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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 968 - (hide annotations) (download)
Sat May 26 14:22:52 2012 UTC (2 years, 6 months ago) by ph10
File size: 115845 byte(s)
Spelling corrections in documentation.

1 ph10 964 .TH PCREAPI 3 "04 May 2012" "PCRE 8.31"
2 nigel 63 .SH NAME
3     PCRE - Perl-compatible regular expressions
4 ph10 856 .sp
5     .B #include <pcre.h>
6     .
7     .
8 ph10 678 .SH "PCRE NATIVE API BASIC FUNCTIONS"
9 nigel 63 .rs
10     .sp
11     .SM
12 nigel 75 .B pcre *pcre_compile(const char *\fIpattern\fP, int \fIoptions\fP,
13 nigel 63 .ti +5n
14 nigel 75 .B const char **\fIerrptr\fP, int *\fIerroffset\fP,
15 nigel 63 .ti +5n
16 nigel 75 .B const unsigned char *\fItableptr\fP);
17 nigel 63 .PP
18 nigel 77 .B pcre *pcre_compile2(const char *\fIpattern\fP, int \fIoptions\fP,
19     .ti +5n
20     .B int *\fIerrorcodeptr\fP,
21     .ti +5n
22     .B const char **\fIerrptr\fP, int *\fIerroffset\fP,
23     .ti +5n
24     .B const unsigned char *\fItableptr\fP);
25     .PP
26 nigel 75 .B pcre_extra *pcre_study(const pcre *\fIcode\fP, int \fIoptions\fP,
27 nigel 63 .ti +5n
28 nigel 75 .B const char **\fIerrptr\fP);
29 nigel 63 .PP
30 ph10 678 .B void pcre_free_study(pcre_extra *\fIextra\fP);
31     .PP
32 nigel 75 .B int pcre_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
33 nigel 63 .ti +5n
34 nigel 75 .B "const char *\fIsubject\fP," int \fIlength\fP, int \fIstartoffset\fP,
35 nigel 63 .ti +5n
36 nigel 75 .B int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP);
37 nigel 63 .PP
38 nigel 77 .B int pcre_dfa_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
39     .ti +5n
40     .B "const char *\fIsubject\fP," int \fIlength\fP, int \fIstartoffset\fP,
41     .ti +5n
42     .B int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP,
43     .ti +5n
44     .B int *\fIworkspace\fP, int \fIwscount\fP);
45 ph10 856 .
46     .
47     .SH "PCRE NATIVE API STRING EXTRACTION FUNCTIONS"
48     .rs
49     .sp
50 nigel 75 .B int pcre_copy_named_substring(const pcre *\fIcode\fP,
51 nigel 63 .ti +5n
52 nigel 75 .B const char *\fIsubject\fP, int *\fIovector\fP,
53 nigel 63 .ti +5n
54 nigel 75 .B int \fIstringcount\fP, const char *\fIstringname\fP,
55 nigel 63 .ti +5n
56 nigel 75 .B char *\fIbuffer\fP, int \fIbuffersize\fP);
57 nigel 63 .PP
58 nigel 75 .B int pcre_copy_substring(const char *\fIsubject\fP, int *\fIovector\fP,
59 nigel 63 .ti +5n
60 nigel 75 .B int \fIstringcount\fP, int \fIstringnumber\fP, char *\fIbuffer\fP,
61 nigel 63 .ti +5n
62 nigel 75 .B int \fIbuffersize\fP);
63 nigel 63 .PP
64 nigel 75 .B int pcre_get_named_substring(const pcre *\fIcode\fP,
65 nigel 63 .ti +5n
66 nigel 75 .B const char *\fIsubject\fP, int *\fIovector\fP,
67 nigel 63 .ti +5n
68 nigel 75 .B int \fIstringcount\fP, const char *\fIstringname\fP,
69 nigel 63 .ti +5n
70 nigel 75 .B const char **\fIstringptr\fP);
71 nigel 63 .PP
72 nigel 75 .B int pcre_get_stringnumber(const pcre *\fIcode\fP,
73 nigel 63 .ti +5n
74 nigel 75 .B const char *\fIname\fP);
75 nigel 63 .PP
76 nigel 91 .B int pcre_get_stringtable_entries(const pcre *\fIcode\fP,
77     .ti +5n
78     .B const char *\fIname\fP, char **\fIfirst\fP, char **\fIlast\fP);
79     .PP
80 nigel 75 .B int pcre_get_substring(const char *\fIsubject\fP, int *\fIovector\fP,
81 nigel 63 .ti +5n
82 nigel 75 .B int \fIstringcount\fP, int \fIstringnumber\fP,
83 nigel 63 .ti +5n
84 nigel 75 .B const char **\fIstringptr\fP);
85 nigel 63 .PP
86 nigel 75 .B int pcre_get_substring_list(const char *\fIsubject\fP,
87 nigel 63 .ti +5n
88 nigel 75 .B int *\fIovector\fP, int \fIstringcount\fP, "const char ***\fIlistptr\fP);"
89 nigel 63 .PP
90 nigel 75 .B void pcre_free_substring(const char *\fIstringptr\fP);
91 nigel 63 .PP
92 nigel 75 .B void pcre_free_substring_list(const char **\fIstringptr\fP);
93 ph10 856 .
94     .
95     .SH "PCRE NATIVE API AUXILIARY FUNCTIONS"
96     .rs
97     .sp
98     .B pcre_jit_stack *pcre_jit_stack_alloc(int \fIstartsize\fP, int \fImaxsize\fP);
99 nigel 63 .PP
100 ph10 856 .B void pcre_jit_stack_free(pcre_jit_stack *\fIstack\fP);
101     .PP
102     .B void pcre_assign_jit_stack(pcre_extra *\fIextra\fP,
103     .ti +5n
104     .B pcre_jit_callback \fIcallback\fP, void *\fIdata\fP);
105     .PP
106 nigel 63 .B const unsigned char *pcre_maketables(void);
107     .PP
108 nigel 75 .B int pcre_fullinfo(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
109 nigel 63 .ti +5n
110 nigel 75 .B int \fIwhat\fP, void *\fIwhere\fP);
111 nigel 63 .PP
112 nigel 77 .B int pcre_refcount(pcre *\fIcode\fP, int \fIadjust\fP);
113     .PP
114 nigel 75 .B int pcre_config(int \fIwhat\fP, void *\fIwhere\fP);
115 nigel 63 .PP
116 ph10 856 .B const char *pcre_version(void);
117     .PP
118     .B int pcre_pattern_to_host_byte_order(pcre *\fIcode\fP,
119     .ti +5n
120     .B pcre_extra *\fIextra\fP, const unsigned char *\fItables\fP);
121 ph10 686 .
122     .
123     .SH "PCRE NATIVE API INDIRECTED FUNCTIONS"
124     .rs
125     .sp
126 nigel 63 .B void *(*pcre_malloc)(size_t);
127     .PP
128     .B void (*pcre_free)(void *);
129     .PP
130 nigel 73 .B void *(*pcre_stack_malloc)(size_t);
131     .PP
132     .B void (*pcre_stack_free)(void *);
133     .PP
134 nigel 63 .B int (*pcre_callout)(pcre_callout_block *);
135 nigel 75 .
136     .
137 ph10 856 .SH "PCRE 8-BIT AND 16-BIT LIBRARIES"
138     .rs
139     .sp
140     From release 8.30, PCRE can be compiled as a library for handling 16-bit
141     character strings as well as, or instead of, the original library that handles
142     8-bit character strings. To avoid too much complication, this document
143     describes the 8-bit versions of the functions, with only occasional references
144     to the 16-bit library.
145     .P
146     The 16-bit functions operate in the same way as their 8-bit counterparts; they
147     just use different data types for their arguments and results, and their names
148     start with \fBpcre16_\fP instead of \fBpcre_\fP. For every option that has UTF8
149     in its name (for example, PCRE_UTF8), there is a corresponding 16-bit name with
150     UTF8 replaced by UTF16. This facility is in fact just cosmetic; the 16-bit
151 ph10 903 option names define the same bit values.
152 ph10 856 .P
153     References to bytes and UTF-8 in this document should be read as references to
154     16-bit data quantities and UTF-16 when using the 16-bit library, unless
155     specified otherwise. More details of the specific differences for the 16-bit
156     library are given in the
157     .\" HREF
158     \fBpcre16\fP
159     .\"
160 ph10 903 page.
161 ph10 856 .
162     .
163 nigel 75 .SH "PCRE API OVERVIEW"
164 nigel 63 .rs
165     .sp
166 nigel 93 PCRE has its own native API, which is described in this document. There are
167 ph10 856 also some wrapper functions (for the 8-bit library only) that correspond to the
168     POSIX regular expression API, but they do not give access to all the
169     functionality. They are described in the
170 nigel 75 .\" HREF
171     \fBpcreposix\fP
172     .\"
173 nigel 77 documentation. Both of these APIs define a set of C function calls. A C++
174 ph10 856 wrapper (again for the 8-bit library only) is also distributed with PCRE. It is
175     documented in the
176 nigel 77 .\" HREF
177     \fBpcrecpp\fP
178     .\"
179     page.
180 nigel 75 .P
181 nigel 77 The native API C function prototypes are defined in the header file
182 ph10 856 \fBpcre.h\fP, and on Unix-like systems the (8-bit) library itself is called
183     \fBlibpcre\fP. It can normally be accessed by adding \fB-lpcre\fP to the
184     command for linking an application that uses PCRE. The header file defines the
185     macros PCRE_MAJOR and PCRE_MINOR to contain the major and minor release numbers
186     for the library. Applications can use these to include support for different
187     releases of PCRE.
188 nigel 75 .P
189 ph10 525 In a Windows environment, if you want to statically link an application program
190     against a non-dll \fBpcre.a\fP file, you must define PCRE_STATIC before
191 ph10 526 including \fBpcre.h\fP or \fBpcrecpp.h\fP, because otherwise the
192     \fBpcre_malloc()\fP and \fBpcre_free()\fP exported functions will be declared
193 ph10 525 \fB__declspec(dllimport)\fP, with unwanted results.
194     .P
195 nigel 77 The functions \fBpcre_compile()\fP, \fBpcre_compile2()\fP, \fBpcre_study()\fP,
196     and \fBpcre_exec()\fP are used for compiling and matching regular expressions
197     in a Perl-compatible manner. A sample program that demonstrates the simplest
198 ph10 429 way of using them is provided in the file called \fIpcredemo.c\fP in the PCRE
199     source distribution. A listing of this program is given in the
200 nigel 75 .\" HREF
201 ph10 429 \fBpcredemo\fP
202     .\"
203     documentation, and the
204     .\" HREF
205 nigel 75 \fBpcresample\fP
206     .\"
207 ph10 312 documentation describes how to compile and run it.
208 nigel 75 .P
209 ph10 678 Just-in-time compiler support is an optional feature of PCRE that can be built
210 ph10 686 in appropriate hardware environments. It greatly speeds up the matching
211     performance of many patterns. Simple programs can easily request that it be
212     used if available, by setting an option that is ignored when it is not
213     relevant. More complicated programs might need to make use of the functions
214     \fBpcre_jit_stack_alloc()\fP, \fBpcre_jit_stack_free()\fP, and
215     \fBpcre_assign_jit_stack()\fP in order to control the JIT code's memory usage.
216 ph10 678 These functions are discussed in the
217     .\" HREF
218     \fBpcrejit\fP
219     .\"
220     documentation.
221     .P
222 nigel 77 A second matching function, \fBpcre_dfa_exec()\fP, which is not
223     Perl-compatible, is also provided. This uses a different algorithm for the
224 nigel 91 matching. The alternative algorithm finds all possible matches (at a given
225 ph10 435 point in the subject), and scans the subject just once (unless there are
226     lookbehind assertions). However, this algorithm does not return captured
227     substrings. A description of the two matching algorithms and their advantages
228     and disadvantages is given in the
229 nigel 77 .\" HREF
230     \fBpcrematching\fP
231     .\"
232     documentation.
233     .P
234 nigel 75 In addition to the main compiling and matching functions, there are convenience
235 nigel 77 functions for extracting captured substrings from a subject string that is
236     matched by \fBpcre_exec()\fP. They are:
237 nigel 75 .sp
238     \fBpcre_copy_substring()\fP
239     \fBpcre_copy_named_substring()\fP
240     \fBpcre_get_substring()\fP
241     \fBpcre_get_named_substring()\fP
242     \fBpcre_get_substring_list()\fP
243     \fBpcre_get_stringnumber()\fP
244 nigel 91 \fBpcre_get_stringtable_entries()\fP
245 nigel 75 .sp
246     \fBpcre_free_substring()\fP and \fBpcre_free_substring_list()\fP are also
247 nigel 63 provided, to free the memory used for extracted strings.
248 nigel 75 .P
249     The function \fBpcre_maketables()\fP is used to build a set of character tables
250 nigel 77 in the current locale for passing to \fBpcre_compile()\fP, \fBpcre_exec()\fP,
251     or \fBpcre_dfa_exec()\fP. This is an optional facility that is provided for
252     specialist use. Most commonly, no special tables are passed, in which case
253     internal tables that are generated when PCRE is built are used.
254 nigel 75 .P
255     The function \fBpcre_fullinfo()\fP is used to find out information about a
256 ph10 856 compiled pattern. The function \fBpcre_version()\fP returns a pointer to a
257     string containing the version of PCRE and its date of release.
258 nigel 75 .P
259 nigel 77 The function \fBpcre_refcount()\fP maintains a reference count in a data block
260     containing a compiled pattern. This is provided for the benefit of
261     object-oriented applications.
262     .P
263 nigel 75 The global variables \fBpcre_malloc\fP and \fBpcre_free\fP initially contain
264     the entry points of the standard \fBmalloc()\fP and \fBfree()\fP functions,
265 nigel 63 respectively. PCRE calls the memory management functions via these variables,
266     so a calling program can replace them if it wishes to intercept the calls. This
267     should be done before calling any PCRE functions.
268 nigel 75 .P
269     The global variables \fBpcre_stack_malloc\fP and \fBpcre_stack_free\fP are also
270 nigel 73 indirections to memory management functions. These special functions are used
271     only when PCRE is compiled to use the heap for remembering data, instead of
272 nigel 91 recursive function calls, when running the \fBpcre_exec()\fP function. See the
273     .\" HREF
274     \fBpcrebuild\fP
275     .\"
276     documentation for details of how to do this. It is a non-standard way of
277     building PCRE, for use in environments that have limited stacks. Because of the
278     greater use of memory management, it runs more slowly. Separate functions are
279     provided so that special-purpose external code can be used for this case. When
280     used, these functions are always called in a stack-like manner (last obtained,
281     first freed), and always for memory blocks of the same size. There is a
282     discussion about PCRE's stack usage in the
283     .\" HREF
284     \fBpcrestack\fP
285     .\"
286     documentation.
287 nigel 75 .P
288     The global variable \fBpcre_callout\fP initially contains NULL. It can be set
289 nigel 63 by the caller to a "callout" function, which PCRE will then call at specified
290 nigel 75 points during a matching operation. Details are given in the
291     .\" HREF
292     \fBpcrecallout\fP
293     .\"
294 nigel 63 documentation.
295 nigel 75 .
296     .
297 ph10 227 .\" HTML <a name="newlines"></a>
298 nigel 91 .SH NEWLINES
299 nigel 93 .rs
300 nigel 91 .sp
301 ph10 149 PCRE supports five different conventions for indicating line breaks in
302 nigel 93 strings: a single CR (carriage return) character, a single LF (linefeed)
303 ph10 149 character, the two-character sequence CRLF, any of the three preceding, or any
304     Unicode newline sequence. The Unicode newline sequences are the three just
305 ph10 968 mentioned, plus the single characters VT (vertical tab, U+000B), FF (form feed,
306 ph10 149 U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
307     (paragraph separator, U+2029).
308 nigel 93 .P
309     Each of the first three conventions is used by at least one operating system as
310     its standard newline sequence. When PCRE is built, a default can be specified.
311     The default default is LF, which is the Unix standard. When PCRE is run, the
312     default can be overridden, either when a pattern is compiled, or when it is
313     matched.
314     .P
315 ph10 227 At compile time, the newline convention can be specified by the \fIoptions\fP
316     argument of \fBpcre_compile()\fP, or it can be specified by special text at the
317     start of the pattern itself; this overrides any other settings. See the
318     .\" HREF
319     \fBpcrepattern\fP
320     .\"
321     page for details of the special character sequences.
322     .P
323 nigel 91 In the PCRE documentation the word "newline" is used to mean "the character or
324 nigel 93 pair of characters that indicate a line break". The choice of newline
325     convention affects the handling of the dot, circumflex, and dollar
326     metacharacters, the handling of #-comments in /x mode, and, when CRLF is a
327     recognized line ending sequence, the match position advancement for a
328 ph10 226 non-anchored pattern. There is more detail about this in the
329     .\" HTML <a href="#execoptions">
330     .\" </a>
331     section on \fBpcre_exec()\fP options
332     .\"
333 ph10 231 below.
334     .P
335     The choice of newline convention does not affect the interpretation of
336     the \en or \er escape sequences, nor does it affect what \eR matches, which is
337     controlled in a similar way, but by separate options.
338 nigel 91 .
339     .
340 nigel 63 .SH MULTITHREADING
341     .rs
342     .sp
343     The PCRE functions can be used in multi-threading applications, with the
344 nigel 75 proviso that the memory management functions pointed to by \fBpcre_malloc\fP,
345     \fBpcre_free\fP, \fBpcre_stack_malloc\fP, and \fBpcre_stack_free\fP, and the
346     callout function pointed to by \fBpcre_callout\fP, are shared by all threads.
347     .P
348 nigel 63 The compiled form of a regular expression is not altered during matching, so
349     the same compiled pattern can safely be used by several threads at once.
350 ph10 678 .P
351 ph10 691 If the just-in-time optimization feature is being used, it needs separate
352 ph10 678 memory stack areas for each thread. See the
353     .\" HREF
354     \fBpcrejit\fP
355     .\"
356     documentation for more details.
357 nigel 75 .
358     .
359     .SH "SAVING PRECOMPILED PATTERNS FOR LATER USE"
360 nigel 63 .rs
361     .sp
362 nigel 75 The compiled form of a regular expression can be saved and re-used at a later
363     time, possibly by a different program, and even on a host other than the one on
364     which it was compiled. Details are given in the
365     .\" HREF
366     \fBpcreprecompile\fP
367     .\"
368 ph10 856 documentation, which includes a description of the
369     \fBpcre_pattern_to_host_byte_order()\fP function. However, compiling a regular
370     expression with one version of PCRE for use with a different version is not
371     guaranteed to work and may cause crashes.
372 nigel 75 .
373     .
374     .SH "CHECKING BUILD-TIME OPTIONS"
375     .rs
376     .sp
377     .B int pcre_config(int \fIwhat\fP, void *\fIwhere\fP);
378 nigel 63 .PP
379 nigel 75 The function \fBpcre_config()\fP makes it possible for a PCRE client to
380 nigel 63 discover which optional features have been compiled into the PCRE library. The
381     .\" HREF
382 nigel 75 \fBpcrebuild\fP
383 nigel 63 .\"
384     documentation has more details about these optional features.
385 nigel 75 .P
386     The first argument for \fBpcre_config()\fP is an integer, specifying which
387 nigel 63 information is required; the second argument is a pointer to a variable into
388 ph10 856 which the information is placed. The returned value is zero on success, or the
389     negative error code PCRE_ERROR_BADOPTION if the value in the first argument is
390     not recognized. The following information is available:
391 nigel 75 .sp
392 nigel 63 PCRE_CONFIG_UTF8
393 nigel 75 .sp
394 nigel 63 The output is an integer that is set to one if UTF-8 support is available;
395 ph10 903 otherwise it is set to zero. If this option is given to the 16-bit version of
396 ph10 856 this function, \fBpcre16_config()\fP, the result is PCRE_ERROR_BADOPTION.
397 nigel 75 .sp
398 ph10 856 PCRE_CONFIG_UTF16
399     .sp
400     The output is an integer that is set to one if UTF-16 support is available;
401     otherwise it is set to zero. This value should normally be given to the 16-bit
402     version of this function, \fBpcre16_config()\fP. If it is given to the 8-bit
403     version of this function, the result is PCRE_ERROR_BADOPTION.
404     .sp
405 nigel 75 PCRE_CONFIG_UNICODE_PROPERTIES
406     .sp
407     The output is an integer that is set to one if support for Unicode character
408     properties is available; otherwise it is set to zero.
409     .sp
410 ph10 678 PCRE_CONFIG_JIT
411     .sp
412 ph10 691 The output is an integer that is set to one if support for just-in-time
413     compiling is available; otherwise it is set to zero.
414 ph10 678 .sp
415 ph10 887 PCRE_CONFIG_JITTARGET
416     .sp
417     The output is a pointer to a zero-terminated "const char *" string. If JIT
418 ph10 903 support is available, the string contains the name of the architecture for
419     which the JIT compiler is configured, for example "x86 32bit (little endian +
420 ph10 887 unaligned)". If JIT support is not available, the result is NULL.
421     .sp
422 nigel 63 PCRE_CONFIG_NEWLINE
423 nigel 75 .sp
424 nigel 91 The output is an integer whose value specifies the default character sequence
425 nigel 93 that is recognized as meaning "newline". The four values that are supported
426 ph10 392 are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for ANYCRLF, and -1 for ANY.
427 ph10 391 Though they are derived from ASCII, the same values are returned in EBCDIC
428     environments. The default should normally correspond to the standard sequence
429     for your operating system.
430 nigel 75 .sp
431 ph10 231 PCRE_CONFIG_BSR
432     .sp
433     The output is an integer whose value indicates what character sequences the \eR
434     escape sequence matches by default. A value of 0 means that \eR matches any
435     Unicode line ending sequence; a value of 1 means that \eR matches only CR, LF,
436     or CRLF. The default can be overridden when a pattern is compiled or matched.
437     .sp
438 nigel 63 PCRE_CONFIG_LINK_SIZE
439 nigel 75 .sp
440 nigel 63 The output is an integer that contains the number of bytes used for internal
441 ph10 856 linkage in compiled regular expressions. For the 8-bit library, the value can
442     be 2, 3, or 4. For the 16-bit library, the value is either 2 or 4 and is still
443     a number of bytes. The default value of 2 is sufficient for all but the most
444     massive patterns, since it allows the compiled pattern to be up to 64K in size.
445     Larger values allow larger regular expressions to be compiled, at the expense
446     of slower matching.
447 nigel 75 .sp
448 nigel 63 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
449 nigel 75 .sp
450 nigel 63 The output is an integer that contains the threshold above which the POSIX
451 nigel 75 interface uses \fBmalloc()\fP for output vectors. Further details are given in
452     the
453     .\" HREF
454     \fBpcreposix\fP
455     .\"
456     documentation.
457     .sp
458 nigel 63 PCRE_CONFIG_MATCH_LIMIT
459 nigel 75 .sp
460 ph10 376 The output is a long integer that gives the default limit for the number of
461 nigel 75 internal matching function calls in a \fBpcre_exec()\fP execution. Further
462     details are given with \fBpcre_exec()\fP below.
463     .sp
464 nigel 87 PCRE_CONFIG_MATCH_LIMIT_RECURSION
465     .sp
466 ph10 376 The output is a long integer that gives the default limit for the depth of
467 nigel 87 recursion when calling the internal matching function in a \fBpcre_exec()\fP
468     execution. Further details are given with \fBpcre_exec()\fP below.
469     .sp
470 nigel 73 PCRE_CONFIG_STACKRECURSE
471 nigel 75 .sp
472 nigel 77 The output is an integer that is set to one if internal recursion when running
473     \fBpcre_exec()\fP is implemented by recursive function calls that use the stack
474     to remember their state. This is the usual way that PCRE is compiled. The
475     output is zero if PCRE was compiled to use blocks of data on the heap instead
476     of recursive function calls. In this case, \fBpcre_stack_malloc\fP and
477     \fBpcre_stack_free\fP are called to manage memory blocks on the heap, thus
478     avoiding the use of the stack.
479 nigel 75 .
480     .
481     .SH "COMPILING A PATTERN"
482 nigel 63 .rs
483     .sp
484 nigel 75 .B pcre *pcre_compile(const char *\fIpattern\fP, int \fIoptions\fP,
485 nigel 63 .ti +5n
486 nigel 75 .B const char **\fIerrptr\fP, int *\fIerroffset\fP,
487 nigel 63 .ti +5n
488 nigel 75 .B const unsigned char *\fItableptr\fP);
489 nigel 77 .sp
490     .B pcre *pcre_compile2(const char *\fIpattern\fP, int \fIoptions\fP,
491     .ti +5n
492     .B int *\fIerrorcodeptr\fP,
493     .ti +5n
494     .B const char **\fIerrptr\fP, int *\fIerroffset\fP,
495     .ti +5n
496     .B const unsigned char *\fItableptr\fP);
497 nigel 75 .P
498 nigel 77 Either of the functions \fBpcre_compile()\fP or \fBpcre_compile2()\fP can be
499     called to compile a pattern into an internal form. The only difference between
500     the two interfaces is that \fBpcre_compile2()\fP has an additional argument,
501 ph10 461 \fIerrorcodeptr\fP, via which a numerical error code can be returned. To avoid
502     too much repetition, we refer just to \fBpcre_compile()\fP below, but the
503 ph10 456 information applies equally to \fBpcre_compile2()\fP.
504 nigel 75 .P
505 nigel 77 The pattern is a C string terminated by a binary zero, and is passed in the
506     \fIpattern\fP argument. A pointer to a single block of memory that is obtained
507     via \fBpcre_malloc\fP is returned. This contains the compiled code and related
508     data. The \fBpcre\fP type is defined for the returned block; this is a typedef
509     for a structure whose contents are not externally defined. It is up to the
510 nigel 91 caller to free the memory (via \fBpcre_free\fP) when it is no longer required.
511 nigel 77 .P
512 nigel 63 Although the compiled code of a PCRE regex is relocatable, that is, it does not
513 nigel 75 depend on memory location, the complete \fBpcre\fP data block is not
514     fully relocatable, because it may contain a copy of the \fItableptr\fP
515     argument, which is an address (see below).
516     .P
517 nigel 93 The \fIoptions\fP argument contains various bit settings that affect the
518 nigel 75 compilation. It should be zero if no options are required. The available
519 ph10 412 options are described below. Some of them (in particular, those that are
520 ph10 456 compatible with Perl, but some others as well) can also be set and unset from
521 ph10 412 within the pattern (see the detailed description in the
522 nigel 75 .\" HREF
523     \fBpcrepattern\fP
524     .\"
525 ph10 412 documentation). For those options that can be different in different parts of
526 ph10 456 the pattern, the contents of the \fIoptions\fP argument specifies their
527 ph10 461 settings at the start of compilation and execution. The PCRE_ANCHORED,
528 ph10 576 PCRE_BSR_\fIxxx\fP, PCRE_NEWLINE_\fIxxx\fP, PCRE_NO_UTF8_CHECK, and
529 ph10 930 PCRE_NO_START_OPTIMIZE options can be set at the time of matching as well as at
530 ph10 576 compile time.
531 nigel 75 .P
532     If \fIerrptr\fP is NULL, \fBpcre_compile()\fP returns NULL immediately.
533     Otherwise, if compilation of a pattern fails, \fBpcre_compile()\fP returns
534     NULL, and sets the variable pointed to by \fIerrptr\fP to point to a textual
535 nigel 87 error message. This is a static string that is part of the library. You must
536 ph10 598 not try to free it. Normally, the offset from the start of the pattern to the
537     byte that was being processed when the error was discovered is placed in the
538     variable pointed to by \fIerroffset\fP, which must not be NULL (if it is, an
539     immediate error is given). However, for an invalid UTF-8 string, the offset is
540 ph10 856 that of the first byte of the failing character.
541 nigel 75 .P
542 ph10 856 Some errors are not detected until the whole pattern has been scanned; in these
543     cases, the offset passed back is the length of the pattern. Note that the
544     offset is in bytes, not characters, even in UTF-8 mode. It may sometimes point
545     into the middle of a UTF-8 character.
546 ph10 569 .P
547 nigel 77 If \fBpcre_compile2()\fP is used instead of \fBpcre_compile()\fP, and the
548     \fIerrorcodeptr\fP argument is not NULL, a non-zero error code number is
549     returned via this argument in the event of an error. This is in addition to the
550     textual error message. Error codes and messages are listed below.
551     .P
552 nigel 75 If the final argument, \fItableptr\fP, is NULL, PCRE uses a default set of
553     character tables that are built when PCRE is compiled, using the default C
554     locale. Otherwise, \fItableptr\fP must be an address that is the result of a
555     call to \fBpcre_maketables()\fP. This value is stored with the compiled
556     pattern, and used again by \fBpcre_exec()\fP, unless another table pointer is
557     passed to it. For more discussion, see the section on locale support below.
558     .P
559     This code fragment shows a typical straightforward call to \fBpcre_compile()\fP:
560     .sp
561 nigel 63 pcre *re;
562     const char *error;
563     int erroffset;
564     re = pcre_compile(
565     "^A.*Z", /* the pattern */
566     0, /* default options */
567     &error, /* for error message */
568     &erroffset, /* for error offset */
569     NULL); /* use default character tables */
570 nigel 75 .sp
571     The following names for option bits are defined in the \fBpcre.h\fP header
572     file:
573     .sp
574 nigel 63 PCRE_ANCHORED
575 nigel 75 .sp
576 nigel 63 If this bit is set, the pattern is forced to be "anchored", that is, it is
577 nigel 75 constrained to match only at the first matching point in the string that is
578 nigel 63 being searched (the "subject string"). This effect can also be achieved by
579     appropriate constructs in the pattern itself, which is the only way to do it in
580     Perl.
581 nigel 75 .sp
582     PCRE_AUTO_CALLOUT
583     .sp
584     If this bit is set, \fBpcre_compile()\fP automatically inserts callout items,
585     all with number 255, before each pattern item. For discussion of the callout
586     facility, see the
587     .\" HREF
588     \fBpcrecallout\fP
589     .\"
590     documentation.
591     .sp
592 ph10 231 PCRE_BSR_ANYCRLF
593     PCRE_BSR_UNICODE
594     .sp
595     These options (which are mutually exclusive) control what the \eR escape
596     sequence matches. The choice is either to match only CR, LF, or CRLF, or to
597     match any Unicode newline sequence. The default is specified when PCRE is
598     built. It can be overridden from within the pattern, or by setting an option
599     when a compiled pattern is matched.
600     .sp
601 nigel 63 PCRE_CASELESS
602 nigel 75 .sp
603 nigel 63 If this bit is set, letters in the pattern match both upper and lower case
604     letters. It is equivalent to Perl's /i option, and it can be changed within a
605 nigel 77 pattern by a (?i) option setting. In UTF-8 mode, PCRE always understands the
606     concept of case for characters whose values are less than 128, so caseless
607     matching is always possible. For characters with higher values, the concept of
608     case is supported if PCRE is compiled with Unicode property support, but not
609     otherwise. If you want to use caseless matching for characters 128 and above,
610     you must ensure that PCRE is compiled with Unicode property support as well as
611     with UTF-8 support.
612 nigel 75 .sp
613 nigel 63 PCRE_DOLLAR_ENDONLY
614 nigel 75 .sp
615 nigel 63 If this bit is set, a dollar metacharacter in the pattern matches only at the
616     end of the subject string. Without this option, a dollar also matches
617 nigel 91 immediately before a newline at the end of the string (but not before any other
618     newlines). The PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
619     There is no equivalent to this option in Perl, and no way to set it within a
620     pattern.
621 nigel 75 .sp
622 nigel 63 PCRE_DOTALL
623 nigel 75 .sp
624 ph10 572 If this bit is set, a dot metacharacter in the pattern matches a character of
625     any value, including one that indicates a newline. However, it only ever
626     matches one character, even if newlines are coded as CRLF. Without this option,
627     a dot does not match when the current position is at a newline. This option is
628     equivalent to Perl's /s option, and it can be changed within a pattern by a
629     (?s) option setting. A negative class such as [^a] always matches newline
630     characters, independent of the setting of this option.
631 nigel 75 .sp
632 nigel 91 PCRE_DUPNAMES
633     .sp
634     If this bit is set, names used to identify capturing subpatterns need not be
635     unique. This can be helpful for certain types of pattern when it is known that
636     only one instance of the named subpattern can ever be matched. There are more
637     details of named subpatterns below; see also the
638     .\" HREF
639     \fBpcrepattern\fP
640     .\"
641     documentation.
642     .sp
643 nigel 63 PCRE_EXTENDED
644 nigel 75 .sp
645 ph10 968 If this bit is set, white space data characters in the pattern are totally
646     ignored except when escaped or inside a character class. White space does not
647 nigel 63 include the VT character (code 11). In addition, characters between an
648 nigel 91 unescaped # outside a character class and the next newline, inclusive, are also
649     ignored. This is equivalent to Perl's /x option, and it can be changed within a
650     pattern by a (?x) option setting.
651 nigel 75 .P
652 ph10 598 Which characters are interpreted as newlines is controlled by the options
653     passed to \fBpcre_compile()\fP or by a special sequence at the start of the
654     pattern, as described in the section entitled
655 ph10 572 .\" HTML <a href="pcrepattern.html#newlines">
656     .\" </a>
657     "Newline conventions"
658     .\"
659     in the \fBpcrepattern\fP documentation. Note that the end of this type of
660     comment is a literal newline sequence in the pattern; escape sequences that
661     happen to represent a newline do not count.
662     .P
663 nigel 63 This option makes it possible to include comments inside complicated patterns.
664 ph10 968 Note, however, that this applies only to data characters. White space characters
665 nigel 63 may never appear within special character sequences in a pattern, for example
666 ph10 572 within the sequence (?( that introduces a conditional subpattern.
667 nigel 75 .sp
668 nigel 63 PCRE_EXTRA
669 nigel 75 .sp
670 nigel 63 This option was invented in order to turn on additional functionality of PCRE
671     that is incompatible with Perl, but it is currently of very little use. When
672     set, any backslash in a pattern that is followed by a letter that has no
673     special meaning causes an error, thus reserving these combinations for future
674     expansion. By default, as in Perl, a backslash followed by a letter with no
675 nigel 91 special meaning is treated as a literal. (Perl can, however, be persuaded to
676 ph10 513 give an error for this, by running it with the -w option.) There are at present
677     no other features controlled by this option. It can also be set by a (?X)
678     option setting within a pattern.
679 nigel 75 .sp
680 nigel 77 PCRE_FIRSTLINE
681     .sp
682     If this option is set, an unanchored pattern is required to match before or at
683 nigel 91 the first newline in the subject string, though the matched text may continue
684     over the newline.
685 nigel 77 .sp
686 ph10 336 PCRE_JAVASCRIPT_COMPAT
687     .sp
688 ph10 345 If this option is set, PCRE's behaviour is changed in some ways so that it is
689 ph10 336 compatible with JavaScript rather than Perl. The changes are as follows:
690     .P
691     (1) A lone closing square bracket in a pattern causes a compile-time error,
692     because this is illegal in JavaScript (by default it is treated as a data
693     character). Thus, the pattern AB]CD becomes illegal when this option is set.
694     .P
695     (2) At run time, a back reference to an unset subpattern group matches an empty
696 ph10 345 string (by default this causes the current matching alternative to fail). A
697     pattern such as (\e1)(a) succeeds when this option is set (assuming it can find
698     an "a" in the subject), whereas it fails by default, for Perl compatibility.
699 ph10 745 .P
700 ph10 836 (3) \eU matches an upper case "U" character; by default \eU causes a compile
701 ph10 745 time error (Perl uses \eU to upper case subsequent characters).
702     .P
703 ph10 836 (4) \eu matches a lower case "u" character unless it is followed by four
704     hexadecimal digits, in which case the hexadecimal number defines the code point
705     to match. By default, \eu causes a compile time error (Perl uses it to upper
706 ph10 745 case the following character).
707     .P
708 ph10 836 (5) \ex matches a lower case "x" character unless it is followed by two
709     hexadecimal digits, in which case the hexadecimal number defines the code point
710     to match. By default, as in Perl, a hexadecimal number is always expected after
711     \ex, but it may have zero, one, or two digits (so, for example, \exz matches a
712 ph10 745 binary zero character followed by z).
713 ph10 336 .sp
714 nigel 63 PCRE_MULTILINE
715 nigel 75 .sp
716     By default, PCRE treats the subject string as consisting of a single line of
717     characters (even if it actually contains newlines). The "start of line"
718 nigel 63 metacharacter (^) matches only at the start of the string, while the "end of
719     line" metacharacter ($) matches only at the end of the string, or before a
720     terminating newline (unless PCRE_DOLLAR_ENDONLY is set). This is the same as
721     Perl.
722 nigel 75 .P
723 nigel 63 When PCRE_MULTILINE it is set, the "start of line" and "end of line" constructs
724 nigel 91 match immediately following or immediately before internal newlines in the
725     subject string, respectively, as well as at the very start and end. This is
726     equivalent to Perl's /m option, and it can be changed within a pattern by a
727     (?m) option setting. If there are no newlines in a subject string, or no
728 nigel 63 occurrences of ^ or $ in a pattern, setting PCRE_MULTILINE has no effect.
729 nigel 75 .sp
730 nigel 91 PCRE_NEWLINE_CR
731     PCRE_NEWLINE_LF
732     PCRE_NEWLINE_CRLF
733 ph10 150 PCRE_NEWLINE_ANYCRLF
734 nigel 93 PCRE_NEWLINE_ANY
735 nigel 91 .sp
736     These options override the default newline definition that was chosen when PCRE
737     was built. Setting the first or the second specifies that a newline is
738 nigel 93 indicated by a single character (CR or LF, respectively). Setting
739     PCRE_NEWLINE_CRLF specifies that a newline is indicated by the two-character
740 ph10 149 CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies that any of the three
741     preceding sequences should be recognized. Setting PCRE_NEWLINE_ANY specifies
742     that any Unicode newline sequence should be recognized. The Unicode newline
743     sequences are the three just mentioned, plus the single characters VT (vertical
744 ph10 968 tab, U+000B), FF (form feed, U+000C), NEL (next line, U+0085), LS (line
745 ph10 903 separator, U+2028), and PS (paragraph separator, U+2029). For the 8-bit
746 ph10 856 library, the last two are recognized only in UTF-8 mode.
747 nigel 91 .P
748 nigel 93 The newline setting in the options word uses three bits that are treated
749 ph10 149 as a number, giving eight possibilities. Currently only six are used (default
750     plus the five values above). This means that if you set more than one newline
751 nigel 93 option, the combination may or may not be sensible. For example,
752     PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to PCRE_NEWLINE_CRLF, but
753 ph10 149 other combinations may yield unused numbers and cause an error.
754 nigel 93 .P
755 ph10 572 The only time that a line break in a pattern is specially recognized when
756 ph10 968 compiling is when PCRE_EXTENDED is set. CR and LF are white space characters,
757 ph10 572 and so are ignored in this mode. Also, an unescaped # outside a character class
758     indicates a comment that lasts until after the next line break sequence. In
759     other circumstances, line break sequences in patterns are treated as literal
760     data.
761 nigel 93 .P
762     The newline option that is set at compile time becomes the default that is used
763     for \fBpcre_exec()\fP and \fBpcre_dfa_exec()\fP, but it can be overridden.
764 nigel 91 .sp
765 nigel 63 PCRE_NO_AUTO_CAPTURE
766 nigel 75 .sp
767 nigel 63 If this option is set, it disables the use of numbered capturing parentheses in
768     the pattern. Any opening parenthesis that is not followed by ? behaves as if it
769     were followed by ?: but named parentheses can still be used for capturing (and
770     they acquire numbers in the usual way). There is no equivalent of this option
771     in Perl.
772 nigel 75 .sp
773 ph10 576 NO_START_OPTIMIZE
774     .sp
775     This is an option that acts at matching time; that is, it is really an option
776     for \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP. If it is set at compile time,
777     it is remembered with the compiled pattern and assumed at matching time. For
778     details see the discussion of PCRE_NO_START_OPTIMIZE
779     .\" HTML <a href="#execoptions">
780     .\" </a>
781     below.
782     .\"
783     .sp
784 ph10 518 PCRE_UCP
785     .sp
786 ph10 572 This option changes the way PCRE processes \eB, \eb, \eD, \ed, \eS, \es, \eW,
787     \ew, and some of the POSIX character classes. By default, only ASCII characters
788     are recognized, but if PCRE_UCP is set, Unicode properties are used instead to
789     classify characters. More details are given in the section on
790 ph10 518 .\" HTML <a href="pcre.html#genericchartypes">
791     .\" </a>
792     generic character types
793     .\"
794     in the
795     .\" HREF
796     \fBpcrepattern\fP
797     .\"
798     page. If you set PCRE_UCP, matching one of the items it affects takes much
799     longer. The option is available only if PCRE has been compiled with Unicode
800     property support.
801     .sp
802 nigel 63 PCRE_UNGREEDY
803 nigel 75 .sp
804 nigel 63 This option inverts the "greediness" of the quantifiers so that they are not
805     greedy by default, but become greedy if followed by "?". It is not compatible
806     with Perl. It can also be set by a (?U) option setting within the pattern.
807 nigel 75 .sp
808 nigel 63 PCRE_UTF8
809 nigel 75 .sp
810 nigel 63 This option causes PCRE to regard both the pattern and the subject as strings
811 ph10 856 of UTF-8 characters instead of single-byte strings. However, it is available
812     only when PCRE is built to include UTF support. If not, the use of this option
813     provokes an error. Details of how this option changes the behaviour of PCRE are
814     given in the
815 nigel 63 .\" HREF
816 ph10 678 \fBpcreunicode\fP
817 nigel 63 .\"
818     page.
819 nigel 75 .sp
820 nigel 71 PCRE_NO_UTF8_CHECK
821 nigel 75 .sp
822 ph10 903 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8
823 ph10 856 string is automatically checked. There is a discussion about the
824     .\" HTML <a href="pcreunicode.html#utf8strings">
825 ph10 211 .\" </a>
826 ph10 903 validity of UTF-8 strings
827 ph10 211 .\"
828 ph10 856 in the
829 ph10 211 .\" HREF
830 ph10 856 \fBpcreunicode\fP
831 ph10 211 .\"
832 ph10 856 page. If an invalid UTF-8 sequence is found, \fBpcre_compile()\fP returns an
833     error. If you already know that your pattern is valid, and you want to skip
834     this check for performance reasons, you can set the PCRE_NO_UTF8_CHECK option.
835     When it is set, the effect of passing an invalid UTF-8 string as a pattern is
836     undefined. It may cause your program to crash. Note that this option can also
837     be passed to \fBpcre_exec()\fP and \fBpcre_dfa_exec()\fP, to suppress the
838     validity checking of subject strings.
839 nigel 75 .
840     .
841 nigel 77 .SH "COMPILATION ERROR CODES"
842     .rs
843     .sp
844     The following table lists the error codes than may be returned by
845     \fBpcre_compile2()\fP, along with the error messages that may be returned by
846 ph10 903 both compiling functions. Note that error messages are always 8-bit ASCII
847 ph10 856 strings, even in 16-bit mode. As PCRE has developed, some error codes have
848     fallen out of use. To avoid confusion, they have not been re-used.
849 nigel 77 .sp
850     0 no error
851     1 \e at end of pattern
852     2 \ec at end of pattern
853     3 unrecognized character follows \e
854     4 numbers out of order in {} quantifier
855     5 number too big in {} quantifier
856     6 missing terminating ] for character class
857     7 invalid escape sequence in character class
858     8 range out of order in character class
859     9 nothing to repeat
860 nigel 93 10 [this code is not in use]
861 nigel 77 11 internal error: unexpected repeat
862 ph10 290 12 unrecognized character after (? or (?-
863 nigel 77 13 POSIX named classes are supported only within a class
864     14 missing )
865     15 reference to non-existent subpattern
866     16 erroffset passed as NULL
867     17 unknown option bit(s) set
868     18 missing ) after comment
869 nigel 93 19 [this code is not in use]
870 ph10 290 20 regular expression is too large
871 nigel 77 21 failed to get memory
872     22 unmatched parentheses
873     23 internal error: code overflow
874     24 unrecognized character after (?<
875     25 lookbehind assertion is not fixed length
876 nigel 91 26 malformed number or name after (?(
877 nigel 77 27 conditional group contains more than two branches
878     28 assertion expected after (?(
879 ph10 181 29 (?R or (?[+-]digits must be followed by )
880 nigel 77 30 unknown POSIX class name
881     31 POSIX collating elements are not supported
882 ph10 856 32 this version of PCRE is compiled without UTF support
883 nigel 93 33 [this code is not in use]
884 nigel 77 34 character value in \ex{...} sequence is too large
885     35 invalid condition (?(0)
886     36 \eC not allowed in lookbehind assertion
887 ph10 656 37 PCRE does not support \eL, \el, \eN{name}, \eU, or \eu
888 nigel 77 38 number after (?C is > 255
889     39 closing ) for (?C expected
890     40 recursive call could loop indefinitely
891     41 unrecognized character after (?P
892 nigel 93 42 syntax error in subpattern name (missing terminator)
893 nigel 91 43 two named subpatterns have the same name
894 ph10 856 44 invalid UTF-8 string (specifically UTF-8)
895 nigel 77 45 support for \eP, \ep, and \eX has not been compiled
896     46 malformed \eP or \ep sequence
897     47 unknown property name after \eP or \ep
898 nigel 91 48 subpattern name is too long (maximum 32 characters)
899 ph10 290 49 too many named subpatterns (maximum 10000)
900 ph10 202 50 [this code is not in use]
901 ph10 856 51 octal value is greater than \e377 in 8-bit non-UTF-8 mode
902 nigel 93 52 internal error: overran compiling workspace
903 ph10 548 53 internal error: previously-checked referenced subpattern
904 ph10 546 not found
905 nigel 93 54 DEFINE group contains more than one branch
906     55 repeating a DEFINE group is not allowed
907 ph10 231 56 inconsistent NEWLINE options
908 ph10 345 57 \eg is not followed by a braced, angle-bracketed, or quoted
909     name/number or by a plain number
910 ph10 336 58 a numbered reference must not be zero
911 ph10 510 59 an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)
912 ph10 290 60 (*VERB) not recognized
913     61 number is too big
914     62 subpattern name expected
915 ph10 292 63 digit expected after (?+
916 ph10 336 64 ] is an invalid data character in JavaScript compatibility mode
917 ph10 548 65 different names for subpatterns of the same number are
918 ph10 546 not allowed
919 ph10 510 66 (*MARK) must have an argument
920 ph10 903 67 this version of PCRE is not compiled with Unicode property
921 ph10 856 support
922 ph10 656 68 \ec must be followed by an ASCII character
923 ph10 659 69 \ek is not followed by a braced, angle-bracketed, or quoted name
924 ph10 856 70 internal error: unknown opcode in find_fixedlength()
925     71 \eN is not supported in a class
926     72 too many forward references
927 ph10 903 73 disallowed Unicode code point (>= 0xd800 && <= 0xdfff)
928 ph10 856 74 invalid UTF-16 string (specifically UTF-16)
929 ph10 964 75 name is too long in (*MARK), (*PRUNE), (*SKIP), or (*THEN)
930 ph10 290 .sp
931 ph10 292 The numbers 32 and 10000 in errors 48 and 49 are defaults; different values may
932 ph10 290 be used if the limits were changed when PCRE was built.
933 nigel 77 .
934     .
935 ph10 656 .\" HTML <a name="studyingapattern"></a>
936 nigel 75 .SH "STUDYING A PATTERN"
937 nigel 63 .rs
938     .sp
939 nigel 77 .B pcre_extra *pcre_study(const pcre *\fIcode\fP, int \fIoptions\fP
940 nigel 63 .ti +5n
941 nigel 75 .B const char **\fIerrptr\fP);
942 nigel 63 .PP
943 nigel 75 If a compiled pattern is going to be used several times, it is worth spending
944     more time analyzing it in order to speed up the time taken for matching. The
945     function \fBpcre_study()\fP takes a pointer to a compiled pattern as its first
946     argument. If studying the pattern produces additional information that will
947     help speed up matching, \fBpcre_study()\fP returns a pointer to a
948     \fBpcre_extra\fP block, in which the \fIstudy_data\fP field points to the
949     results of the study.
950     .P
951     The returned value from \fBpcre_study()\fP can be passed directly to
952 ph10 455 \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP. However, a \fBpcre_extra\fP block
953     also contains other fields that can be set by the caller before the block is
954     passed; these are described
955 nigel 75 .\" HTML <a href="#extradata">
956     .\" </a>
957     below
958     .\"
959     in the section on matching a pattern.
960     .P
961 ph10 455 If studying the pattern does not produce any useful information,
962 nigel 75 \fBpcre_study()\fP returns NULL. In that circumstance, if the calling program
963 ph10 461 wants to pass any of the other fields to \fBpcre_exec()\fP or
964 ph10 455 \fBpcre_dfa_exec()\fP, it must set up its own \fBpcre_extra\fP block.
965 nigel 75 .P
966 ph10 921 The second argument of \fBpcre_study()\fP contains option bits. There are three
967     options:
968     .sp
969     PCRE_STUDY_JIT_COMPILE
970     PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE
971     PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE
972     .sp
973     If any of these are set, and the just-in-time compiler is available, the
974     pattern is further compiled into machine code that executes much faster than
975     the \fBpcre_exec()\fP interpretive matching function. If the just-in-time
976     compiler is not available, these options are ignored. All other bits in the
977     \fIoptions\fP argument must be zero.
978 nigel 75 .P
979 ph10 691 JIT compilation is a heavyweight optimization. It can take some time for
980 ph10 678 patterns to be analyzed, and for one-off matches and simple patterns the
981     benefit of faster execution might be offset by a much slower study time.
982 ph10 691 Not all patterns can be optimized by the JIT compiler. For those that cannot be
983 ph10 678 handled, matching automatically falls back to the \fBpcre_exec()\fP
984     interpreter. For more details, see the
985     .\" HREF
986     \fBpcrejit\fP
987     .\"
988     documentation.
989     .P
990 nigel 75 The third argument for \fBpcre_study()\fP is a pointer for an error message. If
991 nigel 63 studying succeeds (even if no data is returned), the variable it points to is
992 nigel 87 set to NULL. Otherwise it is set to point to a textual error message. This is a
993     static string that is part of the library. You must not try to free it. You
994     should test the error pointer for NULL after calling \fBpcre_study()\fP, to be
995     sure that it has run successfully.
996 nigel 75 .P
997 ph10 678 When you are finished with a pattern, you can free the memory used for the
998     study data by calling \fBpcre_free_study()\fP. This function was added to the
999     API for release 8.20. For earlier versions, the memory could be freed with
1000     \fBpcre_free()\fP, just like the pattern itself. This will still work in cases
1001 ph10 921 where JIT optimization is not used, but it is advisable to change to the new
1002     function when convenient.
1003 ph10 678 .P
1004 ph10 691 This is a typical way in which \fBpcre_study\fP() is used (except that in a
1005 ph10 678 real application there should be tests for errors):
1006 nigel 75 .sp
1007 ph10 678 int rc;
1008     pcre *re;
1009     pcre_extra *sd;
1010     re = pcre_compile("pattern", 0, &error, &erroroffset, NULL);
1011     sd = pcre_study(
1012 nigel 63 re, /* result of pcre_compile() */
1013 ph10 678 0, /* no options */
1014 nigel 63 &error); /* set to NULL or points to a message */
1015 ph10 678 rc = pcre_exec( /* see below for details of pcre_exec() options */
1016 ph10 691 re, sd, "subject", 7, 0, 0, ovector, 30);
1017 ph10 678 ...
1018     pcre_free_study(sd);
1019 ph10 691 pcre_free(re);
1020 nigel 75 .sp
1021 ph10 455 Studying a pattern does two things: first, a lower bound for the length of
1022 ph10 461 subject string that is needed to match the pattern is computed. This does not
1023     mean that there are any strings of that length that match, but it does
1024     guarantee that no shorter strings match. The value is used by
1025     \fBpcre_exec()\fP and \fBpcre_dfa_exec()\fP to avoid wasting time by trying to
1026     match strings that are shorter than the lower bound. You can find out the value
1027 ph10 455 in a calling program via the \fBpcre_fullinfo()\fP function.
1028     .P
1029     Studying a pattern is also useful for non-anchored patterns that do not have a
1030     single fixed starting character. A bitmap of possible starting bytes is
1031 ph10 461 created. This speeds up finding a position in the subject at which to start
1032 ph10 856 matching. (In 16-bit mode, the bitmap is used for 16-bit values less than 256.)
1033 ph10 547 .P
1034 ph10 691 These two optimizations apply to both \fBpcre_exec()\fP and
1035 ph10 921 \fBpcre_dfa_exec()\fP, and the information is also used by the JIT compiler.
1036     The optimizations can be disabled by setting the PCRE_NO_START_OPTIMIZE option
1037     when calling \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP, but if this is done,
1038     JIT execution is also disabled. You might want to do this if your pattern
1039     contains callouts or (*MARK) and you want to make use of these facilities in
1040     cases where matching fails. See the discussion of PCRE_NO_START_OPTIMIZE
1041 ph10 547 .\" HTML <a href="#execoptions">
1042     .\" </a>
1043     below.
1044     .\"
1045 nigel 75 .
1046     .
1047 nigel 63 .\" HTML <a name="localesupport"></a>
1048 nigel 75 .SH "LOCALE SUPPORT"
1049 nigel 63 .rs
1050     .sp
1051 ph10 139 PCRE handles caseless matching, and determines whether characters are letters,
1052 nigel 75 digits, or whatever, by reference to a set of tables, indexed by character
1053 ph10 856 value. When running in UTF-8 mode, this applies only to characters
1054     with codes less than 128. By default, higher-valued codes never match escapes
1055     such as \ew or \ed, but they can be tested with \ep if PCRE is built with
1056     Unicode character property support. Alternatively, the PCRE_UCP option can be
1057     set at compile time; this causes \ew and friends to use Unicode property
1058     support instead of built-in tables. The use of locales with Unicode is
1059     discouraged. If you are handling characters with codes greater than 128, you
1060     should either use UTF-8 and Unicode, or use locales, but not try to mix the
1061     two.
1062 nigel 75 .P
1063 ph10 139 PCRE contains an internal set of tables that are used when the final argument
1064     of \fBpcre_compile()\fP is NULL. These are sufficient for many applications.
1065 ph10 142 Normally, the internal tables recognize only ASCII characters. However, when
1066 ph10 139 PCRE is built, it is possible to cause the internal tables to be rebuilt in the
1067     default "C" locale of the local system, which may cause them to be different.
1068 nigel 75 .P
1069 ph10 139 The internal tables can always be overridden by tables supplied by the
1070     application that calls PCRE. These may be created in a different locale from
1071     the default. As more and more applications change to using Unicode, the need
1072     for this locale support is expected to die away.
1073     .P
1074 nigel 75 External tables are built by calling the \fBpcre_maketables()\fP function,
1075     which has no arguments, in the relevant locale. The result can then be passed
1076     to \fBpcre_compile()\fP or \fBpcre_exec()\fP as often as necessary. For
1077     example, to build and use tables that are appropriate for the French locale
1078     (where accented characters with values greater than 128 are treated as letters),
1079     the following code could be used:
1080     .sp
1081     setlocale(LC_CTYPE, "fr_FR");
1082 nigel 63 tables = pcre_maketables();
1083     re = pcre_compile(..., tables);
1084 nigel 75 .sp
1085 ph10 142 The locale name "fr_FR" is used on Linux and other Unix-like systems; if you
1086 ph10 139 are using Windows, the name for the French locale is "french".
1087     .P
1088 nigel 75 When \fBpcre_maketables()\fP runs, the tables are built in memory that is
1089     obtained via \fBpcre_malloc\fP. It is the caller's responsibility to ensure
1090     that the memory containing the tables remains available for as long as it is
1091     needed.
1092     .P
1093     The pointer that is passed to \fBpcre_compile()\fP is saved with the compiled
1094     pattern, and the same tables are used via this pointer by \fBpcre_study()\fP
1095     and normally also by \fBpcre_exec()\fP. Thus, by default, for any single
1096     pattern, compilation, studying and matching all happen in the same locale, but
1097     different patterns can be compiled in different locales.
1098     .P
1099     It is possible to pass a table pointer or NULL (indicating the use of the
1100     internal tables) to \fBpcre_exec()\fP. Although not intended for this purpose,
1101     this facility could be used to match a pattern in a different locale from the
1102     one in which it was compiled. Passing table pointers at run time is discussed
1103     below in the section on matching a pattern.
1104     .
1105     .
1106 ph10 598 .\" HTML <a name="infoaboutpattern"></a>
1107 nigel 75 .SH "INFORMATION ABOUT A PATTERN"
1108 nigel 63 .rs
1109     .sp
1110 nigel 75 .B int pcre_fullinfo(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
1111 nigel 63 .ti +5n
1112 nigel 75 .B int \fIwhat\fP, void *\fIwhere\fP);
1113 nigel 63 .PP
1114 nigel 75 The \fBpcre_fullinfo()\fP function returns information about a compiled
1115 ph10 856 pattern. It replaces the \fBpcre_info()\fP function, which was removed from the
1116     library at version 8.30, after more than 10 years of obsolescence.
1117 nigel 75 .P
1118     The first argument for \fBpcre_fullinfo()\fP is a pointer to the compiled
1119     pattern. The second argument is the result of \fBpcre_study()\fP, or NULL if
1120 nigel 63 the pattern was not studied. The third argument specifies which piece of
1121     information is required, and the fourth argument is a pointer to a variable
1122     to receive the data. The yield of the function is zero for success, or one of
1123     the following negative numbers:
1124 nigel 75 .sp
1125 ph10 856 PCRE_ERROR_NULL the argument \fIcode\fP was NULL
1126     the argument \fIwhere\fP was NULL
1127     PCRE_ERROR_BADMAGIC the "magic number" was not found
1128 ph10 903 PCRE_ERROR_BADENDIANNESS the pattern was compiled with different
1129 ph10 856 endianness
1130     PCRE_ERROR_BADOPTION the value of \fIwhat\fP was invalid
1131 nigel 75 .sp
1132     The "magic number" is placed at the start of each compiled pattern as an simple
1133 ph10 903 check against passing an arbitrary memory pointer. The endianness error can
1134 ph10 856 occur if a compiled pattern is saved and reloaded on a different host. Here is
1135     a typical call of \fBpcre_fullinfo()\fP, to obtain the length of the compiled
1136     pattern:
1137 nigel 75 .sp
1138 nigel 63 int rc;
1139 nigel 91 size_t length;
1140 nigel 63 rc = pcre_fullinfo(
1141     re, /* result of pcre_compile() */
1142 ph10 678 sd, /* result of pcre_study(), or NULL */
1143 nigel 63 PCRE_INFO_SIZE, /* what is required */
1144     &length); /* where to put the data */
1145 nigel 75 .sp
1146     The possible values for the third argument are defined in \fBpcre.h\fP, and are
1147 nigel 63 as follows:
1148 nigel 75 .sp
1149 nigel 63 PCRE_INFO_BACKREFMAX
1150 nigel 75 .sp
1151 nigel 63 Return the number of the highest back reference in the pattern. The fourth
1152 nigel 75 argument should point to an \fBint\fP variable. Zero is returned if there are
1153 nigel 63 no back references.
1154 nigel 75 .sp
1155 nigel 63 PCRE_INFO_CAPTURECOUNT
1156 nigel 75 .sp
1157 nigel 63 Return the number of capturing subpatterns in the pattern. The fourth argument
1158 nigel 75 should point to an \fBint\fP variable.
1159     .sp
1160 nigel 77 PCRE_INFO_DEFAULT_TABLES
1161 nigel 75 .sp
1162     Return a pointer to the internal default character tables within PCRE. The
1163     fourth argument should point to an \fBunsigned char *\fP variable. This
1164     information call is provided for internal use by the \fBpcre_study()\fP
1165     function. External callers can cause PCRE to use its internal tables by passing
1166     a NULL table pointer.
1167     .sp
1168 nigel 63 PCRE_INFO_FIRSTBYTE
1169 nigel 75 .sp
1170 ph10 856 Return information about the first data unit of any matched string, for a
1171     non-anchored pattern. (The name of this option refers to the 8-bit library,
1172     where data units are bytes.) The fourth argument should point to an \fBint\fP
1173     variable.
1174 nigel 75 .P
1175 ph10 856 If there is a fixed first value, for example, the letter "c" from a pattern
1176 ph10 903 such as (cat|cow|coyote), its value is returned. In the 8-bit library, the
1177     value is always less than 256; in the 16-bit library the value can be up to
1178 ph10 856 0xffff.
1179     .P
1180     If there is no fixed first value, and if either
1181 nigel 75 .sp
1182 nigel 63 (a) the pattern was compiled with the PCRE_MULTILINE option, and every branch
1183     starts with "^", or
1184 nigel 75 .sp
1185 nigel 63 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not set
1186     (if it were set, the pattern would be anchored),
1187 nigel 75 .sp
1188 nigel 63 -1 is returned, indicating that the pattern matches only at the start of a
1189     subject string or after any newline within the string. Otherwise -2 is
1190     returned. For anchored patterns, -2 is returned.
1191 nigel 75 .sp
1192 nigel 63 PCRE_INFO_FIRSTTABLE
1193 nigel 75 .sp
1194 nigel 63 If the pattern was studied, and this resulted in the construction of a 256-bit
1195 ph10 856 table indicating a fixed set of values for the first data unit in any matching
1196 nigel 63 string, a pointer to the table is returned. Otherwise NULL is returned. The
1197 nigel 75 fourth argument should point to an \fBunsigned char *\fP variable.
1198     .sp
1199 ph10 226 PCRE_INFO_HASCRORLF
1200     .sp
1201 ph10 227 Return 1 if the pattern contains any explicit matches for CR or LF characters,
1202 ph10 243 otherwise 0. The fourth argument should point to an \fBint\fP variable. An
1203 ph10 231 explicit match is either a literal CR or LF character, or \er or \en.
1204 ph10 226 .sp
1205 ph10 169 PCRE_INFO_JCHANGED
1206     .sp
1207 ph10 278 Return 1 if the (?J) or (?-J) option setting is used in the pattern, otherwise
1208     0. The fourth argument should point to an \fBint\fP variable. (?J) and
1209     (?-J) set and unset the local PCRE_DUPNAMES option, respectively.
1210 ph10 169 .sp
1211 ph10 678 PCRE_INFO_JIT
1212     .sp
1213 ph10 921 Return 1 if the pattern was studied with one of the JIT options, and
1214 ph10 678 just-in-time compiling was successful. The fourth argument should point to an
1215     \fBint\fP variable. A return value of 0 means that JIT support is not available
1216 ph10 921 in this version of PCRE, or that the pattern was not studied with a JIT option,
1217     or that the JIT compiler could not handle this particular pattern. See the
1218 ph10 678 .\" HREF
1219     \fBpcrejit\fP
1220     .\"
1221     documentation for details of what can and cannot be handled.
1222     .sp
1223 ph10 836 PCRE_INFO_JITSIZE
1224     .sp
1225 ph10 921 If the pattern was successfully studied with a JIT option, return the size of
1226     the JIT compiled code, otherwise return zero. The fourth argument should point
1227     to a \fBsize_t\fP variable.
1228 ph10 836 .sp
1229 nigel 63 PCRE_INFO_LASTLITERAL
1230 nigel 75 .sp
1231 ph10 856 Return the value of the rightmost literal data unit that must exist in any
1232     matched string, other than at its start, if such a value has been recorded. The
1233     fourth argument should point to an \fBint\fP variable. If there is no such
1234     value, -1 is returned. For anchored patterns, a last literal value is recorded
1235     only if it follows something of variable length. For example, for the pattern
1236 nigel 75 /^a\ed+z\ed+/ the returned value is "z", but for /^a\edz\ed/ the returned value
1237 nigel 65 is -1.
1238 nigel 75 .sp
1239 ph10 932 PCRE_INFO_MAXLOOKBEHIND
1240     .sp
1241     Return the number of characters (NB not bytes) in the longest lookbehind
1242     assertion in the pattern. Note that the simple assertions \eb and \eB require a
1243     one-character lookbehind. This information is useful when doing multi-segment
1244     matching using the partial matching facilities.
1245     .sp
1246 ph10 455 PCRE_INFO_MINLENGTH
1247     .sp
1248     If the pattern was studied and a minimum length for matching subject strings
1249     was computed, its value is returned. Otherwise the returned value is -1. The
1250 ph10 856 value is a number of characters, which in UTF-8 mode may be different from the
1251     number of bytes. The fourth argument should point to an \fBint\fP variable. A
1252 ph10 455 non-negative value is a lower bound to the length of any matching string. There
1253     may not be any strings of that length that do actually match, but every string
1254     that does match is at least that long.
1255     .sp
1256 nigel 63 PCRE_INFO_NAMECOUNT
1257     PCRE_INFO_NAMEENTRYSIZE
1258     PCRE_INFO_NAMETABLE
1259 nigel 75 .sp
1260 nigel 63 PCRE supports the use of named as well as numbered capturing parentheses. The
1261     names are just an additional way of identifying the parentheses, which still
1262 nigel 91 acquire numbers. Several convenience functions such as
1263     \fBpcre_get_named_substring()\fP are provided for extracting captured
1264     substrings by name. It is also possible to extract the data directly, by first
1265     converting the name to a number in order to access the correct pointers in the
1266     output vector (described with \fBpcre_exec()\fP below). To do the conversion,
1267     you need to use the name-to-number map, which is described by these three
1268     values.
1269 nigel 75 .P
1270 nigel 63 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT gives
1271     the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size of each
1272 nigel 75 entry; both of these return an \fBint\fP value. The entry size depends on the
1273 nigel 63 length of the longest name. PCRE_INFO_NAMETABLE returns a pointer to the first
1274 ph10 856 entry of the table. This is a pointer to \fBchar\fP in the 8-bit library, where
1275     the first two bytes of each entry are the number of the capturing parenthesis,
1276     most significant byte first. In the 16-bit library, the pointer points to
1277     16-bit data units, the first of which contains the parenthesis number. The rest
1278     of the entry is the corresponding name, zero terminated.
1279 ph10 457 .P
1280     The names are in alphabetical order. Duplicate names may appear if (?| is used
1281     to create multiple groups with the same number, as described in the
1282     .\" HTML <a href="pcrepattern.html#dupsubpatternnumber">
1283     .\" </a>
1284     section on duplicate subpattern numbers
1285     .\"
1286     in the
1287     .\" HREF
1288     \fBpcrepattern\fP
1289     .\"
1290 ph10 461 page. Duplicate names for subpatterns with different numbers are permitted only
1291     if PCRE_DUPNAMES is set. In all cases of duplicate names, they appear in the
1292     table in the order in which they were found in the pattern. In the absence of
1293     (?| this is the order of increasing number; when (?| is used this is not
1294 ph10 457 necessarily the case because later subpatterns may have lower numbers.
1295     .P
1296     As a simple example of the name/number table, consider the following pattern
1297 ph10 856 after compilation by the 8-bit library (assume PCRE_EXTENDED is set, so white
1298     space - including newlines - is ignored):
1299 nigel 75 .sp
1300     .\" JOIN
1301 nigel 93 (?<date> (?<year>(\ed\ed)?\ed\ed) -
1302     (?<month>\ed\ed) - (?<day>\ed\ed) )
1303 nigel 75 .sp
1304 nigel 63 There are four named subpatterns, so the table has four entries, and each entry
1305     in the table is eight bytes long. The table is as follows, with non-printing
1306 nigel 75 bytes shows in hexadecimal, and undefined bytes shown as ??:
1307     .sp
1308 nigel 63 00 01 d a t e 00 ??
1309     00 05 d a y 00 ?? ??
1310     00 04 m o n t h 00
1311     00 02 y e a r 00 ??
1312 nigel 75 .sp
1313     When writing code to extract data from named subpatterns using the
1314 nigel 91 name-to-number map, remember that the length of the entries is likely to be
1315 nigel 75 different for each compiled pattern.
1316     .sp
1317 ph10 169 PCRE_INFO_OKPARTIAL
1318     .sp
1319 ph10 435 Return 1 if the pattern can be used for partial matching with
1320     \fBpcre_exec()\fP, otherwise 0. The fourth argument should point to an
1321     \fBint\fP variable. From release 8.00, this always returns 1, because the
1322     restrictions that previously applied to partial matching have been lifted. The
1323 ph10 169 .\" HREF
1324     \fBpcrepartial\fP
1325     .\"
1326 ph10 426 documentation gives details of partial matching.
1327 ph10 169 .sp
1328 nigel 63 PCRE_INFO_OPTIONS
1329 nigel 75 .sp
1330 nigel 63 Return a copy of the options with which the pattern was compiled. The fourth
1331 nigel 75 argument should point to an \fBunsigned long int\fP variable. These option bits
1332     are those specified in the call to \fBpcre_compile()\fP, modified by any
1333 ph10 196 top-level option settings at the start of the pattern itself. In other words,
1334     they are the options that will be in force when matching starts. For example,
1335     if the pattern /(?im)abc(?-i)d/ is compiled with the PCRE_EXTENDED option, the
1336     result is PCRE_CASELESS, PCRE_MULTILINE, and PCRE_EXTENDED.
1337 nigel 75 .P
1338 nigel 63 A pattern is automatically anchored by PCRE if all of its top-level
1339     alternatives begin with one of the following:
1340 nigel 75 .sp
1341 nigel 63 ^ unless PCRE_MULTILINE is set
1342 nigel 75 \eA always
1343     \eG always
1344     .\" JOIN
1345 nigel 63 .* if PCRE_DOTALL is set and there are no back
1346     references to the subpattern in which .* appears
1347 nigel 75 .sp
1348 nigel 63 For such patterns, the PCRE_ANCHORED bit is set in the options returned by
1349 nigel 75 \fBpcre_fullinfo()\fP.
1350     .sp
1351 nigel 63 PCRE_INFO_SIZE
1352 nigel 75 .sp
1353 ph10 856 Return the size of the compiled pattern in bytes (for both libraries). The
1354     fourth argument should point to a \fBsize_t\fP variable. This value does not
1355     include the size of the \fBpcre\fP structure that is returned by
1356     \fBpcre_compile()\fP. The value that is passed as the argument to
1357     \fBpcre_malloc()\fP when \fBpcre_compile()\fP is getting memory in which to
1358     place the compiled data is the value returned by this option plus the size of
1359     the \fBpcre\fP structure. Studying a compiled pattern, with or without JIT,
1360     does not alter the value returned by this option.
1361 nigel 75 .sp
1362 nigel 63 PCRE_INFO_STUDYSIZE
1363 nigel 75 .sp
1364 ph10 856 Return the size in bytes of the data block pointed to by the \fIstudy_data\fP
1365     field in a \fBpcre_extra\fP block. If \fBpcre_extra\fP is NULL, or there is no
1366     study data, zero is returned. The fourth argument should point to a
1367     \fBsize_t\fP variable. The \fIstudy_data\fP field is set by \fBpcre_study()\fP
1368     to record information that will speed up matching (see the section entitled
1369 ph10 656 .\" HTML <a href="#studyingapattern">
1370     .\" </a>
1371 ph10 659 "Studying a pattern"
1372 ph10 656 .\"
1373     above). The format of the \fIstudy_data\fP block is private, but its length
1374     is made available via this option so that it can be saved and restored (see the
1375     .\" HREF
1376     \fBpcreprecompile\fP
1377     .\"
1378     documentation for details).
1379 nigel 75 .
1380     .
1381 nigel 77 .SH "REFERENCE COUNTS"
1382 nigel 63 .rs
1383     .sp
1384 nigel 77 .B int pcre_refcount(pcre *\fIcode\fP, int \fIadjust\fP);
1385     .PP
1386     The \fBpcre_refcount()\fP function is used to maintain a reference count in the
1387     data block that contains a compiled pattern. It is provided for the benefit of
1388     applications that operate in an object-oriented manner, where different parts
1389     of the application may be using the same compiled pattern, but you want to free
1390     the block when they are all done.
1391     .P
1392     When a pattern is compiled, the reference count field is initialized to zero.
1393     It is changed only by calling this function, whose action is to add the
1394     \fIadjust\fP value (which may be positive or negative) to it. The yield of the
1395     function is the new value. However, the value of the count is constrained to
1396     lie between 0 and 65535, inclusive. If the new value is outside these limits,
1397     it is forced to the appropriate limit value.
1398     .P
1399     Except when it is zero, the reference count is not correctly preserved if a
1400     pattern is compiled on one host and then transferred to a host whose byte-order
1401     is different. (This seems a highly unlikely scenario.)
1402     .
1403     .
1404     .SH "MATCHING A PATTERN: THE TRADITIONAL FUNCTION"
1405     .rs
1406     .sp
1407 nigel 75 .B int pcre_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
1408 nigel 63 .ti +5n
1409 nigel 75 .B "const char *\fIsubject\fP," int \fIlength\fP, int \fIstartoffset\fP,
1410 nigel 63 .ti +5n
1411 nigel 75 .B int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP);
1412     .P
1413     The function \fBpcre_exec()\fP is called to match a subject string against a
1414     compiled pattern, which is passed in the \fIcode\fP argument. If the
1415 ph10 455 pattern was studied, the result of the study should be passed in the
1416 ph10 707 \fIextra\fP argument. You can call \fBpcre_exec()\fP with the same \fIcode\fP
1417     and \fIextra\fP arguments as many times as you like, in order to match
1418     different subject strings with the same pattern.
1419     .P
1420     This function is the main matching facility of the library, and it operates in
1421     a Perl-like manner. For specialist use there is also an alternative matching
1422     function, which is described
1423 nigel 77 .\" HTML <a href="#dfamatch">
1424     .\" </a>
1425     below
1426     .\"
1427     in the section about the \fBpcre_dfa_exec()\fP function.
1428 nigel 75 .P
1429     In most applications, the pattern will have been compiled (and optionally
1430     studied) in the same process that calls \fBpcre_exec()\fP. However, it is
1431     possible to save compiled patterns and study data, and then use them later
1432     in different processes, possibly even on different hosts. For a discussion
1433     about this, see the
1434     .\" HREF
1435     \fBpcreprecompile\fP
1436     .\"
1437     documentation.
1438     .P
1439     Here is an example of a simple call to \fBpcre_exec()\fP:
1440     .sp
1441 nigel 63 int rc;
1442     int ovector[30];
1443     rc = pcre_exec(
1444     re, /* result of pcre_compile() */
1445     NULL, /* we didn't study the pattern */
1446     "some string", /* the subject string */
1447     11, /* the length of the subject string */
1448     0, /* start at offset 0 in the subject */
1449     0, /* default options */
1450 nigel 75 ovector, /* vector of integers for substring information */
1451 nigel 77 30); /* number of elements (NOT size in bytes) */
1452 nigel 75 .
1453 ph10 656 .
1454 nigel 75 .\" HTML <a name="extradata"></a>
1455     .SS "Extra data for \fBpcre_exec()\fR"
1456     .rs
1457     .sp
1458     If the \fIextra\fP argument is not NULL, it must point to a \fBpcre_extra\fP
1459     data block. The \fBpcre_study()\fP function returns such a block (when it
1460 nigel 63 doesn't return NULL), but you can also create one for yourself, and pass
1461 nigel 87 additional information in it. The \fBpcre_extra\fP block contains the following
1462     fields (not necessarily in this order):
1463 nigel 75 .sp
1464     unsigned long int \fIflags\fP;
1465     void *\fIstudy_data\fP;
1466 ph10 691 void *\fIexecutable_jit\fP;
1467 nigel 75 unsigned long int \fImatch_limit\fP;
1468 nigel 87 unsigned long int \fImatch_limit_recursion\fP;
1469 nigel 75 void *\fIcallout_data\fP;
1470     const unsigned char *\fItables\fP;
1471 ph10 512 unsigned char **\fImark\fP;
1472 nigel 75 .sp
1473 ph10 903 In the 16-bit version of this structure, the \fImark\fP field has type
1474 ph10 859 "PCRE_UCHAR16 **".
1475 ph10 856 .P
1476 ph10 926 The \fIflags\fP field is used to specify which of the other fields are set. The
1477     flag bits are:
1478 nigel 75 .sp
1479 ph10 922 PCRE_EXTRA_CALLOUT_DATA
1480 ph10 691 PCRE_EXTRA_EXECUTABLE_JIT
1481 ph10 922 PCRE_EXTRA_MARK
1482 nigel 63 PCRE_EXTRA_MATCH_LIMIT
1483 nigel 87 PCRE_EXTRA_MATCH_LIMIT_RECURSION
1484 ph10 922 PCRE_EXTRA_STUDY_DATA
1485 nigel 75 PCRE_EXTRA_TABLES
1486     .sp
1487 ph10 678 Other flag bits should be set to zero. The \fIstudy_data\fP field and sometimes
1488     the \fIexecutable_jit\fP field are set in the \fBpcre_extra\fP block that is
1489     returned by \fBpcre_study()\fP, together with the appropriate flag bits. You
1490 ph10 922 should not set these yourself, but you may add to the block by setting other
1491     fields and their corresponding flag bits.
1492 nigel 75 .P
1493     The \fImatch_limit\fP field provides a means of preventing PCRE from using up a
1494 nigel 63 vast amount of resources when running patterns that are not going to match,
1495     but which have a very large number of possibilities in their search trees. The
1496 ph10 456 classic example is a pattern that uses nested unlimited repeats.
1497 nigel 75 .P
1498 ph10 678 Internally, \fBpcre_exec()\fP uses a function called \fBmatch()\fP, which it
1499     calls repeatedly (sometimes recursively). The limit set by \fImatch_limit\fP is
1500     imposed on the number of times this function is called during a match, which
1501     has the effect of limiting the amount of backtracking that can take place. For
1502     patterns that are not anchored, the count restarts from zero for each position
1503 ph10 691 in the subject string.
1504 nigel 75 .P
1505 ph10 691 When \fBpcre_exec()\fP is called with a pattern that was successfully studied
1506 ph10 921 with a JIT option, the way that the matching is executed is entirely different.
1507     However, there is still the possibility of runaway matching that goes on for a
1508     very long time, and so the \fImatch_limit\fP value is also used in this case
1509     (but in a different way) to limit how long the matching can continue.
1510 ph10 678 .P
1511 nigel 87 The default value for the limit can be set when PCRE is built; the default
1512 nigel 63 default is 10 million, which handles all but the most extreme cases. You can
1513 nigel 87 override the default by suppling \fBpcre_exec()\fP with a \fBpcre_extra\fP
1514     block in which \fImatch_limit\fP is set, and PCRE_EXTRA_MATCH_LIMIT is set in
1515     the \fIflags\fP field. If the limit is exceeded, \fBpcre_exec()\fP returns
1516     PCRE_ERROR_MATCHLIMIT.
1517 nigel 75 .P
1518 nigel 87 The \fImatch_limit_recursion\fP field is similar to \fImatch_limit\fP, but
1519     instead of limiting the total number of times that \fBmatch()\fP is called, it
1520     limits the depth of recursion. The recursion depth is a smaller number than the
1521     total number of calls, because not all calls to \fBmatch()\fP are recursive.
1522 ph10 691 This limit is of use only if it is set smaller than \fImatch_limit\fP.
1523 nigel 87 .P
1524 ph10 678 Limiting the recursion depth limits the amount of machine stack that can be
1525     used, or, when PCRE has been compiled to use memory on the heap instead of the
1526 ph10 686 stack, the amount of heap memory that can be used. This limit is not relevant,
1527 ph10 921 and is ignored, when matching is done using JIT compiled code.
1528 nigel 87 .P
1529     The default value for \fImatch_limit_recursion\fP can be set when PCRE is
1530     built; the default default is the same value as the default for
1531     \fImatch_limit\fP. You can override the default by suppling \fBpcre_exec()\fP
1532     with a \fBpcre_extra\fP block in which \fImatch_limit_recursion\fP is set, and
1533     PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the \fIflags\fP field. If the limit
1534     is exceeded, \fBpcre_exec()\fP returns PCRE_ERROR_RECURSIONLIMIT.
1535     .P
1536 ph10 440 The \fIcallout_data\fP field is used in conjunction with the "callout" feature,
1537     and is described in the
1538 nigel 75 .\" HREF
1539     \fBpcrecallout\fP
1540     .\"
1541     documentation.
1542     .P
1543     The \fItables\fP field is used to pass a character tables pointer to
1544     \fBpcre_exec()\fP; this overrides the value that is stored with the compiled
1545     pattern. A non-NULL value is stored with the compiled pattern only if custom
1546     tables were supplied to \fBpcre_compile()\fP via its \fItableptr\fP argument.
1547     If NULL is passed to \fBpcre_exec()\fP using this mechanism, it forces PCRE's
1548     internal tables to be used. This facility is helpful when re-using patterns
1549     that have been saved after compiling with an external set of tables, because
1550     the external tables might be at a different address when \fBpcre_exec()\fP is
1551     called. See the
1552     .\" HREF
1553     \fBpcreprecompile\fP
1554     .\"
1555     documentation for a discussion of saving compiled patterns for later use.
1556 ph10 510 .P
1557     If PCRE_EXTRA_MARK is set in the \fIflags\fP field, the \fImark\fP field must
1558 ph10 856 be set to point to a suitable variable. If the pattern contains any
1559 ph10 510 backtracking control verbs such as (*MARK:NAME), and the execution ends up with
1560     a name to pass back, a pointer to the name string (zero terminated) is placed
1561     in the variable pointed to by the \fImark\fP field. The names are within the
1562     compiled pattern; if you wish to retain such a name you must copy it before
1563     freeing the memory of a compiled pattern. If there is no name to pass back, the
1564 ph10 856 variable pointed to by the \fImark\fP field is set to NULL. For details of the
1565 ph10 510 backtracking control verbs, see the section entitled
1566     .\" HTML <a href="pcrepattern#backtrackcontrol">
1567     .\" </a>
1568     "Backtracking control"
1569     .\"
1570     in the
1571     .\" HREF
1572     \fBpcrepattern\fP
1573     .\"
1574     documentation.
1575 nigel 75 .
1576 ph10 510 .
1577 ph10 226 .\" HTML <a name="execoptions"></a>
1578 nigel 75 .SS "Option bits for \fBpcre_exec()\fP"
1579     .rs
1580     .sp
1581     The unused bits of the \fIoptions\fP argument for \fBpcre_exec()\fP must be
1582 nigel 91 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_\fIxxx\fP,
1583 ph10 442 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
1584 ph10 921 PCRE_NO_START_OPTIMIZE, PCRE_NO_UTF8_CHECK, PCRE_PARTIAL_HARD, and
1585     PCRE_PARTIAL_SOFT.
1586 ph10 686 .P
1587 ph10 921 If the pattern was successfully studied with one of the just-in-time (JIT)
1588     compile options, the only supported options for JIT execution are
1589     PCRE_NO_UTF8_CHECK, PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY,
1590     PCRE_NOTEMPTY_ATSTART, PCRE_PARTIAL_HARD, and PCRE_PARTIAL_SOFT. If an
1591     unsupported option is used, JIT execution is disabled and the normal
1592     interpretive code in \fBpcre_exec()\fP is run.
1593 nigel 75 .sp
1594     PCRE_ANCHORED
1595     .sp
1596     The PCRE_ANCHORED option limits \fBpcre_exec()\fP to matching at the first
1597     matching position. If a pattern was compiled with PCRE_ANCHORED, or turned out
1598     to be anchored by virtue of its contents, it cannot be made unachored at
1599     matching time.
1600     .sp
1601 ph10 231 PCRE_BSR_ANYCRLF
1602     PCRE_BSR_UNICODE
1603     .sp
1604     These options (which are mutually exclusive) control what the \eR escape
1605     sequence matches. The choice is either to match only CR, LF, or CRLF, or to
1606     match any Unicode newline sequence. These options override the choice that was
1607     made or defaulted when the pattern was compiled.
1608     .sp
1609 nigel 91 PCRE_NEWLINE_CR
1610     PCRE_NEWLINE_LF
1611     PCRE_NEWLINE_CRLF
1612 ph10 150 PCRE_NEWLINE_ANYCRLF
1613 nigel 93 PCRE_NEWLINE_ANY
1614 nigel 91 .sp
1615     These options override the newline definition that was chosen or defaulted when
1616 nigel 93 the pattern was compiled. For details, see the description of
1617     \fBpcre_compile()\fP above. During matching, the newline choice affects the
1618     behaviour of the dot, circumflex, and dollar metacharacters. It may also alter
1619     the way the match position is advanced after a match failure for an unanchored
1620 ph10 227 pattern.
1621 ph10 225 .P
1622 ph10 226 When PCRE_NEWLINE_CRLF, PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is set, and a
1623     match attempt for an unanchored pattern fails when the current position is at a
1624 ph10 230 CRLF sequence, and the pattern contains no explicit matches for CR or LF
1625 ph10 226 characters, the match position is advanced by two characters instead of one, in
1626     other words, to after the CRLF.
1627     .P
1628 ph10 227 The above rule is a compromise that makes the most common cases work as
1629     expected. For example, if the pattern is .+A (and the PCRE_DOTALL option is not
1630     set), it does not match the string "\er\enA" because, after failing at the
1631     start, it skips both the CR and the LF before retrying. However, the pattern
1632     [\er\en]A does match that string, because it contains an explicit CR or LF
1633 ph10 226 reference, and so advances only by one character after the first failure.
1634     .P
1635 ph10 231 An explicit match for CR of LF is either a literal appearance of one of those
1636     characters, or one of the \er or \en escape sequences. Implicit matches such as
1637 ph10 230 [^X] do not count, nor does \es (which includes CR and LF in the characters
1638     that it matches).
1639     .P
1640 ph10 226 Notwithstanding the above, anomalous effects may still occur when CRLF is a
1641     valid newline sequence and explicit \er or \en escapes appear in the pattern.
1642 nigel 91 .sp
1643 nigel 63 PCRE_NOTBOL
1644 nigel 75 .sp
1645     This option specifies that first character of the subject string is not the
1646     beginning of a line, so the circumflex metacharacter should not match before
1647     it. Setting this without PCRE_MULTILINE (at compile time) causes circumflex
1648     never to match. This option affects only the behaviour of the circumflex
1649     metacharacter. It does not affect \eA.
1650     .sp
1651 nigel 63 PCRE_NOTEOL
1652 nigel 75 .sp
1653     This option specifies that the end of the subject string is not the end of a
1654     line, so the dollar metacharacter should not match it nor (except in multiline
1655     mode) a newline immediately before it. Setting this without PCRE_MULTILINE (at
1656     compile time) causes dollar never to match. This option affects only the
1657     behaviour of the dollar metacharacter. It does not affect \eZ or \ez.
1658     .sp
1659 nigel 63 PCRE_NOTEMPTY
1660 nigel 75 .sp
1661 nigel 63 An empty string is not considered to be a valid match if this option is set. If
1662     there are alternatives in the pattern, they are tried. If all the alternatives
1663     match the empty string, the entire match fails. For example, if the pattern
1664 nigel 75 .sp
1665 nigel 63 a?b?
1666 nigel 75 .sp
1667 ph10 442 is applied to a string not beginning with "a" or "b", it matches an empty
1668 nigel 63 string at the start of the subject. With PCRE_NOTEMPTY set, this match is not
1669     valid, so PCRE searches further into the string for occurrences of "a" or "b".
1670 ph10 442 .sp
1671     PCRE_NOTEMPTY_ATSTART
1672     .sp
1673 ph10 461 This is like PCRE_NOTEMPTY, except that an empty string match that is not at
1674 ph10 442 the start of the subject is permitted. If the pattern is anchored, such a match
1675     can occur only if the pattern contains \eK.
1676 nigel 75 .P
1677 ph10 442 Perl has no direct equivalent of PCRE_NOTEMPTY or PCRE_NOTEMPTY_ATSTART, but it
1678     does make a special case of a pattern match of the empty string within its
1679     \fBsplit()\fP function, and when using the /g modifier. It is possible to
1680     emulate Perl's behaviour after matching a null string by first trying the match
1681     again at the same offset with PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED, and then
1682     if that fails, by advancing the starting offset (see below) and trying an
1683     ordinary match again. There is some code that demonstrates how to do this in
1684     the
1685 ph10 429 .\" HREF
1686     \fBpcredemo\fP
1687     .\"
1688 ph10 572 sample program. In the most general case, you have to check to see if the
1689     newline convention recognizes CRLF as a newline, and if so, and the current
1690 ph10 566 character is CR followed by LF, advance the starting offset by two characters
1691     instead of one.
1692 nigel 75 .sp
1693 ph10 389 PCRE_NO_START_OPTIMIZE
1694     .sp
1695 ph10 392 There are a number of optimizations that \fBpcre_exec()\fP uses at the start of
1696 ph10 542 a match, in order to speed up the process. For example, if it is known that an
1697     unanchored match must start with a specific character, it searches the subject
1698     for that character, and fails immediately if it cannot find it, without
1699 ph10 545 actually running the main matching function. This means that a special item
1700 ph10 542 such as (*COMMIT) at the start of a pattern is not considered until after a
1701 ph10 548 suitable starting point for the match has been found. When callouts or (*MARK)
1702 ph10 546 items are in use, these "start-up" optimizations can cause them to be skipped
1703     if the pattern is never actually used. The start-up optimizations are in effect
1704     a pre-scan of the subject that takes place before the pattern is run.
1705     .P
1706     The PCRE_NO_START_OPTIMIZE option disables the start-up optimizations, possibly
1707     causing performance to suffer, but ensuring that in cases where the result is
1708     "no match", the callouts do occur, and that items such as (*COMMIT) and (*MARK)
1709 ph10 579 are considered at every possible starting position in the subject string. If
1710     PCRE_NO_START_OPTIMIZE is set at compile time, it cannot be unset at matching
1711 ph10 921 time. The use of PCRE_NO_START_OPTIMIZE disables JIT execution; when it is set,
1712     matching is always done using interpretively.
1713 ph10 576 .P
1714 ph10 546 Setting PCRE_NO_START_OPTIMIZE can change the outcome of a matching operation.
1715     Consider the pattern
1716 ph10 389 .sp
1717 ph10 546 (*COMMIT)ABC
1718     .sp
1719     When this is compiled, PCRE records the fact that a match must start with the
1720 ph10 548 character "A". Suppose the subject string is "DEFABC". The start-up
1721     optimization scans along the subject, finds "A" and runs the first match
1722 ph10 546 attempt from there. The (*COMMIT) item means that the pattern must match the
1723 ph10 548 current starting position, which in this case, it does. However, if the same
1724     match is run with PCRE_NO_START_OPTIMIZE set, the initial scan along the
1725     subject string does not happen. The first match attempt is run starting from
1726 ph10 546 "D" and when this fails, (*COMMIT) prevents any further matches being tried, so
1727     the overall result is "no match". If the pattern is studied, more start-up
1728     optimizations may be used. For example, a minimum length for the subject may be
1729     recorded. Consider the pattern
1730     .sp
1731     (*MARK:A)(X|Y)
1732     .sp
1733 ph10 548 The minimum length for a match is one character. If the subject is "ABC", there
1734     will be attempts to match "ABC", "BC", "C", and then finally an empty string.
1735     If the pattern is studied, the final attempt does not take place, because PCRE
1736     knows that the subject is too short, and so the (*MARK) is never encountered.
1737     In this case, studying the pattern does not affect the overall match result,
1738     which is still "no match", but it does affect the auxiliary information that is
1739 ph10 546 returned.
1740     .sp
1741 nigel 75 PCRE_NO_UTF8_CHECK
1742     .sp
1743     When PCRE_UTF8 is set at compile time, the validity of the subject as a UTF-8
1744     string is automatically checked when \fBpcre_exec()\fP is subsequently called.
1745 ph10 959 The entire string is checked before any other processing takes place. The value
1746     of \fIstartoffset\fP is also checked to ensure that it points to the start of a
1747     UTF-8 character. There is a discussion about the
1748     .\" HTML <a href="pcreunicode.html#utf8strings">
1749     .\" </a>
1750     validity of UTF-8 strings
1751     .\"
1752     in the
1753 ph10 211 .\" HREF
1754 ph10 856 \fBpcreunicode\fP
1755 ph10 211 .\"
1756 ph10 856 page. If an invalid sequence of bytes is found, \fBpcre_exec()\fP returns the
1757     error PCRE_ERROR_BADUTF8 or, if PCRE_PARTIAL_HARD is set and the problem is a
1758     truncated character at the end of the subject, PCRE_ERROR_SHORTUTF8. In both
1759     cases, information about the precise nature of the error may also be returned
1760     (see the descriptions of these errors in the section entitled \fIError return
1761     values from\fP \fBpcre_exec()\fP
1762 ph10 598 .\" HTML <a href="#errorlist">
1763     .\" </a>
1764     below).
1765     .\"
1766     If \fIstartoffset\fP contains a value that does not point to the start of a
1767     UTF-8 character (or to the end of the subject), PCRE_ERROR_BADUTF8_OFFSET is
1768 ph10 569 returned.
1769 nigel 75 .P
1770     If you already know that your subject is valid, and you want to skip these
1771     checks for performance reasons, you can set the PCRE_NO_UTF8_CHECK option when
1772     calling \fBpcre_exec()\fP. You might want to do this for the second and
1773     subsequent calls to \fBpcre_exec()\fP if you are making repeated calls to find
1774     all the matches in a single subject string. However, you should be sure that
1775 ph10 856 the value of \fIstartoffset\fP points to the start of a character (or the end
1776     of the subject). When PCRE_NO_UTF8_CHECK is set, the effect of passing an
1777     invalid string as a subject or an invalid value of \fIstartoffset\fP is
1778 ph10 567 undefined. Your program may crash.
1779 nigel 75 .sp
1780 ph10 461 PCRE_PARTIAL_HARD
1781 ph10 428 PCRE_PARTIAL_SOFT
1782 nigel 75 .sp
1783 ph10 428 These options turn on the partial matching feature. For backwards
1784     compatibility, PCRE_PARTIAL is a synonym for PCRE_PARTIAL_SOFT. A partial match
1785     occurs if the end of the subject string is reached successfully, but there are
1786     not enough subject characters to complete the match. If this happens when
1787 ph10 553 PCRE_PARTIAL_SOFT (but not PCRE_PARTIAL_HARD) is set, matching continues by
1788     testing any remaining alternatives. Only if no complete match can be found is
1789     PCRE_ERROR_PARTIAL returned instead of PCRE_ERROR_NOMATCH. In other words,
1790     PCRE_PARTIAL_SOFT says that the caller is prepared to handle a partial match,
1791     but only if no complete match can be found.
1792     .P
1793     If PCRE_PARTIAL_HARD is set, it overrides PCRE_PARTIAL_SOFT. In this case, if a
1794     partial match is found, \fBpcre_exec()\fP immediately returns
1795     PCRE_ERROR_PARTIAL, without considering any other alternatives. In other words,
1796 ph10 572 when PCRE_PARTIAL_HARD is set, a partial match is considered to be more
1797 ph10 553 important that an alternative complete match.
1798     .P
1799     In both cases, the portion of the string that was inspected when the partial
1800     match was found is set as the first matching string. There is a more detailed
1801     discussion of partial and multi-segment matching, with examples, in the
1802 nigel 75 .\" HREF
1803     \fBpcrepartial\fP
1804     .\"
1805     documentation.
1806     .
1807 ph10 567 .
1808 nigel 75 .SS "The string to be matched by \fBpcre_exec()\fP"
1809     .rs
1810     .sp
1811     The subject string is passed to \fBpcre_exec()\fP as a pointer in
1812 ph10 856 \fIsubject\fP, a length in bytes in \fIlength\fP, and a starting byte offset
1813 ph10 572 in \fIstartoffset\fP. If this is negative or greater than the length of the
1814     subject, \fBpcre_exec()\fP returns PCRE_ERROR_BADOFFSET. When the starting
1815     offset is zero, the search for a match starts at the beginning of the subject,
1816     and this is by far the most common case. In UTF-8 mode, the byte offset must
1817     point to the start of a UTF-8 character (or the end of the subject). Unlike the
1818     pattern string, the subject may contain binary zero bytes.
1819 nigel 75 .P
1820 nigel 63 A non-zero starting offset is useful when searching for another match in the
1821 nigel 75 same subject by calling \fBpcre_exec()\fP again after a previous success.
1822     Setting \fIstartoffset\fP differs from just passing over a shortened string and
1823 nigel 63 setting PCRE_NOTBOL in the case of a pattern that begins with any kind of
1824     lookbehind. For example, consider the pattern
1825 nigel 75 .sp
1826     \eBiss\eB
1827     .sp
1828     which finds occurrences of "iss" in the middle of words. (\eB matches only if
1829 nigel 63 the current position in the subject is not a word boundary.) When applied to
1830 nigel 75 the string "Mississipi" the first call to \fBpcre_exec()\fP finds the first
1831     occurrence. If \fBpcre_exec()\fP is called again with just the remainder of the
1832     subject, namely "issipi", it does not match, because \eB is always false at the
1833 nigel 63 start of the subject, which is deemed to be a word boundary. However, if
1834 nigel 75 \fBpcre_exec()\fP is passed the entire string again, but with \fIstartoffset\fP
1835 nigel 63 set to 4, it finds the second occurrence of "iss" because it is able to look
1836     behind the starting point to discover that it is preceded by a letter.
1837 nigel 75 .P
1838 ph10 567 Finding all the matches in a subject is tricky when the pattern can match an
1839     empty string. It is possible to emulate Perl's /g behaviour by first trying the
1840     match again at the same offset, with the PCRE_NOTEMPTY_ATSTART and
1841     PCRE_ANCHORED options, and then if that fails, advancing the starting offset
1842     and trying an ordinary match again. There is some code that demonstrates how to
1843     do this in the
1844     .\" HREF
1845     \fBpcredemo\fP
1846     .\"
1847 ph10 572 sample program. In the most general case, you have to check to see if the
1848     newline convention recognizes CRLF as a newline, and if so, and the current
1849 ph10 567 character is CR followed by LF, advance the starting offset by two characters
1850     instead of one.
1851     .P
1852 nigel 63 If a non-zero starting offset is passed when the pattern is anchored, one
1853 nigel 75 attempt to match at the given offset is made. This can only succeed if the
1854 nigel 63 pattern does not require the match to be at the start of the subject.
1855 nigel 75 .
1856 ph10 567 .
1857 nigel 75 .SS "How \fBpcre_exec()\fP returns captured substrings"
1858     .rs
1859     .sp
1860 nigel 63 In general, a pattern matches a certain portion of the subject, and in
1861     addition, further substrings from the subject may be picked out by parts of the
1862     pattern. Following the usage in Jeffrey Friedl's book, this is called
1863     "capturing" in what follows, and the phrase "capturing subpattern" is used for
1864     a fragment of a pattern that picks out a substring. PCRE supports several other
1865     kinds of parenthesized subpattern that do not cause substrings to be captured.
1866 nigel 75 .P
1867 ph10 368 Captured substrings are returned to the caller via a vector of integers whose
1868     address is passed in \fIovector\fP. The number of elements in the vector is
1869     passed in \fIovecsize\fP, which must be a non-negative number. \fBNote\fP: this
1870     argument is NOT the size of \fIovector\fP in bytes.
1871 nigel 75 .P
1872     The first two-thirds of the vector is used to pass back captured substrings,
1873     each substring using a pair of integers. The remaining third of the vector is
1874     used as workspace by \fBpcre_exec()\fP while matching capturing subpatterns,
1875 ph10 368 and is not available for passing back information. The number passed in
1876 nigel 75 \fIovecsize\fP should always be a multiple of three. If it is not, it is
1877     rounded down.
1878     .P
1879     When a match is successful, information about captured substrings is returned
1880     in pairs of integers, starting at the beginning of \fIovector\fP, and
1881 ph10 371 continuing up to two-thirds of its length at the most. The first element of
1882 ph10 368 each pair is set to the byte offset of the first character in a substring, and
1883     the second is set to the byte offset of the first character after the end of a
1884     substring. \fBNote\fP: these values are always byte offsets, even in UTF-8
1885     mode. They are not character counts.
1886 nigel 75 .P
1887 ph10 368 The first pair of integers, \fIovector[0]\fP and \fIovector[1]\fP, identify the
1888     portion of the subject string matched by the entire pattern. The next pair is
1889     used for the first capturing subpattern, and so on. The value returned by
1890     \fBpcre_exec()\fP is one more than the highest numbered pair that has been set.
1891     For example, if two substrings have been captured, the returned value is 3. If
1892     there are no capturing subpatterns, the return value from a successful match is
1893     1, indicating that just the first pair of offsets has been set.
1894     .P
1895 nigel 63 If a capturing subpattern is matched repeatedly, it is the last portion of the
1896 nigel 75 string that it matched that is returned.
1897     .P
1898     If the vector is too small to hold all the captured substring offsets, it is
1899     used as far as possible (up to two-thirds of its length), and the function
1900 ph10 950 returns a value of zero. If neither the actual string matched nor any captured
1901 ph10 686 substrings are of interest, \fBpcre_exec()\fP may be called with \fIovector\fP
1902     passed as NULL and \fIovecsize\fP as zero. However, if the pattern contains
1903     back references and the \fIovector\fP is not big enough to remember the related
1904     substrings, PCRE has to get additional memory for use during matching. Thus it
1905     is usually advisable to supply an \fIovector\fP of reasonable size.
1906 nigel 75 .P
1907 ph10 686 There are some cases where zero is returned (indicating vector overflow) when
1908     in fact the vector is exactly the right size for the final match. For example,
1909     consider the pattern
1910     .sp
1911     (a)(?:(b)c|bd)
1912     .sp
1913     If a vector of 6 elements (allowing for only 1 captured substring) is given
1914     with subject string "abd", \fBpcre_exec()\fP will try to set the second
1915     captured string, thereby recording a vector overflow, before failing to match
1916 ph10 691 "c" and backing up to try the second alternative. The zero return, however,
1917 ph10 686 does correctly indicate that the maximum number of slots (namely 2) have been
1918     filled. In similar cases where there is temporary overflow, but the final
1919     number of used slots is actually less than the maximum, a non-zero value is
1920     returned.
1921     .P
1922 ph10 456 The \fBpcre_fullinfo()\fP function can be used to find out how many capturing
1923 nigel 63 subpatterns there are in a compiled pattern. The smallest size for
1924 nigel 75 \fIovector\fP that will allow for \fIn\fP captured substrings, in addition to
1925     the offsets of the substring matched by the whole pattern, is (\fIn\fP+1)*3.
1926 nigel 91 .P
1927     It is possible for capturing subpattern number \fIn+1\fP to match some part of
1928     the subject when subpattern \fIn\fP has not been used at all. For example, if
1929     the string "abc" is matched against the pattern (a|(z))(bc) the return from the
1930     function is 4, and subpatterns 1 and 3 are matched, but 2 is not. When this
1931     happens, both values in the offset pairs corresponding to unused subpatterns
1932     are set to -1.
1933     .P
1934     Offset values that correspond to unused subpatterns at the end of the
1935     expression are also set to -1. For example, if the string "abc" is matched
1936     against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not matched. The
1937     return from the function is 2, because the highest used capturing subpattern
1938 ph10 568 number is 1, and the offsets for for the second and third capturing subpatterns
1939     (assuming the vector is large enough, of course) are set to -1.
1940 nigel 91 .P
1941 ph10 686 \fBNote\fP: Elements in the first two-thirds of \fIovector\fP that do not
1942 ph10 683 correspond to capturing parentheses in the pattern are never changed. That is,
1943     if a pattern contains \fIn\fP capturing parentheses, no more than
1944     \fIovector[0]\fP to \fIovector[2n+1]\fP are set by \fBpcre_exec()\fP. The other
1945     elements (in the first two-thirds) retain whatever values they previously had.
1946 ph10 568 .P
1947 nigel 91 Some convenience functions are provided for extracting the captured substrings
1948     as separate strings. These are described below.
1949 nigel 75 .
1950 ph10 598 .
1951 nigel 77 .\" HTML <a name="errorlist"></a>
1952 nigel 91 .SS "Error return values from \fBpcre_exec()\fP"
1953 nigel 75 .rs
1954     .sp
1955     If \fBpcre_exec()\fP fails, it returns a negative number. The following are
1956 nigel 63 defined in the header file:
1957 nigel 75 .sp
1958 nigel 63 PCRE_ERROR_NOMATCH (-1)
1959 nigel 75 .sp
1960 nigel 63 The subject string did not match the pattern.
1961 nigel 75 .sp
1962 nigel 63 PCRE_ERROR_NULL (-2)
1963 nigel 75 .sp
1964     Either \fIcode\fP or \fIsubject\fP was passed as NULL, or \fIovector\fP was
1965     NULL and \fIovecsize\fP was not zero.
1966     .sp
1967 nigel 63 PCRE_ERROR_BADOPTION (-3)
1968 nigel 75 .sp
1969     An unrecognized bit was set in the \fIoptions\fP argument.
1970     .sp
1971 nigel 63 PCRE_ERROR_BADMAGIC (-4)
1972 nigel 75 .sp
1973 nigel 63 PCRE stores a 4-byte "magic number" at the start of the compiled code, to catch
1974 nigel 75 the case when it is passed a junk pointer and to detect when a pattern that was
1975     compiled in an environment of one endianness is run in an environment with the
1976     other endianness. This is the error that PCRE gives when the magic number is
1977     not present.
1978     .sp
1979 nigel 93 PCRE_ERROR_UNKNOWN_OPCODE (-5)
1980 nigel 75 .sp
1981 nigel 63 While running the pattern match, an unknown item was encountered in the
1982     compiled pattern. This error could be caused by a bug in PCRE or by overwriting
1983     of the compiled pattern.
1984 nigel 75 .sp
1985 nigel 63 PCRE_ERROR_NOMEMORY (-6)
1986 nigel 75 .sp
1987     If a pattern contains back references, but the \fIovector\fP that is passed to
1988     \fBpcre_exec()\fP is not big enough to remember the referenced substrings, PCRE
1989 nigel 63 gets a block of memory at the start of matching to use for this purpose. If the
1990 nigel 75 call via \fBpcre_malloc()\fP fails, this error is given. The memory is
1991     automatically freed at the end of matching.
1992 ph10 531 .P
1993 ph10 535 This error is also given if \fBpcre_stack_malloc()\fP fails in
1994 ph10 531 \fBpcre_exec()\fP. This can happen only when PCRE has been compiled with
1995     \fB--disable-stack-for-recursion\fP.
1996 nigel 75 .sp
1997 nigel 63 PCRE_ERROR_NOSUBSTRING (-7)
1998 nigel 75 .sp
1999     This error is used by the \fBpcre_copy_substring()\fP,
2000     \fBpcre_get_substring()\fP, and \fBpcre_get_substring_list()\fP functions (see
2001     below). It is never returned by \fBpcre_exec()\fP.
2002     .sp
2003 nigel 63 PCRE_ERROR_MATCHLIMIT (-8)
2004 nigel 75 .sp
2005 nigel 87 The backtracking limit, as specified by the \fImatch_limit\fP field in a
2006     \fBpcre_extra\fP structure (or defaulted) was reached. See the description
2007     above.
2008     .sp
2009 nigel 63 PCRE_ERROR_CALLOUT (-9)
2010 nigel 75 .sp
2011     This error is never generated by \fBpcre_exec()\fP itself. It is provided for
2012 nigel 63 use by callout functions that want to yield a distinctive error code. See the
2013 nigel 75 .\" HREF
2014     \fBpcrecallout\fP
2015     .\"
2016     documentation for details.
2017     .sp
2018 nigel 73 PCRE_ERROR_BADUTF8 (-10)
2019 nigel 75 .sp
2020 ph10 598 A string that contains an invalid UTF-8 byte sequence was passed as a subject,
2021     and the PCRE_NO_UTF8_CHECK option was not set. If the size of the output vector
2022     (\fIovecsize\fP) is at least 2, the byte offset to the start of the the invalid
2023     UTF-8 character is placed in the first element, and a reason code is placed in
2024     the second element. The reason codes are listed in the
2025     .\" HTML <a href="#badutf8reasons">
2026     .\" </a>
2027     following section.
2028     .\"
2029     For backward compatibility, if PCRE_PARTIAL_HARD is set and the problem is a
2030     truncated UTF-8 character at the end of the subject (reason codes 1 to 5),
2031     PCRE_ERROR_SHORTUTF8 is returned instead of PCRE_ERROR_BADUTF8.
2032 nigel 75 .sp
2033 nigel 73 PCRE_ERROR_BADUTF8_OFFSET (-11)
2034 nigel 75 .sp
2035 ph10 654 The UTF-8 byte sequence that was passed as a subject was checked and found to
2036 ph10 598 be valid (the PCRE_NO_UTF8_CHECK option was not set), but the value of
2037     \fIstartoffset\fP did not point to the beginning of a UTF-8 character or the
2038 ph10 569 end of the subject.
2039 nigel 75 .sp
2040 nigel 77 PCRE_ERROR_PARTIAL (-12)
2041 nigel 75 .sp
2042     The subject string did not match, but it did match partially. See the
2043     .\" HREF
2044     \fBpcrepartial\fP
2045     .\"
2046     documentation for details of partial matching.
2047     .sp
2048 nigel 77 PCRE_ERROR_BADPARTIAL (-13)
2049 nigel 75 .sp
2050 ph10 426 This code is no longer in use. It was formerly returned when the PCRE_PARTIAL
2051     option was used with a compiled pattern containing items that were not
2052 ph10 461 supported for partial matching. From release 8.00 onwards, there are no
2053 ph10 426 restrictions on partial matching.
2054 nigel 75 .sp
2055 nigel 77 PCRE_ERROR_INTERNAL (-14)
2056 nigel 75 .sp
2057     An unexpected internal error has occurred. This error could be caused by a bug
2058     in PCRE or by overwriting of the compiled pattern.
2059     .sp
2060 nigel 77 PCRE_ERROR_BADCOUNT (-15)
2061 nigel 75 .sp
2062     This error is given if the value of the \fIovecsize\fP argument is negative.
2063 nigel 93 .sp
2064     PCRE_ERROR_RECURSIONLIMIT (-21)
2065     .sp
2066     The internal recursion limit, as specified by the \fImatch_limit_recursion\fP
2067     field in a \fBpcre_extra\fP structure (or defaulted) was reached. See the
2068     description above.
2069     .sp
2070     PCRE_ERROR_BADNEWLINE (-23)
2071     .sp
2072     An invalid combination of PCRE_NEWLINE_\fIxxx\fP options was given.
2073 ph10 567 .sp
2074     PCRE_ERROR_BADOFFSET (-24)
2075     .sp
2076 ph10 572 The value of \fIstartoffset\fP was negative or greater than the length of the
2077 ph10 567 subject, that is, the value in \fIlength\fP.
2078 ph10 569 .sp
2079     PCRE_ERROR_SHORTUTF8 (-25)
2080     .sp
2081 ph10 598 This error is returned instead of PCRE_ERROR_BADUTF8 when the subject string
2082     ends with a truncated UTF-8 character and the PCRE_PARTIAL_HARD option is set.
2083     Information about the failure is returned as for PCRE_ERROR_BADUTF8. It is in
2084     fact sufficient to detect this case, but this special error code for
2085     PCRE_PARTIAL_HARD precedes the implementation of returned information; it is
2086     retained for backwards compatibility.
2087 ph10 642 .sp
2088     PCRE_ERROR_RECURSELOOP (-26)
2089     .sp
2090 ph10 654 This error is returned when \fBpcre_exec()\fP detects a recursion loop within
2091     the pattern. Specifically, it means that either the whole pattern or a
2092     subpattern has been called recursively for the second time at the same position
2093 ph10 642 in the subject string. Some simple patterns that might do this are detected and
2094     faulted at compile time, but more complicated cases, in particular mutual
2095     recursions between two different subpatterns, cannot be detected until run
2096     time.
2097 ph10 678 .sp
2098     PCRE_ERROR_JIT_STACKLIMIT (-27)
2099     .sp
2100 ph10 922 This error is returned when a pattern that was successfully studied using a
2101     JIT compile option is being matched, but the memory available for the
2102     just-in-time processing stack is not large enough. See the
2103 ph10 678 .\" HREF
2104     \fBpcrejit\fP
2105     .\"
2106 ph10 691 documentation for more details.
2107 ph10 856 .sp
2108 ph10 960 PCRE_ERROR_BADMODE (-28)
2109 ph10 856 .sp
2110 ph10 903 This error is given if a pattern that was compiled by the 8-bit library is
2111 ph10 856 passed to a 16-bit library function, or vice versa.
2112     .sp
2113 ph10 960 PCRE_ERROR_BADENDIANNESS (-29)
2114 ph10 903 .sp
2115     This error is given if a pattern that was compiled and saved is reloaded on a
2116     host with different endianness. The utility function
2117     \fBpcre_pattern_to_host_byte_order()\fP can be used to convert such a pattern
2118 ph10 856 so that it runs on the new host.
2119 nigel 93 .P
2120 ph10 960 Error numbers -16 to -20, -22, and -30 are not used by \fBpcre_exec()\fP.
2121 nigel 75 .
2122     .
2123 ph10 598 .\" HTML <a name="badutf8reasons"></a>
2124     .SS "Reason codes for invalid UTF-8 strings"
2125     .rs
2126     .sp
2127 ph10 903 This section applies only to the 8-bit library. The corresponding information
2128 ph10 856 for the 16-bit library is given in the
2129     .\" HREF
2130     \fBpcre16\fP
2131     .\"
2132     page.
2133     .P
2134 ph10 654 When \fBpcre_exec()\fP returns either PCRE_ERROR_BADUTF8 or
2135     PCRE_ERROR_SHORTUTF8, and the size of the output vector (\fIovecsize\fP) is at
2136     least 2, the offset of the start of the invalid UTF-8 character is placed in
2137     the first output vector element (\fIovector[0]\fP) and a reason code is placed
2138 ph10 598 in the second element (\fIovector[1]\fP). The reason codes are given names in
2139     the \fBpcre.h\fP header file:
2140     .sp
2141     PCRE_UTF8_ERR1
2142     PCRE_UTF8_ERR2
2143     PCRE_UTF8_ERR3
2144     PCRE_UTF8_ERR4
2145     PCRE_UTF8_ERR5
2146     .sp
2147 ph10 654 The string ends with a truncated UTF-8 character; the code specifies how many
2148 ph10 598 bytes are missing (1 to 5). Although RFC 3629 restricts UTF-8 characters to be
2149     no longer than 4 bytes, the encoding scheme (originally defined by RFC 2279)
2150 ph10 654 allows for up to 6 bytes, and this is checked first; hence the possibility of
2151 ph10 598 4 or 5 missing bytes.
2152     .sp
2153     PCRE_UTF8_ERR6
2154     PCRE_UTF8_ERR7
2155     PCRE_UTF8_ERR8
2156     PCRE_UTF8_ERR9
2157     PCRE_UTF8_ERR10
2158     .sp
2159 ph10 654 The two most significant bits of the 2nd, 3rd, 4th, 5th, or 6th byte of the
2160 ph10 598 character do not have the binary value 0b10 (that is, either the most
2161     significant bit is 0, or the next bit is 1).
2162 ph10 654 .sp
2163 ph10 598 PCRE_UTF8_ERR11
2164     PCRE_UTF8_ERR12
2165     .sp
2166 ph10 654 A character that is valid by the RFC 2279 rules is either 5 or 6 bytes long;
2167     these code points are excluded by RFC 3629.
2168     .sp
2169 ph10 598 PCRE_UTF8_ERR13
2170     .sp
2171 ph10 654 A 4-byte character has a value greater than 0x10fff; these code points are
2172 ph10 598 excluded by RFC 3629.
2173 ph10 654 .sp
2174 ph10 598 PCRE_UTF8_ERR14
2175     .sp
2176     A 3-byte character has a value in the range 0xd800 to 0xdfff; this range of
2177 ph10 654 code points are reserved by RFC 3629 for use with UTF-16, and so are excluded
2178 ph10 598 from UTF-8.
2179 ph10 654 .sp
2180 ph10 598 PCRE_UTF8_ERR15
2181     PCRE_UTF8_ERR16
2182     PCRE_UTF8_ERR17
2183     PCRE_UTF8_ERR18
2184     PCRE_UTF8_ERR19
2185     .sp
2186 ph10 654 A 2-, 3-, 4-, 5-, or 6-byte character is "overlong", that is, it codes for a
2187     value that can be represented by fewer bytes, which is invalid. For example,
2188 ph10 598 the two bytes 0xc0, 0xae give the value 0x2e, whose correct coding uses just
2189     one byte.
2190     .sp
2191     PCRE_UTF8_ERR20
2192     .sp
2193 ph10 654 The two most significant bits of the first byte of a character have the binary
2194     value 0b10 (that is, the most significant bit is 1 and the second is 0). Such a
2195 ph10 598 byte can only validly occur as the second or subsequent byte of a multi-byte
2196     character.
2197     .sp
2198     PCRE_UTF8_ERR21
2199     .sp
2200     The first byte of a character has the value 0xfe or 0xff. These values can
2201     never occur in a valid UTF-8 string.
2202     .
2203     .
2204 nigel 75 .SH "EXTRACTING CAPTURED SUBSTRINGS BY NUMBER"
2205 nigel 63 .rs
2206     .sp
2207 nigel 75 .B int pcre_copy_substring(const char *\fIsubject\fP, int *\fIovector\fP,
2208 nigel 63 .ti +5n
2209 nigel 75 .B int \fIstringcount\fP, int \fIstringnumber\fP, char *\fIbuffer\fP,
2210 nigel 63 .ti +5n
2211 nigel 75 .B int \fIbuffersize\fP);
2212 nigel 63 .PP
2213 nigel 75 .B int pcre_get_substring(const char *\fIsubject\fP, int *\fIovector\fP,
2214 nigel 63 .ti +5n
2215 nigel 75 .B int \fIstringcount\fP, int \fIstringnumber\fP,
2216 nigel 63 .ti +5n
2217 nigel 75 .B const char **\fIstringptr\fP);
2218 nigel 63 .PP
2219 nigel 75 .B int pcre_get_substring_list(const char *\fIsubject\fP,
2220 nigel 63 .ti +5n
2221 nigel 75 .B int *\fIovector\fP, int \fIstringcount\fP, "const char ***\fIlistptr\fP);"
2222 nigel 63 .PP
2223     Captured substrings can be accessed directly by using the offsets returned by
2224 nigel 75 \fBpcre_exec()\fP in \fIovector\fP. For convenience, the functions
2225     \fBpcre_copy_substring()\fP, \fBpcre_get_substring()\fP, and
2226     \fBpcre_get_substring_list()\fP are provided for extracting captured substrings
2227 nigel 63 as new, separate, zero-terminated strings. These functions identify substrings
2228     by number. The next section describes functions for extracting named
2229 nigel 91 substrings.
2230 nigel 75 .P
2231 nigel 91 A substring that contains a binary zero is correctly extracted and has a
2232     further zero added on the end, but the result is not, of course, a C string.
2233     However, you can process such a string by referring to the length that is
2234     returned by \fBpcre_copy_substring()\fP and \fBpcre_get_substring()\fP.
2235     Unfortunately, the interface to \fBpcre_get_substring_list()\fP is not adequate
2236     for handling strings containing binary zeros, because the end of the final
2237     string is not independently indicated.
2238     .P
2239 nigel 63 The first three arguments are the same for all three of these functions:
2240 nigel 75 \fIsubject\fP is the subject string that has just been successfully matched,
2241     \fIovector\fP is a pointer to the vector of integer offsets that was passed to
2242     \fBpcre_exec()\fP, and \fIstringcount\fP is the number of substrings that were
2243 nigel 63 captured by the match, including the substring that matched the entire regular
2244 nigel 75 expression. This is the value returned by \fBpcre_exec()\fP if it is greater
2245     than zero. If \fBpcre_exec()\fP returned zero, indicating that it ran out of
2246     space in \fIovector\fP, the value passed as \fIstringcount\fP should be the
2247     number of elements in the vector divided by three.
2248     .P
2249     The functions \fBpcre_copy_substring()\fP and \fBpcre_get_substring()\fP
2250     extract a single substring, whose number is given as \fIstringnumber\fP. A
2251     value of zero extracts the substring that matched the entire pattern, whereas
2252     higher values extract the captured substrings. For \fBpcre_copy_substring()\fP,
2253     the string is placed in \fIbuffer\fP, whose length is given by
2254     \fIbuffersize\fP, while for \fBpcre_get_substring()\fP a new block of memory is
2255     obtained via \fBpcre_malloc\fP, and its address is returned via
2256     \fIstringptr\fP. The yield of the function is the length of the string, not
2257 nigel 93 including the terminating zero, or one of these error codes:
2258 nigel 75 .sp
2259 nigel 63 PCRE_ERROR_NOMEMORY (-6)
2260 nigel 75 .sp
2261     The buffer was too small for \fBpcre_copy_substring()\fP, or the attempt to get
2262     memory failed for \fBpcre_get_substring()\fP.
2263     .sp
2264 nigel 63 PCRE_ERROR_NOSUBSTRING (-7)
2265 nigel 75 .sp
2266     There is no substring whose number is \fIstringnumber\fP.
2267     .P
2268     The \fBpcre_get_substring_list()\fP function extracts all available substrings
2269 nigel 63 and builds a list of pointers to them. All this is done in a single block of
2270 nigel 75 memory that is obtained via \fBpcre_malloc\fP. The address of the memory block
2271     is returned via \fIlistptr\fP, which is also the start of the list of string
2272 nigel 63 pointers. The end of the list is marked by a NULL pointer. The yield of the
2273 nigel 93 function is zero if all went well, or the error code
2274 nigel 75 .sp
2275 nigel 63 PCRE_ERROR_NOMEMORY (-6)
2276 nigel 75 .sp
2277 nigel 63 if the attempt to get the memory block failed.
2278 nigel 75 .P
2279 nigel 63 When any of these functions encounter a substring that is unset, which can
2280 nigel 75 happen when capturing subpattern number \fIn+1\fP matches some part of the
2281     subject, but subpattern \fIn\fP has not been used at all, they return an empty
2282 nigel 63 string. This can be distinguished from a genuine zero-length substring by
2283 nigel 75 inspecting the appropriate offset in \fIovector\fP, which is negative for unset
2284 nigel 63 substrings.
2285 nigel 75 .P
2286     The two convenience functions \fBpcre_free_substring()\fP and
2287     \fBpcre_free_substring_list()\fP can be used to free the memory returned by
2288     a previous call of \fBpcre_get_substring()\fP or
2289     \fBpcre_get_substring_list()\fP, respectively. They do nothing more than call
2290     the function pointed to by \fBpcre_free\fP, which of course could be called
2291 nigel 63 directly from a C program. However, PCRE is used in some situations where it is
2292 nigel 91 linked via a special interface to another programming language that cannot use
2293 nigel 75 \fBpcre_free\fP directly; it is for these cases that the functions are
2294 nigel 63 provided.
2295 nigel 75 .
2296     .
2297     .SH "EXTRACTING CAPTURED SUBSTRINGS BY NAME"
2298 nigel 63 .rs
2299     .sp
2300 nigel 75 .B int pcre_get_stringnumber(const pcre *\fIcode\fP,
2301 nigel 63 .ti +5n
2302 nigel 75 .B const char *\fIname\fP);
2303 nigel 63 .PP
2304 nigel 75 .B int pcre_copy_named_substring(const pcre *\fIcode\fP,
2305 nigel 63 .ti +5n
2306 nigel 75 .B const char *\fIsubject\fP, int *\fIovector\fP,
2307     .ti +5n
2308     .B int \fIstringcount\fP, const char *\fIstringname\fP,
2309     .ti +5n
2310     .B char *\fIbuffer\fP, int \fIbuffersize\fP);
2311 nigel 63 .PP
2312 nigel 75 .B int pcre_get_named_substring(const pcre *\fIcode\fP,
2313 nigel 63 .ti +5n
2314 nigel 75 .B const char *\fIsubject\fP, int *\fIovector\fP,
2315 nigel 63 .ti +5n
2316 nigel 75 .B int \fIstringcount\fP, const char *\fIstringname\fP,
2317 nigel 63 .ti +5n
2318 nigel 75 .B const char **\fIstringptr\fP);
2319 nigel 63 .PP
2320 nigel 75 To extract a substring by name, you first have to find associated number.
2321     For example, for this pattern
2322     .sp
2323 nigel 93 (a+)b(?<xxx>\ed+)...
2324 nigel 75 .sp
2325 nigel 91 the number of the subpattern called "xxx" is 2. If the name is known to be
2326     unique (PCRE_DUPNAMES was not set), you can find the number from the name by
2327     calling \fBpcre_get_stringnumber()\fP. The first argument is the compiled
2328     pattern, and the second is the name. The yield of the function is the
2329 nigel 75 subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no subpattern of
2330     that name.
2331     .P
2332     Given the number, you can extract the substring directly, or use one of the
2333     functions described in the previous section. For convenience, there are also
2334     two functions that do the whole job.
2335     .P
2336 nigel 91 Most of the arguments of \fBpcre_copy_named_substring()\fP and
2337     \fBpcre_get_named_substring()\fP are the same as those for the similarly named
2338 nigel 75 functions that extract by number. As these are described in the previous
2339     section, they are not re-described here. There are just two differences:
2340     .P
2341 nigel 63 First, instead of a substring number, a substring name is given. Second, there
2342     is an extra argument, given at the start, which is a pointer to the compiled
2343     pattern. This is needed in order to gain access to the name-to-number
2344     translation table.
2345 nigel 75 .P
2346     These functions call \fBpcre_get_stringnumber()\fP, and if it succeeds, they
2347 ph10 127 then call \fBpcre_copy_substring()\fP or \fBpcre_get_substring()\fP, as
2348     appropriate. \fBNOTE:\fP If PCRE_DUPNAMES is set and there are duplicate names,
2349 ph10 128 the behaviour may not be what you want (see the next section).
2350 ph10 385 .P
2351 ph10 457 \fBWarning:\fP If the pattern uses the (?| feature to set up multiple
2352     subpatterns with the same number, as described in the
2353     .\" HTML <a href="pcrepattern.html#dupsubpatternnumber">
2354     .\" </a>
2355     section on duplicate subpattern numbers
2356     .\"
2357     in the
2358     .\" HREF
2359     \fBpcrepattern\fP
2360     .\"
2361     page, you cannot use names to distinguish the different subpatterns, because
2362     names are not included in the compiled code. The matching process uses only
2363     numbers. For this reason, the use of different names for subpatterns of the
2364     same number causes an error at compile time.
2365 nigel 77 .
2366 ph10 686 .
2367 nigel 91 .SH "DUPLICATE SUBPATTERN NAMES"
2368     .rs
2369     .sp
2370     .B int pcre_get_stringtable_entries(const pcre *\fIcode\fP,
2371     .ti +5n
2372     .B const char *\fIname\fP, char **\fIfirst\fP, char **\fIlast\fP);
2373     .PP
2374     When a pattern is compiled with the PCRE_DUPNAMES option, names for subpatterns
2375 ph10 457 are not required to be unique. (Duplicate names are always allowed for
2376     subpatterns with the same number, created by using the (?| feature. Indeed, if
2377     such subpatterns are named, they are required to use the same names.)
2378     .P
2379     Normally, patterns with duplicate names are such that in any one match, only
2380     one of the named subpatterns participates. An example is shown in the
2381 nigel 91 .\" HREF
2382     \fBpcrepattern\fP
2383     .\"
2384 ph10 208 documentation.
2385 ph10 203 .P
2386     When duplicates are present, \fBpcre_copy_named_substring()\fP and
2387     \fBpcre_get_named_substring()\fP return the first substring corresponding to
2388     the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING (-7) is
2389     returned; no data is returned. The \fBpcre_get_stringnumber()\fP function
2390     returns one of the numbers that are associated with the name, but it is not
2391     defined which it is.
2392     .P
2393 nigel 91 If you want to get full details of all captured substrings for a given name,
2394     you must use the \fBpcre_get_stringtable_entries()\fP function. The first
2395     argument is the compiled pattern, and the second is the name. The third and
2396     fourth are pointers to variables which are updated by the function. After it
2397     has run, they point to the first and last entries in the name-to-number table
2398     for the given name. The function itself returns the length of each entry, or
2399 nigel 93 PCRE_ERROR_NOSUBSTRING (-7) if there are none. The format of the table is
2400 ph10 598 described above in the section entitled \fIInformation about a pattern\fP
2401     .\" HTML <a href="#infoaboutpattern">
2402     .\" </a>
2403     above.
2404     .\"
2405 nigel 93 Given all the relevant entries for the name, you can extract each of their
2406     numbers, and hence the captured data, if any.
2407 nigel 91 .
2408     .
2409 nigel 77 .SH "FINDING ALL POSSIBLE MATCHES"
2410     .rs
2411     .sp
2412     The traditional matching function uses a similar algorithm to Perl, which stops
2413     when it finds the first match, starting at a given point in the subject. If you
2414     want to find all possible matches, or the longest possible match, consider
2415     using the alternative matching function (see below) instead. If you cannot use
2416     the alternative function, but still need to find all possible matches, you
2417     can kludge it up by making use of the callout facility, which is described in
2418     the
2419     .\" HREF
2420     \fBpcrecallout\fP
2421     .\"
2422     documentation.
2423 nigel 75 .P
2424 nigel 77 What you have to do is to insert a callout right at the end of the pattern.
2425     When your callout function is called, extract and save the current matched
2426     substring. Then return 1, which forces \fBpcre_exec()\fP to backtrack and try
2427     other alternatives. Ultimately, when it runs out of matches, \fBpcre_exec()\fP
2428     will yield PCRE_ERROR_NOMATCH.
2429     .
2430     .
2431 ph10 901 .SH "OBTAINING AN ESTIMATE OF STACK USAGE"
2432     .rs
2433     .sp
2434     Matching certain patterns using \fBpcre_exec()\fP can use a lot of process
2435 ph10 903 stack, which in certain environments can be rather limited in size. Some users
2436     find it helpful to have an estimate of the amount of stack that is used by
2437 ph10 901 \fBpcre_exec()\fP, to help them set recursion limits, as described in the
2438     .\" HREF
2439     \fBpcrestack\fP
2440     .\"
2441 ph10 903 documentation. The estimate that is output by \fBpcretest\fP when called with
2442     the \fB-m\fP and \fB-C\fP options is obtained by calling \fBpcre_exec\fP with
2443 ph10 901 the values NULL, NULL, NULL, -999, and -999 for its first five arguments.
2444     .P
2445     Normally, if its first argument is NULL, \fBpcre_exec()\fP immediately returns
2446     the negative error code PCRE_ERROR_NULL, but with this special combination of
2447     arguments, it returns instead a negative number whose absolute value is the
2448     approximate stack frame size in bytes. (A negative number is used so that it is
2449     clear that no match has happened.) The value is approximate because in some
2450 ph10 903 cases, recursive calls to \fBpcre_exec()\fP occur when there are one or two
2451 ph10 901 additional variables on the stack.
2452     .P
2453 ph10 903 If PCRE has been compiled to use the heap instead of the stack for recursion,
2454 ph10 901 the value returned is the size of each block that is obtained from the heap.
2455     .
2456     .
2457 nigel 77 .\" HTML <a name="dfamatch"></a>
2458     .SH "MATCHING A PATTERN: THE ALTERNATIVE FUNCTION"
2459     .rs
2460     .sp
2461     .B int pcre_dfa_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
2462     .ti +5n
2463     .B "const char *\fIsubject\fP," int \fIlength\fP, int \fIstartoffset\fP,
2464     .ti +5n
2465     .B int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP,
2466     .ti +5n
2467     .B int *\fIworkspace\fP, int \fIwscount\fP);
2468     .P
2469     The function \fBpcre_dfa_exec()\fP is called to match a subject string against
2470 nigel 93 a compiled pattern, using a matching algorithm that scans the subject string
2471     just once, and does not backtrack. This has different characteristics to the
2472     normal algorithm, and is not compatible with Perl. Some of the features of PCRE
2473     patterns are not supported. Nevertheless, there are times when this kind of
2474 ph10 461 matching can be useful. For a discussion of the two matching algorithms, and a
2475 ph10 435 list of features that \fBpcre_dfa_exec()\fP does not support, see the
2476 nigel 77 .\" HREF
2477     \fBpcrematching\fP
2478     .\"
2479     documentation.
2480     .P
2481     The arguments for the \fBpcre_dfa_exec()\fP function are the same as for
2482     \fBpcre_exec()\fP, plus two extras. The \fIovector\fP argument is used in a
2483     different way, and this is described below. The other common arguments are used
2484     in the same way as for \fBpcre_exec()\fP, so their description is not repeated
2485     here.
2486     .P
2487     The two additional arguments provide workspace for the function. The workspace
2488     vector should contain at least 20 elements. It is used for keeping track of
2489     multiple paths through the pattern tree. More workspace will be needed for
2490 nigel 91 patterns and subjects where there are a lot of potential matches.
2491 nigel 77 .P
2492 nigel 87 Here is an example of a simple call to \fBpcre_dfa_exec()\fP:
2493 nigel 77 .sp
2494     int rc;
2495     int ovector[10];
2496     int wspace[20];
2497 nigel 87 rc = pcre_dfa_exec(
2498 nigel 77 re, /* result of pcre_compile() */
2499     NULL, /* we didn't study the pattern */
2500     "some string", /* the subject string */
2501     11, /* the length of the subject string */
2502     0, /* start at offset 0 in the subject */
2503     0, /* default options */
2504     ovector, /* vector of integers for substring information */
2505     10, /* number of elements (NOT size in bytes) */
2506     wspace, /* working space vector */
2507     20); /* number of elements (NOT size in bytes) */
2508     .
2509     .SS "Option bits for \fBpcre_dfa_exec()\fP"
2510     .rs
2511     .sp
2512     The unused bits of the \fIoptions\fP argument for \fBpcre_dfa_exec()\fP must be
2513 nigel 91 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_\fIxxx\fP,
2514 ph10 442 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
2515 ph10 542 PCRE_NO_UTF8_CHECK, PCRE_BSR_ANYCRLF, PCRE_BSR_UNICODE, PCRE_NO_START_OPTIMIZE,
2516     PCRE_PARTIAL_HARD, PCRE_PARTIAL_SOFT, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART.
2517     All but the last four of these are exactly the same as for \fBpcre_exec()\fP,
2518     so their description is not repeated here.
2519 nigel 77 .sp
2520 ph10 428 PCRE_PARTIAL_HARD
2521 ph10 461 PCRE_PARTIAL_SOFT
2522 nigel 77 .sp
2523 ph10 428 These have the same general effect as they do for \fBpcre_exec()\fP, but the
2524     details are slightly different. When PCRE_PARTIAL_HARD is set for
2525     \fBpcre_dfa_exec()\fP, it returns PCRE_ERROR_PARTIAL if the end of the subject
2526     is reached and there is still at least one matching possibility that requires
2527     additional characters. This happens even if some complete matches have also
2528     been found. When PCRE_PARTIAL_SOFT is set, the return code PCRE_ERROR_NOMATCH
2529     is converted into PCRE_ERROR_PARTIAL if the end of the subject is reached,
2530     there have been no complete matches, but there is still at least one matching
2531 ph10 435 possibility. The portion of the string that was inspected when the longest
2532     partial match was found is set as the first matching string in both cases.
2533 ph10 553 There is a more detailed discussion of partial and multi-segment matching, with
2534     examples, in the
2535     .\" HREF
2536     \fBpcrepartial\fP
2537     .\"
2538     documentation.
2539 nigel 77 .sp
2540     PCRE_DFA_SHORTEST
2541     .sp
2542     Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to stop as
2543 nigel 93 soon as it has found one match. Because of the way the alternative algorithm
2544     works, this is necessarily the shortest possible match at the first possible
2545     matching point in the subject string.
2546 nigel 77 .sp
2547     PCRE_DFA_RESTART
2548     .sp
2549 ph10 428 When \fBpcre_dfa_exec()\fP returns a partial match, it is possible to call it
2550     again, with additional subject characters, and have it continue with the same
2551     match. The PCRE_DFA_RESTART option requests this action; when it is set, the
2552     \fIworkspace\fP and \fIwscount\fP options must reference the same vector as
2553     before because data about the match so far is left in them after a partial
2554     match. There is more discussion of this facility in the
2555 nigel 77 .\" HREF
2556     \fBpcrepartial\fP
2557     .\"
2558     documentation.
2559     .
2560 ph10 598 .
2561 nigel 77 .SS "Successful returns from \fBpcre_dfa_exec()\fP"
2562     .rs
2563     .sp
2564     When \fBpcre_dfa_exec()\fP succeeds, it may have matched more than one
2565     substring in the subject. Note, however, that all the matches from one run of
2566     the function start at the same point in the subject. The shorter matches are
2567     all initial substrings of the longer matches. For example, if the pattern
2568     .sp
2569     <.*>
2570     .sp
2571     is matched against the string
2572     .sp
2573     This is <something> <something else> <something further> no more
2574     .sp
2575     the three matched strings are
2576     .sp
2577     <something>
2578     <something> <something else>
2579     <something> <something else> <something further>
2580     .sp
2581     On success, the yield of the function is a number greater than zero, which is
2582     the number of matched substrings. The substrings themselves are returned in
2583     \fIovector\fP. Each string uses two elements; the first is the offset to the
2584 nigel 93 start, and the second is the offset to the end. In fact, all the strings have
2585     the same start offset. (Space could have been saved by giving this only once,
2586     but it was decided to retain some compatibility with the way \fBpcre_exec()\fP
2587     returns data, even though the meaning of the strings is different.)
2588 nigel 77 .P
2589     The strings are returned in reverse order of length; that is, the longest
2590     matching string is given first. If there were too many matches to fit into
2591     \fIovector\fP, the yield of the function is zero, and the vector is filled with
2592 ph10 691 the longest matches. Unlike \fBpcre_exec()\fP, \fBpcre_dfa_exec()\fP can use
2593 ph10 683 the entire \fIovector\fP for returning matched strings.
2594 nigel 77 .
2595 ph10 598 .
2596 nigel 77 .SS "Error returns from \fBpcre_dfa_exec()\fP"
2597     .rs
2598     .sp
2599     The \fBpcre_dfa_exec()\fP function returns a negative number when it fails.
2600     Many of the errors are the same as for \fBpcre_exec()\fP, and these are
2601     described
2602     .\" HTML <a href="#errorlist">
2603     .\" </a>
2604     above.
2605     .\"
2606     There are in addition the following errors that are specific to
2607     \fBpcre_dfa_exec()\fP:
2608     .sp
2609     PCRE_ERROR_DFA_UITEM (-16)
2610     .sp
2611     This return is given if \fBpcre_dfa_exec()\fP encounters an item in the pattern
2612     that it does not support, for instance, the use of \eC or a back reference.
2613     .sp
2614     PCRE_ERROR_DFA_UCOND (-17)
2615     .sp
2616 nigel 93 This return is given if \fBpcre_dfa_exec()\fP encounters a condition item that
2617     uses a back reference for the condition, or a test for recursion in a specific
2618     group. These are not supported.
2619 nigel 77 .sp
2620     PCRE_ERROR_DFA_UMLIMIT (-18)
2621     .sp
2622     This return is given if \fBpcre_dfa_exec()\fP is called with an \fIextra\fP
2623 ph10 678 block that contains a setting of the \fImatch_limit\fP or
2624     \fImatch_limit_recursion\fP fields. This is not supported (these fields are
2625     meaningless for DFA matching).
2626 nigel 77 .sp
2627     PCRE_ERROR_DFA_WSSIZE (-19)
2628     .sp
2629     This return is given if \fBpcre_dfa_exec()\fP runs out of space in the
2630     \fIworkspace\fP vector.
2631     .sp
2632     PCRE_ERROR_DFA_RECURSE (-20)
2633     .sp
2634     When a recursive subpattern is processed, the matching function calls itself
2635     recursively, using private vectors for \fIovector\fP and \fIworkspace\fP. This
2636     error is given if the output vector is not large enough. This should be
2637     extremely rare, as a vector of size 1000 is used.
2638 ph10 960 .sp
2639     PCRE_ERROR_DFA_BADRESTART (-30)
2640     .sp
2641     When \fBpcre_dfa_exec()\fP is called with the \fBPCRE_DFA_RESTART\fP option,
2642     some plausibility checks are made on the contents of the workspace, which
2643     should contain data about the previous partial match. If any of these checks
2644     fail, this error is given.
2645 nigel 93 .
2646     .
2647     .SH "SEE ALSO"
2648     .rs
2649     .sp
2650 ph10 856 \fBpcre16\fP(3), \fBpcrebuild\fP(3), \fBpcrecallout\fP(3), \fBpcrecpp(3)\fP(3),
2651 nigel 93 \fBpcrematching\fP(3), \fBpcrepartial\fP(3), \fBpcreposix\fP(3),
2652     \fBpcreprecompile\fP(3), \fBpcresample\fP(3), \fBpcrestack\fP(3).
2653 ph10 99 .
2654     .
2655     .SH AUTHOR
2656     .rs
2657     .sp
2658     .nf
2659     Philip Hazel
2660     University Computing Service
2661     Cambridge CB2 3QH, England.
2662     .fi
2663     .
2664     .
2665     .SH REVISION
2666     .rs
2667     .sp
2668     .nf
2669 ph10 964 Last updated: 04 May 2012
2670 ph10 856 Copyright (c) 1997-2012 University of Cambridge.
2671 ph10 99 .fi

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12