caused by a bug in PCRE or by overwriting
2232 of the compiled pattern.
2233 <pre>
2235 </pre>
2236 If a pattern contains back references, but the <i>ovector</i> that is passed to
2237 <b>pcre_exec()</b> is not big enough to remember the referenced substrings, PCRE
2238 gets a block of memory at the start of matching to use for this purpose. If the
2239 call via <b>pcre_malloc()</b> fails, this error is given. The memory is
2240 automatically freed at the end of matching.
2241 </P>
2242 <P>
2243 This error is also given if <b>pcre_stack_malloc()</b> fails in
2244 <b>pcre_exec()</b>. This can happen only when PCRE has been compiled with
2245 <b>--disable-stack-for-recursion</b>.
2246 <pre>
2248 </pre>
2249 This error is used by the <b>pcre_copy_substring()</b>,
2250 <b>pcre_get_substring()</b>, and <b>pcre_get_substring_list()</b> functions (see
2251 below). It is never returned by <b>pcre_exec()</b>.
2252 <pre>
2254 </pre>
2255 The backtracking limit, as specified by the <i>match_limit</i> field in a
2256 <b>pcre_extra</b> structure (or defaulted) was reached. See the description
2257 above.
2258 <pre>
2260 </pre>
2261 This error is never generated by <b>pcre_exec()</b> itself. It is provided for
2262 use by callout functions that want to yield a distinctive error code. See the
2263 <a href="pcrecallout.html"><b>pcrecallout</b></a>
2264 documentation for details.
2265 <pre>
2267 </pre>
2268 A string that contains an invalid UTF-8 byte sequence was passed as a subject,
2269 and the PCRE_NO_UTF8_CHECK option was not set. If the size of the output vector
2270 (<i>ovecsize</i>) is at least 2, the byte offset to the start of the the invalid
2271 UTF-8 character is placed in the first element, and a reason code is placed in
2272 the second element. The reason codes are listed in the
2273 <a href="#badutf8reasons">following section.</a>
2274 For backward compatibility, if PCRE_PARTIAL_HARD is set and the problem is a
2275 truncated UTF-8 character at the end of the subject (reason codes 1 to 5),
2276 PCRE_ERROR_SHORTUTF8 is returned instead of PCRE_ERROR_BADUTF8.
2277 <pre>
2279 </pre>
2280 The UTF-8 byte sequence that was passed as a subject was checked and found to
2281 be valid (the PCRE_NO_UTF8_CHECK option was not set), but the value of
2282 <i>startoffset</i> did not point to the beginning of a UTF-8 character or the
2283 end of the subject.
2284 <pre>
2286 </pre>
2287 The subject string did not match, but it did match partially. See the
2288 <a href="pcrepartial.html"><b>pcrepartial</b></a>
2289 documentation for details of partial matching.
2290 <pre>
2292 </pre>
2293 This code is no longer in use. It was formerly returned when the PCRE_PARTIAL
2294 option was used with a compiled pattern containing items that were not
2295 supported for partial matching. From release 8.00 onwards, there are no
2296 restrictions on partial matching.
2297 <pre>
2299 </pre>
2300 An unexpected internal error has occurred. This error could be caused by a bug
2301 in PCRE or by overwriting of the compiled pattern.
2302 <pre>
2304 </pre>
2305 This error is given if the value of the <i>ovecsize</i> argument is negative.
2306 <pre>
2308 </pre>
2309 The internal recursion limit, as specified by the <i>match_limit_recursion</i>
2310 field in a <b>pcre_extra</b> structure (or defaulted) was reached. See the
2311 description above.
2312 <pre>
2314 </pre>
2315 An invalid combination of PCRE_NEWLINE_<i>xxx</i> options was given.
2316 <pre>
2318 </pre>
2319 The value of <i>startoffset</i> was negative or greater than the length of the
2320 subject, that is, the value in <i>length</i>.
2321 <pre>
2323 </pre>
2324 This error is returned instead of PCRE_ERROR_BADUTF8 when the subject string
2325 ends with a truncated UTF-8 character and the PCRE_PARTIAL_HARD option is set.
2326 Information about the failure is returned as for PCRE_ERROR_BADUTF8. It is in
2327 fact sufficient to detect this case, but this special error code for
2328 PCRE_PARTIAL_HARD precedes the implementation of returned information; it is
2329 retained for backwards compatibility.
2330 <pre>
2332 </pre>
2333 This error is returned when <b>pcre_exec()</b> detects a recursion loop within
2334 the pattern. Specifically, it means that either the whole pattern or a
2335 subpattern has been called recursively for the second time at the same position
2336 in the subject string. Some simple patterns that might do this are detected and
2337 faulted at compile time, but more complicated cases, in particular mutual
2338 recursions between two different subpatterns, cannot be detected until run
2339 time.
2340 <pre>
2342 </pre>
2343 This error is returned when a pattern that was successfully studied using a
2344 JIT compile option is being matched, but the memory available for the
2345 just-in-time processing stack is not large enough. See the
2346 <a href="pcrejit.html"><b>pcrejit</b></a>
2347 documentation for more details.
2348 <pre>
2350 </pre>
2351 This error is given if a pattern that was compiled by the 8-bit library is
2352 passed to a 16-bit or 32-bit library function, or vice versa.
2353 <pre>
2355 </pre>
2356 This error is given if a pattern that was compiled and saved is reloaded on a
2357 host with different endianness. The utility function
2358 <b>pcre_pattern_to_host_byte_order()</b> can be used to convert such a pattern
2359 so that it runs on the new host.
2360 <pre>
2362 </pre>
2363 This error is returned when a pattern that was successfully studied using a JIT
2364 compile option is being matched, but the matching mode (partial or complete
2365 match) does not correspond to any JIT compilation mode. When the JIT fast path
2366 function is used, this error may be also given for invalid options. See the
2367 <a href="pcrejit.html"><b>pcrejit</b></a>
2368 documentation for more details.
2369 <pre>
2371 </pre>
2372 This error is given if <b>pcre_exec()</b> is called with a negative value for
2373 the <i>length</i> argument.
2374 </P>
2375 <P>
2376 Error numbers -16 to -20, -22, and 30 are not used by <b>pcre_exec()</b>.
2377 <a name="badutf8reasons"></a></P>
2378 <br><b>
2379 Reason codes for invalid UTF-8 strings
2380 </b><br>
2381 <P>
2382 This section applies only to the 8-bit library. The corresponding information
2383 for the 16-bit and 32-bit libraries is given in the
2384 <a href="pcre16.html"><b>pcre16</b></a>
2385 and
2386 <a href="pcre32.html"><b>pcre32</b></a>
2387 pages.
2388 </P>
2389 <P>
2390 When <b>pcre_exec()</b> returns either PCRE_ERROR_BADUTF8 or
2391 PCRE_ERROR_SHORTUTF8, and the size of the output vector (<i>ovecsize</i>) is at
2392 least 2, the offset of the start of the invalid UTF-8 character is placed in
2393 the first output vector element (<i>ovector[0]</i>) and a reason code is placed
2394 in the second element (<i>ovector[1]</i>). The reason codes are given names in
2395 the <b>pcre.h</b> header file:
2396 <pre>
2402 </pre>
2403 The string ends with a truncated UTF-8 character; the code specifies how many
2404 bytes are missing (1 to 5). Although RFC 3629 restricts UTF-8 characters to be
2405 no longer than 4 bytes, the encoding scheme (originally defined by RFC 2279)
2406 allows for up to 6 bytes, and this is checked first; hence the possibility of
2407 4 or 5 missing bytes.
2408 <pre>
2413 PCRE_UTF8_ERR10
2414 </pre>
2415 The two most significant bits of the 2nd, 3rd, 4th, 5th, or 6th byte of the
2416 character do not have the binary value 0b10 (that is, either the most
2417 significant bit is 0, or the next bit is 1).
2418 <pre>
2419 PCRE_UTF8_ERR11
2420 PCRE_UTF8_ERR12
2421 </pre>
2422 A character that is valid by the RFC 2279 rules is either 5 or 6 bytes long;
2423 these code points are excluded by RFC 3629.
2424 <pre>
2425 PCRE_UTF8_ERR13
2426 </pre>
2427 A 4-byte character has a value greater than 0x10fff; these code points are
2428 excluded by RFC 3629.
2429 <pre>
2430 PCRE_UTF8_ERR14
2431 </pre>
2432 A 3-byte character has a value in the range 0xd800 to 0xdfff; this range of
2433 code points are reserved by RFC 3629 for use with UTF-16, and so are excluded
2434 from UTF-8.
2435 <pre>
2436 PCRE_UTF8_ERR15
2437 PCRE_UTF8_ERR16
2438 PCRE_UTF8_ERR17
2439 PCRE_UTF8_ERR18
2440 PCRE_UTF8_ERR19
2441 </pre>
2442 A 2-, 3-, 4-, 5-, or 6-byte character is "overlong", that is, it codes for a
2443 value that can be represented by fewer bytes, which is invalid. For example,
2444 the two bytes 0xc0, 0xae give the value 0x2e, whose correct coding uses just
2445 one byte.
2446 <pre>
2447 PCRE_UTF8_ERR20
2448 </pre>
2449 The two most significant bits of the first byte of a character have the binary
2450 value 0b10 (that is, the most significant bit is 1 and the second is 0). Such a
2451 byte can only validly occur as the second or subsequent byte of a multi-byte
2452 character.
2453 <pre>
2454 PCRE_UTF8_ERR21
2455 </pre>
2456 The first byte of a character has the value 0xfe or 0xff. These values can
2457 never occur in a valid UTF-8 string.
2458 <pre>
2459 PCRE_UTF8_ERR22
2460 </pre>
2461 This error code was formerly used when the presence of a so-called
2462 "non-character" caused an error. Unicode corrigendum #9 makes it clear that
2463 such characters should not cause a string to be rejected, and so this code is
2464 no longer in use and is never returned.
2465 </P>
2466 <br><a name="SEC18" href="#TOC1">EXTRACTING CAPTURED SUBSTRINGS BY NUMBER</a><br>
2467 <P>
2468 <b>int pcre_copy_substring(const char *<i>subject</i>, int *<i>ovector</i>,</b>
2469 <b> int <i>stringcount</i>, int <i>stringnumber</i>, char *<i>buffer</i>,</b>
2470 <b> int <i>buffersize</i>);</b>
2471 <br>
2472 <br>
2473 <b>int pcre_get_substring(const char *<i>subject</i>, int *<i>ovector</i>,</b>
2474 <b> int <i>stringcount</i>, int <i>stringnumber</i>,</b>
2475 <b> const char **<i>stringptr</i>);</b>
2476 <br>
2477 <br>
2478 <b>int pcre_get_substring_list(const char *<i>subject</i>,</b>
2479 <b> int *<i>ovector</i>, int <i>stringcount</i>, const char ***<i>listptr</i>);</b>
2480 </P>
2481 <P>
2482 Captured substrings can be accessed directly by using the offsets returned by
2483 <b>pcre_exec()</b> in <i>ovector</i>. For convenience, the functions
2484 <b>pcre_copy_substring()</b>, <b>pcre_get_substring()</b>, and
2485 <b>pcre_get_substring_list()</b> are provided for extracting captured substrings
2486 as new, separate, zero-terminated strings. These functions identify substrings
2487 by number. The next section describes functions for extracting named
2488 substrings.
2489 </P>
2490 <P>
2491 A substring that contains a binary zero is correctly extracted and has a
2492 further zero added on the end, but the result is not, of course, a C string.
2493 However, you can process such a string by referring to the length that is
2494 returned by <b>pcre_copy_substring()</b> and <b>pcre_get_substring()</b>.
2495 Unfortunately, the interface to <b>pcre_get_substring_list()</b> is not adequate
2496 for handling strings containing binary zeros, because the end of the final
2497 string is not independently indicated.
2498 </P>
2499 <P>
2500 The first three arguments are the same for all three of these functions:
2501 <i>subject</i> is the subject string that has just been successfully matched,
2502 <i>ovector</i> is a pointer to the vector of integer offsets that was passed to
2503 <b>pcre_exec()</b>, and <i>stringcount</i> is the number of substrings that were
2504 captured by the match, including the substring that matched the entire regular
2505 expression. This is the value returned by <b>pcre_exec()</b> if it is greater
2506 than zero. If <b>pcre_exec()</b> returned zero, indicating that it ran out of
2507 space in <i>ovector</i>, the value passed as <i>stringcount</i> should be the
2508 number of elements in the vector divided by three.
2509 </P>
2510 <P>
2511 The functions <b>pcre_copy_substring()</b> and <b>pcre_get_substring()</b>
2512 extract a single substring, whose number is given as <i>stringnumber</i>. A
2513 value of zero extracts the substring that matched the entire pattern, whereas
2514 higher values extract the captured substrings. For <b>pcre_copy_substring()</b>,
2515 the string is placed in <i>buffer</i>, whose length is given by
2516 <i>buffersize</i>, while for <b>pcre_get_substring()</b> a new block of memory is
2517 obtained via <b>pcre_malloc</b>, and its address is returned via
2518 <i>stringptr</i>. The yield of the function is the length of the string, not
2519 including the terminating zero, or one of these error codes:
2520 <pre>
2522 </pre>
2523 The buffer was too small for <b>pcre_copy_substring()</b>, or the attempt to get
2524 memory failed for <b>pcre_get_substring()</b>.
2525 <pre>
2527 </pre>
2528 There is no substring whose number is <i>stringnumber</i>.
2529 </P>
2530 <P>
2531 The <b>pcre_get_substring_list()</b> function extracts all available substrings
2532 and builds a list of pointers to them. All this is done in a single block of
2533 memory that is obtained via <b>pcre_malloc</b>. The address of the memory block
2534 is returned via <i>listptr</i>, which is also the start of the list of string
2535 pointers. The end of the list is marked by a NULL pointer. The yield of the
2536 function is zero if all went well, or the error code
2537 <pre>
2539 </pre>
2540 if the attempt to get the memory block failed.
2541 </P>
2542 <P>
2543 When any of these functions encounter a substring that is unset, which can
2544 happen when capturing subpattern number <i>n+1</i> matches some part of the
2545 subject, but subpattern <i>n</i> has not been used at all, they return an empty
2546 string. This can be distinguished from a genuine zero-length substring by
2547 inspecting the appropriate offset in <i>ovector</i>, which is negative for unset
2548 substrings.
2549 </P>
2550 <P>
2551 The two convenience functions <b>pcre_free_substring()</b> and
2552 <b>pcre_free_substring_list()</b> can be used to free the memory returned by
2553 a previous call of <b>pcre_get_substring()</b> or
2554 <b>pcre_get_substring_list()</b>, respectively. They do nothing more than call
2555 the function pointed to by <b>pcre_free</b>, which of course could be called
2556 directly from a C program. However, PCRE is used in some situations where it is
2557 linked via a special interface to another programming language that cannot use
2558 <b>pcre_free</b> directly; it is for these cases that the functions are
2559 provided.
2560 </P>
2561 <br><a name="SEC19" href="#TOC1">EXTRACTING CAPTURED SUBSTRINGS BY NAME</a><br>
2562 <P>
2563 <b>int pcre_get_stringnumber(const pcre *<i>code</i>,</b>
2564 <b> const char *<i>name</i>);</b>
2565 <br>
2566 <br>
2567 <b>int pcre_copy_named_substring(const pcre *<i>code</i>,</b>
2568 <b> const char *<i>subject</i>, int *<i>ovector</i>,</b>
2569 <b> int <i>stringcount</i>, const char *<i>stringname</i>,</b>
2570 <b> char *<i>buffer</i>, int <i>buffersize</i>);</b>
2571 <br>
2572 <br>
2573 <b>int pcre_get_named_substring(const pcre *<i>code</i>,</b>
2574 <b> const char *<i>subject</i>, int *<i>ovector</i>,</b>
2575 <b> int <i>stringcount</i>, const char *<i>stringname</i>,</b>
2576 <b> const char **<i>stringptr</i>);</b>
2577 </P>
2578 <P>
2579 To extract a substring by name, you first have to find associated number.
2580 For example, for this pattern
2581 <pre>
2582 (a+)b(?&#60;xxx&#62;\d+)...
2583 </pre>
2584 the number of the subpattern called "xxx" is 2. If the name is known to be
2585 unique (PCRE_DUPNAMES was not set), you can find the number from the name by
2586 calling <b>pcre_get_stringnumber()</b>. The first argument is the compiled
2587 pattern, and the second is the name. The yield of the function is the
2588 subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no subpattern of
2589 that name.
2590 </P>
2591 <P>
2592 Given the number, you can extract the substring directly, or use one of the
2593 functions described in the previous section. For convenience, there are also
2594 two functions that do the whole job.
2595 </P>
2596 <P>
2597 Most of the arguments of <b>pcre_copy_named_substring()</b> and
2598 <b>pcre_get_named_substring()</b> are the same as those for the similarly named
2599 functions that extract by number. As these are described in the previous
2600 section, they are not re-described here. There are just two differences:
2601 </P>
2602 <P>
2603 First, instead of a substring number, a substring name is given. Second, there
2604 is an extra argument, given at the start, which is a pointer to the compiled
2605 pattern. This is needed in order to gain access to the name-to-number
2606 translation table.
2607 </P>
2608 <P>
2609 These functions call <b>pcre_get_stringnumber()</b>, and if it succeeds, they
2610 then call <b>pcre_copy_substring()</b> or <b>pcre_get_substring()</b>, as
2611 appropriate. <b>NOTE:</b> If PCRE_DUPNAMES is set and there are duplicate names,
2612 the behaviour may not be what you want (see the next section).
2613 </P>
2614 <P>
2615 <b>Warning:</b> If the pattern uses the (?| feature to set up multiple
2616 subpatterns with the same number, as described in the
2617 <a href="pcrepattern.html#dupsubpatternnumber">section on duplicate subpattern numbers</a>
2618 in the
2619 <a href="pcrepattern.html"><b>pcrepattern</b></a>
2620 page, you cannot use names to distinguish the different subpatterns, because
2621 names are not included in the compiled code. The matching process uses only
2622 numbers. For this reason, the use of different names for subpatterns of the
2623 same number causes an error at compile time.
2624 </P>
2625 <br><a name="SEC20" href="#TOC1">DUPLICATE SUBPATTERN NAMES</a><br>
2626 <P>
2627 <b>int pcre_get_stringtable_entries(const pcre *<i>code</i>,</b>
2628 <b> const char *<i>name</i>, char **<i>first</i>, char **<i>last</i>);</b>
2629 </P>
2630 <P>
2631 When a pattern is compiled with the PCRE_DUPNAMES option, names for subpatterns
2632 are not required to be unique. (Duplicate names are always allowed for
2633 subpatterns with the same number, created by using the (?| feature. Indeed, if
2634 such subpatterns are named, they are required to use the same names.)
2635 </P>
2636 <P>
2637 Normally, patterns with duplicate names are such that in any one match, only
2638 one of the named subpatterns participates. An example is shown in the
2639 <a href="pcrepattern.html"><b>pcrepattern</b></a>
2640 documentation.
2641 </P>
2642 <P>
2643 When duplicates are present, <b>pcre_copy_named_substring()</b> and
2644 <b>pcre_get_named_substring()</b> return the first substring corresponding to
2645 the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING (-7) is
2646 returned; no data is returned. The <b>pcre_get_stringnumber()</b> function
2647 returns one of the numbers that are associated with the name, but it is not
2648 defined which it is.
2649 </P>
2650 <P>
2651 If you want to get full details of all captured substrings for a given name,
2652 you must use the <b>pcre_get_stringtable_entries()</b> function. The first
2653 argument is the compiled pattern, and the second is the name. The third and
2654 fourth are pointers to variables which are updated by the function. After it
2655 has run, they point to the first and last entries in the name-to-number table
2656 for the given name. The function itself returns the length of each entry, or
2657 PCRE_ERROR_NOSUBSTRING (-7) if there are none. The format of the table is
2658 described above in the section entitled <i>Information about a pattern</i>
2659 <a href="#infoaboutpattern">above.</a>
2660 Given all the relevant entries for the name, you can extract each of their
2661 numbers, and hence the captured data, if any.
2662 </P>
2663 <br><a name="SEC21" href="#TOC1">FINDING ALL POSSIBLE MATCHES</a><br>
2664 <P>
2665 The traditional matching function uses a similar algorithm to Perl, which stops
2666 when it finds the first match, starting at a given point in the subject. If you
2667 want to find all possible matches, or the longest possible match, consider
2668 using the alternative matching function (see below) instead. If you cannot use
2669 the alternative function, but still need to find all possible matches, you
2670 can kludge it up by making use of the callout facility, which is described in
2671 the
2672 <a href="pcrecallout.html"><b>pcrecallout</b></a>
2673 documentation.
2674 </P>
2675 <P>
2676 What you have to do is to insert a callout right at the end of the pattern.
2677 When your callout function is called, extract and save the current matched
2678 substring. Then return 1, which forces <b>pcre_exec()</b> to backtrack and try
2679 other alternatives. Ultimately, when it runs out of matches, <b>pcre_exec()</b>
2680 will yield PCRE_ERROR_NOMATCH.
2681 </P>
2682 <br><a name="SEC22" href="#TOC1">OBTAINING AN ESTIMATE OF STACK USAGE</a><br>
2683 <P>
2684 Matching certain patterns using <b>pcre_exec()</b> can use a lot of process
2685 stack, which in certain environments can be rather limited in size. Some users
2686 find it helpful to have an estimate of the amount of stack that is used by
2687 <b>pcre_exec()</b>, to help them set recursion limits, as described in the
2688 <a href="pcrestack.html"><b>pcrestack</b></a>
2689 documentation. The estimate that is output by <b>pcretest</b> when called with
2690 the <b>-m</b> and <b>-C</b> options is obtained by calling <b>pcre_exec</b> with
2691 the values NULL, NULL, NULL, -999, and -999 for its first five arguments.
2692 </P>
2693 <P>
2694 Normally, if its first argument is NULL, <b>pcre_exec()</b> immediately returns
2695 the negative error code PCRE_ERROR_NULL, but with this special combination of
2696 arguments, it returns instead a negative number whose absolute value is the
2697 approximate stack frame size in bytes. (A negative number is used so that it is
2698 clear that no match has happened.) The value is approximate because in some
2699 cases, recursive calls to <b>pcre_exec()</b> occur when there are one or two
2700 additional variables on the stack.
2701 </P>
2702 <P>
2703 If PCRE has been compiled to use the heap instead of the stack for recursion,
2704 the value returned is the size of each block that is obtained from the heap.
2705 <a name="dfamatch"></a></P>
2706 <br><a name="SEC23" href="#TOC1">MATCHING A PATTERN: THE ALTERNATIVE FUNCTION</a><br>
2707 <P>
2708 <b>int pcre_dfa_exec(const pcre *<i>code</i>, const pcre_extra *<i>extra</i>,</b>
2709 <b> const char *<i>subject</i>, int <i>length</i>, int <i>startoffset</i>,</b>
2710 <b> int <i>options</i>, int *<i>ovector</i>, int <i>ovecsize</i>,</b>
2711 <b> int *<i>workspace</i>, int <i>wscount</i>);</b>
2712 </P>
2713 <P>
2714 The function <b>pcre_dfa_exec()</b> is called to match a subject string against
2715 a compiled pattern, using a matching algorithm that scans the subject string
2716 just once, and does not backtrack. This has different characteristics to the
2717 normal algorithm, and is not compatible with Perl. Some of the features of PCRE
2718 patterns are not supported. Nevertheless, there are times when this kind of
2719 matching can be useful. For a discussion of the two matching algorithms, and a
2720 list of features that <b>pcre_dfa_exec()</b> does not support, see the
2721 <a href="pcrematching.html"><b>pcrematching</b></a>
2722 documentation.
2723 </P>
2724 <P>
2725 The arguments for the <b>pcre_dfa_exec()</b> function are the same as for
2726 <b>pcre_exec()</b>, plus two extras. The <i>ovector</i> argument is used in a
2727 different way, and this is described below. The other common arguments are used
2728 in the same way as for <b>pcre_exec()</b>, so their description is not repeated
2729 here.
2730 </P>
2731 <P>
2732 The two additional arguments provide workspace for the function. The workspace
2733 vector should contain at least 20 elements. It is used for keeping track of
2734 multiple paths through the pattern tree. More workspace will be needed for
2735 patterns and subjects where there are a lot of potential matches.
2736 </P>
2737 <P>
2738 Here is an example of a simple call to <b>pcre_dfa_exec()</b>:
2739 <pre>
2740 int rc;
2741 int ovector[10];
2742 int wspace[20];
2743 rc = pcre_dfa_exec(
2744 re, /* result of pcre_compile() */
2745 NULL, /* we didn't study the pattern */
2746 "some string", /* the subject string */
2747 11, /* the length of the subject string */
2748 0, /* start at offset 0 in the subject */
2749 0, /* default options */
2750 ovector, /* vector of integers for substring information */
2751 10, /* number of elements (NOT size in bytes) */
2752 wspace, /* working space vector */
2753 20); /* number of elements (NOT size in bytes) */
2754 </PRE>
2755 </P>
2756 <br><b>
2757 Option bits for <b>pcre_dfa_exec()</b>
2758 </b><br>
2759 <P>
2760 The unused bits of the <i>options</i> argument for <b>pcre_dfa_exec()</b> must be
2761 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_<i>xxx</i>,
2765 All but the last four of these are exactly the same as for <b>pcre_exec()</b>,
2766 so their description is not repeated here.
2767 <pre>
2770 </pre>
2771 These have the same general effect as they do for <b>pcre_exec()</b>, but the
2772 details are slightly different. When PCRE_PARTIAL_HARD is set for
2773 <b>pcre_dfa_exec()</b>, it returns PCRE_ERROR_PARTIAL if the end of the subject
2774 is reached and there is still at least one matching possibility that requires
2775 additional characters. This happens even if some complete matches have also
2776 been found. When PCRE_PARTIAL_SOFT is set, the return code PCRE_ERROR_NOMATCH
2777 is converted into PCRE_ERROR_PARTIAL if the end of the subject is reached,
2778 there have been no complete matches, but there is still at least one matching
2779 possibility. The portion of the string that was inspected when the longest
2780 partial match was found is set as the first matching string in both cases.
2781 There is a more detailed discussion of partial and multi-segment matching, with
2782 examples, in the
2783 <a href="pcrepartial.html"><b>pcrepartial</b></a>
2784 documentation.
2785 <pre>
2787 </pre>
2788 Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to stop as
2789 soon as it has found one match. Because of the way the alternative algorithm
2790 works, this is necessarily the shortest possible match at the first possible
2791 matching point in the subject string.
2792 <pre>
2794 </pre>
2795 When <b>pcre_dfa_exec()</b> returns a partial match, it is possible to call it
2796 again, with additional subject characters, and have it continue with the same
2797 match. The PCRE_DFA_RESTART option requests this action; when it is set, the
2798 <i>workspace</i> and <i>wscount</i> options must reference the same vector as
2799 before because data about the match so far is left in them after a partial
2800 match. There is more discussion of this facility in the
2801 <a href="pcrepartial.html"><b>pcrepartial</b></a>
2802 documentation.
2803 </P>
2804 <br><b>
2805 Successful returns from <b>pcre_dfa_exec()</b>
2806 </b><br>
2807 <P>
2808 When <b>pcre_dfa_exec()</b> succeeds, it may have matched more than one
2809 substring in the subject. Note, however, that all the matches from one run of
2810 the function start at the same point in the subject. The shorter matches are
2811 all initial substrings of the longer matches. For example, if the pattern
2812 <pre>
2813 &#60;.*&#62;
2814 </pre>
2815 is matched against the string
2816 <pre>
2817 This is &#60;something&#62; &#60;something else&#62; &#60;something further&#62; no more
2818 </pre>
2819 the three matched strings are
2820 <pre>
2821 &#60;something&#62;
2822 &#60;something&#62; &#60;something else&#62;
2823 &#60;something&#62; &#60;something else&#62; &#60;something further&#62;
2824 </pre>
2825 On success, the yield of the function is a number greater than zero, which is
2826 the number of matched substrings. The substrings themselves are returned in
2827 <i>ovector</i>. Each string uses two elements; the first is the offset to the
2828 start, and the second is the offset to the end. In fact, all the strings have
2829 the same start offset. (Space could have been saved by giving this only once,
2830 but it was decided to retain some compatibility with the way <b>pcre_exec()</b>
2831 returns data, even though the meaning of the strings is different.)
2832 </P>
2833 <P>
2834 The strings are returned in reverse order of length; that is, the longest
2835 matching string is given first. If there were too many matches to fit into
2836 <i>ovector</i>, the yield of the function is zero, and the vector is filled with
2837 the longest matches. Unlike <b>pcre_exec()</b>, <b>pcre_dfa_exec()</b> can use
2838 the entire <i>ovector</i> for returning matched strings.
2839 </P>
2840 <P>
2841 NOTE: PCRE's "auto-possessification" optimization usually applies to character
2842 repeats at the end of a pattern (as well as internally). For example, the
2843 pattern "a\d+" is compiled as if it were "a\d++" because there is no point
2844 even considering the possibility of backtracking into the repeated digits. For
2845 DFA matching, this means that only one possible match is found. If you really
2846 do want multiple matches in such cases, either use an ungreedy repeat
2847 ("a\d+?") or set the PCRE_NO_AUTO_POSSESS option when compiling.
2848 </P>
2849 <br><b>
2850 Error returns from <b>pcre_dfa_exec()</b>
2851 </b><br>
2852 <P>
2853 The <b>pcre_dfa_exec()</b> function returns a negative number when it fails.
2854 Many of the errors are the same as for <b>pcre_exec()</b>, and these are
2855 described
2856 <a href="#errorlist">above.</a>
2857 There are in addition the following errors that are specific to
2858 <b>pcre_dfa_exec()</b>:
2859 <pre>
2861 </pre>
2862 This return is given if <b>pcre_dfa_exec()</b> encounters an item in the pattern
2863 that it does not support, for instance, the use of \C or a back reference.
2864 <pre>
2866 </pre>
2867 This return is given if <b>pcre_dfa_exec()</b> encounters a condition item that
2868 uses a back reference for the condition, or a test for recursion in a specific
2869 group. These are not supported.
2870 <pre>
2872 </pre>
2873 This return is given if <b>pcre_dfa_exec()</b> is called with an <i>extra</i>
2874 block that contains a setting of the <i>match_limit</i> or
2875 <i>match_limit_recursion</i> fields. This is not supported (these fields are
2876 meaningless for DFA matching).
2877 <pre>
2879 </pre>
2880 This return is given if <b>pcre_dfa_exec()</b> runs out of space in the
2881 <i>workspace</i> vector.
2882 <pre>
2884 </pre>
2885 When a recursive subpattern is processed, the matching function calls itself
2886 recursively, using private vectors for <i>ovector</i> and <i>workspace</i>. This
2887 error is given if the output vector is not large enough. This should be
2888 extremely rare, as a vector of size 1000 is used.
2889 <pre>
2891 </pre>
2892 When <b>pcre_dfa_exec()</b> is called with the <b>PCRE_DFA_RESTART</b> option,
2893 some plausibility checks are made on the contents of the workspace, which
2894 should contain data about the previous partial match. If any of these checks
2895 fail, this error is given.
2896 </P>
2897 <br><a name="SEC24" href="#TOC1">SEE ALSO</a><br>
2898 <P>
2899 <b>pcre16</b>(3), <b>pcre32</b>(3), <b>pcrebuild</b>(3), <b>pcrecallout</b>(3),
2900 <b>pcrecpp(3)</b>(3), <b>pcrematching</b>(3), <b>pcrepartial</b>(3),
2901 <b>pcreposix</b>(3), <b>pcreprecompile</b>(3), <b>pcresample</b>(3),
2902 <b>pcrestack</b>(3).
2903 </P>
2904 <br><a name="SEC25" href="#TOC1">AUTHOR</a><br>
2905 <P>
2906 Philip Hazel
2907 <br>
2908 University Computing Service
2909 <br>
2910 Cambridge CB2 3QH, England.
2911 <br>
2912 </P>
2913 <br><a name="SEC26" href="#TOC1">REVISION</a><br>
2914 <P>
2915 Last updated: 18 December 2015
2916 <br>
2917 Copyright &copy; 1997-2015 University of Cambridge.
2918 <br>
2919 <p>
2920 Return to the <a href="index.html">PCRE index page</a>.
2921 </p>


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

  ViewVC Help
Powered by ViewVC 1.1.5