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) 1997-2014 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_", "_pcre16_" or "_pcre32_"
44 depending on the PRIV macro. */
45
46 #ifndef PCRE_INTERNAL_H
47 #define PCRE_INTERNAL_H
48
49 /* Define PCRE_DEBUG to get debugging output on stdout. */
50
51 #if 0
52 #define PCRE_DEBUG
53 #endif
54
55 /* PCRE is compiled as an 8 bit library if it is not requested otherwise. */
56
57 #if !defined COMPILE_PCRE16 && !defined COMPILE_PCRE32
58 #define COMPILE_PCRE8
59 #endif
60
61 /* If SUPPORT_UCP is defined, SUPPORT_UTF must also be defined. The
62 "configure" script ensures this, but not everybody uses "configure". */
63
64 #if defined SUPPORT_UCP && !(defined SUPPORT_UTF)
65 #define SUPPORT_UTF 1
66 #endif
67
68 /* We define SUPPORT_UTF if SUPPORT_UTF8 is enabled for compatibility
69 reasons with existing code. */
70
71 #if defined SUPPORT_UTF8 && !(defined SUPPORT_UTF)
72 #define SUPPORT_UTF 1
73 #endif
74
75 /* Fixme: SUPPORT_UTF8 should be eventually disappear from the code.
76 Until then we define it if SUPPORT_UTF is defined. */
77
78 #if defined SUPPORT_UTF && !(defined SUPPORT_UTF8)
79 #define SUPPORT_UTF8 1
80 #endif
81
82 /* We do not support both EBCDIC and UTF-8/16/32 at the same time. The "configure"
83 script prevents both being selected, but not everybody uses "configure". */
84
85 #if defined EBCDIC && defined SUPPORT_UTF
86 #error The use of both EBCDIC and SUPPORT_UTF is not supported.
87 #endif
88
89 /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef
90 inline, and there are *still* stupid compilers about that don't like indented
91 pre-processor statements, or at least there were when I first wrote this. After
92 all, it had only been about 10 years then...
93
94 It turns out that the Mac Debugging.h header also defines the macro DPRINTF, so
95 be absolutely sure we get our version. */
96
97 #undef DPRINTF
98 #ifdef PCRE_DEBUG
99 #define DPRINTF(p) printf p
100 #else
101 #define DPRINTF(p) /* Nothing */
102 #endif
103
104
105 /* Standard C headers plus the external interface definition. The only time
106 setjmp and stdarg are used is when NO_RECURSE is set. */
107
108 #include <ctype.h>
109 #include <limits.h>
110 #include <stddef.h>
111 #include <stdio.h>
112 #include <stdlib.h>
113 #include <string.h>
114
115 /* Valgrind (memcheck) support */
116
117 #ifdef SUPPORT_VALGRIND
118 #include <valgrind/memcheck.h>
119 #endif
120
121 /* When compiling a DLL for Windows, the exported symbols have to be declared
122 using some MS magic. I found some useful information on this web page:
123 http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
124 information there, using __declspec(dllexport) without "extern" we have a
125 definition; with "extern" we have a declaration. The settings here override the
126 setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
127 which is all that is needed for applications (they just import the symbols). We
128 use:
129
130 PCRE_EXP_DECL for declarations
131 PCRE_EXP_DEFN for definitions of exported functions
132 PCRE_EXP_DATA_DEFN for definitions of exported variables
133
134 The reason for the two DEFN macros is that in non-Windows environments, one
135 does not want to have "extern" before variable definitions because it leads to
136 compiler warnings. So we distinguish between functions and variables. In
137 Windows, the two should always be the same.
138
139 The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
140 which is an application, but needs to import this file in order to "peek" at
141 internals, can #include pcre.h first to get an application's-eye view.
142
143 In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
144 special-purpose environments) might want to stick other stuff in front of
145 exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
146 PCRE_EXP_DATA_DEFN only if they are not already set. */
147
148 #ifndef PCRE_EXP_DECL
149 # ifdef _WIN32
150 # ifndef PCRE_STATIC
151 # define PCRE_EXP_DECL extern __declspec(dllexport)
152 # define PCRE_EXP_DEFN __declspec(dllexport)
153 # define PCRE_EXP_DATA_DEFN __declspec(dllexport)
154 # else
155 # define PCRE_EXP_DECL extern
156 # define PCRE_EXP_DEFN
157 # define PCRE_EXP_DATA_DEFN
158 # endif
159 # else
160 # ifdef __cplusplus
161 # define PCRE_EXP_DECL extern "C"
162 # else
163 # define PCRE_EXP_DECL extern
164 # endif
165 # ifndef PCRE_EXP_DEFN
166 # define PCRE_EXP_DEFN PCRE_EXP_DECL
167 # endif
168 # ifndef PCRE_EXP_DATA_DEFN
169 # define PCRE_EXP_DATA_DEFN
170 # endif
171 # endif
172 #endif
173
174 /* When compiling with the MSVC compiler, it is sometimes necessary to include
175 a "calling convention" before exported function names. (This is secondhand
176 information; I know nothing about MSVC myself). For example, something like
177
178 void __cdecl function(....)
179
180 might be needed. In order so make this easy, all the exported functions have
181 PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
182 set, we ensure here that it has no effect. */
183
184 #ifndef PCRE_CALL_CONVENTION
185 #define PCRE_CALL_CONVENTION
186 #endif
187
188 /* We need to have types that specify unsigned 8, 16 and 32-bit integers. We
189 cannot determine these outside the compilation (e.g. by running a program as
190 part of "configure") because PCRE is often cross-compiled for use on other
191 systems. Instead we make use of the maximum sizes that are available at
192 preprocessor time in standard C environments. */
193
194 typedef unsigned char pcre_uint8;
195
196 #if USHRT_MAX == 65535
197 typedef unsigned short pcre_uint16;
198 typedef short pcre_int16;
199 #define PCRE_UINT16_MAX USHRT_MAX
200 #define PCRE_INT16_MAX SHRT_MAX
201 #elif UINT_MAX == 65535
202 typedef unsigned int pcre_uint16;
203 typedef int pcre_int16;
204 #define PCRE_UINT16_MAX UINT_MAX
205 #define PCRE_INT16_MAX INT_MAX
206 #else
207 #error Cannot determine a type for 16-bit integers
208 #endif
209
210 #if UINT_MAX == 4294967295U
211 typedef unsigned int pcre_uint32;
212 typedef int pcre_int32;
213 #define PCRE_UINT32_MAX UINT_MAX
214 #define PCRE_INT32_MAX INT_MAX
215 #elif ULONG_MAX == 4294967295UL
216 typedef unsigned long int pcre_uint32;
217 typedef long int pcre_int32;
218 #define PCRE_UINT32_MAX ULONG_MAX
219 #define PCRE_INT32_MAX LONG_MAX
220 #else
221 #error Cannot determine a type for 32-bit integers
222 #endif
223
224 /* When checking for integer overflow in pcre_compile(), we need to handle
225 large integers. If a 64-bit integer type is available, we can use that.
226 Otherwise we have to cast to double, which of course requires floating point
227 arithmetic. Handle this by defining a macro for the appropriate type. If
228 stdint.h is available, include it; it may define INT64_MAX. Systems that do not
229 have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
230 by "configure". */
231
232 #ifdef PHP_WIN32
233 #include "win32/php_stdint.h"
234 #else
235 #if defined HAVE_STDINT_H
236 #include <stdint.h>
237 #elif defined HAVE_INTTYPES_H
238 #include <inttypes.h>
239 #endif
240 #endif
241
242 #if defined INT64_MAX || defined int64_t
243 #define INT64_OR_DOUBLE int64_t
244 #else
245 #define INT64_OR_DOUBLE double
246 #endif
247
248 /* All character handling must be done as unsigned characters. Otherwise there
249 are problems with top-bit-set characters and functions such as isspace().
250 However, we leave the interface to the outside world as char * or short *,
251 because that should make things easier for callers. This character type is
252 called pcre_uchar.
253
254 The IN_UCHARS macro multiply its argument with the byte size of the current
255 pcre_uchar type. Useful for memcpy and such operations, whose require the
256 byte size of their input/output buffers.
257
258 The MAX_255 macro checks whether its pcre_uchar input is less than 256.
259
260 The TABLE_GET macro is designed for accessing elements of tables whose contain
261 exactly 256 items. When the character is able to contain more than 256
262 items, some check is needed before accessing these tables.
263 */
264
265 #if defined COMPILE_PCRE8
266
267 typedef unsigned char pcre_uchar;
268 #define IN_UCHARS(x) (x)
269 #define MAX_255(c) 1
270 #define TABLE_GET(c, table, default) ((table)[c])
271
272 #elif defined COMPILE_PCRE16
273
274 #if USHRT_MAX != 65535
275 /* This is a warning message. Change PCRE_UCHAR16 to a 16 bit data type in
276 pcre.h(.in) and disable (comment out) this message. */
277 #error Warning: PCRE_UCHAR16 is not a 16 bit data type.
278 #endif
279
280 typedef pcre_uint16 pcre_uchar;
281 #define UCHAR_SHIFT (1)
282 #define IN_UCHARS(x) ((x) << UCHAR_SHIFT)
283 #define MAX_255(c) ((c) <= 255u)
284 #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
285
286 #elif defined COMPILE_PCRE32
287
288 typedef pcre_uint32 pcre_uchar;
289 #define UCHAR_SHIFT (2)
290 #define IN_UCHARS(x) ((x) << UCHAR_SHIFT)
291 #define MAX_255(c) ((c) <= 255u)
292 #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
293
294 #else
295 #error Unsupported compiling mode
296 #endif /* COMPILE_PCRE[8|16|32] */
297
298 /* This is an unsigned int value that no character can ever have. UTF-8
299 characters only go up to 0x7fffffff (though Unicode doesn't go beyond
300 0x0010ffff). */
301
302 #define NOTACHAR 0xffffffff
303
304 /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
305 "any" and "anycrlf" at present). The following macros are used to package up
306 testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
307 modules to indicate in which datablock the parameters exist, and what the
308 start/end of string field names are. */
309
310 #define NLTYPE_FIXED 0 /* Newline is a fixed length string */
311 #define NLTYPE_ANY 1 /* Newline is any Unicode line ending */
312 #define NLTYPE_ANYCRLF 2 /* Newline is CR, LF, or CRLF */
313
314 /* This macro checks for a newline at the given position */
315
316 #define IS_NEWLINE(p) \
317 ((NLBLOCK->nltype != NLTYPE_FIXED)? \
318 ((p) < NLBLOCK->PSEND && \
319 PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, \
320 &(NLBLOCK->nllen), utf)) \
321 : \
322 ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
323 UCHAR21TEST(p) == NLBLOCK->nl[0] && \
324 (NLBLOCK->nllen == 1 || UCHAR21TEST(p+1) == NLBLOCK->nl[1]) \
325 ) \
326 )
327
328 /* This macro checks for a newline immediately preceding the given position */
329
330 #define WAS_NEWLINE(p) \
331 ((NLBLOCK->nltype != NLTYPE_FIXED)? \
332 ((p) > NLBLOCK->PSSTART && \
333 PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
334 &(NLBLOCK->nllen), utf)) \
335 : \
336 ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
337 UCHAR21TEST(p - NLBLOCK->nllen) == NLBLOCK->nl[0] && \
338 (NLBLOCK->nllen == 1 || UCHAR21TEST(p - NLBLOCK->nllen + 1) == NLBLOCK->nl[1]) \
339 ) \
340 )
341
342 /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
343 with a custom type. This makes it possible, for example, to allow pcre_exec()
344 to process subject strings that are discontinuous by using a smart pointer
345 class. It must always be possible to inspect all of the subject string in
346 pcre_exec() because of the way it backtracks. Two macros are required in the
347 normal case, for sign-unspecified and unsigned char pointers. The former is
348 used for the external interface and appears in pcre.h, which is why its name
349 must begin with PCRE_. */
350
351 #ifdef CUSTOM_SUBJECT_PTR
352 #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
353 #else
354 #define PCRE_PUCHAR const pcre_uchar *
355 #endif
356
357 /* Include the public PCRE header and the definitions of UCP character property
358 values. */
359
360 #include "pcre.h"
361 #include "ucp.h"
362
363 #ifdef COMPILE_PCRE32
364 /* Assert that the public PCRE_UCHAR32 is a 32-bit type */
365 typedef int __assert_pcre_uchar32_size[sizeof(PCRE_UCHAR32) == 4 ? 1 : -1];
366 #endif
367
368 /* When compiling for use with the Virtual Pascal compiler, these functions
369 need to have their names changed. PCRE must be compiled with the -DVPCOMPAT
370 option on the command line. */
371
372 #ifdef VPCOMPAT
373 #define strlen(s) _strlen(s)
374 #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
375 #define memcmp(s,c,n) _memcmp(s,c,n)
376 #define memcpy(d,s,n) _memcpy(d,s,n)
377 #define memmove(d,s,n) _memmove(d,s,n)
378 #define memset(s,c,n) _memset(s,c,n)
379 #else /* VPCOMPAT */
380
381 /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
382 define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
383 is set. Otherwise, include an emulating function for those systems that have
384 neither (there some non-Unix environments where this is the case). */
385
386 #ifndef HAVE_MEMMOVE
387 #undef memmove /* some systems may have a macro */
388 #ifdef HAVE_BCOPY
389 #define memmove(a, b, c) bcopy(b, a, c)
390 #else /* HAVE_BCOPY */
391 static void *
pcre_memmove(void * d,const void * s,size_t n)392 pcre_memmove(void *d, const void *s, size_t n)
393 {
394 size_t i;
395 unsigned char *dest = (unsigned char *)d;
396 const unsigned char *src = (const unsigned char *)s;
397 if (dest > src)
398 {
399 dest += n;
400 src += n;
401 for (i = 0; i < n; ++i) *(--dest) = *(--src);
402 return (void *)dest;
403 }
404 else
405 {
406 for (i = 0; i < n; ++i) *dest++ = *src++;
407 return (void *)(dest - n);
408 }
409 }
410 #define memmove(a, b, c) pcre_memmove(a, b, c)
411 #endif /* not HAVE_BCOPY */
412 #endif /* not HAVE_MEMMOVE */
413 #endif /* not VPCOMPAT */
414
415
416 /* PCRE keeps offsets in its compiled code as 2-byte quantities (always stored
417 in big-endian order) by default. These are used, for example, to link from the
418 start of a subpattern to its alternatives and its end. The use of 2 bytes per
419 offset limits the size of the compiled regex to around 64K, which is big enough
420 for almost everybody. However, I received a request for an even bigger limit.
421 For this reason, and also to make the code easier to maintain, the storing and
422 loading of offsets from the byte string is now handled by the macros that are
423 defined here.
424
425 The macros are controlled by the value of LINK_SIZE. This defaults to 2 in
426 the config.h file, but can be overridden by using -D on the command line. This
427 is automated on Unix systems via the "configure" command. */
428
429 #if defined COMPILE_PCRE8
430
431 #if LINK_SIZE == 2
432
433 #define PUT(a,n,d) \
434 (a[n] = (d) >> 8), \
435 (a[(n)+1] = (d) & 255)
436
437 #define GET(a,n) \
438 (((a)[n] << 8) | (a)[(n)+1])
439
440 #define MAX_PATTERN_SIZE (1 << 16)
441
442
443 #elif LINK_SIZE == 3
444
445 #define PUT(a,n,d) \
446 (a[n] = (d) >> 16), \
447 (a[(n)+1] = (d) >> 8), \
448 (a[(n)+2] = (d) & 255)
449
450 #define GET(a,n) \
451 (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2])
452
453 #define MAX_PATTERN_SIZE (1 << 24)
454
455
456 #elif LINK_SIZE == 4
457
458 #define PUT(a,n,d) \
459 (a[n] = (d) >> 24), \
460 (a[(n)+1] = (d) >> 16), \
461 (a[(n)+2] = (d) >> 8), \
462 (a[(n)+3] = (d) & 255)
463
464 #define GET(a,n) \
465 (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
466
467 /* Keep it positive */
468 #define MAX_PATTERN_SIZE (1 << 30)
469
470 #else
471 #error LINK_SIZE must be either 2, 3, or 4
472 #endif
473
474 #elif defined COMPILE_PCRE16
475
476 #if LINK_SIZE == 2
477
478 /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
479 #undef LINK_SIZE
480 #define LINK_SIZE 1
481
482 #define PUT(a,n,d) \
483 (a[n] = (d))
484
485 #define GET(a,n) \
486 (a[n])
487
488 #define MAX_PATTERN_SIZE (1 << 16)
489
490 #elif LINK_SIZE == 3 || LINK_SIZE == 4
491
492 /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
493 #undef LINK_SIZE
494 #define LINK_SIZE 2
495
496 #define PUT(a,n,d) \
497 (a[n] = (d) >> 16), \
498 (a[(n)+1] = (d) & 65535)
499
500 #define GET(a,n) \
501 (((a)[n] << 16) | (a)[(n)+1])
502
503 /* Keep it positive */
504 #define MAX_PATTERN_SIZE (1 << 30)
505
506 #else
507 #error LINK_SIZE must be either 2, 3, or 4
508 #endif
509
510 #elif defined COMPILE_PCRE32
511
512 /* Only supported LINK_SIZE is 4 */
513 /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
514 #undef LINK_SIZE
515 #define LINK_SIZE 1
516
517 #define PUT(a,n,d) \
518 (a[n] = (d))
519
520 #define GET(a,n) \
521 (a[n])
522
523 /* Keep it positive */
524 #define MAX_PATTERN_SIZE (1 << 30)
525
526 #else
527 #error Unsupported compiling mode
528 #endif /* COMPILE_PCRE[8|16|32] */
529
530 /* Convenience macro defined in terms of the others */
531
532 #define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE
533
534
535 /* PCRE uses some other 2-byte quantities that do not change when the size of
536 offsets changes. There are used for repeat counts and for other things such as
537 capturing parenthesis numbers in back references. */
538
539 #if defined COMPILE_PCRE8
540
541 #define IMM2_SIZE 2
542
543 #define PUT2(a,n,d) \
544 a[n] = (d) >> 8; \
545 a[(n)+1] = (d) & 255
546
547 /* For reasons that I do not understand, the expression in this GET2 macro is
548 treated by gcc as a signed expression, even when a is declared as unsigned. It
549 seems that any kind of arithmetic results in a signed value. */
550
551 #define GET2(a,n) \
552 (unsigned int)(((a)[n] << 8) | (a)[(n)+1])
553
554 #elif defined COMPILE_PCRE16
555
556 #define IMM2_SIZE 1
557
558 #define PUT2(a,n,d) \
559 a[n] = d
560
561 #define GET2(a,n) \
562 a[n]
563
564 #elif defined COMPILE_PCRE32
565
566 #define IMM2_SIZE 1
567
568 #define PUT2(a,n,d) \
569 a[n] = d
570
571 #define GET2(a,n) \
572 a[n]
573
574 #else
575 #error Unsupported compiling mode
576 #endif /* COMPILE_PCRE[8|16|32] */
577
578 #define PUT2INC(a,n,d) PUT2(a,n,d), a += IMM2_SIZE
579
580 /* The maximum length of a MARK name is currently one data unit; it may be
581 changed in future to be a fixed number of bytes or to depend on LINK_SIZE. */
582
583 #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
584 #define MAX_MARK ((1u << 16) - 1)
585 #else
586 #define MAX_MARK ((1u << 8) - 1)
587 #endif
588
589 /* There is a proposed future special "UTF-21" mode, in which only the lowest
590 21 bits of a 32-bit character are interpreted as UTF, with the remaining 11
591 high-order bits available to the application for other uses. In preparation for
592 the future implementation of this mode, there are macros that load a data item
593 and, if in this special mode, mask it to 21 bits. These macros all have names
594 starting with UCHAR21. In all other modes, including the normal 32-bit
595 library, the macros all have the same simple definitions. When the new mode is
596 implemented, it is expected that these definitions will be varied appropriately
597 using #ifdef when compiling the library that supports the special mode. */
598
599 #define UCHAR21(eptr) (*(eptr))
600 #define UCHAR21TEST(eptr) (*(eptr))
601 #define UCHAR21INC(eptr) (*(eptr)++)
602 #define UCHAR21INCTEST(eptr) (*(eptr)++)
603
604 /* When UTF encoding is being used, a character is no longer just a single
605 byte in 8-bit mode or a single short in 16-bit mode. The macros for character
606 handling generate simple sequences when used in the basic mode, and more
607 complicated ones for UTF characters. GETCHARLENTEST and other macros are not
608 used when UTF is not supported. To make sure they can never even appear when
609 UTF support is omitted, we don't even define them. */
610
611 #ifndef SUPPORT_UTF
612
613 /* #define MAX_VALUE_FOR_SINGLE_CHAR */
614 /* #define HAS_EXTRALEN(c) */
615 /* #define GET_EXTRALEN(c) */
616 /* #define NOT_FIRSTCHAR(c) */
617 #define GETCHAR(c, eptr) c = *eptr;
618 #define GETCHARTEST(c, eptr) c = *eptr;
619 #define GETCHARINC(c, eptr) c = *eptr++;
620 #define GETCHARINCTEST(c, eptr) c = *eptr++;
621 #define GETCHARLEN(c, eptr, len) c = *eptr;
622 /* #define GETCHARLENTEST(c, eptr, len) */
623 /* #define BACKCHAR(eptr) */
624 /* #define FORWARDCHAR(eptr) */
625 /* #define ACROSSCHAR(condition, eptr, action) */
626
627 #else /* SUPPORT_UTF */
628
629 /* Tests whether the code point needs extra characters to decode. */
630
631 #define HASUTF8EXTRALEN(c) ((c) >= 0xc0)
632
633 /* Base macro to pick up the remaining bytes of a UTF-8 character, not
634 advancing the pointer. */
635
636 #define GETUTF8(c, eptr) \
637 { \
638 if ((c & 0x20) == 0) \
639 c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
640 else if ((c & 0x10) == 0) \
641 c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
642 else if ((c & 0x08) == 0) \
643 c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
644 ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
645 else if ((c & 0x04) == 0) \
646 c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
647 ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
648 (eptr[4] & 0x3f); \
649 else \
650 c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
651 ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
652 ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
653 }
654
655 /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
656 the pointer. */
657
658 #define GETUTF8INC(c, eptr) \
659 { \
660 if ((c & 0x20) == 0) \
661 c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
662 else if ((c & 0x10) == 0) \
663 { \
664 c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
665 eptr += 2; \
666 } \
667 else if ((c & 0x08) == 0) \
668 { \
669 c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
670 ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
671 eptr += 3; \
672 } \
673 else if ((c & 0x04) == 0) \
674 { \
675 c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
676 ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
677 (eptr[3] & 0x3f); \
678 eptr += 4; \
679 } \
680 else \
681 { \
682 c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
683 ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
684 ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
685 eptr += 5; \
686 } \
687 }
688
689 #if defined COMPILE_PCRE8
690
691 /* These macros were originally written in the form of loops that used data
692 from the tables whose names start with PRIV(utf8_table). They were rewritten by
693 a user so as not to use loops, because in some environments this gives a
694 significant performance advantage, and it seems never to do any harm. */
695
696 /* Tells the biggest code point which can be encoded as a single character. */
697
698 #define MAX_VALUE_FOR_SINGLE_CHAR 127
699
700 /* Tests whether the code point needs extra characters to decode. */
701
702 #define HAS_EXTRALEN(c) ((c) >= 0xc0)
703
704 /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
705 Otherwise it has an undefined behaviour. */
706
707 #define GET_EXTRALEN(c) (PRIV(utf8_table4)[(c) & 0x3f])
708
709 /* Returns TRUE, if the given character is not the first character
710 of a UTF sequence. */
711
712 #define NOT_FIRSTCHAR(c) (((c) & 0xc0) == 0x80)
713
714 /* Get the next UTF-8 character, not advancing the pointer. This is called when
715 we know we are in UTF-8 mode. */
716
717 #define GETCHAR(c, eptr) \
718 c = *eptr; \
719 if (c >= 0xc0) GETUTF8(c, eptr);
720
721 /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
722 pointer. */
723
724 #define GETCHARTEST(c, eptr) \
725 c = *eptr; \
726 if (utf && c >= 0xc0) GETUTF8(c, eptr);
727
728 /* Get the next UTF-8 character, advancing the pointer. This is called when we
729 know we are in UTF-8 mode. */
730
731 #define GETCHARINC(c, eptr) \
732 c = *eptr++; \
733 if (c >= 0xc0) GETUTF8INC(c, eptr);
734
735 /* Get the next character, testing for UTF-8 mode, and advancing the pointer.
736 This is called when we don't know if we are in UTF-8 mode. */
737
738 #define GETCHARINCTEST(c, eptr) \
739 c = *eptr++; \
740 if (utf && c >= 0xc0) GETUTF8INC(c, eptr);
741
742 /* Base macro to pick up the remaining bytes of a UTF-8 character, not
743 advancing the pointer, incrementing the length. */
744
745 #define GETUTF8LEN(c, eptr, len) \
746 { \
747 if ((c & 0x20) == 0) \
748 { \
749 c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
750 len++; \
751 } \
752 else if ((c & 0x10) == 0) \
753 { \
754 c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
755 len += 2; \
756 } \
757 else if ((c & 0x08) == 0) \
758 {\
759 c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
760 ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
761 len += 3; \
762 } \
763 else if ((c & 0x04) == 0) \
764 { \
765 c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
766 ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
767 (eptr[4] & 0x3f); \
768 len += 4; \
769 } \
770 else \
771 {\
772 c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
773 ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
774 ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
775 len += 5; \
776 } \
777 }
778
779 /* Get the next UTF-8 character, not advancing the pointer, incrementing length
780 if there are extra bytes. This is called when we know we are in UTF-8 mode. */
781
782 #define GETCHARLEN(c, eptr, len) \
783 c = *eptr; \
784 if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
785
786 /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
787 pointer, incrementing length if there are extra bytes. This is called when we
788 do not know if we are in UTF-8 mode. */
789
790 #define GETCHARLENTEST(c, eptr, len) \
791 c = *eptr; \
792 if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len);
793
794 /* If the pointer is not at the start of a character, move it back until
795 it is. This is called only in UTF-8 mode - we don't put a test within the macro
796 because almost all calls are already within a block of UTF-8 only code. */
797
798 #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
799
800 /* Same as above, just in the other direction. */
801 #define FORWARDCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr++
802
803 /* Same as above, but it allows a fully customizable form. */
804 #define ACROSSCHAR(condition, eptr, action) \
805 while((condition) && ((eptr) & 0xc0) == 0x80) action
806
807 #elif defined COMPILE_PCRE16
808
809 /* Tells the biggest code point which can be encoded as a single character. */
810
811 #define MAX_VALUE_FOR_SINGLE_CHAR 65535
812
813 /* Tests whether the code point needs extra characters to decode. */
814
815 #define HAS_EXTRALEN(c) (((c) & 0xfc00) == 0xd800)
816
817 /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
818 Otherwise it has an undefined behaviour. */
819
820 #define GET_EXTRALEN(c) 1
821
822 /* Returns TRUE, if the given character is not the first character
823 of a UTF sequence. */
824
825 #define NOT_FIRSTCHAR(c) (((c) & 0xfc00) == 0xdc00)
826
827 /* Base macro to pick up the low surrogate of a UTF-16 character, not
828 advancing the pointer. */
829
830 #define GETUTF16(c, eptr) \
831 { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; }
832
833 /* Get the next UTF-16 character, not advancing the pointer. This is called when
834 we know we are in UTF-16 mode. */
835
836 #define GETCHAR(c, eptr) \
837 c = *eptr; \
838 if ((c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
839
840 /* Get the next UTF-16 character, testing for UTF-16 mode, and not advancing the
841 pointer. */
842
843 #define GETCHARTEST(c, eptr) \
844 c = *eptr; \
845 if (utf && (c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
846
847 /* Base macro to pick up the low surrogate of a UTF-16 character, advancing
848 the pointer. */
849
850 #define GETUTF16INC(c, eptr) \
851 { c = (((c & 0x3ff) << 10) | (*eptr++ & 0x3ff)) + 0x10000; }
852
853 /* Get the next UTF-16 character, advancing the pointer. This is called when we
854 know we are in UTF-16 mode. */
855
856 #define GETCHARINC(c, eptr) \
857 c = *eptr++; \
858 if ((c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
859
860 /* Get the next character, testing for UTF-16 mode, and advancing the pointer.
861 This is called when we don't know if we are in UTF-16 mode. */
862
863 #define GETCHARINCTEST(c, eptr) \
864 c = *eptr++; \
865 if (utf && (c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
866
867 /* Base macro to pick up the low surrogate of a UTF-16 character, not
868 advancing the pointer, incrementing the length. */
869
870 #define GETUTF16LEN(c, eptr, len) \
871 { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; len++; }
872
873 /* Get the next UTF-16 character, not advancing the pointer, incrementing
874 length if there is a low surrogate. This is called when we know we are in
875 UTF-16 mode. */
876
877 #define GETCHARLEN(c, eptr, len) \
878 c = *eptr; \
879 if ((c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
880
881 /* Get the next UTF-816character, testing for UTF-16 mode, not advancing the
882 pointer, incrementing length if there is a low surrogate. This is called when
883 we do not know if we are in UTF-16 mode. */
884
885 #define GETCHARLENTEST(c, eptr, len) \
886 c = *eptr; \
887 if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
888
889 /* If the pointer is not at the start of a character, move it back until
890 it is. This is called only in UTF-16 mode - we don't put a test within the
891 macro because almost all calls are already within a block of UTF-16 only
892 code. */
893
894 #define BACKCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr--
895
896 /* Same as above, just in the other direction. */
897 #define FORWARDCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr++
898
899 /* Same as above, but it allows a fully customizable form. */
900 #define ACROSSCHAR(condition, eptr, action) \
901 if ((condition) && ((eptr) & 0xfc00) == 0xdc00) action
902
903 #elif defined COMPILE_PCRE32
904
905 /* These are trivial for the 32-bit library, since all UTF-32 characters fit
906 into one pcre_uchar unit. */
907 #define MAX_VALUE_FOR_SINGLE_CHAR (0x10ffffu)
908 #define HAS_EXTRALEN(c) (0)
909 #define GET_EXTRALEN(c) (0)
910 #define NOT_FIRSTCHAR(c) (0)
911
912 /* Get the next UTF-32 character, not advancing the pointer. This is called when
913 we know we are in UTF-32 mode. */
914
915 #define GETCHAR(c, eptr) \
916 c = *(eptr);
917
918 /* Get the next UTF-32 character, testing for UTF-32 mode, and not advancing the
919 pointer. */
920
921 #define GETCHARTEST(c, eptr) \
922 c = *(eptr);
923
924 /* Get the next UTF-32 character, advancing the pointer. This is called when we
925 know we are in UTF-32 mode. */
926
927 #define GETCHARINC(c, eptr) \
928 c = *((eptr)++);
929
930 /* Get the next character, testing for UTF-32 mode, and advancing the pointer.
931 This is called when we don't know if we are in UTF-32 mode. */
932
933 #define GETCHARINCTEST(c, eptr) \
934 c = *((eptr)++);
935
936 /* Get the next UTF-32 character, not advancing the pointer, not incrementing
937 length (since all UTF-32 is of length 1). This is called when we know we are in
938 UTF-32 mode. */
939
940 #define GETCHARLEN(c, eptr, len) \
941 GETCHAR(c, eptr)
942
943 /* Get the next UTF-32character, testing for UTF-32 mode, not advancing the
944 pointer, not incrementing the length (since all UTF-32 is of length 1).
945 This is called when we do not know if we are in UTF-32 mode. */
946
947 #define GETCHARLENTEST(c, eptr, len) \
948 GETCHARTEST(c, eptr)
949
950 /* If the pointer is not at the start of a character, move it back until
951 it is. This is called only in UTF-32 mode - we don't put a test within the
952 macro because almost all calls are already within a block of UTF-32 only
953 code.
954 These are all no-ops since all UTF-32 characters fit into one pcre_uchar. */
955
956 #define BACKCHAR(eptr) do { } while (0)
957
958 /* Same as above, just in the other direction. */
959 #define FORWARDCHAR(eptr) do { } while (0)
960
961 /* Same as above, but it allows a fully customizable form. */
962 #define ACROSSCHAR(condition, eptr, action) do { } while (0)
963
964 #else
965 #error Unsupported compiling mode
966 #endif /* COMPILE_PCRE[8|16|32] */
967
968 #endif /* SUPPORT_UTF */
969
970 /* Tests for Unicode horizontal and vertical whitespace characters must check a
971 number of different values. Using a switch statement for this generates the
972 fastest code (no loop, no memory access), and there are several places in the
973 interpreter code where this happens. In order to ensure that all the case lists
974 remain in step, we use macros so that there is only one place where the lists
975 are defined.
976
977 These values are also required as lists in pcre_compile.c when processing \h,
978 \H, \v and \V in a character class. The lists are defined in pcre_tables.c, but
979 macros that define the values are here so that all the definitions are
980 together. The lists must be in ascending character order, terminated by
981 NOTACHAR (which is 0xffffffff).
982
983 Any changes should ensure that the various macros are kept in step with each
984 other. NOTE: The values also appear in pcre_jit_compile.c. */
985
986 /* ------ ASCII/Unicode environments ------ */
987
988 #ifndef EBCDIC
989
990 #define HSPACE_LIST \
991 CHAR_HT, CHAR_SPACE, 0xa0, \
992 0x1680, 0x180e, 0x2000, 0x2001, 0x2002, 0x2003, 0x2004, 0x2005, \
993 0x2006, 0x2007, 0x2008, 0x2009, 0x200A, 0x202f, 0x205f, 0x3000, \
994 NOTACHAR
995
996 #define HSPACE_MULTIBYTE_CASES \
997 case 0x1680: /* OGHAM SPACE MARK */ \
998 case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ \
999 case 0x2000: /* EN QUAD */ \
1000 case 0x2001: /* EM QUAD */ \
1001 case 0x2002: /* EN SPACE */ \
1002 case 0x2003: /* EM SPACE */ \
1003 case 0x2004: /* THREE-PER-EM SPACE */ \
1004 case 0x2005: /* FOUR-PER-EM SPACE */ \
1005 case 0x2006: /* SIX-PER-EM SPACE */ \
1006 case 0x2007: /* FIGURE SPACE */ \
1007 case 0x2008: /* PUNCTUATION SPACE */ \
1008 case 0x2009: /* THIN SPACE */ \
1009 case 0x200A: /* HAIR SPACE */ \
1010 case 0x202f: /* NARROW NO-BREAK SPACE */ \
1011 case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ \
1012 case 0x3000 /* IDEOGRAPHIC SPACE */
1013
1014 #define HSPACE_BYTE_CASES \
1015 case CHAR_HT: \
1016 case CHAR_SPACE: \
1017 case 0xa0 /* NBSP */
1018
1019 #define HSPACE_CASES \
1020 HSPACE_BYTE_CASES: \
1021 HSPACE_MULTIBYTE_CASES
1022
1023 #define VSPACE_LIST \
1024 CHAR_LF, CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, 0x2028, 0x2029, NOTACHAR
1025
1026 #define VSPACE_MULTIBYTE_CASES \
1027 case 0x2028: /* LINE SEPARATOR */ \
1028 case 0x2029 /* PARAGRAPH SEPARATOR */
1029
1030 #define VSPACE_BYTE_CASES \
1031 case CHAR_LF: \
1032 case CHAR_VT: \
1033 case CHAR_FF: \
1034 case CHAR_CR: \
1035 case CHAR_NEL
1036
1037 #define VSPACE_CASES \
1038 VSPACE_BYTE_CASES: \
1039 VSPACE_MULTIBYTE_CASES
1040
1041 /* ------ EBCDIC environments ------ */
1042
1043 #else
1044 #define HSPACE_LIST CHAR_HT, CHAR_SPACE
1045
1046 #define HSPACE_BYTE_CASES \
1047 case CHAR_HT: \
1048 case CHAR_SPACE
1049
1050 #define HSPACE_CASES HSPACE_BYTE_CASES
1051
1052 #ifdef EBCDIC_NL25
1053 #define VSPACE_LIST \
1054 CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, CHAR_LF, NOTACHAR
1055 #else
1056 #define VSPACE_LIST \
1057 CHAR_VT, CHAR_FF, CHAR_CR, CHAR_LF, CHAR_NEL, NOTACHAR
1058 #endif
1059
1060 #define VSPACE_BYTE_CASES \
1061 case CHAR_LF: \
1062 case CHAR_VT: \
1063 case CHAR_FF: \
1064 case CHAR_CR: \
1065 case CHAR_NEL
1066
1067 #define VSPACE_CASES VSPACE_BYTE_CASES
1068 #endif /* EBCDIC */
1069
1070 /* ------ End of whitespace macros ------ */
1071
1072
1073
1074 /* Private flags containing information about the compiled regex. They used to
1075 live at the top end of the options word, but that got almost full, so they were
1076 moved to a 16-bit flags word - which got almost full, so now they are in a
1077 32-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as the
1078 restrictions on partial matching have been lifted. It remains for backwards
1079 compatibility. */
1080
1081 #define PCRE_MODE8 0x00000001 /* compiled in 8 bit mode */
1082 #define PCRE_MODE16 0x00000002 /* compiled in 16 bit mode */
1083 #define PCRE_MODE32 0x00000004 /* compiled in 32 bit mode */
1084 #define PCRE_FIRSTSET 0x00000010 /* first_char is set */
1085 #define PCRE_FCH_CASELESS 0x00000020 /* caseless first char */
1086 #define PCRE_REQCHSET 0x00000040 /* req_byte is set */
1087 #define PCRE_RCH_CASELESS 0x00000080 /* caseless requested char */
1088 #define PCRE_STARTLINE 0x00000100 /* start after \n for multiline */
1089 #define PCRE_NOPARTIAL 0x00000200 /* can't use partial with this regex */
1090 #define PCRE_JCHANGED 0x00000400 /* j option used in regex */
1091 #define PCRE_HASCRORLF 0x00000800 /* explicit \r or \n in pattern */
1092 #define PCRE_HASTHEN 0x00001000 /* pattern contains (*THEN) */
1093 #define PCRE_MLSET 0x00002000 /* match limit set by regex */
1094 #define PCRE_RLSET 0x00004000 /* recursion limit set by regex */
1095 #define PCRE_MATCH_EMPTY 0x00008000 /* pattern can match empty string */
1096
1097 #if defined COMPILE_PCRE8
1098 #define PCRE_MODE PCRE_MODE8
1099 #elif defined COMPILE_PCRE16
1100 #define PCRE_MODE PCRE_MODE16
1101 #elif defined COMPILE_PCRE32
1102 #define PCRE_MODE PCRE_MODE32
1103 #endif
1104 #define PCRE_MODE_MASK (PCRE_MODE8 | PCRE_MODE16 | PCRE_MODE32)
1105
1106 /* Flags for the "extra" block produced by pcre_study(). */
1107
1108 #define PCRE_STUDY_MAPPED 0x0001 /* a map of starting chars exists */
1109 #define PCRE_STUDY_MINLEN 0x0002 /* a minimum length field exists */
1110
1111 /* Masks for identifying the public options that are permitted at compile
1112 time, run time, or study time, respectively. */
1113
1114 #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
1115 PCRE_NEWLINE_ANYCRLF)
1116
1117 #define PUBLIC_COMPILE_OPTIONS \
1118 (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
1119 PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
1120 PCRE_NO_AUTO_CAPTURE|PCRE_NO_AUTO_POSSESS| \
1121 PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
1122 PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
1123 PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE|PCRE_NEVER_UTF)
1124
1125 #define PUBLIC_EXEC_OPTIONS \
1126 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
1127 PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
1128 PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
1129
1130 #define PUBLIC_DFA_EXEC_OPTIONS \
1131 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
1132 PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
1133 PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
1134 PCRE_NO_START_OPTIMIZE)
1135
1136 #define PUBLIC_STUDY_OPTIONS \
1137 (PCRE_STUDY_JIT_COMPILE|PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE| \
1138 PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE|PCRE_STUDY_EXTRA_NEEDED)
1139
1140 #define PUBLIC_JIT_EXEC_OPTIONS \
1141 (PCRE_NO_UTF8_CHECK|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|\
1142 PCRE_NOTEMPTY_ATSTART|PCRE_PARTIAL_SOFT|PCRE_PARTIAL_HARD)
1143
1144 /* Magic number to provide a small check against being handed junk. */
1145
1146 #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
1147
1148 /* This variable is used to detect a loaded regular expression
1149 in different endianness. */
1150
1151 #define REVERSED_MAGIC_NUMBER 0x45524350UL /* 'ERCP' */
1152
1153 /* The maximum remaining length of subject we are prepared to search for a
1154 req_byte match. */
1155
1156 #define REQ_BYTE_MAX 1000
1157
1158 /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
1159 environments where these macros are defined elsewhere. Unfortunately, there
1160 is no way to do the same for the typedef. */
1161
1162 typedef int BOOL;
1163
1164 #ifndef FALSE
1165 #define FALSE 0
1166 #define TRUE 1
1167 #endif
1168
1169 /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
1170 character constants like '*' because the compiler would emit their EBCDIC code,
1171 which is different from their ASCII/UTF-8 code. Instead we define macros for
1172 the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
1173 is enabled. When UTF-8 support is not enabled, the definitions use character
1174 literals. Both character and string versions of each character are needed, and
1175 there are some longer strings as well.
1176
1177 This means that, on EBCDIC platforms, the PCRE library can handle either
1178 EBCDIC, or UTF-8, but not both. To support both in the same compiled library
1179 would need different lookups depending on whether PCRE_UTF8 was set or not.
1180 This would make it impossible to use characters in switch/case statements,
1181 which would reduce performance. For a theoretical use (which nobody has asked
1182 for) in a minority area (EBCDIC platforms), this is not sensible. Any
1183 application that did need both could compile two versions of the library, using
1184 macros to give the functions distinct names. */
1185
1186 #ifndef SUPPORT_UTF
1187
1188 /* UTF-8 support is not enabled; use the platform-dependent character literals
1189 so that PCRE works in both ASCII and EBCDIC environments, but only in non-UTF
1190 mode. Newline characters are problematic in EBCDIC. Though it has CR and LF
1191 characters, a common practice has been to use its NL (0x15) character as the
1192 line terminator in C-like processing environments. However, sometimes the LF
1193 (0x25) character is used instead, according to this Unicode document:
1194
1195 http://unicode.org/standard/reports/tr13/tr13-5.html
1196
1197 PCRE defaults EBCDIC NL to 0x15, but has a build-time option to select 0x25
1198 instead. Whichever is *not* chosen is defined as NEL.
1199
1200 In both ASCII and EBCDIC environments, CHAR_NL and CHAR_LF are synonyms for the
1201 same code point. */
1202
1203 #ifdef EBCDIC
1204
1205 #ifndef EBCDIC_NL25
1206 #define CHAR_NL '\x15'
1207 #define CHAR_NEL '\x25'
1208 #define STR_NL "\x15"
1209 #define STR_NEL "\x25"
1210 #else
1211 #define CHAR_NL '\x25'
1212 #define CHAR_NEL '\x15'
1213 #define STR_NL "\x25"
1214 #define STR_NEL "\x15"
1215 #endif
1216
1217 #define CHAR_LF CHAR_NL
1218 #define STR_LF STR_NL
1219
1220 #define CHAR_ESC '\047'
1221 #define CHAR_DEL '\007'
1222 #define STR_ESC "\047"
1223 #define STR_DEL "\007"
1224
1225 #else /* Not EBCDIC */
1226
1227 /* In ASCII/Unicode, linefeed is '\n' and we equate this to NL for
1228 compatibility. NEL is the Unicode newline character; make sure it is
1229 a positive value. */
1230
1231 #define CHAR_LF '\n'
1232 #define CHAR_NL CHAR_LF
1233 #define CHAR_NEL ((unsigned char)'\x85')
1234 #define CHAR_ESC '\033'
1235 #define CHAR_DEL '\177'
1236
1237 #define STR_LF "\n"
1238 #define STR_NL STR_LF
1239 #define STR_NEL "\x85"
1240 #define STR_ESC "\033"
1241 #define STR_DEL "\177"
1242
1243 #endif /* EBCDIC */
1244
1245 /* The remaining definitions work in both environments. */
1246
1247 #define CHAR_NULL '\0'
1248 #define CHAR_HT '\t'
1249 #define CHAR_VT '\v'
1250 #define CHAR_FF '\f'
1251 #define CHAR_CR '\r'
1252 #define CHAR_BS '\b'
1253 #define CHAR_BEL '\a'
1254
1255 #define CHAR_SPACE ' '
1256 #define CHAR_EXCLAMATION_MARK '!'
1257 #define CHAR_QUOTATION_MARK '"'
1258 #define CHAR_NUMBER_SIGN '#'
1259 #define CHAR_DOLLAR_SIGN '$'
1260 #define CHAR_PERCENT_SIGN '%'
1261 #define CHAR_AMPERSAND '&'
1262 #define CHAR_APOSTROPHE '\''
1263 #define CHAR_LEFT_PARENTHESIS '('
1264 #define CHAR_RIGHT_PARENTHESIS ')'
1265 #define CHAR_ASTERISK '*'
1266 #define CHAR_PLUS '+'
1267 #define CHAR_COMMA ','
1268 #define CHAR_MINUS '-'
1269 #define CHAR_DOT '.'
1270 #define CHAR_SLASH '/'
1271 #define CHAR_0 '0'
1272 #define CHAR_1 '1'
1273 #define CHAR_2 '2'
1274 #define CHAR_3 '3'
1275 #define CHAR_4 '4'
1276 #define CHAR_5 '5'
1277 #define CHAR_6 '6'
1278 #define CHAR_7 '7'
1279 #define CHAR_8 '8'
1280 #define CHAR_9 '9'
1281 #define CHAR_COLON ':'
1282 #define CHAR_SEMICOLON ';'
1283 #define CHAR_LESS_THAN_SIGN '<'
1284 #define CHAR_EQUALS_SIGN '='
1285 #define CHAR_GREATER_THAN_SIGN '>'
1286 #define CHAR_QUESTION_MARK '?'
1287 #define CHAR_COMMERCIAL_AT '@'
1288 #define CHAR_A 'A'
1289 #define CHAR_B 'B'
1290 #define CHAR_C 'C'
1291 #define CHAR_D 'D'
1292 #define CHAR_E 'E'
1293 #define CHAR_F 'F'
1294 #define CHAR_G 'G'
1295 #define CHAR_H 'H'
1296 #define CHAR_I 'I'
1297 #define CHAR_J 'J'
1298 #define CHAR_K 'K'
1299 #define CHAR_L 'L'
1300 #define CHAR_M 'M'
1301 #define CHAR_N 'N'
1302 #define CHAR_O 'O'
1303 #define CHAR_P 'P'
1304 #define CHAR_Q 'Q'
1305 #define CHAR_R 'R'
1306 #define CHAR_S 'S'
1307 #define CHAR_T 'T'
1308 #define CHAR_U 'U'
1309 #define CHAR_V 'V'
1310 #define CHAR_W 'W'
1311 #define CHAR_X 'X'
1312 #define CHAR_Y 'Y'
1313 #define CHAR_Z 'Z'
1314 #define CHAR_LEFT_SQUARE_BRACKET '['
1315 #define CHAR_BACKSLASH '\\'
1316 #define CHAR_RIGHT_SQUARE_BRACKET ']'
1317 #define CHAR_CIRCUMFLEX_ACCENT '^'
1318 #define CHAR_UNDERSCORE '_'
1319 #define CHAR_GRAVE_ACCENT '`'
1320 #define CHAR_a 'a'
1321 #define CHAR_b 'b'
1322 #define CHAR_c 'c'
1323 #define CHAR_d 'd'
1324 #define CHAR_e 'e'
1325 #define CHAR_f 'f'
1326 #define CHAR_g 'g'
1327 #define CHAR_h 'h'
1328 #define CHAR_i 'i'
1329 #define CHAR_j 'j'
1330 #define CHAR_k 'k'
1331 #define CHAR_l 'l'
1332 #define CHAR_m 'm'
1333 #define CHAR_n 'n'
1334 #define CHAR_o 'o'
1335 #define CHAR_p 'p'
1336 #define CHAR_q 'q'
1337 #define CHAR_r 'r'
1338 #define CHAR_s 's'
1339 #define CHAR_t 't'
1340 #define CHAR_u 'u'
1341 #define CHAR_v 'v'
1342 #define CHAR_w 'w'
1343 #define CHAR_x 'x'
1344 #define CHAR_y 'y'
1345 #define CHAR_z 'z'
1346 #define CHAR_LEFT_CURLY_BRACKET '{'
1347 #define CHAR_VERTICAL_LINE '|'
1348 #define CHAR_RIGHT_CURLY_BRACKET '}'
1349 #define CHAR_TILDE '~'
1350
1351 #define STR_HT "\t"
1352 #define STR_VT "\v"
1353 #define STR_FF "\f"
1354 #define STR_CR "\r"
1355 #define STR_BS "\b"
1356 #define STR_BEL "\a"
1357
1358 #define STR_SPACE " "
1359 #define STR_EXCLAMATION_MARK "!"
1360 #define STR_QUOTATION_MARK "\""
1361 #define STR_NUMBER_SIGN "#"
1362 #define STR_DOLLAR_SIGN "$"
1363 #define STR_PERCENT_SIGN "%"
1364 #define STR_AMPERSAND "&"
1365 #define STR_APOSTROPHE "'"
1366 #define STR_LEFT_PARENTHESIS "("
1367 #define STR_RIGHT_PARENTHESIS ")"
1368 #define STR_ASTERISK "*"
1369 #define STR_PLUS "+"
1370 #define STR_COMMA ","
1371 #define STR_MINUS "-"
1372 #define STR_DOT "."
1373 #define STR_SLASH "/"
1374 #define STR_0 "0"
1375 #define STR_1 "1"
1376 #define STR_2 "2"
1377 #define STR_3 "3"
1378 #define STR_4 "4"
1379 #define STR_5 "5"
1380 #define STR_6 "6"
1381 #define STR_7 "7"
1382 #define STR_8 "8"
1383 #define STR_9 "9"
1384 #define STR_COLON ":"
1385 #define STR_SEMICOLON ";"
1386 #define STR_LESS_THAN_SIGN "<"
1387 #define STR_EQUALS_SIGN "="
1388 #define STR_GREATER_THAN_SIGN ">"
1389 #define STR_QUESTION_MARK "?"
1390 #define STR_COMMERCIAL_AT "@"
1391 #define STR_A "A"
1392 #define STR_B "B"
1393 #define STR_C "C"
1394 #define STR_D "D"
1395 #define STR_E "E"
1396 #define STR_F "F"
1397 #define STR_G "G"
1398 #define STR_H "H"
1399 #define STR_I "I"
1400 #define STR_J "J"
1401 #define STR_K "K"
1402 #define STR_L "L"
1403 #define STR_M "M"
1404 #define STR_N "N"
1405 #define STR_O "O"
1406 #define STR_P "P"
1407 #define STR_Q "Q"
1408 #define STR_R "R"
1409 #define STR_S "S"
1410 #define STR_T "T"
1411 #define STR_U "U"
1412 #define STR_V "V"
1413 #define STR_W "W"
1414 #define STR_X "X"
1415 #define STR_Y "Y"
1416 #define STR_Z "Z"
1417 #define STR_LEFT_SQUARE_BRACKET "["
1418 #define STR_BACKSLASH "\\"
1419 #define STR_RIGHT_SQUARE_BRACKET "]"
1420 #define STR_CIRCUMFLEX_ACCENT "^"
1421 #define STR_UNDERSCORE "_"
1422 #define STR_GRAVE_ACCENT "`"
1423 #define STR_a "a"
1424 #define STR_b "b"
1425 #define STR_c "c"
1426 #define STR_d "d"
1427 #define STR_e "e"
1428 #define STR_f "f"
1429 #define STR_g "g"
1430 #define STR_h "h"
1431 #define STR_i "i"
1432 #define STR_j "j"
1433 #define STR_k "k"
1434 #define STR_l "l"
1435 #define STR_m "m"
1436 #define STR_n "n"
1437 #define STR_o "o"
1438 #define STR_p "p"
1439 #define STR_q "q"
1440 #define STR_r "r"
1441 #define STR_s "s"
1442 #define STR_t "t"
1443 #define STR_u "u"
1444 #define STR_v "v"
1445 #define STR_w "w"
1446 #define STR_x "x"
1447 #define STR_y "y"
1448 #define STR_z "z"
1449 #define STR_LEFT_CURLY_BRACKET "{"
1450 #define STR_VERTICAL_LINE "|"
1451 #define STR_RIGHT_CURLY_BRACKET "}"
1452 #define STR_TILDE "~"
1453
1454 #define STRING_ACCEPT0 "ACCEPT\0"
1455 #define STRING_COMMIT0 "COMMIT\0"
1456 #define STRING_F0 "F\0"
1457 #define STRING_FAIL0 "FAIL\0"
1458 #define STRING_MARK0 "MARK\0"
1459 #define STRING_PRUNE0 "PRUNE\0"
1460 #define STRING_SKIP0 "SKIP\0"
1461 #define STRING_THEN "THEN"
1462
1463 #define STRING_alpha0 "alpha\0"
1464 #define STRING_lower0 "lower\0"
1465 #define STRING_upper0 "upper\0"
1466 #define STRING_alnum0 "alnum\0"
1467 #define STRING_ascii0 "ascii\0"
1468 #define STRING_blank0 "blank\0"
1469 #define STRING_cntrl0 "cntrl\0"
1470 #define STRING_digit0 "digit\0"
1471 #define STRING_graph0 "graph\0"
1472 #define STRING_print0 "print\0"
1473 #define STRING_punct0 "punct\0"
1474 #define STRING_space0 "space\0"
1475 #define STRING_word0 "word\0"
1476 #define STRING_xdigit "xdigit"
1477
1478 #define STRING_DEFINE "DEFINE"
1479 #define STRING_WEIRD_STARTWORD "[:<:]]"
1480 #define STRING_WEIRD_ENDWORD "[:>:]]"
1481
1482 #define STRING_CR_RIGHTPAR "CR)"
1483 #define STRING_LF_RIGHTPAR "LF)"
1484 #define STRING_CRLF_RIGHTPAR "CRLF)"
1485 #define STRING_ANY_RIGHTPAR "ANY)"
1486 #define STRING_ANYCRLF_RIGHTPAR "ANYCRLF)"
1487 #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
1488 #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
1489 #define STRING_UTF8_RIGHTPAR "UTF8)"
1490 #define STRING_UTF16_RIGHTPAR "UTF16)"
1491 #define STRING_UTF32_RIGHTPAR "UTF32)"
1492 #define STRING_UTF_RIGHTPAR "UTF)"
1493 #define STRING_UCP_RIGHTPAR "UCP)"
1494 #define STRING_NO_AUTO_POSSESS_RIGHTPAR "NO_AUTO_POSSESS)"
1495 #define STRING_NO_START_OPT_RIGHTPAR "NO_START_OPT)"
1496 #define STRING_LIMIT_MATCH_EQ "LIMIT_MATCH="
1497 #define STRING_LIMIT_RECURSION_EQ "LIMIT_RECURSION="
1498
1499 #else /* SUPPORT_UTF */
1500
1501 /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
1502 works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
1503 only. */
1504
1505 #define CHAR_HT '\011'
1506 #define CHAR_VT '\013'
1507 #define CHAR_FF '\014'
1508 #define CHAR_CR '\015'
1509 #define CHAR_LF '\012'
1510 #define CHAR_NL CHAR_LF
1511 #define CHAR_NEL ((unsigned char)'\x85')
1512 #define CHAR_BS '\010'
1513 #define CHAR_BEL '\007'
1514 #define CHAR_ESC '\033'
1515 #define CHAR_DEL '\177'
1516
1517 #define CHAR_NULL '\0'
1518 #define CHAR_SPACE '\040'
1519 #define CHAR_EXCLAMATION_MARK '\041'
1520 #define CHAR_QUOTATION_MARK '\042'
1521 #define CHAR_NUMBER_SIGN '\043'
1522 #define CHAR_DOLLAR_SIGN '\044'
1523 #define CHAR_PERCENT_SIGN '\045'
1524 #define CHAR_AMPERSAND '\046'
1525 #define CHAR_APOSTROPHE '\047'
1526 #define CHAR_LEFT_PARENTHESIS '\050'
1527 #define CHAR_RIGHT_PARENTHESIS '\051'
1528 #define CHAR_ASTERISK '\052'
1529 #define CHAR_PLUS '\053'
1530 #define CHAR_COMMA '\054'
1531 #define CHAR_MINUS '\055'
1532 #define CHAR_DOT '\056'
1533 #define CHAR_SLASH '\057'
1534 #define CHAR_0 '\060'
1535 #define CHAR_1 '\061'
1536 #define CHAR_2 '\062'
1537 #define CHAR_3 '\063'
1538 #define CHAR_4 '\064'
1539 #define CHAR_5 '\065'
1540 #define CHAR_6 '\066'
1541 #define CHAR_7 '\067'
1542 #define CHAR_8 '\070'
1543 #define CHAR_9 '\071'
1544 #define CHAR_COLON '\072'
1545 #define CHAR_SEMICOLON '\073'
1546 #define CHAR_LESS_THAN_SIGN '\074'
1547 #define CHAR_EQUALS_SIGN '\075'
1548 #define CHAR_GREATER_THAN_SIGN '\076'
1549 #define CHAR_QUESTION_MARK '\077'
1550 #define CHAR_COMMERCIAL_AT '\100'
1551 #define CHAR_A '\101'
1552 #define CHAR_B '\102'
1553 #define CHAR_C '\103'
1554 #define CHAR_D '\104'
1555 #define CHAR_E '\105'
1556 #define CHAR_F '\106'
1557 #define CHAR_G '\107'
1558 #define CHAR_H '\110'
1559 #define CHAR_I '\111'
1560 #define CHAR_J '\112'
1561 #define CHAR_K '\113'
1562 #define CHAR_L '\114'
1563 #define CHAR_M '\115'
1564 #define CHAR_N '\116'
1565 #define CHAR_O '\117'
1566 #define CHAR_P '\120'
1567 #define CHAR_Q '\121'
1568 #define CHAR_R '\122'
1569 #define CHAR_S '\123'
1570 #define CHAR_T '\124'
1571 #define CHAR_U '\125'
1572 #define CHAR_V '\126'
1573 #define CHAR_W '\127'
1574 #define CHAR_X '\130'
1575 #define CHAR_Y '\131'
1576 #define CHAR_Z '\132'
1577 #define CHAR_LEFT_SQUARE_BRACKET '\133'
1578 #define CHAR_BACKSLASH '\134'
1579 #define CHAR_RIGHT_SQUARE_BRACKET '\135'
1580 #define CHAR_CIRCUMFLEX_ACCENT '\136'
1581 #define CHAR_UNDERSCORE '\137'
1582 #define CHAR_GRAVE_ACCENT '\140'
1583 #define CHAR_a '\141'
1584 #define CHAR_b '\142'
1585 #define CHAR_c '\143'
1586 #define CHAR_d '\144'
1587 #define CHAR_e '\145'
1588 #define CHAR_f '\146'
1589 #define CHAR_g '\147'
1590 #define CHAR_h '\150'
1591 #define CHAR_i '\151'
1592 #define CHAR_j '\152'
1593 #define CHAR_k '\153'
1594 #define CHAR_l '\154'
1595 #define CHAR_m '\155'
1596 #define CHAR_n '\156'
1597 #define CHAR_o '\157'
1598 #define CHAR_p '\160'
1599 #define CHAR_q '\161'
1600 #define CHAR_r '\162'
1601 #define CHAR_s '\163'
1602 #define CHAR_t '\164'
1603 #define CHAR_u '\165'
1604 #define CHAR_v '\166'
1605 #define CHAR_w '\167'
1606 #define CHAR_x '\170'
1607 #define CHAR_y '\171'
1608 #define CHAR_z '\172'
1609 #define CHAR_LEFT_CURLY_BRACKET '\173'
1610 #define CHAR_VERTICAL_LINE '\174'
1611 #define CHAR_RIGHT_CURLY_BRACKET '\175'
1612 #define CHAR_TILDE '\176'
1613
1614 #define STR_HT "\011"
1615 #define STR_VT "\013"
1616 #define STR_FF "\014"
1617 #define STR_CR "\015"
1618 #define STR_NL "\012"
1619 #define STR_BS "\010"
1620 #define STR_BEL "\007"
1621 #define STR_ESC "\033"
1622 #define STR_DEL "\177"
1623
1624 #define STR_SPACE "\040"
1625 #define STR_EXCLAMATION_MARK "\041"
1626 #define STR_QUOTATION_MARK "\042"
1627 #define STR_NUMBER_SIGN "\043"
1628 #define STR_DOLLAR_SIGN "\044"
1629 #define STR_PERCENT_SIGN "\045"
1630 #define STR_AMPERSAND "\046"
1631 #define STR_APOSTROPHE "\047"
1632 #define STR_LEFT_PARENTHESIS "\050"
1633 #define STR_RIGHT_PARENTHESIS "\051"
1634 #define STR_ASTERISK "\052"
1635 #define STR_PLUS "\053"
1636 #define STR_COMMA "\054"
1637 #define STR_MINUS "\055"
1638 #define STR_DOT "\056"
1639 #define STR_SLASH "\057"
1640 #define STR_0 "\060"
1641 #define STR_1 "\061"
1642 #define STR_2 "\062"
1643 #define STR_3 "\063"
1644 #define STR_4 "\064"
1645 #define STR_5 "\065"
1646 #define STR_6 "\066"
1647 #define STR_7 "\067"
1648 #define STR_8 "\070"
1649 #define STR_9 "\071"
1650 #define STR_COLON "\072"
1651 #define STR_SEMICOLON "\073"
1652 #define STR_LESS_THAN_SIGN "\074"
1653 #define STR_EQUALS_SIGN "\075"
1654 #define STR_GREATER_THAN_SIGN "\076"
1655 #define STR_QUESTION_MARK "\077"
1656 #define STR_COMMERCIAL_AT "\100"
1657 #define STR_A "\101"
1658 #define STR_B "\102"
1659 #define STR_C "\103"
1660 #define STR_D "\104"
1661 #define STR_E "\105"
1662 #define STR_F "\106"
1663 #define STR_G "\107"
1664 #define STR_H "\110"
1665 #define STR_I "\111"
1666 #define STR_J "\112"
1667 #define STR_K "\113"
1668 #define STR_L "\114"
1669 #define STR_M "\115"
1670 #define STR_N "\116"
1671 #define STR_O "\117"
1672 #define STR_P "\120"
1673 #define STR_Q "\121"
1674 #define STR_R "\122"
1675 #define STR_S "\123"
1676 #define STR_T "\124"
1677 #define STR_U "\125"
1678 #define STR_V "\126"
1679 #define STR_W "\127"
1680 #define STR_X "\130"
1681 #define STR_Y "\131"
1682 #define STR_Z "\132"
1683 #define STR_LEFT_SQUARE_BRACKET "\133"
1684 #define STR_BACKSLASH "\134"
1685 #define STR_RIGHT_SQUARE_BRACKET "\135"
1686 #define STR_CIRCUMFLEX_ACCENT "\136"
1687 #define STR_UNDERSCORE "\137"
1688 #define STR_GRAVE_ACCENT "\140"
1689 #define STR_a "\141"
1690 #define STR_b "\142"
1691 #define STR_c "\143"
1692 #define STR_d "\144"
1693 #define STR_e "\145"
1694 #define STR_f "\146"
1695 #define STR_g "\147"
1696 #define STR_h "\150"
1697 #define STR_i "\151"
1698 #define STR_j "\152"
1699 #define STR_k "\153"
1700 #define STR_l "\154"
1701 #define STR_m "\155"
1702 #define STR_n "\156"
1703 #define STR_o "\157"
1704 #define STR_p "\160"
1705 #define STR_q "\161"
1706 #define STR_r "\162"
1707 #define STR_s "\163"
1708 #define STR_t "\164"
1709 #define STR_u "\165"
1710 #define STR_v "\166"
1711 #define STR_w "\167"
1712 #define STR_x "\170"
1713 #define STR_y "\171"
1714 #define STR_z "\172"
1715 #define STR_LEFT_CURLY_BRACKET "\173"
1716 #define STR_VERTICAL_LINE "\174"
1717 #define STR_RIGHT_CURLY_BRACKET "\175"
1718 #define STR_TILDE "\176"
1719
1720 #define STRING_ACCEPT0 STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1721 #define STRING_COMMIT0 STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1722 #define STRING_F0 STR_F "\0"
1723 #define STRING_FAIL0 STR_F STR_A STR_I STR_L "\0"
1724 #define STRING_MARK0 STR_M STR_A STR_R STR_K "\0"
1725 #define STRING_PRUNE0 STR_P STR_R STR_U STR_N STR_E "\0"
1726 #define STRING_SKIP0 STR_S STR_K STR_I STR_P "\0"
1727 #define STRING_THEN STR_T STR_H STR_E STR_N
1728
1729 #define STRING_alpha0 STR_a STR_l STR_p STR_h STR_a "\0"
1730 #define STRING_lower0 STR_l STR_o STR_w STR_e STR_r "\0"
1731 #define STRING_upper0 STR_u STR_p STR_p STR_e STR_r "\0"
1732 #define STRING_alnum0 STR_a STR_l STR_n STR_u STR_m "\0"
1733 #define STRING_ascii0 STR_a STR_s STR_c STR_i STR_i "\0"
1734 #define STRING_blank0 STR_b STR_l STR_a STR_n STR_k "\0"
1735 #define STRING_cntrl0 STR_c STR_n STR_t STR_r STR_l "\0"
1736 #define STRING_digit0 STR_d STR_i STR_g STR_i STR_t "\0"
1737 #define STRING_graph0 STR_g STR_r STR_a STR_p STR_h "\0"
1738 #define STRING_print0 STR_p STR_r STR_i STR_n STR_t "\0"
1739 #define STRING_punct0 STR_p STR_u STR_n STR_c STR_t "\0"
1740 #define STRING_space0 STR_s STR_p STR_a STR_c STR_e "\0"
1741 #define STRING_word0 STR_w STR_o STR_r STR_d "\0"
1742 #define STRING_xdigit STR_x STR_d STR_i STR_g STR_i STR_t
1743
1744 #define STRING_DEFINE STR_D STR_E STR_F STR_I STR_N STR_E
1745 #define STRING_WEIRD_STARTWORD STR_LEFT_SQUARE_BRACKET STR_COLON STR_LESS_THAN_SIGN STR_COLON STR_RIGHT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET
1746 #define STRING_WEIRD_ENDWORD STR_LEFT_SQUARE_BRACKET STR_COLON STR_GREATER_THAN_SIGN STR_COLON STR_RIGHT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET
1747
1748 #define STRING_CR_RIGHTPAR STR_C STR_R STR_RIGHT_PARENTHESIS
1749 #define STRING_LF_RIGHTPAR STR_L STR_F STR_RIGHT_PARENTHESIS
1750 #define STRING_CRLF_RIGHTPAR STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1751 #define STRING_ANY_RIGHTPAR STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1752 #define STRING_ANYCRLF_RIGHTPAR STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1753 #define STRING_BSR_ANYCRLF_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1754 #define STRING_BSR_UNICODE_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS
1755 #define STRING_UTF8_RIGHTPAR STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1756 #define STRING_UTF16_RIGHTPAR STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS
1757 #define STRING_UTF32_RIGHTPAR STR_U STR_T STR_F STR_3 STR_2 STR_RIGHT_PARENTHESIS
1758 #define STRING_UTF_RIGHTPAR STR_U STR_T STR_F STR_RIGHT_PARENTHESIS
1759 #define STRING_UCP_RIGHTPAR STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1760 #define STRING_NO_AUTO_POSSESS_RIGHTPAR STR_N STR_O STR_UNDERSCORE STR_A STR_U STR_T STR_O STR_UNDERSCORE STR_P STR_O STR_S STR_S STR_E STR_S STR_S STR_RIGHT_PARENTHESIS
1761 #define STRING_NO_START_OPT_RIGHTPAR STR_N STR_O STR_UNDERSCORE STR_S STR_T STR_A STR_R STR_T STR_UNDERSCORE STR_O STR_P STR_T STR_RIGHT_PARENTHESIS
1762 #define STRING_LIMIT_MATCH_EQ STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_M STR_A STR_T STR_C STR_H STR_EQUALS_SIGN
1763 #define STRING_LIMIT_RECURSION_EQ STR_L STR_I STR_M STR_I STR_T STR_UNDERSCORE STR_R STR_E STR_C STR_U STR_R STR_S STR_I STR_O STR_N STR_EQUALS_SIGN
1764
1765 #endif /* SUPPORT_UTF */
1766
1767 /* Escape items that are just an encoding of a particular data value. */
1768
1769 #ifndef ESC_e
1770 #define ESC_e CHAR_ESC
1771 #endif
1772
1773 #ifndef ESC_f
1774 #define ESC_f CHAR_FF
1775 #endif
1776
1777 #ifndef ESC_n
1778 #define ESC_n CHAR_LF
1779 #endif
1780
1781 #ifndef ESC_r
1782 #define ESC_r CHAR_CR
1783 #endif
1784
1785 /* We can't officially use ESC_t because it is a POSIX reserved identifier
1786 (presumably because of all the others like size_t). */
1787
1788 #ifndef ESC_tee
1789 #define ESC_tee CHAR_HT
1790 #endif
1791
1792 /* Codes for different types of Unicode property */
1793
1794 #define PT_ANY 0 /* Any property - matches all chars */
1795 #define PT_LAMP 1 /* L& - the union of Lu, Ll, Lt */
1796 #define PT_GC 2 /* Specified general characteristic (e.g. L) */
1797 #define PT_PC 3 /* Specified particular characteristic (e.g. Lu) */
1798 #define PT_SC 4 /* Script (e.g. Han) */
1799 #define PT_ALNUM 5 /* Alphanumeric - the union of L and N */
1800 #define PT_SPACE 6 /* Perl space - Z plus 9,10,12,13 */
1801 #define PT_PXSPACE 7 /* POSIX space - Z plus 9,10,11,12,13 */
1802 #define PT_WORD 8 /* Word - L plus N plus underscore */
1803 #define PT_CLIST 9 /* Pseudo-property: match character list */
1804 #define PT_UCNC 10 /* Universal Character nameable character */
1805 #define PT_TABSIZE 11 /* Size of square table for autopossessify tests */
1806
1807 /* The following special properties are used only in XCLASS items, when POSIX
1808 classes are specified and PCRE_UCP is set - in other words, for Unicode
1809 handling of these classes. They are not available via the \p or \P escapes like
1810 those in the above list, and so they do not take part in the autopossessifying
1811 table. */
1812
1813 #define PT_PXGRAPH 11 /* [:graph:] - characters that mark the paper */
1814 #define PT_PXPRINT 12 /* [:print:] - [:graph:] plus non-control spaces */
1815 #define PT_PXPUNCT 13 /* [:punct:] - punctuation characters */
1816
1817 /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
1818 contain characters with values greater than 255. */
1819
1820 #define XCL_NOT 0x01 /* Flag: this is a negative class */
1821 #define XCL_MAP 0x02 /* Flag: a 32-byte map is present */
1822 #define XCL_HASPROP 0x04 /* Flag: property checks are present. */
1823
1824 #define XCL_END 0 /* Marks end of individual items */
1825 #define XCL_SINGLE 1 /* Single item (one multibyte char) follows */
1826 #define XCL_RANGE 2 /* A range (two multibyte chars) follows */
1827 #define XCL_PROP 3 /* Unicode property (2-byte property code follows) */
1828 #define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */
1829
1830 /* These are escaped items that aren't just an encoding of a particular data
1831 value such as \n. They must have non-zero values, as check_escape() returns 0
1832 for a data character. Also, they must appear in the same order as in the
1833 opcode definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it
1834 corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1835 used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In
1836 non-DOTALL mode, "." behaves like \N.
1837
1838 The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1839 when PCRE_UCP is set and replacement of \d etc by \p sequences is required.
1840 They must be contiguous, and remain in order so that the replacements can be
1841 looked up from a table.
1842
1843 Negative numbers are used to encode a backreference (\1, \2, \3, etc.) in
1844 check_escape(). There are two tests in the code for an escape
1845 greater than ESC_b and less than ESC_Z to detect the types that may be
1846 repeated. These are the types that consume characters. If any new escapes are
1847 put in between that don't consume a character, that code will have to change.
1848 */
1849
1850 enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
1851 ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1852 ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1853 ESC_E, ESC_Q, ESC_g, ESC_k,
1854 ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };
1855
1856
1857 /********************** Opcode definitions ******************/
1858
1859 /****** NOTE NOTE NOTE ******
1860
1861 Starting from 1 (i.e. after OP_END), the values up to OP_EOD must correspond in
1862 order to the list of escapes immediately above. Furthermore, values up to
1863 OP_DOLLM must not be changed without adjusting the table called autoposstab in
1864 pcre_compile.c
1865
1866 Whenever this list is updated, the two macro definitions that follow must be
1867 updated to match. The possessification table called "opcode_possessify" in
1868 pcre_compile.c must also be updated, and also the tables called "coptable"
1869 and "poptable" in pcre_dfa_exec.c.
1870
1871 ****** NOTE NOTE NOTE ******/
1872
1873
1874 /* The values between FIRST_AUTOTAB_OP and LAST_AUTOTAB_RIGHT_OP, inclusive,
1875 are used in a table for deciding whether a repeated character type can be
1876 auto-possessified. */
1877
1878 #define FIRST_AUTOTAB_OP OP_NOT_DIGIT
1879 #define LAST_AUTOTAB_LEFT_OP OP_EXTUNI
1880 #define LAST_AUTOTAB_RIGHT_OP OP_DOLLM
1881
1882 enum {
1883 OP_END, /* 0 End of pattern */
1884
1885 /* Values corresponding to backslashed metacharacters */
1886
1887 OP_SOD, /* 1 Start of data: \A */
1888 OP_SOM, /* 2 Start of match (subject + offset): \G */
1889 OP_SET_SOM, /* 3 Set start of match (\K) */
1890 OP_NOT_WORD_BOUNDARY, /* 4 \B */
1891 OP_WORD_BOUNDARY, /* 5 \b */
1892 OP_NOT_DIGIT, /* 6 \D */
1893 OP_DIGIT, /* 7 \d */
1894 OP_NOT_WHITESPACE, /* 8 \S */
1895 OP_WHITESPACE, /* 9 \s */
1896 OP_NOT_WORDCHAR, /* 10 \W */
1897 OP_WORDCHAR, /* 11 \w */
1898
1899 OP_ANY, /* 12 Match any character except newline (\N) */
1900 OP_ALLANY, /* 13 Match any character */
1901 OP_ANYBYTE, /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1902 OP_NOTPROP, /* 15 \P (not Unicode property) */
1903 OP_PROP, /* 16 \p (Unicode property) */
1904 OP_ANYNL, /* 17 \R (any newline sequence) */
1905 OP_NOT_HSPACE, /* 18 \H (not horizontal whitespace) */
1906 OP_HSPACE, /* 19 \h (horizontal whitespace) */
1907 OP_NOT_VSPACE, /* 20 \V (not vertical whitespace) */
1908 OP_VSPACE, /* 21 \v (vertical whitespace) */
1909 OP_EXTUNI, /* 22 \X (extended Unicode sequence */
1910 OP_EODN, /* 23 End of data or \n at end of data (\Z) */
1911 OP_EOD, /* 24 End of data (\z) */
1912
1913 /* Line end assertions */
1914
1915 OP_DOLL, /* 25 End of line - not multiline */
1916 OP_DOLLM, /* 26 End of line - multiline */
1917 OP_CIRC, /* 27 Start of line - not multiline */
1918 OP_CIRCM, /* 28 Start of line - multiline */
1919
1920 /* Single characters; caseful must precede the caseless ones */
1921
1922 OP_CHAR, /* 29 Match one character, casefully */
1923 OP_CHARI, /* 30 Match one character, caselessly */
1924 OP_NOT, /* 31 Match one character, not the given one, casefully */
1925 OP_NOTI, /* 32 Match one character, not the given one, caselessly */
1926
1927 /* The following sets of 13 opcodes must always be kept in step because
1928 the offset from the first one is used to generate the others. */
1929
1930 /* Repeated characters; caseful must precede the caseless ones */
1931
1932 OP_STAR, /* 33 The maximizing and minimizing versions of */
1933 OP_MINSTAR, /* 34 these six opcodes must come in pairs, with */
1934 OP_PLUS, /* 35 the minimizing one second. */
1935 OP_MINPLUS, /* 36 */
1936 OP_QUERY, /* 37 */
1937 OP_MINQUERY, /* 38 */
1938
1939 OP_UPTO, /* 39 From 0 to n matches of one character, caseful*/
1940 OP_MINUPTO, /* 40 */
1941 OP_EXACT, /* 41 Exactly n matches */
1942
1943 OP_POSSTAR, /* 42 Possessified star, caseful */
1944 OP_POSPLUS, /* 43 Possessified plus, caseful */
1945 OP_POSQUERY, /* 44 Posesssified query, caseful */
1946 OP_POSUPTO, /* 45 Possessified upto, caseful */
1947
1948 /* Repeated characters; caseless must follow the caseful ones */
1949
1950 OP_STARI, /* 46 */
1951 OP_MINSTARI, /* 47 */
1952 OP_PLUSI, /* 48 */
1953 OP_MINPLUSI, /* 49 */
1954 OP_QUERYI, /* 50 */
1955 OP_MINQUERYI, /* 51 */
1956
1957 OP_UPTOI, /* 52 From 0 to n matches of one character, caseless */
1958 OP_MINUPTOI, /* 53 */
1959 OP_EXACTI, /* 54 */
1960
1961 OP_POSSTARI, /* 55 Possessified star, caseless */
1962 OP_POSPLUSI, /* 56 Possessified plus, caseless */
1963 OP_POSQUERYI, /* 57 Posesssified query, caseless */
1964 OP_POSUPTOI, /* 58 Possessified upto, caseless */
1965
1966 /* The negated ones must follow the non-negated ones, and match them */
1967 /* Negated repeated character, caseful; must precede the caseless ones */
1968
1969 OP_NOTSTAR, /* 59 The maximizing and minimizing versions of */
1970 OP_NOTMINSTAR, /* 60 these six opcodes must come in pairs, with */
1971 OP_NOTPLUS, /* 61 the minimizing one second. They must be in */
1972 OP_NOTMINPLUS, /* 62 exactly the same order as those above. */
1973 OP_NOTQUERY, /* 63 */
1974 OP_NOTMINQUERY, /* 64 */
1975
1976 OP_NOTUPTO, /* 65 From 0 to n matches, caseful */
1977 OP_NOTMINUPTO, /* 66 */
1978 OP_NOTEXACT, /* 67 Exactly n matches */
1979
1980 OP_NOTPOSSTAR, /* 68 Possessified versions, caseful */
1981 OP_NOTPOSPLUS, /* 69 */
1982 OP_NOTPOSQUERY, /* 70 */
1983 OP_NOTPOSUPTO, /* 71 */
1984
1985 /* Negated repeated character, caseless; must follow the caseful ones */
1986
1987 OP_NOTSTARI, /* 72 */
1988 OP_NOTMINSTARI, /* 73 */
1989 OP_NOTPLUSI, /* 74 */
1990 OP_NOTMINPLUSI, /* 75 */
1991 OP_NOTQUERYI, /* 76 */
1992 OP_NOTMINQUERYI, /* 77 */
1993
1994 OP_NOTUPTOI, /* 78 From 0 to n matches, caseless */
1995 OP_NOTMINUPTOI, /* 79 */
1996 OP_NOTEXACTI, /* 80 Exactly n matches */
1997
1998 OP_NOTPOSSTARI, /* 81 Possessified versions, caseless */
1999 OP_NOTPOSPLUSI, /* 82 */
2000 OP_NOTPOSQUERYI, /* 83 */
2001 OP_NOTPOSUPTOI, /* 84 */
2002
2003 /* Character types */
2004
2005 OP_TYPESTAR, /* 85 The maximizing and minimizing versions of */
2006 OP_TYPEMINSTAR, /* 86 these six opcodes must come in pairs, with */
2007 OP_TYPEPLUS, /* 87 the minimizing one second. These codes must */
2008 OP_TYPEMINPLUS, /* 88 be in exactly the same order as those above. */
2009 OP_TYPEQUERY, /* 89 */
2010 OP_TYPEMINQUERY, /* 90 */
2011
2012 OP_TYPEUPTO, /* 91 From 0 to n matches */
2013 OP_TYPEMINUPTO, /* 92 */
2014 OP_TYPEEXACT, /* 93 Exactly n matches */
2015
2016 OP_TYPEPOSSTAR, /* 94 Possessified versions */
2017 OP_TYPEPOSPLUS, /* 95 */
2018 OP_TYPEPOSQUERY, /* 96 */
2019 OP_TYPEPOSUPTO, /* 97 */
2020
2021 /* These are used for character classes and back references; only the
2022 first six are the same as the sets above. */
2023
2024 OP_CRSTAR, /* 98 The maximizing and minimizing versions of */
2025 OP_CRMINSTAR, /* 99 all these opcodes must come in pairs, with */
2026 OP_CRPLUS, /* 100 the minimizing one second. These codes must */
2027 OP_CRMINPLUS, /* 101 be in exactly the same order as those above. */
2028 OP_CRQUERY, /* 102 */
2029 OP_CRMINQUERY, /* 103 */
2030
2031 OP_CRRANGE, /* 104 These are different to the three sets above. */
2032 OP_CRMINRANGE, /* 105 */
2033
2034 OP_CRPOSSTAR, /* 106 Possessified versions */
2035 OP_CRPOSPLUS, /* 107 */
2036 OP_CRPOSQUERY, /* 108 */
2037 OP_CRPOSRANGE, /* 109 */
2038
2039 /* End of quantifier opcodes */
2040
2041 OP_CLASS, /* 110 Match a character class, chars < 256 only */
2042 OP_NCLASS, /* 111 Same, but the bitmap was created from a negative
2043 class - the difference is relevant only when a
2044 character > 255 is encountered. */
2045 OP_XCLASS, /* 112 Extended class for handling > 255 chars within the
2046 class. This does both positive and negative. */
2047 OP_REF, /* 113 Match a back reference, casefully */
2048 OP_REFI, /* 114 Match a back reference, caselessly */
2049 OP_DNREF, /* 115 Match a duplicate name backref, casefully */
2050 OP_DNREFI, /* 116 Match a duplicate name backref, caselessly */
2051 OP_RECURSE, /* 117 Match a numbered subpattern (possibly recursive) */
2052 OP_CALLOUT, /* 118 Call out to external function if provided */
2053
2054 OP_ALT, /* 119 Start of alternation */
2055 OP_KET, /* 120 End of group that doesn't have an unbounded repeat */
2056 OP_KETRMAX, /* 121 These two must remain together and in this */
2057 OP_KETRMIN, /* 122 order. They are for groups the repeat for ever. */
2058 OP_KETRPOS, /* 123 Possessive unlimited repeat. */
2059
2060 /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
2061 asserts must remain in order. */
2062
2063 OP_REVERSE, /* 124 Move pointer back - used in lookbehind assertions */
2064 OP_ASSERT, /* 125 Positive lookahead */
2065 OP_ASSERT_NOT, /* 126 Negative lookahead */
2066 OP_ASSERTBACK, /* 127 Positive lookbehind */
2067 OP_ASSERTBACK_NOT, /* 128 Negative lookbehind */
2068
2069 /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
2070 after the assertions, with ONCE first, as there's a test for >= ONCE for a
2071 subpattern that isn't an assertion. The POS versions must immediately follow
2072 the non-POS versions in each case. */
2073
2074 OP_ONCE, /* 129 Atomic group, contains captures */
2075 OP_ONCE_NC, /* 130 Atomic group containing no captures */
2076 OP_BRA, /* 131 Start of non-capturing bracket */
2077 OP_BRAPOS, /* 132 Ditto, with unlimited, possessive repeat */
2078 OP_CBRA, /* 133 Start of capturing bracket */
2079 OP_CBRAPOS, /* 134 Ditto, with unlimited, possessive repeat */
2080 OP_COND, /* 135 Conditional group */
2081
2082 /* These five must follow the previous five, in the same order. There's a
2083 check for >= SBRA to distinguish the two sets. */
2084
2085 OP_SBRA, /* 136 Start of non-capturing bracket, check empty */
2086 OP_SBRAPOS, /* 137 Ditto, with unlimited, possessive repeat */
2087 OP_SCBRA, /* 138 Start of capturing bracket, check empty */
2088 OP_SCBRAPOS, /* 139 Ditto, with unlimited, possessive repeat */
2089 OP_SCOND, /* 140 Conditional group, check empty */
2090
2091 /* The next two pairs must (respectively) be kept together. */
2092
2093 OP_CREF, /* 141 Used to hold a capture number as condition */
2094 OP_DNCREF, /* 142 Used to point to duplicate names as a condition */
2095 OP_RREF, /* 143 Used to hold a recursion number as condition */
2096 OP_DNRREF, /* 144 Used to point to duplicate names as a condition */
2097 OP_DEF, /* 145 The DEFINE condition */
2098
2099 OP_BRAZERO, /* 146 These two must remain together and in this */
2100 OP_BRAMINZERO, /* 147 order. */
2101 OP_BRAPOSZERO, /* 148 */
2102
2103 /* These are backtracking control verbs */
2104
2105 OP_MARK, /* 149 always has an argument */
2106 OP_PRUNE, /* 150 */
2107 OP_PRUNE_ARG, /* 151 same, but with argument */
2108 OP_SKIP, /* 152 */
2109 OP_SKIP_ARG, /* 153 same, but with argument */
2110 OP_THEN, /* 154 */
2111 OP_THEN_ARG, /* 155 same, but with argument */
2112 OP_COMMIT, /* 156 */
2113
2114 /* These are forced failure and success verbs */
2115
2116 OP_FAIL, /* 157 */
2117 OP_ACCEPT, /* 158 */
2118 OP_ASSERT_ACCEPT, /* 159 Used inside assertions */
2119 OP_CLOSE, /* 160 Used before OP_ACCEPT to close open captures */
2120
2121 /* This is used to skip a subpattern with a {0} quantifier */
2122
2123 OP_SKIPZERO, /* 161 */
2124
2125 /* This is not an opcode, but is used to check that tables indexed by opcode
2126 are the correct length, in order to catch updating errors - there have been
2127 some in the past. */
2128
2129 OP_TABLE_LENGTH
2130 };
2131
2132 /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
2133 definitions that follow must also be updated to match. There are also tables
2134 called "opcode_possessify" in pcre_compile.c and "coptable" and "poptable" in
2135 pcre_dfa_exec.c that must be updated. */
2136
2137
2138 /* This macro defines textual names for all the opcodes. These are used only
2139 for debugging, and some of them are only partial names. The macro is referenced
2140 only in pcre_printint.c, which fills out the full names in many cases (and in
2141 some cases doesn't actually use these names at all). */
2142
2143 #define OP_NAME_LIST \
2144 "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d", \
2145 "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte", \
2146 "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v", \
2147 "extuni", "\\Z", "\\z", \
2148 "$", "$", "^", "^", "char", "chari", "not", "noti", \
2149 "*", "*?", "+", "+?", "?", "??", \
2150 "{", "{", "{", \
2151 "*+","++", "?+", "{", \
2152 "*", "*?", "+", "+?", "?", "??", \
2153 "{", "{", "{", \
2154 "*+","++", "?+", "{", \
2155 "*", "*?", "+", "+?", "?", "??", \
2156 "{", "{", "{", \
2157 "*+","++", "?+", "{", \
2158 "*", "*?", "+", "+?", "?", "??", \
2159 "{", "{", "{", \
2160 "*+","++", "?+", "{", \
2161 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
2162 "*+","++", "?+", "{", \
2163 "*", "*?", "+", "+?", "?", "??", "{", "{", \
2164 "*+","++", "?+", "{", \
2165 "class", "nclass", "xclass", "Ref", "Refi", "DnRef", "DnRefi", \
2166 "Recurse", "Callout", \
2167 "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos", \
2168 "Reverse", "Assert", "Assert not", "AssertB", "AssertB not", \
2169 "Once", "Once_NC", \
2170 "Bra", "BraPos", "CBra", "CBraPos", \
2171 "Cond", \
2172 "SBra", "SBraPos", "SCBra", "SCBraPos", \
2173 "SCond", \
2174 "Cond ref", "Cond dnref", "Cond rec", "Cond dnrec", "Cond def", \
2175 "Brazero", "Braminzero", "Braposzero", \
2176 "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP", \
2177 "*THEN", "*THEN", "*COMMIT", "*FAIL", \
2178 "*ACCEPT", "*ASSERT_ACCEPT", \
2179 "Close", "Skip zero"
2180
2181
2182 /* This macro defines the length of fixed length operations in the compiled
2183 regex. The lengths are used when searching for specific things, and also in the
2184 debugging printing of a compiled regex. We use a macro so that it can be
2185 defined close to the definitions of the opcodes themselves.
2186
2187 As things have been extended, some of these are no longer fixed lenths, but are
2188 minima instead. For example, the length of a single-character repeat may vary
2189 in UTF-8 mode. The code that uses this table must know about such things. */
2190
2191 #define OP_LENGTHS \
2192 1, /* End */ \
2193 1, 1, 1, 1, 1, /* \A, \G, \K, \B, \b */ \
2194 1, 1, 1, 1, 1, 1, /* \D, \d, \S, \s, \W, \w */ \
2195 1, 1, 1, /* Any, AllAny, Anybyte */ \
2196 3, 3, /* \P, \p */ \
2197 1, 1, 1, 1, 1, /* \R, \H, \h, \V, \v */ \
2198 1, /* \X */ \
2199 1, 1, 1, 1, 1, 1, /* \Z, \z, $, $M ^, ^M */ \
2200 2, /* Char - the minimum length */ \
2201 2, /* Chari - the minimum length */ \
2202 2, /* not */ \
2203 2, /* noti */ \
2204 /* Positive single-char repeats ** These are */ \
2205 2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** minima in */ \
2206 2+IMM2_SIZE, 2+IMM2_SIZE, /* upto, minupto ** mode */ \
2207 2+IMM2_SIZE, /* exact */ \
2208 2, 2, 2, 2+IMM2_SIZE, /* *+, ++, ?+, upto+ */ \
2209 2, 2, 2, 2, 2, 2, /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8 */ \
2210 2+IMM2_SIZE, 2+IMM2_SIZE, /* upto I, minupto I */ \
2211 2+IMM2_SIZE, /* exact I */ \
2212 2, 2, 2, 2+IMM2_SIZE, /* *+I, ++I, ?+I, upto+I */ \
2213 /* Negative single-char repeats - only for chars < 256 */ \
2214 2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \
2215 2+IMM2_SIZE, 2+IMM2_SIZE, /* NOT upto, minupto */ \
2216 2+IMM2_SIZE, /* NOT exact */ \
2217 2, 2, 2, 2+IMM2_SIZE, /* Possessive NOT *, +, ?, upto */ \
2218 2, 2, 2, 2, 2, 2, /* NOT *I, *?I, +I, +?I, ?I, ??I */ \
2219 2+IMM2_SIZE, 2+IMM2_SIZE, /* NOT upto I, minupto I */ \
2220 2+IMM2_SIZE, /* NOT exact I */ \
2221 2, 2, 2, 2+IMM2_SIZE, /* Possessive NOT *I, +I, ?I, upto I */ \
2222 /* Positive type repeats */ \
2223 2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \
2224 2+IMM2_SIZE, 2+IMM2_SIZE, /* Type upto, minupto */ \
2225 2+IMM2_SIZE, /* Type exact */ \
2226 2, 2, 2, 2+IMM2_SIZE, /* Possessive *+, ++, ?+, upto+ */ \
2227 /* Character class & ref repeats */ \
2228 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \
2229 1+2*IMM2_SIZE, 1+2*IMM2_SIZE, /* CRRANGE, CRMINRANGE */ \
2230 1, 1, 1, 1+2*IMM2_SIZE, /* Possessive *+, ++, ?+, CRPOSRANGE */ \
2231 1+(32/sizeof(pcre_uchar)), /* CLASS */ \
2232 1+(32/sizeof(pcre_uchar)), /* NCLASS */ \
2233 0, /* XCLASS - variable length */ \
2234 1+IMM2_SIZE, /* REF */ \
2235 1+IMM2_SIZE, /* REFI */ \
2236 1+2*IMM2_SIZE, /* DNREF */ \
2237 1+2*IMM2_SIZE, /* DNREFI */ \
2238 1+LINK_SIZE, /* RECURSE */ \
2239 2+2*LINK_SIZE, /* CALLOUT */ \
2240 1+LINK_SIZE, /* Alt */ \
2241 1+LINK_SIZE, /* Ket */ \
2242 1+LINK_SIZE, /* KetRmax */ \
2243 1+LINK_SIZE, /* KetRmin */ \
2244 1+LINK_SIZE, /* KetRpos */ \
2245 1+LINK_SIZE, /* Reverse */ \
2246 1+LINK_SIZE, /* Assert */ \
2247 1+LINK_SIZE, /* Assert not */ \
2248 1+LINK_SIZE, /* Assert behind */ \
2249 1+LINK_SIZE, /* Assert behind not */ \
2250 1+LINK_SIZE, /* ONCE */ \
2251 1+LINK_SIZE, /* ONCE_NC */ \
2252 1+LINK_SIZE, /* BRA */ \
2253 1+LINK_SIZE, /* BRAPOS */ \
2254 1+LINK_SIZE+IMM2_SIZE, /* CBRA */ \
2255 1+LINK_SIZE+IMM2_SIZE, /* CBRAPOS */ \
2256 1+LINK_SIZE, /* COND */ \
2257 1+LINK_SIZE, /* SBRA */ \
2258 1+LINK_SIZE, /* SBRAPOS */ \
2259 1+LINK_SIZE+IMM2_SIZE, /* SCBRA */ \
2260 1+LINK_SIZE+IMM2_SIZE, /* SCBRAPOS */ \
2261 1+LINK_SIZE, /* SCOND */ \
2262 1+IMM2_SIZE, 1+2*IMM2_SIZE, /* CREF, DNCREF */ \
2263 1+IMM2_SIZE, 1+2*IMM2_SIZE, /* RREF, DNRREF */ \
2264 1, /* DEF */ \
2265 1, 1, 1, /* BRAZERO, BRAMINZERO, BRAPOSZERO */ \
2266 3, 1, 3, /* MARK, PRUNE, PRUNE_ARG */ \
2267 1, 3, /* SKIP, SKIP_ARG */ \
2268 1, 3, /* THEN, THEN_ARG */ \
2269 1, 1, 1, 1, /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT */ \
2270 1+IMM2_SIZE, 1 /* CLOSE, SKIPZERO */
2271
2272 /* A magic value for OP_RREF to indicate the "any recursion" condition. */
2273
2274 #define RREF_ANY 0xffff
2275
2276 /* Compile time error code numbers. They are given names so that they can more
2277 easily be tracked. When a new number is added, the table called eint in
2278 pcreposix.c must be updated. */
2279
2280 enum { ERR0, ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9,
2281 ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
2282 ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
2283 ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
2284 ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
2285 ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
2286 ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
2287 ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERR79,
2288 ERR80, ERR81, ERR82, ERR83, ERR84, ERR85, ERR86, ERRCOUNT };
2289
2290 /* JIT compiling modes. The function list is indexed by them. */
2291
2292 enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
2293 JIT_NUMBER_OF_COMPILE_MODES };
2294
2295 /* The real format of the start of the pcre block; the index of names and the
2296 code vector run on as long as necessary after the end. We store an explicit
2297 offset to the name table so that if a regex is compiled on one host, saved, and
2298 then run on another where the size of pointers is different, all might still
2299 be well.
2300
2301 The size of the structure must be a multiple of 8 bytes. For the case of
2302 compiled-on-4 and run-on-8, we include an extra pointer that is always NULL so
2303 that there are an even number of pointers which therefore are a multiple of 8
2304 bytes.
2305
2306 It is necessary to fork the struct for the 32 bit library, since it needs to
2307 use pcre_uint32 for first_char and req_char. We can't put an ifdef inside the
2308 typedef because pcretest needs access to the struct of the 8-, 16- and 32-bit
2309 variants.
2310
2311 *** WARNING ***
2312 When new fields are added to these structures, remember to adjust the code in
2313 pcre_byte_order.c that is concerned with swapping the byte order of the fields
2314 when a compiled regex is reloaded on a host with different endianness.
2315 *** WARNING ***
2316 There is also similar byte-flipping code in pcretest.c, which is used for
2317 testing the byte-flipping features. It must also be kept in step.
2318 *** WARNING ***
2319 */
2320
2321 typedef struct real_pcre8_or_16 {
2322 pcre_uint32 magic_number;
2323 pcre_uint32 size; /* Total that was malloced */
2324 pcre_uint32 options; /* Public options */
2325 pcre_uint32 flags; /* Private flags */
2326 pcre_uint32 limit_match; /* Limit set from regex */
2327 pcre_uint32 limit_recursion; /* Limit set from regex */
2328 pcre_uint16 first_char; /* Starting character */
2329 pcre_uint16 req_char; /* This character must be seen */
2330 pcre_uint16 max_lookbehind; /* Longest lookbehind (characters) */
2331 pcre_uint16 top_bracket; /* Highest numbered group */
2332 pcre_uint16 top_backref; /* Highest numbered back reference */
2333 pcre_uint16 name_table_offset; /* Offset to name table that follows */
2334 pcre_uint16 name_entry_size; /* Size of any name items */
2335 pcre_uint16 name_count; /* Number of name items */
2336 pcre_uint16 ref_count; /* Reference count */
2337 pcre_uint16 dummy1; /* To ensure size is a multiple of 8 */
2338 pcre_uint16 dummy2; /* To ensure size is a multiple of 8 */
2339 pcre_uint16 dummy3; /* To ensure size is a multiple of 8 */
2340 const pcre_uint8 *tables; /* Pointer to tables or NULL for std */
2341 void *nullpad; /* NULL padding */
2342 } real_pcre8_or_16;
2343
2344 typedef struct real_pcre8_or_16 real_pcre;
2345 typedef struct real_pcre8_or_16 real_pcre16;
2346
2347 typedef struct real_pcre32 {
2348 pcre_uint32 magic_number;
2349 pcre_uint32 size; /* Total that was malloced */
2350 pcre_uint32 options; /* Public options */
2351 pcre_uint32 flags; /* Private flags */
2352 pcre_uint32 limit_match; /* Limit set from regex */
2353 pcre_uint32 limit_recursion; /* Limit set from regex */
2354 pcre_uint32 first_char; /* Starting character */
2355 pcre_uint32 req_char; /* This character must be seen */
2356 pcre_uint16 max_lookbehind; /* Longest lookbehind (characters) */
2357 pcre_uint16 top_bracket; /* Highest numbered group */
2358 pcre_uint16 top_backref; /* Highest numbered back reference */
2359 pcre_uint16 name_table_offset; /* Offset to name table that follows */
2360 pcre_uint16 name_entry_size; /* Size of any name items */
2361 pcre_uint16 name_count; /* Number of name items */
2362 pcre_uint16 ref_count; /* Reference count */
2363 pcre_uint16 dummy; /* To ensure size is a multiple of 8 */
2364 const pcre_uint8 *tables; /* Pointer to tables or NULL for std */
2365 void *nullpad; /* NULL padding */
2366 } real_pcre32;
2367
2368 #if defined COMPILE_PCRE8
2369 #define REAL_PCRE real_pcre
2370 #elif defined COMPILE_PCRE16
2371 #define REAL_PCRE real_pcre16
2372 #elif defined COMPILE_PCRE32
2373 #define REAL_PCRE real_pcre32
2374 #endif
2375
2376 /* Assert that the size of REAL_PCRE is divisible by 8 */
2377 typedef int __assert_real_pcre_size_divisible_8[(sizeof(REAL_PCRE) % 8) == 0 ? 1 : -1];
2378
2379 /* Needed in pcretest to access some fields in the real_pcre* structures
2380 * directly. They're unified for 8/16/32 bits since the structs only differ
2381 * after these fields; if that ever changes, need to fork those defines into
2382 * 8/16 and 32 bit versions. */
2383 #define REAL_PCRE_MAGIC(re) (((REAL_PCRE*)re)->magic_number)
2384 #define REAL_PCRE_SIZE(re) (((REAL_PCRE*)re)->size)
2385 #define REAL_PCRE_OPTIONS(re) (((REAL_PCRE*)re)->options)
2386 #define REAL_PCRE_FLAGS(re) (((REAL_PCRE*)re)->flags)
2387
2388 /* The format of the block used to store data from pcre_study(). The same
2389 remark (see NOTE above) about extending this structure applies. */
2390
2391 typedef struct pcre_study_data {
2392 pcre_uint32 size; /* Total that was malloced */
2393 pcre_uint32 flags; /* Private flags */
2394 pcre_uint8 start_bits[32]; /* Starting char bits */
2395 pcre_uint32 minlength; /* Minimum subject length */
2396 } pcre_study_data;
2397
2398 /* Structure for building a chain of open capturing subpatterns during
2399 compiling, so that instructions to close them can be compiled when (*ACCEPT) is
2400 encountered. This is also used to identify subpatterns that contain recursive
2401 back references to themselves, so that they can be made atomic. */
2402
2403 typedef struct open_capitem {
2404 struct open_capitem *next; /* Chain link */
2405 pcre_uint16 number; /* Capture number */
2406 pcre_uint16 flag; /* Set TRUE if recursive back ref */
2407 } open_capitem;
2408
2409 /* Structure for building a list of named groups during the first pass of
2410 compiling. */
2411
2412 typedef struct named_group {
2413 const pcre_uchar *name; /* Points to the name in the pattern */
2414 int length; /* Length of the name */
2415 pcre_uint32 number; /* Group number */
2416 } named_group;
2417
2418 /* Structure for passing "static" information around between the functions
2419 doing the compiling, so that they are thread-safe. */
2420
2421 typedef struct compile_data {
2422 const pcre_uint8 *lcc; /* Points to lower casing table */
2423 const pcre_uint8 *fcc; /* Points to case-flipping table */
2424 const pcre_uint8 *cbits; /* Points to character type table */
2425 const pcre_uint8 *ctypes; /* Points to table of type maps */
2426 const pcre_uchar *start_workspace;/* The start of working space */
2427 const pcre_uchar *start_code; /* The start of the compiled code */
2428 const pcre_uchar *start_pattern; /* The start of the pattern */
2429 const pcre_uchar *end_pattern; /* The end of the pattern */
2430 pcre_uchar *hwm; /* High watermark of workspace */
2431 open_capitem *open_caps; /* Chain of open capture items */
2432 named_group *named_groups; /* Points to vector in pre-compile */
2433 pcre_uchar *name_table; /* The name/number table */
2434 int names_found; /* Number of entries so far */
2435 int name_entry_size; /* Size of each entry */
2436 int named_group_list_size; /* Number of entries in the list */
2437 int workspace_size; /* Size of workspace */
2438 unsigned int bracount; /* Count of capturing parens as we compile */
2439 int final_bracount; /* Saved value after first pass */
2440 int max_lookbehind; /* Maximum lookbehind (characters) */
2441 int top_backref; /* Maximum back reference */
2442 unsigned int backref_map; /* Bitmap of low back refs */
2443 unsigned int namedrefcount; /* Number of backreferences by name */
2444 int parens_depth; /* Depth of nested parentheses */
2445 int assert_depth; /* Depth of nested assertions */
2446 pcre_uint32 external_options; /* External (initial) options */
2447 pcre_uint32 external_flags; /* External flag bits to be set */
2448 int req_varyopt; /* "After variable item" flag for reqbyte */
2449 BOOL had_accept; /* (*ACCEPT) encountered */
2450 BOOL had_pruneorskip; /* (*PRUNE) or (*SKIP) encountered */
2451 BOOL check_lookbehind; /* Lookbehinds need later checking */
2452 BOOL dupnames; /* Duplicate names exist */
2453 BOOL iscondassert; /* Next assert is a condition */
2454 int nltype; /* Newline type */
2455 int nllen; /* Newline string length */
2456 pcre_uchar nl[4]; /* Newline string when fixed length */
2457 } compile_data;
2458
2459 /* Structure for maintaining a chain of pointers to the currently incomplete
2460 branches, for testing for left recursion while compiling. */
2461
2462 typedef struct branch_chain {
2463 struct branch_chain *outer;
2464 pcre_uchar *current_branch;
2465 } branch_chain;
2466
2467 /* Structure for mutual recursion detection. */
2468
2469 typedef struct recurse_check {
2470 struct recurse_check *prev;
2471 const pcre_uchar *group;
2472 } recurse_check;
2473
2474 /* Structure for items in a linked list that represents an explicit recursive
2475 call within the pattern; used by pcre_exec(). */
2476
2477 typedef struct recursion_info {
2478 struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
2479 unsigned int group_num; /* Number of group that was called */
2480 int *offset_save; /* Pointer to start of saved offsets */
2481 int saved_max; /* Number of saved offsets */
2482 int saved_capture_last; /* Last capture number */
2483 PCRE_PUCHAR subject_position; /* Position at start of recursion */
2484 } recursion_info;
2485
2486 /* A similar structure for pcre_dfa_exec(). */
2487
2488 typedef struct dfa_recursion_info {
2489 struct dfa_recursion_info *prevrec;
2490 int group_num;
2491 PCRE_PUCHAR subject_position;
2492 } dfa_recursion_info;
2493
2494 /* Structure for building a chain of data for holding the values of the subject
2495 pointer at the start of each subpattern, so as to detect when an empty string
2496 has been matched by a subpattern - to break infinite loops; used by
2497 pcre_exec(). */
2498
2499 typedef struct eptrblock {
2500 struct eptrblock *epb_prev;
2501 PCRE_PUCHAR epb_saved_eptr;
2502 } eptrblock;
2503
2504
2505 /* Structure for passing "static" information around between the functions
2506 doing traditional NFA matching, so that they are thread-safe. */
2507
2508 typedef struct match_data {
2509 unsigned long int match_call_count; /* As it says */
2510 unsigned long int match_limit; /* As it says */
2511 unsigned long int match_limit_recursion; /* As it says */
2512 int *offset_vector; /* Offset vector */
2513 int offset_end; /* One past the end */
2514 int offset_max; /* The maximum usable for return data */
2515 int nltype; /* Newline type */
2516 int nllen; /* Newline string length */
2517 int name_count; /* Number of names in name table */
2518 int name_entry_size; /* Size of entry in names table */
2519 unsigned int skip_arg_count; /* For counting SKIP_ARGs */
2520 unsigned int ignore_skip_arg; /* For re-run when SKIP arg name not found */
2521 pcre_uchar *name_table; /* Table of names */
2522 pcre_uchar nl[4]; /* Newline string when fixed */
2523 const pcre_uint8 *lcc; /* Points to lower casing table */
2524 const pcre_uint8 *fcc; /* Points to case-flipping table */
2525 const pcre_uint8 *ctypes; /* Points to table of type maps */
2526 BOOL notbol; /* NOTBOL flag */
2527 BOOL noteol; /* NOTEOL flag */
2528 BOOL utf; /* UTF-8 / UTF-16 flag */
2529 BOOL jscript_compat; /* JAVASCRIPT_COMPAT flag */
2530 BOOL use_ucp; /* PCRE_UCP flag */
2531 BOOL endonly; /* Dollar not before final \n */
2532 BOOL notempty; /* Empty string match not wanted */
2533 BOOL notempty_atstart; /* Empty string match at start not wanted */
2534 BOOL hitend; /* Hit the end of the subject at some point */
2535 BOOL bsr_anycrlf; /* \R is just any CRLF, not full Unicode */
2536 BOOL hasthen; /* Pattern contains (*THEN) */
2537 const pcre_uchar *start_code; /* For use when recursing */
2538 PCRE_PUCHAR start_subject; /* Start of the subject string */
2539 PCRE_PUCHAR end_subject; /* End of the subject string */
2540 PCRE_PUCHAR start_match_ptr; /* Start of matched string */
2541 PCRE_PUCHAR end_match_ptr; /* Subject position at end match */
2542 PCRE_PUCHAR start_used_ptr; /* Earliest consulted character */
2543 int partial; /* PARTIAL options */
2544 int end_offset_top; /* Highwater mark at end of match */
2545 pcre_int32 capture_last; /* Most recent capture number + overflow flag */
2546 int start_offset; /* The start offset value */
2547 int match_function_type; /* Set for certain special calls of MATCH() */
2548 eptrblock *eptrchain; /* Chain of eptrblocks for tail recursions */
2549 int eptrn; /* Next free eptrblock */
2550 recursion_info *recursive; /* Linked list of recursion data */
2551 void *callout_data; /* To pass back to callouts */
2552 const pcre_uchar *mark; /* Mark pointer to pass back on success */
2553 const pcre_uchar *nomatch_mark;/* Mark pointer to pass back on failure */
2554 const pcre_uchar *once_target; /* Where to back up to for atomic groups */
2555 #ifdef NO_RECURSE
2556 void *match_frames_base; /* For remembering malloc'd frames */
2557 #endif
2558 } match_data;
2559
2560 /* A similar structure is used for the same purpose by the DFA matching
2561 functions. */
2562
2563 typedef struct dfa_match_data {
2564 const pcre_uchar *start_code; /* Start of the compiled pattern */
2565 const pcre_uchar *start_subject ; /* Start of the subject string */
2566 const pcre_uchar *end_subject; /* End of subject string */
2567 const pcre_uchar *start_used_ptr; /* Earliest consulted character */
2568 const pcre_uint8 *tables; /* Character tables */
2569 int start_offset; /* The start offset value */
2570 int moptions; /* Match options */
2571 int poptions; /* Pattern options */
2572 int nltype; /* Newline type */
2573 int nllen; /* Newline string length */
2574 pcre_uchar nl[4]; /* Newline string when fixed */
2575 void *callout_data; /* To pass back to callouts */
2576 dfa_recursion_info *recursive; /* Linked list of recursion data */
2577 } dfa_match_data;
2578
2579 /* Bit definitions for entries in the pcre_ctypes table. */
2580
2581 #define ctype_space 0x01
2582 #define ctype_letter 0x02
2583 #define ctype_digit 0x04
2584 #define ctype_xdigit 0x08
2585 #define ctype_word 0x10 /* alphanumeric or '_' */
2586 #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
2587
2588 /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
2589 of bits for a class map. Some classes are built by combining these tables. */
2590
2591 #define cbit_space 0 /* [:space:] or \s */
2592 #define cbit_xdigit 32 /* [:xdigit:] */
2593 #define cbit_digit 64 /* [:digit:] or \d */
2594 #define cbit_upper 96 /* [:upper:] */
2595 #define cbit_lower 128 /* [:lower:] */
2596 #define cbit_word 160 /* [:word:] or \w */
2597 #define cbit_graph 192 /* [:graph:] */
2598 #define cbit_print 224 /* [:print:] */
2599 #define cbit_punct 256 /* [:punct:] */
2600 #define cbit_cntrl 288 /* [:cntrl:] */
2601 #define cbit_length 320 /* Length of the cbits table */
2602
2603 /* Offsets of the various tables from the base tables pointer, and
2604 total length. */
2605
2606 #define lcc_offset 0
2607 #define fcc_offset 256
2608 #define cbits_offset 512
2609 #define ctypes_offset (cbits_offset + cbit_length)
2610 #define tables_length (ctypes_offset + 256)
2611
2612 /* Internal function and data prefixes. */
2613
2614 #if defined COMPILE_PCRE8
2615 #ifndef PUBL
2616 #define PUBL(name) pcre_##name
2617 #endif
2618 #ifndef PRIV
2619 #define PRIV(name) _pcre_##name
2620 #endif
2621 #elif defined COMPILE_PCRE16
2622 #ifndef PUBL
2623 #define PUBL(name) pcre16_##name
2624 #endif
2625 #ifndef PRIV
2626 #define PRIV(name) _pcre16_##name
2627 #endif
2628 #elif defined COMPILE_PCRE32
2629 #ifndef PUBL
2630 #define PUBL(name) pcre32_##name
2631 #endif
2632 #ifndef PRIV
2633 #define PRIV(name) _pcre32_##name
2634 #endif
2635 #else
2636 #error Unsupported compiling mode
2637 #endif /* COMPILE_PCRE[8|16|32] */
2638
2639 /* Layout of the UCP type table that translates property names into types and
2640 codes. Each entry used to point directly to a name, but to reduce the number of
2641 relocations in shared libraries, it now has an offset into a single string
2642 instead. */
2643
2644 typedef struct {
2645 pcre_uint16 name_offset;
2646 pcre_uint16 type;
2647 pcre_uint16 value;
2648 } ucp_type_table;
2649
2650
2651 /* Internal shared data tables. These are tables that are used by more than one
2652 of the exported public functions. They have to be "external" in the C sense,
2653 but are not part of the PCRE public API. The data for these tables is in the
2654 pcre_tables.c module. */
2655
2656 #ifdef COMPILE_PCRE8
2657 extern const int PRIV(utf8_table1)[];
2658 extern const int PRIV(utf8_table1_size);
2659 extern const int PRIV(utf8_table2)[];
2660 extern const int PRIV(utf8_table3)[];
2661 extern const pcre_uint8 PRIV(utf8_table4)[];
2662 #endif /* COMPILE_PCRE8 */
2663
2664 extern const char PRIV(utt_names)[];
2665 extern const ucp_type_table PRIV(utt)[];
2666 extern const int PRIV(utt_size);
2667
2668 extern const pcre_uint8 PRIV(OP_lengths)[];
2669 extern const pcre_uint8 PRIV(default_tables)[];
2670
2671 extern const pcre_uint32 PRIV(hspace_list)[];
2672 extern const pcre_uint32 PRIV(vspace_list)[];
2673
2674
2675 /* Internal shared functions. These are functions that are used by more than
2676 one of the exported public functions. They have to be "external" in the C
2677 sense, but are not part of the PCRE public API. */
2678
2679 /* String comparison functions. */
2680 #if defined COMPILE_PCRE8
2681
2682 #define STRCMP_UC_UC(str1, str2) \
2683 strcmp((char *)(str1), (char *)(str2))
2684 #define STRCMP_UC_C8(str1, str2) \
2685 strcmp((char *)(str1), (str2))
2686 #define STRNCMP_UC_UC(str1, str2, num) \
2687 strncmp((char *)(str1), (char *)(str2), (num))
2688 #define STRNCMP_UC_C8(str1, str2, num) \
2689 strncmp((char *)(str1), (str2), (num))
2690 #define STRLEN_UC(str) strlen((const char *)str)
2691
2692 #elif defined COMPILE_PCRE16 || defined COMPILE_PCRE32
2693
2694 extern int PRIV(strcmp_uc_uc)(const pcre_uchar *,
2695 const pcre_uchar *);
2696 extern int PRIV(strcmp_uc_c8)(const pcre_uchar *,
2697 const char *);
2698 extern int PRIV(strncmp_uc_uc)(const pcre_uchar *,
2699 const pcre_uchar *, unsigned int num);
2700 extern int PRIV(strncmp_uc_c8)(const pcre_uchar *,
2701 const char *, unsigned int num);
2702 extern unsigned int PRIV(strlen_uc)(const pcre_uchar *str);
2703
2704 #define STRCMP_UC_UC(str1, str2) \
2705 PRIV(strcmp_uc_uc)((str1), (str2))
2706 #define STRCMP_UC_C8(str1, str2) \
2707 PRIV(strcmp_uc_c8)((str1), (str2))
2708 #define STRNCMP_UC_UC(str1, str2, num) \
2709 PRIV(strncmp_uc_uc)((str1), (str2), (num))
2710 #define STRNCMP_UC_C8(str1, str2, num) \
2711 PRIV(strncmp_uc_c8)((str1), (str2), (num))
2712 #define STRLEN_UC(str) PRIV(strlen_uc)(str)
2713
2714 #endif /* COMPILE_PCRE[8|16|32] */
2715
2716 #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
2717
2718 #define STRCMP_UC_UC_TEST(str1, str2) STRCMP_UC_UC(str1, str2)
2719 #define STRCMP_UC_C8_TEST(str1, str2) STRCMP_UC_C8(str1, str2)
2720
2721 #elif defined COMPILE_PCRE32
2722
2723 extern int PRIV(strcmp_uc_uc_utf)(const pcre_uchar *,
2724 const pcre_uchar *);
2725 extern int PRIV(strcmp_uc_c8_utf)(const pcre_uchar *,
2726 const char *);
2727
2728 #define STRCMP_UC_UC_TEST(str1, str2) \
2729 (utf ? PRIV(strcmp_uc_uc_utf)((str1), (str2)) : PRIV(strcmp_uc_uc)((str1), (str2)))
2730 #define STRCMP_UC_C8_TEST(str1, str2) \
2731 (utf ? PRIV(strcmp_uc_c8_utf)((str1), (str2)) : PRIV(strcmp_uc_c8)((str1), (str2)))
2732
2733 #endif /* COMPILE_PCRE[8|16|32] */
2734
2735 extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2736 extern BOOL PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2737 int *, BOOL);
2738 extern unsigned int PRIV(ord2utf)(pcre_uint32, pcre_uchar *);
2739 extern int PRIV(valid_utf)(PCRE_PUCHAR, int, int *);
2740 extern BOOL PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2741 int *, BOOL);
2742 extern BOOL PRIV(xclass)(pcre_uint32, const pcre_uchar *, BOOL);
2743
2744 #ifdef SUPPORT_JIT
2745 extern void PRIV(jit_compile)(const REAL_PCRE *,
2746 PUBL(extra) *, int);
2747 extern int PRIV(jit_exec)(const PUBL(extra) *,
2748 const pcre_uchar *, int, int, int, int *, int);
2749 extern void PRIV(jit_free)(void *);
2750 extern int PRIV(jit_get_size)(void *);
2751 extern const char* PRIV(jit_get_target)(void);
2752 #endif
2753
2754 /* Unicode character database (UCD) */
2755
2756 typedef struct {
2757 pcre_uint8 script; /* ucp_Arabic, etc. */
2758 pcre_uint8 chartype; /* ucp_Cc, etc. (general categories) */
2759 pcre_uint8 gbprop; /* ucp_gbControl, etc. (grapheme break property) */
2760 pcre_uint8 caseset; /* offset to multichar other cases or zero */
2761 pcre_int32 other_case; /* offset to other case, or zero if none */
2762 } ucd_record;
2763
2764 extern const pcre_uint32 PRIV(ucd_caseless_sets)[];
2765 extern const ucd_record PRIV(ucd_records)[];
2766 extern const pcre_uint8 PRIV(ucd_stage1)[];
2767 extern const pcre_uint16 PRIV(ucd_stage2)[];
2768 extern const pcre_uint32 PRIV(ucp_gentype)[];
2769 extern const pcre_uint32 PRIV(ucp_gbtable)[];
2770 #ifdef SUPPORT_JIT
2771 extern const int PRIV(ucp_typerange)[];
2772 #endif
2773
2774 #ifdef SUPPORT_UCP
2775 /* UCD access macros */
2776
2777 #define UCD_BLOCK_SIZE 128
2778 #define GET_UCD(ch) (PRIV(ucd_records) + \
2779 PRIV(ucd_stage2)[PRIV(ucd_stage1)[(int)(ch) / UCD_BLOCK_SIZE] * \
2780 UCD_BLOCK_SIZE + (int)(ch) % UCD_BLOCK_SIZE])
2781
2782 #define UCD_CHARTYPE(ch) GET_UCD(ch)->chartype
2783 #define UCD_SCRIPT(ch) GET_UCD(ch)->script
2784 #define UCD_CATEGORY(ch) PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2785 #define UCD_GRAPHBREAK(ch) GET_UCD(ch)->gbprop
2786 #define UCD_CASESET(ch) GET_UCD(ch)->caseset
2787 #define UCD_OTHERCASE(ch) ((pcre_uint32)((int)ch + (int)(GET_UCD(ch)->other_case)))
2788
2789 #endif /* SUPPORT_UCP */
2790
2791 #endif
2792
2793 /* End of pcre_internal.h */
2794
