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

Contents of /code/trunk/pcre_internal.h

Parent Directory Parent Directory | Revision Log Revision Log


Revision 118 - (show annotations) (download)
Fri Mar 9 16:17:46 2007 UTC (7 years, 5 months ago) by ph10
File MIME type: text/plain
File size: 42923 byte(s)
Fix bug in emulated memmove() for downward moves.

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) 1997-2007 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 /* Get the definitions provided by running "configure" */
71
72 #include "config.h"
73
74 /* Standard C headers plus the external interface definition. The only time
75 setjmp and stdarg are used is when NO_RECURSE is set. */
76
77 #include <ctype.h>
78 #include <limits.h>
79 #include <setjmp.h>
80 #include <stdarg.h>
81 #include <stddef.h>
82 #include <stdio.h>
83 #include <stdlib.h>
84 #include <string.h>
85
86 #ifndef PCRE_SPY
87 #define PCRE_DEFINITION /* Win32 __declspec(export) trigger for .dll */
88 #endif
89
90 /* We need to have types that specify unsigned 16-bit and 32-bit integers. We
91 cannot determine these outside the compilation (e.g. by running a program as
92 part of "configure") because PCRE is often cross-compiled for use on other
93 systems. Instead we make use of the maximum sizes that are available at
94 preprocessor time in standard C environments. */
95
96 #if USHRT_MAX == 65535
97 typedef unsigned short pcre_uint16;
98 #elif UINT_MAX == 65535
99 typedef unsigned int pcre_uint16;
100 #else
101 #error Cannot determine a type for 16-bit unsigned integers
102 #endif
103
104 #if UINT_MAX == 4294967295
105 typedef unsigned int pcre_uint32;
106 #elif ULONG_MAX == 4294967295
107 typedef unsigned long int pcre_uint32;
108 #else
109 #error Cannot determine a type for 32-bit unsigned integers
110 #endif
111
112 /* All character handling must be done as unsigned characters. Otherwise there
113 are problems with top-bit-set characters and functions such as isspace().
114 However, we leave the interface to the outside world as char *, because that
115 should make things easier for callers. We define a short type for unsigned char
116 to save lots of typing. I tried "uchar", but it causes problems on Digital
117 Unix, where it is defined in sys/types, so use "uschar" instead. */
118
119 typedef unsigned char uschar;
120
121 /* This is an unsigned int value that no character can ever have. UTF-8
122 characters only go up to 0x7fffffff (though Unicode doesn't go beyond
123 0x0010ffff). */
124
125 #define NOTACHAR 0xffffffff
126
127 /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
128 and "all" at present). The following macros are used to package up testing for
129 newlines. NLBLOCK, PSSTART, and PSEND are defined in the various modules to
130 indicate in which datablock the parameters exist, and what the start/end of
131 string field names are. */
132
133 #define NLTYPE_FIXED 0 /* Newline is a fixed length string */
134 #define NLTYPE_ANY 1 /* Newline is any Unicode line ending */
135
136 /* This macro checks for a newline at the given position */
137
138 #define IS_NEWLINE(p) \
139 ((NLBLOCK->nltype != NLTYPE_FIXED)? \
140 ((p) < NLBLOCK->PSEND && \
141 _pcre_is_newline((p), NLBLOCK->PSEND, &(NLBLOCK->nllen), utf8) \
142 ) \
143 : \
144 ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
145 (p)[0] == NLBLOCK->nl[0] && \
146 (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
147 ) \
148 )
149
150 /* This macro checks for a newline immediately preceding the given position */
151
152 #define WAS_NEWLINE(p) \
153 ((NLBLOCK->nltype != NLTYPE_FIXED)? \
154 ((p) > NLBLOCK->PSSTART && \
155 _pcre_was_newline((p), NLBLOCK->PSSTART, &(NLBLOCK->nllen), utf8) \
156 ) \
157 : \
158 ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
159 (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
160 (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
161 ) \
162 )
163
164 /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
165 with a custom type. This makes it possible, for example, to allow pcre_exec()
166 to process subject strings that are discontinuous by using a smart pointer
167 class. It must always be possible to inspect all of the subject string in
168 pcre_exec() because of the way it backtracks. Two macros are required in the
169 normal case, for sign-unspecified and unsigned char pointers. The former is
170 used for the external interface and appears in pcre.h, which is why its name
171 must begin with PCRE_. */
172
173 #ifdef CUSTOM_SUBJECT_PTR
174 #define PCRE_SPTR CUSTOM_SUBJECT_PTR
175 #define USPTR CUSTOM_SUBJECT_PTR
176 #else
177 #define PCRE_SPTR const char *
178 #define USPTR const unsigned char *
179 #endif
180
181 /* Include the public PCRE header and the definitions of UCP character property
182 values. */
183
184 #include "pcre.h"
185 #include "ucp.h"
186
187 /* When compiling for use with the Virtual Pascal compiler, these functions
188 need to have their names changed. PCRE must be compiled with the -DVPCOMPAT
189 option on the command line. */
190
191 #ifdef VPCOMPAT
192 #define strlen(s) _strlen(s)
193 #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
194 #define memcmp(s,c,n) _memcmp(s,c,n)
195 #define memcpy(d,s,n) _memcpy(d,s,n)
196 #define memmove(d,s,n) _memmove(d,s,n)
197 #define memset(s,c,n) _memset(s,c,n)
198 #else /* VPCOMPAT */
199
200 /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
201 define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
202 is set. Otherwise, include an emulating function for those systems that have
203 neither (there some non-Unix environments where this is the case). */
204
205 #if ! HAVE_MEMMOVE
206 #undef memmove /* some systems may have a macro */
207 #if HAVE_BCOPY
208 #define memmove(a, b, c) bcopy(b, a, c)
209 #else /* HAVE_BCOPY */
210 static void *
211 pcre_memmove(void *d, const void *s, size_t n)
212 {
213 size_t i;
214 unsigned char *dest = (unsigned char *)d;
215 const unsigned char *src = (const unsigned char *)s;
216 if (dest > src)
217 {
218 dest += n;
219 src += n;
220 for (i = 0; i < n; ++i) *(--dest) = *(--src);
221 return (void *)dest;
222 }
223 else
224 {
225 for (i = 0; i < n; ++i) *dest++ = *src++;
226 return (void *)(dest - n);
227 }
228 }
229 #define memmove(a, b, c) pcre_memmove(a, b, c)
230 #endif /* not HAVE_BCOPY */
231 #endif /* not HAVE_MEMMOVE */
232 #endif /* not VPCOMPAT */
233
234
235 /* PCRE keeps offsets in its compiled code as 2-byte quantities (always stored
236 in big-endian order) by default. These are used, for example, to link from the
237 start of a subpattern to its alternatives and its end. The use of 2 bytes per
238 offset limits the size of the compiled regex to around 64K, which is big enough
239 for almost everybody. However, I received a request for an even bigger limit.
240 For this reason, and also to make the code easier to maintain, the storing and
241 loading of offsets from the byte string is now handled by the macros that are
242 defined here.
243
244 The macros are controlled by the value of LINK_SIZE. This defaults to 2 in
245 the config.h file, but can be overridden by using -D on the command line. This
246 is automated on Unix systems via the "configure" command. */
247
248 #if LINK_SIZE == 2
249
250 #define PUT(a,n,d) \
251 (a[n] = (d) >> 8), \
252 (a[(n)+1] = (d) & 255)
253
254 #define GET(a,n) \
255 (((a)[n] << 8) | (a)[(n)+1])
256
257 #define MAX_PATTERN_SIZE (1 << 16)
258
259
260 #elif LINK_SIZE == 3
261
262 #define PUT(a,n,d) \
263 (a[n] = (d) >> 16), \
264 (a[(n)+1] = (d) >> 8), \
265 (a[(n)+2] = (d) & 255)
266
267 #define GET(a,n) \
268 (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2])
269
270 #define MAX_PATTERN_SIZE (1 << 24)
271
272
273 #elif LINK_SIZE == 4
274
275 #define PUT(a,n,d) \
276 (a[n] = (d) >> 24), \
277 (a[(n)+1] = (d) >> 16), \
278 (a[(n)+2] = (d) >> 8), \
279 (a[(n)+3] = (d) & 255)
280
281 #define GET(a,n) \
282 (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
283
284 #define MAX_PATTERN_SIZE (1 << 30) /* Keep it positive */
285
286
287 #else
288 #error LINK_SIZE must be either 2, 3, or 4
289 #endif
290
291
292 /* Convenience macro defined in terms of the others */
293
294 #define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE
295
296
297 /* PCRE uses some other 2-byte quantities that do not change when the size of
298 offsets changes. There are used for repeat counts and for other things such as
299 capturing parenthesis numbers in back references. */
300
301 #define PUT2(a,n,d) \
302 a[n] = (d) >> 8; \
303 a[(n)+1] = (d) & 255
304
305 #define GET2(a,n) \
306 (((a)[n] << 8) | (a)[(n)+1])
307
308 #define PUT2INC(a,n,d) PUT2(a,n,d), a += 2
309
310
311 /* When UTF-8 encoding is being used, a character is no longer just a single
312 byte. The macros for character handling generate simple sequences when used in
313 byte-mode, and more complicated ones for UTF-8 characters. */
314
315 #ifndef SUPPORT_UTF8
316 #define GETCHAR(c, eptr) c = *eptr;
317 #define GETCHARTEST(c, eptr) c = *eptr;
318 #define GETCHARINC(c, eptr) c = *eptr++;
319 #define GETCHARINCTEST(c, eptr) c = *eptr++;
320 #define GETCHARLEN(c, eptr, len) c = *eptr;
321 #define BACKCHAR(eptr)
322
323 #else /* SUPPORT_UTF8 */
324
325 /* Get the next UTF-8 character, not advancing the pointer. This is called when
326 we know we are in UTF-8 mode. */
327
328 #define GETCHAR(c, eptr) \
329 c = *eptr; \
330 if (c >= 0xc0) \
331 { \
332 int gcii; \
333 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
334 int gcss = 6*gcaa; \
335 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
336 for (gcii = 1; gcii <= gcaa; gcii++) \
337 { \
338 gcss -= 6; \
339 c |= (eptr[gcii] & 0x3f) << gcss; \
340 } \
341 }
342
343 /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
344 pointer. */
345
346 #define GETCHARTEST(c, eptr) \
347 c = *eptr; \
348 if (utf8 && c >= 0xc0) \
349 { \
350 int gcii; \
351 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
352 int gcss = 6*gcaa; \
353 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
354 for (gcii = 1; gcii <= gcaa; gcii++) \
355 { \
356 gcss -= 6; \
357 c |= (eptr[gcii] & 0x3f) << gcss; \
358 } \
359 }
360
361 /* Get the next UTF-8 character, advancing the pointer. This is called when we
362 know we are in UTF-8 mode. */
363
364 #define GETCHARINC(c, eptr) \
365 c = *eptr++; \
366 if (c >= 0xc0) \
367 { \
368 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
369 int gcss = 6*gcaa; \
370 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
371 while (gcaa-- > 0) \
372 { \
373 gcss -= 6; \
374 c |= (*eptr++ & 0x3f) << gcss; \
375 } \
376 }
377
378 /* Get the next character, testing for UTF-8 mode, and advancing the pointer */
379
380 #define GETCHARINCTEST(c, eptr) \
381 c = *eptr++; \
382 if (utf8 && c >= 0xc0) \
383 { \
384 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
385 int gcss = 6*gcaa; \
386 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
387 while (gcaa-- > 0) \
388 { \
389 gcss -= 6; \
390 c |= (*eptr++ & 0x3f) << gcss; \
391 } \
392 }
393
394 /* Get the next UTF-8 character, not advancing the pointer, incrementing length
395 if there are extra bytes. This is called when we know we are in UTF-8 mode. */
396
397 #define GETCHARLEN(c, eptr, len) \
398 c = *eptr; \
399 if (c >= 0xc0) \
400 { \
401 int gcii; \
402 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
403 int gcss = 6*gcaa; \
404 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
405 for (gcii = 1; gcii <= gcaa; gcii++) \
406 { \
407 gcss -= 6; \
408 c |= (eptr[gcii] & 0x3f) << gcss; \
409 } \
410 len += gcaa; \
411 }
412
413 /* If the pointer is not at the start of a character, move it back until
414 it is. Called only in UTF-8 mode. */
415
416 #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;
417
418 #endif
419
420
421 /* In case there is no definition of offsetof() provided - though any proper
422 Standard C system should have one. */
423
424 #ifndef offsetof
425 #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
426 #endif
427
428
429 /* These are the public options that can change during matching. */
430
431 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
432
433 /* Private options flags start at the most significant end of the four bytes.
434 The public options defined in pcre.h start at the least significant end. Make
435 sure they don't overlap! The bits are getting a bit scarce now -- when we run
436 out, there is a dummy word in the structure that could be used for the private
437 bits. */
438
439 #define PCRE_NOPARTIAL 0x80000000 /* can't use partial with this regex */
440 #define PCRE_FIRSTSET 0x40000000 /* first_byte is set */
441 #define PCRE_REQCHSET 0x20000000 /* req_byte is set */
442 #define PCRE_STARTLINE 0x10000000 /* start after \n for multiline */
443 #define PCRE_JCHANGED 0x08000000 /* j option changes within regex */
444
445 /* Options for the "extra" block produced by pcre_study(). */
446
447 #define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */
448
449 /* Masks for identifying the public options that are permitted at compile
450 time, run time, or study time, respectively. */
451
452 #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY)
453
454 #define PUBLIC_OPTIONS \
455 (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
456 PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
457 PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
458 PCRE_DUPNAMES|PCRE_NEWLINE_BITS)
459
460 #define PUBLIC_EXEC_OPTIONS \
461 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
462 PCRE_PARTIAL|PCRE_NEWLINE_BITS)
463
464 #define PUBLIC_DFA_EXEC_OPTIONS \
465 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
466 PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)
467
468 #define PUBLIC_STUDY_OPTIONS 0 /* None defined */
469
470 /* Magic number to provide a small check against being handed junk. Also used
471 to detect whether a pattern was compiled on a host of different endianness. */
472
473 #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
474
475 /* Negative values for the firstchar and reqchar variables */
476
477 #define REQ_UNSET (-2)
478 #define REQ_NONE (-1)
479
480 /* The maximum remaining length of subject we are prepared to search for a
481 req_byte match. */
482
483 #define REQ_BYTE_MAX 1000
484
485 /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a
486 variable-length repeat, or a anything other than literal characters. */
487
488 #define REQ_CASELESS 0x0100 /* indicates caselessness */
489 #define REQ_VARY 0x0200 /* reqbyte followed non-literal item */
490
491 /* Miscellaneous definitions */
492
493 typedef int BOOL;
494
495 #define FALSE 0
496 #define TRUE 1
497
498 /* Escape items that are just an encoding of a particular data value. */
499
500 #ifndef ESC_e
501 #define ESC_e 27
502 #endif
503
504 #ifndef ESC_f
505 #define ESC_f '\f'
506 #endif
507
508 #ifndef ESC_n
509 #define ESC_n '\n'
510 #endif
511
512 #ifndef ESC_r
513 #define ESC_r '\r'
514 #endif
515
516 /* We can't officially use ESC_t because it is a POSIX reserved identifier
517 (presumably because of all the others like size_t). */
518
519 #ifndef ESC_tee
520 #define ESC_tee '\t'
521 #endif
522
523 /* Codes for different types of Unicode property */
524
525 #define PT_ANY 0 /* Any property - matches all chars */
526 #define PT_LAMP 1 /* L& - the union of Lu, Ll, Lt */
527 #define PT_GC 2 /* General characteristic (e.g. L) */
528 #define PT_PC 3 /* Particular characteristic (e.g. Lu) */
529 #define PT_SC 4 /* Script (e.g. Han) */
530
531 /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
532 contain UTF-8 characters with values greater than 255. */
533
534 #define XCL_NOT 0x01 /* Flag: this is a negative class */
535 #define XCL_MAP 0x02 /* Flag: a 32-byte map is present */
536
537 #define XCL_END 0 /* Marks end of individual items */
538 #define XCL_SINGLE 1 /* Single item (one multibyte char) follows */
539 #define XCL_RANGE 2 /* A range (two multibyte chars) follows */
540 #define XCL_PROP 3 /* Unicode property (2-byte property code follows) */
541 #define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */
542
543 /* These are escaped items that aren't just an encoding of a particular data
544 value such as \n. They must have non-zero values, as check_escape() returns
545 their negation. Also, they must appear in the same order as in the opcode
546 definitions below, up to ESC_z. There's a dummy for OP_ANY because it
547 corresponds to "." rather than an escape sequence. The final one must be
548 ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).
549 There are two tests in the code for an escape greater than ESC_b and less than
550 ESC_Z to detect the types that may be repeated. These are the types that
551 consume characters. If any new escapes are put in between that don't consume a
552 character, that code will have to change. */
553
554 enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,
555 ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_X, ESC_Z, ESC_z,
556 ESC_E, ESC_Q, ESC_k, ESC_REF };
557
558
559 /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
560 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
561 OP_EOD must correspond in order to the list of escapes immediately above.
562
563 To keep stored, compiled patterns compatible, new opcodes should be added
564 immediately before OP_BRA, where (since release 7.0) a gap is left for this
565 purpose.
566
567 *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
568 that follow must also be updated to match. There is also a table called
569 "coptable" in pcre_dfa_exec.c that must be updated. */
570
571 enum {
572 OP_END, /* 0 End of pattern */
573
574 /* Values corresponding to backslashed metacharacters */
575
576 OP_SOD, /* 1 Start of data: \A */
577 OP_SOM, /* 2 Start of match (subject + offset): \G */
578 OP_NOT_WORD_BOUNDARY, /* 3 \B */
579 OP_WORD_BOUNDARY, /* 4 \b */
580 OP_NOT_DIGIT, /* 5 \D */
581 OP_DIGIT, /* 6 \d */
582 OP_NOT_WHITESPACE, /* 7 \S */
583 OP_WHITESPACE, /* 8 \s */
584 OP_NOT_WORDCHAR, /* 9 \W */
585 OP_WORDCHAR, /* 10 \w */
586 OP_ANY, /* 11 Match any character */
587 OP_ANYBYTE, /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */
588 OP_NOTPROP, /* 13 \P (not Unicode property) */
589 OP_PROP, /* 14 \p (Unicode property) */
590 OP_ANYNL, /* 15 \R (any newline sequence) */
591 OP_EXTUNI, /* 16 \X (extended Unicode sequence */
592 OP_EODN, /* 17 End of data or \n at end of data: \Z. */
593 OP_EOD, /* 18 End of data: \z */
594
595 OP_OPT, /* 19 Set runtime options */
596 OP_CIRC, /* 20 Start of line - varies with multiline switch */
597 OP_DOLL, /* 21 End of line - varies with multiline switch */
598 OP_CHAR, /* 22 Match one character, casefully */
599 OP_CHARNC, /* 23 Match one character, caselessly */
600 OP_NOT, /* 24 Match one character, not the following one */
601
602 OP_STAR, /* 25 The maximizing and minimizing versions of */
603 OP_MINSTAR, /* 26 these six opcodes must come in pairs, with */
604 OP_PLUS, /* 27 the minimizing one second. */
605 OP_MINPLUS, /* 28 This first set applies to single characters.*/
606 OP_QUERY, /* 29 */
607 OP_MINQUERY, /* 30 */
608
609 OP_UPTO, /* 31 From 0 to n matches */
610 OP_MINUPTO, /* 32 */
611 OP_EXACT, /* 33 Exactly n matches */
612
613 OP_POSSTAR, /* 34 Possessified star */
614 OP_POSPLUS, /* 35 Possessified plus */
615 OP_POSQUERY, /* 36 Posesssified query */
616 OP_POSUPTO, /* 37 Possessified upto */
617
618 OP_NOTSTAR, /* 38 The maximizing and minimizing versions of */
619 OP_NOTMINSTAR, /* 39 these six opcodes must come in pairs, with */
620 OP_NOTPLUS, /* 40 the minimizing one second. They must be in */
621 OP_NOTMINPLUS, /* 41 exactly the same order as those above. */
622 OP_NOTQUERY, /* 42 This set applies to "not" single characters. */
623 OP_NOTMINQUERY, /* 43 */
624
625 OP_NOTUPTO, /* 44 From 0 to n matches */
626 OP_NOTMINUPTO, /* 45 */
627 OP_NOTEXACT, /* 46 Exactly n matches */
628
629 OP_NOTPOSSTAR, /* 47 Possessified versions */
630 OP_NOTPOSPLUS, /* 48 */
631 OP_NOTPOSQUERY, /* 49 */
632 OP_NOTPOSUPTO, /* 50 */
633
634 OP_TYPESTAR, /* 51 The maximizing and minimizing versions of */
635 OP_TYPEMINSTAR, /* 52 these six opcodes must come in pairs, with */
636 OP_TYPEPLUS, /* 53 the minimizing one second. These codes must */
637 OP_TYPEMINPLUS, /* 54 be in exactly the same order as those above. */
638 OP_TYPEQUERY, /* 55 This set applies to character types such as \d */
639 OP_TYPEMINQUERY, /* 56 */
640
641 OP_TYPEUPTO, /* 57 From 0 to n matches */
642 OP_TYPEMINUPTO, /* 58 */
643 OP_TYPEEXACT, /* 59 Exactly n matches */
644
645 OP_TYPEPOSSTAR, /* 60 Possessified versions */
646 OP_TYPEPOSPLUS, /* 61 */
647 OP_TYPEPOSQUERY, /* 62 */
648 OP_TYPEPOSUPTO, /* 63 */
649
650 OP_CRSTAR, /* 64 The maximizing and minimizing versions of */
651 OP_CRMINSTAR, /* 65 all these opcodes must come in pairs, with */
652 OP_CRPLUS, /* 66 the minimizing one second. These codes must */
653 OP_CRMINPLUS, /* 67 be in exactly the same order as those above. */
654 OP_CRQUERY, /* 68 These are for character classes and back refs */
655 OP_CRMINQUERY, /* 69 */
656 OP_CRRANGE, /* 70 These are different to the three sets above. */
657 OP_CRMINRANGE, /* 71 */
658
659 OP_CLASS, /* 72 Match a character class, chars < 256 only */
660 OP_NCLASS, /* 73 Same, but the bitmap was created from a negative
661 class - the difference is relevant only when a UTF-8
662 character > 255 is encountered. */
663
664 OP_XCLASS, /* 74 Extended class for handling UTF-8 chars within the
665 class. This does both positive and negative. */
666
667 OP_REF, /* 75 Match a back reference */
668 OP_RECURSE, /* 76 Match a numbered subpattern (possibly recursive) */
669 OP_CALLOUT, /* 77 Call out to external function if provided */
670
671 OP_ALT, /* 78 Start of alternation */
672 OP_KET, /* 79 End of group that doesn't have an unbounded repeat */
673 OP_KETRMAX, /* 80 These two must remain together and in this */
674 OP_KETRMIN, /* 81 order. They are for groups the repeat for ever. */
675
676 /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
677
678 OP_ASSERT, /* 82 Positive lookahead */
679 OP_ASSERT_NOT, /* 83 Negative lookahead */
680 OP_ASSERTBACK, /* 84 Positive lookbehind */
681 OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */
682 OP_REVERSE, /* 86 Move pointer back - used in lookbehind assertions */
683
684 /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
685 as there's a test for >= ONCE for a subpattern that isn't an assertion. */
686
687 OP_ONCE, /* 87 Atomic group */
688 OP_BRA, /* 88 Start of non-capturing bracket */
689 OP_CBRA, /* 89 Start of capturing bracket */
690 OP_COND, /* 90 Conditional group */
691
692 /* These three must follow the previous three, in the same order. There's a
693 check for >= SBRA to distinguish the two sets. */
694
695 OP_SBRA, /* 91 Start of non-capturing bracket, check empty */
696 OP_SCBRA, /* 92 Start of capturing bracket, check empty */
697 OP_SCOND, /* 93 Conditional group, check empty */
698
699 OP_CREF, /* 94 Used to hold a capture number as condition */
700 OP_RREF, /* 95 Used to hold a recursion number as condition */
701 OP_DEF, /* 96 The DEFINE condition */
702
703 OP_BRAZERO, /* 97 These two must remain together and in this */
704 OP_BRAMINZERO /* 98 order. */
705 };
706
707
708 /* This macro defines textual names for all the opcodes. These are used only
709 for debugging. The macro is referenced only in pcre_printint.c. */
710
711 #define OP_NAME_LIST \
712 "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d", \
713 "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte", \
714 "notprop", "prop", "anynl", "extuni", \
715 "\\Z", "\\z", \
716 "Opt", "^", "$", "char", "charnc", "not", \
717 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
718 "*+","++", "?+", "{", \
719 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
720 "*+","++", "?+", "{", \
721 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
722 "*+","++", "?+", "{", \
723 "*", "*?", "+", "+?", "?", "??", "{", "{", \
724 "class", "nclass", "xclass", "Ref", "Recurse", "Callout", \
725 "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not", \
726 "AssertB", "AssertB not", "Reverse", \
727 "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond", \
728 "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"
729
730
731 /* This macro defines the length of fixed length operations in the compiled
732 regex. The lengths are used when searching for specific things, and also in the
733 debugging printing of a compiled regex. We use a macro so that it can be
734 defined close to the definitions of the opcodes themselves.
735
736 As things have been extended, some of these are no longer fixed lenths, but are
737 minima instead. For example, the length of a single-character repeat may vary
738 in UTF-8 mode. The code that uses this table must know about such things. */
739
740 #define OP_LENGTHS \
741 1, /* End */ \
742 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \
743 1, 1, /* Any, Anybyte */ \
744 3, 3, 1, 1, /* NOTPROP, PROP, EXTUNI, ANYNL */ \
745 1, 1, 2, 1, 1, /* \Z, \z, Opt, ^, $ */ \
746 2, /* Char - the minimum length */ \
747 2, /* Charnc - the minimum length */ \
748 2, /* not */ \
749 /* Positive single-char repeats ** These are */ \
750 2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** minima in */ \
751 4, 4, 4, /* upto, minupto, exact ** UTF-8 mode */ \
752 2, 2, 2, 4, /* *+, ++, ?+, upto+ */ \
753 /* Negative single-char repeats - only for chars < 256 */ \
754 2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \
755 4, 4, 4, /* NOT upto, minupto, exact */ \
756 2, 2, 2, 4, /* Possessive *, +, ?, upto */ \
757 /* Positive type repeats */ \
758 2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \
759 4, 4, 4, /* Type upto, minupto, exact */ \
760 2, 2, 2, 4, /* Possessive *+, ++, ?+, upto+ */ \
761 /* Character class & ref repeats */ \
762 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \
763 5, 5, /* CRRANGE, CRMINRANGE */ \
764 33, /* CLASS */ \
765 33, /* NCLASS */ \
766 0, /* XCLASS - variable length */ \
767 3, /* REF */ \
768 1+LINK_SIZE, /* RECURSE */ \
769 2+2*LINK_SIZE, /* CALLOUT */ \
770 1+LINK_SIZE, /* Alt */ \
771 1+LINK_SIZE, /* Ket */ \
772 1+LINK_SIZE, /* KetRmax */ \
773 1+LINK_SIZE, /* KetRmin */ \
774 1+LINK_SIZE, /* Assert */ \
775 1+LINK_SIZE, /* Assert not */ \
776 1+LINK_SIZE, /* Assert behind */ \
777 1+LINK_SIZE, /* Assert behind not */ \
778 1+LINK_SIZE, /* Reverse */ \
779 1+LINK_SIZE, /* ONCE */ \
780 1+LINK_SIZE, /* BRA */ \
781 3+LINK_SIZE, /* CBRA */ \
782 1+LINK_SIZE, /* COND */ \
783 1+LINK_SIZE, /* SBRA */ \
784 3+LINK_SIZE, /* SCBRA */ \
785 1+LINK_SIZE, /* SCOND */ \
786 3, /* CREF */ \
787 3, /* RREF */ \
788 1, /* DEF */ \
789 1, 1, /* BRAZERO, BRAMINZERO */ \
790
791
792 /* A magic value for OP_RREF to indicate the "any recursion" condition. */
793
794 #define RREF_ANY 0xffff
795
796 /* Error code numbers. They are given names so that they can more easily be
797 tracked. */
798
799 enum { ERR0, ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9,
800 ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
801 ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
802 ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
803 ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
804 ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57 };
805
806 /* The real format of the start of the pcre block; the index of names and the
807 code vector run on as long as necessary after the end. We store an explicit
808 offset to the name table so that if a regex is compiled on one host, saved, and
809 then run on another where the size of pointers is different, all might still
810 be well. For the case of compiled-on-4 and run-on-8, we include an extra
811 pointer that is always NULL. For future-proofing, a few dummy fields were
812 originally included - even though you can never get this planning right - but
813 there is only one left now.
814
815 NOTE NOTE NOTE:
816 Because people can now save and re-use compiled patterns, any additions to this
817 structure should be made at the end, and something earlier (e.g. a new
818 flag in the options or one of the dummy fields) should indicate that the new
819 fields are present. Currently PCRE always sets the dummy fields to zero.
820 NOTE NOTE NOTE:
821 */
822
823 typedef struct real_pcre {
824 pcre_uint32 magic_number;
825 pcre_uint32 size; /* Total that was malloced */
826 pcre_uint32 options;
827 pcre_uint32 dummy1; /* For future use, maybe */
828
829 pcre_uint16 top_bracket;
830 pcre_uint16 top_backref;
831 pcre_uint16 first_byte;
832 pcre_uint16 req_byte;
833 pcre_uint16 name_table_offset; /* Offset to name table that follows */
834 pcre_uint16 name_entry_size; /* Size of any name items */
835 pcre_uint16 name_count; /* Number of name items */
836 pcre_uint16 ref_count; /* Reference count */
837
838 const unsigned char *tables; /* Pointer to tables or NULL for std */
839 const unsigned char *nullpad; /* NULL padding */
840 } real_pcre;
841
842 /* The format of the block used to store data from pcre_study(). The same
843 remark (see NOTE above) about extending this structure applies. */
844
845 typedef struct pcre_study_data {
846 pcre_uint32 size; /* Total that was malloced */
847 pcre_uint32 options;
848 uschar start_bits[32];
849 } pcre_study_data;
850
851 /* Structure for passing "static" information around between the functions
852 doing the compiling, so that they are thread-safe. */
853
854 typedef struct compile_data {
855 const uschar *lcc; /* Points to lower casing table */
856 const uschar *fcc; /* Points to case-flipping table */
857 const uschar *cbits; /* Points to character type table */
858 const uschar *ctypes; /* Points to table of type maps */
859 const uschar *start_workspace;/* The start of working space */
860 const uschar *start_code; /* The start of the compiled code */
861 const uschar *start_pattern; /* The start of the pattern */
862 const uschar *end_pattern; /* The end of the pattern */
863 uschar *hwm; /* High watermark of workspace */
864 uschar *name_table; /* The name/number table */
865 int names_found; /* Number of entries so far */
866 int name_entry_size; /* Size of each entry */
867 int bracount; /* Count of capturing parens */
868 int top_backref; /* Maximum back reference */
869 unsigned int backref_map; /* Bitmap of low back refs */
870 int external_options; /* External (initial) options */
871 int req_varyopt; /* "After variable item" flag for reqbyte */
872 BOOL nopartial; /* Set TRUE if partial won't work */
873 int nltype; /* Newline type */
874 int nllen; /* Newline string length */
875 uschar nl[4]; /* Newline string when fixed length */
876 } compile_data;
877
878 /* Structure for maintaining a chain of pointers to the currently incomplete
879 branches, for testing for left recursion. */
880
881 typedef struct branch_chain {
882 struct branch_chain *outer;
883 uschar *current;
884 } branch_chain;
885
886 /* Structure for items in a linked list that represents an explicit recursive
887 call within the pattern. */
888
889 typedef struct recursion_info {
890 struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
891 int group_num; /* Number of group that was called */
892 const uschar *after_call; /* "Return value": points after the call in the expr */
893 USPTR save_start; /* Old value of md->start_match */
894 int *offset_save; /* Pointer to start of saved offsets */
895 int saved_max; /* Number of saved offsets */
896 } recursion_info;
897
898 /* When compiling in a mode that doesn't use recursive calls to match(),
899 a structure is used to remember local variables on the heap. It is defined in
900 pcre_exec.c, close to the match() function, so that it is easy to keep it in
901 step with any changes of local variable. However, the pointer to the current
902 frame must be saved in some "static" place over a longjmp(). We declare the
903 structure here so that we can put a pointer in the match_data structure. NOTE:
904 This isn't used for a "normal" compilation of pcre. */
905
906 struct heapframe;
907
908 /* Structure for building a chain of data for holding the values of the subject
909 pointer at the start of each subpattern, so as to detect when an empty string
910 has been matched by a subpattern - to break infinite loops. */
911
912 typedef struct eptrblock {
913 struct eptrblock *epb_prev;
914 USPTR epb_saved_eptr;
915 } eptrblock;
916
917
918 /* Structure for passing "static" information around between the functions
919 doing traditional NFA matching, so that they are thread-safe. */
920
921 typedef struct match_data {
922 unsigned long int match_call_count; /* As it says */
923 unsigned long int match_limit; /* As it says */
924 unsigned long int match_limit_recursion; /* As it says */
925 int *offset_vector; /* Offset vector */
926 int offset_end; /* One past the end */
927 int offset_max; /* The maximum usable for return data */
928 int nltype; /* Newline type */
929 int nllen; /* Newline string length */
930 uschar nl[4]; /* Newline string when fixed */
931 const uschar *lcc; /* Points to lower casing table */
932 const uschar *ctypes; /* Points to table of type maps */
933 BOOL offset_overflow; /* Set if too many extractions */
934 BOOL notbol; /* NOTBOL flag */
935 BOOL noteol; /* NOTEOL flag */
936 BOOL utf8; /* UTF8 flag */
937 BOOL endonly; /* Dollar not before final \n */
938 BOOL notempty; /* Empty string match not wanted */
939 BOOL partial; /* PARTIAL flag */
940 BOOL hitend; /* Hit the end of the subject at some point */
941 const uschar *start_code; /* For use when recursing */
942 USPTR start_subject; /* Start of the subject string */
943 USPTR end_subject; /* End of the subject string */
944 USPTR start_match; /* Start of this match attempt */
945 USPTR end_match_ptr; /* Subject position at end match */
946 int end_offset_top; /* Highwater mark at end of match */
947 int capture_last; /* Most recent capture number */
948 int start_offset; /* The start offset value */
949 eptrblock *eptrchain; /* Chain of eptrblocks for tail recursions */
950 int eptrn; /* Next free eptrblock */
951 recursion_info *recursive; /* Linked list of recursion data */
952 void *callout_data; /* To pass back to callouts */
953 struct heapframe *thisframe; /* Used only when compiling for no recursion */
954 } match_data;
955
956 /* A similar structure is used for the same purpose by the DFA matching
957 functions. */
958
959 typedef struct dfa_match_data {
960 const uschar *start_code; /* Start of the compiled pattern */
961 const uschar *start_subject; /* Start of the subject string */
962 const uschar *end_subject; /* End of subject string */
963 const uschar *tables; /* Character tables */
964 int moptions; /* Match options */
965 int poptions; /* Pattern options */
966 int nltype; /* Newline type */
967 int nllen; /* Newline string length */
968 uschar nl[4]; /* Newline string when fixed */
969 void *callout_data; /* To pass back to callouts */
970 } dfa_match_data;
971
972 /* Bit definitions for entries in the pcre_ctypes table. */
973
974 #define ctype_space 0x01
975 #define ctype_letter 0x02
976 #define ctype_digit 0x04
977 #define ctype_xdigit 0x08
978 #define ctype_word 0x10 /* alphameric or '_' */
979 #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
980
981 /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
982 of bits for a class map. Some classes are built by combining these tables. */
983
984 #define cbit_space 0 /* [:space:] or \s */
985 #define cbit_xdigit 32 /* [:xdigit:] */
986 #define cbit_digit 64 /* [:digit:] or \d */
987 #define cbit_upper 96 /* [:upper:] */
988 #define cbit_lower 128 /* [:lower:] */
989 #define cbit_word 160 /* [:word:] or \w */
990 #define cbit_graph 192 /* [:graph:] */
991 #define cbit_print 224 /* [:print:] */
992 #define cbit_punct 256 /* [:punct:] */
993 #define cbit_cntrl 288 /* [:cntrl:] */
994 #define cbit_length 320 /* Length of the cbits table */
995
996 /* Offsets of the various tables from the base tables pointer, and
997 total length. */
998
999 #define lcc_offset 0
1000 #define fcc_offset 256
1001 #define cbits_offset 512
1002 #define ctypes_offset (cbits_offset + cbit_length)
1003 #define tables_length (ctypes_offset + 256)
1004
1005 /* Layout of the UCP type table that translates property names into types and
1006 codes. */
1007
1008 typedef struct {
1009 const char *name;
1010 pcre_uint16 type;
1011 pcre_uint16 value;
1012 } ucp_type_table;
1013
1014
1015 /* Internal shared data tables. These are tables that are used by more than one
1016 of the exported public functions. They have to be "external" in the C sense,
1017 but are not part of the PCRE public API. The data for these tables is in the
1018 pcre_tables.c module. */
1019
1020 extern const int _pcre_utf8_table1[];
1021 extern const int _pcre_utf8_table2[];
1022 extern const int _pcre_utf8_table3[];
1023 extern const uschar _pcre_utf8_table4[];
1024
1025 extern const int _pcre_utf8_table1_size;
1026
1027 extern const ucp_type_table _pcre_utt[];
1028 extern const int _pcre_utt_size;
1029
1030 extern const uschar _pcre_default_tables[];
1031
1032 extern const uschar _pcre_OP_lengths[];
1033
1034
1035 /* Internal shared functions. These are functions that are used by more than
1036 one of the exported public functions. They have to be "external" in the C
1037 sense, but are not part of the PCRE public API. */
1038
1039 extern BOOL _pcre_is_newline(const uschar *, const uschar *, int *,
1040 BOOL);
1041 extern int _pcre_ord2utf8(int, uschar *);
1042 extern real_pcre *_pcre_try_flipped(const real_pcre *, real_pcre *,
1043 const pcre_study_data *, pcre_study_data *);
1044 extern int _pcre_ucp_findprop(const unsigned int, int *, int *);
1045 extern unsigned int _pcre_ucp_othercase(const unsigned int);
1046 extern int _pcre_valid_utf8(const uschar *, int);
1047 extern BOOL _pcre_was_newline(const uschar *, const uschar *, int *,
1048 BOOL);
1049 extern BOOL _pcre_xclass(int, const uschar *);
1050
1051 #endif
1052
1053 /* 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