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

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

Parent Directory Parent Directory | Revision Log Revision Log


Revision 887 - (hide annotations) (download)
Tue Jan 17 14:32:32 2012 UTC (2 years, 6 months ago) by ph10
File size: 113740 byte(s)
Document PCRE_CONFIG_JITTARGET

1 nigel 79 .TH PCREAPI 3
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     option names define the same bit values.
152     .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     page.
161     .
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     mentioned, plus the single characters VT (vertical tab, U+000B), FF (formfeed,
306     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 856 otherwise it is set to zero. If this option is given to the 16-bit version of
396     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     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     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     PCRE_NO_START_OPT options can be set at the time of matching as well as at
530     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 nigel 63 If this bit is set, whitespace data characters in the pattern are totally
646     ignored except when escaped or inside a character class. Whitespace does not
647     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     Note, however, that this applies only to data characters. Whitespace characters
665     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     tab, U+000B), FF (formfeed, U+000C), NEL (next line, U+0085), LS (line
745 ph10 856 separator, U+2028), and PS (paragraph separator, U+2029). For the 8-bit
746     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     compiling is when PCRE_EXTENDED is set. CR and LF are whitespace characters,
757     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 856 When PCRE_UTF8 is set, the validity of the pattern as a UTF-8
823     string is automatically checked. There is a discussion about the
824     .\" HTML <a href="pcreunicode.html#utf8strings">
825 ph10 211 .\" </a>
826 ph10 856 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 856 both compiling functions. Note that error messages are always 8-bit ASCII
847     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 856 67 this version of PCRE is not compiled with Unicode property
921     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     73 disallowed Unicode code point (>= 0xd800 && <= 0xdfff)
928     74 invalid UTF-16 string (specifically UTF-16)
929 ph10 290 .sp
930 ph10 292 The numbers 32 and 10000 in errors 48 and 49 are defaults; different values may
931 ph10 290 be used if the limits were changed when PCRE was built.
932 nigel 77 .
933     .
934 ph10 656 .\" HTML <a name="studyingapattern"></a>
935 nigel 75 .SH "STUDYING A PATTERN"
936 nigel 63 .rs
937     .sp
938 nigel 77 .B pcre_extra *pcre_study(const pcre *\fIcode\fP, int \fIoptions\fP
939 nigel 63 .ti +5n
940 nigel 75 .B const char **\fIerrptr\fP);
941 nigel 63 .PP
942 nigel 75 If a compiled pattern is going to be used several times, it is worth spending
943     more time analyzing it in order to speed up the time taken for matching. The
944     function \fBpcre_study()\fP takes a pointer to a compiled pattern as its first
945     argument. If studying the pattern produces additional information that will
946     help speed up matching, \fBpcre_study()\fP returns a pointer to a
947     \fBpcre_extra\fP block, in which the \fIstudy_data\fP field points to the
948     results of the study.
949     .P
950     The returned value from \fBpcre_study()\fP can be passed directly to
951 ph10 455 \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP. However, a \fBpcre_extra\fP block
952     also contains other fields that can be set by the caller before the block is
953     passed; these are described
954 nigel 75 .\" HTML <a href="#extradata">
955     .\" </a>
956     below
957     .\"
958     in the section on matching a pattern.
959     .P
960 ph10 455 If studying the pattern does not produce any useful information,
961 nigel 75 \fBpcre_study()\fP returns NULL. In that circumstance, if the calling program
962 ph10 461 wants to pass any of the other fields to \fBpcre_exec()\fP or
963 ph10 455 \fBpcre_dfa_exec()\fP, it must set up its own \fBpcre_extra\fP block.
964 nigel 75 .P
965 ph10 691 The second argument of \fBpcre_study()\fP contains option bits. There is only
966     one option: PCRE_STUDY_JIT_COMPILE. If this is set, and the just-in-time
967     compiler is available, the pattern is further compiled into machine code that
968     executes much faster than the \fBpcre_exec()\fP matching function. If
969     the just-in-time compiler is not available, this option is ignored. All other
970 ph10 678 bits in the \fIoptions\fP argument must be zero.
971 nigel 75 .P
972 ph10 691 JIT compilation is a heavyweight optimization. It can take some time for
973 ph10 678 patterns to be analyzed, and for one-off matches and simple patterns the
974     benefit of faster execution might be offset by a much slower study time.
975 ph10 691 Not all patterns can be optimized by the JIT compiler. For those that cannot be
976 ph10 678 handled, matching automatically falls back to the \fBpcre_exec()\fP
977     interpreter. For more details, see the
978     .\" HREF
979     \fBpcrejit\fP
980     .\"
981     documentation.
982     .P
983 nigel 75 The third argument for \fBpcre_study()\fP is a pointer for an error message. If
984 nigel 63 studying succeeds (even if no data is returned), the variable it points to is
985 nigel 87 set to NULL. Otherwise it is set to point to a textual error message. This is a
986     static string that is part of the library. You must not try to free it. You
987     should test the error pointer for NULL after calling \fBpcre_study()\fP, to be
988     sure that it has run successfully.
989 nigel 75 .P
990 ph10 678 When you are finished with a pattern, you can free the memory used for the
991     study data by calling \fBpcre_free_study()\fP. This function was added to the
992     API for release 8.20. For earlier versions, the memory could be freed with
993     \fBpcre_free()\fP, just like the pattern itself. This will still work in cases
994     where PCRE_STUDY_JIT_COMPILE is not used, but it is advisable to change to the
995     new function when convenient.
996     .P
997 ph10 691 This is a typical way in which \fBpcre_study\fP() is used (except that in a
998 ph10 678 real application there should be tests for errors):
999 nigel 75 .sp
1000 ph10 678 int rc;
1001     pcre *re;
1002     pcre_extra *sd;
1003     re = pcre_compile("pattern", 0, &error, &erroroffset, NULL);
1004     sd = pcre_study(
1005 nigel 63 re, /* result of pcre_compile() */
1006 ph10 678 0, /* no options */
1007 nigel 63 &error); /* set to NULL or points to a message */
1008 ph10 678 rc = pcre_exec( /* see below for details of pcre_exec() options */
1009 ph10 691 re, sd, "subject", 7, 0, 0, ovector, 30);
1010 ph10 678 ...
1011     pcre_free_study(sd);
1012 ph10 691 pcre_free(re);
1013 nigel 75 .sp
1014 ph10 455 Studying a pattern does two things: first, a lower bound for the length of
1015 ph10 461 subject string that is needed to match the pattern is computed. This does not
1016     mean that there are any strings of that length that match, but it does
1017     guarantee that no shorter strings match. The value is used by
1018     \fBpcre_exec()\fP and \fBpcre_dfa_exec()\fP to avoid wasting time by trying to
1019     match strings that are shorter than the lower bound. You can find out the value
1020 ph10 455 in a calling program via the \fBpcre_fullinfo()\fP function.
1021     .P
1022     Studying a pattern is also useful for non-anchored patterns that do not have a
1023     single fixed starting character. A bitmap of possible starting bytes is
1024 ph10 461 created. This speeds up finding a position in the subject at which to start
1025 ph10 856 matching. (In 16-bit mode, the bitmap is used for 16-bit values less than 256.)
1026 ph10 547 .P
1027 ph10 691 These two optimizations apply to both \fBpcre_exec()\fP and
1028     \fBpcre_dfa_exec()\fP. However, they are not used by \fBpcre_exec()\fP if
1029 ph10 678 \fBpcre_study()\fP is called with the PCRE_STUDY_JIT_COMPILE option, and
1030     just-in-time compiling is successful. The optimizations can be disabled by
1031     setting the PCRE_NO_START_OPTIMIZE option when calling \fBpcre_exec()\fP or
1032 ph10 548 \fBpcre_dfa_exec()\fP. You might want to do this if your pattern contains
1033 ph10 678 callouts or (*MARK) (which cannot be handled by the JIT compiler), and you want
1034     to make use of these facilities in cases where matching fails. See the
1035     discussion of PCRE_NO_START_OPTIMIZE
1036 ph10 547 .\" HTML <a href="#execoptions">
1037     .\" </a>
1038     below.
1039     .\"
1040 nigel 75 .
1041     .
1042 nigel 63 .\" HTML <a name="localesupport"></a>
1043 nigel 75 .SH "LOCALE SUPPORT"
1044 nigel 63 .rs
1045     .sp
1046 ph10 139 PCRE handles caseless matching, and determines whether characters are letters,
1047 nigel 75 digits, or whatever, by reference to a set of tables, indexed by character
1048 ph10 856 value. When running in UTF-8 mode, this applies only to characters
1049     with codes less than 128. By default, higher-valued codes never match escapes
1050     such as \ew or \ed, but they can be tested with \ep if PCRE is built with
1051     Unicode character property support. Alternatively, the PCRE_UCP option can be
1052     set at compile time; this causes \ew and friends to use Unicode property
1053     support instead of built-in tables. The use of locales with Unicode is
1054     discouraged. If you are handling characters with codes greater than 128, you
1055     should either use UTF-8 and Unicode, or use locales, but not try to mix the
1056     two.
1057 nigel 75 .P
1058 ph10 139 PCRE contains an internal set of tables that are used when the final argument
1059     of \fBpcre_compile()\fP is NULL. These are sufficient for many applications.
1060 ph10 142 Normally, the internal tables recognize only ASCII characters. However, when
1061 ph10 139 PCRE is built, it is possible to cause the internal tables to be rebuilt in the
1062     default "C" locale of the local system, which may cause them to be different.
1063 nigel 75 .P
1064 ph10 139 The internal tables can always be overridden by tables supplied by the
1065     application that calls PCRE. These may be created in a different locale from
1066     the default. As more and more applications change to using Unicode, the need
1067     for this locale support is expected to die away.
1068     .P
1069 nigel 75 External tables are built by calling the \fBpcre_maketables()\fP function,
1070     which has no arguments, in the relevant locale. The result can then be passed
1071     to \fBpcre_compile()\fP or \fBpcre_exec()\fP as often as necessary. For
1072     example, to build and use tables that are appropriate for the French locale
1073     (where accented characters with values greater than 128 are treated as letters),
1074     the following code could be used:
1075     .sp
1076     setlocale(LC_CTYPE, "fr_FR");
1077 nigel 63 tables = pcre_maketables();
1078     re = pcre_compile(..., tables);
1079 nigel 75 .sp
1080 ph10 142 The locale name "fr_FR" is used on Linux and other Unix-like systems; if you
1081 ph10 139 are using Windows, the name for the French locale is "french".
1082     .P
1083 nigel 75 When \fBpcre_maketables()\fP runs, the tables are built in memory that is
1084     obtained via \fBpcre_malloc\fP. It is the caller's responsibility to ensure
1085     that the memory containing the tables remains available for as long as it is
1086     needed.
1087     .P
1088     The pointer that is passed to \fBpcre_compile()\fP is saved with the compiled
1089     pattern, and the same tables are used via this pointer by \fBpcre_study()\fP
1090     and normally also by \fBpcre_exec()\fP. Thus, by default, for any single
1091     pattern, compilation, studying and matching all happen in the same locale, but
1092     different patterns can be compiled in different locales.
1093     .P
1094     It is possible to pass a table pointer or NULL (indicating the use of the
1095     internal tables) to \fBpcre_exec()\fP. Although not intended for this purpose,
1096     this facility could be used to match a pattern in a different locale from the
1097     one in which it was compiled. Passing table pointers at run time is discussed
1098     below in the section on matching a pattern.
1099     .
1100     .
1101 ph10 598 .\" HTML <a name="infoaboutpattern"></a>
1102 nigel 75 .SH "INFORMATION ABOUT A PATTERN"
1103 nigel 63 .rs
1104     .sp
1105 nigel 75 .B int pcre_fullinfo(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
1106 nigel 63 .ti +5n
1107 nigel 75 .B int \fIwhat\fP, void *\fIwhere\fP);
1108 nigel 63 .PP
1109 nigel 75 The \fBpcre_fullinfo()\fP function returns information about a compiled
1110 ph10 856 pattern. It replaces the \fBpcre_info()\fP function, which was removed from the
1111     library at version 8.30, after more than 10 years of obsolescence.
1112 nigel 75 .P
1113     The first argument for \fBpcre_fullinfo()\fP is a pointer to the compiled
1114     pattern. The second argument is the result of \fBpcre_study()\fP, or NULL if
1115 nigel 63 the pattern was not studied. The third argument specifies which piece of
1116     information is required, and the fourth argument is a pointer to a variable
1117     to receive the data. The yield of the function is zero for success, or one of
1118     the following negative numbers:
1119 nigel 75 .sp
1120 ph10 856 PCRE_ERROR_NULL the argument \fIcode\fP was NULL
1121     the argument \fIwhere\fP was NULL
1122     PCRE_ERROR_BADMAGIC the "magic number" was not found
1123     PCRE_ERROR_BADENDIANNESS the pattern was compiled with different
1124     endianness
1125     PCRE_ERROR_BADOPTION the value of \fIwhat\fP was invalid
1126 nigel 75 .sp
1127     The "magic number" is placed at the start of each compiled pattern as an simple
1128 ph10 856 check against passing an arbitrary memory pointer. The endianness error can
1129     occur if a compiled pattern is saved and reloaded on a different host. Here is
1130     a typical call of \fBpcre_fullinfo()\fP, to obtain the length of the compiled
1131     pattern:
1132 nigel 75 .sp
1133 nigel 63 int rc;
1134 nigel 91 size_t length;
1135 nigel 63 rc = pcre_fullinfo(
1136     re, /* result of pcre_compile() */
1137 ph10 678 sd, /* result of pcre_study(), or NULL */
1138 nigel 63 PCRE_INFO_SIZE, /* what is required */
1139     &length); /* where to put the data */
1140 nigel 75 .sp
1141     The possible values for the third argument are defined in \fBpcre.h\fP, and are
1142 nigel 63 as follows:
1143 nigel 75 .sp
1144 nigel 63 PCRE_INFO_BACKREFMAX
1145 nigel 75 .sp
1146 nigel 63 Return the number of the highest back reference in the pattern. The fourth
1147 nigel 75 argument should point to an \fBint\fP variable. Zero is returned if there are
1148 nigel 63 no back references.
1149 nigel 75 .sp
1150 nigel 63 PCRE_INFO_CAPTURECOUNT
1151 nigel 75 .sp
1152 nigel 63 Return the number of capturing subpatterns in the pattern. The fourth argument
1153 nigel 75 should point to an \fBint\fP variable.
1154     .sp
1155 nigel 77 PCRE_INFO_DEFAULT_TABLES
1156 nigel 75 .sp
1157     Return a pointer to the internal default character tables within PCRE. The
1158     fourth argument should point to an \fBunsigned char *\fP variable. This
1159     information call is provided for internal use by the \fBpcre_study()\fP
1160     function. External callers can cause PCRE to use its internal tables by passing
1161     a NULL table pointer.
1162     .sp
1163 nigel 63 PCRE_INFO_FIRSTBYTE
1164 nigel 75 .sp
1165 ph10 856 Return information about the first data unit of any matched string, for a
1166     non-anchored pattern. (The name of this option refers to the 8-bit library,
1167     where data units are bytes.) The fourth argument should point to an \fBint\fP
1168     variable.
1169 nigel 75 .P
1170 ph10 856 If there is a fixed first value, for example, the letter "c" from a pattern
1171     such as (cat|cow|coyote), its value is returned. In the 8-bit library, the
1172     value is always less than 256; in the 16-bit library the value can be up to
1173     0xffff.
1174     .P
1175     If there is no fixed first value, and if either
1176 nigel 75 .sp
1177 nigel 63 (a) the pattern was compiled with the PCRE_MULTILINE option, and every branch
1178     starts with "^", or
1179 nigel 75 .sp
1180 nigel 63 (b) every branch of the pattern starts with ".*" and PCRE_DOTALL is not set
1181     (if it were set, the pattern would be anchored),
1182 nigel 75 .sp
1183 nigel 63 -1 is returned, indicating that the pattern matches only at the start of a
1184     subject string or after any newline within the string. Otherwise -2 is
1185     returned. For anchored patterns, -2 is returned.
1186 nigel 75 .sp
1187 nigel 63 PCRE_INFO_FIRSTTABLE
1188 nigel 75 .sp
1189 nigel 63 If the pattern was studied, and this resulted in the construction of a 256-bit
1190 ph10 856 table indicating a fixed set of values for the first data unit in any matching
1191 nigel 63 string, a pointer to the table is returned. Otherwise NULL is returned. The
1192 nigel 75 fourth argument should point to an \fBunsigned char *\fP variable.
1193     .sp
1194 ph10 226 PCRE_INFO_HASCRORLF
1195     .sp
1196 ph10 227 Return 1 if the pattern contains any explicit matches for CR or LF characters,
1197 ph10 243 otherwise 0. The fourth argument should point to an \fBint\fP variable. An
1198 ph10 231 explicit match is either a literal CR or LF character, or \er or \en.
1199 ph10 226 .sp
1200 ph10 169 PCRE_INFO_JCHANGED
1201     .sp
1202 ph10 278 Return 1 if the (?J) or (?-J) option setting is used in the pattern, otherwise
1203     0. The fourth argument should point to an \fBint\fP variable. (?J) and
1204     (?-J) set and unset the local PCRE_DUPNAMES option, respectively.
1205 ph10 169 .sp
1206 ph10 678 PCRE_INFO_JIT
1207     .sp
1208 ph10 691 Return 1 if the pattern was studied with the PCRE_STUDY_JIT_COMPILE option, and
1209 ph10 678 just-in-time compiling was successful. The fourth argument should point to an
1210     \fBint\fP variable. A return value of 0 means that JIT support is not available
1211 ph10 691 in this version of PCRE, or that the pattern was not studied with the
1212     PCRE_STUDY_JIT_COMPILE option, or that the JIT compiler could not handle this
1213 ph10 678 particular pattern. See the
1214     .\" HREF
1215     \fBpcrejit\fP
1216     .\"
1217     documentation for details of what can and cannot be handled.
1218     .sp
1219 ph10 836 PCRE_INFO_JITSIZE
1220     .sp
1221     If the pattern was successfully studied with the PCRE_STUDY_JIT_COMPILE option,
1222     return the size of the JIT compiled code, otherwise return zero. The fourth
1223     argument should point to a \fBsize_t\fP variable.
1224     .sp
1225 nigel 63 PCRE_INFO_LASTLITERAL
1226 nigel 75 .sp
1227 ph10 856 Return the value of the rightmost literal data unit that must exist in any
1228     matched string, other than at its start, if such a value has been recorded. The
1229     fourth argument should point to an \fBint\fP variable. If there is no such
1230     value, -1 is returned. For anchored patterns, a last literal value is recorded
1231     only if it follows something of variable length. For example, for the pattern
1232 nigel 75 /^a\ed+z\ed+/ the returned value is "z", but for /^a\edz\ed/ the returned value
1233 nigel 65 is -1.
1234 nigel 75 .sp
1235 ph10 455 PCRE_INFO_MINLENGTH
1236     .sp
1237     If the pattern was studied and a minimum length for matching subject strings
1238     was computed, its value is returned. Otherwise the returned value is -1. The
1239 ph10 856 value is a number of characters, which in UTF-8 mode may be different from the
1240     number of bytes. The fourth argument should point to an \fBint\fP variable. A
1241 ph10 455 non-negative value is a lower bound to the length of any matching string. There
1242     may not be any strings of that length that do actually match, but every string
1243     that does match is at least that long.
1244     .sp
1245 nigel 63 PCRE_INFO_NAMECOUNT
1246     PCRE_INFO_NAMEENTRYSIZE
1247     PCRE_INFO_NAMETABLE
1248 nigel 75 .sp
1249 nigel 63 PCRE supports the use of named as well as numbered capturing parentheses. The
1250     names are just an additional way of identifying the parentheses, which still
1251 nigel 91 acquire numbers. Several convenience functions such as
1252     \fBpcre_get_named_substring()\fP are provided for extracting captured
1253     substrings by name. It is also possible to extract the data directly, by first
1254     converting the name to a number in order to access the correct pointers in the
1255     output vector (described with \fBpcre_exec()\fP below). To do the conversion,
1256     you need to use the name-to-number map, which is described by these three
1257     values.
1258 nigel 75 .P
1259 nigel 63 The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT gives
1260     the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size of each
1261 nigel 75 entry; both of these return an \fBint\fP value. The entry size depends on the
1262 nigel 63 length of the longest name. PCRE_INFO_NAMETABLE returns a pointer to the first
1263 ph10 856 entry of the table. This is a pointer to \fBchar\fP in the 8-bit library, where
1264     the first two bytes of each entry are the number of the capturing parenthesis,
1265     most significant byte first. In the 16-bit library, the pointer points to
1266     16-bit data units, the first of which contains the parenthesis number. The rest
1267     of the entry is the corresponding name, zero terminated.
1268 ph10 457 .P
1269     The names are in alphabetical order. Duplicate names may appear if (?| is used
1270     to create multiple groups with the same number, as described in the
1271     .\" HTML <a href="pcrepattern.html#dupsubpatternnumber">
1272     .\" </a>
1273     section on duplicate subpattern numbers
1274     .\"
1275     in the
1276     .\" HREF
1277     \fBpcrepattern\fP
1278     .\"
1279 ph10 461 page. Duplicate names for subpatterns with different numbers are permitted only
1280     if PCRE_DUPNAMES is set. In all cases of duplicate names, they appear in the
1281     table in the order in which they were found in the pattern. In the absence of
1282     (?| this is the order of increasing number; when (?| is used this is not
1283 ph10 457 necessarily the case because later subpatterns may have lower numbers.
1284     .P
1285     As a simple example of the name/number table, consider the following pattern
1286 ph10 856 after compilation by the 8-bit library (assume PCRE_EXTENDED is set, so white
1287     space - including newlines - is ignored):
1288 nigel 75 .sp
1289     .\" JOIN
1290 nigel 93 (?<date> (?<year>(\ed\ed)?\ed\ed) -
1291     (?<month>\ed\ed) - (?<day>\ed\ed) )
1292 nigel 75 .sp
1293 nigel 63 There are four named subpatterns, so the table has four entries, and each entry
1294     in the table is eight bytes long. The table is as follows, with non-printing
1295 nigel 75 bytes shows in hexadecimal, and undefined bytes shown as ??:
1296     .sp
1297 nigel 63 00 01 d a t e 00 ??
1298     00 05 d a y 00 ?? ??
1299     00 04 m o n t h 00
1300     00 02 y e a r 00 ??
1301 nigel 75 .sp
1302     When writing code to extract data from named subpatterns using the
1303 nigel 91 name-to-number map, remember that the length of the entries is likely to be
1304 nigel 75 different for each compiled pattern.
1305     .sp
1306 ph10 169 PCRE_INFO_OKPARTIAL
1307     .sp
1308 ph10 435 Return 1 if the pattern can be used for partial matching with
1309     \fBpcre_exec()\fP, otherwise 0. The fourth argument should point to an
1310     \fBint\fP variable. From release 8.00, this always returns 1, because the
1311     restrictions that previously applied to partial matching have been lifted. The
1312 ph10 169 .\" HREF
1313     \fBpcrepartial\fP
1314     .\"
1315 ph10 426 documentation gives details of partial matching.
1316 ph10 169 .sp
1317 nigel 63 PCRE_INFO_OPTIONS
1318 nigel 75 .sp
1319 nigel 63 Return a copy of the options with which the pattern was compiled. The fourth
1320 nigel 75 argument should point to an \fBunsigned long int\fP variable. These option bits
1321     are those specified in the call to \fBpcre_compile()\fP, modified by any
1322 ph10 196 top-level option settings at the start of the pattern itself. In other words,
1323     they are the options that will be in force when matching starts. For example,
1324     if the pattern /(?im)abc(?-i)d/ is compiled with the PCRE_EXTENDED option, the
1325     result is PCRE_CASELESS, PCRE_MULTILINE, and PCRE_EXTENDED.
1326 nigel 75 .P
1327 nigel 63 A pattern is automatically anchored by PCRE if all of its top-level
1328     alternatives begin with one of the following:
1329 nigel 75 .sp
1330 nigel 63 ^ unless PCRE_MULTILINE is set
1331 nigel 75 \eA always
1332     \eG always
1333     .\" JOIN
1334 nigel 63 .* if PCRE_DOTALL is set and there are no back
1335     references to the subpattern in which .* appears
1336 nigel 75 .sp
1337 nigel 63 For such patterns, the PCRE_ANCHORED bit is set in the options returned by
1338 nigel 75 \fBpcre_fullinfo()\fP.
1339     .sp
1340 nigel 63 PCRE_INFO_SIZE
1341 nigel 75 .sp
1342 ph10 856 Return the size of the compiled pattern in bytes (for both libraries). The
1343     fourth argument should point to a \fBsize_t\fP variable. This value does not
1344     include the size of the \fBpcre\fP structure that is returned by
1345     \fBpcre_compile()\fP. The value that is passed as the argument to
1346     \fBpcre_malloc()\fP when \fBpcre_compile()\fP is getting memory in which to
1347     place the compiled data is the value returned by this option plus the size of
1348     the \fBpcre\fP structure. Studying a compiled pattern, with or without JIT,
1349     does not alter the value returned by this option.
1350 nigel 75 .sp
1351 nigel 63 PCRE_INFO_STUDYSIZE
1352 nigel 75 .sp
1353 ph10 856 Return the size in bytes of the data block pointed to by the \fIstudy_data\fP
1354     field in a \fBpcre_extra\fP block. If \fBpcre_extra\fP is NULL, or there is no
1355     study data, zero is returned. The fourth argument should point to a
1356     \fBsize_t\fP variable. The \fIstudy_data\fP field is set by \fBpcre_study()\fP
1357     to record information that will speed up matching (see the section entitled
1358 ph10 656 .\" HTML <a href="#studyingapattern">
1359     .\" </a>
1360 ph10 659 "Studying a pattern"
1361 ph10 656 .\"
1362     above). The format of the \fIstudy_data\fP block is private, but its length
1363     is made available via this option so that it can be saved and restored (see the
1364     .\" HREF
1365     \fBpcreprecompile\fP
1366     .\"
1367     documentation for details).
1368 nigel 75 .
1369     .
1370 nigel 77 .SH "REFERENCE COUNTS"
1371 nigel 63 .rs
1372     .sp
1373 nigel 77 .B int pcre_refcount(pcre *\fIcode\fP, int \fIadjust\fP);
1374     .PP
1375     The \fBpcre_refcount()\fP function is used to maintain a reference count in the
1376     data block that contains a compiled pattern. It is provided for the benefit of
1377     applications that operate in an object-oriented manner, where different parts
1378     of the application may be using the same compiled pattern, but you want to free
1379     the block when they are all done.
1380     .P
1381     When a pattern is compiled, the reference count field is initialized to zero.
1382     It is changed only by calling this function, whose action is to add the
1383     \fIadjust\fP value (which may be positive or negative) to it. The yield of the
1384     function is the new value. However, the value of the count is constrained to
1385     lie between 0 and 65535, inclusive. If the new value is outside these limits,
1386     it is forced to the appropriate limit value.
1387     .P
1388     Except when it is zero, the reference count is not correctly preserved if a
1389     pattern is compiled on one host and then transferred to a host whose byte-order
1390     is different. (This seems a highly unlikely scenario.)
1391     .
1392     .
1393     .SH "MATCHING A PATTERN: THE TRADITIONAL FUNCTION"
1394     .rs
1395     .sp
1396 nigel 75 .B int pcre_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
1397 nigel 63 .ti +5n
1398 nigel 75 .B "const char *\fIsubject\fP," int \fIlength\fP, int \fIstartoffset\fP,
1399 nigel 63 .ti +5n
1400 nigel 75 .B int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP);
1401     .P
1402     The function \fBpcre_exec()\fP is called to match a subject string against a
1403     compiled pattern, which is passed in the \fIcode\fP argument. If the
1404 ph10 455 pattern was studied, the result of the study should be passed in the
1405 ph10 707 \fIextra\fP argument. You can call \fBpcre_exec()\fP with the same \fIcode\fP
1406     and \fIextra\fP arguments as many times as you like, in order to match
1407     different subject strings with the same pattern.
1408     .P
1409     This function is the main matching facility of the library, and it operates in
1410     a Perl-like manner. For specialist use there is also an alternative matching
1411     function, which is described
1412 nigel 77 .\" HTML <a href="#dfamatch">
1413     .\" </a>
1414     below
1415     .\"
1416     in the section about the \fBpcre_dfa_exec()\fP function.
1417 nigel 75 .P
1418     In most applications, the pattern will have been compiled (and optionally
1419     studied) in the same process that calls \fBpcre_exec()\fP. However, it is
1420     possible to save compiled patterns and study data, and then use them later
1421     in different processes, possibly even on different hosts. For a discussion
1422     about this, see the
1423     .\" HREF
1424     \fBpcreprecompile\fP
1425     .\"
1426     documentation.
1427     .P
1428     Here is an example of a simple call to \fBpcre_exec()\fP:
1429     .sp
1430 nigel 63 int rc;
1431     int ovector[30];
1432     rc = pcre_exec(
1433     re, /* result of pcre_compile() */
1434     NULL, /* we didn't study the pattern */
1435     "some string", /* the subject string */
1436     11, /* the length of the subject string */
1437     0, /* start at offset 0 in the subject */
1438     0, /* default options */
1439 nigel 75 ovector, /* vector of integers for substring information */
1440 nigel 77 30); /* number of elements (NOT size in bytes) */
1441 nigel 75 .
1442 ph10 656 .
1443 nigel 75 .\" HTML <a name="extradata"></a>
1444     .SS "Extra data for \fBpcre_exec()\fR"
1445     .rs
1446     .sp
1447     If the \fIextra\fP argument is not NULL, it must point to a \fBpcre_extra\fP
1448     data block. The \fBpcre_study()\fP function returns such a block (when it
1449 nigel 63 doesn't return NULL), but you can also create one for yourself, and pass
1450 nigel 87 additional information in it. The \fBpcre_extra\fP block contains the following
1451     fields (not necessarily in this order):
1452 nigel 75 .sp
1453     unsigned long int \fIflags\fP;
1454     void *\fIstudy_data\fP;
1455 ph10 691 void *\fIexecutable_jit\fP;
1456 nigel 75 unsigned long int \fImatch_limit\fP;
1457 nigel 87 unsigned long int \fImatch_limit_recursion\fP;
1458 nigel 75 void *\fIcallout_data\fP;
1459     const unsigned char *\fItables\fP;
1460 ph10 512 unsigned char **\fImark\fP;
1461 nigel 75 .sp
1462 ph10 856 In the 16-bit version of this structure, the \fImark\fP field has type
1463 ph10 859 "PCRE_UCHAR16 **".
1464 ph10 856 .P
1465 nigel 75 The \fIflags\fP field is a bitmap that specifies which of the other fields
1466 nigel 63 are set. The flag bits are:
1467 nigel 75 .sp
1468 nigel 63 PCRE_EXTRA_STUDY_DATA
1469 ph10 691 PCRE_EXTRA_EXECUTABLE_JIT
1470 nigel 63 PCRE_EXTRA_MATCH_LIMIT
1471 nigel 87 PCRE_EXTRA_MATCH_LIMIT_RECURSION
1472 nigel 63 PCRE_EXTRA_CALLOUT_DATA
1473 nigel 75 PCRE_EXTRA_TABLES
1474 ph10 512 PCRE_EXTRA_MARK
1475 nigel 75 .sp
1476 ph10 678 Other flag bits should be set to zero. The \fIstudy_data\fP field and sometimes
1477     the \fIexecutable_jit\fP field are set in the \fBpcre_extra\fP block that is
1478     returned by \fBpcre_study()\fP, together with the appropriate flag bits. You
1479     should not set these yourself, but you may add to the block by setting the
1480     other fields and their corresponding flag bits.
1481 nigel 75 .P
1482     The \fImatch_limit\fP field provides a means of preventing PCRE from using up a
1483 nigel 63 vast amount of resources when running patterns that are not going to match,
1484     but which have a very large number of possibilities in their search trees. The
1485 ph10 456 classic example is a pattern that uses nested unlimited repeats.
1486 nigel 75 .P
1487 ph10 678 Internally, \fBpcre_exec()\fP uses a function called \fBmatch()\fP, which it
1488     calls repeatedly (sometimes recursively). The limit set by \fImatch_limit\fP is
1489     imposed on the number of times this function is called during a match, which
1490     has the effect of limiting the amount of backtracking that can take place. For
1491     patterns that are not anchored, the count restarts from zero for each position
1492 ph10 691 in the subject string.
1493 nigel 75 .P
1494 ph10 691 When \fBpcre_exec()\fP is called with a pattern that was successfully studied
1495     with the PCRE_STUDY_JIT_COMPILE option, the way that the matching is executed
1496     is entirely different. However, there is still the possibility of runaway
1497 ph10 678 matching that goes on for a very long time, and so the \fImatch_limit\fP value
1498     is also used in this case (but in a different way) to limit how long the
1499     matching can continue.
1500     .P
1501 nigel 87 The default value for the limit can be set when PCRE is built; the default
1502 nigel 63 default is 10 million, which handles all but the most extreme cases. You can
1503 nigel 87 override the default by suppling \fBpcre_exec()\fP with a \fBpcre_extra\fP
1504     block in which \fImatch_limit\fP is set, and PCRE_EXTRA_MATCH_LIMIT is set in
1505     the \fIflags\fP field. If the limit is exceeded, \fBpcre_exec()\fP returns
1506     PCRE_ERROR_MATCHLIMIT.
1507 nigel 75 .P
1508 nigel 87 The \fImatch_limit_recursion\fP field is similar to \fImatch_limit\fP, but
1509     instead of limiting the total number of times that \fBmatch()\fP is called, it
1510     limits the depth of recursion. The recursion depth is a smaller number than the
1511     total number of calls, because not all calls to \fBmatch()\fP are recursive.
1512 ph10 691 This limit is of use only if it is set smaller than \fImatch_limit\fP.
1513 nigel 87 .P
1514 ph10 678 Limiting the recursion depth limits the amount of machine stack that can be
1515     used, or, when PCRE has been compiled to use memory on the heap instead of the
1516 ph10 686 stack, the amount of heap memory that can be used. This limit is not relevant,
1517     and is ignored, if the pattern was successfully studied with
1518 ph10 678 PCRE_STUDY_JIT_COMPILE.
1519 nigel 87 .P
1520     The default value for \fImatch_limit_recursion\fP can be set when PCRE is
1521     built; the default default is the same value as the default for
1522     \fImatch_limit\fP. You can override the default by suppling \fBpcre_exec()\fP
1523     with a \fBpcre_extra\fP block in which \fImatch_limit_recursion\fP is set, and
1524     PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the \fIflags\fP field. If the limit
1525     is exceeded, \fBpcre_exec()\fP returns PCRE_ERROR_RECURSIONLIMIT.
1526     .P
1527 ph10 440 The \fIcallout_data\fP field is used in conjunction with the "callout" feature,
1528     and is described in the
1529 nigel 75 .\" HREF
1530     \fBpcrecallout\fP
1531     .\"
1532     documentation.
1533     .P
1534     The \fItables\fP field is used to pass a character tables pointer to
1535     \fBpcre_exec()\fP; this overrides the value that is stored with the compiled
1536     pattern. A non-NULL value is stored with the compiled pattern only if custom
1537     tables were supplied to \fBpcre_compile()\fP via its \fItableptr\fP argument.
1538     If NULL is passed to \fBpcre_exec()\fP using this mechanism, it forces PCRE's
1539     internal tables to be used. This facility is helpful when re-using patterns
1540     that have been saved after compiling with an external set of tables, because
1541     the external tables might be at a different address when \fBpcre_exec()\fP is
1542     called. See the
1543     .\" HREF
1544     \fBpcreprecompile\fP
1545     .\"
1546     documentation for a discussion of saving compiled patterns for later use.
1547 ph10 510 .P
1548     If PCRE_EXTRA_MARK is set in the \fIflags\fP field, the \fImark\fP field must
1549 ph10 856 be set to point to a suitable variable. If the pattern contains any
1550 ph10 510 backtracking control verbs such as (*MARK:NAME), and the execution ends up with
1551     a name to pass back, a pointer to the name string (zero terminated) is placed
1552     in the variable pointed to by the \fImark\fP field. The names are within the
1553     compiled pattern; if you wish to retain such a name you must copy it before
1554     freeing the memory of a compiled pattern. If there is no name to pass back, the
1555 ph10 856 variable pointed to by the \fImark\fP field is set to NULL. For details of the
1556 ph10 510 backtracking control verbs, see the section entitled
1557     .\" HTML <a href="pcrepattern#backtrackcontrol">
1558     .\" </a>
1559     "Backtracking control"
1560     .\"
1561     in the
1562     .\" HREF
1563     \fBpcrepattern\fP
1564     .\"
1565     documentation.
1566 nigel 75 .
1567 ph10 510 .
1568 ph10 226 .\" HTML <a name="execoptions"></a>
1569 nigel 75 .SS "Option bits for \fBpcre_exec()\fP"
1570     .rs
1571     .sp
1572     The unused bits of the \fIoptions\fP argument for \fBpcre_exec()\fP must be
1573 nigel 91 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_\fIxxx\fP,
1574 ph10 442 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
1575     PCRE_NO_START_OPTIMIZE, PCRE_NO_UTF8_CHECK, PCRE_PARTIAL_SOFT, and
1576     PCRE_PARTIAL_HARD.
1577 ph10 686 .P
1578     If the pattern was successfully studied with the PCRE_STUDY_JIT_COMPILE option,
1579     the only supported options for JIT execution are PCRE_NO_UTF8_CHECK,
1580     PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, and PCRE_NOTEMPTY_ATSTART. Note in
1581 ph10 691 particular that partial matching is not supported. If an unsupported option is
1582     used, JIT execution is disabled and the normal interpretive code in
1583 ph10 686 \fBpcre_exec()\fP is run.
1584 nigel 75 .sp
1585     PCRE_ANCHORED
1586     .sp
1587     The PCRE_ANCHORED option limits \fBpcre_exec()\fP to matching at the first
1588     matching position. If a pattern was compiled with PCRE_ANCHORED, or turned out
1589     to be anchored by virtue of its contents, it cannot be made unachored at
1590     matching time.
1591     .sp
1592 ph10 231 PCRE_BSR_ANYCRLF
1593     PCRE_BSR_UNICODE
1594     .sp
1595     These options (which are mutually exclusive) control what the \eR escape
1596     sequence matches. The choice is either to match only CR, LF, or CRLF, or to
1597     match any Unicode newline sequence. These options override the choice that was
1598     made or defaulted when the pattern was compiled.
1599     .sp
1600 nigel 91 PCRE_NEWLINE_CR
1601     PCRE_NEWLINE_LF
1602     PCRE_NEWLINE_CRLF
1603 ph10 150 PCRE_NEWLINE_ANYCRLF
1604 nigel 93 PCRE_NEWLINE_ANY
1605 nigel 91 .sp
1606     These options override the newline definition that was chosen or defaulted when
1607 nigel 93 the pattern was compiled. For details, see the description of
1608     \fBpcre_compile()\fP above. During matching, the newline choice affects the
1609     behaviour of the dot, circumflex, and dollar metacharacters. It may also alter
1610     the way the match position is advanced after a match failure for an unanchored
1611 ph10 227 pattern.
1612 ph10 225 .P
1613 ph10 226 When PCRE_NEWLINE_CRLF, PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is set, and a
1614     match attempt for an unanchored pattern fails when the current position is at a
1615 ph10 230 CRLF sequence, and the pattern contains no explicit matches for CR or LF
1616 ph10 226 characters, the match position is advanced by two characters instead of one, in
1617     other words, to after the CRLF.
1618     .P
1619 ph10 227 The above rule is a compromise that makes the most common cases work as
1620     expected. For example, if the pattern is .+A (and the PCRE_DOTALL option is not
1621     set), it does not match the string "\er\enA" because, after failing at the
1622     start, it skips both the CR and the LF before retrying. However, the pattern
1623     [\er\en]A does match that string, because it contains an explicit CR or LF
1624 ph10 226 reference, and so advances only by one character after the first failure.
1625     .P
1626 ph10 231 An explicit match for CR of LF is either a literal appearance of one of those
1627     characters, or one of the \er or \en escape sequences. Implicit matches such as
1628 ph10 230 [^X] do not count, nor does \es (which includes CR and LF in the characters
1629     that it matches).
1630     .P
1631 ph10 226 Notwithstanding the above, anomalous effects may still occur when CRLF is a
1632     valid newline sequence and explicit \er or \en escapes appear in the pattern.
1633 nigel 91 .sp
1634 nigel 63 PCRE_NOTBOL
1635 nigel 75 .sp
1636     This option specifies that first character of the subject string is not the
1637     beginning of a line, so the circumflex metacharacter should not match before
1638     it. Setting this without PCRE_MULTILINE (at compile time) causes circumflex
1639     never to match. This option affects only the behaviour of the circumflex
1640     metacharacter. It does not affect \eA.
1641     .sp
1642 nigel 63 PCRE_NOTEOL
1643 nigel 75 .sp
1644     This option specifies that the end of the subject string is not the end of a
1645     line, so the dollar metacharacter should not match it nor (except in multiline
1646     mode) a newline immediately before it. Setting this without PCRE_MULTILINE (at
1647     compile time) causes dollar never to match. This option affects only the
1648     behaviour of the dollar metacharacter. It does not affect \eZ or \ez.
1649     .sp
1650 nigel 63 PCRE_NOTEMPTY
1651 nigel 75 .sp
1652 nigel 63 An empty string is not considered to be a valid match if this option is set. If
1653     there are alternatives in the pattern, they are tried. If all the alternatives
1654     match the empty string, the entire match fails. For example, if the pattern
1655 nigel 75 .sp
1656 nigel 63 a?b?
1657 nigel 75 .sp
1658 ph10 442 is applied to a string not beginning with "a" or "b", it matches an empty
1659 nigel 63 string at the start of the subject. With PCRE_NOTEMPTY set, this match is not
1660     valid, so PCRE searches further into the string for occurrences of "a" or "b".
1661 ph10 442 .sp
1662     PCRE_NOTEMPTY_ATSTART
1663     .sp
1664 ph10 461 This is like PCRE_NOTEMPTY, except that an empty string match that is not at
1665 ph10 442 the start of the subject is permitted. If the pattern is anchored, such a match
1666     can occur only if the pattern contains \eK.
1667 nigel 75 .P
1668 ph10 442 Perl has no direct equivalent of PCRE_NOTEMPTY or PCRE_NOTEMPTY_ATSTART, but it
1669     does make a special case of a pattern match of the empty string within its
1670     \fBsplit()\fP function, and when using the /g modifier. It is possible to
1671     emulate Perl's behaviour after matching a null string by first trying the match
1672     again at the same offset with PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED, and then
1673     if that fails, by advancing the starting offset (see below) and trying an
1674     ordinary match again. There is some code that demonstrates how to do this in
1675     the
1676 ph10 429 .\" HREF
1677     \fBpcredemo\fP
1678     .\"
1679 ph10 572 sample program. In the most general case, you have to check to see if the
1680     newline convention recognizes CRLF as a newline, and if so, and the current
1681 ph10 566 character is CR followed by LF, advance the starting offset by two characters
1682     instead of one.
1683 nigel 75 .sp
1684 ph10 389 PCRE_NO_START_OPTIMIZE
1685     .sp
1686 ph10 392 There are a number of optimizations that \fBpcre_exec()\fP uses at the start of
1687 ph10 542 a match, in order to speed up the process. For example, if it is known that an
1688     unanchored match must start with a specific character, it searches the subject
1689     for that character, and fails immediately if it cannot find it, without
1690 ph10 545 actually running the main matching function. This means that a special item
1691 ph10 542 such as (*COMMIT) at the start of a pattern is not considered until after a
1692 ph10 548 suitable starting point for the match has been found. When callouts or (*MARK)
1693 ph10 546 items are in use, these "start-up" optimizations can cause them to be skipped
1694     if the pattern is never actually used. The start-up optimizations are in effect
1695     a pre-scan of the subject that takes place before the pattern is run.
1696     .P
1697     The PCRE_NO_START_OPTIMIZE option disables the start-up optimizations, possibly
1698     causing performance to suffer, but ensuring that in cases where the result is
1699     "no match", the callouts do occur, and that items such as (*COMMIT) and (*MARK)
1700 ph10 579 are considered at every possible starting position in the subject string. If
1701     PCRE_NO_START_OPTIMIZE is set at compile time, it cannot be unset at matching
1702 ph10 576 time.
1703     .P
1704 ph10 546 Setting PCRE_NO_START_OPTIMIZE can change the outcome of a matching operation.
1705     Consider the pattern
1706 ph10 389 .sp
1707 ph10 546 (*COMMIT)ABC
1708     .sp
1709     When this is compiled, PCRE records the fact that a match must start with the
1710 ph10 548 character "A". Suppose the subject string is "DEFABC". The start-up
1711     optimization scans along the subject, finds "A" and runs the first match
1712 ph10 546 attempt from there. The (*COMMIT) item means that the pattern must match the
1713 ph10 548 current starting position, which in this case, it does. However, if the same
1714     match is run with PCRE_NO_START_OPTIMIZE set, the initial scan along the
1715     subject string does not happen. The first match attempt is run starting from
1716 ph10 546 "D" and when this fails, (*COMMIT) prevents any further matches being tried, so
1717     the overall result is "no match". If the pattern is studied, more start-up
1718     optimizations may be used. For example, a minimum length for the subject may be
1719     recorded. Consider the pattern
1720     .sp
1721     (*MARK:A)(X|Y)
1722     .sp
1723 ph10 548 The minimum length for a match is one character. If the subject is "ABC", there
1724     will be attempts to match "ABC", "BC", "C", and then finally an empty string.
1725     If the pattern is studied, the final attempt does not take place, because PCRE
1726     knows that the subject is too short, and so the (*MARK) is never encountered.
1727     In this case, studying the pattern does not affect the overall match result,
1728     which is still "no match", but it does affect the auxiliary information that is
1729 ph10 546 returned.
1730     .sp
1731 nigel 75 PCRE_NO_UTF8_CHECK
1732     .sp
1733     When PCRE_UTF8 is set at compile time, the validity of the subject as a UTF-8
1734     string is automatically checked when \fBpcre_exec()\fP is subsequently called.
1735 ph10 211 The value of \fIstartoffset\fP is also checked to ensure that it points to the
1736     start of a UTF-8 character. There is a discussion about the validity of UTF-8
1737     strings in the
1738     .\" HREF
1739 ph10 856 \fBpcreunicode\fP
1740 ph10 211 .\"
1741 ph10 856 page. If an invalid sequence of bytes is found, \fBpcre_exec()\fP returns the
1742     error PCRE_ERROR_BADUTF8 or, if PCRE_PARTIAL_HARD is set and the problem is a
1743     truncated character at the end of the subject, PCRE_ERROR_SHORTUTF8. In both
1744     cases, information about the precise nature of the error may also be returned
1745     (see the descriptions of these errors in the section entitled \fIError return
1746     values from\fP \fBpcre_exec()\fP
1747 ph10 598 .\" HTML <a href="#errorlist">
1748     .\" </a>
1749     below).
1750     .\"
1751     If \fIstartoffset\fP contains a value that does not point to the start of a
1752     UTF-8 character (or to the end of the subject), PCRE_ERROR_BADUTF8_OFFSET is
1753 ph10 569 returned.
1754 nigel 75 .P
1755     If you already know that your subject is valid, and you want to skip these
1756     checks for performance reasons, you can set the PCRE_NO_UTF8_CHECK option when
1757     calling \fBpcre_exec()\fP. You might want to do this for the second and
1758     subsequent calls to \fBpcre_exec()\fP if you are making repeated calls to find
1759     all the matches in a single subject string. However, you should be sure that
1760 ph10 856 the value of \fIstartoffset\fP points to the start of a character (or the end
1761     of the subject). When PCRE_NO_UTF8_CHECK is set, the effect of passing an
1762     invalid string as a subject or an invalid value of \fIstartoffset\fP is
1763 ph10 567 undefined. Your program may crash.
1764 nigel 75 .sp
1765 ph10 461 PCRE_PARTIAL_HARD
1766 ph10 428 PCRE_PARTIAL_SOFT
1767 nigel 75 .sp
1768 ph10 428 These options turn on the partial matching feature. For backwards
1769     compatibility, PCRE_PARTIAL is a synonym for PCRE_PARTIAL_SOFT. A partial match
1770     occurs if the end of the subject string is reached successfully, but there are
1771     not enough subject characters to complete the match. If this happens when
1772 ph10 553 PCRE_PARTIAL_SOFT (but not PCRE_PARTIAL_HARD) is set, matching continues by
1773     testing any remaining alternatives. Only if no complete match can be found is
1774     PCRE_ERROR_PARTIAL returned instead of PCRE_ERROR_NOMATCH. In other words,
1775     PCRE_PARTIAL_SOFT says that the caller is prepared to handle a partial match,
1776     but only if no complete match can be found.
1777     .P
1778     If PCRE_PARTIAL_HARD is set, it overrides PCRE_PARTIAL_SOFT. In this case, if a
1779     partial match is found, \fBpcre_exec()\fP immediately returns
1780     PCRE_ERROR_PARTIAL, without considering any other alternatives. In other words,
1781 ph10 572 when PCRE_PARTIAL_HARD is set, a partial match is considered to be more
1782 ph10 553 important that an alternative complete match.
1783     .P
1784     In both cases, the portion of the string that was inspected when the partial
1785     match was found is set as the first matching string. There is a more detailed
1786     discussion of partial and multi-segment matching, with examples, in the
1787 nigel 75 .\" HREF
1788     \fBpcrepartial\fP
1789     .\"
1790     documentation.
1791     .
1792 ph10 567 .
1793 nigel 75 .SS "The string to be matched by \fBpcre_exec()\fP"
1794     .rs
1795     .sp
1796     The subject string is passed to \fBpcre_exec()\fP as a pointer in
1797 ph10 856 \fIsubject\fP, a length in bytes in \fIlength\fP, and a starting byte offset
1798 ph10 572 in \fIstartoffset\fP. If this is negative or greater than the length of the
1799     subject, \fBpcre_exec()\fP returns PCRE_ERROR_BADOFFSET. When the starting
1800     offset is zero, the search for a match starts at the beginning of the subject,
1801     and this is by far the most common case. In UTF-8 mode, the byte offset must
1802     point to the start of a UTF-8 character (or the end of the subject). Unlike the
1803     pattern string, the subject may contain binary zero bytes.
1804 nigel 75 .P
1805 nigel 63 A non-zero starting offset is useful when searching for another match in the
1806 nigel 75 same subject by calling \fBpcre_exec()\fP again after a previous success.
1807     Setting \fIstartoffset\fP differs from just passing over a shortened string and
1808 nigel 63 setting PCRE_NOTBOL in the case of a pattern that begins with any kind of
1809     lookbehind. For example, consider the pattern
1810 nigel 75 .sp
1811     \eBiss\eB
1812     .sp
1813     which finds occurrences of "iss" in the middle of words. (\eB matches only if
1814 nigel 63 the current position in the subject is not a word boundary.) When applied to
1815 nigel 75 the string "Mississipi" the first call to \fBpcre_exec()\fP finds the first
1816     occurrence. If \fBpcre_exec()\fP is called again with just the remainder of the
1817     subject, namely "issipi", it does not match, because \eB is always false at the
1818 nigel 63 start of the subject, which is deemed to be a word boundary. However, if
1819 nigel 75 \fBpcre_exec()\fP is passed the entire string again, but with \fIstartoffset\fP
1820 nigel 63 set to 4, it finds the second occurrence of "iss" because it is able to look
1821     behind the starting point to discover that it is preceded by a letter.
1822 nigel 75 .P
1823 ph10 567 Finding all the matches in a subject is tricky when the pattern can match an
1824     empty string. It is possible to emulate Perl's /g behaviour by first trying the
1825     match again at the same offset, with the PCRE_NOTEMPTY_ATSTART and
1826     PCRE_ANCHORED options, and then if that fails, advancing the starting offset
1827     and trying an ordinary match again. There is some code that demonstrates how to
1828     do this in the
1829     .\" HREF
1830     \fBpcredemo\fP
1831     .\"
1832 ph10 572 sample program. In the most general case, you have to check to see if the
1833     newline convention recognizes CRLF as a newline, and if so, and the current
1834 ph10 567 character is CR followed by LF, advance the starting offset by two characters
1835     instead of one.
1836     .P
1837 nigel 63 If a non-zero starting offset is passed when the pattern is anchored, one
1838 nigel 75 attempt to match at the given offset is made. This can only succeed if the
1839 nigel 63 pattern does not require the match to be at the start of the subject.
1840 nigel 75 .
1841 ph10 567 .
1842 nigel 75 .SS "How \fBpcre_exec()\fP returns captured substrings"
1843     .rs
1844     .sp
1845 nigel 63 In general, a pattern matches a certain portion of the subject, and in
1846     addition, further substrings from the subject may be picked out by parts of the
1847     pattern. Following the usage in Jeffrey Friedl's book, this is called
1848     "capturing" in what follows, and the phrase "capturing subpattern" is used for
1849     a fragment of a pattern that picks out a substring. PCRE supports several other
1850     kinds of parenthesized subpattern that do not cause substrings to be captured.
1851 nigel 75 .P
1852 ph10 368 Captured substrings are returned to the caller via a vector of integers whose
1853     address is passed in \fIovector\fP. The number of elements in the vector is
1854     passed in \fIovecsize\fP, which must be a non-negative number. \fBNote\fP: this
1855     argument is NOT the size of \fIovector\fP in bytes.
1856 nigel 75 .P
1857     The first two-thirds of the vector is used to pass back captured substrings,
1858     each substring using a pair of integers. The remaining third of the vector is
1859     used as workspace by \fBpcre_exec()\fP while matching capturing subpatterns,
1860 ph10 368 and is not available for passing back information. The number passed in
1861 nigel 75 \fIovecsize\fP should always be a multiple of three. If it is not, it is
1862     rounded down.
1863     .P
1864     When a match is successful, information about captured substrings is returned
1865     in pairs of integers, starting at the beginning of \fIovector\fP, and
1866 ph10 371 continuing up to two-thirds of its length at the most. The first element of
1867 ph10 368 each pair is set to the byte offset of the first character in a substring, and
1868     the second is set to the byte offset of the first character after the end of a
1869     substring. \fBNote\fP: these values are always byte offsets, even in UTF-8
1870     mode. They are not character counts.
1871 nigel 75 .P
1872 ph10 368 The first pair of integers, \fIovector[0]\fP and \fIovector[1]\fP, identify the
1873     portion of the subject string matched by the entire pattern. The next pair is
1874     used for the first capturing subpattern, and so on. The value returned by
1875     \fBpcre_exec()\fP is one more than the highest numbered pair that has been set.
1876     For example, if two substrings have been captured, the returned value is 3. If
1877     there are no capturing subpatterns, the return value from a successful match is
1878     1, indicating that just the first pair of offsets has been set.
1879     .P
1880 nigel 63 If a capturing subpattern is matched repeatedly, it is the last portion of the
1881 nigel 75 string that it matched that is returned.
1882     .P
1883     If the vector is too small to hold all the captured substring offsets, it is
1884     used as far as possible (up to two-thirds of its length), and the function
1885 ph10 686 returns a value of zero. If neither the actual string matched not any captured
1886     substrings are of interest, \fBpcre_exec()\fP may be called with \fIovector\fP
1887     passed as NULL and \fIovecsize\fP as zero. However, if the pattern contains
1888     back references and the \fIovector\fP is not big enough to remember the related
1889     substrings, PCRE has to get additional memory for use during matching. Thus it
1890     is usually advisable to supply an \fIovector\fP of reasonable size.
1891 nigel 75 .P
1892 ph10 686 There are some cases where zero is returned (indicating vector overflow) when
1893     in fact the vector is exactly the right size for the final match. For example,
1894     consider the pattern
1895     .sp
1896     (a)(?:(b)c|bd)
1897     .sp
1898     If a vector of 6 elements (allowing for only 1 captured substring) is given
1899     with subject string "abd", \fBpcre_exec()\fP will try to set the second
1900     captured string, thereby recording a vector overflow, before failing to match
1901 ph10 691 "c" and backing up to try the second alternative. The zero return, however,
1902 ph10 686 does correctly indicate that the maximum number of slots (namely 2) have been
1903     filled. In similar cases where there is temporary overflow, but the final
1904     number of used slots is actually less than the maximum, a non-zero value is
1905     returned.
1906     .P
1907 ph10 456 The \fBpcre_fullinfo()\fP function can be used to find out how many capturing
1908 nigel 63 subpatterns there are in a compiled pattern. The smallest size for
1909 nigel 75 \fIovector\fP that will allow for \fIn\fP captured substrings, in addition to
1910     the offsets of the substring matched by the whole pattern, is (\fIn\fP+1)*3.
1911 nigel 91 .P
1912     It is possible for capturing subpattern number \fIn+1\fP to match some part of
1913     the subject when subpattern \fIn\fP has not been used at all. For example, if
1914     the string "abc" is matched against the pattern (a|(z))(bc) the return from the
1915     function is 4, and subpatterns 1 and 3 are matched, but 2 is not. When this
1916     happens, both values in the offset pairs corresponding to unused subpatterns
1917     are set to -1.
1918     .P
1919     Offset values that correspond to unused subpatterns at the end of the
1920     expression are also set to -1. For example, if the string "abc" is matched
1921     against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not matched. The
1922     return from the function is 2, because the highest used capturing subpattern
1923 ph10 568 number is 1, and the offsets for for the second and third capturing subpatterns
1924     (assuming the vector is large enough, of course) are set to -1.
1925 nigel 91 .P
1926 ph10 686 \fBNote\fP: Elements in the first two-thirds of \fIovector\fP that do not
1927 ph10 683 correspond to capturing parentheses in the pattern are never changed. That is,
1928     if a pattern contains \fIn\fP capturing parentheses, no more than
1929     \fIovector[0]\fP to \fIovector[2n+1]\fP are set by \fBpcre_exec()\fP. The other
1930     elements (in the first two-thirds) retain whatever values they previously had.
1931 ph10 568 .P
1932 nigel 91 Some convenience functions are provided for extracting the captured substrings
1933     as separate strings. These are described below.
1934 nigel 75 .
1935 ph10 598 .
1936 nigel 77 .\" HTML <a name="errorlist"></a>
1937 nigel 91 .SS "Error return values from \fBpcre_exec()\fP"
1938 nigel 75 .rs
1939     .sp
1940     If \fBpcre_exec()\fP fails, it returns a negative number. The following are
1941 nigel 63 defined in the header file:
1942 nigel 75 .sp
1943 nigel 63 PCRE_ERROR_NOMATCH (-1)
1944 nigel 75 .sp
1945 nigel 63 The subject string did not match the pattern.
1946 nigel 75 .sp
1947 nigel 63 PCRE_ERROR_NULL (-2)
1948 nigel 75 .sp
1949     Either \fIcode\fP or \fIsubject\fP was passed as NULL, or \fIovector\fP was
1950     NULL and \fIovecsize\fP was not zero.
1951     .sp
1952 nigel 63 PCRE_ERROR_BADOPTION (-3)
1953 nigel 75 .sp
1954     An unrecognized bit was set in the \fIoptions\fP argument.
1955     .sp
1956 nigel 63 PCRE_ERROR_BADMAGIC (-4)
1957 nigel 75 .sp
1958 nigel 63 PCRE stores a 4-byte "magic number" at the start of the compiled code, to catch
1959 nigel 75 the case when it is passed a junk pointer and to detect when a pattern that was
1960     compiled in an environment of one endianness is run in an environment with the
1961     other endianness. This is the error that PCRE gives when the magic number is
1962     not present.
1963     .sp
1964 nigel 93 PCRE_ERROR_UNKNOWN_OPCODE (-5)
1965 nigel 75 .sp
1966 nigel 63 While running the pattern match, an unknown item was encountered in the
1967     compiled pattern. This error could be caused by a bug in PCRE or by overwriting
1968     of the compiled pattern.
1969 nigel 75 .sp
1970 nigel 63 PCRE_ERROR_NOMEMORY (-6)
1971 nigel 75 .sp
1972     If a pattern contains back references, but the \fIovector\fP that is passed to
1973     \fBpcre_exec()\fP is not big enough to remember the referenced substrings, PCRE
1974 nigel 63 gets a block of memory at the start of matching to use for this purpose. If the
1975 nigel 75 call via \fBpcre_malloc()\fP fails, this error is given. The memory is
1976     automatically freed at the end of matching.
1977 ph10 531 .P
1978 ph10 535 This error is also given if \fBpcre_stack_malloc()\fP fails in
1979 ph10 531 \fBpcre_exec()\fP. This can happen only when PCRE has been compiled with
1980     \fB--disable-stack-for-recursion\fP.
1981 nigel 75 .sp
1982 nigel 63 PCRE_ERROR_NOSUBSTRING (-7)
1983 nigel 75 .sp
1984     This error is used by the \fBpcre_copy_substring()\fP,
1985     \fBpcre_get_substring()\fP, and \fBpcre_get_substring_list()\fP functions (see
1986     below). It is never returned by \fBpcre_exec()\fP.
1987     .sp
1988 nigel 63 PCRE_ERROR_MATCHLIMIT (-8)
1989 nigel 75 .sp
1990 nigel 87 The backtracking limit, as specified by the \fImatch_limit\fP field in a
1991     \fBpcre_extra\fP structure (or defaulted) was reached. See the description
1992     above.
1993     .sp
1994 nigel 63 PCRE_ERROR_CALLOUT (-9)
1995 nigel 75 .sp
1996     This error is never generated by \fBpcre_exec()\fP itself. It is provided for
1997 nigel 63 use by callout functions that want to yield a distinctive error code. See the
1998 nigel 75 .\" HREF
1999     \fBpcrecallout\fP
2000     .\"
2001     documentation for details.
2002     .sp
2003 nigel 73 PCRE_ERROR_BADUTF8 (-10)
2004 nigel 75 .sp
2005 ph10 598 A string that contains an invalid UTF-8 byte sequence was passed as a subject,
2006     and the PCRE_NO_UTF8_CHECK option was not set. If the size of the output vector
2007     (\fIovecsize\fP) is at least 2, the byte offset to the start of the the invalid
2008     UTF-8 character is placed in the first element, and a reason code is placed in
2009     the second element. The reason codes are listed in the
2010     .\" HTML <a href="#badutf8reasons">
2011     .\" </a>
2012     following section.
2013     .\"
2014     For backward compatibility, if PCRE_PARTIAL_HARD is set and the problem is a
2015     truncated UTF-8 character at the end of the subject (reason codes 1 to 5),
2016     PCRE_ERROR_SHORTUTF8 is returned instead of PCRE_ERROR_BADUTF8.
2017 nigel 75 .sp
2018 nigel 73 PCRE_ERROR_BADUTF8_OFFSET (-11)
2019 nigel 75 .sp
2020 ph10 654 The UTF-8 byte sequence that was passed as a subject was checked and found to
2021 ph10 598 be valid (the PCRE_NO_UTF8_CHECK option was not set), but the value of
2022     \fIstartoffset\fP did not point to the beginning of a UTF-8 character or the
2023 ph10 569 end of the subject.
2024 nigel 75 .sp
2025 nigel 77 PCRE_ERROR_PARTIAL (-12)
2026 nigel 75 .sp
2027     The subject string did not match, but it did match partially. See the
2028     .\" HREF
2029     \fBpcrepartial\fP
2030     .\"
2031     documentation for details of partial matching.
2032     .sp
2033 nigel 77 PCRE_ERROR_BADPARTIAL (-13)
2034 nigel 75 .sp
2035 ph10 426 This code is no longer in use. It was formerly returned when the PCRE_PARTIAL
2036     option was used with a compiled pattern containing items that were not
2037 ph10 461 supported for partial matching. From release 8.00 onwards, there are no
2038 ph10 426 restrictions on partial matching.
2039 nigel 75 .sp
2040 nigel 77 PCRE_ERROR_INTERNAL (-14)
2041 nigel 75 .sp
2042     An unexpected internal error has occurred. This error could be caused by a bug
2043     in PCRE or by overwriting of the compiled pattern.
2044     .sp
2045 nigel 77 PCRE_ERROR_BADCOUNT (-15)
2046 nigel 75 .sp
2047     This error is given if the value of the \fIovecsize\fP argument is negative.
2048 nigel 93 .sp
2049     PCRE_ERROR_RECURSIONLIMIT (-21)
2050     .sp
2051     The internal recursion limit, as specified by the \fImatch_limit_recursion\fP
2052     field in a \fBpcre_extra\fP structure (or defaulted) was reached. See the
2053     description above.
2054     .sp
2055     PCRE_ERROR_BADNEWLINE (-23)
2056     .sp
2057     An invalid combination of PCRE_NEWLINE_\fIxxx\fP options was given.
2058 ph10 567 .sp
2059     PCRE_ERROR_BADOFFSET (-24)
2060     .sp
2061 ph10 572 The value of \fIstartoffset\fP was negative or greater than the length of the
2062 ph10 567 subject, that is, the value in \fIlength\fP.
2063 ph10 569 .sp
2064     PCRE_ERROR_SHORTUTF8 (-25)
2065     .sp
2066 ph10 598 This error is returned instead of PCRE_ERROR_BADUTF8 when the subject string
2067     ends with a truncated UTF-8 character and the PCRE_PARTIAL_HARD option is set.
2068     Information about the failure is returned as for PCRE_ERROR_BADUTF8. It is in
2069     fact sufficient to detect this case, but this special error code for
2070     PCRE_PARTIAL_HARD precedes the implementation of returned information; it is
2071     retained for backwards compatibility.
2072 ph10 642 .sp
2073     PCRE_ERROR_RECURSELOOP (-26)
2074     .sp
2075 ph10 654 This error is returned when \fBpcre_exec()\fP detects a recursion loop within
2076     the pattern. Specifically, it means that either the whole pattern or a
2077     subpattern has been called recursively for the second time at the same position
2078 ph10 642 in the subject string. Some simple patterns that might do this are detected and
2079     faulted at compile time, but more complicated cases, in particular mutual
2080     recursions between two different subpatterns, cannot be detected until run
2081     time.
2082 ph10 678 .sp
2083     PCRE_ERROR_JIT_STACKLIMIT (-27)
2084     .sp
2085 ph10 691 This error is returned when a pattern that was successfully studied using the
2086 ph10 686 PCRE_STUDY_JIT_COMPILE option is being matched, but the memory available for
2087     the just-in-time processing stack is not large enough. See the
2088 ph10 678 .\" HREF
2089     \fBpcrejit\fP
2090     .\"
2091 ph10 691 documentation for more details.
2092 ph10 856 .sp
2093     PCRE_ERROR_BADMODE (-28)
2094     .sp
2095     This error is given if a pattern that was compiled by the 8-bit library is
2096     passed to a 16-bit library function, or vice versa.
2097     .sp
2098     PCRE_ERROR_BADENDIANNESS (-29)
2099     .sp
2100     This error is given if a pattern that was compiled and saved is reloaded on a
2101     host with different endianness. The utility function
2102     \fBpcre_pattern_to_host_byte_order()\fP can be used to convert such a pattern
2103     so that it runs on the new host.
2104 nigel 93 .P
2105 ph10 197 Error numbers -16 to -20 and -22 are not used by \fBpcre_exec()\fP.
2106 nigel 75 .
2107     .
2108 ph10 598 .\" HTML <a name="badutf8reasons"></a>
2109     .SS "Reason codes for invalid UTF-8 strings"
2110     .rs
2111     .sp
2112 ph10 856 This section applies only to the 8-bit library. The corresponding information
2113     for the 16-bit library is given in the
2114     .\" HREF
2115     \fBpcre16\fP
2116     .\"
2117     page.
2118     .P
2119 ph10 654 When \fBpcre_exec()\fP returns either PCRE_ERROR_BADUTF8 or
2120     PCRE_ERROR_SHORTUTF8, and the size of the output vector (\fIovecsize\fP) is at
2121     least 2, the offset of the start of the invalid UTF-8 character is placed in
2122     the first output vector element (\fIovector[0]\fP) and a reason code is placed
2123 ph10 598 in the second element (\fIovector[1]\fP). The reason codes are given names in
2124     the \fBpcre.h\fP header file:
2125     .sp
2126     PCRE_UTF8_ERR1
2127     PCRE_UTF8_ERR2
2128     PCRE_UTF8_ERR3
2129     PCRE_UTF8_ERR4
2130     PCRE_UTF8_ERR5
2131     .sp
2132 ph10 654 The string ends with a truncated UTF-8 character; the code specifies how many
2133 ph10 598 bytes are missing (1 to 5). Although RFC 3629 restricts UTF-8 characters to be
2134     no longer than 4 bytes, the encoding scheme (originally defined by RFC 2279)
2135 ph10 654 allows for up to 6 bytes, and this is checked first; hence the possibility of
2136 ph10 598 4 or 5 missing bytes.
2137     .sp
2138     PCRE_UTF8_ERR6
2139     PCRE_UTF8_ERR7
2140     PCRE_UTF8_ERR8
2141     PCRE_UTF8_ERR9
2142     PCRE_UTF8_ERR10
2143     .sp
2144 ph10 654 The two most significant bits of the 2nd, 3rd, 4th, 5th, or 6th byte of the
2145 ph10 598 character do not have the binary value 0b10 (that is, either the most
2146     significant bit is 0, or the next bit is 1).
2147 ph10 654 .sp
2148 ph10 598 PCRE_UTF8_ERR11
2149     PCRE_UTF8_ERR12
2150     .sp
2151 ph10 654 A character that is valid by the RFC 2279 rules is either 5 or 6 bytes long;
2152     these code points are excluded by RFC 3629.
2153     .sp
2154 ph10 598 PCRE_UTF8_ERR13
2155     .sp
2156 ph10 654 A 4-byte character has a value greater than 0x10fff; these code points are
2157 ph10 598 excluded by RFC 3629.
2158 ph10 654 .sp
2159 ph10 598 PCRE_UTF8_ERR14
2160     .sp
2161     A 3-byte character has a value in the range 0xd800 to 0xdfff; this range of
2162 ph10 654 code points are reserved by RFC 3629 for use with UTF-16, and so are excluded
2163 ph10 598 from UTF-8.
2164 ph10 654 .sp
2165 ph10 598 PCRE_UTF8_ERR15
2166     PCRE_UTF8_ERR16
2167     PCRE_UTF8_ERR17
2168     PCRE_UTF8_ERR18
2169     PCRE_UTF8_ERR19
2170     .sp
2171 ph10 654 A 2-, 3-, 4-, 5-, or 6-byte character is "overlong", that is, it codes for a
2172     value that can be represented by fewer bytes, which is invalid. For example,
2173 ph10 598 the two bytes 0xc0, 0xae give the value 0x2e, whose correct coding uses just
2174     one byte.
2175     .sp
2176     PCRE_UTF8_ERR20
2177     .sp
2178 ph10 654 The two most significant bits of the first byte of a character have the binary
2179     value 0b10 (that is, the most significant bit is 1 and the second is 0). Such a
2180 ph10 598 byte can only validly occur as the second or subsequent byte of a multi-byte
2181     character.
2182     .sp
2183     PCRE_UTF8_ERR21
2184     .sp
2185     The first byte of a character has the value 0xfe or 0xff. These values can
2186     never occur in a valid UTF-8 string.
2187     .
2188     .
2189 nigel 75 .SH "EXTRACTING CAPTURED SUBSTRINGS BY NUMBER"
2190 nigel 63 .rs
2191     .sp
2192 nigel 75 .B int pcre_copy_substring(const char *\fIsubject\fP, int *\fIovector\fP,
2193 nigel 63 .ti +5n
2194 nigel 75 .B int \fIstringcount\fP, int \fIstringnumber\fP, char *\fIbuffer\fP,
2195 nigel 63 .ti +5n
2196 nigel 75 .B int \fIbuffersize\fP);
2197 nigel 63 .PP
2198 nigel 75 .B int pcre_get_substring(const char *\fIsubject\fP, int *\fIovector\fP,
2199 nigel 63 .ti +5n
2200 nigel 75 .B int \fIstringcount\fP, int \fIstringnumber\fP,
2201 nigel 63 .ti +5n
2202 nigel 75 .B const char **\fIstringptr\fP);
2203 nigel 63 .PP
2204 nigel 75 .B int pcre_get_substring_list(const char *\fIsubject\fP,
2205 nigel 63 .ti +5n
2206 nigel 75 .B int *\fIovector\fP, int \fIstringcount\fP, "const char ***\fIlistptr\fP);"
2207 nigel 63 .PP
2208     Captured substrings can be accessed directly by using the offsets returned by
2209 nigel 75 \fBpcre_exec()\fP in \fIovector\fP. For convenience, the functions
2210     \fBpcre_copy_substring()\fP, \fBpcre_get_substring()\fP, and
2211     \fBpcre_get_substring_list()\fP are provided for extracting captured substrings
2212 nigel 63 as new, separate, zero-terminated strings. These functions identify substrings
2213     by number. The next section describes functions for extracting named
2214 nigel 91 substrings.
2215 nigel 75 .P
2216 nigel 91 A substring that contains a binary zero is correctly extracted and has a
2217     further zero added on the end, but the result is not, of course, a C string.
2218     However, you can process such a string by referring to the length that is
2219     returned by \fBpcre_copy_substring()\fP and \fBpcre_get_substring()\fP.
2220     Unfortunately, the interface to \fBpcre_get_substring_list()\fP is not adequate
2221     for handling strings containing binary zeros, because the end of the final
2222     string is not independently indicated.
2223     .P
2224 nigel 63 The first three arguments are the same for all three of these functions:
2225 nigel 75 \fIsubject\fP is the subject string that has just been successfully matched,
2226     \fIovector\fP is a pointer to the vector of integer offsets that was passed to
2227     \fBpcre_exec()\fP, and \fIstringcount\fP is the number of substrings that were
2228 nigel 63 captured by the match, including the substring that matched the entire regular
2229 nigel 75 expression. This is the value returned by \fBpcre_exec()\fP if it is greater
2230     than zero. If \fBpcre_exec()\fP returned zero, indicating that it ran out of
2231     space in \fIovector\fP, the value passed as \fIstringcount\fP should be the
2232     number of elements in the vector divided by three.
2233     .P
2234     The functions \fBpcre_copy_substring()\fP and \fBpcre_get_substring()\fP
2235     extract a single substring, whose number is given as \fIstringnumber\fP. A
2236     value of zero extracts the substring that matched the entire pattern, whereas
2237     higher values extract the captured substrings. For \fBpcre_copy_substring()\fP,
2238     the string is placed in \fIbuffer\fP, whose length is given by
2239     \fIbuffersize\fP, while for \fBpcre_get_substring()\fP a new block of memory is
2240     obtained via \fBpcre_malloc\fP, and its address is returned via
2241     \fIstringptr\fP. The yield of the function is the length of the string, not
2242 nigel 93 including the terminating zero, or one of these error codes:
2243 nigel 75 .sp
2244 nigel 63 PCRE_ERROR_NOMEMORY (-6)
2245 nigel 75 .sp
2246     The buffer was too small for \fBpcre_copy_substring()\fP, or the attempt to get
2247     memory failed for \fBpcre_get_substring()\fP.
2248     .sp
2249 nigel 63 PCRE_ERROR_NOSUBSTRING (-7)
2250 nigel 75 .sp
2251     There is no substring whose number is \fIstringnumber\fP.
2252     .P
2253     The \fBpcre_get_substring_list()\fP function extracts all available substrings
2254 nigel 63 and builds a list of pointers to them. All this is done in a single block of
2255 nigel 75 memory that is obtained via \fBpcre_malloc\fP. The address of the memory block
2256     is returned via \fIlistptr\fP, which is also the start of the list of string
2257 nigel 63 pointers. The end of the list is marked by a NULL pointer. The yield of the
2258 nigel 93 function is zero if all went well, or the error code
2259 nigel 75 .sp
2260 nigel 63 PCRE_ERROR_NOMEMORY (-6)
2261 nigel 75 .sp
2262 nigel 63 if the attempt to get the memory block failed.
2263 nigel 75 .P
2264 nigel 63 When any of these functions encounter a substring that is unset, which can
2265 nigel 75 happen when capturing subpattern number \fIn+1\fP matches some part of the
2266     subject, but subpattern \fIn\fP has not been used at all, they return an empty
2267 nigel 63 string. This can be distinguished from a genuine zero-length substring by
2268 nigel 75 inspecting the appropriate offset in \fIovector\fP, which is negative for unset
2269 nigel 63 substrings.
2270 nigel 75 .P
2271     The two convenience functions \fBpcre_free_substring()\fP and
2272     \fBpcre_free_substring_list()\fP can be used to free the memory returned by
2273     a previous call of \fBpcre_get_substring()\fP or
2274     \fBpcre_get_substring_list()\fP, respectively. They do nothing more than call
2275     the function pointed to by \fBpcre_free\fP, which of course could be called
2276 nigel 63 directly from a C program. However, PCRE is used in some situations where it is
2277 nigel 91 linked via a special interface to another programming language that cannot use
2278 nigel 75 \fBpcre_free\fP directly; it is for these cases that the functions are
2279 nigel 63 provided.
2280 nigel 75 .
2281     .
2282     .SH "EXTRACTING CAPTURED SUBSTRINGS BY NAME"
2283 nigel 63 .rs
2284     .sp
2285 nigel 75 .B int pcre_get_stringnumber(const pcre *\fIcode\fP,
2286 nigel 63 .ti +5n
2287 nigel 75 .B const char *\fIname\fP);
2288 nigel 63 .PP
2289 nigel 75 .B int pcre_copy_named_substring(const pcre *\fIcode\fP,
2290 nigel 63 .ti +5n
2291 nigel 75 .B const char *\fIsubject\fP, int *\fIovector\fP,
2292     .ti +5n
2293     .B int \fIstringcount\fP, const char *\fIstringname\fP,
2294     .ti +5n
2295     .B char *\fIbuffer\fP, int \fIbuffersize\fP);
2296 nigel 63 .PP
2297 nigel 75 .B int pcre_get_named_substring(const pcre *\fIcode\fP,
2298 nigel 63 .ti +5n
2299 nigel 75 .B const char *\fIsubject\fP, int *\fIovector\fP,
2300 nigel 63 .ti +5n
2301 nigel 75 .B int \fIstringcount\fP, const char *\fIstringname\fP,
2302 nigel 63 .ti +5n
2303 nigel 75 .B const char **\fIstringptr\fP);
2304 nigel 63 .PP
2305 nigel 75 To extract a substring by name, you first have to find associated number.
2306     For example, for this pattern
2307     .sp
2308 nigel 93 (a+)b(?<xxx>\ed+)...
2309 nigel 75 .sp
2310 nigel 91 the number of the subpattern called "xxx" is 2. If the name is known to be
2311     unique (PCRE_DUPNAMES was not set), you can find the number from the name by
2312     calling \fBpcre_get_stringnumber()\fP. The first argument is the compiled
2313     pattern, and the second is the name. The yield of the function is the
2314 nigel 75 subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no subpattern of
2315     that name.
2316     .P
2317     Given the number, you can extract the substring directly, or use one of the
2318     functions described in the previous section. For convenience, there are also
2319     two functions that do the whole job.
2320     .P
2321 nigel 91 Most of the arguments of \fBpcre_copy_named_substring()\fP and
2322     \fBpcre_get_named_substring()\fP are the same as those for the similarly named
2323 nigel 75 functions that extract by number. As these are described in the previous
2324     section, they are not re-described here. There are just two differences:
2325     .P
2326 nigel 63 First, instead of a substring number, a substring name is given. Second, there
2327     is an extra argument, given at the start, which is a pointer to the compiled
2328     pattern. This is needed in order to gain access to the name-to-number
2329     translation table.
2330 nigel 75 .P
2331     These functions call \fBpcre_get_stringnumber()\fP, and if it succeeds, they
2332 ph10 127 then call \fBpcre_copy_substring()\fP or \fBpcre_get_substring()\fP, as
2333     appropriate. \fBNOTE:\fP If PCRE_DUPNAMES is set and there are duplicate names,
2334 ph10 128 the behaviour may not be what you want (see the next section).
2335 ph10 385 .P
2336 ph10 457 \fBWarning:\fP If the pattern uses the (?| feature to set up multiple
2337     subpatterns with the same number, as described in the
2338     .\" HTML <a href="pcrepattern.html#dupsubpatternnumber">
2339     .\" </a>
2340     section on duplicate subpattern numbers
2341     .\"
2342     in the
2343     .\" HREF
2344     \fBpcrepattern\fP
2345     .\"
2346     page, you cannot use names to distinguish the different subpatterns, because
2347     names are not included in the compiled code. The matching process uses only
2348     numbers. For this reason, the use of different names for subpatterns of the
2349     same number causes an error at compile time.
2350 nigel 77 .
2351 ph10 686 .
2352 nigel 91 .SH "DUPLICATE SUBPATTERN NAMES"
2353     .rs
2354     .sp
2355     .B int pcre_get_stringtable_entries(const pcre *\fIcode\fP,
2356     .ti +5n
2357     .B const char *\fIname\fP, char **\fIfirst\fP, char **\fIlast\fP);
2358     .PP
2359     When a pattern is compiled with the PCRE_DUPNAMES option, names for subpatterns
2360 ph10 457 are not required to be unique. (Duplicate names are always allowed for
2361     subpatterns with the same number, created by using the (?| feature. Indeed, if
2362     such subpatterns are named, they are required to use the same names.)
2363     .P
2364     Normally, patterns with duplicate names are such that in any one match, only
2365     one of the named subpatterns participates. An example is shown in the
2366 nigel 91 .\" HREF
2367     \fBpcrepattern\fP
2368     .\"
2369 ph10 208 documentation.
2370 ph10 203 .P
2371     When duplicates are present, \fBpcre_copy_named_substring()\fP and
2372     \fBpcre_get_named_substring()\fP return the first substring corresponding to
2373     the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING (-7) is
2374     returned; no data is returned. The \fBpcre_get_stringnumber()\fP function
2375     returns one of the numbers that are associated with the name, but it is not
2376     defined which it is.
2377     .P
2378 nigel 91 If you want to get full details of all captured substrings for a given name,
2379     you must use the \fBpcre_get_stringtable_entries()\fP function. The first
2380     argument is the compiled pattern, and the second is the name. The third and
2381     fourth are pointers to variables which are updated by the function. After it
2382     has run, they point to the first and last entries in the name-to-number table
2383     for the given name. The function itself returns the length of each entry, or
2384 nigel 93 PCRE_ERROR_NOSUBSTRING (-7) if there are none. The format of the table is
2385 ph10 598 described above in the section entitled \fIInformation about a pattern\fP
2386     .\" HTML <a href="#infoaboutpattern">
2387     .\" </a>
2388     above.
2389     .\"
2390 nigel 93 Given all the relevant entries for the name, you can extract each of their
2391     numbers, and hence the captured data, if any.
2392 nigel 91 .
2393     .
2394 nigel 77 .SH "FINDING ALL POSSIBLE MATCHES"
2395     .rs
2396     .sp
2397     The traditional matching function uses a similar algorithm to Perl, which stops
2398     when it finds the first match, starting at a given point in the subject. If you
2399     want to find all possible matches, or the longest possible match, consider
2400     using the alternative matching function (see below) instead. If you cannot use
2401     the alternative function, but still need to find all possible matches, you
2402     can kludge it up by making use of the callout facility, which is described in
2403     the
2404     .\" HREF
2405     \fBpcrecallout\fP
2406     .\"
2407     documentation.
2408 nigel 75 .P
2409 nigel 77 What you have to do is to insert a callout right at the end of the pattern.
2410     When your callout function is called, extract and save the current matched
2411     substring. Then return 1, which forces \fBpcre_exec()\fP to backtrack and try
2412     other alternatives. Ultimately, when it runs out of matches, \fBpcre_exec()\fP
2413     will yield PCRE_ERROR_NOMATCH.
2414     .
2415     .
2416     .\" HTML <a name="dfamatch"></a>
2417     .SH "MATCHING A PATTERN: THE ALTERNATIVE FUNCTION"
2418     .rs
2419     .sp
2420     .B int pcre_dfa_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
2421     .ti +5n
2422     .B "const char *\fIsubject\fP," int \fIlength\fP, int \fIstartoffset\fP,
2423     .ti +5n
2424     .B int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP,
2425     .ti +5n
2426     .B int *\fIworkspace\fP, int \fIwscount\fP);
2427     .P
2428     The function \fBpcre_dfa_exec()\fP is called to match a subject string against
2429 nigel 93 a compiled pattern, using a matching algorithm that scans the subject string
2430     just once, and does not backtrack. This has different characteristics to the
2431     normal algorithm, and is not compatible with Perl. Some of the features of PCRE
2432     patterns are not supported. Nevertheless, there are times when this kind of
2433 ph10 461 matching can be useful. For a discussion of the two matching algorithms, and a
2434 ph10 435 list of features that \fBpcre_dfa_exec()\fP does not support, see the
2435 nigel 77 .\" HREF
2436     \fBpcrematching\fP
2437     .\"
2438     documentation.
2439     .P
2440     The arguments for the \fBpcre_dfa_exec()\fP function are the same as for
2441     \fBpcre_exec()\fP, plus two extras. The \fIovector\fP argument is used in a
2442     different way, and this is described below. The other common arguments are used
2443     in the same way as for \fBpcre_exec()\fP, so their description is not repeated
2444     here.
2445     .P
2446     The two additional arguments provide workspace for the function. The workspace
2447     vector should contain at least 20 elements. It is used for keeping track of
2448     multiple paths through the pattern tree. More workspace will be needed for
2449 nigel 91 patterns and subjects where there are a lot of potential matches.
2450 nigel 77 .P
2451 nigel 87 Here is an example of a simple call to \fBpcre_dfa_exec()\fP:
2452 nigel 77 .sp
2453     int rc;
2454     int ovector[10];
2455     int wspace[20];
2456 nigel 87 rc = pcre_dfa_exec(
2457 nigel 77 re, /* result of pcre_compile() */
2458     NULL, /* we didn't study the pattern */
2459     "some string", /* the subject string */
2460     11, /* the length of the subject string */
2461     0, /* start at offset 0 in the subject */
2462     0, /* default options */
2463     ovector, /* vector of integers for substring information */
2464     10, /* number of elements (NOT size in bytes) */
2465     wspace, /* working space vector */
2466     20); /* number of elements (NOT size in bytes) */
2467     .
2468     .SS "Option bits for \fBpcre_dfa_exec()\fP"
2469     .rs
2470     .sp
2471     The unused bits of the \fIoptions\fP argument for \fBpcre_dfa_exec()\fP must be
2472 nigel 91 zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_\fIxxx\fP,
2473 ph10 442 PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
2474 ph10 542 PCRE_NO_UTF8_CHECK, PCRE_BSR_ANYCRLF, PCRE_BSR_UNICODE, PCRE_NO_START_OPTIMIZE,
2475     PCRE_PARTIAL_HARD, PCRE_PARTIAL_SOFT, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART.
2476     All but the last four of these are exactly the same as for \fBpcre_exec()\fP,
2477     so their description is not repeated here.
2478 nigel 77 .sp
2479 ph10 428 PCRE_PARTIAL_HARD
2480 ph10 461 PCRE_PARTIAL_SOFT
2481 nigel 77 .sp
2482 ph10 428 These have the same general effect as they do for \fBpcre_exec()\fP, but the
2483     details are slightly different. When PCRE_PARTIAL_HARD is set for
2484     \fBpcre_dfa_exec()\fP, it returns PCRE_ERROR_PARTIAL if the end of the subject
2485     is reached and there is still at least one matching possibility that requires
2486     additional characters. This happens even if some complete matches have also
2487     been found. When PCRE_PARTIAL_SOFT is set, the return code PCRE_ERROR_NOMATCH
2488     is converted into PCRE_ERROR_PARTIAL if the end of the subject is reached,
2489     there have been no complete matches, but there is still at least one matching
2490 ph10 435 possibility. The portion of the string that was inspected when the longest
2491     partial match was found is set as the first matching string in both cases.
2492 ph10 553 There is a more detailed discussion of partial and multi-segment matching, with
2493     examples, in the
2494     .\" HREF
2495     \fBpcrepartial\fP
2496     .\"
2497     documentation.
2498 nigel 77 .sp
2499     PCRE_DFA_SHORTEST
2500     .sp
2501     Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to stop as
2502 nigel 93 soon as it has found one match. Because of the way the alternative algorithm
2503     works, this is necessarily the shortest possible match at the first possible
2504     matching point in the subject string.
2505 nigel 77 .sp
2506     PCRE_DFA_RESTART
2507     .sp
2508 ph10 428 When \fBpcre_dfa_exec()\fP returns a partial match, it is possible to call it
2509     again, with additional subject characters, and have it continue with the same
2510     match. The PCRE_DFA_RESTART option requests this action; when it is set, the
2511     \fIworkspace\fP and \fIwscount\fP options must reference the same vector as
2512     before because data about the match so far is left in them after a partial
2513     match. There is more discussion of this facility in the
2514 nigel 77 .\" HREF
2515     \fBpcrepartial\fP
2516     .\"
2517     documentation.
2518     .
2519 ph10 598 .
2520 nigel 77 .SS "Successful returns from \fBpcre_dfa_exec()\fP"
2521     .rs
2522     .sp
2523     When \fBpcre_dfa_exec()\fP succeeds, it may have matched more than one
2524     substring in the subject. Note, however, that all the matches from one run of
2525     the function start at the same point in the subject. The shorter matches are
2526     all initial substrings of the longer matches. For example, if the pattern
2527     .sp
2528     <.*>
2529     .sp
2530     is matched against the string
2531     .sp
2532     This is <something> <something else> <something further> no more
2533     .sp
2534     the three matched strings are
2535     .sp
2536     <something>
2537     <something> <something else>
2538     <something> <something else> <something further>
2539     .sp
2540     On success, the yield of the function is a number greater than zero, which is
2541     the number of matched substrings. The substrings themselves are returned in
2542     \fIovector\fP. Each string uses two elements; the first is the offset to the
2543 nigel 93 start, and the second is the offset to the end. In fact, all the strings have
2544     the same start offset. (Space could have been saved by giving this only once,
2545     but it was decided to retain some compatibility with the way \fBpcre_exec()\fP
2546     returns data, even though the meaning of the strings is different.)
2547 nigel 77 .P
2548     The strings are returned in reverse order of length; that is, the longest
2549     matching string is given first. If there were too many matches to fit into
2550     \fIovector\fP, the yield of the function is zero, and the vector is filled with
2551 ph10 691 the longest matches. Unlike \fBpcre_exec()\fP, \fBpcre_dfa_exec()\fP can use
2552 ph10 683 the entire \fIovector\fP for returning matched strings.
2553 nigel 77 .
2554 ph10 598 .
2555 nigel 77 .SS "Error returns from \fBpcre_dfa_exec()\fP"
2556     .rs
2557     .sp
2558     The \fBpcre_dfa_exec()\fP function returns a negative number when it fails.
2559     Many of the errors are the same as for \fBpcre_exec()\fP, and these are
2560     described
2561     .\" HTML <a href="#errorlist">
2562     .\" </a>
2563     above.
2564     .\"
2565     There are in addition the following errors that are specific to
2566     \fBpcre_dfa_exec()\fP:
2567     .sp
2568     PCRE_ERROR_DFA_UITEM (-16)
2569     .sp
2570     This return is given if \fBpcre_dfa_exec()\fP encounters an item in the pattern
2571     that it does not support, for instance, the use of \eC or a back reference.
2572     .sp
2573     PCRE_ERROR_DFA_UCOND (-17)
2574     .sp
2575 nigel 93 This return is given if \fBpcre_dfa_exec()\fP encounters a condition item that
2576     uses a back reference for the condition, or a test for recursion in a specific
2577     group. These are not supported.
2578 nigel 77 .sp
2579     PCRE_ERROR_DFA_UMLIMIT (-18)
2580     .sp
2581     This return is given if \fBpcre_dfa_exec()\fP is called with an \fIextra\fP
2582 ph10 678 block that contains a setting of the \fImatch_limit\fP or
2583     \fImatch_limit_recursion\fP fields. This is not supported (these fields are
2584     meaningless for DFA matching).
2585 nigel 77 .sp
2586     PCRE_ERROR_DFA_WSSIZE (-19)
2587     .sp
2588     This return is given if \fBpcre_dfa_exec()\fP runs out of space in the
2589     \fIworkspace\fP vector.
2590     .sp
2591     PCRE_ERROR_DFA_RECURSE (-20)
2592     .sp
2593     When a recursive subpattern is processed, the matching function calls itself
2594     recursively, using private vectors for \fIovector\fP and \fIworkspace\fP. This
2595     error is given if the output vector is not large enough. This should be
2596     extremely rare, as a vector of size 1000 is used.
2597 nigel 93 .
2598     .
2599     .SH "SEE ALSO"
2600     .rs
2601     .sp
2602 ph10 856 \fBpcre16\fP(3), \fBpcrebuild\fP(3), \fBpcrecallout\fP(3), \fBpcrecpp(3)\fP(3),
2603 nigel 93 \fBpcrematching\fP(3), \fBpcrepartial\fP(3), \fBpcreposix\fP(3),
2604     \fBpcreprecompile\fP(3), \fBpcresample\fP(3), \fBpcrestack\fP(3).
2605 ph10 99 .
2606     .
2607     .SH AUTHOR
2608     .rs
2609     .sp
2610     .nf
2611     Philip Hazel
2612     University Computing Service
2613     Cambridge CB2 3QH, England.
2614     .fi
2615     .
2616     .
2617     .SH REVISION
2618     .rs
2619     .sp
2620     .nf
2621 ph10 887 Last updated: 17 January 2012
2622 ph10 856 Copyright (c) 1997-2012 University of Cambridge.
2623 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