/[pcre]/code/trunk/pcre_internal.h
ViewVC logotype

Contents of /code/trunk/pcre_internal.h

Parent Directory Parent Directory | Revision Log Revision Log


Revision 199 - (hide annotations) (download)
Tue Jul 31 14:39:09 2007 UTC (7 years, 1 month ago) by ph10
File MIME type: text/plain
File size: 44944 byte(s)
Daniel's patch for config.h and Windows DLL declarations (not fully working).

1 nigel 77 /*************************************************
2     * Perl-Compatible Regular Expressions *
3     *************************************************/
4    
5    
6     /* PCRE is a library of functions to support regular expressions whose syntax
7     and semantics are as close as possible to those of the Perl 5 language.
8    
9     Written by Philip Hazel
10 ph10 117 Copyright (c) 1997-2007 University of Cambridge
11 nigel 77
12     -----------------------------------------------------------------------------
13     Redistribution and use in source and binary forms, with or without
14     modification, are permitted provided that the following conditions are met:
15    
16     * Redistributions of source code must retain the above copyright notice,
17     this list of conditions and the following disclaimer.
18    
19     * Redistributions in binary form must reproduce the above copyright
20     notice, this list of conditions and the following disclaimer in the
21     documentation and/or other materials provided with the distribution.
22    
23     * Neither the name of the University of Cambridge nor the names of its
24     contributors may be used to endorse or promote products derived from
25     this software without specific prior written permission.
26    
27     THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
28     AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
29     IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
30     ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
31     LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
32     CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
33     SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
34     INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
35     CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
36     ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
37     POSSIBILITY OF SUCH DAMAGE.
38     -----------------------------------------------------------------------------
39     */
40    
41     /* This header contains definitions that are shared between the different
42     modules, but which are not relevant to the exported API. This includes some
43     functions whose names all begin with "_pcre_". */
44    
45 nigel 85 #ifndef PCRE_INTERNAL_H
46     #define PCRE_INTERNAL_H
47 nigel 77
48     /* Define DEBUG to get debugging output on stdout. */
49    
50 nigel 85 #if 0
51 nigel 77 #define DEBUG
52 nigel 85 #endif
53 nigel 77
54     /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef
55     inline, and there are *still* stupid compilers about that don't like indented
56     pre-processor statements, or at least there were when I first wrote this. After
57 nigel 93 all, it had only been about 10 years then...
58 nigel 77
59 nigel 93 It turns out that the Mac Debugging.h header also defines the macro DPRINTF, so
60     be absolutely sure we get our version. */
61    
62     #undef DPRINTF
63 nigel 77 #ifdef DEBUG
64     #define DPRINTF(p) printf p
65     #else
66 nigel 93 #define DPRINTF(p) /* Nothing */
67 nigel 77 #endif
68    
69    
70     /* Standard C headers plus the external interface definition. The only time
71     setjmp and stdarg are used is when NO_RECURSE is set. */
72    
73     #include <ctype.h>
74     #include <limits.h>
75     #include <setjmp.h>
76     #include <stdarg.h>
77     #include <stddef.h>
78     #include <stdio.h>
79     #include <stdlib.h>
80     #include <string.h>
81    
82 ph10 145 /* When compiling a DLL for Windows, the exported symbols have to be declared
83     using some MS magic. I found some useful information on this web page:
84     http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
85 ph10 154 information there, using __declspec(dllexport) without "extern" we have a
86 ph10 145 definition; with "extern" we have a declaration. The settings here override the
87     setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
88 ph10 154 which is all that is needed for applications (they just import the symbols). We
89     use:
90 ph10 145
91     PCRE_EXP_DECL for declarations
92     PCRE_EXP_DEFN for definitions of exported functions
93     PCRE_EXP_DATA_DEFN for definitions of exported variables
94    
95     The reason for the two DEFN macros is that in non-Windows environments, one
96     does not want to have "extern" before variable definitions because it leads to
97     compiler warnings. So we distinguish between functions and variables. In
98     Windows, the two should always be the same.
99    
100     The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
101     which is an application, but needs to import this file in order to "peek" at
102 ph10 154 internals, can #include pcre.h first to get an application's-eye view.
103 ph10 145
104     In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
105     special-purpose environments) might want to stick other stuff in front of
106     exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
107     PCRE_EXP_DATA_DEFN only if they are not already set. */
108    
109     #ifndef PCRE_EXP_DECL
110     # ifdef _WIN32
111 ph10 199 # ifndef PCRE_STATIC
112 ph10 145 # define PCRE_EXP_DECL extern __declspec(dllexport)
113     # define PCRE_EXP_DEFN __declspec(dllexport)
114     # define PCRE_EXP_DATA_DEFN __declspec(dllexport)
115     # else
116     # define PCRE_EXP_DECL extern
117     # define PCRE_EXP_DEFN
118     # define PCRE_EXP_DATA_DEFN
119     # endif
120     # else
121     # ifdef __cplusplus
122     # define PCRE_EXP_DECL extern "C"
123     # else
124     # define PCRE_EXP_DECL extern
125     # endif
126     # ifndef PCRE_EXP_DEFN
127     # define PCRE_EXP_DEFN PCRE_EXP_DECL
128     # endif
129     # ifndef PCRE_EXP_DATA_DEFN
130     # define PCRE_EXP_DATA_DEFN
131     # endif
132     # endif
133 nigel 77 #endif
134    
135     /* We need to have types that specify unsigned 16-bit and 32-bit integers. We
136     cannot determine these outside the compilation (e.g. by running a program as
137     part of "configure") because PCRE is often cross-compiled for use on other
138     systems. Instead we make use of the maximum sizes that are available at
139     preprocessor time in standard C environments. */
140    
141     #if USHRT_MAX == 65535
142     typedef unsigned short pcre_uint16;
143     #elif UINT_MAX == 65535
144     typedef unsigned int pcre_uint16;
145     #else
146     #error Cannot determine a type for 16-bit unsigned integers
147     #endif
148    
149     #if UINT_MAX == 4294967295
150     typedef unsigned int pcre_uint32;
151     #elif ULONG_MAX == 4294967295
152     typedef unsigned long int pcre_uint32;
153     #else
154     #error Cannot determine a type for 32-bit unsigned integers
155     #endif
156    
157     /* All character handling must be done as unsigned characters. Otherwise there
158     are problems with top-bit-set characters and functions such as isspace().
159     However, we leave the interface to the outside world as char *, because that
160     should make things easier for callers. We define a short type for unsigned char
161     to save lots of typing. I tried "uchar", but it causes problems on Digital
162     Unix, where it is defined in sys/types, so use "uschar" instead. */
163    
164     typedef unsigned char uschar;
165    
166 nigel 93 /* This is an unsigned int value that no character can ever have. UTF-8
167     characters only go up to 0x7fffffff (though Unicode doesn't go beyond
168     0x0010ffff). */
169 nigel 91
170 nigel 93 #define NOTACHAR 0xffffffff
171    
172     /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
173 ph10 149 "any" and "anycrlf" at present). The following macros are used to package up
174     testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
175     modules to indicate in which datablock the parameters exist, and what the
176     start/end of string field names are. */
177 nigel 93
178 ph10 149 #define NLTYPE_FIXED 0 /* Newline is a fixed length string */
179     #define NLTYPE_ANY 1 /* Newline is any Unicode line ending */
180     #define NLTYPE_ANYCRLF 2 /* Newline is CR, LF, or CRLF */
181 nigel 93
182     /* This macro checks for a newline at the given position */
183    
184 nigel 91 #define IS_NEWLINE(p) \
185 nigel 93 ((NLBLOCK->nltype != NLTYPE_FIXED)? \
186     ((p) < NLBLOCK->PSEND && \
187 ph10 149 _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
188     utf8)) \
189 nigel 93 : \
190     ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
191     (p)[0] == NLBLOCK->nl[0] && \
192     (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
193     ) \
194     )
195 nigel 91
196 nigel 93 /* This macro checks for a newline immediately preceding the given position */
197    
198     #define WAS_NEWLINE(p) \
199     ((NLBLOCK->nltype != NLTYPE_FIXED)? \
200     ((p) > NLBLOCK->PSSTART && \
201 ph10 149 _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
202     &(NLBLOCK->nllen), utf8)) \
203 nigel 93 : \
204     ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
205     (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
206     (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
207     ) \
208     )
209    
210 nigel 87 /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
211     with a custom type. This makes it possible, for example, to allow pcre_exec()
212     to process subject strings that are discontinuous by using a smart pointer
213     class. It must always be possible to inspect all of the subject string in
214     pcre_exec() because of the way it backtracks. Two macros are required in the
215     normal case, for sign-unspecified and unsigned char pointers. The former is
216     used for the external interface and appears in pcre.h, which is why its name
217     must begin with PCRE_. */
218    
219     #ifdef CUSTOM_SUBJECT_PTR
220     #define PCRE_SPTR CUSTOM_SUBJECT_PTR
221     #define USPTR CUSTOM_SUBJECT_PTR
222     #else
223     #define PCRE_SPTR const char *
224     #define USPTR const unsigned char *
225     #endif
226    
227 ph10 145
228    
229 nigel 85 /* Include the public PCRE header and the definitions of UCP character property
230     values. */
231 nigel 77
232 ph10 189 #include "pcre.h"
233 nigel 77 #include "ucp.h"
234    
235     /* When compiling for use with the Virtual Pascal compiler, these functions
236     need to have their names changed. PCRE must be compiled with the -DVPCOMPAT
237     option on the command line. */
238    
239     #ifdef VPCOMPAT
240 ph10 101 #define strlen(s) _strlen(s)
241 nigel 77 #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
242 ph10 101 #define memcmp(s,c,n) _memcmp(s,c,n)
243 nigel 77 #define memcpy(d,s,n) _memcpy(d,s,n)
244     #define memmove(d,s,n) _memmove(d,s,n)
245     #define memset(s,c,n) _memset(s,c,n)
246     #else /* VPCOMPAT */
247    
248     /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
249     define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
250     is set. Otherwise, include an emulating function for those systems that have
251 ph10 118 neither (there some non-Unix environments where this is the case). */
252 nigel 77
253 ph10 137 #ifndef HAVE_MEMMOVE
254 nigel 77 #undef memmove /* some systems may have a macro */
255 ph10 137 #ifdef HAVE_BCOPY
256 nigel 77 #define memmove(a, b, c) bcopy(b, a, c)
257     #else /* HAVE_BCOPY */
258 nigel 91 static void *
259 ph10 108 pcre_memmove(void *d, const void *s, size_t n)
260 nigel 77 {
261 nigel 87 size_t i;
262 ph10 108 unsigned char *dest = (unsigned char *)d;
263     const unsigned char *src = (const unsigned char *)s;
264 ph10 118 if (dest > src)
265     {
266     dest += n;
267     src += n;
268     for (i = 0; i < n; ++i) *(--dest) = *(--src);
269     return (void *)dest;
270     }
271     else
272     {
273     for (i = 0; i < n; ++i) *dest++ = *src++;
274 ph10 123 return (void *)(dest - n);
275     }
276 nigel 77 }
277     #define memmove(a, b, c) pcre_memmove(a, b, c)
278     #endif /* not HAVE_BCOPY */
279     #endif /* not HAVE_MEMMOVE */
280     #endif /* not VPCOMPAT */
281    
282    
283     /* PCRE keeps offsets in its compiled code as 2-byte quantities (always stored
284     in big-endian order) by default. These are used, for example, to link from the
285     start of a subpattern to its alternatives and its end. The use of 2 bytes per
286     offset limits the size of the compiled regex to around 64K, which is big enough
287     for almost everybody. However, I received a request for an even bigger limit.
288     For this reason, and also to make the code easier to maintain, the storing and
289     loading of offsets from the byte string is now handled by the macros that are
290     defined here.
291    
292     The macros are controlled by the value of LINK_SIZE. This defaults to 2 in
293     the config.h file, but can be overridden by using -D on the command line. This
294     is automated on Unix systems via the "configure" command. */
295    
296     #if LINK_SIZE == 2
297    
298     #define PUT(a,n,d) \
299     (a[n] = (d) >> 8), \
300     (a[(n)+1] = (d) & 255)
301    
302     #define GET(a,n) \
303     (((a)[n] << 8) | (a)[(n)+1])
304    
305     #define MAX_PATTERN_SIZE (1 << 16)
306    
307    
308     #elif LINK_SIZE == 3
309    
310     #define PUT(a,n,d) \
311     (a[n] = (d) >> 16), \
312     (a[(n)+1] = (d) >> 8), \
313     (a[(n)+2] = (d) & 255)
314    
315     #define GET(a,n) \
316     (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2])
317    
318     #define MAX_PATTERN_SIZE (1 << 24)
319    
320    
321     #elif LINK_SIZE == 4
322    
323     #define PUT(a,n,d) \
324     (a[n] = (d) >> 24), \
325     (a[(n)+1] = (d) >> 16), \
326     (a[(n)+2] = (d) >> 8), \
327     (a[(n)+3] = (d) & 255)
328    
329     #define GET(a,n) \
330     (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
331    
332     #define MAX_PATTERN_SIZE (1 << 30) /* Keep it positive */
333    
334    
335     #else
336     #error LINK_SIZE must be either 2, 3, or 4
337     #endif
338    
339    
340     /* Convenience macro defined in terms of the others */
341    
342     #define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE
343    
344    
345     /* PCRE uses some other 2-byte quantities that do not change when the size of
346     offsets changes. There are used for repeat counts and for other things such as
347     capturing parenthesis numbers in back references. */
348    
349     #define PUT2(a,n,d) \
350     a[n] = (d) >> 8; \
351     a[(n)+1] = (d) & 255
352    
353     #define GET2(a,n) \
354     (((a)[n] << 8) | (a)[(n)+1])
355    
356     #define PUT2INC(a,n,d) PUT2(a,n,d), a += 2
357    
358    
359     /* When UTF-8 encoding is being used, a character is no longer just a single
360     byte. The macros for character handling generate simple sequences when used in
361     byte-mode, and more complicated ones for UTF-8 characters. */
362    
363     #ifndef SUPPORT_UTF8
364     #define GETCHAR(c, eptr) c = *eptr;
365     #define GETCHARTEST(c, eptr) c = *eptr;
366     #define GETCHARINC(c, eptr) c = *eptr++;
367     #define GETCHARINCTEST(c, eptr) c = *eptr++;
368     #define GETCHARLEN(c, eptr, len) c = *eptr;
369     #define BACKCHAR(eptr)
370    
371     #else /* SUPPORT_UTF8 */
372    
373     /* Get the next UTF-8 character, not advancing the pointer. This is called when
374     we know we are in UTF-8 mode. */
375    
376     #define GETCHAR(c, eptr) \
377     c = *eptr; \
378 nigel 93 if (c >= 0xc0) \
379 nigel 77 { \
380     int gcii; \
381     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
382     int gcss = 6*gcaa; \
383     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
384     for (gcii = 1; gcii <= gcaa; gcii++) \
385     { \
386     gcss -= 6; \
387     c |= (eptr[gcii] & 0x3f) << gcss; \
388     } \
389     }
390    
391     /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
392     pointer. */
393    
394     #define GETCHARTEST(c, eptr) \
395     c = *eptr; \
396 nigel 93 if (utf8 && c >= 0xc0) \
397 nigel 77 { \
398     int gcii; \
399     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
400     int gcss = 6*gcaa; \
401     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
402     for (gcii = 1; gcii <= gcaa; gcii++) \
403     { \
404     gcss -= 6; \
405     c |= (eptr[gcii] & 0x3f) << gcss; \
406     } \
407     }
408    
409     /* Get the next UTF-8 character, advancing the pointer. This is called when we
410     know we are in UTF-8 mode. */
411    
412     #define GETCHARINC(c, eptr) \
413     c = *eptr++; \
414 nigel 93 if (c >= 0xc0) \
415 nigel 77 { \
416     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
417     int gcss = 6*gcaa; \
418     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
419     while (gcaa-- > 0) \
420     { \
421     gcss -= 6; \
422     c |= (*eptr++ & 0x3f) << gcss; \
423     } \
424     }
425    
426     /* Get the next character, testing for UTF-8 mode, and advancing the pointer */
427    
428     #define GETCHARINCTEST(c, eptr) \
429     c = *eptr++; \
430 nigel 93 if (utf8 && c >= 0xc0) \
431 nigel 77 { \
432     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
433     int gcss = 6*gcaa; \
434     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
435     while (gcaa-- > 0) \
436     { \
437     gcss -= 6; \
438     c |= (*eptr++ & 0x3f) << gcss; \
439     } \
440     }
441    
442     /* Get the next UTF-8 character, not advancing the pointer, incrementing length
443     if there are extra bytes. This is called when we know we are in UTF-8 mode. */
444    
445     #define GETCHARLEN(c, eptr, len) \
446     c = *eptr; \
447 nigel 93 if (c >= 0xc0) \
448 nigel 77 { \
449     int gcii; \
450     int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
451     int gcss = 6*gcaa; \
452     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
453     for (gcii = 1; gcii <= gcaa; gcii++) \
454     { \
455     gcss -= 6; \
456     c |= (eptr[gcii] & 0x3f) << gcss; \
457     } \
458     len += gcaa; \
459     }
460    
461     /* If the pointer is not at the start of a character, move it back until
462     it is. Called only in UTF-8 mode. */
463    
464     #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;
465    
466     #endif
467    
468    
469     /* In case there is no definition of offsetof() provided - though any proper
470     Standard C system should have one. */
471    
472     #ifndef offsetof
473     #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
474     #endif
475    
476    
477     /* These are the public options that can change during matching. */
478    
479     #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
480    
481 nigel 91 /* Private options flags start at the most significant end of the four bytes.
482     The public options defined in pcre.h start at the least significant end. Make
483     sure they don't overlap! The bits are getting a bit scarce now -- when we run
484     out, there is a dummy word in the structure that could be used for the private
485     bits. */
486 nigel 77
487 nigel 91 #define PCRE_NOPARTIAL 0x80000000 /* can't use partial with this regex */
488 nigel 77 #define PCRE_FIRSTSET 0x40000000 /* first_byte is set */
489     #define PCRE_REQCHSET 0x20000000 /* req_byte is set */
490     #define PCRE_STARTLINE 0x10000000 /* start after \n for multiline */
491 nigel 91 #define PCRE_JCHANGED 0x08000000 /* j option changes within regex */
492 nigel 77
493     /* Options for the "extra" block produced by pcre_study(). */
494    
495     #define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */
496    
497     /* Masks for identifying the public options that are permitted at compile
498     time, run time, or study time, respectively. */
499    
500 ph10 149 #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
501     PCRE_NEWLINE_ANYCRLF)
502 nigel 93
503 nigel 77 #define PUBLIC_OPTIONS \
504     (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
505     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
506 nigel 91 PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
507 nigel 93 PCRE_DUPNAMES|PCRE_NEWLINE_BITS)
508 nigel 77
509     #define PUBLIC_EXEC_OPTIONS \
510     (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
511 nigel 93 PCRE_PARTIAL|PCRE_NEWLINE_BITS)
512 nigel 77
513     #define PUBLIC_DFA_EXEC_OPTIONS \
514     (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
515 nigel 93 PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)
516 nigel 77
517     #define PUBLIC_STUDY_OPTIONS 0 /* None defined */
518    
519     /* Magic number to provide a small check against being handed junk. Also used
520     to detect whether a pattern was compiled on a host of different endianness. */
521    
522     #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
523    
524     /* Negative values for the firstchar and reqchar variables */
525    
526     #define REQ_UNSET (-2)
527     #define REQ_NONE (-1)
528    
529     /* The maximum remaining length of subject we are prepared to search for a
530     req_byte match. */
531    
532     #define REQ_BYTE_MAX 1000
533    
534     /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a
535     variable-length repeat, or a anything other than literal characters. */
536    
537     #define REQ_CASELESS 0x0100 /* indicates caselessness */
538     #define REQ_VARY 0x0200 /* reqbyte followed non-literal item */
539    
540     /* Miscellaneous definitions */
541    
542     typedef int BOOL;
543    
544     #define FALSE 0
545     #define TRUE 1
546    
547 nigel 93 /* Escape items that are just an encoding of a particular data value. */
548 nigel 77
549     #ifndef ESC_e
550     #define ESC_e 27
551     #endif
552    
553     #ifndef ESC_f
554     #define ESC_f '\f'
555     #endif
556    
557     #ifndef ESC_n
558 nigel 93 #define ESC_n '\n'
559 nigel 77 #endif
560    
561     #ifndef ESC_r
562     #define ESC_r '\r'
563     #endif
564    
565     /* We can't officially use ESC_t because it is a POSIX reserved identifier
566     (presumably because of all the others like size_t). */
567    
568     #ifndef ESC_tee
569     #define ESC_tee '\t'
570     #endif
571    
572 nigel 87 /* Codes for different types of Unicode property */
573    
574     #define PT_ANY 0 /* Any property - matches all chars */
575     #define PT_LAMP 1 /* L& - the union of Lu, Ll, Lt */
576     #define PT_GC 2 /* General characteristic (e.g. L) */
577     #define PT_PC 3 /* Particular characteristic (e.g. Lu) */
578     #define PT_SC 4 /* Script (e.g. Han) */
579    
580     /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
581     contain UTF-8 characters with values greater than 255. */
582    
583     #define XCL_NOT 0x01 /* Flag: this is a negative class */
584     #define XCL_MAP 0x02 /* Flag: a 32-byte map is present */
585    
586     #define XCL_END 0 /* Marks end of individual items */
587     #define XCL_SINGLE 1 /* Single item (one multibyte char) follows */
588     #define XCL_RANGE 2 /* A range (two multibyte chars) follows */
589     #define XCL_PROP 3 /* Unicode property (2-byte property code follows) */
590     #define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */
591    
592 nigel 77 /* These are escaped items that aren't just an encoding of a particular data
593     value such as \n. They must have non-zero values, as check_escape() returns
594     their negation. Also, they must appear in the same order as in the opcode
595     definitions below, up to ESC_z. There's a dummy for OP_ANY because it
596     corresponds to "." rather than an escape sequence. The final one must be
597 nigel 93 ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).
598     There are two tests in the code for an escape greater than ESC_b and less than
599     ESC_Z to detect the types that may be repeated. These are the types that
600     consume characters. If any new escapes are put in between that don't consume a
601 nigel 77 character, that code will have to change. */
602    
603 ph10 172 enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
604 ph10 178 ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h,
605     ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_k, ESC_REF };
606 nigel 77
607 nigel 93
608 nigel 77 /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
609     that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
610     OP_EOD must correspond in order to the list of escapes immediately above.
611    
612 nigel 93 To keep stored, compiled patterns compatible, new opcodes should be added
613     immediately before OP_BRA, where (since release 7.0) a gap is left for this
614     purpose.
615    
616     *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
617     that follow must also be updated to match. There is also a table called
618     "coptable" in pcre_dfa_exec.c that must be updated. */
619    
620 nigel 77 enum {
621     OP_END, /* 0 End of pattern */
622    
623     /* Values corresponding to backslashed metacharacters */
624    
625     OP_SOD, /* 1 Start of data: \A */
626     OP_SOM, /* 2 Start of match (subject + offset): \G */
627 ph10 172 OP_SET_SOM, /* 3 Set start of match (\K) */
628 ph10 168 OP_NOT_WORD_BOUNDARY, /* 4 \B */
629     OP_WORD_BOUNDARY, /* 5 \b */
630     OP_NOT_DIGIT, /* 6 \D */
631     OP_DIGIT, /* 7 \d */
632     OP_NOT_WHITESPACE, /* 8 \S */
633     OP_WHITESPACE, /* 9 \s */
634     OP_NOT_WORDCHAR, /* 10 \W */
635     OP_WORDCHAR, /* 11 \w */
636     OP_ANY, /* 12 Match any character */
637     OP_ANYBYTE, /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */
638     OP_NOTPROP, /* 14 \P (not Unicode property) */
639     OP_PROP, /* 15 \p (Unicode property) */
640     OP_ANYNL, /* 16 \R (any newline sequence) */
641 ph10 178 OP_NOT_HSPACE, /* 17 \H (not horizontal whitespace) */
642     OP_HSPACE, /* 18 \h (horizontal whitespace) */
643     OP_NOT_VSPACE, /* 19 \V (not vertical whitespace) */
644     OP_VSPACE, /* 20 \v (vertical whitespace) */
645     OP_EXTUNI, /* 21 \X (extended Unicode sequence */
646     OP_EODN, /* 22 End of data or \n at end of data: \Z. */
647     OP_EOD, /* 23 End of data: \z */
648 nigel 77
649 ph10 178 OP_OPT, /* 24 Set runtime options */
650     OP_CIRC, /* 25 Start of line - varies with multiline switch */
651     OP_DOLL, /* 26 End of line - varies with multiline switch */
652     OP_CHAR, /* 27 Match one character, casefully */
653     OP_CHARNC, /* 28 Match one character, caselessly */
654     OP_NOT, /* 29 Match one character, not the following one */
655 nigel 77
656 ph10 178 OP_STAR, /* 30 The maximizing and minimizing versions of */
657     OP_MINSTAR, /* 31 these six opcodes must come in pairs, with */
658     OP_PLUS, /* 32 the minimizing one second. */
659     OP_MINPLUS, /* 33 This first set applies to single characters.*/
660     OP_QUERY, /* 34 */
661     OP_MINQUERY, /* 35 */
662 nigel 77
663 ph10 178 OP_UPTO, /* 36 From 0 to n matches */
664     OP_MINUPTO, /* 37 */
665     OP_EXACT, /* 38 Exactly n matches */
666 nigel 77
667 ph10 178 OP_POSSTAR, /* 39 Possessified star */
668     OP_POSPLUS, /* 40 Possessified plus */
669     OP_POSQUERY, /* 41 Posesssified query */
670     OP_POSUPTO, /* 42 Possessified upto */
671 nigel 77
672 ph10 178 OP_NOTSTAR, /* 43 The maximizing and minimizing versions of */
673     OP_NOTMINSTAR, /* 44 these six opcodes must come in pairs, with */
674     OP_NOTPLUS, /* 45 the minimizing one second. They must be in */
675     OP_NOTMINPLUS, /* 46 exactly the same order as those above. */
676     OP_NOTQUERY, /* 47 This set applies to "not" single characters. */
677     OP_NOTMINQUERY, /* 48 */
678 nigel 77
679 ph10 178 OP_NOTUPTO, /* 49 From 0 to n matches */
680     OP_NOTMINUPTO, /* 50 */
681     OP_NOTEXACT, /* 51 Exactly n matches */
682 nigel 93
683 ph10 178 OP_NOTPOSSTAR, /* 52 Possessified versions */
684     OP_NOTPOSPLUS, /* 53 */
685     OP_NOTPOSQUERY, /* 54 */
686     OP_NOTPOSUPTO, /* 55 */
687 nigel 93
688 ph10 178 OP_TYPESTAR, /* 56 The maximizing and minimizing versions of */
689     OP_TYPEMINSTAR, /* 57 these six opcodes must come in pairs, with */
690     OP_TYPEPLUS, /* 58 the minimizing one second. These codes must */
691     OP_TYPEMINPLUS, /* 59 be in exactly the same order as those above. */
692     OP_TYPEQUERY, /* 60 This set applies to character types such as \d */
693     OP_TYPEMINQUERY, /* 61 */
694 nigel 93
695 ph10 178 OP_TYPEUPTO, /* 62 From 0 to n matches */
696     OP_TYPEMINUPTO, /* 63 */
697     OP_TYPEEXACT, /* 64 Exactly n matches */
698 nigel 93
699 ph10 178 OP_TYPEPOSSTAR, /* 65 Possessified versions */
700     OP_TYPEPOSPLUS, /* 66 */
701     OP_TYPEPOSQUERY, /* 67 */
702     OP_TYPEPOSUPTO, /* 68 */
703 nigel 93
704 ph10 178 OP_CRSTAR, /* 69 The maximizing and minimizing versions of */
705     OP_CRMINSTAR, /* 70 all these opcodes must come in pairs, with */
706     OP_CRPLUS, /* 71 the minimizing one second. These codes must */
707     OP_CRMINPLUS, /* 72 be in exactly the same order as those above. */
708     OP_CRQUERY, /* 73 These are for character classes and back refs */
709     OP_CRMINQUERY, /* 74 */
710     OP_CRRANGE, /* 75 These are different to the three sets above. */
711     OP_CRMINRANGE, /* 76 */
712 nigel 93
713 ph10 178 OP_CLASS, /* 77 Match a character class, chars < 256 only */
714     OP_NCLASS, /* 78 Same, but the bitmap was created from a negative
715 nigel 77 class - the difference is relevant only when a UTF-8
716     character > 255 is encountered. */
717    
718 ph10 178 OP_XCLASS, /* 79 Extended class for handling UTF-8 chars within the
719 nigel 77 class. This does both positive and negative. */
720    
721 ph10 178 OP_REF, /* 80 Match a back reference */
722     OP_RECURSE, /* 81 Match a numbered subpattern (possibly recursive) */
723     OP_CALLOUT, /* 82 Call out to external function if provided */
724 nigel 77
725 ph10 178 OP_ALT, /* 83 Start of alternation */
726     OP_KET, /* 84 End of group that doesn't have an unbounded repeat */
727     OP_KETRMAX, /* 85 These two must remain together and in this */
728     OP_KETRMIN, /* 86 order. They are for groups the repeat for ever. */
729 nigel 77
730 nigel 93 /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
731 nigel 77
732 ph10 178 OP_ASSERT, /* 87 Positive lookahead */
733     OP_ASSERT_NOT, /* 88 Negative lookahead */
734     OP_ASSERTBACK, /* 89 Positive lookbehind */
735     OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */
736     OP_REVERSE, /* 91 Move pointer back - used in lookbehind assertions */
737 nigel 77
738 nigel 93 /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
739     as there's a test for >= ONCE for a subpattern that isn't an assertion. */
740 nigel 77
741 ph10 178 OP_ONCE, /* 92 Atomic group */
742 ph10 190 OP_BRA, /* 93 Start of non-capturing bracket */
743 ph10 178 OP_CBRA, /* 94 Start of capturing bracket */
744     OP_COND, /* 95 Conditional group */
745 nigel 77
746 nigel 93 /* These three must follow the previous three, in the same order. There's a
747     check for >= SBRA to distinguish the two sets. */
748 nigel 77
749 ph10 178 OP_SBRA, /* 96 Start of non-capturing bracket, check empty */
750     OP_SCBRA, /* 97 Start of capturing bracket, check empty */
751     OP_SCOND, /* 98 Conditional group, check empty */
752 nigel 77
753 ph10 178 OP_CREF, /* 99 Used to hold a capture number as condition */
754     OP_RREF, /* 100 Used to hold a recursion number as condition */
755     OP_DEF, /* 101 The DEFINE condition */
756 nigel 93
757 ph10 178 OP_BRAZERO, /* 102 These two must remain together and in this */
758     OP_BRAMINZERO /* 103 order. */
759 nigel 77 };
760    
761    
762     /* This macro defines textual names for all the opcodes. These are used only
763     for debugging. The macro is referenced only in pcre_printint.c. */
764    
765     #define OP_NAME_LIST \
766 ph10 168 "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d", \
767 nigel 77 "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte", \
768 ph10 178 "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v", \
769     "extuni", "\\Z", "\\z", \
770 nigel 77 "Opt", "^", "$", "char", "charnc", "not", \
771     "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
772 nigel 93 "*+","++", "?+", "{", \
773 nigel 77 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
774 nigel 93 "*+","++", "?+", "{", \
775 nigel 77 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
776 nigel 93 "*+","++", "?+", "{", \
777 nigel 77 "*", "*?", "+", "+?", "?", "??", "{", "{", \
778     "class", "nclass", "xclass", "Ref", "Recurse", "Callout", \
779     "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not", \
780 nigel 93 "AssertB", "AssertB not", "Reverse", \
781 ph10 192 "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond", \
782 nigel 93 "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"
783 nigel 77
784    
785     /* This macro defines the length of fixed length operations in the compiled
786     regex. The lengths are used when searching for specific things, and also in the
787     debugging printing of a compiled regex. We use a macro so that it can be
788     defined close to the definitions of the opcodes themselves.
789    
790     As things have been extended, some of these are no longer fixed lenths, but are
791     minima instead. For example, the length of a single-character repeat may vary
792     in UTF-8 mode. The code that uses this table must know about such things. */
793    
794     #define OP_LENGTHS \
795     1, /* End */ \
796 ph10 168 1, 1, 1, 1, 1, /* \A, \G, \K, \B, \b */ \
797     1, 1, 1, 1, 1, 1, /* \D, \d, \S, \s, \W, \w */ \
798 nigel 77 1, 1, /* Any, Anybyte */ \
799 ph10 178 3, 3, 1, /* NOTPROP, PROP, EXTUNI */ \
800     1, 1, 1, 1, 1, /* \R, \H, \h, \V, \v */ \
801 nigel 77 1, 1, 2, 1, 1, /* \Z, \z, Opt, ^, $ */ \
802     2, /* Char - the minimum length */ \
803     2, /* Charnc - the minimum length */ \
804     2, /* not */ \
805     /* Positive single-char repeats ** These are */ \
806     2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** minima in */ \
807     4, 4, 4, /* upto, minupto, exact ** UTF-8 mode */ \
808 nigel 93 2, 2, 2, 4, /* *+, ++, ?+, upto+ */ \
809 nigel 77 /* Negative single-char repeats - only for chars < 256 */ \
810     2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \
811     4, 4, 4, /* NOT upto, minupto, exact */ \
812 nigel 93 2, 2, 2, 4, /* Possessive *, +, ?, upto */ \
813 nigel 77 /* Positive type repeats */ \
814     2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \
815     4, 4, 4, /* Type upto, minupto, exact */ \
816 nigel 93 2, 2, 2, 4, /* Possessive *+, ++, ?+, upto+ */ \
817 nigel 77 /* Character class & ref repeats */ \
818     1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \
819     5, 5, /* CRRANGE, CRMINRANGE */ \
820     33, /* CLASS */ \
821     33, /* NCLASS */ \
822     0, /* XCLASS - variable length */ \
823     3, /* REF */ \
824     1+LINK_SIZE, /* RECURSE */ \
825     2+2*LINK_SIZE, /* CALLOUT */ \
826     1+LINK_SIZE, /* Alt */ \
827     1+LINK_SIZE, /* Ket */ \
828     1+LINK_SIZE, /* KetRmax */ \
829     1+LINK_SIZE, /* KetRmin */ \
830     1+LINK_SIZE, /* Assert */ \
831     1+LINK_SIZE, /* Assert not */ \
832     1+LINK_SIZE, /* Assert behind */ \
833     1+LINK_SIZE, /* Assert behind not */ \
834     1+LINK_SIZE, /* Reverse */ \
835 nigel 93 1+LINK_SIZE, /* ONCE */ \
836     1+LINK_SIZE, /* BRA */ \
837     3+LINK_SIZE, /* CBRA */ \
838 nigel 77 1+LINK_SIZE, /* COND */ \
839 nigel 93 1+LINK_SIZE, /* SBRA */ \
840     3+LINK_SIZE, /* SCBRA */ \
841     1+LINK_SIZE, /* SCOND */ \
842 nigel 77 3, /* CREF */ \
843 nigel 93 3, /* RREF */ \
844     1, /* DEF */ \
845 nigel 77 1, 1, /* BRAZERO, BRAMINZERO */ \
846    
847    
848 nigel 93 /* A magic value for OP_RREF to indicate the "any recursion" condition. */
849 nigel 77
850 nigel 93 #define RREF_ANY 0xffff
851 nigel 77
852     /* Error code numbers. They are given names so that they can more easily be
853     tracked. */
854    
855     enum { ERR0, ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9,
856     ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
857     ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
858     ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
859 nigel 91 ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
860 ph10 166 ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58 };
861 nigel 77
862     /* The real format of the start of the pcre block; the index of names and the
863     code vector run on as long as necessary after the end. We store an explicit
864     offset to the name table so that if a regex is compiled on one host, saved, and
865     then run on another where the size of pointers is different, all might still
866     be well. For the case of compiled-on-4 and run-on-8, we include an extra
867     pointer that is always NULL. For future-proofing, a few dummy fields were
868     originally included - even though you can never get this planning right - but
869     there is only one left now.
870    
871     NOTE NOTE NOTE:
872     Because people can now save and re-use compiled patterns, any additions to this
873     structure should be made at the end, and something earlier (e.g. a new
874     flag in the options or one of the dummy fields) should indicate that the new
875     fields are present. Currently PCRE always sets the dummy fields to zero.
876     NOTE NOTE NOTE:
877     */
878    
879     typedef struct real_pcre {
880     pcre_uint32 magic_number;
881     pcre_uint32 size; /* Total that was malloced */
882     pcre_uint32 options;
883     pcre_uint32 dummy1; /* For future use, maybe */
884    
885     pcre_uint16 top_bracket;
886     pcre_uint16 top_backref;
887     pcre_uint16 first_byte;
888     pcre_uint16 req_byte;
889     pcre_uint16 name_table_offset; /* Offset to name table that follows */
890     pcre_uint16 name_entry_size; /* Size of any name items */
891     pcre_uint16 name_count; /* Number of name items */
892     pcre_uint16 ref_count; /* Reference count */
893    
894     const unsigned char *tables; /* Pointer to tables or NULL for std */
895     const unsigned char *nullpad; /* NULL padding */
896     } real_pcre;
897    
898     /* The format of the block used to store data from pcre_study(). The same
899     remark (see NOTE above) about extending this structure applies. */
900    
901     typedef struct pcre_study_data {
902     pcre_uint32 size; /* Total that was malloced */
903     pcre_uint32 options;
904     uschar start_bits[32];
905     } pcre_study_data;
906    
907     /* Structure for passing "static" information around between the functions
908     doing the compiling, so that they are thread-safe. */
909    
910     typedef struct compile_data {
911     const uschar *lcc; /* Points to lower casing table */
912     const uschar *fcc; /* Points to case-flipping table */
913     const uschar *cbits; /* Points to character type table */
914     const uschar *ctypes; /* Points to table of type maps */
915 nigel 93 const uschar *start_workspace;/* The start of working space */
916 nigel 77 const uschar *start_code; /* The start of the compiled code */
917     const uschar *start_pattern; /* The start of the pattern */
918 nigel 93 const uschar *end_pattern; /* The end of the pattern */
919     uschar *hwm; /* High watermark of workspace */
920 nigel 77 uschar *name_table; /* The name/number table */
921     int names_found; /* Number of entries so far */
922     int name_entry_size; /* Size of each entry */
923 nigel 93 int bracount; /* Count of capturing parens */
924 nigel 77 int top_backref; /* Maximum back reference */
925     unsigned int backref_map; /* Bitmap of low back refs */
926 nigel 93 int external_options; /* External (initial) options */
927 nigel 77 int req_varyopt; /* "After variable item" flag for reqbyte */
928     BOOL nopartial; /* Set TRUE if partial won't work */
929 nigel 93 int nltype; /* Newline type */
930     int nllen; /* Newline string length */
931     uschar nl[4]; /* Newline string when fixed length */
932 nigel 77 } compile_data;
933    
934     /* Structure for maintaining a chain of pointers to the currently incomplete
935     branches, for testing for left recursion. */
936    
937     typedef struct branch_chain {
938     struct branch_chain *outer;
939     uschar *current;
940     } branch_chain;
941    
942     /* Structure for items in a linked list that represents an explicit recursive
943     call within the pattern. */
944    
945     typedef struct recursion_info {
946     struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
947     int group_num; /* Number of group that was called */
948     const uschar *after_call; /* "Return value": points after the call in the expr */
949 ph10 168 USPTR save_start; /* Old value of mstart */
950 nigel 77 int *offset_save; /* Pointer to start of saved offsets */
951     int saved_max; /* Number of saved offsets */
952     } recursion_info;
953    
954 nigel 93 /* Structure for building a chain of data for holding the values of the subject
955     pointer at the start of each subpattern, so as to detect when an empty string
956     has been matched by a subpattern - to break infinite loops. */
957    
958     typedef struct eptrblock {
959     struct eptrblock *epb_prev;
960     USPTR epb_saved_eptr;
961     } eptrblock;
962    
963    
964 nigel 77 /* Structure for passing "static" information around between the functions
965     doing traditional NFA matching, so that they are thread-safe. */
966    
967     typedef struct match_data {
968 nigel 87 unsigned long int match_call_count; /* As it says */
969     unsigned long int match_limit; /* As it says */
970     unsigned long int match_limit_recursion; /* As it says */
971 nigel 77 int *offset_vector; /* Offset vector */
972     int offset_end; /* One past the end */
973     int offset_max; /* The maximum usable for return data */
974 nigel 93 int nltype; /* Newline type */
975     int nllen; /* Newline string length */
976     uschar nl[4]; /* Newline string when fixed */
977 nigel 77 const uschar *lcc; /* Points to lower casing table */
978     const uschar *ctypes; /* Points to table of type maps */
979     BOOL offset_overflow; /* Set if too many extractions */
980     BOOL notbol; /* NOTBOL flag */
981     BOOL noteol; /* NOTEOL flag */
982     BOOL utf8; /* UTF8 flag */
983     BOOL endonly; /* Dollar not before final \n */
984     BOOL notempty; /* Empty string match not wanted */
985     BOOL partial; /* PARTIAL flag */
986     BOOL hitend; /* Hit the end of the subject at some point */
987     const uschar *start_code; /* For use when recursing */
988 nigel 87 USPTR start_subject; /* Start of the subject string */
989     USPTR end_subject; /* End of the subject string */
990 ph10 168 USPTR start_match_ptr; /* Start of matched string */
991 nigel 87 USPTR end_match_ptr; /* Subject position at end match */
992 nigel 77 int end_offset_top; /* Highwater mark at end of match */
993     int capture_last; /* Most recent capture number */
994     int start_offset; /* The start offset value */
995 nigel 93 eptrblock *eptrchain; /* Chain of eptrblocks for tail recursions */
996     int eptrn; /* Next free eptrblock */
997 nigel 77 recursion_info *recursive; /* Linked list of recursion data */
998     void *callout_data; /* To pass back to callouts */
999     } match_data;
1000    
1001     /* A similar structure is used for the same purpose by the DFA matching
1002     functions. */
1003    
1004     typedef struct dfa_match_data {
1005     const uschar *start_code; /* Start of the compiled pattern */
1006     const uschar *start_subject; /* Start of the subject string */
1007     const uschar *end_subject; /* End of subject string */
1008     const uschar *tables; /* Character tables */
1009     int moptions; /* Match options */
1010     int poptions; /* Pattern options */
1011 nigel 93 int nltype; /* Newline type */
1012     int nllen; /* Newline string length */
1013     uschar nl[4]; /* Newline string when fixed */
1014 nigel 77 void *callout_data; /* To pass back to callouts */
1015     } dfa_match_data;
1016    
1017     /* Bit definitions for entries in the pcre_ctypes table. */
1018    
1019     #define ctype_space 0x01
1020     #define ctype_letter 0x02
1021     #define ctype_digit 0x04
1022     #define ctype_xdigit 0x08
1023     #define ctype_word 0x10 /* alphameric or '_' */
1024     #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
1025    
1026     /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
1027     of bits for a class map. Some classes are built by combining these tables. */
1028    
1029     #define cbit_space 0 /* [:space:] or \s */
1030     #define cbit_xdigit 32 /* [:xdigit:] */
1031     #define cbit_digit 64 /* [:digit:] or \d */
1032     #define cbit_upper 96 /* [:upper:] */
1033     #define cbit_lower 128 /* [:lower:] */
1034     #define cbit_word 160 /* [:word:] or \w */
1035     #define cbit_graph 192 /* [:graph:] */
1036     #define cbit_print 224 /* [:print:] */
1037     #define cbit_punct 256 /* [:punct:] */
1038     #define cbit_cntrl 288 /* [:cntrl:] */
1039     #define cbit_length 320 /* Length of the cbits table */
1040    
1041     /* Offsets of the various tables from the base tables pointer, and
1042     total length. */
1043    
1044     #define lcc_offset 0
1045     #define fcc_offset 256
1046     #define cbits_offset 512
1047     #define ctypes_offset (cbits_offset + cbit_length)
1048     #define tables_length (ctypes_offset + 256)
1049    
1050 nigel 87 /* Layout of the UCP type table that translates property names into types and
1051     codes. */
1052 nigel 77
1053     typedef struct {
1054     const char *name;
1055 nigel 87 pcre_uint16 type;
1056     pcre_uint16 value;
1057 nigel 77 } ucp_type_table;
1058    
1059    
1060     /* Internal shared data tables. These are tables that are used by more than one
1061     of the exported public functions. They have to be "external" in the C sense,
1062     but are not part of the PCRE public API. The data for these tables is in the
1063     pcre_tables.c module. */
1064    
1065     extern const int _pcre_utf8_table1[];
1066     extern const int _pcre_utf8_table2[];
1067     extern const int _pcre_utf8_table3[];
1068     extern const uschar _pcre_utf8_table4[];
1069    
1070     extern const int _pcre_utf8_table1_size;
1071    
1072     extern const ucp_type_table _pcre_utt[];
1073     extern const int _pcre_utt_size;
1074    
1075     extern const uschar _pcre_default_tables[];
1076    
1077     extern const uschar _pcre_OP_lengths[];
1078    
1079    
1080     /* Internal shared functions. These are functions that are used by more than
1081     one of the exported public functions. They have to be "external" in the C
1082     sense, but are not part of the PCRE public API. */
1083    
1084 ph10 150 extern BOOL _pcre_is_newline(const uschar *, int, const uschar *,
1085 ph10 149 int *, BOOL);
1086 nigel 93 extern int _pcre_ord2utf8(int, uschar *);
1087     extern real_pcre *_pcre_try_flipped(const real_pcre *, real_pcre *,
1088     const pcre_study_data *, pcre_study_data *);
1089     extern int _pcre_ucp_findprop(const unsigned int, int *, int *);
1090     extern unsigned int _pcre_ucp_othercase(const unsigned int);
1091     extern int _pcre_valid_utf8(const uschar *, int);
1092 ph10 150 extern BOOL _pcre_was_newline(const uschar *, int, const uschar *,
1093 ph10 149 int *, BOOL);
1094 nigel 93 extern BOOL _pcre_xclass(int, const uschar *);
1095 nigel 77
1096 nigel 85 #endif
1097    
1098 nigel 77 /* End of pcre_internal.h */

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12