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

Contents of /code/trunk/pcre_internal.h

Parent Directory Parent Directory | Revision Log Revision Log


Revision 101 - (show annotations) (download)
Tue Mar 6 15:19:44 2007 UTC (7 years, 5 months ago) by ph10
File MIME type: text/plain
File size: 42824 byte(s)
Updated the support (such as it is) for Virtual Pascal, thanks to Stefan 
Weber: (1) pcre_internal.h was missing some function renames; (2) updated 
makevp.bat for the current PCRE, using the additional files !compile.txt, 
!linklib.txt, and pcregexp.pas.

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) 1997-2006 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). This assumes
204 that all calls to memmove are moving strings upwards in store, which is the
205 case in PCRE. */
206
207 #if ! HAVE_MEMMOVE
208 #undef memmove /* some systems may have a macro */
209 #if HAVE_BCOPY
210 #define memmove(a, b, c) bcopy(b, a, c)
211 #else /* HAVE_BCOPY */
212 static void *
213 pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)
214 {
215 size_t i;
216 dest += n;
217 src += n;
218 for (i = 0; i < n; ++i) *(--dest) = *(--src);
219 return dest;
220 }
221 #define memmove(a, b, c) pcre_memmove(a, b, c)
222 #endif /* not HAVE_BCOPY */
223 #endif /* not HAVE_MEMMOVE */
224 #endif /* not VPCOMPAT */
225
226
227 /* PCRE keeps offsets in its compiled code as 2-byte quantities (always stored
228 in big-endian order) by default. These are used, for example, to link from the
229 start of a subpattern to its alternatives and its end. The use of 2 bytes per
230 offset limits the size of the compiled regex to around 64K, which is big enough
231 for almost everybody. However, I received a request for an even bigger limit.
232 For this reason, and also to make the code easier to maintain, the storing and
233 loading of offsets from the byte string is now handled by the macros that are
234 defined here.
235
236 The macros are controlled by the value of LINK_SIZE. This defaults to 2 in
237 the config.h file, but can be overridden by using -D on the command line. This
238 is automated on Unix systems via the "configure" command. */
239
240 #if LINK_SIZE == 2
241
242 #define PUT(a,n,d) \
243 (a[n] = (d) >> 8), \
244 (a[(n)+1] = (d) & 255)
245
246 #define GET(a,n) \
247 (((a)[n] << 8) | (a)[(n)+1])
248
249 #define MAX_PATTERN_SIZE (1 << 16)
250
251
252 #elif LINK_SIZE == 3
253
254 #define PUT(a,n,d) \
255 (a[n] = (d) >> 16), \
256 (a[(n)+1] = (d) >> 8), \
257 (a[(n)+2] = (d) & 255)
258
259 #define GET(a,n) \
260 (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2])
261
262 #define MAX_PATTERN_SIZE (1 << 24)
263
264
265 #elif LINK_SIZE == 4
266
267 #define PUT(a,n,d) \
268 (a[n] = (d) >> 24), \
269 (a[(n)+1] = (d) >> 16), \
270 (a[(n)+2] = (d) >> 8), \
271 (a[(n)+3] = (d) & 255)
272
273 #define GET(a,n) \
274 (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
275
276 #define MAX_PATTERN_SIZE (1 << 30) /* Keep it positive */
277
278
279 #else
280 #error LINK_SIZE must be either 2, 3, or 4
281 #endif
282
283
284 /* Convenience macro defined in terms of the others */
285
286 #define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE
287
288
289 /* PCRE uses some other 2-byte quantities that do not change when the size of
290 offsets changes. There are used for repeat counts and for other things such as
291 capturing parenthesis numbers in back references. */
292
293 #define PUT2(a,n,d) \
294 a[n] = (d) >> 8; \
295 a[(n)+1] = (d) & 255
296
297 #define GET2(a,n) \
298 (((a)[n] << 8) | (a)[(n)+1])
299
300 #define PUT2INC(a,n,d) PUT2(a,n,d), a += 2
301
302
303 /* When UTF-8 encoding is being used, a character is no longer just a single
304 byte. The macros for character handling generate simple sequences when used in
305 byte-mode, and more complicated ones for UTF-8 characters. */
306
307 #ifndef SUPPORT_UTF8
308 #define GETCHAR(c, eptr) c = *eptr;
309 #define GETCHARTEST(c, eptr) c = *eptr;
310 #define GETCHARINC(c, eptr) c = *eptr++;
311 #define GETCHARINCTEST(c, eptr) c = *eptr++;
312 #define GETCHARLEN(c, eptr, len) c = *eptr;
313 #define BACKCHAR(eptr)
314
315 #else /* SUPPORT_UTF8 */
316
317 /* Get the next UTF-8 character, not advancing the pointer. This is called when
318 we know we are in UTF-8 mode. */
319
320 #define GETCHAR(c, eptr) \
321 c = *eptr; \
322 if (c >= 0xc0) \
323 { \
324 int gcii; \
325 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
326 int gcss = 6*gcaa; \
327 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
328 for (gcii = 1; gcii <= gcaa; gcii++) \
329 { \
330 gcss -= 6; \
331 c |= (eptr[gcii] & 0x3f) << gcss; \
332 } \
333 }
334
335 /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
336 pointer. */
337
338 #define GETCHARTEST(c, eptr) \
339 c = *eptr; \
340 if (utf8 && c >= 0xc0) \
341 { \
342 int gcii; \
343 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
344 int gcss = 6*gcaa; \
345 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
346 for (gcii = 1; gcii <= gcaa; gcii++) \
347 { \
348 gcss -= 6; \
349 c |= (eptr[gcii] & 0x3f) << gcss; \
350 } \
351 }
352
353 /* Get the next UTF-8 character, advancing the pointer. This is called when we
354 know we are in UTF-8 mode. */
355
356 #define GETCHARINC(c, eptr) \
357 c = *eptr++; \
358 if (c >= 0xc0) \
359 { \
360 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
361 int gcss = 6*gcaa; \
362 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
363 while (gcaa-- > 0) \
364 { \
365 gcss -= 6; \
366 c |= (*eptr++ & 0x3f) << gcss; \
367 } \
368 }
369
370 /* Get the next character, testing for UTF-8 mode, and advancing the pointer */
371
372 #define GETCHARINCTEST(c, eptr) \
373 c = *eptr++; \
374 if (utf8 && c >= 0xc0) \
375 { \
376 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
377 int gcss = 6*gcaa; \
378 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
379 while (gcaa-- > 0) \
380 { \
381 gcss -= 6; \
382 c |= (*eptr++ & 0x3f) << gcss; \
383 } \
384 }
385
386 /* Get the next UTF-8 character, not advancing the pointer, incrementing length
387 if there are extra bytes. This is called when we know we are in UTF-8 mode. */
388
389 #define GETCHARLEN(c, eptr, len) \
390 c = *eptr; \
391 if (c >= 0xc0) \
392 { \
393 int gcii; \
394 int gcaa = _pcre_utf8_table4[c & 0x3f]; /* Number of additional bytes */ \
395 int gcss = 6*gcaa; \
396 c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
397 for (gcii = 1; gcii <= gcaa; gcii++) \
398 { \
399 gcss -= 6; \
400 c |= (eptr[gcii] & 0x3f) << gcss; \
401 } \
402 len += gcaa; \
403 }
404
405 /* If the pointer is not at the start of a character, move it back until
406 it is. Called only in UTF-8 mode. */
407
408 #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;
409
410 #endif
411
412
413 /* In case there is no definition of offsetof() provided - though any proper
414 Standard C system should have one. */
415
416 #ifndef offsetof
417 #define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
418 #endif
419
420
421 /* These are the public options that can change during matching. */
422
423 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
424
425 /* Private options flags start at the most significant end of the four bytes.
426 The public options defined in pcre.h start at the least significant end. Make
427 sure they don't overlap! The bits are getting a bit scarce now -- when we run
428 out, there is a dummy word in the structure that could be used for the private
429 bits. */
430
431 #define PCRE_NOPARTIAL 0x80000000 /* can't use partial with this regex */
432 #define PCRE_FIRSTSET 0x40000000 /* first_byte is set */
433 #define PCRE_REQCHSET 0x20000000 /* req_byte is set */
434 #define PCRE_STARTLINE 0x10000000 /* start after \n for multiline */
435 #define PCRE_JCHANGED 0x08000000 /* j option changes within regex */
436
437 /* Options for the "extra" block produced by pcre_study(). */
438
439 #define PCRE_STUDY_MAPPED 0x01 /* a map of starting chars exists */
440
441 /* Masks for identifying the public options that are permitted at compile
442 time, run time, or study time, respectively. */
443
444 #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY)
445
446 #define PUBLIC_OPTIONS \
447 (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
448 PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
449 PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
450 PCRE_DUPNAMES|PCRE_NEWLINE_BITS)
451
452 #define PUBLIC_EXEC_OPTIONS \
453 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
454 PCRE_PARTIAL|PCRE_NEWLINE_BITS)
455
456 #define PUBLIC_DFA_EXEC_OPTIONS \
457 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
458 PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)
459
460 #define PUBLIC_STUDY_OPTIONS 0 /* None defined */
461
462 /* Magic number to provide a small check against being handed junk. Also used
463 to detect whether a pattern was compiled on a host of different endianness. */
464
465 #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
466
467 /* Negative values for the firstchar and reqchar variables */
468
469 #define REQ_UNSET (-2)
470 #define REQ_NONE (-1)
471
472 /* The maximum remaining length of subject we are prepared to search for a
473 req_byte match. */
474
475 #define REQ_BYTE_MAX 1000
476
477 /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a
478 variable-length repeat, or a anything other than literal characters. */
479
480 #define REQ_CASELESS 0x0100 /* indicates caselessness */
481 #define REQ_VARY 0x0200 /* reqbyte followed non-literal item */
482
483 /* Miscellaneous definitions */
484
485 typedef int BOOL;
486
487 #define FALSE 0
488 #define TRUE 1
489
490 /* Escape items that are just an encoding of a particular data value. */
491
492 #ifndef ESC_e
493 #define ESC_e 27
494 #endif
495
496 #ifndef ESC_f
497 #define ESC_f '\f'
498 #endif
499
500 #ifndef ESC_n
501 #define ESC_n '\n'
502 #endif
503
504 #ifndef ESC_r
505 #define ESC_r '\r'
506 #endif
507
508 /* We can't officially use ESC_t because it is a POSIX reserved identifier
509 (presumably because of all the others like size_t). */
510
511 #ifndef ESC_tee
512 #define ESC_tee '\t'
513 #endif
514
515 /* Codes for different types of Unicode property */
516
517 #define PT_ANY 0 /* Any property - matches all chars */
518 #define PT_LAMP 1 /* L& - the union of Lu, Ll, Lt */
519 #define PT_GC 2 /* General characteristic (e.g. L) */
520 #define PT_PC 3 /* Particular characteristic (e.g. Lu) */
521 #define PT_SC 4 /* Script (e.g. Han) */
522
523 /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
524 contain UTF-8 characters with values greater than 255. */
525
526 #define XCL_NOT 0x01 /* Flag: this is a negative class */
527 #define XCL_MAP 0x02 /* Flag: a 32-byte map is present */
528
529 #define XCL_END 0 /* Marks end of individual items */
530 #define XCL_SINGLE 1 /* Single item (one multibyte char) follows */
531 #define XCL_RANGE 2 /* A range (two multibyte chars) follows */
532 #define XCL_PROP 3 /* Unicode property (2-byte property code follows) */
533 #define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */
534
535 /* These are escaped items that aren't just an encoding of a particular data
536 value such as \n. They must have non-zero values, as check_escape() returns
537 their negation. Also, they must appear in the same order as in the opcode
538 definitions below, up to ESC_z. There's a dummy for OP_ANY because it
539 corresponds to "." rather than an escape sequence. The final one must be
540 ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).
541 There are two tests in the code for an escape greater than ESC_b and less than
542 ESC_Z to detect the types that may be repeated. These are the types that
543 consume characters. If any new escapes are put in between that don't consume a
544 character, that code will have to change. */
545
546 enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,
547 ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_X, ESC_Z, ESC_z,
548 ESC_E, ESC_Q, ESC_k, ESC_REF };
549
550
551 /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
552 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
553 OP_EOD must correspond in order to the list of escapes immediately above.
554
555 To keep stored, compiled patterns compatible, new opcodes should be added
556 immediately before OP_BRA, where (since release 7.0) a gap is left for this
557 purpose.
558
559 *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
560 that follow must also be updated to match. There is also a table called
561 "coptable" in pcre_dfa_exec.c that must be updated. */
562
563 enum {
564 OP_END, /* 0 End of pattern */
565
566 /* Values corresponding to backslashed metacharacters */
567
568 OP_SOD, /* 1 Start of data: \A */
569 OP_SOM, /* 2 Start of match (subject + offset): \G */
570 OP_NOT_WORD_BOUNDARY, /* 3 \B */
571 OP_WORD_BOUNDARY, /* 4 \b */
572 OP_NOT_DIGIT, /* 5 \D */
573 OP_DIGIT, /* 6 \d */
574 OP_NOT_WHITESPACE, /* 7 \S */
575 OP_WHITESPACE, /* 8 \s */
576 OP_NOT_WORDCHAR, /* 9 \W */
577 OP_WORDCHAR, /* 10 \w */
578 OP_ANY, /* 11 Match any character */
579 OP_ANYBYTE, /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */
580 OP_NOTPROP, /* 13 \P (not Unicode property) */
581 OP_PROP, /* 14 \p (Unicode property) */
582 OP_ANYNL, /* 15 \R (any newline sequence) */
583 OP_EXTUNI, /* 16 \X (extended Unicode sequence */
584 OP_EODN, /* 17 End of data or \n at end of data: \Z. */
585 OP_EOD, /* 18 End of data: \z */
586
587 OP_OPT, /* 19 Set runtime options */
588 OP_CIRC, /* 20 Start of line - varies with multiline switch */
589 OP_DOLL, /* 21 End of line - varies with multiline switch */
590 OP_CHAR, /* 22 Match one character, casefully */
591 OP_CHARNC, /* 23 Match one character, caselessly */
592 OP_NOT, /* 24 Match one character, not the following one */
593
594 OP_STAR, /* 25 The maximizing and minimizing versions of */
595 OP_MINSTAR, /* 26 these six opcodes must come in pairs, with */
596 OP_PLUS, /* 27 the minimizing one second. */
597 OP_MINPLUS, /* 28 This first set applies to single characters.*/
598 OP_QUERY, /* 29 */
599 OP_MINQUERY, /* 30 */
600
601 OP_UPTO, /* 31 From 0 to n matches */
602 OP_MINUPTO, /* 32 */
603 OP_EXACT, /* 33 Exactly n matches */
604
605 OP_POSSTAR, /* 34 Possessified star */
606 OP_POSPLUS, /* 35 Possessified plus */
607 OP_POSQUERY, /* 36 Posesssified query */
608 OP_POSUPTO, /* 37 Possessified upto */
609
610 OP_NOTSTAR, /* 38 The maximizing and minimizing versions of */
611 OP_NOTMINSTAR, /* 39 these six opcodes must come in pairs, with */
612 OP_NOTPLUS, /* 40 the minimizing one second. They must be in */
613 OP_NOTMINPLUS, /* 41 exactly the same order as those above. */
614 OP_NOTQUERY, /* 42 This set applies to "not" single characters. */
615 OP_NOTMINQUERY, /* 43 */
616
617 OP_NOTUPTO, /* 44 From 0 to n matches */
618 OP_NOTMINUPTO, /* 45 */
619 OP_NOTEXACT, /* 46 Exactly n matches */
620
621 OP_NOTPOSSTAR, /* 47 Possessified versions */
622 OP_NOTPOSPLUS, /* 48 */
623 OP_NOTPOSQUERY, /* 49 */
624 OP_NOTPOSUPTO, /* 50 */
625
626 OP_TYPESTAR, /* 51 The maximizing and minimizing versions of */
627 OP_TYPEMINSTAR, /* 52 these six opcodes must come in pairs, with */
628 OP_TYPEPLUS, /* 53 the minimizing one second. These codes must */
629 OP_TYPEMINPLUS, /* 54 be in exactly the same order as those above. */
630 OP_TYPEQUERY, /* 55 This set applies to character types such as \d */
631 OP_TYPEMINQUERY, /* 56 */
632
633 OP_TYPEUPTO, /* 57 From 0 to n matches */
634 OP_TYPEMINUPTO, /* 58 */
635 OP_TYPEEXACT, /* 59 Exactly n matches */
636
637 OP_TYPEPOSSTAR, /* 60 Possessified versions */
638 OP_TYPEPOSPLUS, /* 61 */
639 OP_TYPEPOSQUERY, /* 62 */
640 OP_TYPEPOSUPTO, /* 63 */
641
642 OP_CRSTAR, /* 64 The maximizing and minimizing versions of */
643 OP_CRMINSTAR, /* 65 all these opcodes must come in pairs, with */
644 OP_CRPLUS, /* 66 the minimizing one second. These codes must */
645 OP_CRMINPLUS, /* 67 be in exactly the same order as those above. */
646 OP_CRQUERY, /* 68 These are for character classes and back refs */
647 OP_CRMINQUERY, /* 69 */
648 OP_CRRANGE, /* 70 These are different to the three sets above. */
649 OP_CRMINRANGE, /* 71 */
650
651 OP_CLASS, /* 72 Match a character class, chars < 256 only */
652 OP_NCLASS, /* 73 Same, but the bitmap was created from a negative
653 class - the difference is relevant only when a UTF-8
654 character > 255 is encountered. */
655
656 OP_XCLASS, /* 74 Extended class for handling UTF-8 chars within the
657 class. This does both positive and negative. */
658
659 OP_REF, /* 75 Match a back reference */
660 OP_RECURSE, /* 76 Match a numbered subpattern (possibly recursive) */
661 OP_CALLOUT, /* 77 Call out to external function if provided */
662
663 OP_ALT, /* 78 Start of alternation */
664 OP_KET, /* 79 End of group that doesn't have an unbounded repeat */
665 OP_KETRMAX, /* 80 These two must remain together and in this */
666 OP_KETRMIN, /* 81 order. They are for groups the repeat for ever. */
667
668 /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
669
670 OP_ASSERT, /* 82 Positive lookahead */
671 OP_ASSERT_NOT, /* 83 Negative lookahead */
672 OP_ASSERTBACK, /* 84 Positive lookbehind */
673 OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */
674 OP_REVERSE, /* 86 Move pointer back - used in lookbehind assertions */
675
676 /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
677 as there's a test for >= ONCE for a subpattern that isn't an assertion. */
678
679 OP_ONCE, /* 87 Atomic group */
680 OP_BRA, /* 88 Start of non-capturing bracket */
681 OP_CBRA, /* 89 Start of capturing bracket */
682 OP_COND, /* 90 Conditional group */
683
684 /* These three must follow the previous three, in the same order. There's a
685 check for >= SBRA to distinguish the two sets. */
686
687 OP_SBRA, /* 91 Start of non-capturing bracket, check empty */
688 OP_SCBRA, /* 92 Start of capturing bracket, check empty */
689 OP_SCOND, /* 93 Conditional group, check empty */
690
691 OP_CREF, /* 94 Used to hold a capture number as condition */
692 OP_RREF, /* 95 Used to hold a recursion number as condition */
693 OP_DEF, /* 96 The DEFINE condition */
694
695 OP_BRAZERO, /* 97 These two must remain together and in this */
696 OP_BRAMINZERO /* 98 order. */
697 };
698
699
700 /* This macro defines textual names for all the opcodes. These are used only
701 for debugging. The macro is referenced only in pcre_printint.c. */
702
703 #define OP_NAME_LIST \
704 "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d", \
705 "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte", \
706 "notprop", "prop", "anynl", "extuni", \
707 "\\Z", "\\z", \
708 "Opt", "^", "$", "char", "charnc", "not", \
709 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
710 "*+","++", "?+", "{", \
711 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
712 "*+","++", "?+", "{", \
713 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
714 "*+","++", "?+", "{", \
715 "*", "*?", "+", "+?", "?", "??", "{", "{", \
716 "class", "nclass", "xclass", "Ref", "Recurse", "Callout", \
717 "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not", \
718 "AssertB", "AssertB not", "Reverse", \
719 "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond", \
720 "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"
721
722
723 /* This macro defines the length of fixed length operations in the compiled
724 regex. The lengths are used when searching for specific things, and also in the
725 debugging printing of a compiled regex. We use a macro so that it can be
726 defined close to the definitions of the opcodes themselves.
727
728 As things have been extended, some of these are no longer fixed lenths, but are
729 minima instead. For example, the length of a single-character repeat may vary
730 in UTF-8 mode. The code that uses this table must know about such things. */
731
732 #define OP_LENGTHS \
733 1, /* End */ \
734 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \
735 1, 1, /* Any, Anybyte */ \
736 3, 3, 1, 1, /* NOTPROP, PROP, EXTUNI, ANYNL */ \
737 1, 1, 2, 1, 1, /* \Z, \z, Opt, ^, $ */ \
738 2, /* Char - the minimum length */ \
739 2, /* Charnc - the minimum length */ \
740 2, /* not */ \
741 /* Positive single-char repeats ** These are */ \
742 2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** minima in */ \
743 4, 4, 4, /* upto, minupto, exact ** UTF-8 mode */ \
744 2, 2, 2, 4, /* *+, ++, ?+, upto+ */ \
745 /* Negative single-char repeats - only for chars < 256 */ \
746 2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \
747 4, 4, 4, /* NOT upto, minupto, exact */ \
748 2, 2, 2, 4, /* Possessive *, +, ?, upto */ \
749 /* Positive type repeats */ \
750 2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \
751 4, 4, 4, /* Type upto, minupto, exact */ \
752 2, 2, 2, 4, /* Possessive *+, ++, ?+, upto+ */ \
753 /* Character class & ref repeats */ \
754 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \
755 5, 5, /* CRRANGE, CRMINRANGE */ \
756 33, /* CLASS */ \
757 33, /* NCLASS */ \
758 0, /* XCLASS - variable length */ \
759 3, /* REF */ \
760 1+LINK_SIZE, /* RECURSE */ \
761 2+2*LINK_SIZE, /* CALLOUT */ \
762 1+LINK_SIZE, /* Alt */ \
763 1+LINK_SIZE, /* Ket */ \
764 1+LINK_SIZE, /* KetRmax */ \
765 1+LINK_SIZE, /* KetRmin */ \
766 1+LINK_SIZE, /* Assert */ \
767 1+LINK_SIZE, /* Assert not */ \
768 1+LINK_SIZE, /* Assert behind */ \
769 1+LINK_SIZE, /* Assert behind not */ \
770 1+LINK_SIZE, /* Reverse */ \
771 1+LINK_SIZE, /* ONCE */ \
772 1+LINK_SIZE, /* BRA */ \
773 3+LINK_SIZE, /* CBRA */ \
774 1+LINK_SIZE, /* COND */ \
775 1+LINK_SIZE, /* SBRA */ \
776 3+LINK_SIZE, /* SCBRA */ \
777 1+LINK_SIZE, /* SCOND */ \
778 3, /* CREF */ \
779 3, /* RREF */ \
780 1, /* DEF */ \
781 1, 1, /* BRAZERO, BRAMINZERO */ \
782
783
784 /* A magic value for OP_RREF to indicate the "any recursion" condition. */
785
786 #define RREF_ANY 0xffff
787
788 /* Error code numbers. They are given names so that they can more easily be
789 tracked. */
790
791 enum { ERR0, ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9,
792 ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
793 ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
794 ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
795 ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
796 ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57 };
797
798 /* The real format of the start of the pcre block; the index of names and the
799 code vector run on as long as necessary after the end. We store an explicit
800 offset to the name table so that if a regex is compiled on one host, saved, and
801 then run on another where the size of pointers is different, all might still
802 be well. For the case of compiled-on-4 and run-on-8, we include an extra
803 pointer that is always NULL. For future-proofing, a few dummy fields were
804 originally included - even though you can never get this planning right - but
805 there is only one left now.
806
807 NOTE NOTE NOTE:
808 Because people can now save and re-use compiled patterns, any additions to this
809 structure should be made at the end, and something earlier (e.g. a new
810 flag in the options or one of the dummy fields) should indicate that the new
811 fields are present. Currently PCRE always sets the dummy fields to zero.
812 NOTE NOTE NOTE:
813 */
814
815 typedef struct real_pcre {
816 pcre_uint32 magic_number;
817 pcre_uint32 size; /* Total that was malloced */
818 pcre_uint32 options;
819 pcre_uint32 dummy1; /* For future use, maybe */
820
821 pcre_uint16 top_bracket;
822 pcre_uint16 top_backref;
823 pcre_uint16 first_byte;
824 pcre_uint16 req_byte;
825 pcre_uint16 name_table_offset; /* Offset to name table that follows */
826 pcre_uint16 name_entry_size; /* Size of any name items */
827 pcre_uint16 name_count; /* Number of name items */
828 pcre_uint16 ref_count; /* Reference count */
829
830 const unsigned char *tables; /* Pointer to tables or NULL for std */
831 const unsigned char *nullpad; /* NULL padding */
832 } real_pcre;
833
834 /* The format of the block used to store data from pcre_study(). The same
835 remark (see NOTE above) about extending this structure applies. */
836
837 typedef struct pcre_study_data {
838 pcre_uint32 size; /* Total that was malloced */
839 pcre_uint32 options;
840 uschar start_bits[32];
841 } pcre_study_data;
842
843 /* Structure for passing "static" information around between the functions
844 doing the compiling, so that they are thread-safe. */
845
846 typedef struct compile_data {
847 const uschar *lcc; /* Points to lower casing table */
848 const uschar *fcc; /* Points to case-flipping table */
849 const uschar *cbits; /* Points to character type table */
850 const uschar *ctypes; /* Points to table of type maps */
851 const uschar *start_workspace;/* The start of working space */
852 const uschar *start_code; /* The start of the compiled code */
853 const uschar *start_pattern; /* The start of the pattern */
854 const uschar *end_pattern; /* The end of the pattern */
855 uschar *hwm; /* High watermark of workspace */
856 uschar *name_table; /* The name/number table */
857 int names_found; /* Number of entries so far */
858 int name_entry_size; /* Size of each entry */
859 int bracount; /* Count of capturing parens */
860 int top_backref; /* Maximum back reference */
861 unsigned int backref_map; /* Bitmap of low back refs */
862 int external_options; /* External (initial) options */
863 int req_varyopt; /* "After variable item" flag for reqbyte */
864 BOOL nopartial; /* Set TRUE if partial won't work */
865 int nltype; /* Newline type */
866 int nllen; /* Newline string length */
867 uschar nl[4]; /* Newline string when fixed length */
868 } compile_data;
869
870 /* Structure for maintaining a chain of pointers to the currently incomplete
871 branches, for testing for left recursion. */
872
873 typedef struct branch_chain {
874 struct branch_chain *outer;
875 uschar *current;
876 } branch_chain;
877
878 /* Structure for items in a linked list that represents an explicit recursive
879 call within the pattern. */
880
881 typedef struct recursion_info {
882 struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
883 int group_num; /* Number of group that was called */
884 const uschar *after_call; /* "Return value": points after the call in the expr */
885 USPTR save_start; /* Old value of md->start_match */
886 int *offset_save; /* Pointer to start of saved offsets */
887 int saved_max; /* Number of saved offsets */
888 } recursion_info;
889
890 /* When compiling in a mode that doesn't use recursive calls to match(),
891 a structure is used to remember local variables on the heap. It is defined in
892 pcre_exec.c, close to the match() function, so that it is easy to keep it in
893 step with any changes of local variable. However, the pointer to the current
894 frame must be saved in some "static" place over a longjmp(). We declare the
895 structure here so that we can put a pointer in the match_data structure. NOTE:
896 This isn't used for a "normal" compilation of pcre. */
897
898 struct heapframe;
899
900 /* Structure for building a chain of data for holding the values of the subject
901 pointer at the start of each subpattern, so as to detect when an empty string
902 has been matched by a subpattern - to break infinite loops. */
903
904 typedef struct eptrblock {
905 struct eptrblock *epb_prev;
906 USPTR epb_saved_eptr;
907 } eptrblock;
908
909
910 /* Structure for passing "static" information around between the functions
911 doing traditional NFA matching, so that they are thread-safe. */
912
913 typedef struct match_data {
914 unsigned long int match_call_count; /* As it says */
915 unsigned long int match_limit; /* As it says */
916 unsigned long int match_limit_recursion; /* As it says */
917 int *offset_vector; /* Offset vector */
918 int offset_end; /* One past the end */
919 int offset_max; /* The maximum usable for return data */
920 int nltype; /* Newline type */
921 int nllen; /* Newline string length */
922 uschar nl[4]; /* Newline string when fixed */
923 const uschar *lcc; /* Points to lower casing table */
924 const uschar *ctypes; /* Points to table of type maps */
925 BOOL offset_overflow; /* Set if too many extractions */
926 BOOL notbol; /* NOTBOL flag */
927 BOOL noteol; /* NOTEOL flag */
928 BOOL utf8; /* UTF8 flag */
929 BOOL endonly; /* Dollar not before final \n */
930 BOOL notempty; /* Empty string match not wanted */
931 BOOL partial; /* PARTIAL flag */
932 BOOL hitend; /* Hit the end of the subject at some point */
933 const uschar *start_code; /* For use when recursing */
934 USPTR start_subject; /* Start of the subject string */
935 USPTR end_subject; /* End of the subject string */
936 USPTR start_match; /* Start of this match attempt */
937 USPTR end_match_ptr; /* Subject position at end match */
938 int end_offset_top; /* Highwater mark at end of match */
939 int capture_last; /* Most recent capture number */
940 int start_offset; /* The start offset value */
941 eptrblock *eptrchain; /* Chain of eptrblocks for tail recursions */
942 int eptrn; /* Next free eptrblock */
943 recursion_info *recursive; /* Linked list of recursion data */
944 void *callout_data; /* To pass back to callouts */
945 struct heapframe *thisframe; /* Used only when compiling for no recursion */
946 } match_data;
947
948 /* A similar structure is used for the same purpose by the DFA matching
949 functions. */
950
951 typedef struct dfa_match_data {
952 const uschar *start_code; /* Start of the compiled pattern */
953 const uschar *start_subject; /* Start of the subject string */
954 const uschar *end_subject; /* End of subject string */
955 const uschar *tables; /* Character tables */
956 int moptions; /* Match options */
957 int poptions; /* Pattern options */
958 int nltype; /* Newline type */
959 int nllen; /* Newline string length */
960 uschar nl[4]; /* Newline string when fixed */
961 void *callout_data; /* To pass back to callouts */
962 } dfa_match_data;
963
964 /* Bit definitions for entries in the pcre_ctypes table. */
965
966 #define ctype_space 0x01
967 #define ctype_letter 0x02
968 #define ctype_digit 0x04
969 #define ctype_xdigit 0x08
970 #define ctype_word 0x10 /* alphameric or '_' */
971 #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
972
973 /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
974 of bits for a class map. Some classes are built by combining these tables. */
975
976 #define cbit_space 0 /* [:space:] or \s */
977 #define cbit_xdigit 32 /* [:xdigit:] */
978 #define cbit_digit 64 /* [:digit:] or \d */
979 #define cbit_upper 96 /* [:upper:] */
980 #define cbit_lower 128 /* [:lower:] */
981 #define cbit_word 160 /* [:word:] or \w */
982 #define cbit_graph 192 /* [:graph:] */
983 #define cbit_print 224 /* [:print:] */
984 #define cbit_punct 256 /* [:punct:] */
985 #define cbit_cntrl 288 /* [:cntrl:] */
986 #define cbit_length 320 /* Length of the cbits table */
987
988 /* Offsets of the various tables from the base tables pointer, and
989 total length. */
990
991 #define lcc_offset 0
992 #define fcc_offset 256
993 #define cbits_offset 512
994 #define ctypes_offset (cbits_offset + cbit_length)
995 #define tables_length (ctypes_offset + 256)
996
997 /* Layout of the UCP type table that translates property names into types and
998 codes. */
999
1000 typedef struct {
1001 const char *name;
1002 pcre_uint16 type;
1003 pcre_uint16 value;
1004 } ucp_type_table;
1005
1006
1007 /* Internal shared data tables. These are tables that are used by more than one
1008 of the exported public functions. They have to be "external" in the C sense,
1009 but are not part of the PCRE public API. The data for these tables is in the
1010 pcre_tables.c module. */
1011
1012 extern const int _pcre_utf8_table1[];
1013 extern const int _pcre_utf8_table2[];
1014 extern const int _pcre_utf8_table3[];
1015 extern const uschar _pcre_utf8_table4[];
1016
1017 extern const int _pcre_utf8_table1_size;
1018
1019 extern const ucp_type_table _pcre_utt[];
1020 extern const int _pcre_utt_size;
1021
1022 extern const uschar _pcre_default_tables[];
1023
1024 extern const uschar _pcre_OP_lengths[];
1025
1026
1027 /* Internal shared functions. These are functions that are used by more than
1028 one of the exported public functions. They have to be "external" in the C
1029 sense, but are not part of the PCRE public API. */
1030
1031 extern BOOL _pcre_is_newline(const uschar *, const uschar *, int *,
1032 BOOL);
1033 extern int _pcre_ord2utf8(int, uschar *);
1034 extern real_pcre *_pcre_try_flipped(const real_pcre *, real_pcre *,
1035 const pcre_study_data *, pcre_study_data *);
1036 extern int _pcre_ucp_findprop(const unsigned int, int *, int *);
1037 extern unsigned int _pcre_ucp_othercase(const unsigned int);
1038 extern int _pcre_valid_utf8(const uschar *, int);
1039 extern BOOL _pcre_was_newline(const uschar *, const uschar *, int *,
1040 BOOL);
1041 extern BOOL _pcre_xclass(int, const uschar *);
1042
1043 #endif
1044
1045 /* 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