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

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

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 75 by nigel, Sat Feb 24 21:40:37 2007 UTC revision 745 by ph10, Mon Nov 14 11:41:03 2011 UTC
# Line 1  Line 1 
1  .TH PCRE 3  .TH PCREAPI 3
2  .SH NAME  .SH NAME
3  PCRE - Perl-compatible regular expressions  PCRE - Perl-compatible regular expressions
4  .SH "PCRE NATIVE API"  .SH "PCRE NATIVE API BASIC FUNCTIONS"
5  .rs  .rs
6  .sp  .sp
7  .B #include <pcre.h>  .B #include <pcre.h>
8  .PP  .PP
9  .SM  .SM
 .br  
10  .B pcre *pcre_compile(const char *\fIpattern\fP, int \fIoptions\fP,  .B pcre *pcre_compile(const char *\fIpattern\fP, int \fIoptions\fP,
11  .ti +5n  .ti +5n
12  .B const char **\fIerrptr\fP, int *\fIerroffset\fP,  .B const char **\fIerrptr\fP, int *\fIerroffset\fP,
13  .ti +5n  .ti +5n
14  .B const unsigned char *\fItableptr\fP);  .B const unsigned char *\fItableptr\fP);
15  .PP  .PP
16  .br  .B pcre *pcre_compile2(const char *\fIpattern\fP, int \fIoptions\fP,
17    .ti +5n
18    .B int *\fIerrorcodeptr\fP,
19    .ti +5n
20    .B const char **\fIerrptr\fP, int *\fIerroffset\fP,
21    .ti +5n
22    .B const unsigned char *\fItableptr\fP);
23    .PP
24  .B pcre_extra *pcre_study(const pcre *\fIcode\fP, int \fIoptions\fP,  .B pcre_extra *pcre_study(const pcre *\fIcode\fP, int \fIoptions\fP,
25  .ti +5n  .ti +5n
26  .B const char **\fIerrptr\fP);  .B const char **\fIerrptr\fP);
27  .PP  .PP
28  .br  .B void pcre_free_study(pcre_extra *\fIextra\fP);
29    .PP
30  .B int pcre_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"  .B int pcre_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
31  .ti +5n  .ti +5n
32  .B "const char *\fIsubject\fP," int \fIlength\fP, int \fIstartoffset\fP,  .B "const char *\fIsubject\fP," int \fIlength\fP, int \fIstartoffset\fP,
33  .ti +5n  .ti +5n
34  .B int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP);  .B int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP);
35    .
36    .
37    .SH "PCRE NATIVE API AUXILIARY FUNCTIONS"
38    .rs
39    .sp
40    .B pcre_jit_stack *pcre_jit_stack_alloc(int \fIstartsize\fP, int \fImaxsize\fP);
41    .PP
42    .B void pcre_jit_stack_free(pcre_jit_stack *\fIstack\fP);
43    .PP
44    .B void pcre_assign_jit_stack(pcre_extra *\fIextra\fP,
45    .ti +5n
46    .B pcre_jit_callback \fIcallback\fP, void *\fIdata\fP);
47    .PP
48    .B int pcre_dfa_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
49    .ti +5n
50    .B "const char *\fIsubject\fP," int \fIlength\fP, int \fIstartoffset\fP,
51    .ti +5n
52    .B int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP,
53    .ti +5n
54    .B int *\fIworkspace\fP, int \fIwscount\fP);
55  .PP  .PP
 .br  
56  .B int pcre_copy_named_substring(const pcre *\fIcode\fP,  .B int pcre_copy_named_substring(const pcre *\fIcode\fP,
57  .ti +5n  .ti +5n
58  .B const char *\fIsubject\fP, int *\fIovector\fP,  .B const char *\fIsubject\fP, int *\fIovector\fP,
# Line 35  PCRE - Perl-compatible regular expressio Line 61  PCRE - Perl-compatible regular expressio
61  .ti +5n  .ti +5n
62  .B char *\fIbuffer\fP, int \fIbuffersize\fP);  .B char *\fIbuffer\fP, int \fIbuffersize\fP);
63  .PP  .PP
 .br  
64  .B int pcre_copy_substring(const char *\fIsubject\fP, int *\fIovector\fP,  .B int pcre_copy_substring(const char *\fIsubject\fP, int *\fIovector\fP,
65  .ti +5n  .ti +5n
66  .B int \fIstringcount\fP, int \fIstringnumber\fP, char *\fIbuffer\fP,  .B int \fIstringcount\fP, int \fIstringnumber\fP, char *\fIbuffer\fP,
67  .ti +5n  .ti +5n
68  .B int \fIbuffersize\fP);  .B int \fIbuffersize\fP);
69  .PP  .PP
 .br  
70  .B int pcre_get_named_substring(const pcre *\fIcode\fP,  .B int pcre_get_named_substring(const pcre *\fIcode\fP,
71  .ti +5n  .ti +5n
72  .B const char *\fIsubject\fP, int *\fIovector\fP,  .B const char *\fIsubject\fP, int *\fIovector\fP,
# Line 51  PCRE - Perl-compatible regular expressio Line 75  PCRE - Perl-compatible regular expressio
75  .ti +5n  .ti +5n
76  .B const char **\fIstringptr\fP);  .B const char **\fIstringptr\fP);
77  .PP  .PP
 .br  
78  .B int pcre_get_stringnumber(const pcre *\fIcode\fP,  .B int pcre_get_stringnumber(const pcre *\fIcode\fP,
79  .ti +5n  .ti +5n
80  .B const char *\fIname\fP);  .B const char *\fIname\fP);
81  .PP  .PP
82  .br  .B int pcre_get_stringtable_entries(const pcre *\fIcode\fP,
83    .ti +5n
84    .B const char *\fIname\fP, char **\fIfirst\fP, char **\fIlast\fP);
85    .PP
86  .B int pcre_get_substring(const char *\fIsubject\fP, int *\fIovector\fP,  .B int pcre_get_substring(const char *\fIsubject\fP, int *\fIovector\fP,
87  .ti +5n  .ti +5n
88  .B int \fIstringcount\fP, int \fIstringnumber\fP,  .B int \fIstringcount\fP, int \fIstringnumber\fP,
89  .ti +5n  .ti +5n
90  .B const char **\fIstringptr\fP);  .B const char **\fIstringptr\fP);
91  .PP  .PP
 .br  
92  .B int pcre_get_substring_list(const char *\fIsubject\fP,  .B int pcre_get_substring_list(const char *\fIsubject\fP,
93  .ti +5n  .ti +5n
94  .B int *\fIovector\fP, int \fIstringcount\fP, "const char ***\fIlistptr\fP);"  .B int *\fIovector\fP, int \fIstringcount\fP, "const char ***\fIlistptr\fP);"
95  .PP  .PP
 .br  
96  .B void pcre_free_substring(const char *\fIstringptr\fP);  .B void pcre_free_substring(const char *\fIstringptr\fP);
97  .PP  .PP
 .br  
98  .B void pcre_free_substring_list(const char **\fIstringptr\fP);  .B void pcre_free_substring_list(const char **\fIstringptr\fP);
99  .PP  .PP
 .br  
100  .B const unsigned char *pcre_maketables(void);  .B const unsigned char *pcre_maketables(void);
101  .PP  .PP
 .br  
102  .B int pcre_fullinfo(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"  .B int pcre_fullinfo(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
103  .ti +5n  .ti +5n
104  .B int \fIwhat\fP, void *\fIwhere\fP);  .B int \fIwhat\fP, void *\fIwhere\fP);
105  .PP  .PP
 .br  
106  .B int pcre_info(const pcre *\fIcode\fP, int *\fIoptptr\fP, int  .B int pcre_info(const pcre *\fIcode\fP, int *\fIoptptr\fP, int
107  .B *\fIfirstcharptr\fP);  .B *\fIfirstcharptr\fP);
108  .PP  .PP
109  .br  .B int pcre_refcount(pcre *\fIcode\fP, int \fIadjust\fP);
110    .PP
111  .B int pcre_config(int \fIwhat\fP, void *\fIwhere\fP);  .B int pcre_config(int \fIwhat\fP, void *\fIwhere\fP);
112  .PP  .PP
 .br  
113  .B char *pcre_version(void);  .B char *pcre_version(void);
114  .PP  .
115  .br  .
116    .SH "PCRE NATIVE API INDIRECTED FUNCTIONS"
117    .rs
118    .sp
119  .B void *(*pcre_malloc)(size_t);  .B void *(*pcre_malloc)(size_t);
120  .PP  .PP
 .br  
121  .B void (*pcre_free)(void *);  .B void (*pcre_free)(void *);
122  .PP  .PP
 .br  
123  .B void *(*pcre_stack_malloc)(size_t);  .B void *(*pcre_stack_malloc)(size_t);
124  .PP  .PP
 .br  
125  .B void (*pcre_stack_free)(void *);  .B void (*pcre_stack_free)(void *);
126  .PP  .PP
 .br  
127  .B int (*pcre_callout)(pcre_callout_block *);  .B int (*pcre_callout)(pcre_callout_block *);
128  .  .
129  .  .
130  .SH "PCRE API OVERVIEW"  .SH "PCRE API OVERVIEW"
131  .rs  .rs
132  .sp  .sp
133  PCRE has its own native API, which is described in this document. There is also  PCRE has its own native API, which is described in this document. There are
134  a set of wrapper functions that correspond to the POSIX regular expression API.  also some wrapper functions that correspond to the POSIX regular expression
135  These are described in the  API, but they do not give access to all the functionality. They are described
136    in the
137  .\" HREF  .\" HREF
138  \fBpcreposix\fP  \fBpcreposix\fP
139  .\"  .\"
140  documentation.  documentation. Both of these APIs define a set of C function calls. A C++
141    wrapper is also distributed with PCRE. It is documented in the
142    .\" HREF
143    \fBpcrecpp\fP
144    .\"
145    page.
146  .P  .P
147  The native API function prototypes are defined in the header file \fBpcre.h\fP,  The native API C function prototypes are defined in the header file
148  and on Unix systems the library itself is called \fBlibpcre\fP. It can  \fBpcre.h\fP, and on Unix systems the library itself is called \fBlibpcre\fP.
149  normally be accessed by adding \fB-lpcre\fP to the command for linking an  It can normally be accessed by adding \fB-lpcre\fP to the command for linking
150  application that uses PCRE. The header file defines the macros PCRE_MAJOR and  an application that uses PCRE. The header file defines the macros PCRE_MAJOR
151  PCRE_MINOR to contain the major and minor release numbers for the library.  and PCRE_MINOR to contain the major and minor release numbers for the library.
152  Applications can use these to include support for different releases of PCRE.  Applications can use these to include support for different releases of PCRE.
153  .P  .P
154  The functions \fBpcre_compile()\fP, \fBpcre_study()\fP, and \fBpcre_exec()\fP  In a Windows environment, if you want to statically link an application program
155  are used for compiling and matching regular expressions. A sample program that  against a non-dll \fBpcre.a\fP file, you must define PCRE_STATIC before
156  demonstrates the simplest way of using them is provided in the file called  including \fBpcre.h\fP or \fBpcrecpp.h\fP, because otherwise the
157  \fIpcredemo.c\fP in the source distribution. The  \fBpcre_malloc()\fP and \fBpcre_free()\fP exported functions will be declared
158    \fB__declspec(dllimport)\fP, with unwanted results.
159    .P
160    The functions \fBpcre_compile()\fP, \fBpcre_compile2()\fP, \fBpcre_study()\fP,
161    and \fBpcre_exec()\fP are used for compiling and matching regular expressions
162    in a Perl-compatible manner. A sample program that demonstrates the simplest
163    way of using them is provided in the file called \fIpcredemo.c\fP in the PCRE
164    source distribution. A listing of this program is given in the
165    .\" HREF
166    \fBpcredemo\fP
167    .\"
168    documentation, and the
169  .\" HREF  .\" HREF
170  \fBpcresample\fP  \fBpcresample\fP
171  .\"  .\"
172  documentation describes how to run it.  documentation describes how to compile and run it.
173    .P
174    Just-in-time compiler support is an optional feature of PCRE that can be built
175    in appropriate hardware environments. It greatly speeds up the matching
176    performance of many patterns. Simple programs can easily request that it be
177    used if available, by setting an option that is ignored when it is not
178    relevant. More complicated programs might need to make use of the functions
179    \fBpcre_jit_stack_alloc()\fP, \fBpcre_jit_stack_free()\fP, and
180    \fBpcre_assign_jit_stack()\fP in order to control the JIT code's memory usage.
181    These functions are discussed in the
182    .\" HREF
183    \fBpcrejit\fP
184    .\"
185    documentation.
186    .P
187    A second matching function, \fBpcre_dfa_exec()\fP, which is not
188    Perl-compatible, is also provided. This uses a different algorithm for the
189    matching. The alternative algorithm finds all possible matches (at a given
190    point in the subject), and scans the subject just once (unless there are
191    lookbehind assertions). However, this algorithm does not return captured
192    substrings. A description of the two matching algorithms and their advantages
193    and disadvantages is given in the
194    .\" HREF
195    \fBpcrematching\fP
196    .\"
197    documentation.
198  .P  .P
199  In addition to the main compiling and matching functions, there are convenience  In addition to the main compiling and matching functions, there are convenience
200  functions for extracting captured substrings from a matched subject string.  functions for extracting captured substrings from a subject string that is
201  They are:  matched by \fBpcre_exec()\fP. They are:
202  .sp  .sp
203    \fBpcre_copy_substring()\fP    \fBpcre_copy_substring()\fP
204    \fBpcre_copy_named_substring()\fP    \fBpcre_copy_named_substring()\fP
# Line 145  They are: Line 206  They are:
206    \fBpcre_get_named_substring()\fP    \fBpcre_get_named_substring()\fP
207    \fBpcre_get_substring_list()\fP    \fBpcre_get_substring_list()\fP
208    \fBpcre_get_stringnumber()\fP    \fBpcre_get_stringnumber()\fP
209      \fBpcre_get_stringtable_entries()\fP
210  .sp  .sp
211  \fBpcre_free_substring()\fP and \fBpcre_free_substring_list()\fP are also  \fBpcre_free_substring()\fP and \fBpcre_free_substring_list()\fP are also
212  provided, to free the memory used for extracted strings.  provided, to free the memory used for extracted strings.
213  .P  .P
214  The function \fBpcre_maketables()\fP is used to build a set of character tables  The function \fBpcre_maketables()\fP is used to build a set of character tables
215  in the current locale for passing to \fBpcre_compile()\fP or \fBpcre_exec()\fP.  in the current locale for passing to \fBpcre_compile()\fP, \fBpcre_exec()\fP,
216  This is an optional facility that is provided for specialist use. Most  or \fBpcre_dfa_exec()\fP. This is an optional facility that is provided for
217  commonly, no special tables are passed, in which case internal tables that are  specialist use. Most commonly, no special tables are passed, in which case
218  generated when PCRE is built are used.  internal tables that are generated when PCRE is built are used.
219  .P  .P
220  The function \fBpcre_fullinfo()\fP is used to find out information about a  The function \fBpcre_fullinfo()\fP is used to find out information about a
221  compiled pattern; \fBpcre_info()\fP is an obsolete version that returns only  compiled pattern; \fBpcre_info()\fP is an obsolete version that returns only
# Line 161  some of the available information, but i Line 223  some of the available information, but i
223  The function \fBpcre_version()\fP returns a pointer to a string containing the  The function \fBpcre_version()\fP returns a pointer to a string containing the
224  version of PCRE and its date of release.  version of PCRE and its date of release.
225  .P  .P
226    The function \fBpcre_refcount()\fP maintains a reference count in a data block
227    containing a compiled pattern. This is provided for the benefit of
228    object-oriented applications.
229    .P
230  The global variables \fBpcre_malloc\fP and \fBpcre_free\fP initially contain  The global variables \fBpcre_malloc\fP and \fBpcre_free\fP initially contain
231  the entry points of the standard \fBmalloc()\fP and \fBfree()\fP functions,  the entry points of the standard \fBmalloc()\fP and \fBfree()\fP functions,
232  respectively. PCRE calls the memory management functions via these variables,  respectively. PCRE calls the memory management functions via these variables,
# Line 170  should be done before calling any PCRE f Line 236  should be done before calling any PCRE f
236  The global variables \fBpcre_stack_malloc\fP and \fBpcre_stack_free\fP are also  The global variables \fBpcre_stack_malloc\fP and \fBpcre_stack_free\fP are also
237  indirections to memory management functions. These special functions are used  indirections to memory management functions. These special functions are used
238  only when PCRE is compiled to use the heap for remembering data, instead of  only when PCRE is compiled to use the heap for remembering data, instead of
239  recursive function calls. This is a non-standard way of building PCRE, for use  recursive function calls, when running the \fBpcre_exec()\fP function. See the
240  in environments that have limited stacks. Because of the greater use of memory  .\" HREF
241  management, it runs more slowly. Separate functions are provided so that  \fBpcrebuild\fP
242  special-purpose external code can be used for this case. When used, these  .\"
243  functions are always called in a stack-like manner (last obtained, first  documentation for details of how to do this. It is a non-standard way of
244  freed), and always for memory blocks of the same size.  building PCRE, for use in environments that have limited stacks. Because of the
245    greater use of memory management, it runs more slowly. Separate functions are
246    provided so that special-purpose external code can be used for this case. When
247    used, these functions are always called in a stack-like manner (last obtained,
248    first freed), and always for memory blocks of the same size. There is a
249    discussion about PCRE's stack usage in the
250    .\" HREF
251    \fBpcrestack\fP
252    .\"
253    documentation.
254  .P  .P
255  The global variable \fBpcre_callout\fP initially contains NULL. It can be set  The global variable \fBpcre_callout\fP initially contains NULL. It can be set
256  by the caller to a "callout" function, which PCRE will then call at specified  by the caller to a "callout" function, which PCRE will then call at specified
# Line 186  points during a matching operation. Deta Line 261  points during a matching operation. Deta
261  documentation.  documentation.
262  .  .
263  .  .
264    .\" HTML <a name="newlines"></a>
265    .SH NEWLINES
266    .rs
267    .sp
268    PCRE supports five different conventions for indicating line breaks in
269    strings: a single CR (carriage return) character, a single LF (linefeed)
270    character, the two-character sequence CRLF, any of the three preceding, or any
271    Unicode newline sequence. The Unicode newline sequences are the three just
272    mentioned, plus the single characters VT (vertical tab, U+000B), FF (formfeed,
273    U+000C), NEL (next line, U+0085), LS (line separator, U+2028), and PS
274    (paragraph separator, U+2029).
275    .P
276    Each of the first three conventions is used by at least one operating system as
277    its standard newline sequence. When PCRE is built, a default can be specified.
278    The default default is LF, which is the Unix standard. When PCRE is run, the
279    default can be overridden, either when a pattern is compiled, or when it is
280    matched.
281    .P
282    At compile time, the newline convention can be specified by the \fIoptions\fP
283    argument of \fBpcre_compile()\fP, or it can be specified by special text at the
284    start of the pattern itself; this overrides any other settings. See the
285    .\" HREF
286    \fBpcrepattern\fP
287    .\"
288    page for details of the special character sequences.
289    .P
290    In the PCRE documentation the word "newline" is used to mean "the character or
291    pair of characters that indicate a line break". The choice of newline
292    convention affects the handling of the dot, circumflex, and dollar
293    metacharacters, the handling of #-comments in /x mode, and, when CRLF is a
294    recognized line ending sequence, the match position advancement for a
295    non-anchored pattern. There is more detail about this in the
296    .\" HTML <a href="#execoptions">
297    .\" </a>
298    section on \fBpcre_exec()\fP options
299    .\"
300    below.
301    .P
302    The choice of newline convention does not affect the interpretation of
303    the \en or \er escape sequences, nor does it affect what \eR matches, which is
304    controlled in a similar way, but by separate options.
305    .
306    .
307  .SH MULTITHREADING  .SH MULTITHREADING
308  .rs  .rs
309  .sp  .sp
# Line 196  callout function pointed to by \fBpcre_c Line 314  callout function pointed to by \fBpcre_c
314  .P  .P
315  The compiled form of a regular expression is not altered during matching, so  The compiled form of a regular expression is not altered during matching, so
316  the same compiled pattern can safely be used by several threads at once.  the same compiled pattern can safely be used by several threads at once.
317    .P
318    If the just-in-time optimization feature is being used, it needs separate
319    memory stack areas for each thread. See the
320    .\" HREF
321    \fBpcrejit\fP
322    .\"
323    documentation for more details.
324  .  .
325  .  .
326  .SH "SAVING PRECOMPILED PATTERNS FOR LATER USE"  .SH "SAVING PRECOMPILED PATTERNS FOR LATER USE"
# Line 207  which it was compiled. Details are given Line 332  which it was compiled. Details are given
332  .\" HREF  .\" HREF
333  \fBpcreprecompile\fP  \fBpcreprecompile\fP
334  .\"  .\"
335  documentation.  documentation. However, compiling a regular expression with one version of PCRE
336    for use with a different version is not guaranteed to work and may cause
337    crashes.
338  .  .
339  .  .
340  .SH "CHECKING BUILD-TIME OPTIONS"  .SH "CHECKING BUILD-TIME OPTIONS"
# Line 236  otherwise it is set to zero. Line 363  otherwise it is set to zero.
363  The output is an integer that is set to one if support for Unicode character  The output is an integer that is set to one if support for Unicode character
364  properties is available; otherwise it is set to zero.  properties is available; otherwise it is set to zero.
365  .sp  .sp
366      PCRE_CONFIG_JIT
367    .sp
368    The output is an integer that is set to one if support for just-in-time
369    compiling is available; otherwise it is set to zero.
370    .sp
371    PCRE_CONFIG_NEWLINE    PCRE_CONFIG_NEWLINE
372  .sp  .sp
373  The output is an integer that is set to the value of the code that is used for  The output is an integer whose value specifies the default character sequence
374  the newline character. It is either linefeed (10) or carriage return (13), and  that is recognized as meaning "newline". The four values that are supported
375  should normally be the standard character for your operating system.  are: 10 for LF, 13 for CR, 3338 for CRLF, -2 for ANYCRLF, and -1 for ANY.
376    Though they are derived from ASCII, the same values are returned in EBCDIC
377    environments. The default should normally correspond to the standard sequence
378    for your operating system.
379    .sp
380      PCRE_CONFIG_BSR
381    .sp
382    The output is an integer whose value indicates what character sequences the \eR
383    escape sequence matches by default. A value of 0 means that \eR matches any
384    Unicode line ending sequence; a value of 1 means that \eR matches only CR, LF,
385    or CRLF. The default can be overridden when a pattern is compiled or matched.
386  .sp  .sp
387    PCRE_CONFIG_LINK_SIZE    PCRE_CONFIG_LINK_SIZE
388  .sp  .sp
# Line 262  documentation. Line 404  documentation.
404  .sp  .sp
405    PCRE_CONFIG_MATCH_LIMIT    PCRE_CONFIG_MATCH_LIMIT
406  .sp  .sp
407  The output is an integer that gives the default limit for the number of  The output is a long integer that gives the default limit for the number of
408  internal matching function calls in a \fBpcre_exec()\fP execution. Further  internal matching function calls in a \fBpcre_exec()\fP execution. Further
409  details are given with \fBpcre_exec()\fP below.  details are given with \fBpcre_exec()\fP below.
410  .sp  .sp
411      PCRE_CONFIG_MATCH_LIMIT_RECURSION
412    .sp
413    The output is a long integer that gives the default limit for the depth of
414    recursion when calling the internal matching function in a \fBpcre_exec()\fP
415    execution. Further details are given with \fBpcre_exec()\fP below.
416    .sp
417    PCRE_CONFIG_STACKRECURSE    PCRE_CONFIG_STACKRECURSE
418  .sp  .sp
419  The output is an integer that is set to one if internal recursion is  The output is an integer that is set to one if internal recursion when running
420  implemented by recursive function calls that use the stack to remember their  \fBpcre_exec()\fP is implemented by recursive function calls that use the stack
421  state. This is the usual way that PCRE is compiled. The output is zero if PCRE  to remember their state. This is the usual way that PCRE is compiled. The
422  was compiled to use blocks of data on the heap instead of recursive function  output is zero if PCRE was compiled to use blocks of data on the heap instead
423  calls. In this case, \fBpcre_stack_malloc\fP and \fBpcre_stack_free\fP are  of recursive function calls. In this case, \fBpcre_stack_malloc\fP and
424  called to manage memory blocks on the heap, thus avoiding the use of the stack.  \fBpcre_stack_free\fP are called to manage memory blocks on the heap, thus
425    avoiding the use of the stack.
426  .  .
427  .  .
428  .SH "COMPILING A PATTERN"  .SH "COMPILING A PATTERN"
# Line 284  called to manage memory blocks on the he Line 433  called to manage memory blocks on the he
433  .B const char **\fIerrptr\fP, int *\fIerroffset\fP,  .B const char **\fIerrptr\fP, int *\fIerroffset\fP,
434  .ti +5n  .ti +5n
435  .B const unsigned char *\fItableptr\fP);  .B const unsigned char *\fItableptr\fP);
436    .sp
437    .B pcre *pcre_compile2(const char *\fIpattern\fP, int \fIoptions\fP,
438    .ti +5n
439    .B int *\fIerrorcodeptr\fP,
440    .ti +5n
441    .B const char **\fIerrptr\fP, int *\fIerroffset\fP,
442    .ti +5n
443    .B const unsigned char *\fItableptr\fP);
444  .P  .P
445  The function \fBpcre_compile()\fP is called to compile a pattern into an  Either of the functions \fBpcre_compile()\fP or \fBpcre_compile2()\fP can be
446  internal form. The pattern is a C string terminated by a binary zero, and  called to compile a pattern into an internal form. The only difference between
447  is passed in the \fIpattern\fP argument. A pointer to a single block of memory  the two interfaces is that \fBpcre_compile2()\fP has an additional argument,
448  that is obtained via \fBpcre_malloc\fP is returned. This contains the compiled  \fIerrorcodeptr\fP, via which a numerical error code can be returned. To avoid
449  code and related data. The \fBpcre\fP type is defined for the returned block;  too much repetition, we refer just to \fBpcre_compile()\fP below, but the
450  this is a typedef for a structure whose contents are not externally defined. It  information applies equally to \fBpcre_compile2()\fP.
451  is up to the caller to free the memory when it is no longer required.  .P
452    The pattern is a C string terminated by a binary zero, and is passed in the
453    \fIpattern\fP argument. A pointer to a single block of memory that is obtained
454    via \fBpcre_malloc\fP is returned. This contains the compiled code and related
455    data. The \fBpcre\fP type is defined for the returned block; this is a typedef
456    for a structure whose contents are not externally defined. It is up to the
457    caller to free the memory (via \fBpcre_free\fP) when it is no longer required.
458  .P  .P
459  Although the compiled code of a PCRE regex is relocatable, that is, it does not  Although the compiled code of a PCRE regex is relocatable, that is, it does not
460  depend on memory location, the complete \fBpcre\fP data block is not  depend on memory location, the complete \fBpcre\fP data block is not
461  fully relocatable, because it may contain a copy of the \fItableptr\fP  fully relocatable, because it may contain a copy of the \fItableptr\fP
462  argument, which is an address (see below).  argument, which is an address (see below).
463  .P  .P
464  The \fIoptions\fP argument contains independent bits that affect the  The \fIoptions\fP argument contains various bit settings that affect the
465  compilation. It should be zero if no options are required. The available  compilation. It should be zero if no options are required. The available
466  options are described below. Some of them, in particular, those that are  options are described below. Some of them (in particular, those that are
467  compatible with Perl, can also be set and unset from within the pattern (see  compatible with Perl, but some others as well) can also be set and unset from
468  the detailed description in the  within the pattern (see the detailed description in the
469  .\" HREF  .\" HREF
470  \fBpcrepattern\fP  \fBpcrepattern\fP
471  .\"  .\"
472  documentation). For these options, the contents of the \fIoptions\fP argument  documentation). For those options that can be different in different parts of
473  specifies their initial settings at the start of compilation and execution. The  the pattern, the contents of the \fIoptions\fP argument specifies their
474  PCRE_ANCHORED option can be set at the time of matching as well as at compile  settings at the start of compilation and execution. The PCRE_ANCHORED,
475  time.  PCRE_BSR_\fIxxx\fP, PCRE_NEWLINE_\fIxxx\fP, PCRE_NO_UTF8_CHECK, and
476    PCRE_NO_START_OPT options can be set at the time of matching as well as at
477    compile time.
478  .P  .P
479  If \fIerrptr\fP is NULL, \fBpcre_compile()\fP returns NULL immediately.  If \fIerrptr\fP is NULL, \fBpcre_compile()\fP returns NULL immediately.
480  Otherwise, if compilation of a pattern fails, \fBpcre_compile()\fP returns  Otherwise, if compilation of a pattern fails, \fBpcre_compile()\fP returns
481  NULL, and sets the variable pointed to by \fIerrptr\fP to point to a textual  NULL, and sets the variable pointed to by \fIerrptr\fP to point to a textual
482  error message. The offset from the start of the pattern to the character where  error message. This is a static string that is part of the library. You must
483  the error was discovered is placed in the variable pointed to by  not try to free it. Normally, the offset from the start of the pattern to the
484  \fIerroffset\fP, which must not be NULL. If it is, an immediate error is given.  byte that was being processed when the error was discovered is placed in the
485    variable pointed to by \fIerroffset\fP, which must not be NULL (if it is, an
486    immediate error is given). However, for an invalid UTF-8 string, the offset is
487    that of the first byte of the failing character. Also, some errors are not
488    detected until checks are carried out when the whole pattern has been scanned;
489    in these cases the offset passed back is the length of the pattern.
490    .P
491    Note that the offset is in bytes, not characters, even in UTF-8 mode. It may
492    sometimes point into the middle of a UTF-8 character.
493    .P
494    If \fBpcre_compile2()\fP is used instead of \fBpcre_compile()\fP, and the
495    \fIerrorcodeptr\fP argument is not NULL, a non-zero error code number is
496    returned via this argument in the event of an error. This is in addition to the
497    textual error message. Error codes and messages are listed below.
498  .P  .P
499  If the final argument, \fItableptr\fP, is NULL, PCRE uses a default set of  If the final argument, \fItableptr\fP, is NULL, PCRE uses a default set of
500  character tables that are built when PCRE is compiled, using the default C  character tables that are built when PCRE is compiled, using the default C
# Line 358  facility, see the Line 536  facility, see the
536  .\"  .\"
537  documentation.  documentation.
538  .sp  .sp
539      PCRE_BSR_ANYCRLF
540      PCRE_BSR_UNICODE
541    .sp
542    These options (which are mutually exclusive) control what the \eR escape
543    sequence matches. The choice is either to match only CR, LF, or CRLF, or to
544    match any Unicode newline sequence. The default is specified when PCRE is
545    built. It can be overridden from within the pattern, or by setting an option
546    when a compiled pattern is matched.
547    .sp
548    PCRE_CASELESS    PCRE_CASELESS
549  .sp  .sp
550  If this bit is set, letters in the pattern match both upper and lower case  If this bit is set, letters in the pattern match both upper and lower case
551  letters. It is equivalent to Perl's /i option, and it can be changed within a  letters. It is equivalent to Perl's /i option, and it can be changed within a
552  pattern by a (?i) option setting. When running in UTF-8 mode, case support for  pattern by a (?i) option setting. In UTF-8 mode, PCRE always understands the
553  high-valued characters is available only when PCRE is built with Unicode  concept of case for characters whose values are less than 128, so caseless
554  character property support.  matching is always possible. For characters with higher values, the concept of
555    case is supported if PCRE is compiled with Unicode property support, but not
556    otherwise. If you want to use caseless matching for characters 128 and above,
557    you must ensure that PCRE is compiled with Unicode property support as well as
558    with UTF-8 support.
559  .sp  .sp
560    PCRE_DOLLAR_ENDONLY    PCRE_DOLLAR_ENDONLY
561  .sp  .sp
562  If this bit is set, a dollar metacharacter in the pattern matches only at the  If this bit is set, a dollar metacharacter in the pattern matches only at the
563  end of the subject string. Without this option, a dollar also matches  end of the subject string. Without this option, a dollar also matches
564  immediately before the final character if it is a newline (but not before any  immediately before a newline at the end of the string (but not before any other
565  other newlines). The PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is  newlines). The PCRE_DOLLAR_ENDONLY option is ignored if PCRE_MULTILINE is set.
566  set. There is no equivalent to this option in Perl, and no way to set it within  There is no equivalent to this option in Perl, and no way to set it within a
567  a pattern.  pattern.
568  .sp  .sp
569    PCRE_DOTALL    PCRE_DOTALL
570  .sp  .sp
571  If this bit is set, a dot metacharater in the pattern matches all characters,  If this bit is set, a dot metacharacter in the pattern matches a character of
572  including newlines. Without it, newlines are excluded. This option is  any value, including one that indicates a newline. However, it only ever
573    matches one character, even if newlines are coded as CRLF. Without this option,
574    a dot does not match when the current position is at a newline. This option is
575  equivalent to Perl's /s option, and it can be changed within a pattern by a  equivalent to Perl's /s option, and it can be changed within a pattern by a
576  (?s) option setting. A negative class such as [^a] always matches a newline  (?s) option setting. A negative class such as [^a] always matches newline
577  character, independent of the setting of this option.  characters, independent of the setting of this option.
578    .sp
579      PCRE_DUPNAMES
580    .sp
581    If this bit is set, names used to identify capturing subpatterns need not be
582    unique. This can be helpful for certain types of pattern when it is known that
583    only one instance of the named subpattern can ever be matched. There are more
584    details of named subpatterns below; see also the
585    .\" HREF
586    \fBpcrepattern\fP
587    .\"
588    documentation.
589  .sp  .sp
590    PCRE_EXTENDED    PCRE_EXTENDED
591  .sp  .sp
592  If this bit is set, whitespace data characters in the pattern are totally  If this bit is set, whitespace data characters in the pattern are totally
593  ignored except when escaped or inside a character class. Whitespace does not  ignored except when escaped or inside a character class. Whitespace does not
594  include the VT character (code 11). In addition, characters between an  include the VT character (code 11). In addition, characters between an
595  unescaped # outside a character class and the next newline character,  unescaped # outside a character class and the next newline, inclusive, are also
596  inclusive, are also ignored. This is equivalent to Perl's /x option, and it can  ignored. This is equivalent to Perl's /x option, and it can be changed within a
597  be changed within a pattern by a (?x) option setting.  pattern by a (?x) option setting.
598    .P
599    Which characters are interpreted as newlines is controlled by the options
600    passed to \fBpcre_compile()\fP or by a special sequence at the start of the
601    pattern, as described in the section entitled
602    .\" HTML <a href="pcrepattern.html#newlines">
603    .\" </a>
604    "Newline conventions"
605    .\"
606    in the \fBpcrepattern\fP documentation. Note that the end of this type of
607    comment is a literal newline sequence in the pattern; escape sequences that
608    happen to represent a newline do not count.
609  .P  .P
610  This option makes it possible to include comments inside complicated patterns.  This option makes it possible to include comments inside complicated patterns.
611  Note, however, that this applies only to data characters. Whitespace characters  Note, however, that this applies only to data characters. Whitespace characters
612  may never appear within special character sequences in a pattern, for example  may never appear within special character sequences in a pattern, for example
613  within the sequence (?( which introduces a conditional subpattern.  within the sequence (?( that introduces a conditional subpattern.
614  .sp  .sp
615    PCRE_EXTRA    PCRE_EXTRA
616  .sp  .sp
# Line 404  that is incompatible with Perl, but it i Line 619  that is incompatible with Perl, but it i
619  set, any backslash in a pattern that is followed by a letter that has no  set, any backslash in a pattern that is followed by a letter that has no
620  special meaning causes an error, thus reserving these combinations for future  special meaning causes an error, thus reserving these combinations for future
621  expansion. By default, as in Perl, a backslash followed by a letter with no  expansion. By default, as in Perl, a backslash followed by a letter with no
622  special meaning is treated as a literal. There are at present no other features  special meaning is treated as a literal. (Perl can, however, be persuaded to
623  controlled by this option. It can also be set by a (?X) option setting within a  give an error for this, by running it with the -w option.) There are at present
624  pattern.  no other features controlled by this option. It can also be set by a (?X)
625    option setting within a pattern.
626    .sp
627      PCRE_FIRSTLINE
628    .sp
629    If this option is set, an unanchored pattern is required to match before or at
630    the first newline in the subject string, though the matched text may continue
631    over the newline.
632    .sp
633      PCRE_JAVASCRIPT_COMPAT
634    .sp
635    If this option is set, PCRE's behaviour is changed in some ways so that it is
636    compatible with JavaScript rather than Perl. The changes are as follows:
637    .P
638    (1) A lone closing square bracket in a pattern causes a compile-time error,
639    because this is illegal in JavaScript (by default it is treated as a data
640    character). Thus, the pattern AB]CD becomes illegal when this option is set.
641    .P
642    (2) At run time, a back reference to an unset subpattern group matches an empty
643    string (by default this causes the current matching alternative to fail). A
644    pattern such as (\e1)(a) succeeds when this option is set (assuming it can find
645    an "a" in the subject), whereas it fails by default, for Perl compatibility.
646    .P
647    (3) \eU matches an upper case "U" character; by default \eU causes a compile
648    time error (Perl uses \eU to upper case subsequent characters).
649    .P
650    (4) \eu matches a lower case "u" character unless it is followed by four
651    hexadecimal digits, in which case the hexadecimal number defines the code point
652    to match. By default, \eu causes a compile time error (Perl uses it to upper
653    case the following character).
654    .P
655    (5) \ex matches a lower case "x" character unless it is followed by two
656    hexadecimal digits, in which case the hexadecimal number defines the code point
657    to match. By default, as in Perl, a hexadecimal number is always expected after
658    \ex, but it may have zero, one, or two digits (so, for example, \exz matches a
659    binary zero character followed by z).
660  .sp  .sp
661    PCRE_MULTILINE    PCRE_MULTILINE
662  .sp  .sp
# Line 418  terminating newline (unless PCRE_DOLLAR_ Line 668  terminating newline (unless PCRE_DOLLAR_
668  Perl.  Perl.
669  .P  .P
670  When PCRE_MULTILINE it is set, the "start of line" and "end of line" constructs  When PCRE_MULTILINE it is set, the "start of line" and "end of line" constructs
671  match immediately following or immediately before any newline in the subject  match immediately following or immediately before internal newlines in the
672  string, respectively, as well as at the very start and end. This is equivalent  subject string, respectively, as well as at the very start and end. This is
673  to Perl's /m option, and it can be changed within a pattern by a (?m) option  equivalent to Perl's /m option, and it can be changed within a pattern by a
674  setting. If there are no "\en" characters in a subject string, or no  (?m) option setting. If there are no newlines in a subject string, or no
675  occurrences of ^ or $ in a pattern, setting PCRE_MULTILINE has no effect.  occurrences of ^ or $ in a pattern, setting PCRE_MULTILINE has no effect.
676  .sp  .sp
677      PCRE_NEWLINE_CR
678      PCRE_NEWLINE_LF
679      PCRE_NEWLINE_CRLF
680      PCRE_NEWLINE_ANYCRLF
681      PCRE_NEWLINE_ANY
682    .sp
683    These options override the default newline definition that was chosen when PCRE
684    was built. Setting the first or the second specifies that a newline is
685    indicated by a single character (CR or LF, respectively). Setting
686    PCRE_NEWLINE_CRLF specifies that a newline is indicated by the two-character
687    CRLF sequence. Setting PCRE_NEWLINE_ANYCRLF specifies that any of the three
688    preceding sequences should be recognized. Setting PCRE_NEWLINE_ANY specifies
689    that any Unicode newline sequence should be recognized. The Unicode newline
690    sequences are the three just mentioned, plus the single characters VT (vertical
691    tab, U+000B), FF (formfeed, U+000C), NEL (next line, U+0085), LS (line
692    separator, U+2028), and PS (paragraph separator, U+2029). The last two are
693    recognized only in UTF-8 mode.
694    .P
695    The newline setting in the options word uses three bits that are treated
696    as a number, giving eight possibilities. Currently only six are used (default
697    plus the five values above). This means that if you set more than one newline
698    option, the combination may or may not be sensible. For example,
699    PCRE_NEWLINE_CR with PCRE_NEWLINE_LF is equivalent to PCRE_NEWLINE_CRLF, but
700    other combinations may yield unused numbers and cause an error.
701    .P
702    The only time that a line break in a pattern is specially recognized when
703    compiling is when PCRE_EXTENDED is set. CR and LF are whitespace characters,
704    and so are ignored in this mode. Also, an unescaped # outside a character class
705    indicates a comment that lasts until after the next line break sequence. In
706    other circumstances, line break sequences in patterns are treated as literal
707    data.
708    .P
709    The newline option that is set at compile time becomes the default that is used
710    for \fBpcre_exec()\fP and \fBpcre_dfa_exec()\fP, but it can be overridden.
711    .sp
712    PCRE_NO_AUTO_CAPTURE    PCRE_NO_AUTO_CAPTURE
713  .sp  .sp
714  If this option is set, it disables the use of numbered capturing parentheses in  If this option is set, it disables the use of numbered capturing parentheses in
# Line 432  were followed by ?: but named parenthese Line 717  were followed by ?: but named parenthese
717  they acquire numbers in the usual way). There is no equivalent of this option  they acquire numbers in the usual way). There is no equivalent of this option
718  in Perl.  in Perl.
719  .sp  .sp
720      NO_START_OPTIMIZE
721    .sp
722    This is an option that acts at matching time; that is, it is really an option
723    for \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP. If it is set at compile time,
724    it is remembered with the compiled pattern and assumed at matching time. For
725    details see the discussion of PCRE_NO_START_OPTIMIZE
726    .\" HTML <a href="#execoptions">
727    .\" </a>
728    below.
729    .\"
730    .sp
731      PCRE_UCP
732    .sp
733    This option changes the way PCRE processes \eB, \eb, \eD, \ed, \eS, \es, \eW,
734    \ew, and some of the POSIX character classes. By default, only ASCII characters
735    are recognized, but if PCRE_UCP is set, Unicode properties are used instead to
736    classify characters. More details are given in the section on
737    .\" HTML <a href="pcre.html#genericchartypes">
738    .\" </a>
739    generic character types
740    .\"
741    in the
742    .\" HREF
743    \fBpcrepattern\fP
744    .\"
745    page. If you set PCRE_UCP, matching one of the items it affects takes much
746    longer. The option is available only if PCRE has been compiled with Unicode
747    property support.
748    .sp
749    PCRE_UNGREEDY    PCRE_UNGREEDY
750  .sp  .sp
751  This option inverts the "greediness" of the quantifiers so that they are not  This option inverts the "greediness" of the quantifiers so that they are not
# Line 445  of UTF-8 characters instead of single-by Line 759  of UTF-8 characters instead of single-by
759  available only when PCRE is built to include UTF-8 support. If not, the use  available only when PCRE is built to include UTF-8 support. If not, the use
760  of this option provokes an error. Details of how this option changes the  of this option provokes an error. Details of how this option changes the
761  behaviour of PCRE are given in the  behaviour of PCRE are given in the
762  .\" HTML <a href="pcre.html#utf8support">  .\" HREF
763    \fBpcreunicode\fP
764    .\"
765    page.
766    .sp
767      PCRE_NO_UTF8_CHECK
768    .sp
769    When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is
770    automatically checked. There is a discussion about the
771    .\" HTML <a href="pcre.html#utf8strings">
772  .\" </a>  .\" </a>
773  section on UTF-8 support  validity of UTF-8 strings
774  .\"  .\"
775  in the main  in the main
776  .\" HREF  .\" HREF
777  \fBpcre\fP  \fBpcre\fP
778  .\"  .\"
779  page.  page. If an invalid UTF-8 sequence of bytes is found, \fBpcre_compile()\fP
780    returns an error. If you already know that your pattern is valid, and you want
781    to skip this check for performance reasons, you can set the PCRE_NO_UTF8_CHECK
782    option. When it is set, the effect of passing an invalid UTF-8 string as a
783    pattern is undefined. It may cause your program to crash. Note that this option
784    can also be passed to \fBpcre_exec()\fP and \fBpcre_dfa_exec()\fP, to suppress
785    the UTF-8 validity checking of subject strings.
786    .
787    .
788    .SH "COMPILATION ERROR CODES"
789    .rs
790  .sp  .sp
791    PCRE_NO_UTF8_CHECK  The following table lists the error codes than may be returned by
792    \fBpcre_compile2()\fP, along with the error messages that may be returned by
793    both compiling functions. As PCRE has developed, some error codes have fallen
794    out of use. To avoid confusion, they have not been re-used.
795    .sp
796       0  no error
797       1  \e at end of pattern
798       2  \ec at end of pattern
799       3  unrecognized character follows \e
800       4  numbers out of order in {} quantifier
801       5  number too big in {} quantifier
802       6  missing terminating ] for character class
803       7  invalid escape sequence in character class
804       8  range out of order in character class
805       9  nothing to repeat
806      10  [this code is not in use]
807      11  internal error: unexpected repeat
808      12  unrecognized character after (? or (?-
809      13  POSIX named classes are supported only within a class
810      14  missing )
811      15  reference to non-existent subpattern
812      16  erroffset passed as NULL
813      17  unknown option bit(s) set
814      18  missing ) after comment
815      19  [this code is not in use]
816      20  regular expression is too large
817      21  failed to get memory
818      22  unmatched parentheses
819      23  internal error: code overflow
820      24  unrecognized character after (?<
821      25  lookbehind assertion is not fixed length
822      26  malformed number or name after (?(
823      27  conditional group contains more than two branches
824      28  assertion expected after (?(
825      29  (?R or (?[+-]digits must be followed by )
826      30  unknown POSIX class name
827      31  POSIX collating elements are not supported
828      32  this version of PCRE is not compiled with PCRE_UTF8 support
829      33  [this code is not in use]
830      34  character value in \ex{...} sequence is too large
831      35  invalid condition (?(0)
832      36  \eC not allowed in lookbehind assertion
833      37  PCRE does not support \eL, \el, \eN{name}, \eU, or \eu
834      38  number after (?C is > 255
835      39  closing ) for (?C expected
836      40  recursive call could loop indefinitely
837      41  unrecognized character after (?P
838      42  syntax error in subpattern name (missing terminator)
839      43  two named subpatterns have the same name
840      44  invalid UTF-8 string
841      45  support for \eP, \ep, and \eX has not been compiled
842      46  malformed \eP or \ep sequence
843      47  unknown property name after \eP or \ep
844      48  subpattern name is too long (maximum 32 characters)
845      49  too many named subpatterns (maximum 10000)
846      50  [this code is not in use]
847      51  octal value is greater than \e377 (not in UTF-8 mode)
848      52  internal error: overran compiling workspace
849      53  internal error: previously-checked referenced subpattern
850            not found
851      54  DEFINE group contains more than one branch
852      55  repeating a DEFINE group is not allowed
853      56  inconsistent NEWLINE options
854      57  \eg is not followed by a braced, angle-bracketed, or quoted
855            name/number or by a plain number
856      58  a numbered reference must not be zero
857      59  an argument is not allowed for (*ACCEPT), (*FAIL), or (*COMMIT)
858      60  (*VERB) not recognized
859      61  number is too big
860      62  subpattern name expected
861      63  digit expected after (?+
862      64  ] is an invalid data character in JavaScript compatibility mode
863      65  different names for subpatterns of the same number are
864            not allowed
865      66  (*MARK) must have an argument
866      67  this version of PCRE is not compiled with PCRE_UCP support
867      68  \ec must be followed by an ASCII character
868      69  \ek is not followed by a braced, angle-bracketed, or quoted name
869  .sp  .sp
870  When PCRE_UTF8 is set, the validity of the pattern as a UTF-8 string is  The numbers 32 and 10000 in errors 48 and 49 are defaults; different values may
871  automatically checked. If an invalid UTF-8 sequence of bytes is found,  be used if the limits were changed when PCRE was built.
 \fBpcre_compile()\fP returns an error. If you already know that your pattern is  
 valid, and you want to skip this check for performance reasons, you can set the  
 PCRE_NO_UTF8_CHECK option. When it is set, the effect of passing an invalid  
 UTF-8 string as a pattern is undefined. It may cause your program to crash.  
 Note that this option can also be passed to \fBpcre_exec()\fP, to suppress the  
 UTF-8 validity checking of subject strings.  
872  .  .
873  .  .
874    .\" HTML <a name="studyingapattern"></a>
875  .SH "STUDYING A PATTERN"  .SH "STUDYING A PATTERN"
876  .rs  .rs
877  .sp  .sp
878  .B pcre_extra *pcre_study(const pcre *\fIcode\fP, int \fIoptions\fP,  .B pcre_extra *pcre_study(const pcre *\fIcode\fP, int \fIoptions\fP
879  .ti +5n  .ti +5n
880  .B const char **\fIerrptr\fP);  .B const char **\fIerrptr\fP);
881  .PP  .PP
# Line 483  help speed up matching, \fBpcre_study()\ Line 888  help speed up matching, \fBpcre_study()\
888  results of the study.  results of the study.
889  .P  .P
890  The returned value from \fBpcre_study()\fP can be passed directly to  The returned value from \fBpcre_study()\fP can be passed directly to
891  \fBpcre_exec()\fP. However, a \fBpcre_extra\fP block also contains other  \fBpcre_exec()\fP or \fBpcre_dfa_exec()\fP. However, a \fBpcre_extra\fP block
892  fields that can be set by the caller before the block is passed; these are  also contains other fields that can be set by the caller before the block is
893  described  passed; these are described
894  .\" HTML <a href="#extradata">  .\" HTML <a href="#extradata">
895  .\" </a>  .\" </a>
896  below  below
897  .\"  .\"
898  in the section on matching a pattern.  in the section on matching a pattern.
899  .P  .P
900  If studying the pattern does not produce any additional information,  If studying the pattern does not produce any useful information,
901  \fBpcre_study()\fP returns NULL. In that circumstance, if the calling program  \fBpcre_study()\fP returns NULL. In that circumstance, if the calling program
902  wants to pass any of the other fields to \fBpcre_exec()\fP, it must set up its  wants to pass any of the other fields to \fBpcre_exec()\fP or
903  own \fBpcre_extra\fP block.  \fBpcre_dfa_exec()\fP, it must set up its own \fBpcre_extra\fP block.
904  .P  .P
905  The second argument of \fBpcre_study()\fP contains option bits. At present, no  The second argument of \fBpcre_study()\fP contains option bits. There is only
906  options are defined, and this argument should always be zero.  one option: PCRE_STUDY_JIT_COMPILE. If this is set, and the just-in-time
907    compiler is available, the pattern is further compiled into machine code that
908    executes much faster than the \fBpcre_exec()\fP matching function. If
909    the just-in-time compiler is not available, this option is ignored. All other
910    bits in the \fIoptions\fP argument must be zero.
911    .P
912    JIT compilation is a heavyweight optimization. It can take some time for
913    patterns to be analyzed, and for one-off matches and simple patterns the
914    benefit of faster execution might be offset by a much slower study time.
915    Not all patterns can be optimized by the JIT compiler. For those that cannot be
916    handled, matching automatically falls back to the \fBpcre_exec()\fP
917    interpreter. For more details, see the
918    .\" HREF
919    \fBpcrejit\fP
920    .\"
921    documentation.
922  .P  .P
923  The third argument for \fBpcre_study()\fP is a pointer for an error message. If  The third argument for \fBpcre_study()\fP is a pointer for an error message. If
924  studying succeeds (even if no data is returned), the variable it points to is  studying succeeds (even if no data is returned), the variable it points to is
925  set to NULL. Otherwise it points to a textual error message. You should  set to NULL. Otherwise it is set to point to a textual error message. This is a
926  therefore test the error pointer for NULL after calling \fBpcre_study()\fP, to  static string that is part of the library. You must not try to free it. You
927  be sure that it has run successfully.  should test the error pointer for NULL after calling \fBpcre_study()\fP, to be
928    sure that it has run successfully.
929    .P
930    When you are finished with a pattern, you can free the memory used for the
931    study data by calling \fBpcre_free_study()\fP. This function was added to the
932    API for release 8.20. For earlier versions, the memory could be freed with
933    \fBpcre_free()\fP, just like the pattern itself. This will still work in cases
934    where PCRE_STUDY_JIT_COMPILE is not used, but it is advisable to change to the
935    new function when convenient.
936  .P  .P
937  This is a typical call to \fBpcre_study\fP():  This is a typical way in which \fBpcre_study\fP() is used (except that in a
938    real application there should be tests for errors):
939  .sp  .sp
940    pcre_extra *pe;    int rc;
941    pe = pcre_study(    pcre *re;
942      pcre_extra *sd;
943      re = pcre_compile("pattern", 0, &error, &erroroffset, NULL);
944      sd = pcre_study(
945      re,             /* result of pcre_compile() */      re,             /* result of pcre_compile() */
946      0,              /* no options exist */      0,              /* no options */
947      &error);        /* set to NULL or points to a message */      &error);        /* set to NULL or points to a message */
948  .sp    rc = pcre_exec(   /* see below for details of pcre_exec() options */
949  At present, studying a pattern is useful only for non-anchored patterns that do      re, sd, "subject", 7, 0, 0, ovector, 30);
950  not have a single fixed starting character. A bitmap of possible starting    ...
951  bytes is created.    pcre_free_study(sd);
952      pcre_free(re);
953    .sp
954    Studying a pattern does two things: first, a lower bound for the length of
955    subject string that is needed to match the pattern is computed. This does not
956    mean that there are any strings of that length that match, but it does
957    guarantee that no shorter strings match. The value is used by
958    \fBpcre_exec()\fP and \fBpcre_dfa_exec()\fP to avoid wasting time by trying to
959    match strings that are shorter than the lower bound. You can find out the value
960    in a calling program via the \fBpcre_fullinfo()\fP function.
961    .P
962    Studying a pattern is also useful for non-anchored patterns that do not have a
963    single fixed starting character. A bitmap of possible starting bytes is
964    created. This speeds up finding a position in the subject at which to start
965    matching.
966    .P
967    These two optimizations apply to both \fBpcre_exec()\fP and
968    \fBpcre_dfa_exec()\fP. However, they are not used by \fBpcre_exec()\fP if
969    \fBpcre_study()\fP is called with the PCRE_STUDY_JIT_COMPILE option, and
970    just-in-time compiling is successful. The optimizations can be disabled by
971    setting the PCRE_NO_START_OPTIMIZE option when calling \fBpcre_exec()\fP or
972    \fBpcre_dfa_exec()\fP. You might want to do this if your pattern contains
973    callouts or (*MARK) (which cannot be handled by the JIT compiler), and you want
974    to make use of these facilities in cases where matching fails. See the
975    discussion of PCRE_NO_START_OPTIMIZE
976    .\" HTML <a href="#execoptions">
977    .\" </a>
978    below.
979    .\"
980  .  .
981  .  .
982  .\" HTML <a name="localesupport"></a>  .\" HTML <a name="localesupport"></a>
# Line 525  bytes is created. Line 985  bytes is created.
985  .sp  .sp
986  PCRE handles caseless matching, and determines whether characters are letters,  PCRE handles caseless matching, and determines whether characters are letters,
987  digits, or whatever, by reference to a set of tables, indexed by character  digits, or whatever, by reference to a set of tables, indexed by character
988  value. (When running in UTF-8 mode, this applies only to characters with codes  value. When running in UTF-8 mode, this applies only to characters with codes
989  less than 128. Higher-valued codes never match escapes such as \ew or \ed, but  less than 128. By default, higher-valued codes never match escapes such as \ew
990  can be tested with \ep if PCRE is built with Unicode character property  or \ed, but they can be tested with \ep if PCRE is built with Unicode character
991  support.)  property support. Alternatively, the PCRE_UCP option can be set at compile
992  .P  time; this causes \ew and friends to use Unicode property support instead of
993  An internal set of tables is created in the default C locale when PCRE is  built-in tables. The use of locales with Unicode is discouraged. If you are
994  built. This is used when the final argument of \fBpcre_compile()\fP is NULL,  handling characters with codes greater than 128, you should either use UTF-8
995  and is sufficient for many applications. An alternative set of tables can,  and Unicode, or use locales, but not try to mix the two.
996  however, be supplied. These may be created in a different locale from the  .P
997  default. As more and more applications change to using Unicode, the need for  PCRE contains an internal set of tables that are used when the final argument
998  this locale support is expected to die away.  of \fBpcre_compile()\fP is NULL. These are sufficient for many applications.
999    Normally, the internal tables recognize only ASCII characters. However, when
1000    PCRE is built, it is possible to cause the internal tables to be rebuilt in the
1001    default "C" locale of the local system, which may cause them to be different.
1002    .P
1003    The internal tables can always be overridden by tables supplied by the
1004    application that calls PCRE. These may be created in a different locale from
1005    the default. As more and more applications change to using Unicode, the need
1006    for this locale support is expected to die away.
1007  .P  .P
1008  External tables are built by calling the \fBpcre_maketables()\fP function,  External tables are built by calling the \fBpcre_maketables()\fP function,
1009  which has no arguments, in the relevant locale. The result can then be passed  which has no arguments, in the relevant locale. The result can then be passed
# Line 548  the following code could be used: Line 1016  the following code could be used:
1016    tables = pcre_maketables();    tables = pcre_maketables();
1017    re = pcre_compile(..., tables);    re = pcre_compile(..., tables);
1018  .sp  .sp
1019    The locale name "fr_FR" is used on Linux and other Unix-like systems; if you
1020    are using Windows, the name for the French locale is "french".
1021    .P
1022  When \fBpcre_maketables()\fP runs, the tables are built in memory that is  When \fBpcre_maketables()\fP runs, the tables are built in memory that is
1023  obtained via \fBpcre_malloc\fP. It is the caller's responsibility to ensure  obtained via \fBpcre_malloc\fP. It is the caller's responsibility to ensure
1024  that the memory containing the tables remains available for as long as it is  that the memory containing the tables remains available for as long as it is
# Line 566  one in which it was compiled. Passing ta Line 1037  one in which it was compiled. Passing ta
1037  below in the section on matching a pattern.  below in the section on matching a pattern.
1038  .  .
1039  .  .
1040    .\" HTML <a name="infoaboutpattern"></a>
1041  .SH "INFORMATION ABOUT A PATTERN"  .SH "INFORMATION ABOUT A PATTERN"
1042  .rs  .rs
1043  .sp  .sp
# Line 594  check against passing an arbitrary memor Line 1066  check against passing an arbitrary memor
1066  \fBpcre_fullinfo()\fP, to obtain the length of the compiled pattern:  \fBpcre_fullinfo()\fP, to obtain the length of the compiled pattern:
1067  .sp  .sp
1068    int rc;    int rc;
1069    unsigned long int length;    size_t length;
1070    rc = pcre_fullinfo(    rc = pcre_fullinfo(
1071      re,               /* result of pcre_compile() */      re,               /* result of pcre_compile() */
1072      pe,               /* result of pcre_study(), or NULL */      sd,               /* result of pcre_study(), or NULL */
1073      PCRE_INFO_SIZE,   /* what is required */      PCRE_INFO_SIZE,   /* what is required */
1074      &length);         /* where to put the data */      &length);         /* where to put the data */
1075  .sp  .sp
# Line 615  no back references. Line 1087  no back references.
1087  Return the number of capturing subpatterns in the pattern. The fourth argument  Return the number of capturing subpatterns in the pattern. The fourth argument
1088  should point to an \fBint\fP variable.  should point to an \fBint\fP variable.
1089  .sp  .sp
1090    PCRE_INFO_DEFAULTTABLES    PCRE_INFO_DEFAULT_TABLES
1091  .sp  .sp
1092  Return a pointer to the internal default character tables within PCRE. The  Return a pointer to the internal default character tables within PCRE. The
1093  fourth argument should point to an \fBunsigned char *\fP variable. This  fourth argument should point to an \fBunsigned char *\fP variable. This
# Line 626  a NULL table pointer. Line 1098  a NULL table pointer.
1098    PCRE_INFO_FIRSTBYTE    PCRE_INFO_FIRSTBYTE
1099  .sp  .sp
1100  Return information about the first byte of any matched string, for a  Return information about the first byte of any matched string, for a
1101  non-anchored pattern. (This option used to be called PCRE_INFO_FIRSTCHAR; the  non-anchored pattern. The fourth argument should point to an \fBint\fP
1102  old name is still recognized for backwards compatibility.)  variable. (This option used to be called PCRE_INFO_FIRSTCHAR; the old name is
1103    still recognized for backwards compatibility.)
1104  .P  .P
1105  If there is a fixed first byte, for example, from a pattern such as  If there is a fixed first byte, for example, from a pattern such as
1106  (cat|cow|coyote), it is returned in the integer pointed to by \fIwhere\fP.  (cat|cow|coyote), its value is returned. Otherwise, if either
 Otherwise, if either  
1107  .sp  .sp
1108  (a) the pattern was compiled with the PCRE_MULTILINE option, and every branch  (a) the pattern was compiled with the PCRE_MULTILINE option, and every branch
1109  starts with "^", or  starts with "^", or
# Line 650  table indicating a fixed set of bytes fo Line 1122  table indicating a fixed set of bytes fo
1122  string, a pointer to the table is returned. Otherwise NULL is returned. The  string, a pointer to the table is returned. Otherwise NULL is returned. The
1123  fourth argument should point to an \fBunsigned char *\fP variable.  fourth argument should point to an \fBunsigned char *\fP variable.
1124  .sp  .sp
1125      PCRE_INFO_HASCRORLF
1126    .sp
1127    Return 1 if the pattern contains any explicit matches for CR or LF characters,
1128    otherwise 0. The fourth argument should point to an \fBint\fP variable. An
1129    explicit match is either a literal CR or LF character, or \er or \en.
1130    .sp
1131      PCRE_INFO_JCHANGED
1132    .sp
1133    Return 1 if the (?J) or (?-J) option setting is used in the pattern, otherwise
1134    0. The fourth argument should point to an \fBint\fP variable. (?J) and
1135    (?-J) set and unset the local PCRE_DUPNAMES option, respectively.
1136    .sp
1137      PCRE_INFO_JIT
1138    .sp
1139    Return 1 if the pattern was studied with the PCRE_STUDY_JIT_COMPILE option, and
1140    just-in-time compiling was successful. The fourth argument should point to an
1141    \fBint\fP variable. A return value of 0 means that JIT support is not available
1142    in this version of PCRE, or that the pattern was not studied with the
1143    PCRE_STUDY_JIT_COMPILE option, or that the JIT compiler could not handle this
1144    particular pattern. See the
1145    .\" HREF
1146    \fBpcrejit\fP
1147    .\"
1148    documentation for details of what can and cannot be handled.
1149    .sp
1150    PCRE_INFO_LASTLITERAL    PCRE_INFO_LASTLITERAL
1151  .sp  .sp
1152  Return the value of the rightmost literal byte that must exist in any matched  Return the value of the rightmost literal byte that must exist in any matched
# Line 660  follows something of variable length. Fo Line 1157  follows something of variable length. Fo
1157  /^a\ed+z\ed+/ the returned value is "z", but for /^a\edz\ed/ the returned value  /^a\ed+z\ed+/ the returned value is "z", but for /^a\edz\ed/ the returned value
1158  is -1.  is -1.
1159  .sp  .sp
1160      PCRE_INFO_MINLENGTH
1161    .sp
1162    If the pattern was studied and a minimum length for matching subject strings
1163    was computed, its value is returned. Otherwise the returned value is -1. The
1164    value is a number of characters, not bytes (this may be relevant in UTF-8
1165    mode). The fourth argument should point to an \fBint\fP variable. A
1166    non-negative value is a lower bound to the length of any matching string. There
1167    may not be any strings of that length that do actually match, but every string
1168    that does match is at least that long.
1169    .sp
1170    PCRE_INFO_NAMECOUNT    PCRE_INFO_NAMECOUNT
1171    PCRE_INFO_NAMEENTRYSIZE    PCRE_INFO_NAMEENTRYSIZE
1172    PCRE_INFO_NAMETABLE    PCRE_INFO_NAMETABLE
1173  .sp  .sp
1174  PCRE supports the use of named as well as numbered capturing parentheses. The  PCRE supports the use of named as well as numbered capturing parentheses. The
1175  names are just an additional way of identifying the parentheses, which still  names are just an additional way of identifying the parentheses, which still
1176  acquire numbers. A convenience function called \fBpcre_get_named_substring()\fP  acquire numbers. Several convenience functions such as
1177  is provided for extracting an individual captured substring by name. It is also  \fBpcre_get_named_substring()\fP are provided for extracting captured
1178  possible to extract the data directly, by first converting the name to a number  substrings by name. It is also possible to extract the data directly, by first
1179  in order to access the correct pointers in the output vector (described with  converting the name to a number in order to access the correct pointers in the
1180  \fBpcre_exec()\fP below). To do the conversion, you need to use the  output vector (described with \fBpcre_exec()\fP below). To do the conversion,
1181  name-to-number map, which is described by these three values.  you need to use the name-to-number map, which is described by these three
1182    values.
1183  .P  .P
1184  The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT gives  The map consists of a number of fixed-size entries. PCRE_INFO_NAMECOUNT gives
1185  the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size of each  the number of entries, and PCRE_INFO_NAMEENTRYSIZE gives the size of each
# Line 679  entry; both of these return an \fBint\fP Line 1187  entry; both of these return an \fBint\fP
1187  length of the longest name. PCRE_INFO_NAMETABLE returns a pointer to the first  length of the longest name. PCRE_INFO_NAMETABLE returns a pointer to the first
1188  entry of the table (a pointer to \fBchar\fP). The first two bytes of each entry  entry of the table (a pointer to \fBchar\fP). The first two bytes of each entry
1189  are the number of the capturing parenthesis, most significant byte first. The  are the number of the capturing parenthesis, most significant byte first. The
1190  rest of the entry is the corresponding name, zero terminated. The names are in  rest of the entry is the corresponding name, zero terminated.
1191  alphabetical order. For example, consider the following pattern (assume  .P
1192  PCRE_EXTENDED is set, so white space - including newlines - is ignored):  The names are in alphabetical order. Duplicate names may appear if (?| is used
1193    to create multiple groups with the same number, as described in the
1194    .\" HTML <a href="pcrepattern.html#dupsubpatternnumber">
1195    .\" </a>
1196    section on duplicate subpattern numbers
1197    .\"
1198    in the
1199    .\" HREF
1200    \fBpcrepattern\fP
1201    .\"
1202    page. Duplicate names for subpatterns with different numbers are permitted only
1203    if PCRE_DUPNAMES is set. In all cases of duplicate names, they appear in the
1204    table in the order in which they were found in the pattern. In the absence of
1205    (?| this is the order of increasing number; when (?| is used this is not
1206    necessarily the case because later subpatterns may have lower numbers.
1207    .P
1208    As a simple example of the name/number table, consider the following pattern
1209    (assume PCRE_EXTENDED is set, so white space - including newlines - is
1210    ignored):
1211  .sp  .sp
1212  .\" JOIN  .\" JOIN
1213    (?P<date> (?P<year>(\ed\ed)?\ed\ed) -    (?<date> (?<year>(\ed\ed)?\ed\ed) -
1214    (?P<month>\ed\ed) - (?P<day>\ed\ed) )    (?<month>\ed\ed) - (?<day>\ed\ed) )
1215  .sp  .sp
1216  There are four named subpatterns, so the table has four entries, and each entry  There are four named subpatterns, so the table has four entries, and each entry
1217  in the table is eight bytes long. The table is as follows, with non-printing  in the table is eight bytes long. The table is as follows, with non-printing
# Line 697  bytes shows in hexadecimal, and undefine Line 1223  bytes shows in hexadecimal, and undefine
1223    00 02 y  e  a  r  00 ??    00 02 y  e  a  r  00 ??
1224  .sp  .sp
1225  When writing code to extract data from named subpatterns using the  When writing code to extract data from named subpatterns using the
1226  name-to-number map, remember that the length of each entry is likely to be  name-to-number map, remember that the length of the entries is likely to be
1227  different for each compiled pattern.  different for each compiled pattern.
1228  .sp  .sp
1229      PCRE_INFO_OKPARTIAL
1230    .sp
1231    Return 1 if the pattern can be used for partial matching with
1232    \fBpcre_exec()\fP, otherwise 0. The fourth argument should point to an
1233    \fBint\fP variable. From release 8.00, this always returns 1, because the
1234    restrictions that previously applied to partial matching have been lifted. The
1235    .\" HREF
1236    \fBpcrepartial\fP
1237    .\"
1238    documentation gives details of partial matching.
1239    .sp
1240    PCRE_INFO_OPTIONS    PCRE_INFO_OPTIONS
1241  .sp  .sp
1242  Return a copy of the options with which the pattern was compiled. The fourth  Return a copy of the options with which the pattern was compiled. The fourth
1243  argument should point to an \fBunsigned long int\fP variable. These option bits  argument should point to an \fBunsigned long int\fP variable. These option bits
1244  are those specified in the call to \fBpcre_compile()\fP, modified by any  are those specified in the call to \fBpcre_compile()\fP, modified by any
1245  top-level option settings within the pattern itself.  top-level option settings at the start of the pattern itself. In other words,
1246    they are the options that will be in force when matching starts. For example,
1247    if the pattern /(?im)abc(?-i)d/ is compiled with the PCRE_EXTENDED option, the
1248    result is PCRE_CASELESS, PCRE_MULTILINE, and PCRE_EXTENDED.
1249  .P  .P
1250  A pattern is automatically anchored by PCRE if all of its top-level  A pattern is automatically anchored by PCRE if all of its top-level
1251  alternatives begin with one of the following:  alternatives begin with one of the following:
# Line 729  variable. Line 1269  variable.
1269  .sp  .sp
1270    PCRE_INFO_STUDYSIZE    PCRE_INFO_STUDYSIZE
1271  .sp  .sp
1272  Return the size of the data block pointed to by the \fIstudy_data\fP field in  Return the size of the data block pointed to by the \fIstudy_data\fP field in a
1273  a \fBpcre_extra\fP block. That is, it is the value that was passed to  \fBpcre_extra\fP block. If \fBpcre_extra\fP is NULL, or there is no study data,
1274  \fBpcre_malloc()\fP when PCRE was getting memory into which to place the data  zero is returned. The fourth argument should point to a \fBsize_t\fP variable.
1275  created by \fBpcre_study()\fP. The fourth argument should point to a  The \fIstudy_data\fP field is set by \fBpcre_study()\fP to record information
1276  \fBsize_t\fP variable.  that will speed up matching (see the section entitled
1277    .\" HTML <a href="#studyingapattern">
1278    .\" </a>
1279    "Studying a pattern"
1280    .\"
1281    above). The format of the \fIstudy_data\fP block is private, but its length
1282    is made available via this option so that it can be saved and restored (see the
1283    .\" HREF
1284    \fBpcreprecompile\fP
1285    .\"
1286    documentation for details).
1287  .  .
1288  .  .
1289  .SH "OBSOLETE INFO FUNCTION"  .SH "OBSOLETE INFO FUNCTION"
# Line 760  it is used to pass back information abou Line 1310  it is used to pass back information abou
1310  string (see PCRE_INFO_FIRSTBYTE above).  string (see PCRE_INFO_FIRSTBYTE above).
1311  .  .
1312  .  .
1313  .SH "MATCHING A PATTERN"  .SH "REFERENCE COUNTS"
1314    .rs
1315    .sp
1316    .B int pcre_refcount(pcre *\fIcode\fP, int \fIadjust\fP);
1317    .PP
1318    The \fBpcre_refcount()\fP function is used to maintain a reference count in the
1319    data block that contains a compiled pattern. It is provided for the benefit of
1320    applications that operate in an object-oriented manner, where different parts
1321    of the application may be using the same compiled pattern, but you want to free
1322    the block when they are all done.
1323    .P
1324    When a pattern is compiled, the reference count field is initialized to zero.
1325    It is changed only by calling this function, whose action is to add the
1326    \fIadjust\fP value (which may be positive or negative) to it. The yield of the
1327    function is the new value. However, the value of the count is constrained to
1328    lie between 0 and 65535, inclusive. If the new value is outside these limits,
1329    it is forced to the appropriate limit value.
1330    .P
1331    Except when it is zero, the reference count is not correctly preserved if a
1332    pattern is compiled on one host and then transferred to a host whose byte-order
1333    is different. (This seems a highly unlikely scenario.)
1334    .
1335    .
1336    .SH "MATCHING A PATTERN: THE TRADITIONAL FUNCTION"
1337  .rs  .rs
1338  .sp  .sp
1339  .B int pcre_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"  .B int pcre_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
# Line 771  string (see PCRE_INFO_FIRSTBYTE above). Line 1344  string (see PCRE_INFO_FIRSTBYTE above).
1344  .P  .P
1345  The function \fBpcre_exec()\fP is called to match a subject string against a  The function \fBpcre_exec()\fP is called to match a subject string against a
1346  compiled pattern, which is passed in the \fIcode\fP argument. If the  compiled pattern, which is passed in the \fIcode\fP argument. If the
1347  pattern has been studied, the result of the study should be passed in the  pattern was studied, the result of the study should be passed in the
1348  \fIextra\fP argument.  \fIextra\fP argument. You can call \fBpcre_exec()\fP with the same \fIcode\fP
1349    and \fIextra\fP arguments as many times as you like, in order to match
1350    different subject strings with the same pattern.
1351    .P
1352    This function is the main matching facility of the library, and it operates in
1353    a Perl-like manner. For specialist use there is also an alternative matching
1354    function, which is described
1355    .\" HTML <a href="#dfamatch">
1356    .\" </a>
1357    below
1358    .\"
1359    in the section about the \fBpcre_dfa_exec()\fP function.
1360  .P  .P
1361  In most applications, the pattern will have been compiled (and optionally  In most applications, the pattern will have been compiled (and optionally
1362  studied) in the same process that calls \fBpcre_exec()\fP. However, it is  studied) in the same process that calls \fBpcre_exec()\fP. However, it is
# Line 796  Here is an example of a simple call to \ Line 1380  Here is an example of a simple call to \
1380      0,              /* start at offset 0 in the subject */      0,              /* start at offset 0 in the subject */
1381      0,              /* default options */      0,              /* default options */
1382      ovector,        /* vector of integers for substring information */      ovector,        /* vector of integers for substring information */
1383      30);            /* number of elements in the vector (NOT size in bytes) */      30);            /* number of elements (NOT size in bytes) */
1384    .
1385  .  .
1386  .\" HTML <a name="extradata"></a>  .\" HTML <a name="extradata"></a>
1387  .SS "Extra data for \fBpcre_exec()\fR"  .SS "Extra data for \fBpcre_exec()\fR"
# Line 805  Here is an example of a simple call to \ Line 1390  Here is an example of a simple call to \
1390  If the \fIextra\fP argument is not NULL, it must point to a \fBpcre_extra\fP  If the \fIextra\fP argument is not NULL, it must point to a \fBpcre_extra\fP
1391  data block. The \fBpcre_study()\fP function returns such a block (when it  data block. The \fBpcre_study()\fP function returns such a block (when it
1392  doesn't return NULL), but you can also create one for yourself, and pass  doesn't return NULL), but you can also create one for yourself, and pass
1393  additional information in it. The fields in a \fBpcre_extra\fP block are as  additional information in it. The \fBpcre_extra\fP block contains the following
1394  follows:  fields (not necessarily in this order):
1395  .sp  .sp
1396    unsigned long int \fIflags\fP;    unsigned long int \fIflags\fP;
1397    void *\fIstudy_data\fP;    void *\fIstudy_data\fP;
1398      void *\fIexecutable_jit\fP;
1399    unsigned long int \fImatch_limit\fP;    unsigned long int \fImatch_limit\fP;
1400      unsigned long int \fImatch_limit_recursion\fP;
1401    void *\fIcallout_data\fP;    void *\fIcallout_data\fP;
1402    const unsigned char *\fItables\fP;    const unsigned char *\fItables\fP;
1403      unsigned char **\fImark\fP;
1404  .sp  .sp
1405  The \fIflags\fP field is a bitmap that specifies which of the other fields  The \fIflags\fP field is a bitmap that specifies which of the other fields
1406  are set. The flag bits are:  are set. The flag bits are:
1407  .sp  .sp
1408    PCRE_EXTRA_STUDY_DATA    PCRE_EXTRA_STUDY_DATA
1409      PCRE_EXTRA_EXECUTABLE_JIT
1410    PCRE_EXTRA_MATCH_LIMIT    PCRE_EXTRA_MATCH_LIMIT
1411      PCRE_EXTRA_MATCH_LIMIT_RECURSION
1412    PCRE_EXTRA_CALLOUT_DATA    PCRE_EXTRA_CALLOUT_DATA
1413    PCRE_EXTRA_TABLES    PCRE_EXTRA_TABLES
1414      PCRE_EXTRA_MARK
1415  .sp  .sp
1416  Other flag bits should be set to zero. The \fIstudy_data\fP field is set in the  Other flag bits should be set to zero. The \fIstudy_data\fP field and sometimes
1417  \fBpcre_extra\fP block that is returned by \fBpcre_study()\fP, together with  the \fIexecutable_jit\fP field are set in the \fBpcre_extra\fP block that is
1418  the appropriate flag bit. You should not set this yourself, but you may add to  returned by \fBpcre_study()\fP, together with the appropriate flag bits. You
1419  the block by setting the other fields and their corresponding flag bits.  should not set these yourself, but you may add to the block by setting the
1420    other fields and their corresponding flag bits.
1421  .P  .P
1422  The \fImatch_limit\fP field provides a means of preventing PCRE from using up a  The \fImatch_limit\fP field provides a means of preventing PCRE from using up a
1423  vast amount of resources when running patterns that are not going to match,  vast amount of resources when running patterns that are not going to match,
1424  but which have a very large number of possibilities in their search trees. The  but which have a very large number of possibilities in their search trees. The
1425  classic example is the use of nested unlimited repeats.  classic example is a pattern that uses nested unlimited repeats.
1426  .P  .P
1427  Internally, PCRE uses a function called \fBmatch()\fP which it calls repeatedly  Internally, \fBpcre_exec()\fP uses a function called \fBmatch()\fP, which it
1428  (sometimes recursively). The limit is imposed on the number of times this  calls repeatedly (sometimes recursively). The limit set by \fImatch_limit\fP is
1429  function is called during a match, which has the effect of limiting the amount  imposed on the number of times this function is called during a match, which
1430  of recursion and backtracking that can take place. For patterns that are not  has the effect of limiting the amount of backtracking that can take place. For
1431  anchored, the count starts from zero for each position in the subject string.  patterns that are not anchored, the count restarts from zero for each position
1432    in the subject string.
1433    .P
1434    When \fBpcre_exec()\fP is called with a pattern that was successfully studied
1435    with the PCRE_STUDY_JIT_COMPILE option, the way that the matching is executed
1436    is entirely different. However, there is still the possibility of runaway
1437    matching that goes on for a very long time, and so the \fImatch_limit\fP value
1438    is also used in this case (but in a different way) to limit how long the
1439    matching can continue.
1440  .P  .P
1441  The default limit for the library can be set when PCRE is built; the default  The default value for the limit can be set when PCRE is built; the default
1442  default is 10 million, which handles all but the most extreme cases. You can  default is 10 million, which handles all but the most extreme cases. You can
1443  reduce the default by suppling \fBpcre_exec()\fP with a \fBpcre_extra\fP block  override the default by suppling \fBpcre_exec()\fP with a \fBpcre_extra\fP
1444  in which \fImatch_limit\fP is set to a smaller value, and  block in which \fImatch_limit\fP is set, and PCRE_EXTRA_MATCH_LIMIT is set in
1445  PCRE_EXTRA_MATCH_LIMIT is set in the \fIflags\fP field. If the limit is  the \fIflags\fP field. If the limit is exceeded, \fBpcre_exec()\fP returns
1446  exceeded, \fBpcre_exec()\fP returns PCRE_ERROR_MATCHLIMIT.  PCRE_ERROR_MATCHLIMIT.
1447    .P
1448    The \fImatch_limit_recursion\fP field is similar to \fImatch_limit\fP, but
1449    instead of limiting the total number of times that \fBmatch()\fP is called, it
1450    limits the depth of recursion. The recursion depth is a smaller number than the
1451    total number of calls, because not all calls to \fBmatch()\fP are recursive.
1452    This limit is of use only if it is set smaller than \fImatch_limit\fP.
1453    .P
1454    Limiting the recursion depth limits the amount of machine stack that can be
1455    used, or, when PCRE has been compiled to use memory on the heap instead of the
1456    stack, the amount of heap memory that can be used. This limit is not relevant,
1457    and is ignored, if the pattern was successfully studied with
1458    PCRE_STUDY_JIT_COMPILE.
1459    .P
1460    The default value for \fImatch_limit_recursion\fP can be set when PCRE is
1461    built; the default default is the same value as the default for
1462    \fImatch_limit\fP. You can override the default by suppling \fBpcre_exec()\fP
1463    with a \fBpcre_extra\fP block in which \fImatch_limit_recursion\fP is set, and
1464    PCRE_EXTRA_MATCH_LIMIT_RECURSION is set in the \fIflags\fP field. If the limit
1465    is exceeded, \fBpcre_exec()\fP returns PCRE_ERROR_RECURSIONLIMIT.
1466  .P  .P
1467  The \fIpcre_callout\fP field is used in conjunction with the "callout" feature,  The \fIcallout_data\fP field is used in conjunction with the "callout" feature,
1468  which is described in the  and is described in the
1469  .\" HREF  .\" HREF
1470  \fBpcrecallout\fP  \fBpcrecallout\fP
1471  .\"  .\"
# Line 865  called. See the Line 1484  called. See the
1484  \fBpcreprecompile\fP  \fBpcreprecompile\fP
1485  .\"  .\"
1486  documentation for a discussion of saving compiled patterns for later use.  documentation for a discussion of saving compiled patterns for later use.
1487    .P
1488    If PCRE_EXTRA_MARK is set in the \fIflags\fP field, the \fImark\fP field must
1489    be set to point to a \fBchar *\fP variable. If the pattern contains any
1490    backtracking control verbs such as (*MARK:NAME), and the execution ends up with
1491    a name to pass back, a pointer to the name string (zero terminated) is placed
1492    in the variable pointed to by the \fImark\fP field. The names are within the
1493    compiled pattern; if you wish to retain such a name you must copy it before
1494    freeing the memory of a compiled pattern. If there is no name to pass back, the
1495    variable pointed to by the \fImark\fP field set to NULL. For details of the
1496    backtracking control verbs, see the section entitled
1497    .\" HTML <a href="pcrepattern#backtrackcontrol">
1498    .\" </a>
1499    "Backtracking control"
1500    .\"
1501    in the
1502    .\" HREF
1503    \fBpcrepattern\fP
1504    .\"
1505    documentation.
1506    .
1507  .  .
1508    .\" HTML <a name="execoptions"></a>
1509  .SS "Option bits for \fBpcre_exec()\fP"  .SS "Option bits for \fBpcre_exec()\fP"
1510  .rs  .rs
1511  .sp  .sp
1512  The unused bits of the \fIoptions\fP argument for \fBpcre_exec()\fP must be  The unused bits of the \fIoptions\fP argument for \fBpcre_exec()\fP must be
1513  zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NOTBOL,  zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_\fIxxx\fP,
1514  PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NO_UTF8_CHECK and PCRE_PARTIAL.  PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
1515    PCRE_NO_START_OPTIMIZE, PCRE_NO_UTF8_CHECK, PCRE_PARTIAL_SOFT, and
1516    PCRE_PARTIAL_HARD.
1517    .P
1518    If the pattern was successfully studied with the PCRE_STUDY_JIT_COMPILE option,
1519    the only supported options for JIT execution are PCRE_NO_UTF8_CHECK,
1520    PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, and PCRE_NOTEMPTY_ATSTART. Note in
1521    particular that partial matching is not supported. If an unsupported option is
1522    used, JIT execution is disabled and the normal interpretive code in
1523    \fBpcre_exec()\fP is run.
1524  .sp  .sp
1525    PCRE_ANCHORED    PCRE_ANCHORED
1526  .sp  .sp
# Line 880  matching position. If a pattern was comp Line 1529  matching position. If a pattern was comp
1529  to be anchored by virtue of its contents, it cannot be made unachored at  to be anchored by virtue of its contents, it cannot be made unachored at
1530  matching time.  matching time.
1531  .sp  .sp
1532      PCRE_BSR_ANYCRLF
1533      PCRE_BSR_UNICODE
1534    .sp
1535    These options (which are mutually exclusive) control what the \eR escape
1536    sequence matches. The choice is either to match only CR, LF, or CRLF, or to
1537    match any Unicode newline sequence. These options override the choice that was
1538    made or defaulted when the pattern was compiled.
1539    .sp
1540      PCRE_NEWLINE_CR
1541      PCRE_NEWLINE_LF
1542      PCRE_NEWLINE_CRLF
1543      PCRE_NEWLINE_ANYCRLF
1544      PCRE_NEWLINE_ANY
1545    .sp
1546    These options override the newline definition that was chosen or defaulted when
1547    the pattern was compiled. For details, see the description of
1548    \fBpcre_compile()\fP above. During matching, the newline choice affects the
1549    behaviour of the dot, circumflex, and dollar metacharacters. It may also alter
1550    the way the match position is advanced after a match failure for an unanchored
1551    pattern.
1552    .P
1553    When PCRE_NEWLINE_CRLF, PCRE_NEWLINE_ANYCRLF, or PCRE_NEWLINE_ANY is set, and a
1554    match attempt for an unanchored pattern fails when the current position is at a
1555    CRLF sequence, and the pattern contains no explicit matches for CR or LF
1556    characters, the match position is advanced by two characters instead of one, in
1557    other words, to after the CRLF.
1558    .P
1559    The above rule is a compromise that makes the most common cases work as
1560    expected. For example, if the pattern is .+A (and the PCRE_DOTALL option is not
1561    set), it does not match the string "\er\enA" because, after failing at the
1562    start, it skips both the CR and the LF before retrying. However, the pattern
1563    [\er\en]A does match that string, because it contains an explicit CR or LF
1564    reference, and so advances only by one character after the first failure.
1565    .P
1566    An explicit match for CR of LF is either a literal appearance of one of those
1567    characters, or one of the \er or \en escape sequences. Implicit matches such as
1568    [^X] do not count, nor does \es (which includes CR and LF in the characters
1569    that it matches).
1570    .P
1571    Notwithstanding the above, anomalous effects may still occur when CRLF is a
1572    valid newline sequence and explicit \er or \en escapes appear in the pattern.
1573    .sp
1574    PCRE_NOTBOL    PCRE_NOTBOL
1575  .sp  .sp
1576  This option specifies that first character of the subject string is not the  This option specifies that first character of the subject string is not the
# Line 904  match the empty string, the entire match Line 1595  match the empty string, the entire match
1595  .sp  .sp
1596    a?b?    a?b?
1597  .sp  .sp
1598  is applied to a string not beginning with "a" or "b", it matches the empty  is applied to a string not beginning with "a" or "b", it matches an empty
1599  string at the start of the subject. With PCRE_NOTEMPTY set, this match is not  string at the start of the subject. With PCRE_NOTEMPTY set, this match is not
1600  valid, so PCRE searches further into the string for occurrences of "a" or "b".  valid, so PCRE searches further into the string for occurrences of "a" or "b".
1601    .sp
1602      PCRE_NOTEMPTY_ATSTART
1603    .sp
1604    This is like PCRE_NOTEMPTY, except that an empty string match that is not at
1605    the start of the subject is permitted. If the pattern is anchored, such a match
1606    can occur only if the pattern contains \eK.
1607    .P
1608    Perl has no direct equivalent of PCRE_NOTEMPTY or PCRE_NOTEMPTY_ATSTART, but it
1609    does make a special case of a pattern match of the empty string within its
1610    \fBsplit()\fP function, and when using the /g modifier. It is possible to
1611    emulate Perl's behaviour after matching a null string by first trying the match
1612    again at the same offset with PCRE_NOTEMPTY_ATSTART and PCRE_ANCHORED, and then
1613    if that fails, by advancing the starting offset (see below) and trying an
1614    ordinary match again. There is some code that demonstrates how to do this in
1615    the
1616    .\" HREF
1617    \fBpcredemo\fP
1618    .\"
1619    sample program. In the most general case, you have to check to see if the
1620    newline convention recognizes CRLF as a newline, and if so, and the current
1621    character is CR followed by LF, advance the starting offset by two characters
1622    instead of one.
1623    .sp
1624      PCRE_NO_START_OPTIMIZE
1625    .sp
1626    There are a number of optimizations that \fBpcre_exec()\fP uses at the start of
1627    a match, in order to speed up the process. For example, if it is known that an
1628    unanchored match must start with a specific character, it searches the subject
1629    for that character, and fails immediately if it cannot find it, without
1630    actually running the main matching function. This means that a special item
1631    such as (*COMMIT) at the start of a pattern is not considered until after a
1632    suitable starting point for the match has been found. When callouts or (*MARK)
1633    items are in use, these "start-up" optimizations can cause them to be skipped
1634    if the pattern is never actually used. The start-up optimizations are in effect
1635    a pre-scan of the subject that takes place before the pattern is run.
1636    .P
1637    The PCRE_NO_START_OPTIMIZE option disables the start-up optimizations, possibly
1638    causing performance to suffer, but ensuring that in cases where the result is
1639    "no match", the callouts do occur, and that items such as (*COMMIT) and (*MARK)
1640    are considered at every possible starting position in the subject string. If
1641    PCRE_NO_START_OPTIMIZE is set at compile time, it cannot be unset at matching
1642    time.
1643  .P  .P
1644  Perl has no direct equivalent of PCRE_NOTEMPTY, but it does make a special case  Setting PCRE_NO_START_OPTIMIZE can change the outcome of a matching operation.
1645  of a pattern match of the empty string within its \fBsplit()\fP function, and  Consider the pattern
1646  when using the /g modifier. It is possible to emulate Perl's behaviour after  .sp
1647  matching a null string by first trying the match again at the same offset with    (*COMMIT)ABC
1648  PCRE_NOTEMPTY and PCRE_ANCHORED, and then if that fails by advancing the  .sp
1649  starting offset (see below) and trying an ordinary match again. There is some  When this is compiled, PCRE records the fact that a match must start with the
1650  code that demonstrates how to do this in the \fIpcredemo.c\fP sample program.  character "A". Suppose the subject string is "DEFABC". The start-up
1651    optimization scans along the subject, finds "A" and runs the first match
1652    attempt from there. The (*COMMIT) item means that the pattern must match the
1653    current starting position, which in this case, it does. However, if the same
1654    match is run with PCRE_NO_START_OPTIMIZE set, the initial scan along the
1655    subject string does not happen. The first match attempt is run starting from
1656    "D" and when this fails, (*COMMIT) prevents any further matches being tried, so
1657    the overall result is "no match". If the pattern is studied, more start-up
1658    optimizations may be used. For example, a minimum length for the subject may be
1659    recorded. Consider the pattern
1660    .sp
1661      (*MARK:A)(X|Y)
1662    .sp
1663    The minimum length for a match is one character. If the subject is "ABC", there
1664    will be attempts to match "ABC", "BC", "C", and then finally an empty string.
1665    If the pattern is studied, the final attempt does not take place, because PCRE
1666    knows that the subject is too short, and so the (*MARK) is never encountered.
1667    In this case, studying the pattern does not affect the overall match result,
1668    which is still "no match", but it does affect the auxiliary information that is
1669    returned.
1670  .sp  .sp
1671    PCRE_NO_UTF8_CHECK    PCRE_NO_UTF8_CHECK
1672  .sp  .sp
1673  When PCRE_UTF8 is set at compile time, the validity of the subject as a UTF-8  When PCRE_UTF8 is set at compile time, the validity of the subject as a UTF-8
1674  string is automatically checked when \fBpcre_exec()\fP is subsequently called.  string is automatically checked when \fBpcre_exec()\fP is subsequently called.
1675  The value of \fIstartoffset\fP is also checked to ensure that it points to the  The value of \fIstartoffset\fP is also checked to ensure that it points to the
1676  start of a UTF-8 character. If an invalid UTF-8 sequence of bytes is found,  start of a UTF-8 character. There is a discussion about the validity of UTF-8
1677  \fBpcre_exec()\fP returns the error PCRE_ERROR_BADUTF8. If \fIstartoffset\fP  strings in the
1678  contains an invalid value, PCRE_ERROR_BADUTF8_OFFSET is returned.  .\" HTML <a href="pcre.html#utf8strings">
1679    .\" </a>
1680    section on UTF-8 support
1681    .\"
1682    in the main
1683    .\" HREF
1684    \fBpcre\fP
1685    .\"
1686    page. If an invalid UTF-8 sequence of bytes is found, \fBpcre_exec()\fP returns
1687    the error PCRE_ERROR_BADUTF8 or, if PCRE_PARTIAL_HARD is set and the problem is
1688    a truncated UTF-8 character at the end of the subject, PCRE_ERROR_SHORTUTF8. In
1689    both cases, information about the precise nature of the error may also be
1690    returned (see the descriptions of these errors in the section entitled \fIError
1691    return values from\fP \fBpcre_exec()\fP
1692    .\" HTML <a href="#errorlist">
1693    .\" </a>
1694    below).
1695    .\"
1696    If \fIstartoffset\fP contains a value that does not point to the start of a
1697    UTF-8 character (or to the end of the subject), PCRE_ERROR_BADUTF8_OFFSET is
1698    returned.
1699  .P  .P
1700  If you already know that your subject is valid, and you want to skip these  If you already know that your subject is valid, and you want to skip these
1701  checks for performance reasons, you can set the PCRE_NO_UTF8_CHECK option when  checks for performance reasons, you can set the PCRE_NO_UTF8_CHECK option when
1702  calling \fBpcre_exec()\fP. You might want to do this for the second and  calling \fBpcre_exec()\fP. You might want to do this for the second and
1703  subsequent calls to \fBpcre_exec()\fP if you are making repeated calls to find  subsequent calls to \fBpcre_exec()\fP if you are making repeated calls to find
1704  all the matches in a single subject string. However, you should be sure that  all the matches in a single subject string. However, you should be sure that
1705  the value of \fIstartoffset\fP points to the start of a UTF-8 character. When  the value of \fIstartoffset\fP points to the start of a UTF-8 character (or the
1706  PCRE_NO_UTF8_CHECK is set, the effect of passing an invalid UTF-8 string as a  end of the subject). When PCRE_NO_UTF8_CHECK is set, the effect of passing an
1707  subject, or a value of \fIstartoffset\fP that does not point to the start of a  invalid UTF-8 string as a subject or an invalid value of \fIstartoffset\fP is
1708  UTF-8 character, is undefined. Your program may crash.  undefined. Your program may crash.
1709  .sp  .sp
1710    PCRE_PARTIAL    PCRE_PARTIAL_HARD
1711  .sp    PCRE_PARTIAL_SOFT
1712  This option turns on the partial matching feature. If the subject string fails  .sp
1713  to match the pattern, but at some point during the matching process the end of  These options turn on the partial matching feature. For backwards
1714  the subject was reached (that is, the subject partially matches the pattern and  compatibility, PCRE_PARTIAL is a synonym for PCRE_PARTIAL_SOFT. A partial match
1715  the failure to match occurred only because there were not enough subject  occurs if the end of the subject string is reached successfully, but there are
1716  characters), \fBpcre_exec()\fP returns PCRE_ERROR_PARTIAL instead of  not enough subject characters to complete the match. If this happens when
1717  PCRE_ERROR_NOMATCH. When PCRE_PARTIAL is used, there are restrictions on what  PCRE_PARTIAL_SOFT (but not PCRE_PARTIAL_HARD) is set, matching continues by
1718  may appear in the pattern. These are discussed in the  testing any remaining alternatives. Only if no complete match can be found is
1719    PCRE_ERROR_PARTIAL returned instead of PCRE_ERROR_NOMATCH. In other words,
1720    PCRE_PARTIAL_SOFT says that the caller is prepared to handle a partial match,
1721    but only if no complete match can be found.
1722    .P
1723    If PCRE_PARTIAL_HARD is set, it overrides PCRE_PARTIAL_SOFT. In this case, if a
1724    partial match is found, \fBpcre_exec()\fP immediately returns
1725    PCRE_ERROR_PARTIAL, without considering any other alternatives. In other words,
1726    when PCRE_PARTIAL_HARD is set, a partial match is considered to be more
1727    important that an alternative complete match.
1728    .P
1729    In both cases, the portion of the string that was inspected when the partial
1730    match was found is set as the first matching string. There is a more detailed
1731    discussion of partial and multi-segment matching, with examples, in the
1732  .\" HREF  .\" HREF
1733  \fBpcrepartial\fP  \fBpcrepartial\fP
1734  .\"  .\"
1735  documentation.  documentation.
1736  .  .
1737    .
1738  .SS "The string to be matched by \fBpcre_exec()\fP"  .SS "The string to be matched by \fBpcre_exec()\fP"
1739  .rs  .rs
1740  .sp  .sp
1741  The subject string is passed to \fBpcre_exec()\fP as a pointer in  The subject string is passed to \fBpcre_exec()\fP as a pointer in
1742  \fIsubject\fP, a length in \fIlength\fP, and a starting byte offset in  \fIsubject\fP, a length (in bytes) in \fIlength\fP, and a starting byte offset
1743  \fIstartoffset\fP. In UTF-8 mode, the byte offset must point to the start of a  in \fIstartoffset\fP. If this is negative or greater than the length of the
1744  UTF-8 character. Unlike the pattern string, the subject may contain binary zero  subject, \fBpcre_exec()\fP returns PCRE_ERROR_BADOFFSET. When the starting
1745  bytes. When the starting offset is zero, the search for a match starts at the  offset is zero, the search for a match starts at the beginning of the subject,
1746  beginning of the subject, and this is by far the most common case.  and this is by far the most common case. In UTF-8 mode, the byte offset must
1747    point to the start of a UTF-8 character (or the end of the subject). Unlike the
1748    pattern string, the subject may contain binary zero bytes.
1749  .P  .P
1750  A non-zero starting offset is useful when searching for another match in the  A non-zero starting offset is useful when searching for another match in the
1751  same subject by calling \fBpcre_exec()\fP again after a previous success.  same subject by calling \fBpcre_exec()\fP again after a previous success.
# Line 977  start of the subject, which is deemed to Line 1765  start of the subject, which is deemed to
1765  set to 4, it finds the second occurrence of "iss" because it is able to look  set to 4, it finds the second occurrence of "iss" because it is able to look
1766  behind the starting point to discover that it is preceded by a letter.  behind the starting point to discover that it is preceded by a letter.
1767  .P  .P
1768    Finding all the matches in a subject is tricky when the pattern can match an
1769    empty string. It is possible to emulate Perl's /g behaviour by first trying the
1770    match again at the same offset, with the PCRE_NOTEMPTY_ATSTART and
1771    PCRE_ANCHORED options, and then if that fails, advancing the starting offset
1772    and trying an ordinary match again. There is some code that demonstrates how to
1773    do this in the
1774    .\" HREF
1775    \fBpcredemo\fP
1776    .\"
1777    sample program. In the most general case, you have to check to see if the
1778    newline convention recognizes CRLF as a newline, and if so, and the current
1779    character is CR followed by LF, advance the starting offset by two characters
1780    instead of one.
1781    .P
1782  If a non-zero starting offset is passed when the pattern is anchored, one  If a non-zero starting offset is passed when the pattern is anchored, one
1783  attempt to match at the given offset is made. This can only succeed if the  attempt to match at the given offset is made. This can only succeed if the
1784  pattern does not require the match to be at the start of the subject.  pattern does not require the match to be at the start of the subject.
1785  .  .
1786    .
1787  .SS "How \fBpcre_exec()\fP returns captured substrings"  .SS "How \fBpcre_exec()\fP returns captured substrings"
1788  .rs  .rs
1789  .sp  .sp
# Line 991  pattern. Following the usage in Jeffrey Line 1794  pattern. Following the usage in Jeffrey
1794  a fragment of a pattern that picks out a substring. PCRE supports several other  a fragment of a pattern that picks out a substring. PCRE supports several other
1795  kinds of parenthesized subpattern that do not cause substrings to be captured.  kinds of parenthesized subpattern that do not cause substrings to be captured.
1796  .P  .P
1797  Captured substrings are returned to the caller via a vector of integer offsets  Captured substrings are returned to the caller via a vector of integers whose
1798  whose address is passed in \fIovector\fP. The number of elements in the vector  address is passed in \fIovector\fP. The number of elements in the vector is
1799  is passed in \fIovecsize\fP, which must be a non-negative number. \fBNote\fP:  passed in \fIovecsize\fP, which must be a non-negative number. \fBNote\fP: this
1800  this argument is NOT the size of \fIovector\fP in bytes.  argument is NOT the size of \fIovector\fP in bytes.
1801  .P  .P
1802  The first two-thirds of the vector is used to pass back captured substrings,  The first two-thirds of the vector is used to pass back captured substrings,
1803  each substring using a pair of integers. The remaining third of the vector is  each substring using a pair of integers. The remaining third of the vector is
1804  used as workspace by \fBpcre_exec()\fP while matching capturing subpatterns,  used as workspace by \fBpcre_exec()\fP while matching capturing subpatterns,
1805  and is not available for passing back information. The length passed in  and is not available for passing back information. The number passed in
1806  \fIovecsize\fP should always be a multiple of three. If it is not, it is  \fIovecsize\fP should always be a multiple of three. If it is not, it is
1807  rounded down.  rounded down.
1808  .P  .P
1809  When a match is successful, information about captured substrings is returned  When a match is successful, information about captured substrings is returned
1810  in pairs of integers, starting at the beginning of \fIovector\fP, and  in pairs of integers, starting at the beginning of \fIovector\fP, and
1811  continuing up to two-thirds of its length at the most. The first element of a  continuing up to two-thirds of its length at the most. The first element of
1812  pair is set to the offset of the first character in a substring, and the second  each pair is set to the byte offset of the first character in a substring, and
1813  is set to the offset of the first character after the end of a substring. The  the second is set to the byte offset of the first character after the end of a
1814  first pair, \fIovector[0]\fP and \fIovector[1]\fP, identify the portion of the  substring. \fBNote\fP: these values are always byte offsets, even in UTF-8
1815  subject string matched by the entire pattern. The next pair is used for the  mode. They are not character counts.
1816  first capturing subpattern, and so on. The value returned by \fBpcre_exec()\fP  .P
1817  is the number of pairs that have been set. If there are no capturing  The first pair of integers, \fIovector[0]\fP and \fIovector[1]\fP, identify the
1818  subpatterns, the return value from a successful match is 1, indicating that  portion of the subject string matched by the entire pattern. The next pair is
1819  just the first pair of offsets has been set.  used for the first capturing subpattern, and so on. The value returned by
1820  .P  \fBpcre_exec()\fP is one more than the highest numbered pair that has been set.
1821  Some convenience functions are provided for extracting the captured substrings  For example, if two substrings have been captured, the returned value is 3. If
1822  as separate strings. These are described in the following section.  there are no capturing subpatterns, the return value from a successful match is
1823  .P  1, indicating that just the first pair of offsets has been set.
 It is possible for an capturing subpattern number \fIn+1\fP to match some  
 part of the subject when subpattern \fIn\fP has not been used at all. For  
 example, if the string "abc" is matched against the pattern (a|(z))(bc)  
 subpatterns 1 and 3 are matched, but 2 is not. When this happens, both offset  
 values corresponding to the unused subpattern are set to -1.  
1824  .P  .P
1825  If a capturing subpattern is matched repeatedly, it is the last portion of the  If a capturing subpattern is matched repeatedly, it is the last portion of the
1826  string that it matched that is returned.  string that it matched that is returned.
1827  .P  .P
1828  If the vector is too small to hold all the captured substring offsets, it is  If the vector is too small to hold all the captured substring offsets, it is
1829  used as far as possible (up to two-thirds of its length), and the function  used as far as possible (up to two-thirds of its length), and the function
1830  returns a value of zero. In particular, if the substring offsets are not of  returns a value of zero. If neither the actual string matched not any captured
1831  interest, \fBpcre_exec()\fP may be called with \fIovector\fP passed as NULL and  substrings are of interest, \fBpcre_exec()\fP may be called with \fIovector\fP
1832  \fIovecsize\fP as zero. However, if the pattern contains back references and  passed as NULL and \fIovecsize\fP as zero. However, if the pattern contains
1833  the \fIovector\fP is not big enough to remember the related substrings, PCRE  back references and the \fIovector\fP is not big enough to remember the related
1834  has to get additional memory for use during matching. Thus it is usually  substrings, PCRE has to get additional memory for use during matching. Thus it
1835  advisable to supply an \fIovector\fP.  is usually advisable to supply an \fIovector\fP of reasonable size.
1836    .P
1837    There are some cases where zero is returned (indicating vector overflow) when
1838    in fact the vector is exactly the right size for the final match. For example,
1839    consider the pattern
1840    .sp
1841      (a)(?:(b)c|bd)
1842    .sp
1843    If a vector of 6 elements (allowing for only 1 captured substring) is given
1844    with subject string "abd", \fBpcre_exec()\fP will try to set the second
1845    captured string, thereby recording a vector overflow, before failing to match
1846    "c" and backing up to try the second alternative. The zero return, however,
1847    does correctly indicate that the maximum number of slots (namely 2) have been
1848    filled. In similar cases where there is temporary overflow, but the final
1849    number of used slots is actually less than the maximum, a non-zero value is
1850    returned.
1851  .P  .P
1852  Note that \fBpcre_info()\fP can be used to find out how many capturing  The \fBpcre_fullinfo()\fP function can be used to find out how many capturing
1853  subpatterns there are in a compiled pattern. The smallest size for  subpatterns there are in a compiled pattern. The smallest size for
1854  \fIovector\fP that will allow for \fIn\fP captured substrings, in addition to  \fIovector\fP that will allow for \fIn\fP captured substrings, in addition to
1855  the offsets of the substring matched by the whole pattern, is (\fIn\fP+1)*3.  the offsets of the substring matched by the whole pattern, is (\fIn\fP+1)*3.
1856    .P
1857    It is possible for capturing subpattern number \fIn+1\fP to match some part of
1858    the subject when subpattern \fIn\fP has not been used at all. For example, if
1859    the string "abc" is matched against the pattern (a|(z))(bc) the return from the
1860    function is 4, and subpatterns 1 and 3 are matched, but 2 is not. When this
1861    happens, both values in the offset pairs corresponding to unused subpatterns
1862    are set to -1.
1863    .P
1864    Offset values that correspond to unused subpatterns at the end of the
1865    expression are also set to -1. For example, if the string "abc" is matched
1866    against the pattern (abc)(x(yz)?)? subpatterns 2 and 3 are not matched. The
1867    return from the function is 2, because the highest used capturing subpattern
1868    number is 1, and the offsets for for the second and third capturing subpatterns
1869    (assuming the vector is large enough, of course) are set to -1.
1870    .P
1871    \fBNote\fP: Elements in the first two-thirds of \fIovector\fP that do not
1872    correspond to capturing parentheses in the pattern are never changed. That is,
1873    if a pattern contains \fIn\fP capturing parentheses, no more than
1874    \fIovector[0]\fP to \fIovector[2n+1]\fP are set by \fBpcre_exec()\fP. The other
1875    elements (in the first two-thirds) retain whatever values they previously had.
1876    .P
1877    Some convenience functions are provided for extracting the captured substrings
1878    as separate strings. These are described below.
1879    .
1880  .  .
1881  .SS "Return values from \fBpcre_exec()\fP"  .\" HTML <a name="errorlist"></a>
1882    .SS "Error return values from \fBpcre_exec()\fP"
1883  .rs  .rs
1884  .sp  .sp
1885  If \fBpcre_exec()\fP fails, it returns a negative number. The following are  If \fBpcre_exec()\fP fails, it returns a negative number. The following are
# Line 1068  compiled in an environment of one endian Line 1906  compiled in an environment of one endian
1906  other endianness. This is the error that PCRE gives when the magic number is  other endianness. This is the error that PCRE gives when the magic number is
1907  not present.  not present.
1908  .sp  .sp
1909    PCRE_ERROR_UNKNOWN_NODE   (-5)    PCRE_ERROR_UNKNOWN_OPCODE (-5)
1910  .sp  .sp
1911  While running the pattern match, an unknown item was encountered in the  While running the pattern match, an unknown item was encountered in the
1912  compiled pattern. This error could be caused by a bug in PCRE or by overwriting  compiled pattern. This error could be caused by a bug in PCRE or by overwriting
# Line 1081  If a pattern contains back references, b Line 1919  If a pattern contains back references, b
1919  gets a block of memory at the start of matching to use for this purpose. If the  gets a block of memory at the start of matching to use for this purpose. If the
1920  call via \fBpcre_malloc()\fP fails, this error is given. The memory is  call via \fBpcre_malloc()\fP fails, this error is given. The memory is
1921  automatically freed at the end of matching.  automatically freed at the end of matching.
1922    .P
1923    This error is also given if \fBpcre_stack_malloc()\fP fails in
1924    \fBpcre_exec()\fP. This can happen only when PCRE has been compiled with
1925    \fB--disable-stack-for-recursion\fP.
1926  .sp  .sp
1927    PCRE_ERROR_NOSUBSTRING    (-7)    PCRE_ERROR_NOSUBSTRING    (-7)
1928  .sp  .sp
# Line 1090  below). It is never returned by \fBpcre_ Line 1932  below). It is never returned by \fBpcre_
1932  .sp  .sp
1933    PCRE_ERROR_MATCHLIMIT     (-8)    PCRE_ERROR_MATCHLIMIT     (-8)
1934  .sp  .sp
1935  The recursion and backtracking limit, as specified by the \fImatch_limit\fP  The backtracking limit, as specified by the \fImatch_limit\fP field in a
1936  field in a \fBpcre_extra\fP structure (or defaulted) was reached. See the  \fBpcre_extra\fP structure (or defaulted) was reached. See the description
1937  description above.  above.
1938  .sp  .sp
1939    PCRE_ERROR_CALLOUT        (-9)    PCRE_ERROR_CALLOUT        (-9)
1940  .sp  .sp
# Line 1105  documentation for details. Line 1947  documentation for details.
1947  .sp  .sp
1948    PCRE_ERROR_BADUTF8        (-10)    PCRE_ERROR_BADUTF8        (-10)
1949  .sp  .sp
1950  A string that contains an invalid UTF-8 byte sequence was passed as a subject.  A string that contains an invalid UTF-8 byte sequence was passed as a subject,
1951    and the PCRE_NO_UTF8_CHECK option was not set. If the size of the output vector
1952    (\fIovecsize\fP) is at least 2, the byte offset to the start of the the invalid
1953    UTF-8 character is placed in the first element, and a reason code is placed in
1954    the second element. The reason codes are listed in the
1955    .\" HTML <a href="#badutf8reasons">
1956    .\" </a>
1957    following section.
1958    .\"
1959    For backward compatibility, if PCRE_PARTIAL_HARD is set and the problem is a
1960    truncated UTF-8 character at the end of the subject (reason codes 1 to 5),
1961    PCRE_ERROR_SHORTUTF8 is returned instead of PCRE_ERROR_BADUTF8.
1962  .sp  .sp
1963    PCRE_ERROR_BADUTF8_OFFSET (-11)    PCRE_ERROR_BADUTF8_OFFSET (-11)
1964  .sp  .sp
1965  The UTF-8 byte sequence that was passed as a subject was valid, but the value  The UTF-8 byte sequence that was passed as a subject was checked and found to
1966  of \fIstartoffset\fP did not point to the beginning of a UTF-8 character.  be valid (the PCRE_NO_UTF8_CHECK option was not set), but the value of
1967    \fIstartoffset\fP did not point to the beginning of a UTF-8 character or the
1968    end of the subject.
1969  .sp  .sp
1970    PCRE_ERROR_PARTIAL (-12)    PCRE_ERROR_PARTIAL        (-12)
1971  .sp  .sp
1972  The subject string did not match, but it did match partially. See the  The subject string did not match, but it did match partially. See the
1973  .\" HREF  .\" HREF
# Line 1120  The subject string did not match, but it Line 1975  The subject string did not match, but it
1975  .\"  .\"
1976  documentation for details of partial matching.  documentation for details of partial matching.
1977  .sp  .sp
1978    PCRE_ERROR_BAD_PARTIAL (-13)    PCRE_ERROR_BADPARTIAL     (-13)
1979  .sp  .sp
1980  The PCRE_PARTIAL option was used with a compiled pattern containing items that  This code is no longer in use. It was formerly returned when the PCRE_PARTIAL
1981  are not supported for partial matching. See the  option was used with a compiled pattern containing items that were not
1982  .\" HREF  supported for partial matching. From release 8.00 onwards, there are no
1983  \fBpcrepartial\fP  restrictions on partial matching.
 .\"  
 documentation for details of partial matching.  
1984  .sp  .sp
1985    PCRE_ERROR_INTERNAL (-14)    PCRE_ERROR_INTERNAL       (-14)
1986  .sp  .sp
1987  An unexpected internal error has occurred. This error could be caused by a bug  An unexpected internal error has occurred. This error could be caused by a bug
1988  in PCRE or by overwriting of the compiled pattern.  in PCRE or by overwriting of the compiled pattern.
1989  .sp  .sp
1990    PCRE_ERROR_BADCOUNT (-15)    PCRE_ERROR_BADCOUNT       (-15)
1991  .sp  .sp
1992  This error is given if the value of the \fIovecsize\fP argument is negative.  This error is given if the value of the \fIovecsize\fP argument is negative.
1993    .sp
1994      PCRE_ERROR_RECURSIONLIMIT (-21)
1995    .sp
1996    The internal recursion limit, as specified by the \fImatch_limit_recursion\fP
1997    field in a \fBpcre_extra\fP structure (or defaulted) was reached. See the
1998    description above.
1999    .sp
2000      PCRE_ERROR_BADNEWLINE     (-23)
2001    .sp
2002    An invalid combination of PCRE_NEWLINE_\fIxxx\fP options was given.
2003    .sp
2004      PCRE_ERROR_BADOFFSET      (-24)
2005    .sp
2006    The value of \fIstartoffset\fP was negative or greater than the length of the
2007    subject, that is, the value in \fIlength\fP.
2008    .sp
2009      PCRE_ERROR_SHORTUTF8      (-25)
2010    .sp
2011    This error is returned instead of PCRE_ERROR_BADUTF8 when the subject string
2012    ends with a truncated UTF-8 character and the PCRE_PARTIAL_HARD option is set.
2013    Information about the failure is returned as for PCRE_ERROR_BADUTF8. It is in
2014    fact sufficient to detect this case, but this special error code for
2015    PCRE_PARTIAL_HARD precedes the implementation of returned information; it is
2016    retained for backwards compatibility.
2017    .sp
2018      PCRE_ERROR_RECURSELOOP    (-26)
2019    .sp
2020    This error is returned when \fBpcre_exec()\fP detects a recursion loop within
2021    the pattern. Specifically, it means that either the whole pattern or a
2022    subpattern has been called recursively for the second time at the same position
2023    in the subject string. Some simple patterns that might do this are detected and
2024    faulted at compile time, but more complicated cases, in particular mutual
2025    recursions between two different subpatterns, cannot be detected until run
2026    time.
2027    .sp
2028      PCRE_ERROR_JIT_STACKLIMIT (-27)
2029    .sp
2030    This error is returned when a pattern that was successfully studied using the
2031    PCRE_STUDY_JIT_COMPILE option is being matched, but the memory available for
2032    the just-in-time processing stack is not large enough. See the
2033    .\" HREF
2034    \fBpcrejit\fP
2035    .\"
2036    documentation for more details.
2037    .P
2038    Error numbers -16 to -20 and -22 are not used by \fBpcre_exec()\fP.
2039    .
2040    .
2041    .\" HTML <a name="badutf8reasons"></a>
2042    .SS "Reason codes for invalid UTF-8 strings"
2043    .rs
2044    .sp
2045    When \fBpcre_exec()\fP returns either PCRE_ERROR_BADUTF8 or
2046    PCRE_ERROR_SHORTUTF8, and the size of the output vector (\fIovecsize\fP) is at
2047    least 2, the offset of the start of the invalid UTF-8 character is placed in
2048    the first output vector element (\fIovector[0]\fP) and a reason code is placed
2049    in the second element (\fIovector[1]\fP). The reason codes are given names in
2050    the \fBpcre.h\fP header file:
2051    .sp
2052      PCRE_UTF8_ERR1
2053      PCRE_UTF8_ERR2
2054      PCRE_UTF8_ERR3
2055      PCRE_UTF8_ERR4
2056      PCRE_UTF8_ERR5
2057    .sp
2058    The string ends with a truncated UTF-8 character; the code specifies how many
2059    bytes are missing (1 to 5). Although RFC 3629 restricts UTF-8 characters to be
2060    no longer than 4 bytes, the encoding scheme (originally defined by RFC 2279)
2061    allows for up to 6 bytes, and this is checked first; hence the possibility of
2062    4 or 5 missing bytes.
2063    .sp
2064      PCRE_UTF8_ERR6
2065      PCRE_UTF8_ERR7
2066      PCRE_UTF8_ERR8
2067      PCRE_UTF8_ERR9
2068      PCRE_UTF8_ERR10
2069    .sp
2070    The two most significant bits of the 2nd, 3rd, 4th, 5th, or 6th byte of the
2071    character do not have the binary value 0b10 (that is, either the most
2072    significant bit is 0, or the next bit is 1).
2073    .sp
2074      PCRE_UTF8_ERR11
2075      PCRE_UTF8_ERR12
2076    .sp
2077    A character that is valid by the RFC 2279 rules is either 5 or 6 bytes long;
2078    these code points are excluded by RFC 3629.
2079    .sp
2080      PCRE_UTF8_ERR13
2081    .sp
2082    A 4-byte character has a value greater than 0x10fff; these code points are
2083    excluded by RFC 3629.
2084    .sp
2085      PCRE_UTF8_ERR14
2086    .sp
2087    A 3-byte character has a value in the range 0xd800 to 0xdfff; this range of
2088    code points are reserved by RFC 3629 for use with UTF-16, and so are excluded
2089    from UTF-8.
2090    .sp
2091      PCRE_UTF8_ERR15
2092      PCRE_UTF8_ERR16
2093      PCRE_UTF8_ERR17
2094      PCRE_UTF8_ERR18
2095      PCRE_UTF8_ERR19
2096    .sp
2097    A 2-, 3-, 4-, 5-, or 6-byte character is "overlong", that is, it codes for a
2098    value that can be represented by fewer bytes, which is invalid. For example,
2099    the two bytes 0xc0, 0xae give the value 0x2e, whose correct coding uses just
2100    one byte.
2101    .sp
2102      PCRE_UTF8_ERR20
2103    .sp
2104    The two most significant bits of the first byte of a character have the binary
2105    value 0b10 (that is, the most significant bit is 1 and the second is 0). Such a
2106    byte can only validly occur as the second or subsequent byte of a multi-byte
2107    character.
2108    .sp
2109      PCRE_UTF8_ERR21
2110    .sp
2111    The first byte of a character has the value 0xfe or 0xff. These values can
2112    never occur in a valid UTF-8 string.
2113  .  .
2114  .  .
2115  .SH "EXTRACTING CAPTURED SUBSTRINGS BY NUMBER"  .SH "EXTRACTING CAPTURED SUBSTRINGS BY NUMBER"
# Line 1148  This error is given if the value of the Line 2121  This error is given if the value of the
2121  .ti +5n  .ti +5n
2122  .B int \fIbuffersize\fP);  .B int \fIbuffersize\fP);
2123  .PP  .PP
 .br  
2124  .B int pcre_get_substring(const char *\fIsubject\fP, int *\fIovector\fP,  .B int pcre_get_substring(const char *\fIsubject\fP, int *\fIovector\fP,
2125  .ti +5n  .ti +5n
2126  .B int \fIstringcount\fP, int \fIstringnumber\fP,  .B int \fIstringcount\fP, int \fIstringnumber\fP,
2127  .ti +5n  .ti +5n
2128  .B const char **\fIstringptr\fP);  .B const char **\fIstringptr\fP);
2129  .PP  .PP
 .br  
2130  .B int pcre_get_substring_list(const char *\fIsubject\fP,  .B int pcre_get_substring_list(const char *\fIsubject\fP,
2131  .ti +5n  .ti +5n
2132  .B int *\fIovector\fP, int \fIstringcount\fP, "const char ***\fIlistptr\fP);"  .B int *\fIovector\fP, int \fIstringcount\fP, "const char ***\fIlistptr\fP);"
# Line 1166  Captured substrings can be accessed dire Line 2137  Captured substrings can be accessed dire
2137  \fBpcre_get_substring_list()\fP are provided for extracting captured substrings  \fBpcre_get_substring_list()\fP are provided for extracting captured substrings
2138  as new, separate, zero-terminated strings. These functions identify substrings  as new, separate, zero-terminated strings. These functions identify substrings
2139  by number. The next section describes functions for extracting named  by number. The next section describes functions for extracting named
2140  substrings. A substring that contains a binary zero is correctly extracted and  substrings.
2141  has a further zero added on the end, but the result is not, of course,  .P
2142  a C string.  A substring that contains a binary zero is correctly extracted and has a
2143    further zero added on the end, but the result is not, of course, a C string.
2144    However, you can process such a string by referring to the length that is
2145    returned by \fBpcre_copy_substring()\fP and \fBpcre_get_substring()\fP.
2146    Unfortunately, the interface to \fBpcre_get_substring_list()\fP is not adequate
2147    for handling strings containing binary zeros, because the end of the final
2148    string is not independently indicated.
2149  .P  .P
2150  The first three arguments are the same for all three of these functions:  The first three arguments are the same for all three of these functions:
2151  \fIsubject\fP is the subject string that has just been successfully matched,  \fIsubject\fP is the subject string that has just been successfully matched,
# Line 1188  the string is placed in \fIbuffer\fP, wh Line 2165  the string is placed in \fIbuffer\fP, wh
2165  \fIbuffersize\fP, while for \fBpcre_get_substring()\fP a new block of memory is  \fIbuffersize\fP, while for \fBpcre_get_substring()\fP a new block of memory is
2166  obtained via \fBpcre_malloc\fP, and its address is returned via  obtained via \fBpcre_malloc\fP, and its address is returned via
2167  \fIstringptr\fP. The yield of the function is the length of the string, not  \fIstringptr\fP. The yield of the function is the length of the string, not
2168  including the terminating zero, or one of  including the terminating zero, or one of these error codes:
2169  .sp  .sp
2170    PCRE_ERROR_NOMEMORY       (-6)    PCRE_ERROR_NOMEMORY       (-6)
2171  .sp  .sp
# Line 1204  and builds a list of pointers to them. A Line 2181  and builds a list of pointers to them. A
2181  memory that is obtained via \fBpcre_malloc\fP. The address of the memory block  memory that is obtained via \fBpcre_malloc\fP. The address of the memory block
2182  is returned via \fIlistptr\fP, which is also the start of the list of string  is returned via \fIlistptr\fP, which is also the start of the list of string
2183  pointers. The end of the list is marked by a NULL pointer. The yield of the  pointers. The end of the list is marked by a NULL pointer. The yield of the
2184  function is zero if all went well, or  function is zero if all went well, or the error code
2185  .sp  .sp
2186    PCRE_ERROR_NOMEMORY       (-6)    PCRE_ERROR_NOMEMORY       (-6)
2187  .sp  .sp
# Line 1223  a previous call of \fBpcre_get_substring Line 2200  a previous call of \fBpcre_get_substring
2200  \fBpcre_get_substring_list()\fP, respectively. They do nothing more than call  \fBpcre_get_substring_list()\fP, respectively. They do nothing more than call
2201  the function pointed to by \fBpcre_free\fP, which of course could be called  the function pointed to by \fBpcre_free\fP, which of course could be called
2202  directly from a C program. However, PCRE is used in some situations where it is  directly from a C program. However, PCRE is used in some situations where it is
2203  linked via a special interface to another programming language which cannot use  linked via a special interface to another programming language that cannot use
2204  \fBpcre_free\fP directly; it is for these cases that the functions are  \fBpcre_free\fP directly; it is for these cases that the functions are
2205  provided.  provided.
2206  .  .
# Line 1235  provided. Line 2212  provided.
2212  .ti +5n  .ti +5n
2213  .B const char *\fIname\fP);  .B const char *\fIname\fP);
2214  .PP  .PP
 .br  
2215  .B int pcre_copy_named_substring(const pcre *\fIcode\fP,  .B int pcre_copy_named_substring(const pcre *\fIcode\fP,
2216  .ti +5n  .ti +5n
2217  .B const char *\fIsubject\fP, int *\fIovector\fP,  .B const char *\fIsubject\fP, int *\fIovector\fP,
# Line 1244  provided. Line 2220  provided.
2220  .ti +5n  .ti +5n
2221  .B char *\fIbuffer\fP, int \fIbuffersize\fP);  .B char *\fIbuffer\fP, int \fIbuffersize\fP);
2222  .PP  .PP
 .br  
2223  .B int pcre_get_named_substring(const pcre *\fIcode\fP,  .B int pcre_get_named_substring(const pcre *\fIcode\fP,
2224  .ti +5n  .ti +5n
2225  .B const char *\fIsubject\fP, int *\fIovector\fP,  .B const char *\fIsubject\fP, int *\fIovector\fP,
# Line 1258  For example, for this pattern Line 2233  For example, for this pattern
2233  .sp  .sp
2234    (a+)b(?<xxx>\ed+)...    (a+)b(?<xxx>\ed+)...
2235  .sp  .sp
2236  the number of the subpattern called "xxx" is 2. You can find the number from  the number of the subpattern called "xxx" is 2. If the name is known to be
2237  the name by calling \fBpcre_get_stringnumber()\fP. The first argument is the  unique (PCRE_DUPNAMES was not set), you can find the number from the name by
2238  compiled pattern, and the second is the name. The yield of the function is the  calling \fBpcre_get_stringnumber()\fP. The first argument is the compiled
2239    pattern, and the second is the name. The yield of the function is the
2240  subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no subpattern of  subpattern number, or PCRE_ERROR_NOSUBSTRING (-7) if there is no subpattern of
2241  that name.  that name.
2242  .P  .P
# Line 1268  Given the number, you can extract the su Line 2244  Given the number, you can extract the su
2244  functions described in the previous section. For convenience, there are also  functions described in the previous section. For convenience, there are also
2245  two functions that do the whole job.  two functions that do the whole job.
2246  .P  .P
2247  Most of the arguments of \fIpcre_copy_named_substring()\fP and  Most of the arguments of \fBpcre_copy_named_substring()\fP and
2248  \fIpcre_get_named_substring()\fP are the same as those for the similarly named  \fBpcre_get_named_substring()\fP are the same as those for the similarly named
2249  functions that extract by number. As these are described in the previous  functions that extract by number. As these are described in the previous
2250  section, they are not re-described here. There are just two differences:  section, they are not re-described here. There are just two differences:
2251  .P  .P
# Line 1279  pattern. This is needed in order to gain Line 2255  pattern. This is needed in order to gain
2255  translation table.  translation table.
2256  .P  .P
2257  These functions call \fBpcre_get_stringnumber()\fP, and if it succeeds, they  These functions call \fBpcre_get_stringnumber()\fP, and if it succeeds, they
2258  then call \fIpcre_copy_substring()\fP or \fIpcre_get_substring()\fP, as  then call \fBpcre_copy_substring()\fP or \fBpcre_get_substring()\fP, as
2259  appropriate.  appropriate. \fBNOTE:\fP If PCRE_DUPNAMES is set and there are duplicate names,
2260    the behaviour may not be what you want (see the next section).
2261    .P
2262    \fBWarning:\fP If the pattern uses the (?| feature to set up multiple
2263    subpatterns with the same number, as described in the
2264    .\" HTML <a href="pcrepattern.html#dupsubpatternnumber">
2265    .\" </a>
2266    section on duplicate subpattern numbers
2267    .\"
2268    in the
2269    .\" HREF
2270    \fBpcrepattern\fP
2271    .\"
2272    page, you cannot use names to distinguish the different subpatterns, because
2273    names are not included in the compiled code. The matching process uses only
2274    numbers. For this reason, the use of different names for subpatterns of the
2275    same number causes an error at compile time.
2276    .
2277    .
2278    .SH "DUPLICATE SUBPATTERN NAMES"
2279    .rs
2280    .sp
2281    .B int pcre_get_stringtable_entries(const pcre *\fIcode\fP,
2282    .ti +5n
2283    .B const char *\fIname\fP, char **\fIfirst\fP, char **\fIlast\fP);
2284    .PP
2285    When a pattern is compiled with the PCRE_DUPNAMES option, names for subpatterns
2286    are not required to be unique. (Duplicate names are always allowed for
2287    subpatterns with the same number, created by using the (?| feature. Indeed, if
2288    such subpatterns are named, they are required to use the same names.)
2289    .P
2290    Normally, patterns with duplicate names are such that in any one match, only
2291    one of the named subpatterns participates. An example is shown in the
2292    .\" HREF
2293    \fBpcrepattern\fP
2294    .\"
2295    documentation.
2296    .P
2297    When duplicates are present, \fBpcre_copy_named_substring()\fP and
2298    \fBpcre_get_named_substring()\fP return the first substring corresponding to
2299    the given name that is set. If none are set, PCRE_ERROR_NOSUBSTRING (-7) is
2300    returned; no data is returned. The \fBpcre_get_stringnumber()\fP function
2301    returns one of the numbers that are associated with the name, but it is not
2302    defined which it is.
2303    .P
2304    If you want to get full details of all captured substrings for a given name,
2305    you must use the \fBpcre_get_stringtable_entries()\fP function. The first
2306    argument is the compiled pattern, and the second is the name. The third and
2307    fourth are pointers to variables which are updated by the function. After it
2308    has run, they point to the first and last entries in the name-to-number table
2309    for the given name. The function itself returns the length of each entry, or
2310    PCRE_ERROR_NOSUBSTRING (-7) if there are none. The format of the table is
2311    described above in the section entitled \fIInformation about a pattern\fP
2312    .\" HTML <a href="#infoaboutpattern">
2313    .\" </a>
2314    above.
2315    .\"
2316    Given all the relevant entries for the name, you can extract each of their
2317    numbers, and hence the captured data, if any.
2318    .
2319    .
2320    .SH "FINDING ALL POSSIBLE MATCHES"
2321    .rs
2322    .sp
2323    The traditional matching function uses a similar algorithm to Perl, which stops
2324    when it finds the first match, starting at a given point in the subject. If you
2325    want to find all possible matches, or the longest possible match, consider
2326    using the alternative matching function (see below) instead. If you cannot use
2327    the alternative function, but still need to find all possible matches, you
2328    can kludge it up by making use of the callout facility, which is described in
2329    the
2330    .\" HREF
2331    \fBpcrecallout\fP
2332    .\"
2333    documentation.
2334    .P
2335    What you have to do is to insert a callout right at the end of the pattern.
2336    When your callout function is called, extract and save the current matched
2337    substring. Then return 1, which forces \fBpcre_exec()\fP to backtrack and try
2338    other alternatives. Ultimately, when it runs out of matches, \fBpcre_exec()\fP
2339    will yield PCRE_ERROR_NOMATCH.
2340    .
2341    .
2342    .\" HTML <a name="dfamatch"></a>
2343    .SH "MATCHING A PATTERN: THE ALTERNATIVE FUNCTION"
2344    .rs
2345    .sp
2346    .B int pcre_dfa_exec(const pcre *\fIcode\fP, "const pcre_extra *\fIextra\fP,"
2347    .ti +5n
2348    .B "const char *\fIsubject\fP," int \fIlength\fP, int \fIstartoffset\fP,
2349    .ti +5n
2350    .B int \fIoptions\fP, int *\fIovector\fP, int \fIovecsize\fP,
2351    .ti +5n
2352    .B int *\fIworkspace\fP, int \fIwscount\fP);
2353    .P
2354    The function \fBpcre_dfa_exec()\fP is called to match a subject string against
2355    a compiled pattern, using a matching algorithm that scans the subject string
2356    just once, and does not backtrack. This has different characteristics to the
2357    normal algorithm, and is not compatible with Perl. Some of the features of PCRE
2358    patterns are not supported. Nevertheless, there are times when this kind of
2359    matching can be useful. For a discussion of the two matching algorithms, and a
2360    list of features that \fBpcre_dfa_exec()\fP does not support, see the
2361    .\" HREF
2362    \fBpcrematching\fP
2363    .\"
2364    documentation.
2365    .P
2366    The arguments for the \fBpcre_dfa_exec()\fP function are the same as for
2367    \fBpcre_exec()\fP, plus two extras. The \fIovector\fP argument is used in a
2368    different way, and this is described below. The other common arguments are used
2369    in the same way as for \fBpcre_exec()\fP, so their description is not repeated
2370    here.
2371    .P
2372    The two additional arguments provide workspace for the function. The workspace
2373    vector should contain at least 20 elements. It is used for keeping track of
2374    multiple paths through the pattern tree. More workspace will be needed for
2375    patterns and subjects where there are a lot of potential matches.
2376    .P
2377    Here is an example of a simple call to \fBpcre_dfa_exec()\fP:
2378    .sp
2379      int rc;
2380      int ovector[10];
2381      int wspace[20];
2382      rc = pcre_dfa_exec(
2383        re,             /* result of pcre_compile() */
2384        NULL,           /* we didn't study the pattern */
2385        "some string",  /* the subject string */
2386        11,             /* the length of the subject string */
2387        0,              /* start at offset 0 in the subject */
2388        0,              /* default options */
2389        ovector,        /* vector of integers for substring information */
2390        10,             /* number of elements (NOT size in bytes) */
2391        wspace,         /* working space vector */
2392        20);            /* number of elements (NOT size in bytes) */
2393    .
2394    .SS "Option bits for \fBpcre_dfa_exec()\fP"
2395    .rs
2396    .sp
2397    The unused bits of the \fIoptions\fP argument for \fBpcre_dfa_exec()\fP must be
2398    zero. The only bits that may be set are PCRE_ANCHORED, PCRE_NEWLINE_\fIxxx\fP,
2399    PCRE_NOTBOL, PCRE_NOTEOL, PCRE_NOTEMPTY, PCRE_NOTEMPTY_ATSTART,
2400    PCRE_NO_UTF8_CHECK, PCRE_BSR_ANYCRLF, PCRE_BSR_UNICODE, PCRE_NO_START_OPTIMIZE,
2401    PCRE_PARTIAL_HARD, PCRE_PARTIAL_SOFT, PCRE_DFA_SHORTEST, and PCRE_DFA_RESTART.
2402    All but the last four of these are exactly the same as for \fBpcre_exec()\fP,
2403    so their description is not repeated here.
2404    .sp
2405      PCRE_PARTIAL_HARD
2406      PCRE_PARTIAL_SOFT
2407    .sp
2408    These have the same general effect as they do for \fBpcre_exec()\fP, but the
2409    details are slightly different. When PCRE_PARTIAL_HARD is set for
2410    \fBpcre_dfa_exec()\fP, it returns PCRE_ERROR_PARTIAL if the end of the subject
2411    is reached and there is still at least one matching possibility that requires
2412    additional characters. This happens even if some complete matches have also
2413    been found. When PCRE_PARTIAL_SOFT is set, the return code PCRE_ERROR_NOMATCH
2414    is converted into PCRE_ERROR_PARTIAL if the end of the subject is reached,
2415    there have been no complete matches, but there is still at least one matching
2416    possibility. The portion of the string that was inspected when the longest
2417    partial match was found is set as the first matching string in both cases.
2418    There is a more detailed discussion of partial and multi-segment matching, with
2419    examples, in the
2420    .\" HREF
2421    \fBpcrepartial\fP
2422    .\"
2423    documentation.
2424    .sp
2425      PCRE_DFA_SHORTEST
2426    .sp
2427    Setting the PCRE_DFA_SHORTEST option causes the matching algorithm to stop as
2428    soon as it has found one match. Because of the way the alternative algorithm
2429    works, this is necessarily the shortest possible match at the first possible
2430    matching point in the subject string.
2431    .sp
2432      PCRE_DFA_RESTART
2433    .sp
2434    When \fBpcre_dfa_exec()\fP returns a partial match, it is possible to call it
2435    again, with additional subject characters, and have it continue with the same
2436    match. The PCRE_DFA_RESTART option requests this action; when it is set, the
2437    \fIworkspace\fP and \fIwscount\fP options must reference the same vector as
2438    before because data about the match so far is left in them after a partial
2439    match. There is more discussion of this facility in the
2440    .\" HREF
2441    \fBpcrepartial\fP
2442    .\"
2443    documentation.
2444    .
2445    .
2446    .SS "Successful returns from \fBpcre_dfa_exec()\fP"
2447    .rs
2448    .sp
2449    When \fBpcre_dfa_exec()\fP succeeds, it may have matched more than one
2450    substring in the subject. Note, however, that all the matches from one run of
2451    the function start at the same point in the subject. The shorter matches are
2452    all initial substrings of the longer matches. For example, if the pattern
2453    .sp
2454      <.*>
2455    .sp
2456    is matched against the string
2457    .sp
2458      This is <something> <something else> <something further> no more
2459    .sp
2460    the three matched strings are
2461    .sp
2462      <something>
2463      <something> <something else>
2464      <something> <something else> <something further>
2465    .sp
2466    On success, the yield of the function is a number greater than zero, which is
2467    the number of matched substrings. The substrings themselves are returned in
2468    \fIovector\fP. Each string uses two elements; the first is the offset to the
2469    start, and the second is the offset to the end. In fact, all the strings have
2470    the same start offset. (Space could have been saved by giving this only once,
2471    but it was decided to retain some compatibility with the way \fBpcre_exec()\fP
2472    returns data, even though the meaning of the strings is different.)
2473  .P  .P
2474  .in 0  The strings are returned in reverse order of length; that is, the longest
2475  Last updated: 09 September 2004  matching string is given first. If there were too many matches to fit into
2476  .br  \fIovector\fP, the yield of the function is zero, and the vector is filled with
2477  Copyright (c) 1997-2004 University of Cambridge.  the longest matches. Unlike \fBpcre_exec()\fP, \fBpcre_dfa_exec()\fP can use
2478    the entire \fIovector\fP for returning matched strings.
2479    .
2480    .
2481    .SS "Error returns from \fBpcre_dfa_exec()\fP"
2482    .rs
2483    .sp
2484    The \fBpcre_dfa_exec()\fP function returns a negative number when it fails.
2485    Many of the errors are the same as for \fBpcre_exec()\fP, and these are
2486    described
2487    .\" HTML <a href="#errorlist">
2488    .\" </a>
2489    above.
2490    .\"
2491    There are in addition the following errors that are specific to
2492    \fBpcre_dfa_exec()\fP:
2493    .sp
2494      PCRE_ERROR_DFA_UITEM      (-16)
2495    .sp
2496    This return is given if \fBpcre_dfa_exec()\fP encounters an item in the pattern
2497    that it does not support, for instance, the use of \eC or a back reference.
2498    .sp
2499      PCRE_ERROR_DFA_UCOND      (-17)
2500    .sp
2501    This return is given if \fBpcre_dfa_exec()\fP encounters a condition item that
2502    uses a back reference for the condition, or a test for recursion in a specific
2503    group. These are not supported.
2504    .sp
2505      PCRE_ERROR_DFA_UMLIMIT    (-18)
2506    .sp
2507    This return is given if \fBpcre_dfa_exec()\fP is called with an \fIextra\fP
2508    block that contains a setting of the \fImatch_limit\fP or
2509    \fImatch_limit_recursion\fP fields. This is not supported (these fields are
2510    meaningless for DFA matching).
2511    .sp
2512      PCRE_ERROR_DFA_WSSIZE     (-19)
2513    .sp
2514    This return is given if \fBpcre_dfa_exec()\fP runs out of space in the
2515    \fIworkspace\fP vector.
2516    .sp
2517      PCRE_ERROR_DFA_RECURSE    (-20)
2518    .sp
2519    When a recursive subpattern is processed, the matching function calls itself
2520    recursively, using private vectors for \fIovector\fP and \fIworkspace\fP. This
2521    error is given if the output vector is not large enough. This should be
2522    extremely rare, as a vector of size 1000 is used.
2523    .
2524    .
2525    .SH "SEE ALSO"
2526    .rs
2527    .sp
2528    \fBpcrebuild\fP(3), \fBpcrecallout\fP(3), \fBpcrecpp(3)\fP(3),
2529    \fBpcrematching\fP(3), \fBpcrepartial\fP(3), \fBpcreposix\fP(3),
2530    \fBpcreprecompile\fP(3), \fBpcresample\fP(3), \fBpcrestack\fP(3).
2531    .
2532    .
2533    .SH AUTHOR
2534    .rs
2535    .sp
2536    .nf
2537    Philip Hazel
2538    University Computing Service
2539    Cambridge CB2 3QH, England.
2540    .fi
2541    .
2542    .
2543    .SH REVISION
2544    .rs
2545    .sp
2546    .nf
2547    Last updated: 14 November 2011
2548    Copyright (c) 1997-2011 University of Cambridge.
2549    .fi

Legend:
Removed from v.75  
changed lines
  Added in v.745

  ViewVC Help
Powered by ViewVC 1.1.5