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

Contents of /code/trunk/pcre_internal.h

Parent Directory Parent Directory | Revision Log Revision Log


Revision 359 - (show annotations) (download)
Wed Jul 9 16:20:19 2008 UTC (6 years, 3 months ago) by ph10
File MIME type: text/plain
File size: 47903 byte(s)
Add PCRE_CALL_CONVENTION for MSVC users; add some missing PCRE_EXP_DEFNs.

1 /*************************************************
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 Copyright (c) 1997-2008 University of Cambridge
11
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 #ifndef PCRE_INTERNAL_H
46 #define PCRE_INTERNAL_H
47
48 /* Define DEBUG to get debugging output on stdout. */
49
50 #if 0
51 #define DEBUG
52 #endif
53
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 all, it had only been about 10 years then...
58
59 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 #ifdef DEBUG
64 #define DPRINTF(p) printf p
65 #else
66 #define DPRINTF(p) /* Nothing */
67 #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 /* 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 information there, using __declspec(dllexport) without "extern" we have a
86 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 which is all that is needed for applications (they just import the symbols). We
89 use:
90
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 internals, can #include pcre.h first to get an application's-eye view.
103
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 # ifndef PCRE_STATIC
112 # 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 #endif
134
135 /* When compiling with the MSVC compiler, it is sometimes necessary to include
136 a "calling convention" before exported function names. (This is secondhand
137 information; I know nothing about MSVC myself). For example, something like
138
139 void __cdecl function(....)
140
141 might be needed. In order so make this easy, all the exported functions have
142 PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
143 set, we ensure here that it has no effect. */
144
145 #ifndef PCRE_CALL_CONVENTION
146 #define PCRE_CALL_CONVENTION
147 #endif
148
149 /* We need to have types that specify unsigned 16-bit and 32-bit integers. We
150 cannot determine these outside the compilation (e.g. by running a program as
151 part of "configure") because PCRE is often cross-compiled for use on other
152 systems. Instead we make use of the maximum sizes that are available at
153 preprocessor time in standard C environments. */
154
155 #if USHRT_MAX == 65535
156 typedef unsigned short pcre_uint16;
157 typedef short pcre_int16;
158 #elif UINT_MAX == 65535
159 typedef unsigned int pcre_uint16;
160 typedef int pcre_int16;
161 #else
162 #error Cannot determine a type for 16-bit unsigned integers
163 #endif
164
165 #if UINT_MAX == 4294967295
166 typedef unsigned int pcre_uint32;
167 typedef int pcre_int32;
168 #elif ULONG_MAX == 4294967295
169 typedef unsigned long int pcre_uint32;
170 typedef long int pcre_int32;
171 #else
172 #error Cannot determine a type for 32-bit unsigned integers
173 #endif
174
175 /* All character handling must be done as unsigned characters. Otherwise there
176 are problems with top-bit-set characters and functions such as isspace().
177 However, we leave the interface to the outside world as char *, because that
178 should make things easier for callers. We define a short type for unsigned char
179 to save lots of typing. I tried "uchar", but it causes problems on Digital
180 Unix, where it is defined in sys/types, so use "uschar" instead. */
181
182 typedef unsigned char uschar;
183
184 /* This is an unsigned int value that no character can ever have. UTF-8
185 characters only go up to 0x7fffffff (though Unicode doesn't go beyond
186 0x0010ffff). */
187
188 #define NOTACHAR 0xffffffff
189
190 /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
191 "any" and "anycrlf" at present). The following macros are used to package up
192 testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
193 modules to indicate in which datablock the parameters exist, and what the
194 start/end of string field names are. */
195
196 #define NLTYPE_FIXED 0 /* Newline is a fixed length string */
197 #define NLTYPE_ANY 1 /* Newline is any Unicode line ending */
198 #define NLTYPE_ANYCRLF 2 /* Newline is CR, LF, or CRLF */
199
200 /* This macro checks for a newline at the given position */
201
202 #define IS_NEWLINE(p) \
203 ((NLBLOCK->nltype != NLTYPE_FIXED)? \
204 ((p) < NLBLOCK->PSEND && \
205 _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
206 utf8)) \
207 : \
208 ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
209 (p)[0] == NLBLOCK->nl[0] && \
210 (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
211 ) \
212 )
213
214 /* This macro checks for a newline immediately preceding the given position */
215
216 #define WAS_NEWLINE(p) \
217 ((NLBLOCK->nltype != NLTYPE_FIXED)? \
218 ((p) > NLBLOCK->PSSTART && \
219 _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
220 &(NLBLOCK->nllen), utf8)) \
221 : \
222 ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
223 (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
224 (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
225 ) \
226 )
227
228 /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
229 with a custom type. This makes it possible, for example, to allow pcre_exec()
230 to process subject strings that are discontinuous by using a smart pointer
231 class. It must always be possible to inspect all of the subject string in
232 pcre_exec() because of the way it backtracks. Two macros are required in the
233 normal case, for sign-unspecified and unsigned char pointers. The former is
234 used for the external interface and appears in pcre.h, which is why its name
235 must begin with PCRE_. */
236
237 #ifdef CUSTOM_SUBJECT_PTR
238 #define PCRE_SPTR CUSTOM_SUBJECT_PTR
239 #define USPTR CUSTOM_SUBJECT_PTR
240 #else
241 #define PCRE_SPTR const char *
242 #define USPTR const unsigned char *
243 #endif
244
245
246
247 /* Include the public PCRE header and the definitions of UCP character property
248 values. */
249
250 #include "pcre.h"
251 #include "ucp.h"
252
253 /* When compiling for use with the Virtual Pascal compiler, these functions
254 need to have their names changed. PCRE must be compiled with the -DVPCOMPAT
255 option on the command line. */
256
257 #ifdef VPCOMPAT
258 #define strlen(s) _strlen(s)
259 #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
260 #define memcmp(s,c,n) _memcmp(s,c,n)
261 #define memcpy(d,s,n) _memcpy(d,s,n)
262 #define memmove(d,s,n) _memmove(d,s,n)
263 #define memset(s,c,n) _memset(s,c,n)
264 #else /* VPCOMPAT */
265
266 /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
267 define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
268 is set. Otherwise, include an emulating function for those systems that have
269 neither (there some non-Unix environments where this is the case). */
270
271 #ifndef HAVE_MEMMOVE
272 #undef memmove /* some systems may have a macro */
273 #ifdef HAVE_BCOPY
274 #define memmove(a, b, c) bcopy(b, a, c)
275 #else /* HAVE_BCOPY */
276 static void *
277 pcre_memmove(void *d, const void *s, size_t n)
278 {
279 size_t i;
280 unsigned char *dest = (unsigned char *)d;
281 const unsigned char *src = (const unsigned char *)s;
282 if (dest > src)
283 {
284 dest += n;
285 src += n;
286 for (i = 0; i < n; ++i) *(--dest) = *(--src);
287 return (void *)dest;
288 }
289 else
290 {
291 for (i = 0; i < n; ++i) *dest++ = *src++;
292 return (void *)(dest - n);
293 }
294 }
295 #define memmove(a, b, c) pcre_memmove(a, b, c)
296 #endif /* not HAVE_BCOPY */
297 #endif /* not HAVE_MEMMOVE */
298 #endif /* not VPCOMPAT */
299
300
301 /* PCRE keeps offsets in its compiled code as 2-byte quantities (always stored
302 in big-endian order) by default. These are used, for example, to link from the
303 start of a subpattern to its alternatives and its end. The use of 2 bytes per
304 offset limits the size of the compiled regex to around 64K, which is big enough
305 for almost everybody. However, I received a request for an even bigger limit.
306 For this reason, and also to make the code easier to maintain, the storing and
307 loading of offsets from the byte string is now handled by the macros that are
308 defined here.
309
310 The macros are controlled by the value of LINK_SIZE. This defaults to 2 in
311 the config.h file, but can be overridden by using -D on the command line. This
312 is automated on Unix systems via the "configure" command. */
313
314 #if LINK_SIZE == 2
315
316 #define PUT(a,n,d) \
317 (a[n] = (d) >> 8), \
318 (a[(n)+1] = (d) & 255)
319
320 #define GET(a,n) \
321 (((a)[n] << 8) | (a)[(n)+1])
322
323 #define MAX_PATTERN_SIZE (1 << 16)
324
325
326 #elif LINK_SIZE == 3
327
328 #define PUT(a,n,d) \
329 (a[n] = (d) >> 16), \
330 (a[(n)+1] = (d) >> 8), \
331 (a[(n)+2] = (d) & 255)
332
333 #define GET(a,n) \
334 (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2])
335
336 #define MAX_PATTERN_SIZE (1 << 24)
337
338
339 #elif LINK_SIZE == 4
340
341 #define PUT(a,n,d) \
342 (a[n] = (d) >> 24), \
343 (a[(n)+1] = (d) >> 16), \
344 (a[(n)+2] = (d) >> 8), \
345 (a[(n)+3] = (d) & 255)
346
347 #define GET(a,n) \
348 (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
349
350 #define MAX_PATTERN_SIZE (1 << 30) /* Keep it positive */
351
352
353 #else
354 #error LINK_SIZE must be either 2, 3, or 4
355 #endif
356
357
358 /* Convenience macro defined in terms of the others */
359
360 #define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE
361
362
363 /* PCRE uses some other 2-byte quantities that do not change when the size of
364 offsets changes. There are used for repeat counts and for other things such as
365 capturing parenthesis numbers in back references. */
366
367 #define PUT2(a,n,d) \
368 a[n] = (d) >> 8; \
369 a[(n)+1] = (d) & 255
370
371 #define GET2(a,n) \
372 (((a)[n] << 8) | (a)[(n)+1])
373
374 #define PUT2INC(a,n,d) PUT2(a,n,d), a += 2
375
376
377 /* When UTF-8 encoding is being used, a character is no longer just a single
378 byte. The macros for character handling generate simple sequences when used in
379 byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
380 never be called in byte mode. To make sure it can never even appear when UTF-8
381 support is omitted, we don't even define it. */
382
383 #ifndef SUPPORT_UTF8
384 #define NEXTCHAR(p) p++;
385 #define GETCHAR(c, eptr) c = *eptr;
386 #define GETCHARTEST(c, eptr) c = *eptr;
387 #define GETCHARINC(c, eptr) c = *eptr++;
388 #define GETCHARINCTEST(c, eptr) c = *eptr++;
389 #define GETCHARLEN(c, eptr, len) c = *eptr;
390 /* #define BACKCHAR(eptr) */
391
392 #else /* SUPPORT_UTF8 */
393
394 /* Advance a character pointer one byte in non-UTF-8 mode and by one character
395 in UTF-8 mode. */
396
397 #define NEXTCHAR(p) \
398 p++; \
399 if (utf8) { while((*p & 0xc0) == 0x80) p++; }
400
401 /* Get the next UTF-8 character, not advancing the pointer. This is called when
402 we know we are in UTF-8 mode. */
403
404 #define GETCHAR(c, eptr) \
405 c = *eptr; \
406 if (c >= 0xc0) \
407 { \
408 int gcii; \
409 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
410 int gcss = 6*gcaa; \
411 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
412 for (gcii = 1; gcii <= gcaa; gcii++) \
413 { \
414 gcss -= 6; \
415 c |= (eptr[gcii] & 0x3f) << gcss; \
416 } \
417 }
418
419 /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
420 pointer. */
421
422 #define GETCHARTEST(c, eptr) \
423 c = *eptr; \
424 if (utf8 && c >= 0xc0) \
425 { \
426 int gcii; \
427 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
428 int gcss = 6*gcaa; \
429 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
430 for (gcii = 1; gcii <= gcaa; gcii++) \
431 { \
432 gcss -= 6; \
433 c |= (eptr[gcii] & 0x3f) << gcss; \
434 } \
435 }
436
437 /* Get the next UTF-8 character, advancing the pointer. This is called when we
438 know we are in UTF-8 mode. */
439
440 #define GETCHARINC(c, eptr) \
441 c = *eptr++; \
442 if (c >= 0xc0) \
443 { \
444 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
445 int gcss = 6*gcaa; \
446 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
447 while (gcaa-- > 0) \
448 { \
449 gcss -= 6; \
450 c |= (*eptr++ & 0x3f) << gcss; \
451 } \
452 }
453
454 /* Get the next character, testing for UTF-8 mode, and advancing the pointer */
455
456 #define GETCHARINCTEST(c, eptr) \
457 c = *eptr++; \
458 if (utf8 && c >= 0xc0) \
459 { \
460 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
461 int gcss = 6*gcaa; \
462 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
463 while (gcaa-- > 0) \
464 { \
465 gcss -= 6; \
466 c |= (*eptr++ & 0x3f) << gcss; \
467 } \
468 }
469
470 /* Get the next UTF-8 character, not advancing the pointer, incrementing length
471 if there are extra bytes. This is called when we know we are in UTF-8 mode. */
472
473 #define GETCHARLEN(c, eptr, len) \
474 c = *eptr; \
475 if (c >= 0xc0) \
476 { \
477 int gcii; \
478 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
479 int gcss = 6*gcaa; \
480 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
481 for (gcii = 1; gcii <= gcaa; gcii++) \
482 { \
483 gcss -= 6; \
484 c |= (eptr[gcii] & 0x3f) << gcss; \
485 } \
486 len += gcaa; \
487 }
488
489 /* If the pointer is not at the start of a character, move it back until
490 it is. This is called only in UTF-8 mode - we don't put a test within the macro
491 because almost all calls are already within a block of UTF-8 only code. */
492
493 #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
494
495 #endif
496
497
498 /* In case there is no definition of offsetof() provided - though any proper
499 Standard C system should have one. */
500
501 #ifndef offsetof
502 #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
503 #endif
504
505
506 /* These are the public options that can change during matching. */
507
508 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
509
510 /* Private flags containing information about the compiled regex. They used to
511 live at the top end of the options word, but that got almost full, so now they
512 are in a 16-bit flags word. */
513
514 #define PCRE_NOPARTIAL 0x0001 /* can't use partial with this regex */
515 #define PCRE_FIRSTSET 0x0002 /* first_byte is set */
516 #define PCRE_REQCHSET 0x0004 /* req_byte is set */
517 #define PCRE_STARTLINE 0x0008 /* start after \n for multiline */
518 #define PCRE_JCHANGED 0x0010 /* j option used in regex */
519 #define PCRE_HASCRORLF 0x0020 /* explicit \r or \n in pattern */
520
521 /* Options for the "extra" block produced by pcre_study(). */
522
523 #define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */
524
525 /* Masks for identifying the public options that are permitted at compile
526 time, run time, or study time, respectively. */
527
528 #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
529 PCRE_NEWLINE_ANYCRLF)
530
531 #define PUBLIC_OPTIONS \
532 (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
533 PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
534 PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
535 PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
536 PCRE_JAVASCRIPT_COMPAT)
537
538 #define PUBLIC_EXEC_OPTIONS \
539 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
540 PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
541
542 #define PUBLIC_DFA_EXEC_OPTIONS \
543 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
544 PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
545 PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)
546
547 #define PUBLIC_STUDY_OPTIONS 0 /* None defined */
548
549 /* Magic number to provide a small check against being handed junk. Also used
550 to detect whether a pattern was compiled on a host of different endianness. */
551
552 #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
553
554 /* Negative values for the firstchar and reqchar variables */
555
556 #define REQ_UNSET (-2)
557 #define REQ_NONE (-1)
558
559 /* The maximum remaining length of subject we are prepared to search for a
560 req_byte match. */
561
562 #define REQ_BYTE_MAX 1000
563
564 /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a
565 variable-length repeat, or a anything other than literal characters. */
566
567 #define REQ_CASELESS 0x0100 /* indicates caselessness */
568 #define REQ_VARY 0x0200 /* reqbyte followed non-literal item */
569
570 /* Miscellaneous definitions */
571
572 typedef int BOOL;
573
574 #define FALSE 0
575 #define TRUE 1
576
577 /* Escape items that are just an encoding of a particular data value. */
578
579 #ifndef ESC_e
580 #define ESC_e 27
581 #endif
582
583 #ifndef ESC_f
584 #define ESC_f '\f'
585 #endif
586
587 #ifndef ESC_n
588 #define ESC_n '\n'
589 #endif
590
591 #ifndef ESC_r
592 #define ESC_r '\r'
593 #endif
594
595 /* We can't officially use ESC_t because it is a POSIX reserved identifier
596 (presumably because of all the others like size_t). */
597
598 #ifndef ESC_tee
599 #define ESC_tee '\t'
600 #endif
601
602 /* Codes for different types of Unicode property */
603
604 #define PT_ANY 0 /* Any property - matches all chars */
605 #define PT_LAMP 1 /* L& - the union of Lu, Ll, Lt */
606 #define PT_GC 2 /* General characteristic (e.g. L) */
607 #define PT_PC 3 /* Particular characteristic (e.g. Lu) */
608 #define PT_SC 4 /* Script (e.g. Han) */
609
610 /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
611 contain UTF-8 characters with values greater than 255. */
612
613 #define XCL_NOT 0x01 /* Flag: this is a negative class */
614 #define XCL_MAP 0x02 /* Flag: a 32-byte map is present */
615
616 #define XCL_END 0 /* Marks end of individual items */
617 #define XCL_SINGLE 1 /* Single item (one multibyte char) follows */
618 #define XCL_RANGE 2 /* A range (two multibyte chars) follows */
619 #define XCL_PROP 3 /* Unicode property (2-byte property code follows) */
620 #define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */
621
622 /* These are escaped items that aren't just an encoding of a particular data
623 value such as \n. They must have non-zero values, as check_escape() returns
624 their negation. Also, they must appear in the same order as in the opcode
625 definitions below, up to ESC_z. There's a dummy for OP_ANY because it
626 corresponds to "." rather than an escape sequence, and another for OP_ALLANY
627 (which is used for [^] in JavaScript compatibility mode).
628
629 The final escape must be ESC_REF as subsequent values are used for
630 backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
631 greater than ESC_b and less than ESC_Z to detect the types that may be
632 repeated. These are the types that consume characters. If any new escapes are
633 put in between that don't consume a character, that code will have to change.
634 */
635
636 enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
637 ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
638 ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
639 ESC_REF };
640
641
642 /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
643 OP_EOD must correspond in order to the list of escapes immediately above.
644
645 *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
646 that follow must also be updated to match. There is also a table called
647 "coptable" in pcre_dfa_exec.c that must be updated. */
648
649 enum {
650 OP_END, /* 0 End of pattern */
651
652 /* Values corresponding to backslashed metacharacters */
653
654 OP_SOD, /* 1 Start of data: \A */
655 OP_SOM, /* 2 Start of match (subject + offset): \G */
656 OP_SET_SOM, /* 3 Set start of match (\K) */
657 OP_NOT_WORD_BOUNDARY, /* 4 \B */
658 OP_WORD_BOUNDARY, /* 5 \b */
659 OP_NOT_DIGIT, /* 6 \D */
660 OP_DIGIT, /* 7 \d */
661 OP_NOT_WHITESPACE, /* 8 \S */
662 OP_WHITESPACE, /* 9 \s */
663 OP_NOT_WORDCHAR, /* 10 \W */
664 OP_WORDCHAR, /* 11 \w */
665 OP_ANY, /* 12 Match any character (subject to DOTALL) */
666 OP_ALLANY, /* 13 Match any character (not subject to DOTALL) */
667 OP_ANYBYTE, /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
668 OP_NOTPROP, /* 15 \P (not Unicode property) */
669 OP_PROP, /* 16 \p (Unicode property) */
670 OP_ANYNL, /* 17 \R (any newline sequence) */
671 OP_NOT_HSPACE, /* 18 \H (not horizontal whitespace) */
672 OP_HSPACE, /* 19 \h (horizontal whitespace) */
673 OP_NOT_VSPACE, /* 20 \V (not vertical whitespace) */
674 OP_VSPACE, /* 21 \v (vertical whitespace) */
675 OP_EXTUNI, /* 22 \X (extended Unicode sequence */
676 OP_EODN, /* 23 End of data or \n at end of data: \Z. */
677 OP_EOD, /* 24 End of data: \z */
678
679 OP_OPT, /* 25 Set runtime options */
680 OP_CIRC, /* 26 Start of line - varies with multiline switch */
681 OP_DOLL, /* 27 End of line - varies with multiline switch */
682 OP_CHAR, /* 28 Match one character, casefully */
683 OP_CHARNC, /* 29 Match one character, caselessly */
684 OP_NOT, /* 30 Match one character, not the following one */
685
686 OP_STAR, /* 31 The maximizing and minimizing versions of */
687 OP_MINSTAR, /* 32 these six opcodes must come in pairs, with */
688 OP_PLUS, /* 33 the minimizing one second. */
689 OP_MINPLUS, /* 34 This first set applies to single characters.*/
690 OP_QUERY, /* 35 */
691 OP_MINQUERY, /* 36 */
692
693 OP_UPTO, /* 37 From 0 to n matches */
694 OP_MINUPTO, /* 38 */
695 OP_EXACT, /* 39 Exactly n matches */
696
697 OP_POSSTAR, /* 40 Possessified star */
698 OP_POSPLUS, /* 41 Possessified plus */
699 OP_POSQUERY, /* 42 Posesssified query */
700 OP_POSUPTO, /* 43 Possessified upto */
701
702 OP_NOTSTAR, /* 44 The maximizing and minimizing versions of */
703 OP_NOTMINSTAR, /* 45 these six opcodes must come in pairs, with */
704 OP_NOTPLUS, /* 46 the minimizing one second. They must be in */
705 OP_NOTMINPLUS, /* 47 exactly the same order as those above. */
706 OP_NOTQUERY, /* 48 This set applies to "not" single characters. */
707 OP_NOTMINQUERY, /* 49 */
708
709 OP_NOTUPTO, /* 50 From 0 to n matches */
710 OP_NOTMINUPTO, /* 51 */
711 OP_NOTEXACT, /* 52 Exactly n matches */
712
713 OP_NOTPOSSTAR, /* 53 Possessified versions */
714 OP_NOTPOSPLUS, /* 54 */
715 OP_NOTPOSQUERY, /* 55 */
716 OP_NOTPOSUPTO, /* 56 */
717
718 OP_TYPESTAR, /* 57 The maximizing and minimizing versions of */
719 OP_TYPEMINSTAR, /* 58 these six opcodes must come in pairs, with */
720 OP_TYPEPLUS, /* 59 the minimizing one second. These codes must */
721 OP_TYPEMINPLUS, /* 60 be in exactly the same order as those above. */
722 OP_TYPEQUERY, /* 61 This set applies to character types such as \d */
723 OP_TYPEMINQUERY, /* 62 */
724
725 OP_TYPEUPTO, /* 63 From 0 to n matches */
726 OP_TYPEMINUPTO, /* 64 */
727 OP_TYPEEXACT, /* 65 Exactly n matches */
728
729 OP_TYPEPOSSTAR, /* 66 Possessified versions */
730 OP_TYPEPOSPLUS, /* 67 */
731 OP_TYPEPOSQUERY, /* 68 */
732 OP_TYPEPOSUPTO, /* 69 */
733
734 OP_CRSTAR, /* 70 The maximizing and minimizing versions of */
735 OP_CRMINSTAR, /* 71 all these opcodes must come in pairs, with */
736 OP_CRPLUS, /* 72 the minimizing one second. These codes must */
737 OP_CRMINPLUS, /* 73 be in exactly the same order as those above. */
738 OP_CRQUERY, /* 74 These are for character classes and back refs */
739 OP_CRMINQUERY, /* 75 */
740 OP_CRRANGE, /* 76 These are different to the three sets above. */
741 OP_CRMINRANGE, /* 77 */
742
743 OP_CLASS, /* 78 Match a character class, chars < 256 only */
744 OP_NCLASS, /* 79 Same, but the bitmap was created from a negative
745 class - the difference is relevant only when a UTF-8
746 character > 255 is encountered. */
747
748 OP_XCLASS, /* 80 Extended class for handling UTF-8 chars within the
749 class. This does both positive and negative. */
750
751 OP_REF, /* 81 Match a back reference */
752 OP_RECURSE, /* 82 Match a numbered subpattern (possibly recursive) */
753 OP_CALLOUT, /* 83 Call out to external function if provided */
754
755 OP_ALT, /* 84 Start of alternation */
756 OP_KET, /* 85 End of group that doesn't have an unbounded repeat */
757 OP_KETRMAX, /* 86 These two must remain together and in this */
758 OP_KETRMIN, /* 87 order. They are for groups the repeat for ever. */
759
760 /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
761
762 OP_ASSERT, /* 88 Positive lookahead */
763 OP_ASSERT_NOT, /* 89 Negative lookahead */
764 OP_ASSERTBACK, /* 90 Positive lookbehind */
765 OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
766 OP_REVERSE, /* 92 Move pointer back - used in lookbehind assertions */
767
768 /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
769 as there's a test for >= ONCE for a subpattern that isn't an assertion. */
770
771 OP_ONCE, /* 93 Atomic group */
772 OP_BRA, /* 94 Start of non-capturing bracket */
773 OP_CBRA, /* 95 Start of capturing bracket */
774 OP_COND, /* 96 Conditional group */
775
776 /* These three must follow the previous three, in the same order. There's a
777 check for >= SBRA to distinguish the two sets. */
778
779 OP_SBRA, /* 97 Start of non-capturing bracket, check empty */
780 OP_SCBRA, /* 98 Start of capturing bracket, check empty */
781 OP_SCOND, /* 99 Conditional group, check empty */
782
783 OP_CREF, /* 100 Used to hold a capture number as condition */
784 OP_RREF, /* 101 Used to hold a recursion number as condition */
785 OP_DEF, /* 102 The DEFINE condition */
786
787 OP_BRAZERO, /* 103 These two must remain together and in this */
788 OP_BRAMINZERO, /* 104 order. */
789
790 /* These are backtracking control verbs */
791
792 OP_PRUNE, /* 105 */
793 OP_SKIP, /* 106 */
794 OP_THEN, /* 107 */
795 OP_COMMIT, /* 108 */
796
797 /* These are forced failure and success verbs */
798
799 OP_FAIL, /* 109 */
800 OP_ACCEPT, /* 110 */
801
802 /* This is used to skip a subpattern with a {0} quantifier */
803
804 OP_SKIPZERO /* 111 */
805 };
806
807
808 /* This macro defines textual names for all the opcodes. These are used only
809 for debugging. The macro is referenced only in pcre_printint.c. */
810
811 #define OP_NAME_LIST \
812 "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d", \
813 "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte", \
814 "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v", \
815 "extuni", "\\Z", "\\z", \
816 "Opt", "^", "$", "char", "charnc", "not", \
817 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
818 "*+","++", "?+", "{", \
819 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
820 "*+","++", "?+", "{", \
821 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
822 "*+","++", "?+", "{", \
823 "*", "*?", "+", "+?", "?", "??", "{", "{", \
824 "class", "nclass", "xclass", "Ref", "Recurse", "Callout", \
825 "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not", \
826 "AssertB", "AssertB not", "Reverse", \
827 "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond", \
828 "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero", \
829 "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT", \
830 "Skip zero"
831
832
833 /* This macro defines the length of fixed length operations in the compiled
834 regex. The lengths are used when searching for specific things, and also in the
835 debugging printing of a compiled regex. We use a macro so that it can be
836 defined close to the definitions of the opcodes themselves.
837
838 As things have been extended, some of these are no longer fixed lenths, but are
839 minima instead. For example, the length of a single-character repeat may vary
840 in UTF-8 mode. The code that uses this table must know about such things. */
841
842 #define OP_LENGTHS \
843 1, /* End */ \
844 1, 1, 1, 1, 1, /* \A, \G, \K, \B, \b */ \
845 1, 1, 1, 1, 1, 1, /* \D, \d, \S, \s, \W, \w */ \
846 1, 1, 1, /* Any, AllAny, Anybyte */ \
847 3, 3, 1, /* NOTPROP, PROP, EXTUNI */ \
848 1, 1, 1, 1, 1, /* \R, \H, \h, \V, \v */ \
849 1, 1, 2, 1, 1, /* \Z, \z, Opt, ^, $ */ \
850 2, /* Char - the minimum length */ \
851 2, /* Charnc - the minimum length */ \
852 2, /* not */ \
853 /* Positive single-char repeats ** These are */ \
854 2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** minima in */ \
855 4, 4, 4, /* upto, minupto, exact ** UTF-8 mode */ \
856 2, 2, 2, 4, /* *+, ++, ?+, upto+ */ \
857 /* Negative single-char repeats - only for chars < 256 */ \
858 2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \
859 4, 4, 4, /* NOT upto, minupto, exact */ \
860 2, 2, 2, 4, /* Possessive *, +, ?, upto */ \
861 /* Positive type repeats */ \
862 2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \
863 4, 4, 4, /* Type upto, minupto, exact */ \
864 2, 2, 2, 4, /* Possessive *+, ++, ?+, upto+ */ \
865 /* Character class & ref repeats */ \
866 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \
867 5, 5, /* CRRANGE, CRMINRANGE */ \
868 33, /* CLASS */ \
869 33, /* NCLASS */ \
870 0, /* XCLASS - variable length */ \
871 3, /* REF */ \
872 1+LINK_SIZE, /* RECURSE */ \
873 2+2*LINK_SIZE, /* CALLOUT */ \
874 1+LINK_SIZE, /* Alt */ \
875 1+LINK_SIZE, /* Ket */ \
876 1+LINK_SIZE, /* KetRmax */ \
877 1+LINK_SIZE, /* KetRmin */ \
878 1+LINK_SIZE, /* Assert */ \
879 1+LINK_SIZE, /* Assert not */ \
880 1+LINK_SIZE, /* Assert behind */ \
881 1+LINK_SIZE, /* Assert behind not */ \
882 1+LINK_SIZE, /* Reverse */ \
883 1+LINK_SIZE, /* ONCE */ \
884 1+LINK_SIZE, /* BRA */ \
885 3+LINK_SIZE, /* CBRA */ \
886 1+LINK_SIZE, /* COND */ \
887 1+LINK_SIZE, /* SBRA */ \
888 3+LINK_SIZE, /* SCBRA */ \
889 1+LINK_SIZE, /* SCOND */ \
890 3, /* CREF */ \
891 3, /* RREF */ \
892 1, /* DEF */ \
893 1, 1, /* BRAZERO, BRAMINZERO */ \
894 1, 1, 1, 1, /* PRUNE, SKIP, THEN, COMMIT, */ \
895 1, 1, 1 /* FAIL, ACCEPT, SKIPZERO */
896
897
898 /* A magic value for OP_RREF to indicate the "any recursion" condition. */
899
900 #define RREF_ANY 0xffff
901
902 /* Error code numbers. They are given names so that they can more easily be
903 tracked. */
904
905 enum { ERR0, ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9,
906 ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
907 ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
908 ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
909 ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
910 ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
911 ERR60, ERR61, ERR62, ERR63, ERR64 };
912
913 /* The real format of the start of the pcre block; the index of names and the
914 code vector run on as long as necessary after the end. We store an explicit
915 offset to the name table so that if a regex is compiled on one host, saved, and
916 then run on another where the size of pointers is different, all might still
917 be well. For the case of compiled-on-4 and run-on-8, we include an extra
918 pointer that is always NULL. For future-proofing, a few dummy fields were
919 originally included - even though you can never get this planning right - but
920 there is only one left now.
921
922 NOTE NOTE NOTE:
923 Because people can now save and re-use compiled patterns, any additions to this
924 structure should be made at the end, and something earlier (e.g. a new
925 flag in the options or one of the dummy fields) should indicate that the new
926 fields are present. Currently PCRE always sets the dummy fields to zero.
927 NOTE NOTE NOTE:
928 */
929
930 typedef struct real_pcre {
931 pcre_uint32 magic_number;
932 pcre_uint32 size; /* Total that was malloced */
933 pcre_uint32 options; /* Public options */
934 pcre_uint16 flags; /* Private flags */
935 pcre_uint16 dummy1; /* For future use */
936 pcre_uint16 top_bracket;
937 pcre_uint16 top_backref;
938 pcre_uint16 first_byte;
939 pcre_uint16 req_byte;
940 pcre_uint16 name_table_offset; /* Offset to name table that follows */
941 pcre_uint16 name_entry_size; /* Size of any name items */
942 pcre_uint16 name_count; /* Number of name items */
943 pcre_uint16 ref_count; /* Reference count */
944
945 const unsigned char *tables; /* Pointer to tables or NULL for std */
946 const unsigned char *nullpad; /* NULL padding */
947 } real_pcre;
948
949 /* The format of the block used to store data from pcre_study(). The same
950 remark (see NOTE above) about extending this structure applies. */
951
952 typedef struct pcre_study_data {
953 pcre_uint32 size; /* Total that was malloced */
954 pcre_uint32 options;
955 uschar start_bits[32];
956 } pcre_study_data;
957
958 /* Structure for passing "static" information around between the functions
959 doing the compiling, so that they are thread-safe. */
960
961 typedef struct compile_data {
962 const uschar *lcc; /* Points to lower casing table */
963 const uschar *fcc; /* Points to case-flipping table */
964 const uschar *cbits; /* Points to character type table */
965 const uschar *ctypes; /* Points to table of type maps */
966 const uschar *start_workspace;/* The start of working space */
967 const uschar *start_code; /* The start of the compiled code */
968 const uschar *start_pattern; /* The start of the pattern */
969 const uschar *end_pattern; /* The end of the pattern */
970 uschar *hwm; /* High watermark of workspace */
971 uschar *name_table; /* The name/number table */
972 int names_found; /* Number of entries so far */
973 int name_entry_size; /* Size of each entry */
974 int bracount; /* Count of capturing parens as we compile */
975 int final_bracount; /* Saved value after first pass */
976 int top_backref; /* Maximum back reference */
977 unsigned int backref_map; /* Bitmap of low back refs */
978 int external_options; /* External (initial) options */
979 int external_flags; /* External flag bits to be set */
980 int req_varyopt; /* "After variable item" flag for reqbyte */
981 BOOL had_accept; /* (*ACCEPT) encountered */
982 int nltype; /* Newline type */
983 int nllen; /* Newline string length */
984 uschar nl[4]; /* Newline string when fixed length */
985 } compile_data;
986
987 /* Structure for maintaining a chain of pointers to the currently incomplete
988 branches, for testing for left recursion. */
989
990 typedef struct branch_chain {
991 struct branch_chain *outer;
992 uschar *current;
993 } branch_chain;
994
995 /* Structure for items in a linked list that represents an explicit recursive
996 call within the pattern. */
997
998 typedef struct recursion_info {
999 struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1000 int group_num; /* Number of group that was called */
1001 const uschar *after_call; /* "Return value": points after the call in the expr */
1002 USPTR save_start; /* Old value of mstart */
1003 int *offset_save; /* Pointer to start of saved offsets */
1004 int saved_max; /* Number of saved offsets */
1005 } recursion_info;
1006
1007 /* Structure for building a chain of data for holding the values of the subject
1008 pointer at the start of each subpattern, so as to detect when an empty string
1009 has been matched by a subpattern - to break infinite loops. */
1010
1011 typedef struct eptrblock {
1012 struct eptrblock *epb_prev;
1013 USPTR epb_saved_eptr;
1014 } eptrblock;
1015
1016
1017 /* Structure for passing "static" information around between the functions
1018 doing traditional NFA matching, so that they are thread-safe. */
1019
1020 typedef struct match_data {
1021 unsigned long int match_call_count; /* As it says */
1022 unsigned long int match_limit; /* As it says */
1023 unsigned long int match_limit_recursion; /* As it says */
1024 int *offset_vector; /* Offset vector */
1025 int offset_end; /* One past the end */
1026 int offset_max; /* The maximum usable for return data */
1027 int nltype; /* Newline type */
1028 int nllen; /* Newline string length */
1029 uschar nl[4]; /* Newline string when fixed */
1030 const uschar *lcc; /* Points to lower casing table */
1031 const uschar *ctypes; /* Points to table of type maps */
1032 BOOL offset_overflow; /* Set if too many extractions */
1033 BOOL notbol; /* NOTBOL flag */
1034 BOOL noteol; /* NOTEOL flag */
1035 BOOL utf8; /* UTF8 flag */
1036 BOOL jscript_compat; /* JAVASCRIPT_COMPAT flag */
1037 BOOL endonly; /* Dollar not before final \n */
1038 BOOL notempty; /* Empty string match not wanted */
1039 BOOL partial; /* PARTIAL flag */
1040 BOOL hitend; /* Hit the end of the subject at some point */
1041 BOOL bsr_anycrlf; /* \R is just any CRLF, not full Unicode */
1042 const uschar *start_code; /* For use when recursing */
1043 USPTR start_subject; /* Start of the subject string */
1044 USPTR end_subject; /* End of the subject string */
1045 USPTR start_match_ptr; /* Start of matched string */
1046 USPTR end_match_ptr; /* Subject position at end match */
1047 int end_offset_top; /* Highwater mark at end of match */
1048 int capture_last; /* Most recent capture number */
1049 int start_offset; /* The start offset value */
1050 eptrblock *eptrchain; /* Chain of eptrblocks for tail recursions */
1051 int eptrn; /* Next free eptrblock */
1052 recursion_info *recursive; /* Linked list of recursion data */
1053 void *callout_data; /* To pass back to callouts */
1054 } match_data;
1055
1056 /* A similar structure is used for the same purpose by the DFA matching
1057 functions. */
1058
1059 typedef struct dfa_match_data {
1060 const uschar *start_code; /* Start of the compiled pattern */
1061 const uschar *start_subject; /* Start of the subject string */
1062 const uschar *end_subject; /* End of subject string */
1063 const uschar *tables; /* Character tables */
1064 int moptions; /* Match options */
1065 int poptions; /* Pattern options */
1066 int nltype; /* Newline type */
1067 int nllen; /* Newline string length */
1068 uschar nl[4]; /* Newline string when fixed */
1069 void *callout_data; /* To pass back to callouts */
1070 } dfa_match_data;
1071
1072 /* Bit definitions for entries in the pcre_ctypes table. */
1073
1074 #define ctype_space 0x01
1075 #define ctype_letter 0x02
1076 #define ctype_digit 0x04
1077 #define ctype_xdigit 0x08
1078 #define ctype_word 0x10 /* alphanumeric or '_' */
1079 #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
1080
1081 /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
1082 of bits for a class map. Some classes are built by combining these tables. */
1083
1084 #define cbit_space 0 /* [:space:] or \s */
1085 #define cbit_xdigit 32 /* [:xdigit:] */
1086 #define cbit_digit 64 /* [:digit:] or \d */
1087 #define cbit_upper 96 /* [:upper:] */
1088 #define cbit_lower 128 /* [:lower:] */
1089 #define cbit_word 160 /* [:word:] or \w */
1090 #define cbit_graph 192 /* [:graph:] */
1091 #define cbit_print 224 /* [:print:] */
1092 #define cbit_punct 256 /* [:punct:] */
1093 #define cbit_cntrl 288 /* [:cntrl:] */
1094 #define cbit_length 320 /* Length of the cbits table */
1095
1096 /* Offsets of the various tables from the base tables pointer, and
1097 total length. */
1098
1099 #define lcc_offset 0
1100 #define fcc_offset 256
1101 #define cbits_offset 512
1102 #define ctypes_offset (cbits_offset + cbit_length)
1103 #define tables_length (ctypes_offset + 256)
1104
1105 /* Layout of the UCP type table that translates property names into types and
1106 codes. Each entry used to point directly to a name, but to reduce the number of
1107 relocations in shared libraries, it now has an offset into a single string
1108 instead. */
1109
1110 typedef struct {
1111 pcre_uint16 name_offset;
1112 pcre_uint16 type;
1113 pcre_uint16 value;
1114 } ucp_type_table;
1115
1116
1117 /* Internal shared data tables. These are tables that are used by more than one
1118 of the exported public functions. They have to be "external" in the C sense,
1119 but are not part of the PCRE public API. The data for these tables is in the
1120 pcre_tables.c module. */
1121
1122 extern const int _pcre_utf8_table1[];
1123 extern const int _pcre_utf8_table2[];
1124 extern const int _pcre_utf8_table3[];
1125 extern const uschar _pcre_utf8_table4[];
1126
1127 extern const int _pcre_utf8_table1_size;
1128
1129 extern const char _pcre_utt_names[];
1130 extern const ucp_type_table _pcre_utt[];
1131 extern const int _pcre_utt_size;
1132
1133 extern const uschar _pcre_default_tables[];
1134
1135 extern const uschar _pcre_OP_lengths[];
1136
1137
1138 /* Internal shared functions. These are functions that are used by more than
1139 one of the exported public functions. They have to be "external" in the C
1140 sense, but are not part of the PCRE public API. */
1141
1142 extern BOOL _pcre_is_newline(const uschar *, int, const uschar *,
1143 int *, BOOL);
1144 extern int _pcre_ord2utf8(int, uschar *);
1145 extern real_pcre *_pcre_try_flipped(const real_pcre *, real_pcre *,
1146 const pcre_study_data *, pcre_study_data *);
1147 extern int _pcre_valid_utf8(const uschar *, int);
1148 extern BOOL _pcre_was_newline(const uschar *, int, const uschar *,
1149 int *, BOOL);
1150 extern BOOL _pcre_xclass(int, const uschar *);
1151
1152
1153 /* Unicode character database (UCD) */
1154
1155 typedef struct {
1156 uschar script;
1157 uschar chartype;
1158 pcre_int32 other_case;
1159 } ucd_record;
1160
1161 extern const ucd_record _pcre_ucd_records[];
1162 extern const uschar _pcre_ucd_stage1[];
1163 extern const pcre_uint16 _pcre_ucd_stage2[];
1164 extern const int _pcre_ucp_gentype[];
1165
1166
1167 /* UCD access macros */
1168
1169 #define UCD_BLOCK_SIZE 128
1170 #define GET_UCD(ch) (_pcre_ucd_records + \
1171 _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1172 UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1173
1174 #define UCD_CHARTYPE(ch) GET_UCD(ch)->chartype
1175 #define UCD_SCRIPT(ch) GET_UCD(ch)->script
1176 #define UCD_CATEGORY(ch) _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1177 #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1178
1179 #endif
1180
1181 /* 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