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

Contents of /code/trunk/internal.h

Parent Directory Parent Directory | Revision Log Revision Log


Revision 71 - (show annotations) (download)
Sat Feb 24 21:40:24 2007 UTC (7 years, 5 months ago) by nigel
File MIME type: text/plain
File size: 28334 byte(s)
Load pcre-4.4 into code/trunk.

1 /*************************************************
2 * Perl-Compatible Regular Expressions *
3 *************************************************/
4
5
6 /* This 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. See
8 the file Tech.Notes for some information on the internals.
9
10 Written by: Philip Hazel <ph10@cam.ac.uk>
11
12 Copyright (c) 1997-2003 University of Cambridge
13
14 -----------------------------------------------------------------------------
15 Permission is granted to anyone to use this software for any purpose on any
16 computer system, and to redistribute it freely, subject to the following
17 restrictions:
18
19 1. This software is distributed in the hope that it will be useful,
20 but WITHOUT ANY WARRANTY; without even the implied warranty of
21 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
22
23 2. The origin of this software must not be misrepresented, either by
24 explicit claim or by omission.
25
26 3. Altered versions must be plainly marked as such, and must not be
27 misrepresented as being the original software.
28
29 4. If PCRE is embedded in any software that is released under the GNU
30 General Purpose Licence (GPL), then the terms of that licence shall
31 supersede any condition above with which it is incompatible.
32 -----------------------------------------------------------------------------
33 */
34
35 /* This header contains definitions that are shared between the different
36 modules, but which are not relevant to the outside. */
37
38 /* Get the definitions provided by running "configure" */
39
40 #include "config.h"
41
42 /* When compiling for use with the Virtual Pascal compiler, these functions
43 need to have their names changed. PCRE must be compiled with the -DVPCOMPAT
44 option on the command line. */
45
46 #ifdef VPCOMPAT
47 #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
48 #define memcpy(d,s,n) _memcpy(d,s,n)
49 #define memmove(d,s,n) _memmove(d,s,n)
50 #define memset(s,c,n) _memset(s,c,n)
51 #else /* VPCOMPAT */
52
53 /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
54 define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
55 is set. Otherwise, include an emulating function for those systems that have
56 neither (there some non-Unix environments where this is the case). This assumes
57 that all calls to memmove are moving strings upwards in store, which is the
58 case in PCRE. */
59
60 #if ! HAVE_MEMMOVE
61 #undef memmove /* some systems may have a macro */
62 #if HAVE_BCOPY
63 #define memmove(a, b, c) bcopy(b, a, c)
64 #else /* HAVE_BCOPY */
65 void *
66 pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)
67 {
68 int i;
69 dest += n;
70 src += n;
71 for (i = 0; i < n; ++i) *(--dest) = *(--src);
72 }
73 #define memmove(a, b, c) pcre_memmove(a, b, c)
74 #endif /* not HAVE_BCOPY */
75 #endif /* not HAVE_MEMMOVE */
76 #endif /* not VPCOMPAT */
77
78
79 /* PCRE keeps offsets in its compiled code as 2-byte quantities by default.
80 These are used, for example, to link from the start of a subpattern to its
81 alternatives and its end. The use of 2 bytes per offset limits the size of the
82 compiled regex to around 64K, which is big enough for almost everybody.
83 However, I received a request for an even bigger limit. For this reason, and
84 also to make the code easier to maintain, the storing and loading of offsets
85 from the byte string is now handled by the macros that are defined here.
86
87 The macros are controlled by the value of LINK_SIZE. This defaults to 2 in
88 the config.h file, but can be overridden by using -D on the command line. This
89 is automated on Unix systems via the "configure" command. */
90
91 #if LINK_SIZE == 2
92
93 #define PUT(a,n,d) \
94 (a[n] = (d) >> 8), \
95 (a[(n)+1] = (d) & 255)
96
97 #define GET(a,n) \
98 (((a)[n] << 8) | (a)[(n)+1])
99
100 #define MAX_PATTERN_SIZE (1 << 16)
101
102
103 #elif LINK_SIZE == 3
104
105 #define PUT(a,n,d) \
106 (a[n] = (d) >> 16), \
107 (a[(n)+1] = (d) >> 8), \
108 (a[(n)+2] = (d) & 255)
109
110 #define GET(a,n) \
111 (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2])
112
113 #define MAX_PATTERN_SIZE (1 << 24)
114
115
116 #elif LINK_SIZE == 4
117
118 #define PUT(a,n,d) \
119 (a[n] = (d) >> 24), \
120 (a[(n)+1] = (d) >> 16), \
121 (a[(n)+2] = (d) >> 8), \
122 (a[(n)+3] = (d) & 255)
123
124 #define GET(a,n) \
125 (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
126
127 #define MAX_PATTERN_SIZE (1 << 30) /* Keep it positive */
128
129
130 #else
131 #error LINK_SIZE must be either 2, 3, or 4
132 #endif
133
134
135 /* Convenience macro defined in terms of the others */
136
137 #define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE
138
139
140 /* PCRE uses some other 2-byte quantities that do not change when the size of
141 offsets changes. There are used for repeat counts and for other things such as
142 capturing parenthesis numbers in back references. */
143
144 #define PUT2(a,n,d) \
145 a[n] = (d) >> 8; \
146 a[(n)+1] = (d) & 255
147
148 #define GET2(a,n) \
149 (((a)[n] << 8) | (a)[(n)+1])
150
151 #define PUT2INC(a,n,d) PUT2(a,n,d), a += 2
152
153
154 /* Standard C headers plus the external interface definition */
155
156 #include <ctype.h>
157 #include <limits.h>
158 #include <stddef.h>
159 #include <stdio.h>
160 #include <stdlib.h>
161 #include <string.h>
162
163 #ifndef PCRE_SPY
164 #define PCRE_DEFINITION /* Win32 __declspec(export) trigger for .dll */
165 #endif
166
167 #include "pcre.h"
168
169 /* In case there is no definition of offsetof() provided - though any proper
170 Standard C system should have one. */
171
172 #ifndef offsetof
173 #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
174 #endif
175
176 /* These are the public options that can change during matching. */
177
178 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
179
180 /* Private options flags start at the most significant end of the four bytes,
181 but skip the top bit so we can use ints for convenience without getting tangled
182 with negative values. The public options defined in pcre.h start at the least
183 significant end. Make sure they don't overlap, though now that we have expanded
184 to four bytes there is plenty of space. */
185
186 #define PCRE_FIRSTSET 0x40000000 /* first_byte is set */
187 #define PCRE_REQCHSET 0x20000000 /* req_byte is set */
188 #define PCRE_STARTLINE 0x10000000 /* start after \n for multiline */
189 #define PCRE_ICHANGED 0x08000000 /* i option changes within regex */
190
191 /* Options for the "extra" block produced by pcre_study(). */
192
193 #define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */
194
195 /* Masks for identifying the public options which are permitted at compile
196 time, run time or study time, respectively. */
197
198 #define PUBLIC_OPTIONS \
199 (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
200 PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
201 PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK)
202
203 #define PUBLIC_EXEC_OPTIONS \
204 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK)
205
206 #define PUBLIC_STUDY_OPTIONS 0 /* None defined */
207
208 /* Magic number to provide a small check against being handed junk. */
209
210 #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
211
212 /* Negative values for the firstchar and reqchar variables */
213
214 #define REQ_UNSET (-2)
215 #define REQ_NONE (-1)
216
217 /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a
218 variable-length repeat, or a anything other than literal characters. */
219
220 #define REQ_CASELESS 0x0100 /* indicates caselessness */
221 #define REQ_VARY 0x0200 /* reqbyte followed non-literal item */
222
223 /* Miscellaneous definitions */
224
225 typedef int BOOL;
226
227 #define FALSE 0
228 #define TRUE 1
229
230 /* Escape items that are just an encoding of a particular data value. Note that
231 ESC_n is defined as yet another macro, which is set in config.h to either \n
232 (the default) or \r (which some people want). */
233
234 #ifndef ESC_e
235 #define ESC_e 27
236 #endif
237
238 #ifndef ESC_f
239 #define ESC_f '\f'
240 #endif
241
242 #ifndef ESC_n
243 #define ESC_n NEWLINE
244 #endif
245
246 #ifndef ESC_r
247 #define ESC_r '\r'
248 #endif
249
250 /* We can't officially use ESC_t because it is a POSIX reserved identifier
251 (presumably because of all the others like size_t). */
252
253 #ifndef ESC_tee
254 #define ESC_tee '\t'
255 #endif
256
257 /* These are escaped items that aren't just an encoding of a particular data
258 value such as \n. They must have non-zero values, as check_escape() returns
259 their negation. Also, they must appear in the same order as in the opcode
260 definitions below, up to ESC_z. There's a dummy for OP_ANY because it
261 corresponds to "." rather than an escape sequence. The final one must be
262 ESC_REF as subsequent values are used for \1, \2, \3, etc. There is are two
263 tests in the code for an escape greater than ESC_b and less than ESC_Z to
264 detect the types that may be repeated. These are the types that consume a
265 character. If any new escapes are put in between that don't consume a
266 character, that code will have to change. */
267
268 enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,
269 ESC_w, ESC_dum1, ESC_C, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_REF };
270
271 /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
272 contain UTF-8 characters with values greater than 255. */
273
274 #define XCL_NOT 0x01 /* Flag: this is a negative class */
275 #define XCL_MAP 0x02 /* Flag: a 32-byte map is present */
276
277 #define XCL_END 0 /* Marks end of individual items */
278 #define XCL_SINGLE 1 /* Single item (one multibyte char) follows */
279 #define XCL_RANGE 2 /* A range (two multibyte chars) follows */
280
281
282 /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
283 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
284 OP_EOD must correspond in order to the list of escapes immediately above.
285 Note that whenever this list is updated, the two macro definitions that follow
286 must also be updated to match. */
287
288 enum {
289 OP_END, /* 0 End of pattern */
290
291 /* Values corresponding to backslashed metacharacters */
292
293 OP_SOD, /* 1 Start of data: \A */
294 OP_SOM, /* 2 Start of match (subject + offset): \G */
295 OP_NOT_WORD_BOUNDARY, /* 3 \B */
296 OP_WORD_BOUNDARY, /* 4 \b */
297 OP_NOT_DIGIT, /* 5 \D */
298 OP_DIGIT, /* 6 \d */
299 OP_NOT_WHITESPACE, /* 7 \S */
300 OP_WHITESPACE, /* 8 \s */
301 OP_NOT_WORDCHAR, /* 9 \W */
302 OP_WORDCHAR, /* 10 \w */
303 OP_ANY, /* 11 Match any character */
304 OP_ANYBYTE, /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */
305 OP_EODN, /* 13 End of data or \n at end of data: \Z. */
306 OP_EOD, /* 14 End of data: \z */
307
308 OP_OPT, /* 15 Set runtime options */
309 OP_CIRC, /* 16 Start of line - varies with multiline switch */
310 OP_DOLL, /* 17 End of line - varies with multiline switch */
311 OP_CHARS, /* 18 Match string of characters */
312 OP_NOT, /* 19 Match anything but the following char */
313
314 OP_STAR, /* 20 The maximizing and minimizing versions of */
315 OP_MINSTAR, /* 21 all these opcodes must come in pairs, with */
316 OP_PLUS, /* 22 the minimizing one second. */
317 OP_MINPLUS, /* 23 This first set applies to single characters */
318 OP_QUERY, /* 24 */
319 OP_MINQUERY, /* 25 */
320 OP_UPTO, /* 26 From 0 to n matches */
321 OP_MINUPTO, /* 27 */
322 OP_EXACT, /* 28 Exactly n matches */
323
324 OP_NOTSTAR, /* 29 The maximizing and minimizing versions of */
325 OP_NOTMINSTAR, /* 30 all these opcodes must come in pairs, with */
326 OP_NOTPLUS, /* 31 the minimizing one second. */
327 OP_NOTMINPLUS, /* 32 This set applies to "not" single characters */
328 OP_NOTQUERY, /* 33 */
329 OP_NOTMINQUERY, /* 34 */
330 OP_NOTUPTO, /* 35 From 0 to n matches */
331 OP_NOTMINUPTO, /* 36 */
332 OP_NOTEXACT, /* 37 Exactly n matches */
333
334 OP_TYPESTAR, /* 38 The maximizing and minimizing versions of */
335 OP_TYPEMINSTAR, /* 39 all these opcodes must come in pairs, with */
336 OP_TYPEPLUS, /* 40 the minimizing one second. These codes must */
337 OP_TYPEMINPLUS, /* 41 be in exactly the same order as those above. */
338 OP_TYPEQUERY, /* 42 This set applies to character types such as \d */
339 OP_TYPEMINQUERY, /* 43 */
340 OP_TYPEUPTO, /* 44 From 0 to n matches */
341 OP_TYPEMINUPTO, /* 45 */
342 OP_TYPEEXACT, /* 46 Exactly n matches */
343
344 OP_CRSTAR, /* 47 The maximizing and minimizing versions of */
345 OP_CRMINSTAR, /* 48 all these opcodes must come in pairs, with */
346 OP_CRPLUS, /* 49 the minimizing one second. These codes must */
347 OP_CRMINPLUS, /* 50 be in exactly the same order as those above. */
348 OP_CRQUERY, /* 51 These are for character classes and back refs */
349 OP_CRMINQUERY, /* 52 */
350 OP_CRRANGE, /* 53 These are different to the three seta above. */
351 OP_CRMINRANGE, /* 54 */
352
353 OP_CLASS, /* 55 Match a character class, chars < 256 only */
354 OP_NCLASS, /* 56 Same, but the bitmap was created from a negative
355 class - the difference is relevant only when a UTF-8
356 character > 255 is encountered. */
357
358 OP_XCLASS, /* 56 Extended class for handling UTF-8 chars within the
359 class. This does both positive and negative. */
360
361 OP_REF, /* 57 Match a back reference */
362 OP_RECURSE, /* 58 Match a numbered subpattern (possibly recursive) */
363 OP_CALLOUT, /* 59 Call out to external function if provided */
364
365 OP_ALT, /* 60 Start of alternation */
366 OP_KET, /* 61 End of group that doesn't have an unbounded repeat */
367 OP_KETRMAX, /* 62 These two must remain together and in this */
368 OP_KETRMIN, /* 63 order. They are for groups the repeat for ever. */
369
370 /* The assertions must come before ONCE and COND */
371
372 OP_ASSERT, /* 64 Positive lookahead */
373 OP_ASSERT_NOT, /* 65 Negative lookahead */
374 OP_ASSERTBACK, /* 66 Positive lookbehind */
375 OP_ASSERTBACK_NOT, /* 67 Negative lookbehind */
376 OP_REVERSE, /* 68 Move pointer back - used in lookbehind assertions */
377
378 /* ONCE and COND must come after the assertions, with ONCE first, as there's
379 a test for >= ONCE for a subpattern that isn't an assertion. */
380
381 OP_ONCE, /* 69 Once matched, don't back up into the subpattern */
382 OP_COND, /* 70 Conditional group */
383 OP_CREF, /* 71 Used to hold an extraction string number (cond ref) */
384
385 OP_BRAZERO, /* 72 These two must remain together and in this */
386 OP_BRAMINZERO, /* 73 order. */
387
388 OP_BRANUMBER, /* 74 Used for extracting brackets whose number is greater
389 than can fit into an opcode. */
390
391 OP_BRA /* 75 This and greater values are used for brackets that
392 extract substrings up to a basic limit. After that,
393 use is made of OP_BRANUMBER. */
394 };
395
396 /* WARNING: There is an implicit assumption in study.c that all opcodes are
397 less than 128 in value. This makes handling UTF-8 character sequences easier.
398 */
399
400
401 /* This macro defines textual names for all the opcodes. There are used only
402 for debugging, in pcre.c when DEBUG is defined, and also in pcretest.c. The
403 macro is referenced only in printint.c. */
404
405 #define OP_NAME_LIST \
406 "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d", \
407 "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte", "\\Z", "\\z", \
408 "Opt", "^", "$", "chars", "not", \
409 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
410 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
411 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
412 "*", "*?", "+", "+?", "?", "??", "{", "{", \
413 "class", "nclass", "xclass", "Ref", "Recurse", "Callout", \
414 "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not", \
415 "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\
416 "Brazero", "Braminzero", "Branumber", "Bra"
417
418
419 /* This macro defines the length of fixed length operations in the compiled
420 regex. The lengths are used when searching for specific things, and also in the
421 debugging printing of a compiled regex. We use a macro so that it can be
422 incorporated both into pcre.c and pcretest.c without being publicly exposed.
423
424 As things have been extended, some of these are no longer fixed lenths, but are
425 minima instead. For example, the length of a single-character repeat may vary
426 in UTF-8 mode. The code that uses this table must know about such things. */
427
428 #define OP_LENGTHS \
429 1, /* End */ \
430 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \
431 1, 1, 1, 1, 2, 1, 1, /* Any, Anybyte, \Z, \z, Opt, ^, $ */ \
432 2, /* Chars - the minimum length */ \
433 2, /* not */ \
434 /* Positive single-char repeats */ \
435 2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** These are */ \
436 4, 4, 4, /* upto, minupto, exact ** minima */ \
437 /* Negative single-char repeats */ \
438 2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \
439 4, 4, 4, /* NOT upto, minupto, exact */ \
440 /* Positive type repeats */ \
441 2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \
442 4, 4, 4, /* Type upto, minupto, exact */ \
443 /* Character class & ref repeats */ \
444 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \
445 5, 5, /* CRRANGE, CRMINRANGE */ \
446 33, /* CLASS */ \
447 33, /* NCLASS */ \
448 0, /* XCLASS - variable length */ \
449 3, /* REF */ \
450 1+LINK_SIZE, /* RECURSE */ \
451 2, /* CALLOUT */ \
452 1+LINK_SIZE, /* Alt */ \
453 1+LINK_SIZE, /* Ket */ \
454 1+LINK_SIZE, /* KetRmax */ \
455 1+LINK_SIZE, /* KetRmin */ \
456 1+LINK_SIZE, /* Assert */ \
457 1+LINK_SIZE, /* Assert not */ \
458 1+LINK_SIZE, /* Assert behind */ \
459 1+LINK_SIZE, /* Assert behind not */ \
460 1+LINK_SIZE, /* Reverse */ \
461 1+LINK_SIZE, /* Once */ \
462 1+LINK_SIZE, /* COND */ \
463 3, /* CREF */ \
464 1, 1, /* BRAZERO, BRAMINZERO */ \
465 3, /* BRANUMBER */ \
466 1+LINK_SIZE /* BRA */ \
467
468
469 /* The highest extraction number before we have to start using additional
470 bytes. (Originally PCRE didn't have support for extraction counts highter than
471 this number.) The value is limited by the number of opcodes left after OP_BRA,
472 i.e. 255 - OP_BRA. We actually set it a bit lower to leave room for additional
473 opcodes. */
474
475 #define EXTRACT_BASIC_MAX 150
476
477 /* A magic value for OP_CREF to indicate the "in recursion" condition. */
478
479 #define CREF_RECURSE 0xffff
480
481 /* The texts of compile-time error messages are defined as macros here so that
482 they can be accessed by the POSIX wrapper and converted into error codes. Yes,
483 I could have used error codes in the first place, but didn't feel like changing
484 just to accommodate the POSIX wrapper. */
485
486 #define ERR1 "\\ at end of pattern"
487 #define ERR2 "\\c at end of pattern"
488 #define ERR3 "unrecognized character follows \\"
489 #define ERR4 "numbers out of order in {} quantifier"
490 #define ERR5 "number too big in {} quantifier"
491 #define ERR6 "missing terminating ] for character class"
492 #define ERR7 "invalid escape sequence in character class"
493 #define ERR8 "range out of order in character class"
494 #define ERR9 "nothing to repeat"
495 #define ERR10 "operand of unlimited repeat could match the empty string"
496 #define ERR11 "internal error: unexpected repeat"
497 #define ERR12 "unrecognized character after (?"
498 #define ERR13 "POSIX named classes are supported only within a class"
499 #define ERR14 "missing )"
500 #define ERR15 "reference to non-existent subpattern"
501 #define ERR16 "erroffset passed as NULL"
502 #define ERR17 "unknown option bit(s) set"
503 #define ERR18 "missing ) after comment"
504 #define ERR19 "parentheses nested too deeply"
505 #define ERR20 "regular expression too large"
506 #define ERR21 "failed to get memory"
507 #define ERR22 "unmatched parentheses"
508 #define ERR23 "internal error: code overflow"
509 #define ERR24 "unrecognized character after (?<"
510 #define ERR25 "lookbehind assertion is not fixed length"
511 #define ERR26 "malformed number after (?("
512 #define ERR27 "conditional group contains more than two branches"
513 #define ERR28 "assertion expected after (?("
514 #define ERR29 "(?R or (?digits must be followed by )"
515 #define ERR30 "unknown POSIX class name"
516 #define ERR31 "POSIX collating elements are not supported"
517 #define ERR32 "this version of PCRE is not compiled with PCRE_UTF8 support"
518 #define ERR33 "spare error"
519 #define ERR34 "character value in \\x{...} sequence is too large"
520 #define ERR35 "invalid condition (?(0)"
521 #define ERR36 "\\C not allowed in lookbehind assertion"
522 #define ERR37 "PCRE does not support \\L, \\l, \\N, \\P, \\p, \\U, \\u, or \\X"
523 #define ERR38 "number after (?C is > 255"
524 #define ERR39 "closing ) for (?C expected"
525 #define ERR40 "recursive call could loop indefinitely"
526 #define ERR41 "unrecognized character after (?P"
527 #define ERR42 "syntax error after (?P"
528 #define ERR43 "two named groups have the same name"
529 #define ERR44 "invalid UTF-8 string"
530
531 /* All character handling must be done as unsigned characters. Otherwise there
532 are problems with top-bit-set characters and functions such as isspace().
533 However, we leave the interface to the outside world as char *, because that
534 should make things easier for callers. We define a short type for unsigned char
535 to save lots of typing. I tried "uchar", but it causes problems on Digital
536 Unix, where it is defined in sys/types, so use "uschar" instead. */
537
538 typedef unsigned char uschar;
539
540 /* The real format of the start of the pcre block; the index of names and the
541 code vector run on as long as necessary after the end. */
542
543 typedef struct real_pcre {
544 unsigned long int magic_number;
545 size_t size; /* Total that was malloced */
546 const unsigned char *tables; /* Pointer to tables */
547 unsigned long int options;
548 unsigned short int top_bracket;
549 unsigned short int top_backref;
550 unsigned short int first_byte;
551 unsigned short int req_byte;
552 unsigned short int name_entry_size; /* Size of any name items; 0 => none */
553 unsigned short int name_count; /* Number of name items */
554 } real_pcre;
555
556 /* The format of the block used to store data from pcre_study(). */
557
558 typedef struct pcre_study_data {
559 size_t size; /* Total that was malloced */
560 uschar options;
561 uschar start_bits[32];
562 } pcre_study_data;
563
564 /* Structure for passing "static" information around between the functions
565 doing the compiling, so that they are thread-safe. */
566
567 typedef struct compile_data {
568 const uschar *lcc; /* Points to lower casing table */
569 const uschar *fcc; /* Points to case-flipping table */
570 const uschar *cbits; /* Points to character type table */
571 const uschar *ctypes; /* Points to table of type maps */
572 const uschar *start_code; /* The start of the compiled code */
573 uschar *name_table; /* The name/number table */
574 int names_found; /* Number of entries so far */
575 int name_entry_size; /* Size of each entry */
576 int top_backref; /* Maximum back reference */
577 unsigned int backref_map; /* Bitmap of low back refs */
578 int req_varyopt; /* "After variable item" flag for reqbyte */
579 } compile_data;
580
581 /* Structure for maintaining a chain of pointers to the currently incomplete
582 branches, for testing for left recursion. */
583
584 typedef struct branch_chain {
585 struct branch_chain *outer;
586 uschar *current;
587 } branch_chain;
588
589 /* Structure for items in a linked list that represents an explicit recursive
590 call within the pattern. */
591
592 typedef struct recursion_info {
593 struct recursion_info *prev; /* Previous recursion record (or NULL) */
594 int group_num; /* Number of group that was called */
595 const uschar *after_call; /* "Return value": points after the call in the expr */
596 const uschar *save_start; /* Old value of md->start_match */
597 int *offset_save; /* Pointer to start of saved offsets */
598 int saved_max; /* Number of saved offsets */
599 } recursion_info;
600
601 /* Structure for passing "static" information around between the functions
602 doing the matching, so that they are thread-safe. */
603
604 typedef struct match_data {
605 unsigned long int match_call_count; /* As it says */
606 unsigned long int match_limit;/* As it says */
607 int *offset_vector; /* Offset vector */
608 int offset_end; /* One past the end */
609 int offset_max; /* The maximum usable for return data */
610 const uschar *lcc; /* Points to lower casing table */
611 const uschar *ctypes; /* Points to table of type maps */
612 BOOL offset_overflow; /* Set if too many extractions */
613 BOOL notbol; /* NOTBOL flag */
614 BOOL noteol; /* NOTEOL flag */
615 BOOL utf8; /* UTF8 flag */
616 BOOL endonly; /* Dollar not before final \n */
617 BOOL notempty; /* Empty string match not wanted */
618 const uschar *start_code; /* For use when recursing */
619 const uschar *start_subject; /* Start of the subject string */
620 const uschar *end_subject; /* End of the subject string */
621 const uschar *start_match; /* Start of this match attempt */
622 const uschar *end_match_ptr; /* Subject position at end match */
623 int end_offset_top; /* Highwater mark at end of match */
624 int capture_last; /* Most recent capture number */
625 int start_offset; /* The start offset value */
626 recursion_info *recursive; /* Linked list of recursion data */
627 void *callout_data; /* To pass back to callouts */
628 } match_data;
629
630 /* Bit definitions for entries in the pcre_ctypes table. */
631
632 #define ctype_space 0x01
633 #define ctype_letter 0x02
634 #define ctype_digit 0x04
635 #define ctype_xdigit 0x08
636 #define ctype_word 0x10 /* alphameric or '_' */
637 #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
638
639 /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
640 of bits for a class map. Some classes are built by combining these tables. */
641
642 #define cbit_space 0 /* [:space:] or \s */
643 #define cbit_xdigit 32 /* [:xdigit:] */
644 #define cbit_digit 64 /* [:digit:] or \d */
645 #define cbit_upper 96 /* [:upper:] */
646 #define cbit_lower 128 /* [:lower:] */
647 #define cbit_word 160 /* [:word:] or \w */
648 #define cbit_graph 192 /* [:graph:] */
649 #define cbit_print 224 /* [:print:] */
650 #define cbit_punct 256 /* [:punct:] */
651 #define cbit_cntrl 288 /* [:cntrl:] */
652 #define cbit_length 320 /* Length of the cbits table */
653
654 /* Offsets of the various tables from the base tables pointer, and
655 total length. */
656
657 #define lcc_offset 0
658 #define fcc_offset 256
659 #define cbits_offset 512
660 #define ctypes_offset (cbits_offset + cbit_length)
661 #define tables_length (ctypes_offset + 256)
662
663 /* End of internal.h */

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12