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) 1997-2016 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 /* %ExternalCopyright% */
47
48 #ifndef PCRE_INTERNAL_H
49 #define PCRE_INTERNAL_H
50
51 #ifdef ERLANG_INTEGRATION
52 #include "local_config.h"
53 #endif
54
55 /* Define PCRE_DEBUG to get debugging output on stdout. */
56
57 #if 0
58 #define PCRE_DEBUG
59 #endif
60
61 /* PCRE is compiled as an 8 bit library if it is not requested otherwise. */
62
63 #if !defined COMPILE_PCRE16 && !defined COMPILE_PCRE32
64 #define COMPILE_PCRE8
65 #endif
66
67 /* If SUPPORT_UCP is defined, SUPPORT_UTF must also be defined. The
68 "configure" script ensures this, but not everybody uses "configure". */
69
70 #if defined SUPPORT_UCP && !(defined SUPPORT_UTF)
71 #define SUPPORT_UTF 1
72 #endif
73
74 /* We define SUPPORT_UTF if SUPPORT_UTF8 is enabled for compatibility
75 reasons with existing code. */
76
77 #if defined SUPPORT_UTF8 && !(defined SUPPORT_UTF)
78 #define SUPPORT_UTF 1
79 #endif
80
81 /* Fixme: SUPPORT_UTF8 should be eventually disappear from the code.
82 Until then we define it if SUPPORT_UTF is defined. */
83
84 #if defined SUPPORT_UTF && !(defined SUPPORT_UTF8)
85 #define SUPPORT_UTF8 1
86 #endif
87
88 /* We do not support both EBCDIC and UTF-8/16/32 at the same time. The "configure"
89 script prevents both being selected, but not everybody uses "configure". */
90
91 #if defined EBCDIC && defined SUPPORT_UTF
92 #error The use of both EBCDIC and SUPPORT_UTF is not supported.
93 #endif
94
95 /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef
96 inline, and there are *still* stupid compilers about that don't like indented
97 pre-processor statements, or at least there were when I first wrote this. After
98 all, it had only been about 10 years then...
99
100 It turns out that the Mac Debugging.h header also defines the macro DPRINTF, so
101 be absolutely sure we get our version. */
102
103 #undef DPRINTF
104 #ifdef PCRE_DEBUG
105 #define DPRINTF(p) printf p
106 #else
107 #define DPRINTF(p) /* Nothing */
108 #endif
109
110
111 /* Standard C headers plus the external interface definition. The only time
112 setjmp and stdarg are used is when NO_RECURSE is set. */
113
114 #include <ctype.h>
115 #include <limits.h>
116 #include <stddef.h>
117 #include <stdio.h>
118 #include <stdlib.h>
119 #include <string.h>
120
121 /* Valgrind (memcheck) support */
122
123 #ifdef SUPPORT_VALGRIND
124 #include <valgrind/memcheck.h>
125 #endif
126
127 /* When compiling a DLL for Windows, the exported symbols have to be declared
128 using some MS magic. I found some useful information on this web page:
129 http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
130 information there, using __declspec(dllexport) without "extern" we have a
131 definition; with "extern" we have a declaration. The settings here override the
132 setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
133 which is all that is needed for applications (they just import the symbols). We
134 use:
135
136 PCRE_EXP_DECL for declarations
137 PCRE_EXP_DEFN for definitions of exported functions
138 PCRE_EXP_DATA_DEFN for definitions of exported variables
139
140 The reason for the two DEFN macros is that in non-Windows environments, one
141 does not want to have "extern" before variable definitions because it leads to
142 compiler warnings. So we distinguish between functions and variables. In
143 Windows, the two should always be the same.
144
145 The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
146 which is an application, but needs to import this file in order to "peek" at
147 internals, can #include pcre.h first to get an application's-eye view.
148
149 In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
150 special-purpose environments) might want to stick other stuff in front of
151 exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
152 PCRE_EXP_DATA_DEFN only if they are not already set. */
153
154 #ifndef PCRE_EXP_DECL
155 # ifdef _WIN32
156 # ifndef PCRE_STATIC
157 # define PCRE_EXP_DECL extern __declspec(dllexport)
158 # define PCRE_EXP_DEFN __declspec(dllexport)
159 # define PCRE_EXP_DATA_DEFN __declspec(dllexport)
160 # else
161 # define PCRE_EXP_DECL extern
162 # define PCRE_EXP_DEFN
163 # define PCRE_EXP_DATA_DEFN
164 # endif
165 # else
166 # ifdef __cplusplus
167 # define PCRE_EXP_DECL extern "C"
168 # else
169 # define PCRE_EXP_DECL extern
170 # endif
171 # ifndef PCRE_EXP_DEFN
172 # define PCRE_EXP_DEFN PCRE_EXP_DECL
173 # endif
174 # ifndef PCRE_EXP_DATA_DEFN
175 # define PCRE_EXP_DATA_DEFN
176 # endif
177 # endif
178 #endif
179
180 /* When compiling with the MSVC compiler, it is sometimes necessary to include
181 a "calling convention" before exported function names. (This is secondhand
182 information; I know nothing about MSVC myself). For example, something like
183
184 void __cdecl function(....)
185
186 might be needed. In order so make this easy, all the exported functions have
187 PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
188 set, we ensure here that it has no effect. */
189
190 #ifndef PCRE_CALL_CONVENTION
191 #define PCRE_CALL_CONVENTION
192 #endif
193
194 /* We need to have types that specify unsigned 8, 16 and 32-bit integers. We
195 cannot determine these outside the compilation (e.g. by running a program as
196 part of "configure") because PCRE is often cross-compiled for use on other
197 systems. Instead we make use of the maximum sizes that are available at
198 preprocessor time in standard C environments. */
199
200 typedef unsigned char pcre_uint8;
201
202 #if USHRT_MAX == 65535
203 typedef unsigned short pcre_uint16;
204 typedef short pcre_int16;
205 #define PCRE_UINT16_MAX USHRT_MAX
206 #define PCRE_INT16_MAX SHRT_MAX
207 #elif UINT_MAX == 65535
208 typedef unsigned int pcre_uint16;
209 typedef int pcre_int16;
210 #define PCRE_UINT16_MAX UINT_MAX
211 #define PCRE_INT16_MAX INT_MAX
212 #else
213 #error Cannot determine a type for 16-bit integers
214 #endif
215
216 #if UINT_MAX == 4294967295U
217 typedef unsigned int pcre_uint32;
218 typedef int pcre_int32;
219 #define PCRE_UINT32_MAX UINT_MAX
220 #define PCRE_INT32_MAX INT_MAX
221 #elif ULONG_MAX == 4294967295UL
222 typedef unsigned long int pcre_uint32;
223 typedef long int pcre_int32;
224 #define PCRE_UINT32_MAX ULONG_MAX
225 #define PCRE_INT32_MAX LONG_MAX
226 #else
227 #error Cannot determine a type for 32-bit integers
228 #endif
229
230 /* When checking for integer overflow in pcre_compile(), we need to handle
231 large integers. If a 64-bit integer type is available, we can use that.
232 Otherwise we have to cast to double, which of course requires floating point
233 arithmetic. Handle this by defining a macro for the appropriate type. If
234 stdint.h is available, include it; it may define INT64_MAX. Systems that do not
235 have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
236 by "configure". */
237
238 #if defined HAVE_STDINT_H
239 #include <stdint.h>
240 #elif defined HAVE_INTTYPES_H
241 #include <inttypes.h>
242 #endif
243
244 #if defined INT64_MAX || defined int64_t
245 #define INT64_OR_DOUBLE int64_t
246 #else
247 #define INT64_OR_DOUBLE double
248 #endif
249
250 /* All character handling must be done as unsigned characters. Otherwise there
251 are problems with top-bit-set characters and functions such as isspace().
252 However, we leave the interface to the outside world as char * or short *,
253 because that should make things easier for callers. This character type is
254 called pcre_uchar.
255
256 The IN_UCHARS macro multiply its argument with the byte size of the current
257 pcre_uchar type. Useful for memcpy and such operations, whose require the
258 byte size of their input/output buffers.
259
260 The MAX_255 macro checks whether its pcre_uchar input is less than 256.
261
262 The TABLE_GET macro is designed for accessing elements of tables whose contain
263 exactly 256 items. When the character is able to contain more than 256
264 items, some check is needed before accessing these tables.
265 */
266
267 #if defined COMPILE_PCRE8
268
269 typedef unsigned char pcre_uchar;
270 #define IN_UCHARS(x) (x)
271 #define MAX_255(c) 1
272 #define TABLE_GET(c, table, default) ((table)[c])
273
274 #elif defined COMPILE_PCRE16
275
276 #if USHRT_MAX != 65535
277 /* This is a warning message. Change PCRE_UCHAR16 to a 16 bit data type in
278 pcre.h(.in) and disable (comment out) this message. */
279 #error Warning: PCRE_UCHAR16 is not a 16 bit data type.
280 #endif
281
282 typedef pcre_uint16 pcre_uchar;
283 #define UCHAR_SHIFT (1)
284 #define IN_UCHARS(x) ((x) * 2)
285 #define MAX_255(c) ((c) <= 255u)
286 #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
287
288 #elif defined COMPILE_PCRE32
289
290 typedef pcre_uint32 pcre_uchar;
291 #define UCHAR_SHIFT (2)
292 #define IN_UCHARS(x) ((x) * 4)
293 #define MAX_255(c) ((c) <= 255u)
294 #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
295
296 #else
297 #error Unsupported compiling mode
298 #endif /* COMPILE_PCRE[8|16|32] */
299
300 /* This is an unsigned int value that no character can ever have. UTF-8
301 characters only go up to 0x7fffffff (though Unicode doesn't go beyond
302 0x0010ffff). */
303
304 #define NOTACHAR 0xffffffff
305
306 /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
307 "any" and "anycrlf" at present). The following macros are used to package up
308 testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
309 modules to indicate in which datablock the parameters exist, and what the
310 start/end of string field names are. */
311
312 #define NLTYPE_FIXED 0 /* Newline is a fixed length string */
313 #define NLTYPE_ANY 1 /* Newline is any Unicode line ending */
314 #define NLTYPE_ANYCRLF 2 /* Newline is CR, LF, or CRLF */
315
316 /* This macro checks for a newline at the given position */
317
318 #define IS_NEWLINE(p) \
319 ((NLBLOCK->nltype != NLTYPE_FIXED)? \
320 ((p) < NLBLOCK->PSEND && \
321 PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, \
322 &(NLBLOCK->nllen), utf)) \
323 : \
324 ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
325 UCHAR21TEST(p) == NLBLOCK->nl[0] && \
326 (NLBLOCK->nllen == 1 || UCHAR21TEST(p+1) == NLBLOCK->nl[1]) \
327 ) \
328 )
329
330 /* This macro checks for a newline immediately preceding the given position */
331
332 #define WAS_NEWLINE(p) \
333 ((NLBLOCK->nltype != NLTYPE_FIXED)? \
334 ((p) > NLBLOCK->PSSTART && \
335 PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
336 &(NLBLOCK->nllen), utf)) \
337 : \
338 ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
339 UCHAR21TEST(p - NLBLOCK->nllen) == NLBLOCK->nl[0] && \
340 (NLBLOCK->nllen == 1 || UCHAR21TEST(p - NLBLOCK->nllen + 1) == NLBLOCK->nl[1]) \
341 ) \
342 )
343
344 /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
345 with a custom type. This makes it possible, for example, to allow pcre_exec()
346 to process subject strings that are discontinuous by using a smart pointer
347 class. It must always be possible to inspect all of the subject string in
348 pcre_exec() because of the way it backtracks. Two macros are required in the
349 normal case, for sign-unspecified and unsigned char pointers. The former is
350 used for the external interface and appears in pcre.h, which is why its name
351 must begin with PCRE_. */
352
353 #ifdef CUSTOM_SUBJECT_PTR
354 #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
355 #else
356 #define PCRE_PUCHAR const pcre_uchar *
357 #endif
358
359 /* Include the public PCRE header and the definitions of UCP character property
360 values. */
361
362 #include "pcre.h"
363 #include "ucp.h"
364
365 #ifdef COMPILE_PCRE32
366 /* Assert that the public PCRE_UCHAR32 is a 32-bit type */
367 typedef int __assert_pcre_uchar32_size[sizeof(PCRE_UCHAR32) == 4 ? 1 : -1];
368 #endif
369
370 /* When compiling for use with the Virtual Pascal compiler, these functions
371 need to have their names changed. PCRE must be compiled with the -DVPCOMPAT
372 option on the command line. */
373
374 #ifdef VPCOMPAT
375 #define strlen(s) _strlen(s)
376 #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
377 #define memcmp(s,c,n) _memcmp(s,c,n)
378 #define memcpy(d,s,n) _memcpy(d,s,n)
379 #define memmove(d,s,n) _memmove(d,s,n)
380 #define memset(s,c,n) _memset(s,c,n)
381 #else /* VPCOMPAT */
382
383 /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
384 define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
385 is set. Otherwise, include an emulating function for those systems that have
386 neither (there some non-Unix environments where this is the case). */
387
388 #ifndef HAVE_MEMMOVE
389 #undef memmove /* some systems may have a macro */
390 #ifdef HAVE_BCOPY
391 #define memmove(a, b, c) bcopy(b, a, c)
392 #else /* HAVE_BCOPY */
393 static void *
pcre_memmove(void * d,const void * s,size_t n)394 pcre_memmove(void *d, const void *s, size_t n)
395 {
396 size_t i;
397 unsigned char *dest = (unsigned char *)d;
398 const unsigned char *src = (const unsigned char *)s;
399 if (dest > src)
400 {
401 dest += n;
402 src += n;
403 for (i = 0; i < n; ++i) *(--dest) = *(--src);
404 return (void *)dest;
405 }
406 else
407 {
408 for (i = 0; i < n; ++i) *dest++ = *src++;
409 return (void *)(dest - n);
410 }
411 }
412 #define memmove(a, b, c) pcre_memmove(a, b, c)
413 #endif /* not HAVE_BCOPY */
414 #endif /* not HAVE_MEMMOVE */
415 #endif /* not VPCOMPAT */
416
417
418 /* PCRE keeps offsets in its compiled code as 2-byte quantities (always stored
419 in big-endian order) by default. These are used, for example, to link from the
420 start of a subpattern to its alternatives and its end. The use of 2 bytes per
421 offset limits the size of the compiled regex to around 64K, which is big enough
422 for almost everybody. However, I received a request for an even bigger limit.
423 For this reason, and also to make the code easier to maintain, the storing and
424 loading of offsets from the byte string is now handled by the macros that are
425 defined here.
426
427 The macros are controlled by the value of LINK_SIZE. This defaults to 2 in
428 the config.h file, but can be overridden by using -D on the command line. This
429 is automated on Unix systems via the "configure" command. */
430
431 #if defined COMPILE_PCRE8
432
433 #if LINK_SIZE == 2
434
435 #define PUT(a,n,d) \
436 (a[n] = (d) >> 8), \
437 (a[(n)+1] = (d) & 255)
438
439 #define GET(a,n) \
440 (((a)[n] << 8) | (a)[(n)+1])
441
442 #define MAX_PATTERN_SIZE (1 << 16)
443
444
445 #elif LINK_SIZE == 3
446
447 #define PUT(a,n,d) \
448 (a[n] = (d) >> 16), \
449 (a[(n)+1] = (d) >> 8), \
450 (a[(n)+2] = (d) & 255)
451
452 #define GET(a,n) \
453 (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2])
454
455 #define MAX_PATTERN_SIZE (1 << 24)
456
457
458 #elif LINK_SIZE == 4
459
460 #define PUT(a,n,d) \
461 (a[n] = (d) >> 24), \
462 (a[(n)+1] = (d) >> 16), \
463 (a[(n)+2] = (d) >> 8), \
464 (a[(n)+3] = (d) & 255)
465
466 #define GET(a,n) \
467 (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
468
469 /* Keep it positive */
470 #define MAX_PATTERN_SIZE (1 << 30)
471
472 #else
473 #error LINK_SIZE must be either 2, 3, or 4
474 #endif
475
476 #elif defined COMPILE_PCRE16
477
478 #if LINK_SIZE == 2
479
480 /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
481 #undef LINK_SIZE
482 #define LINK_SIZE 1
483
484 #define PUT(a,n,d) \
485 (a[n] = (d))
486
487 #define GET(a,n) \
488 (a[n])
489
490 #define MAX_PATTERN_SIZE (1 << 16)
491
492 #elif LINK_SIZE == 3 || LINK_SIZE == 4
493
494 /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
495 #undef LINK_SIZE
496 #define LINK_SIZE 2
497
498 #define PUT(a,n,d) \
499 (a[n] = (d) >> 16), \
500 (a[(n)+1] = (d) & 65535)
501
502 #define GET(a,n) \
503 (((a)[n] << 16) | (a)[(n)+1])
504
505 /* Keep it positive */
506 #define MAX_PATTERN_SIZE (1 << 30)
507
508 #else
509 #error LINK_SIZE must be either 2, 3, or 4
510 #endif
511
512 #elif defined COMPILE_PCRE32
513
514 /* Only supported LINK_SIZE is 4 */
515 /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
516 #undef LINK_SIZE
517 #define LINK_SIZE 1
518
519 #define PUT(a,n,d) \
520 (a[n] = (d))
521
522 #define GET(a,n) \
523 (a[n])
524
525 /* Keep it positive */
526 #define MAX_PATTERN_SIZE (1 << 30)
527
528 #else
529 #error Unsupported compiling mode
530 #endif /* COMPILE_PCRE[8|16|32] */
531
532 /* Convenience macro defined in terms of the others */
533
534 #define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE
535
536
537 /* PCRE uses some other 2-byte quantities that do not change when the size of
538 offsets changes. There are used for repeat counts and for other things such as
539 capturing parenthesis numbers in back references. */
540
541 #if defined COMPILE_PCRE8
542
543 #define IMM2_SIZE 2
544
545 #define PUT2(a,n,d) \
546 a[n] = (d) >> 8; \
547 a[(n)+1] = (d) & 255
548
549 /* For reasons that I do not understand, the expression in this GET2 macro is
550 treated by gcc as a signed expression, even when a is declared as unsigned. It
551 seems that any kind of arithmetic results in a signed value. */
552
553 #define GET2(a,n) \
554 (unsigned int)(((a)[n] << 8) | (a)[(n)+1])
555
556 #elif defined COMPILE_PCRE16
557
558 #define IMM2_SIZE 1
559
560 #define PUT2(a,n,d) \
561 a[n] = d
562
563 #define GET2(a,n) \
564 a[n]
565
566 #elif defined COMPILE_PCRE32
567
568 #define IMM2_SIZE 1
569
570 #define PUT2(a,n,d) \
571 a[n] = d
572
573 #define GET2(a,n) \
574 a[n]
575
576 #else
577 #error Unsupported compiling mode
578 #endif /* COMPILE_PCRE[8|16|32] */
579
580 #define PUT2INC(a,n,d) PUT2(a,n,d), a += IMM2_SIZE
581
582 /* The maximum length of a MARK name is currently one data unit; it may be
583 changed in future to be a fixed number of bytes or to depend on LINK_SIZE. */
584
585 #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
586 #define MAX_MARK ((1u << 16) - 1)
587 #else
588 #define MAX_MARK ((1u << 8) - 1)
589 #endif
590
591 /* There is a proposed future special "UTF-21" mode, in which only the lowest
592 21 bits of a 32-bit character are interpreted as UTF, with the remaining 11
593 high-order bits available to the application for other uses. In preparation for
594 the future implementation of this mode, there are macros that load a data item
595 and, if in this special mode, mask it to 21 bits. These macros all have names
596 starting with UCHAR21. In all other modes, including the normal 32-bit
597 library, the macros all have the same simple definitions. When the new mode is
598 implemented, it is expected that these definitions will be varied appropriately
599 using #ifdef when compiling the library that supports the special mode. */
600
601 #define UCHAR21(eptr) (*(eptr))
602 #define UCHAR21TEST(eptr) (*(eptr))
603 #define UCHAR21INC(eptr) (*(eptr)++)
604 #define UCHAR21INCTEST(eptr) (*(eptr)++)
605
606 /* When UTF encoding is being used, a character is no longer just a single
607 byte in 8-bit mode or a single short in 16-bit mode. The macros for character
608 handling generate simple sequences when used in the basic mode, and more
609 complicated ones for UTF characters. GETCHARLENTEST and other macros are not
610 used when UTF is not supported. To make sure they can never even appear when
611 UTF support is omitted, we don't even define them. */
612
613 #ifndef SUPPORT_UTF
614
615 /* #define MAX_VALUE_FOR_SINGLE_CHAR */
616 /* #define HAS_EXTRALEN(c) */
617 /* #define GET_EXTRALEN(c) */
618 /* #define NOT_FIRSTCHAR(c) */
619 #define GETCHAR(c, eptr) c = *eptr;
620 #define GETCHARTEST(c, eptr) c = *eptr;
621 #define GETCHARINC(c, eptr) c = *eptr++;
622 #define GETCHARINCTEST(c, eptr) c = *eptr++;
623 #define GETCHARLEN(c, eptr, len) c = *eptr;
624 /* #define GETCHARLENTEST(c, eptr, len) */
625 /* #define BACKCHAR(eptr) */
626 /* #define FORWARDCHAR(eptr) */
627 /* #define ACROSSCHAR(condition, eptr, action) */
628
629 #else /* SUPPORT_UTF */
630
631 /* Tests whether the code point needs extra characters to decode. */
632
633 #define HASUTF8EXTRALEN(c) ((c) >= 0xc0)
634
635 /* Base macro to pick up the remaining bytes of a UTF-8 character, not
636 advancing the pointer. */
637
638 #define GETUTF8(c, eptr) \
639 { \
640 if ((c & 0x20) == 0) \
641 c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
642 else if ((c & 0x10) == 0) \
643 c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
644 else if ((c & 0x08) == 0) \
645 c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
646 ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
647 else if ((c & 0x04) == 0) \
648 c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
649 ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
650 (eptr[4] & 0x3f); \
651 else \
652 c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
653 ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
654 ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
655 }
656
657 /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
658 the pointer. */
659
660 #define GETUTF8INC(c, eptr) \
661 { \
662 if ((c & 0x20) == 0) \
663 c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
664 else if ((c & 0x10) == 0) \
665 { \
666 c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
667 eptr += 2; \
668 } \
669 else if ((c & 0x08) == 0) \
670 { \
671 c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
672 ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
673 eptr += 3; \
674 } \
675 else if ((c & 0x04) == 0) \
676 { \
677 c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
678 ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
679 (eptr[3] & 0x3f); \
680 eptr += 4; \
681 } \
682 else \
683 { \
684 c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
685 ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
686 ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
687 eptr += 5; \
688 } \
689 }
690
691 #if defined COMPILE_PCRE8
692
693 /* These macros were originally written in the form of loops that used data
694 from the tables whose names start with PRIV(utf8_table). They were rewritten by
695 a user so as not to use loops, because in some environments this gives a
696 significant performance advantage, and it seems never to do any harm. */
697
698 /* Tells the biggest code point which can be encoded as a single character. */
699
700 #define MAX_VALUE_FOR_SINGLE_CHAR 127
701
702 /* Tests whether the code point needs extra characters to decode. */
703
704 #define HAS_EXTRALEN(c) ((c) >= 0xc0)
705
706 /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
707 Otherwise it has an undefined behaviour. */
708
709 #define GET_EXTRALEN(c) (PRIV(utf8_table4)[(c) & 0x3f])
710
711 /* Returns TRUE, if the given character is not the first character
712 of a UTF sequence. */
713
714 #define NOT_FIRSTCHAR(c) (((c) & 0xc0) == 0x80)
715
716 /* Get the next UTF-8 character, not advancing the pointer. This is called when
717 we know we are in UTF-8 mode. */
718
719 #define GETCHAR(c, eptr) \
720 c = *eptr; \
721 if (c >= 0xc0) GETUTF8(c, eptr);
722
723 /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
724 pointer. */
725
726 #define GETCHARTEST(c, eptr) \
727 c = *eptr; \
728 if (utf && c >= 0xc0) GETUTF8(c, eptr);
729
730 /* Get the next UTF-8 character, advancing the pointer. This is called when we
731 know we are in UTF-8 mode. */
732
733 #define GETCHARINC(c, eptr) \
734 c = *eptr++; \
735 if (c >= 0xc0) GETUTF8INC(c, eptr);
736
737 /* Get the next character, testing for UTF-8 mode, and advancing the pointer.
738 This is called when we don't know if we are in UTF-8 mode. */
739
740 #define GETCHARINCTEST(c, eptr) \
741 c = *eptr++; \
742 if (utf && c >= 0xc0) GETUTF8INC(c, eptr);
743
744 /* Base macro to pick up the remaining bytes of a UTF-8 character, not
745 advancing the pointer, incrementing the length. */
746
747 #define GETUTF8LEN(c, eptr, len) \
748 { \
749 if ((c & 0x20) == 0) \
750 { \
751 c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
752 len++; \
753 } \
754 else if ((c & 0x10) == 0) \
755 { \
756 c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
757 len += 2; \
758 } \
759 else if ((c & 0x08) == 0) \
760 {\
761 c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
762 ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
763 len += 3; \
764 } \
765 else if ((c & 0x04) == 0) \
766 { \
767 c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
768 ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
769 (eptr[4] & 0x3f); \
770 len += 4; \
771 } \
772 else \
773 {\
774 c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
775 ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
776 ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
777 len += 5; \
778 } \
779 }
780
781 /* Get the next UTF-8 character, not advancing the pointer, incrementing length
782 if there are extra bytes. This is called when we know we are in UTF-8 mode. */
783
784 #define GETCHARLEN(c, eptr, len) \
785 c = *eptr; \
786 if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
787
788 /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
789 pointer, incrementing length if there are extra bytes. This is called when we
790 do not know if we are in UTF-8 mode. */
791
792 #define GETCHARLENTEST(c, eptr, len) \
793 c = *eptr; \
794 if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len);
795
796 /* If the pointer is not at the start of a character, move it back until
797 it is. This is called only in UTF-8 mode - we don't put a test within the macro
798 because almost all calls are already within a block of UTF-8 only code. */
799
800 #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
801
802 /* Same as above, just in the other direction. */
803 #define FORWARDCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr++
804
805 /* Same as above, but it allows a fully customizable form. */
806 #define ACROSSCHAR(condition, eptr, action) \
807 while((condition) && ((eptr) & 0xc0) == 0x80) action
808
809 #elif defined COMPILE_PCRE16
810
811 /* Tells the biggest code point which can be encoded as a single character. */
812
813 #define MAX_VALUE_FOR_SINGLE_CHAR 65535
814
815 /* Tests whether the code point needs extra characters to decode. */
816
817 #define HAS_EXTRALEN(c) (((c) & 0xfc00) == 0xd800)
818
819 /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
820 Otherwise it has an undefined behaviour. */
821
822 #define GET_EXTRALEN(c) 1
823
824 /* Returns TRUE, if the given character is not the first character
825 of a UTF sequence. */
826
827 #define NOT_FIRSTCHAR(c) (((c) & 0xfc00) == 0xdc00)
828
829 /* Base macro to pick up the low surrogate of a UTF-16 character, not
830 advancing the pointer. */
831
832 #define GETUTF16(c, eptr) \
833 { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; }
834
835 /* Get the next UTF-16 character, not advancing the pointer. This is called when
836 we know we are in UTF-16 mode. */
837
838 #define GETCHAR(c, eptr) \
839 c = *eptr; \
840 if ((c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
841
842 /* Get the next UTF-16 character, testing for UTF-16 mode, and not advancing the
843 pointer. */
844
845 #define GETCHARTEST(c, eptr) \
846 c = *eptr; \
847 if (utf && (c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
848
849 /* Base macro to pick up the low surrogate of a UTF-16 character, advancing
850 the pointer. */
851
852 #define GETUTF16INC(c, eptr) \
853 { c = (((c & 0x3ff) << 10) | (*eptr++ & 0x3ff)) + 0x10000; }
854
855 /* Get the next UTF-16 character, advancing the pointer. This is called when we
856 know we are in UTF-16 mode. */
857
858 #define GETCHARINC(c, eptr) \
859 c = *eptr++; \
860 if ((c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
861
862 /* Get the next character, testing for UTF-16 mode, and advancing the pointer.
863 This is called when we don't know if we are in UTF-16 mode. */
864
865 #define GETCHARINCTEST(c, eptr) \
866 c = *eptr++; \
867 if (utf && (c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
868
869 /* Base macro to pick up the low surrogate of a UTF-16 character, not
870 advancing the pointer, incrementing the length. */
871
872 #define GETUTF16LEN(c, eptr, len) \
873 { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; len++; }
874
875 /* Get the next UTF-16 character, not advancing the pointer, incrementing
876 length if there is a low surrogate. This is called when we know we are in
877 UTF-16 mode. */
878
879 #define GETCHARLEN(c, eptr, len) \
880 c = *eptr; \
881 if ((c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
882
883 /* Get the next UTF-816character, testing for UTF-16 mode, not advancing the
884 pointer, incrementing length if there is a low surrogate. This is called when
885 we do not know if we are in UTF-16 mode. */
886
887 #define GETCHARLENTEST(c, eptr, len) \
888 c = *eptr; \
889 if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
890
891 /* If the pointer is not at the start of a character, move it back until
892 it is. This is called only in UTF-16 mode - we don't put a test within the
893 macro because almost all calls are already within a block of UTF-16 only
894 code. */
895
896 #define BACKCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr--
897
898 /* Same as above, just in the other direction. */
899 #define FORWARDCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr++
900
901 /* Same as above, but it allows a fully customizable form. */
902 #define ACROSSCHAR(condition, eptr, action) \
903 if ((condition) && ((eptr) & 0xfc00) == 0xdc00) action
904
905 #elif defined COMPILE_PCRE32
906
907 /* These are trivial for the 32-bit library, since all UTF-32 characters fit
908 into one pcre_uchar unit. */
909 #define MAX_VALUE_FOR_SINGLE_CHAR (0x10ffffu)
910 #define HAS_EXTRALEN(c) (0)
911 #define GET_EXTRALEN(c) (0)
912 #define NOT_FIRSTCHAR(c) (0)
913
914 /* Get the next UTF-32 character, not advancing the pointer. This is called when
915 we know we are in UTF-32 mode. */
916
917 #define GETCHAR(c, eptr) \
918 c = *(eptr);
919
920 /* Get the next UTF-32 character, testing for UTF-32 mode, and not advancing the
921 pointer. */
922
923 #define GETCHARTEST(c, eptr) \
924 c = *(eptr);
925
926 /* Get the next UTF-32 character, advancing the pointer. This is called when we
927 know we are in UTF-32 mode. */
928
929 #define GETCHARINC(c, eptr) \
930 c = *((eptr)++);
931
932 /* Get the next character, testing for UTF-32 mode, and advancing the pointer.
933 This is called when we don't know if we are in UTF-32 mode. */
934
935 #define GETCHARINCTEST(c, eptr) \
936 c = *((eptr)++);
937
938 /* Get the next UTF-32 character, not advancing the pointer, not incrementing
939 length (since all UTF-32 is of length 1). This is called when we know we are in
940 UTF-32 mode. */
941
942 #define GETCHARLEN(c, eptr, len) \
943 GETCHAR(c, eptr)
944
945 /* Get the next UTF-32character, testing for UTF-32 mode, not advancing the
946 pointer, not incrementing the length (since all UTF-32 is of length 1).
947 This is called when we do not know if we are in UTF-32 mode. */
948
949 #define GETCHARLENTEST(c, eptr, len) \
950 GETCHARTEST(c, eptr)
951
952 /* If the pointer is not at the start of a character, move it back until
953 it is. This is called only in UTF-32 mode - we don't put a test within the
954 macro because almost all calls are already within a block of UTF-32 only
955 code.
956 These are all no-ops since all UTF-32 characters fit into one pcre_uchar. */
957
958 #define BACKCHAR(eptr) do { } while (0)
959
960 /* Same as above, just in the other direction. */
961 #define FORWARDCHAR(eptr) do { } while (0)
962
963 /* Same as above, but it allows a fully customizable form. */
964 #define ACROSSCHAR(condition, eptr, action) do { } while (0)
965
966 #else
967 #error Unsupported compiling mode
968 #endif /* COMPILE_PCRE[8|16|32] */
969
970 #endif /* SUPPORT_UTF */
971
972 /* Tests for Unicode horizontal and vertical whitespace characters must check a
973 number of different values. Using a switch statement for this generates the
974 fastest code (no loop, no memory access), and there are several places in the
975 interpreter code where this happens. In order to ensure that all the case lists
976 remain in step, we use macros so that there is only one place where the lists
977 are defined.
978
979 These values are also required as lists in pcre_compile.c when processing \h,
980 \H, \v and \V in a character class. The lists are defined in pcre_tables.c, but
981 macros that define the values are here so that all the definitions are
982 together. The lists must be in ascending character order, terminated by
983 NOTACHAR (which is 0xffffffff).
984
985 Any changes should ensure that the various macros are kept in step with each
986 other. NOTE: The values also appear in pcre_jit_compile.c. */
987
988 /* ------ ASCII/Unicode environments ------ */
989
990 #ifndef EBCDIC
991
992 #define HSPACE_LIST \
993 CHAR_HT, CHAR_SPACE, CHAR_NBSP, \
994 0x1680, 0x180e, 0x2000, 0x2001, 0x2002, 0x2003, 0x2004, 0x2005, \
995 0x2006, 0x2007, 0x2008, 0x2009, 0x200A, 0x202f, 0x205f, 0x3000, \
996 NOTACHAR
997
998 #define HSPACE_MULTIBYTE_CASES \
999 case 0x1680: /* OGHAM SPACE MARK */ \
1000 case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ \
1001 case 0x2000: /* EN QUAD */ \
1002 case 0x2001: /* EM QUAD */ \
1003 case 0x2002: /* EN SPACE */ \
1004 case 0x2003: /* EM SPACE */ \
1005 case 0x2004: /* THREE-PER-EM SPACE */ \
1006 case 0x2005: /* FOUR-PER-EM SPACE */ \
1007 case 0x2006: /* SIX-PER-EM SPACE */ \
1008 case 0x2007: /* FIGURE SPACE */ \
1009 case 0x2008: /* PUNCTUATION SPACE */ \
1010 case 0x2009: /* THIN SPACE */ \
1011 case 0x200A: /* HAIR SPACE */ \
1012 case 0x202f: /* NARROW NO-BREAK SPACE */ \
1013 case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ \
1014 case 0x3000 /* IDEOGRAPHIC SPACE */
1015
1016 #define HSPACE_BYTE_CASES \
1017 case CHAR_HT: \
1018 case CHAR_SPACE: \
1019 case CHAR_NBSP
1020
1021 #define HSPACE_CASES \
1022 HSPACE_BYTE_CASES: \
1023 HSPACE_MULTIBYTE_CASES
1024
1025 #define VSPACE_LIST \
1026 CHAR_LF, CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, 0x2028, 0x2029, NOTACHAR
1027
1028 #define VSPACE_MULTIBYTE_CASES \
1029 case 0x2028: /* LINE SEPARATOR */ \
1030 case 0x2029 /* PARAGRAPH SEPARATOR */
1031
1032 #define VSPACE_BYTE_CASES \
1033 case CHAR_LF: \
1034 case CHAR_VT: \
1035 case CHAR_FF: \
1036 case CHAR_CR: \
1037 case CHAR_NEL
1038
1039 #define VSPACE_CASES \
1040 VSPACE_BYTE_CASES: \
1041 VSPACE_MULTIBYTE_CASES
1042
1043 /* ------ EBCDIC environments ------ */
1044
1045 #else
1046 #define HSPACE_LIST CHAR_HT, CHAR_SPACE, CHAR_NBSP, NOTACHAR
1047
1048 #define HSPACE_BYTE_CASES \
1049 case CHAR_HT: \
1050 case CHAR_SPACE: \
1051 case CHAR_NBSP
1052
1053 #define HSPACE_CASES HSPACE_BYTE_CASES
1054
1055 #ifdef EBCDIC_NL25
1056 #define VSPACE_LIST \
1057 CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, CHAR_LF, NOTACHAR
1058 #else
1059 #define VSPACE_LIST \
1060 CHAR_VT, CHAR_FF, CHAR_CR, CHAR_LF, CHAR_NEL, NOTACHAR
1061 #endif
1062
1063 #define VSPACE_BYTE_CASES \
1064 case CHAR_LF: \
1065 case CHAR_VT: \
1066 case CHAR_FF: \
1067 case CHAR_CR: \
1068 case CHAR_NEL
1069
1070 #define VSPACE_CASES VSPACE_BYTE_CASES
1071 #endif /* EBCDIC */
1072
1073 /* ------ End of whitespace macros ------ */
1074
1075
1076
1077 /* Private flags containing information about the compiled regex. They used to
1078 live at the top end of the options word, but that got almost full, so they were
1079 moved to a 16-bit flags word - which got almost full, so now they are in a
1080 32-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as the
1081 restrictions on partial matching have been lifted. It remains for backwards
1082 compatibility. */
1083
1084 #define PCRE_MODE8 0x00000001 /* compiled in 8 bit mode */
1085 #define PCRE_MODE16 0x00000002 /* compiled in 16 bit mode */
1086 #define PCRE_MODE32 0x00000004 /* compiled in 32 bit mode */
1087 #define PCRE_FIRSTSET 0x00000010 /* first_char is set */
1088 #define PCRE_FCH_CASELESS 0x00000020 /* caseless first char */
1089 #define PCRE_REQCHSET 0x00000040 /* req_byte is set */
1090 #define PCRE_RCH_CASELESS 0x00000080 /* caseless requested char */
1091 #define PCRE_STARTLINE 0x00000100 /* start after \n for multiline */
1092 #define PCRE_NOPARTIAL 0x00000200 /* can't use partial with this regex */
1093 #define PCRE_JCHANGED 0x00000400 /* j option used in regex */
1094 #define PCRE_HASCRORLF 0x00000800 /* explicit \r or \n in pattern */
1095 #define PCRE_HASTHEN 0x00001000 /* pattern contains (*THEN) */
1096 #define PCRE_MLSET 0x00002000 /* match limit set by regex */
1097 #define PCRE_RLSET 0x00004000 /* recursion limit set by regex */
1098 #define PCRE_MATCH_EMPTY 0x00008000 /* pattern can match empty string */
1099
1100 #if defined COMPILE_PCRE8
1101 #define PCRE_MODE PCRE_MODE8
1102 #elif defined COMPILE_PCRE16
1103 #define PCRE_MODE PCRE_MODE16
1104 #elif defined COMPILE_PCRE32
1105 #define PCRE_MODE PCRE_MODE32
1106 #endif
1107 #define PCRE_MODE_MASK (PCRE_MODE8 | PCRE_MODE16 | PCRE_MODE32)
1108
1109 /* Flags for the "extra" block produced by pcre_study(). */
1110
1111 #define PCRE_STUDY_MAPPED 0x0001 /* a map of starting chars exists */
1112 #define PCRE_STUDY_MINLEN 0x0002 /* a minimum length field exists */
1113
1114 /* Masks for identifying the public options that are permitted at compile
1115 time, run time, or study time, respectively. */
1116
1117 #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
1118 PCRE_NEWLINE_ANYCRLF)
1119
1120 #define PUBLIC_COMPILE_OPTIONS \
1121 (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
1122 PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
1123 PCRE_NO_AUTO_CAPTURE|PCRE_NO_AUTO_POSSESS| \
1124 PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
1125 PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
1126 PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE|PCRE_NEVER_UTF)
1127
1128 #define PUBLIC_EXEC_OPTIONS \
1129 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
1130 PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
1131 PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
1132
1133 #define PUBLIC_DFA_EXEC_OPTIONS \
1134 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
1135 PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
1136 PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
1137 PCRE_NO_START_OPTIMIZE)
1138
1139 #define PUBLIC_STUDY_OPTIONS \
1140 (PCRE_STUDY_JIT_COMPILE|PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE| \
1141 PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE|PCRE_STUDY_EXTRA_NEEDED)
1142
1143 #define PUBLIC_JIT_EXEC_OPTIONS \
1144 (PCRE_NO_UTF8_CHECK|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|\
1145 PCRE_NOTEMPTY_ATSTART|PCRE_PARTIAL_SOFT|PCRE_PARTIAL_HARD)
1146
1147 /* Magic number to provide a small check against being handed junk. */
1148
1149 #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
1150
1151 /* This variable is used to detect a loaded regular expression
1152 in different endianness. */
1153
1154 #define REVERSED_MAGIC_NUMBER 0x45524350UL /* 'ERCP' */
1155
1156 /* The maximum remaining length of subject we are prepared to search for a
1157 req_byte match. */
1158
1159 #define REQ_BYTE_MAX 1000
1160
1161 /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
1162 environments where these macros are defined elsewhere. Unfortunately, there
1163 is no way to do the same for the typedef. */
1164
1165 typedef int BOOL;
1166
1167 #ifndef FALSE
1168 #define FALSE 0
1169 #define TRUE 1
1170 #endif
1171
1172 /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
1173 character constants like '*' because the compiler would emit their EBCDIC code,
1174 which is different from their ASCII/UTF-8 code. Instead we define macros for
1175 the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
1176 is enabled. When UTF-8 support is not enabled, the definitions use character
1177 literals. Both character and string versions of each character are needed, and
1178 there are some longer strings as well.
1179
1180 This means that, on EBCDIC platforms, the PCRE library can handle either
1181 EBCDIC, or UTF-8, but not both. To support both in the same compiled library
1182 would need different lookups depending on whether PCRE_UTF8 was set or not.
1183 This would make it impossible to use characters in switch/case statements,
1184 which would reduce performance. For a theoretical use (which nobody has asked
1185 for) in a minority area (EBCDIC platforms), this is not sensible. Any
1186 application that did need both could compile two versions of the library, using
1187 macros to give the functions distinct names. */
1188
1189 #ifndef SUPPORT_UTF
1190
1191 /* UTF-8 support is not enabled; use the platform-dependent character literals
1192 so that PCRE works in both ASCII and EBCDIC environments, but only in non-UTF
1193 mode. Newline characters are problematic in EBCDIC. Though it has CR and LF
1194 characters, a common practice has been to use its NL (0x15) character as the
1195 line terminator in C-like processing environments. However, sometimes the LF
1196 (0x25) character is used instead, according to this Unicode document:
1197
1198 http://unicode.org/standard/reports/tr13/tr13-5.html
1199
1200 PCRE defaults EBCDIC NL to 0x15, but has a build-time option to select 0x25
1201 instead. Whichever is *not* chosen is defined as NEL.
1202
1203 In both ASCII and EBCDIC environments, CHAR_NL and CHAR_LF are synonyms for the
1204 same code point. */
1205
1206 #ifdef EBCDIC
1207
1208 #ifndef EBCDIC_NL25
1209 #define CHAR_NL '\x15'
1210 #define CHAR_NEL '\x25'
1211 #define STR_NL "\x15"
1212 #define STR_NEL "\x25"
1213 #else
1214 #define CHAR_NL '\x25'
1215 #define CHAR_NEL '\x15'
1216 #define STR_NL "\x25"
1217 #define STR_NEL "\x15"
1218 #endif
1219
1220 #define CHAR_LF CHAR_NL
1221 #define STR_LF STR_NL
1222
1223 #define CHAR_ESC '\047'
1224 #define CHAR_DEL '\007'
1225 #define CHAR_NBSP '\x41'
1226 #define STR_ESC "\047"
1227 #define STR_DEL "\007"
1228
1229 #else /* Not EBCDIC */
1230
1231 /* In ASCII/Unicode, linefeed is '\n' and we equate this to NL for
1232 compatibility. NEL is the Unicode newline character; make sure it is
1233 a positive value. */
1234
1235 #define CHAR_LF '\n'
1236 #define CHAR_NL CHAR_LF
1237 #define CHAR_NEL ((unsigned char)'\x85')
1238 #define CHAR_ESC '\033'
1239 #define CHAR_DEL '\177'
1240 #define CHAR_NBSP ((unsigned char)'\xa0')
1241
1242 #define STR_LF "\n"
1243 #define STR_NL STR_LF
1244 #define STR_NEL "\x85"
1245 #define STR_ESC "\033"
1246 #define STR_DEL "\177"
1247
1248 #endif /* EBCDIC */
1249
1250 /* The remaining definitions work in both environments. */
1251
1252 #define CHAR_NULL '\0'
1253 #define CHAR_HT '\t'
1254 #define CHAR_VT '\v'
1255 #define CHAR_FF '\f'
1256 #define CHAR_CR '\r'
1257 #define CHAR_BS '\b'
1258 #define CHAR_BEL '\a'
1259
1260 #define CHAR_SPACE ' '
1261 #define CHAR_EXCLAMATION_MARK '!'
1262 #define CHAR_QUOTATION_MARK '"'
1263 #define CHAR_NUMBER_SIGN '#'
1264 #define CHAR_DOLLAR_SIGN '$'
1265 #define CHAR_PERCENT_SIGN '%'
1266 #define CHAR_AMPERSAND '&'
1267 #define CHAR_APOSTROPHE '\''
1268 #define CHAR_LEFT_PARENTHESIS '('
1269 #define CHAR_RIGHT_PARENTHESIS ')'
1270 #define CHAR_ASTERISK '*'
1271 #define CHAR_PLUS '+'
1272 #define CHAR_COMMA ','
1273 #define CHAR_MINUS '-'
1274 #define CHAR_DOT '.'
1275 #define CHAR_SLASH '/'
1276 #define CHAR_0 '0'
1277 #define CHAR_1 '1'
1278 #define CHAR_2 '2'
1279 #define CHAR_3 '3'
1280 #define CHAR_4 '4'
1281 #define CHAR_5 '5'
1282 #define CHAR_6 '6'
1283 #define CHAR_7 '7'
1284 #define CHAR_8 '8'
1285 #define CHAR_9 '9'
1286 #define CHAR_COLON ':'
1287 #define CHAR_SEMICOLON ';'
1288 #define CHAR_LESS_THAN_SIGN '<'
1289 #define CHAR_EQUALS_SIGN '='
1290 #define CHAR_GREATER_THAN_SIGN '>'
1291 #define CHAR_QUESTION_MARK '?'
1292 #define CHAR_COMMERCIAL_AT '@'
1293 #define CHAR_A 'A'
1294 #define CHAR_B 'B'
1295 #define CHAR_C 'C'
1296 #define CHAR_D 'D'
1297 #define CHAR_E 'E'
1298 #define CHAR_F 'F'
1299 #define CHAR_G 'G'
1300 #define CHAR_H 'H'
1301 #define CHAR_I 'I'
1302 #define CHAR_J 'J'
1303 #define CHAR_K 'K'
1304 #define CHAR_L 'L'
1305 #define CHAR_M 'M'
1306 #define CHAR_N 'N'
1307 #define CHAR_O 'O'
1308 #define CHAR_P 'P'
1309 #define CHAR_Q 'Q'
1310 #define CHAR_R 'R'
1311 #define CHAR_S 'S'
1312 #define CHAR_T 'T'
1313 #define CHAR_U 'U'
1314 #define CHAR_V 'V'
1315 #define CHAR_W 'W'
1316 #define CHAR_X 'X'
1317 #define CHAR_Y 'Y'
1318 #define CHAR_Z 'Z'
1319 #define CHAR_LEFT_SQUARE_BRACKET '['
1320 #define CHAR_BACKSLASH '\\'
1321 #define CHAR_RIGHT_SQUARE_BRACKET ']'
1322 #define CHAR_CIRCUMFLEX_ACCENT '^'
1323 #define CHAR_UNDERSCORE '_'
1324 #define CHAR_GRAVE_ACCENT '`'
1325 #define CHAR_a 'a'
1326 #define CHAR_b 'b'
1327 #define CHAR_c 'c'
1328 #define CHAR_d 'd'
1329 #define CHAR_e 'e'
1330 #define CHAR_f 'f'
1331 #define CHAR_g 'g'
1332 #define CHAR_h 'h'
1333 #define CHAR_i 'i'
1334 #define CHAR_j 'j'
1335 #define CHAR_k 'k'
1336 #define CHAR_l 'l'
1337 #define CHAR_m 'm'
1338 #define CHAR_n 'n'
1339 #define CHAR_o 'o'
1340 #define CHAR_p 'p'
1341 #define CHAR_q 'q'
1342 #define CHAR_r 'r'
1343 #define CHAR_s 's'
1344 #define CHAR_t 't'
1345 #define CHAR_u 'u'
1346 #define CHAR_v 'v'
1347 #define CHAR_w 'w'
1348 #define CHAR_x 'x'
1349 #define CHAR_y 'y'
1350 #define CHAR_z 'z'
1351 #define CHAR_LEFT_CURLY_BRACKET '{'
1352 #define CHAR_VERTICAL_LINE '|'
1353 #define CHAR_RIGHT_CURLY_BRACKET '}'
1354 #define CHAR_TILDE '~'
1355
1356 #define STR_HT "\t"
1357 #define STR_VT "\v"
1358 #define STR_FF "\f"
1359 #define STR_CR "\r"
1360 #define STR_BS "\b"
1361 #define STR_BEL "\a"
1362
1363 #define STR_SPACE " "
1364 #define STR_EXCLAMATION_MARK "!"
1365 #define STR_QUOTATION_MARK "\""
1366 #define STR_NUMBER_SIGN "#"
1367 #define STR_DOLLAR_SIGN "$"
1368 #define STR_PERCENT_SIGN "%"
1369 #define STR_AMPERSAND "&"
1370 #define STR_APOSTROPHE "'"
1371 #define STR_LEFT_PARENTHESIS "("
1372 #define STR_RIGHT_PARENTHESIS ")"
1373 #define STR_ASTERISK "*"
1374 #define STR_PLUS "+"
1375 #define STR_COMMA ","
1376 #define STR_MINUS "-"
1377 #define STR_DOT "."
1378 #define STR_SLASH "/"
1379 #define STR_0 "0"
1380 #define STR_1 "1"
1381 #define STR_2 "2"
1382 #define STR_3 "3"
1383 #define STR_4 "4"
1384 #define STR_5 "5"
1385 #define STR_6 "6"
1386 #define STR_7 "7"
1387 #define STR_8 "8"
1388 #define STR_9 "9"
1389 #define STR_COLON ":"
1390 #define STR_SEMICOLON ";"
1391 #define STR_LESS_THAN_SIGN "<"
1392 #define STR_EQUALS_SIGN "="
1393 #define STR_GREATER_THAN_SIGN ">"
1394 #define STR_QUESTION_MARK "?"
1395 #define STR_COMMERCIAL_AT "@"
1396 #define STR_A "A"
1397 #define STR_B "B"
1398 #define STR_C "C"
1399 #define STR_D "D"
1400 #define STR_E "E"
1401 #define STR_F "F"
1402 #define STR_G "G"
1403 #define STR_H "H"
1404 #define STR_I "I"
1405 #define STR_J "J"
1406 #define STR_K "K"
1407 #define STR_L "L"
1408 #define STR_M "M"
1409 #define STR_N "N"
1410 #define STR_O "O"
1411 #define STR_P "P"
1412 #define STR_Q "Q"
1413 #define STR_R "R"
1414 #define STR_S "S"
1415 #define STR_T "T"
1416 #define STR_U "U"
1417 #define STR_V "V"
1418 #define STR_W "W"
1419 #define STR_X "X"
1420 #define STR_Y "Y"
1421 #define STR_Z "Z"
1422 #define STR_LEFT_SQUARE_BRACKET "["
1423 #define STR_BACKSLASH "\\"
1424 #define STR_RIGHT_SQUARE_BRACKET "]"
1425 #define STR_CIRCUMFLEX_ACCENT "^"
1426 #define STR_UNDERSCORE "_"
1427 #define STR_GRAVE_ACCENT "`"
1428 #define STR_a "a"
1429 #define STR_b "b"
1430 #define STR_c "c"
1431 #define STR_d "d"
1432 #define STR_e "e"
1433 #define STR_f "f"
1434 #define STR_g "g"
1435 #define STR_h "h"
1436 #define STR_i "i"
1437 #define STR_j "j"
1438 #define STR_k "k"
1439 #define STR_l "l"
1440 #define STR_m "m"
1441 #define STR_n "n"
1442 #define STR_o "o"
1443 #define STR_p "p"
1444 #define STR_q "q"
1445 #define STR_r "r"
1446 #define STR_s "s"
1447 #define STR_t "t"
1448 #define STR_u "u"
1449 #define STR_v "v"
1450 #define STR_w "w"
1451 #define STR_x "x"
1452 #define STR_y "y"
1453 #define STR_z "z"
1454 #define STR_LEFT_CURLY_BRACKET "{"
1455 #define STR_VERTICAL_LINE "|"
1456 #define STR_RIGHT_CURLY_BRACKET "}"
1457 #define STR_TILDE "~"
1458
1459 #define STRING_ACCEPT0 "ACCEPT\0"
1460 #define STRING_COMMIT0 "COMMIT\0"
1461 #define STRING_F0 "F\0"
1462 #define STRING_FAIL0 "FAIL\0"
1463 #define STRING_MARK0 "MARK\0"
1464 #define STRING_PRUNE0 "PRUNE\0"
1465 #define STRING_SKIP0 "SKIP\0"
1466 #define STRING_THEN "THEN"
1467
1468 #define STRING_alpha0 "alpha\0"
1469 #define STRING_lower0 "lower\0"
1470 #define STRING_upper0 "upper\0"
1471 #define STRING_alnum0 "alnum\0"
1472 #define STRING_ascii0 "ascii\0"
1473 #define STRING_blank0 "blank\0"
1474 #define STRING_cntrl0 "cntrl\0"
1475 #define STRING_digit0 "digit\0"
1476 #define STRING_graph0 "graph\0"
1477 #define STRING_print0 "print\0"
1478 #define STRING_punct0 "punct\0"
1479 #define STRING_space0 "space\0"
1480 #define STRING_word0 "word\0"
1481 #define STRING_xdigit "xdigit"
1482
1483 #define STRING_DEFINE "DEFINE"
1484 #define STRING_WEIRD_STARTWORD "[:<:]]"
1485 #define STRING_WEIRD_ENDWORD "[:>:]]"
1486
1487 #define STRING_CR_RIGHTPAR "CR)"
1488 #define STRING_LF_RIGHTPAR "LF)"
1489 #define STRING_CRLF_RIGHTPAR "CRLF)"
1490 #define STRING_ANY_RIGHTPAR "ANY)"
1491 #define STRING_ANYCRLF_RIGHTPAR "ANYCRLF)"
1492 #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
1493 #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
1494 #define STRING_UTF8_RIGHTPAR "UTF8)"
1495 #define STRING_UTF16_RIGHTPAR "UTF16)"
1496 #define STRING_UTF32_RIGHTPAR "UTF32)"
1497 #define STRING_UTF_RIGHTPAR "UTF)"
1498 #define STRING_UCP_RIGHTPAR "UCP)"
1499 #define STRING_NO_AUTO_POSSESS_RIGHTPAR "NO_AUTO_POSSESS)"
1500 #define STRING_NO_START_OPT_RIGHTPAR "NO_START_OPT)"
1501 #define STRING_LIMIT_MATCH_EQ "LIMIT_MATCH="
1502 #define STRING_LIMIT_RECURSION_EQ "LIMIT_RECURSION="
1503
1504 #else /* SUPPORT_UTF */
1505
1506 /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
1507 works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
1508 only. */
1509
1510 #define CHAR_HT '\011'
1511 #define CHAR_VT '\013'
1512 #define CHAR_FF '\014'
1513 #define CHAR_CR '\015'
1514 #define CHAR_LF '\012'
1515 #define CHAR_NL CHAR_LF
1516 #define CHAR_NEL ((unsigned char)'\x85')
1517 #define CHAR_BS '\010'
1518 #define CHAR_BEL '\007'
1519 #define CHAR_ESC '\033'
1520 #define CHAR_DEL '\177'
1521
1522 #define CHAR_NULL '\0'
1523 #define CHAR_SPACE '\040'
1524 #define CHAR_EXCLAMATION_MARK '\041'
1525 #define CHAR_QUOTATION_MARK '\042'
1526 #define CHAR_NUMBER_SIGN '\043'
1527 #define CHAR_DOLLAR_SIGN '\044'
1528 #define CHAR_PERCENT_SIGN '\045'
1529 #define CHAR_AMPERSAND '\046'
1530 #define CHAR_APOSTROPHE '\047'
1531 #define CHAR_LEFT_PARENTHESIS '\050'
1532 #define CHAR_RIGHT_PARENTHESIS '\051'
1533 #define CHAR_ASTERISK '\052'
1534 #define CHAR_PLUS '\053'
1535 #define CHAR_COMMA '\054'
1536 #define CHAR_MINUS '\055'
1537 #define CHAR_DOT '\056'
1538 #define CHAR_SLASH '\057'
1539 #define CHAR_0 '\060'
1540 #define CHAR_1 '\061'
1541 #define CHAR_2 '\062'
1542 #define CHAR_3 '\063'
1543 #define CHAR_4 '\064'
1544 #define CHAR_5 '\065'
1545 #define CHAR_6 '\066'
1546 #define CHAR_7 '\067'
1547 #define CHAR_8 '\070'
1548 #define CHAR_9 '\071'
1549 #define CHAR_COLON '\072'
1550 #define CHAR_SEMICOLON '\073'
1551 #define CHAR_LESS_THAN_SIGN '\074'
1552 #define CHAR_EQUALS_SIGN '\075'
1553 #define CHAR_GREATER_THAN_SIGN '\076'
1554 #define CHAR_QUESTION_MARK '\077'
1555 #define CHAR_COMMERCIAL_AT '\100'
1556 #define CHAR_A '\101'
1557 #define CHAR_B '\102'
1558 #define CHAR_C '\103'
1559 #define CHAR_D '\104'
1560 #define CHAR_E '\105'
1561 #define CHAR_F '\106'
1562 #define CHAR_G '\107'
1563 #define CHAR_H '\110'
1564 #define CHAR_I '\111'
1565 #define CHAR_J '\112'
1566 #define CHAR_K '\113'
1567 #define CHAR_L '\114'
1568 #define CHAR_M '\115'
1569 #define CHAR_N '\116'
1570 #define CHAR_O '\117'
1571 #define CHAR_P '\120'
1572 #define CHAR_Q '\121'
1573 #define CHAR_R '\122'
1574 #define CHAR_S '\123'
1575 #define CHAR_T '\124'
1576 #define CHAR_U '\125'
1577 #define CHAR_V '\126'
1578 #define CHAR_W '\127'
1579 #define CHAR_X '\130'
1580 #define CHAR_Y '\131'
1581 #define CHAR_Z '\132'
1582 #define CHAR_LEFT_SQUARE_BRACKET '\133'
1583 #define CHAR_BACKSLASH '\134'
1584 #define CHAR_RIGHT_SQUARE_BRACKET '\135'
1585 #define CHAR_CIRCUMFLEX_ACCENT '\136'
1586 #define CHAR_UNDERSCORE '\137'
1587 #define CHAR_GRAVE_ACCENT '\140'
1588 #define CHAR_a '\141'
1589 #define CHAR_b '\142'
1590 #define CHAR_c '\143'
1591 #define CHAR_d '\144'
1592 #define CHAR_e '\145'
1593 #define CHAR_f '\146'
1594 #define CHAR_g '\147'
1595 #define CHAR_h '\150'
1596 #define CHAR_i '\151'
1597 #define CHAR_j '\152'
1598 #define CHAR_k '\153'
1599 #define CHAR_l '\154'
1600 #define CHAR_m '\155'
1601 #define CHAR_n '\156'
1602 #define CHAR_o '\157'
1603 #define CHAR_p '\160'
1604 #define CHAR_q '\161'
1605 #define CHAR_r '\162'
1606 #define CHAR_s '\163'
1607 #define CHAR_t '\164'
1608 #define CHAR_u '\165'
1609 #define CHAR_v '\166'
1610 #define CHAR_w '\167'
1611 #define CHAR_x '\170'
1612 #define CHAR_y '\171'
1613 #define CHAR_z '\172'
1614 #define CHAR_LEFT_CURLY_BRACKET '\173'
1615 #define CHAR_VERTICAL_LINE '\174'
1616 #define CHAR_RIGHT_CURLY_BRACKET '\175'
1617 #define CHAR_TILDE '\176'
1618 #define CHAR_NBSP ((unsigned char)'\xa0')
1619
1620 #define STR_HT "\011"
1621 #define STR_VT "\013"
1622 #define STR_FF "\014"
1623 #define STR_CR "\015"
1624 #define STR_NL "\012"
1625 #define STR_BS "\010"
1626 #define STR_BEL "\007"
1627 #define STR_ESC "\033"
1628 #define STR_DEL "\177"
1629
1630 #define STR_SPACE "\040"
1631 #define STR_EXCLAMATION_MARK "\041"
1632 #define STR_QUOTATION_MARK "\042"
1633 #define STR_NUMBER_SIGN "\043"
1634 #define STR_DOLLAR_SIGN "\044"
1635 #define STR_PERCENT_SIGN "\045"
1636 #define STR_AMPERSAND "\046"
1637 #define STR_APOSTROPHE "\047"
1638 #define STR_LEFT_PARENTHESIS "\050"
1639 #define STR_RIGHT_PARENTHESIS "\051"
1640 #define STR_ASTERISK "\052"
1641 #define STR_PLUS "\053"
1642 #define STR_COMMA "\054"
1643 #define STR_MINUS "\055"
1644 #define STR_DOT "\056"
1645 #define STR_SLASH "\057"
1646 #define STR_0 "\060"
1647 #define STR_1 "\061"
1648 #define STR_2 "\062"
1649 #define STR_3 "\063"
1650 #define STR_4 "\064"
1651 #define STR_5 "\065"
1652 #define STR_6 "\066"
1653 #define STR_7 "\067"
1654 #define STR_8 "\070"
1655 #define STR_9 "\071"
1656 #define STR_COLON "\072"
1657 #define STR_SEMICOLON "\073"
1658 #define STR_LESS_THAN_SIGN "\074"
1659 #define STR_EQUALS_SIGN "\075"
1660 #define STR_GREATER_THAN_SIGN "\076"
1661 #define STR_QUESTION_MARK "\077"
1662 #define STR_COMMERCIAL_AT "\100"
1663 #define STR_A "\101"
1664 #define STR_B "\102"
1665 #define STR_C "\103"
1666 #define STR_D "\104"
1667 #define STR_E "\105"
1668 #define STR_F "\106"
1669 #define STR_G "\107"
1670 #define STR_H "\110"
1671 #define STR_I "\111"
1672 #define STR_J "\112"
1673 #define STR_K "\113"
1674 #define STR_L "\114"
1675 #define STR_M "\115"
1676 #define STR_N "\116"
1677 #define STR_O "\117"
1678 #define STR_P "\120"
1679 #define STR_Q "\121"
1680 #define STR_R "\122"
1681 #define STR_S "\123"
1682 #define STR_T "\124"
1683 #define STR_U "\125"
1684 #define STR_V "\126"
1685 #define STR_W "\127"
1686 #define STR_X "\130"
1687 #define STR_Y "\131"
1688 #define STR_Z "\132"
1689 #define STR_LEFT_SQUARE_BRACKET "\133"
1690 #define STR_BACKSLASH "\134"
1691 #define STR_RIGHT_SQUARE_BRACKET "\135"
1692 #define STR_CIRCUMFLEX_ACCENT "\136"
1693 #define STR_UNDERSCORE "\137"
1694 #define STR_GRAVE_ACCENT "\140"
1695 #define STR_a "\141"
1696 #define STR_b "\142"
1697 #define STR_c "\143"
1698 #define STR_d "\144"
1699 #define STR_e "\145"
1700 #define STR_f "\146"
1701 #define STR_g "\147"
1702 #define STR_h "\150"
1703 #define STR_i "\151"
1704 #define STR_j "\152"
1705 #define STR_k "\153"
1706 #define STR_l "\154"
1707 #define STR_m "\155"
1708 #define STR_n "\156"
1709 #define STR_o "\157"
1710 #define STR_p "\160"
1711 #define STR_q "\161"
1712 #define STR_r "\162"
1713 #define STR_s "\163"
1714 #define STR_t "\164"
1715 #define STR_u "\165"
1716 #define STR_v "\166"
1717 #define STR_w "\167"
1718 #define STR_x "\170"
1719 #define STR_y "\171"
1720 #define STR_z "\172"
1721 #define STR_LEFT_CURLY_BRACKET "\173"
1722 #define STR_VERTICAL_LINE "\174"
1723 #define STR_RIGHT_CURLY_BRACKET "\175"
1724 #define STR_TILDE "\176"
1725
1726 #define STRING_ACCEPT0 STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1727 #define STRING_COMMIT0 STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1728 #define STRING_F0 STR_F "\0"
1729 #define STRING_FAIL0 STR_F STR_A STR_I STR_L "\0"
1730 #define STRING_MARK0 STR_M STR_A STR_R STR_K "\0"
1731 #define STRING_PRUNE0 STR_P STR_R STR_U STR_N STR_E "\0"
1732 #define STRING_SKIP0 STR_S STR_K STR_I STR_P "\0"
1733 #define STRING_THEN STR_T STR_H STR_E STR_N
1734
1735 #define STRING_alpha0 STR_a STR_l STR_p STR_h STR_a "\0"
1736 #define STRING_lower0 STR_l STR_o STR_w STR_e STR_r "\0"
1737 #define STRING_upper0 STR_u STR_p STR_p STR_e STR_r "\0"
1738 #define STRING_alnum0 STR_a STR_l STR_n STR_u STR_m "\0"
1739 #define STRING_ascii0 STR_a STR_s STR_c STR_i STR_i "\0"
1740 #define STRING_blank0 STR_b STR_l STR_a STR_n STR_k "\0"
1741 #define STRING_cntrl0 STR_c STR_n STR_t STR_r STR_l "\0"
1742 #define STRING_digit0 STR_d STR_i STR_g STR_i STR_t "\0"
1743 #define STRING_graph0 STR_g STR_r STR_a STR_p STR_h "\0"
1744 #define STRING_print0 STR_p STR_r STR_i STR_n STR_t "\0"
1745 #define STRING_punct0 STR_p STR_u STR_n STR_c STR_t "\0"
1746 #define STRING_space0 STR_s STR_p STR_a STR_c STR_e "\0"
1747 #define STRING_word0 STR_w STR_o STR_r STR_d "\0"
1748 #define STRING_xdigit STR_x STR_d STR_i STR_g STR_i STR_t
1749
1750 #define STRING_DEFINE STR_D STR_E STR_F STR_I STR_N STR_E
1751 #define STRING_WEIRD_STARTWORD STR_LEFT_SQUARE_BRACKET STR_COLON STR_LESS_THAN_SIGN STR_COLON STR_RIGHT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET
1752 #define STRING_WEIRD_ENDWORD STR_LEFT_SQUARE_BRACKET STR_COLON STR_GREATER_THAN_SIGN STR_COLON STR_RIGHT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET
1753
1754 #define STRING_CR_RIGHTPAR STR_C STR_R STR_RIGHT_PARENTHESIS
1755 #define STRING_LF_RIGHTPAR STR_L STR_F STR_RIGHT_PARENTHESIS
1756 #define STRING_CRLF_RIGHTPAR STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1757 #define STRING_ANY_RIGHTPAR STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1758 #define STRING_ANYCRLF_RIGHTPAR STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1759 #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
1760 #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
1761 #define STRING_UTF8_RIGHTPAR STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1762 #define STRING_UTF16_RIGHTPAR STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS
1763 #define STRING_UTF32_RIGHTPAR STR_U STR_T STR_F STR_3 STR_2 STR_RIGHT_PARENTHESIS
1764 #define STRING_UTF_RIGHTPAR STR_U STR_T STR_F STR_RIGHT_PARENTHESIS
1765 #define STRING_UCP_RIGHTPAR STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1766 #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
1767 #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
1768 #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
1769 #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
1770
1771 #endif /* SUPPORT_UTF */
1772
1773 /* Escape items that are just an encoding of a particular data value. */
1774
1775 #ifndef ESC_a
1776 #define ESC_a CHAR_BEL
1777 #endif
1778
1779 #ifndef ESC_e
1780 #define ESC_e CHAR_ESC
1781 #endif
1782
1783 #ifndef ESC_f
1784 #define ESC_f CHAR_FF
1785 #endif
1786
1787 #ifndef ESC_n
1788 #define ESC_n CHAR_LF
1789 #endif
1790
1791 #ifndef ESC_r
1792 #define ESC_r CHAR_CR
1793 #endif
1794
1795 /* We can't officially use ESC_t because it is a POSIX reserved identifier
1796 (presumably because of all the others like size_t). */
1797
1798 #ifndef ESC_tee
1799 #define ESC_tee CHAR_HT
1800 #endif
1801
1802 /* Codes for different types of Unicode property */
1803
1804 #define PT_ANY 0 /* Any property - matches all chars */
1805 #define PT_LAMP 1 /* L& - the union of Lu, Ll, Lt */
1806 #define PT_GC 2 /* Specified general characteristic (e.g. L) */
1807 #define PT_PC 3 /* Specified particular characteristic (e.g. Lu) */
1808 #define PT_SC 4 /* Script (e.g. Han) */
1809 #define PT_ALNUM 5 /* Alphanumeric - the union of L and N */
1810 #define PT_SPACE 6 /* Perl space - Z plus 9,10,12,13 */
1811 #define PT_PXSPACE 7 /* POSIX space - Z plus 9,10,11,12,13 */
1812 #define PT_WORD 8 /* Word - L plus N plus underscore */
1813 #define PT_CLIST 9 /* Pseudo-property: match character list */
1814 #define PT_UCNC 10 /* Universal Character nameable character */
1815 #define PT_TABSIZE 11 /* Size of square table for autopossessify tests */
1816
1817 /* The following special properties are used only in XCLASS items, when POSIX
1818 classes are specified and PCRE_UCP is set - in other words, for Unicode
1819 handling of these classes. They are not available via the \p or \P escapes like
1820 those in the above list, and so they do not take part in the autopossessifying
1821 table. */
1822
1823 #define PT_PXGRAPH 11 /* [:graph:] - characters that mark the paper */
1824 #define PT_PXPRINT 12 /* [:print:] - [:graph:] plus non-control spaces */
1825 #define PT_PXPUNCT 13 /* [:punct:] - punctuation characters */
1826
1827 /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
1828 contain characters with values greater than 255. */
1829
1830 #define XCL_NOT 0x01 /* Flag: this is a negative class */
1831 #define XCL_MAP 0x02 /* Flag: a 32-byte map is present */
1832 #define XCL_HASPROP 0x04 /* Flag: property checks are present. */
1833
1834 #define XCL_END 0 /* Marks end of individual items */
1835 #define XCL_SINGLE 1 /* Single item (one multibyte char) follows */
1836 #define XCL_RANGE 2 /* A range (two multibyte chars) follows */
1837 #define XCL_PROP 3 /* Unicode property (2-byte property code follows) */
1838 #define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */
1839
1840 /* These are escaped items that aren't just an encoding of a particular data
1841 value such as \n. They must have non-zero values, as check_escape() returns 0
1842 for a data character. Also, they must appear in the same order as in the
1843 opcode definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it
1844 corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1845 used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In
1846 non-DOTALL mode, "." behaves like \N.
1847
1848 The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1849 when PCRE_UCP is set and replacement of \d etc by \p sequences is required.
1850 They must be contiguous, and remain in order so that the replacements can be
1851 looked up from a table.
1852
1853 Negative numbers are used to encode a backreference (\1, \2, \3, etc.) in
1854 check_escape(). There are two tests in the code for an escape
1855 greater than ESC_b and less than ESC_Z to detect the types that may be
1856 repeated. These are the types that consume characters. If any new escapes are
1857 put in between that don't consume a character, that code will have to change.
1858 */
1859
1860 enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
1861 ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1862 ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1863 ESC_E, ESC_Q, ESC_g, ESC_k,
1864 ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };
1865
1866
1867 /********************** Opcode definitions ******************/
1868
1869 /****** NOTE NOTE NOTE ******
1870
1871 Starting from 1 (i.e. after OP_END), the values up to OP_EOD must correspond in
1872 order to the list of escapes immediately above. Furthermore, values up to
1873 OP_DOLLM must not be changed without adjusting the table called autoposstab in
1874 pcre_compile.c
1875
1876 Whenever this list is updated, the two macro definitions that follow must be
1877 updated to match. The possessification table called "opcode_possessify" in
1878 pcre_compile.c must also be updated, and also the tables called "coptable"
1879 and "poptable" in pcre_dfa_exec.c.
1880
1881 ****** NOTE NOTE NOTE ******/
1882
1883
1884 /* The values between FIRST_AUTOTAB_OP and LAST_AUTOTAB_RIGHT_OP, inclusive,
1885 are used in a table for deciding whether a repeated character type can be
1886 auto-possessified. */
1887
1888 #define FIRST_AUTOTAB_OP OP_NOT_DIGIT
1889 #define LAST_AUTOTAB_LEFT_OP OP_EXTUNI
1890 #define LAST_AUTOTAB_RIGHT_OP OP_DOLLM
1891
1892 enum {
1893 OP_END, /* 0 End of pattern */
1894
1895 /* Values corresponding to backslashed metacharacters */
1896
1897 OP_SOD, /* 1 Start of data: \A */
1898 OP_SOM, /* 2 Start of match (subject + offset): \G */
1899 OP_SET_SOM, /* 3 Set start of match (\K) */
1900 OP_NOT_WORD_BOUNDARY, /* 4 \B */
1901 OP_WORD_BOUNDARY, /* 5 \b */
1902 OP_NOT_DIGIT, /* 6 \D */
1903 OP_DIGIT, /* 7 \d */
1904 OP_NOT_WHITESPACE, /* 8 \S */
1905 OP_WHITESPACE, /* 9 \s */
1906 OP_NOT_WORDCHAR, /* 10 \W */
1907 OP_WORDCHAR, /* 11 \w */
1908
1909 OP_ANY, /* 12 Match any character except newline (\N) */
1910 OP_ALLANY, /* 13 Match any character */
1911 OP_ANYBYTE, /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1912 OP_NOTPROP, /* 15 \P (not Unicode property) */
1913 OP_PROP, /* 16 \p (Unicode property) */
1914 OP_ANYNL, /* 17 \R (any newline sequence) */
1915 OP_NOT_HSPACE, /* 18 \H (not horizontal whitespace) */
1916 OP_HSPACE, /* 19 \h (horizontal whitespace) */
1917 OP_NOT_VSPACE, /* 20 \V (not vertical whitespace) */
1918 OP_VSPACE, /* 21 \v (vertical whitespace) */
1919 OP_EXTUNI, /* 22 \X (extended Unicode sequence */
1920 OP_EODN, /* 23 End of data or \n at end of data (\Z) */
1921 OP_EOD, /* 24 End of data (\z) */
1922
1923 /* Line end assertions */
1924
1925 OP_DOLL, /* 25 End of line - not multiline */
1926 OP_DOLLM, /* 26 End of line - multiline */
1927 OP_CIRC, /* 27 Start of line - not multiline */
1928 OP_CIRCM, /* 28 Start of line - multiline */
1929
1930 /* Single characters; caseful must precede the caseless ones */
1931
1932 OP_CHAR, /* 29 Match one character, casefully */
1933 OP_CHARI, /* 30 Match one character, caselessly */
1934 OP_NOT, /* 31 Match one character, not the given one, casefully */
1935 OP_NOTI, /* 32 Match one character, not the given one, caselessly */
1936
1937 /* The following sets of 13 opcodes must always be kept in step because
1938 the offset from the first one is used to generate the others. */
1939
1940 /* Repeated characters; caseful must precede the caseless ones */
1941
1942 OP_STAR, /* 33 The maximizing and minimizing versions of */
1943 OP_MINSTAR, /* 34 these six opcodes must come in pairs, with */
1944 OP_PLUS, /* 35 the minimizing one second. */
1945 OP_MINPLUS, /* 36 */
1946 OP_QUERY, /* 37 */
1947 OP_MINQUERY, /* 38 */
1948
1949 OP_UPTO, /* 39 From 0 to n matches of one character, caseful*/
1950 OP_MINUPTO, /* 40 */
1951 OP_EXACT, /* 41 Exactly n matches */
1952
1953 OP_POSSTAR, /* 42 Possessified star, caseful */
1954 OP_POSPLUS, /* 43 Possessified plus, caseful */
1955 OP_POSQUERY, /* 44 Posesssified query, caseful */
1956 OP_POSUPTO, /* 45 Possessified upto, caseful */
1957
1958 /* Repeated characters; caseless must follow the caseful ones */
1959
1960 OP_STARI, /* 46 */
1961 OP_MINSTARI, /* 47 */
1962 OP_PLUSI, /* 48 */
1963 OP_MINPLUSI, /* 49 */
1964 OP_QUERYI, /* 50 */
1965 OP_MINQUERYI, /* 51 */
1966
1967 OP_UPTOI, /* 52 From 0 to n matches of one character, caseless */
1968 OP_MINUPTOI, /* 53 */
1969 OP_EXACTI, /* 54 */
1970
1971 OP_POSSTARI, /* 55 Possessified star, caseless */
1972 OP_POSPLUSI, /* 56 Possessified plus, caseless */
1973 OP_POSQUERYI, /* 57 Posesssified query, caseless */
1974 OP_POSUPTOI, /* 58 Possessified upto, caseless */
1975
1976 /* The negated ones must follow the non-negated ones, and match them */
1977 /* Negated repeated character, caseful; must precede the caseless ones */
1978
1979 OP_NOTSTAR, /* 59 The maximizing and minimizing versions of */
1980 OP_NOTMINSTAR, /* 60 these six opcodes must come in pairs, with */
1981 OP_NOTPLUS, /* 61 the minimizing one second. They must be in */
1982 OP_NOTMINPLUS, /* 62 exactly the same order as those above. */
1983 OP_NOTQUERY, /* 63 */
1984 OP_NOTMINQUERY, /* 64 */
1985
1986 OP_NOTUPTO, /* 65 From 0 to n matches, caseful */
1987 OP_NOTMINUPTO, /* 66 */
1988 OP_NOTEXACT, /* 67 Exactly n matches */
1989
1990 OP_NOTPOSSTAR, /* 68 Possessified versions, caseful */
1991 OP_NOTPOSPLUS, /* 69 */
1992 OP_NOTPOSQUERY, /* 70 */
1993 OP_NOTPOSUPTO, /* 71 */
1994
1995 /* Negated repeated character, caseless; must follow the caseful ones */
1996
1997 OP_NOTSTARI, /* 72 */
1998 OP_NOTMINSTARI, /* 73 */
1999 OP_NOTPLUSI, /* 74 */
2000 OP_NOTMINPLUSI, /* 75 */
2001 OP_NOTQUERYI, /* 76 */
2002 OP_NOTMINQUERYI, /* 77 */
2003
2004 OP_NOTUPTOI, /* 78 From 0 to n matches, caseless */
2005 OP_NOTMINUPTOI, /* 79 */
2006 OP_NOTEXACTI, /* 80 Exactly n matches */
2007
2008 OP_NOTPOSSTARI, /* 81 Possessified versions, caseless */
2009 OP_NOTPOSPLUSI, /* 82 */
2010 OP_NOTPOSQUERYI, /* 83 */
2011 OP_NOTPOSUPTOI, /* 84 */
2012
2013 /* Character types */
2014
2015 OP_TYPESTAR, /* 85 The maximizing and minimizing versions of */
2016 OP_TYPEMINSTAR, /* 86 these six opcodes must come in pairs, with */
2017 OP_TYPEPLUS, /* 87 the minimizing one second. These codes must */
2018 OP_TYPEMINPLUS, /* 88 be in exactly the same order as those above. */
2019 OP_TYPEQUERY, /* 89 */
2020 OP_TYPEMINQUERY, /* 90 */
2021
2022 OP_TYPEUPTO, /* 91 From 0 to n matches */
2023 OP_TYPEMINUPTO, /* 92 */
2024 OP_TYPEEXACT, /* 93 Exactly n matches */
2025
2026 OP_TYPEPOSSTAR, /* 94 Possessified versions */
2027 OP_TYPEPOSPLUS, /* 95 */
2028 OP_TYPEPOSQUERY, /* 96 */
2029 OP_TYPEPOSUPTO, /* 97 */
2030
2031 /* These are used for character classes and back references; only the
2032 first six are the same as the sets above. */
2033
2034 OP_CRSTAR, /* 98 The maximizing and minimizing versions of */
2035 OP_CRMINSTAR, /* 99 all these opcodes must come in pairs, with */
2036 OP_CRPLUS, /* 100 the minimizing one second. These codes must */
2037 OP_CRMINPLUS, /* 101 be in exactly the same order as those above. */
2038 OP_CRQUERY, /* 102 */
2039 OP_CRMINQUERY, /* 103 */
2040
2041 OP_CRRANGE, /* 104 These are different to the three sets above. */
2042 OP_CRMINRANGE, /* 105 */
2043
2044 OP_CRPOSSTAR, /* 106 Possessified versions */
2045 OP_CRPOSPLUS, /* 107 */
2046 OP_CRPOSQUERY, /* 108 */
2047 OP_CRPOSRANGE, /* 109 */
2048
2049 /* End of quantifier opcodes */
2050
2051 OP_CLASS, /* 110 Match a character class, chars < 256 only */
2052 OP_NCLASS, /* 111 Same, but the bitmap was created from a negative
2053 class - the difference is relevant only when a
2054 character > 255 is encountered. */
2055 OP_XCLASS, /* 112 Extended class for handling > 255 chars within the
2056 class. This does both positive and negative. */
2057 OP_REF, /* 113 Match a back reference, casefully */
2058 OP_REFI, /* 114 Match a back reference, caselessly */
2059 OP_DNREF, /* 115 Match a duplicate name backref, casefully */
2060 OP_DNREFI, /* 116 Match a duplicate name backref, caselessly */
2061 OP_RECURSE, /* 117 Match a numbered subpattern (possibly recursive) */
2062 OP_CALLOUT, /* 118 Call out to external function if provided */
2063
2064 OP_ALT, /* 119 Start of alternation */
2065 OP_KET, /* 120 End of group that doesn't have an unbounded repeat */
2066 OP_KETRMAX, /* 121 These two must remain together and in this */
2067 OP_KETRMIN, /* 122 order. They are for groups the repeat for ever. */
2068 OP_KETRPOS, /* 123 Possessive unlimited repeat. */
2069
2070 /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
2071 asserts must remain in order. */
2072
2073 OP_REVERSE, /* 124 Move pointer back - used in lookbehind assertions */
2074 OP_ASSERT, /* 125 Positive lookahead */
2075 OP_ASSERT_NOT, /* 126 Negative lookahead */
2076 OP_ASSERTBACK, /* 127 Positive lookbehind */
2077 OP_ASSERTBACK_NOT, /* 128 Negative lookbehind */
2078
2079 /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
2080 after the assertions, with ONCE first, as there's a test for >= ONCE for a
2081 subpattern that isn't an assertion. The POS versions must immediately follow
2082 the non-POS versions in each case. */
2083
2084 OP_ONCE, /* 129 Atomic group, contains captures */
2085 OP_ONCE_NC, /* 130 Atomic group containing no captures */
2086 OP_BRA, /* 131 Start of non-capturing bracket */
2087 OP_BRAPOS, /* 132 Ditto, with unlimited, possessive repeat */
2088 OP_CBRA, /* 133 Start of capturing bracket */
2089 OP_CBRAPOS, /* 134 Ditto, with unlimited, possessive repeat */
2090 OP_COND, /* 135 Conditional group */
2091
2092 /* These five must follow the previous five, in the same order. There's a
2093 check for >= SBRA to distinguish the two sets. */
2094
2095 OP_SBRA, /* 136 Start of non-capturing bracket, check empty */
2096 OP_SBRAPOS, /* 137 Ditto, with unlimited, possessive repeat */
2097 OP_SCBRA, /* 138 Start of capturing bracket, check empty */
2098 OP_SCBRAPOS, /* 139 Ditto, with unlimited, possessive repeat */
2099 OP_SCOND, /* 140 Conditional group, check empty */
2100
2101 /* The next two pairs must (respectively) be kept together. */
2102
2103 OP_CREF, /* 141 Used to hold a capture number as condition */
2104 OP_DNCREF, /* 142 Used to point to duplicate names as a condition */
2105 OP_RREF, /* 143 Used to hold a recursion number as condition */
2106 OP_DNRREF, /* 144 Used to point to duplicate names as a condition */
2107 OP_DEF, /* 145 The DEFINE condition */
2108
2109 OP_BRAZERO, /* 146 These two must remain together and in this */
2110 OP_BRAMINZERO, /* 147 order. */
2111 OP_BRAPOSZERO, /* 148 */
2112
2113 /* These are backtracking control verbs */
2114
2115 OP_MARK, /* 149 always has an argument */
2116 OP_PRUNE, /* 150 */
2117 OP_PRUNE_ARG, /* 151 same, but with argument */
2118 OP_SKIP, /* 152 */
2119 OP_SKIP_ARG, /* 153 same, but with argument */
2120 OP_THEN, /* 154 */
2121 OP_THEN_ARG, /* 155 same, but with argument */
2122 OP_COMMIT, /* 156 */
2123
2124 /* These are forced failure and success verbs */
2125
2126 OP_FAIL, /* 157 */
2127 OP_ACCEPT, /* 158 */
2128 OP_ASSERT_ACCEPT, /* 159 Used inside assertions */
2129 OP_CLOSE, /* 160 Used before OP_ACCEPT to close open captures */
2130
2131 /* This is used to skip a subpattern with a {0} quantifier */
2132
2133 OP_SKIPZERO, /* 161 */
2134
2135 /* This is not an opcode, but is used to check that tables indexed by opcode
2136 are the correct length, in order to catch updating errors - there have been
2137 some in the past. */
2138
2139 OP_TABLE_LENGTH
2140 };
2141
2142 /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
2143 definitions that follow must also be updated to match. There are also tables
2144 called "opcode_possessify" in pcre_compile.c and "coptable" and "poptable" in
2145 pcre_dfa_exec.c that must be updated. */
2146
2147
2148 /* This macro defines textual names for all the opcodes. These are used only
2149 for debugging, and some of them are only partial names. The macro is referenced
2150 only in pcre_printint.c, which fills out the full names in many cases (and in
2151 some cases doesn't actually use these names at all). */
2152
2153 #define OP_NAME_LIST \
2154 "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d", \
2155 "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte", \
2156 "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v", \
2157 "extuni", "\\Z", "\\z", \
2158 "$", "$", "^", "^", "char", "chari", "not", "noti", \
2159 "*", "*?", "+", "+?", "?", "??", \
2160 "{", "{", "{", \
2161 "*+","++", "?+", "{", \
2162 "*", "*?", "+", "+?", "?", "??", \
2163 "{", "{", "{", \
2164 "*+","++", "?+", "{", \
2165 "*", "*?", "+", "+?", "?", "??", \
2166 "{", "{", "{", \
2167 "*+","++", "?+", "{", \
2168 "*", "*?", "+", "+?", "?", "??", \
2169 "{", "{", "{", \
2170 "*+","++", "?+", "{", \
2171 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
2172 "*+","++", "?+", "{", \
2173 "*", "*?", "+", "+?", "?", "??", "{", "{", \
2174 "*+","++", "?+", "{", \
2175 "class", "nclass", "xclass", "Ref", "Refi", "DnRef", "DnRefi", \
2176 "Recurse", "Callout", \
2177 "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos", \
2178 "Reverse", "Assert", "Assert not", "AssertB", "AssertB not", \
2179 "Once", "Once_NC", \
2180 "Bra", "BraPos", "CBra", "CBraPos", \
2181 "Cond", \
2182 "SBra", "SBraPos", "SCBra", "SCBraPos", \
2183 "SCond", \
2184 "Cond ref", "Cond dnref", "Cond rec", "Cond dnrec", "Cond def", \
2185 "Brazero", "Braminzero", "Braposzero", \
2186 "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP", \
2187 "*THEN", "*THEN", "*COMMIT", "*FAIL", \
2188 "*ACCEPT", "*ASSERT_ACCEPT", \
2189 "Close", "Skip zero"
2190
2191
2192 /* This macro defines the length of fixed length operations in the compiled
2193 regex. The lengths are used when searching for specific things, and also in the
2194 debugging printing of a compiled regex. We use a macro so that it can be
2195 defined close to the definitions of the opcodes themselves.
2196
2197 As things have been extended, some of these are no longer fixed lenths, but are
2198 minima instead. For example, the length of a single-character repeat may vary
2199 in UTF-8 mode. The code that uses this table must know about such things. */
2200
2201 #define OP_LENGTHS \
2202 1, /* End */ \
2203 1, 1, 1, 1, 1, /* \A, \G, \K, \B, \b */ \
2204 1, 1, 1, 1, 1, 1, /* \D, \d, \S, \s, \W, \w */ \
2205 1, 1, 1, /* Any, AllAny, Anybyte */ \
2206 3, 3, /* \P, \p */ \
2207 1, 1, 1, 1, 1, /* \R, \H, \h, \V, \v */ \
2208 1, /* \X */ \
2209 1, 1, 1, 1, 1, 1, /* \Z, \z, $, $M ^, ^M */ \
2210 2, /* Char - the minimum length */ \
2211 2, /* Chari - the minimum length */ \
2212 2, /* not */ \
2213 2, /* noti */ \
2214 /* Positive single-char repeats ** These are */ \
2215 2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** minima in */ \
2216 2+IMM2_SIZE, 2+IMM2_SIZE, /* upto, minupto ** mode */ \
2217 2+IMM2_SIZE, /* exact */ \
2218 2, 2, 2, 2+IMM2_SIZE, /* *+, ++, ?+, upto+ */ \
2219 2, 2, 2, 2, 2, 2, /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8 */ \
2220 2+IMM2_SIZE, 2+IMM2_SIZE, /* upto I, minupto I */ \
2221 2+IMM2_SIZE, /* exact I */ \
2222 2, 2, 2, 2+IMM2_SIZE, /* *+I, ++I, ?+I, upto+I */ \
2223 /* Negative single-char repeats - only for chars < 256 */ \
2224 2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \
2225 2+IMM2_SIZE, 2+IMM2_SIZE, /* NOT upto, minupto */ \
2226 2+IMM2_SIZE, /* NOT exact */ \
2227 2, 2, 2, 2+IMM2_SIZE, /* Possessive NOT *, +, ?, upto */ \
2228 2, 2, 2, 2, 2, 2, /* NOT *I, *?I, +I, +?I, ?I, ??I */ \
2229 2+IMM2_SIZE, 2+IMM2_SIZE, /* NOT upto I, minupto I */ \
2230 2+IMM2_SIZE, /* NOT exact I */ \
2231 2, 2, 2, 2+IMM2_SIZE, /* Possessive NOT *I, +I, ?I, upto I */ \
2232 /* Positive type repeats */ \
2233 2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \
2234 2+IMM2_SIZE, 2+IMM2_SIZE, /* Type upto, minupto */ \
2235 2+IMM2_SIZE, /* Type exact */ \
2236 2, 2, 2, 2+IMM2_SIZE, /* Possessive *+, ++, ?+, upto+ */ \
2237 /* Character class & ref repeats */ \
2238 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \
2239 1+2*IMM2_SIZE, 1+2*IMM2_SIZE, /* CRRANGE, CRMINRANGE */ \
2240 1, 1, 1, 1+2*IMM2_SIZE, /* Possessive *+, ++, ?+, CRPOSRANGE */ \
2241 1+(32/sizeof(pcre_uchar)), /* CLASS */ \
2242 1+(32/sizeof(pcre_uchar)), /* NCLASS */ \
2243 0, /* XCLASS - variable length */ \
2244 1+IMM2_SIZE, /* REF */ \
2245 1+IMM2_SIZE, /* REFI */ \
2246 1+2*IMM2_SIZE, /* DNREF */ \
2247 1+2*IMM2_SIZE, /* DNREFI */ \
2248 1+LINK_SIZE, /* RECURSE */ \
2249 2+2*LINK_SIZE, /* CALLOUT */ \
2250 1+LINK_SIZE, /* Alt */ \
2251 1+LINK_SIZE, /* Ket */ \
2252 1+LINK_SIZE, /* KetRmax */ \
2253 1+LINK_SIZE, /* KetRmin */ \
2254 1+LINK_SIZE, /* KetRpos */ \
2255 1+LINK_SIZE, /* Reverse */ \
2256 1+LINK_SIZE, /* Assert */ \
2257 1+LINK_SIZE, /* Assert not */ \
2258 1+LINK_SIZE, /* Assert behind */ \
2259 1+LINK_SIZE, /* Assert behind not */ \
2260 1+LINK_SIZE, /* ONCE */ \
2261 1+LINK_SIZE, /* ONCE_NC */ \
2262 1+LINK_SIZE, /* BRA */ \
2263 1+LINK_SIZE, /* BRAPOS */ \
2264 1+LINK_SIZE+IMM2_SIZE, /* CBRA */ \
2265 1+LINK_SIZE+IMM2_SIZE, /* CBRAPOS */ \
2266 1+LINK_SIZE, /* COND */ \
2267 1+LINK_SIZE, /* SBRA */ \
2268 1+LINK_SIZE, /* SBRAPOS */ \
2269 1+LINK_SIZE+IMM2_SIZE, /* SCBRA */ \
2270 1+LINK_SIZE+IMM2_SIZE, /* SCBRAPOS */ \
2271 1+LINK_SIZE, /* SCOND */ \
2272 1+IMM2_SIZE, 1+2*IMM2_SIZE, /* CREF, DNCREF */ \
2273 1+IMM2_SIZE, 1+2*IMM2_SIZE, /* RREF, DNRREF */ \
2274 1, /* DEF */ \
2275 1, 1, 1, /* BRAZERO, BRAMINZERO, BRAPOSZERO */ \
2276 3, 1, 3, /* MARK, PRUNE, PRUNE_ARG */ \
2277 1, 3, /* SKIP, SKIP_ARG */ \
2278 1, 3, /* THEN, THEN_ARG */ \
2279 1, 1, 1, 1, /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT */ \
2280 1+IMM2_SIZE, 1 /* CLOSE, SKIPZERO */
2281
2282 /* A magic value for OP_RREF to indicate the "any recursion" condition. */
2283
2284 #define RREF_ANY 0xffff
2285
2286 /* Compile time error code numbers. They are given names so that they can more
2287 easily be tracked. When a new number is added, the table called eint in
2288 pcreposix.c must be updated. */
2289
2290 enum { ERR0, ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9,
2291 ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
2292 ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
2293 ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
2294 ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
2295 ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
2296 ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
2297 ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERR79,
2298 ERR80, ERR81, ERR82, ERR83, ERR84, ERR85, ERR86, ERR87, ERRCOUNT };
2299
2300 /* JIT compiling modes. The function list is indexed by them. */
2301
2302 enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
2303 JIT_NUMBER_OF_COMPILE_MODES };
2304
2305 /* The real format of the start of the pcre block; the index of names and the
2306 code vector run on as long as necessary after the end. We store an explicit
2307 offset to the name table so that if a regex is compiled on one host, saved, and
2308 then run on another where the size of pointers is different, all might still
2309 be well.
2310
2311 The size of the structure must be a multiple of 8 bytes. For the case of
2312 compiled-on-4 and run-on-8, we include an extra pointer that is always NULL so
2313 that there are an even number of pointers which therefore are a multiple of 8
2314 bytes.
2315
2316 It is necessary to fork the struct for the 32 bit library, since it needs to
2317 use pcre_uint32 for first_char and req_char. We can't put an ifdef inside the
2318 typedef because pcretest needs access to the struct of the 8-, 16- and 32-bit
2319 variants.
2320
2321 *** WARNING ***
2322 When new fields are added to these structures, remember to adjust the code in
2323 pcre_byte_order.c that is concerned with swapping the byte order of the fields
2324 when a compiled regex is reloaded on a host with different endianness.
2325 *** WARNING ***
2326 There is also similar byte-flipping code in pcretest.c, which is used for
2327 testing the byte-flipping features. It must also be kept in step.
2328 *** WARNING ***
2329 */
2330
2331 typedef struct real_pcre8_or_16 {
2332 pcre_uint32 magic_number;
2333 pcre_uint32 size; /* Total that was malloced */
2334 pcre_uint32 options; /* Public options */
2335 pcre_uint32 flags; /* Private flags */
2336 pcre_uint32 limit_match; /* Limit set from regex */
2337 pcre_uint32 limit_recursion; /* Limit set from regex */
2338 pcre_uint16 first_char; /* Starting character */
2339 pcre_uint16 req_char; /* This character must be seen */
2340 pcre_uint16 max_lookbehind; /* Longest lookbehind (characters) */
2341 pcre_uint16 top_bracket; /* Highest numbered group */
2342 pcre_uint16 top_backref; /* Highest numbered back reference */
2343 pcre_uint16 name_table_offset; /* Offset to name table that follows */
2344 pcre_uint16 name_entry_size; /* Size of any name items */
2345 pcre_uint16 name_count; /* Number of name items */
2346 pcre_uint16 ref_count; /* Reference count */
2347 pcre_uint16 dummy1; /* To ensure size is a multiple of 8 */
2348 pcre_uint16 dummy2; /* To ensure size is a multiple of 8 */
2349 pcre_uint16 dummy3; /* To ensure size is a multiple of 8 */
2350 const pcre_uint8 *tables; /* Pointer to tables or NULL for std */
2351 void *nullpad; /* NULL padding */
2352 } real_pcre8_or_16;
2353
2354 typedef struct real_pcre8_or_16 real_pcre;
2355 typedef struct real_pcre8_or_16 real_pcre16;
2356
2357 typedef struct real_pcre32 {
2358 pcre_uint32 magic_number;
2359 pcre_uint32 size; /* Total that was malloced */
2360 pcre_uint32 options; /* Public options */
2361 pcre_uint32 flags; /* Private flags */
2362 pcre_uint32 limit_match; /* Limit set from regex */
2363 pcre_uint32 limit_recursion; /* Limit set from regex */
2364 pcre_uint32 first_char; /* Starting character */
2365 pcre_uint32 req_char; /* This character must be seen */
2366 pcre_uint16 max_lookbehind; /* Longest lookbehind (characters) */
2367 pcre_uint16 top_bracket; /* Highest numbered group */
2368 pcre_uint16 top_backref; /* Highest numbered back reference */
2369 pcre_uint16 name_table_offset; /* Offset to name table that follows */
2370 pcre_uint16 name_entry_size; /* Size of any name items */
2371 pcre_uint16 name_count; /* Number of name items */
2372 pcre_uint16 ref_count; /* Reference count */
2373 pcre_uint16 dummy; /* To ensure size is a multiple of 8 */
2374 const pcre_uint8 *tables; /* Pointer to tables or NULL for std */
2375 void *nullpad; /* NULL padding */
2376 } real_pcre32;
2377
2378 #if defined COMPILE_PCRE8
2379 #define REAL_PCRE real_pcre
2380 #elif defined COMPILE_PCRE16
2381 #define REAL_PCRE real_pcre16
2382 #elif defined COMPILE_PCRE32
2383 #define REAL_PCRE real_pcre32
2384 #endif
2385
2386 /* Assert that the size of REAL_PCRE is divisible by 8 */
2387 typedef int __assert_real_pcre_size_divisible_8[(sizeof(REAL_PCRE) % 8) == 0 ? 1 : -1];
2388
2389 /* Needed in pcretest to access some fields in the real_pcre* structures
2390 * directly. They're unified for 8/16/32 bits since the structs only differ
2391 * after these fields; if that ever changes, need to fork those defines into
2392 * 8/16 and 32 bit versions. */
2393 #define REAL_PCRE_MAGIC(re) (((REAL_PCRE*)re)->magic_number)
2394 #define REAL_PCRE_SIZE(re) (((REAL_PCRE*)re)->size)
2395 #define REAL_PCRE_OPTIONS(re) (((REAL_PCRE*)re)->options)
2396 #define REAL_PCRE_FLAGS(re) (((REAL_PCRE*)re)->flags)
2397
2398 /* The format of the block used to store data from pcre_study(). The same
2399 remark (see NOTE above) about extending this structure applies. */
2400
2401 typedef struct pcre_study_data {
2402 pcre_uint32 size; /* Total that was malloced */
2403 pcre_uint32 flags; /* Private flags */
2404 pcre_uint8 start_bits[32]; /* Starting char bits */
2405 pcre_uint32 minlength; /* Minimum subject length */
2406 } pcre_study_data;
2407
2408 /* Structure for building a chain of open capturing subpatterns during
2409 compiling, so that instructions to close them can be compiled when (*ACCEPT) is
2410 encountered. This is also used to identify subpatterns that contain recursive
2411 back references to themselves, so that they can be made atomic. */
2412
2413 typedef struct open_capitem {
2414 struct open_capitem *next; /* Chain link */
2415 pcre_uint16 number; /* Capture number */
2416 pcre_uint16 flag; /* Set TRUE if recursive back ref */
2417 } open_capitem;
2418
2419 /* Structure for building a list of named groups during the first pass of
2420 compiling. */
2421
2422 typedef struct named_group {
2423 const pcre_uchar *name; /* Points to the name in the pattern */
2424 int length; /* Length of the name */
2425 pcre_uint32 number; /* Group number */
2426 } named_group;
2427
2428 /* Structure for passing "static" information around between the functions
2429 doing the compiling, so that they are thread-safe. */
2430
2431 typedef struct compile_data {
2432 const pcre_uint8 *lcc; /* Points to lower casing table */
2433 const pcre_uint8 *fcc; /* Points to case-flipping table */
2434 const pcre_uint8 *cbits; /* Points to character type table */
2435 const pcre_uint8 *ctypes; /* Points to table of type maps */
2436 const pcre_uchar *start_workspace;/* The start of working space */
2437 const pcre_uchar *start_code; /* The start of the compiled code */
2438 const pcre_uchar *start_pattern; /* The start of the pattern */
2439 const pcre_uchar *end_pattern; /* The end of the pattern */
2440 pcre_uchar *hwm; /* High watermark of workspace */
2441 open_capitem *open_caps; /* Chain of open capture items */
2442 named_group *named_groups; /* Points to vector in pre-compile */
2443 pcre_uchar *name_table; /* The name/number table */
2444 int names_found; /* Number of entries so far */
2445 int name_entry_size; /* Size of each entry */
2446 int named_group_list_size; /* Number of entries in the list */
2447 int workspace_size; /* Size of workspace */
2448 unsigned int bracount; /* Count of capturing parens as we compile */
2449 int final_bracount; /* Saved value after first pass */
2450 int max_lookbehind; /* Maximum lookbehind (characters) */
2451 int top_backref; /* Maximum back reference */
2452 unsigned int backref_map; /* Bitmap of low back refs */
2453 unsigned int namedrefcount; /* Number of backreferences by name */
2454 int parens_depth; /* Depth of nested parentheses */
2455 int assert_depth; /* Depth of nested assertions */
2456 pcre_uint32 external_options; /* External (initial) options */
2457 pcre_uint32 external_flags; /* External flag bits to be set */
2458 int req_varyopt; /* "After variable item" flag for reqbyte */
2459 BOOL had_accept; /* (*ACCEPT) encountered */
2460 BOOL had_pruneorskip; /* (*PRUNE) or (*SKIP) encountered */
2461 BOOL check_lookbehind; /* Lookbehinds need later checking */
2462 BOOL dupnames; /* Duplicate names exist */
2463 BOOL dupgroups; /* Duplicate groups exist: (?| found */
2464 BOOL iscondassert; /* Next assert is a condition */
2465 int nltype; /* Newline type */
2466 int nllen; /* Newline string length */
2467 pcre_uchar nl[4]; /* Newline string when fixed length */
2468 } compile_data;
2469
2470 /* Structure for maintaining a chain of pointers to the currently incomplete
2471 branches, for testing for left recursion while compiling. */
2472
2473 typedef struct branch_chain {
2474 struct branch_chain *outer;
2475 pcre_uchar *current_branch;
2476 } branch_chain;
2477
2478 /* Structure for mutual recursion detection. */
2479
2480 typedef struct recurse_check {
2481 struct recurse_check *prev;
2482 const pcre_uchar *group;
2483 } recurse_check;
2484
2485 /* Structure for items in a linked list that represents an explicit recursive
2486 call within the pattern; used by pcre_exec(). */
2487
2488 typedef struct recursion_info {
2489 struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
2490 unsigned int group_num; /* Number of group that was called */
2491 int *offset_save; /* Pointer to start of saved offsets */
2492 int saved_max; /* Number of saved offsets */
2493 int saved_capture_last; /* Last capture number */
2494 PCRE_PUCHAR subject_position; /* Position at start of recursion */
2495 } recursion_info;
2496
2497 /* A similar structure for pcre_dfa_exec(). */
2498
2499 typedef struct dfa_recursion_info {
2500 struct dfa_recursion_info *prevrec;
2501 int group_num;
2502 PCRE_PUCHAR subject_position;
2503 } dfa_recursion_info;
2504
2505 /* Structure for building a chain of data for holding the values of the subject
2506 pointer at the start of each subpattern, so as to detect when an empty string
2507 has been matched by a subpattern - to break infinite loops; used by
2508 pcre_exec(). */
2509
2510 typedef struct eptrblock {
2511 struct eptrblock *epb_prev;
2512 PCRE_PUCHAR epb_saved_eptr;
2513 } eptrblock;
2514
2515
2516 /* Structure for passing "static" information around between the functions
2517 doing traditional NFA matching, so that they are thread-safe. */
2518
2519 typedef struct match_data {
2520 #if defined(ERLANG_INTEGRATION)
2521 unsigned long int loop_limit;
2522 void *state_save;
2523 #endif
2524 unsigned long int match_call_count; /* As it says */
2525 unsigned long int match_limit; /* As it says */
2526 unsigned long int match_limit_recursion; /* As it says */
2527 int *offset_vector; /* Offset vector */
2528 int offset_end; /* One past the end */
2529 int offset_max; /* The maximum usable for return data */
2530 int nltype; /* Newline type */
2531 int nllen; /* Newline string length */
2532 int name_count; /* Number of names in name table */
2533 int name_entry_size; /* Size of entry in names table */
2534 unsigned int skip_arg_count; /* For counting SKIP_ARGs */
2535 unsigned int ignore_skip_arg; /* For re-run when SKIP arg name not found */
2536 pcre_uchar *name_table; /* Table of names */
2537 pcre_uchar nl[4]; /* Newline string when fixed */
2538 const pcre_uint8 *lcc; /* Points to lower casing table */
2539 const pcre_uint8 *fcc; /* Points to case-flipping table */
2540 const pcre_uint8 *ctypes; /* Points to table of type maps */
2541 BOOL notbol; /* NOTBOL flag */
2542 BOOL noteol; /* NOTEOL flag */
2543 BOOL utf; /* UTF-8 / UTF-16 flag */
2544 BOOL jscript_compat; /* JAVASCRIPT_COMPAT flag */
2545 BOOL use_ucp; /* PCRE_UCP flag */
2546 BOOL endonly; /* Dollar not before final \n */
2547 BOOL notempty; /* Empty string match not wanted */
2548 BOOL notempty_atstart; /* Empty string match at start not wanted */
2549 BOOL hitend; /* Hit the end of the subject at some point */
2550 BOOL bsr_anycrlf; /* \R is just any CRLF, not full Unicode */
2551 BOOL hasthen; /* Pattern contains (*THEN) */
2552 const pcre_uchar *start_code; /* For use when recursing */
2553 PCRE_PUCHAR start_subject; /* Start of the subject string */
2554 PCRE_PUCHAR end_subject; /* End of the subject string */
2555 PCRE_PUCHAR start_match_ptr; /* Start of matched string */
2556 PCRE_PUCHAR end_match_ptr; /* Subject position at end match */
2557 PCRE_PUCHAR start_used_ptr; /* Earliest consulted character */
2558 int partial; /* PARTIAL options */
2559 int end_offset_top; /* Highwater mark at end of match */
2560 pcre_int32 capture_last; /* Most recent capture number + overflow flag */
2561 int start_offset; /* The start offset value */
2562 int match_function_type; /* Set for certain special calls of MATCH() */
2563 eptrblock *eptrchain; /* Chain of eptrblocks for tail recursions */
2564 int eptrn; /* Next free eptrblock */
2565 recursion_info *recursive; /* Linked list of recursion data */
2566 void *callout_data; /* To pass back to callouts */
2567 const pcre_uchar *mark; /* Mark pointer to pass back on success */
2568 const pcre_uchar *nomatch_mark;/* Mark pointer to pass back on failure */
2569 const pcre_uchar *once_target; /* Where to back up to for atomic groups */
2570 #ifdef NO_RECURSE
2571 void *match_frames_base; /* For remembering malloc'd frames */
2572 #endif
2573 } match_data;
2574
2575 /* A similar structure is used for the same purpose by the DFA matching
2576 functions. */
2577
2578 typedef struct dfa_match_data {
2579 const pcre_uchar *start_code; /* Start of the compiled pattern */
2580 const pcre_uchar *start_subject ; /* Start of the subject string */
2581 const pcre_uchar *end_subject; /* End of subject string */
2582 const pcre_uchar *start_used_ptr; /* Earliest consulted character */
2583 const pcre_uint8 *tables; /* Character tables */
2584 int start_offset; /* The start offset value */
2585 int moptions; /* Match options */
2586 int poptions; /* Pattern options */
2587 int nltype; /* Newline type */
2588 int nllen; /* Newline string length */
2589 pcre_uchar nl[4]; /* Newline string when fixed */
2590 void *callout_data; /* To pass back to callouts */
2591 dfa_recursion_info *recursive; /* Linked list of recursion data */
2592 } dfa_match_data;
2593
2594 /* Bit definitions for entries in the pcre_ctypes table. */
2595
2596 #define ctype_space 0x01
2597 #define ctype_letter 0x02
2598 #define ctype_digit 0x04
2599 #define ctype_xdigit 0x08
2600 #define ctype_word 0x10 /* alphanumeric or '_' */
2601 #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
2602
2603 /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
2604 of bits for a class map. Some classes are built by combining these tables. */
2605
2606 #define cbit_space 0 /* [:space:] or \s */
2607 #define cbit_xdigit 32 /* [:xdigit:] */
2608 #define cbit_digit 64 /* [:digit:] or \d */
2609 #define cbit_upper 96 /* [:upper:] */
2610 #define cbit_lower 128 /* [:lower:] */
2611 #define cbit_word 160 /* [:word:] or \w */
2612 #define cbit_graph 192 /* [:graph:] */
2613 #define cbit_print 224 /* [:print:] */
2614 #define cbit_punct 256 /* [:punct:] */
2615 #define cbit_cntrl 288 /* [:cntrl:] */
2616 #define cbit_length 320 /* Length of the cbits table */
2617
2618 /* Offsets of the various tables from the base tables pointer, and
2619 total length. */
2620
2621 #define lcc_offset 0
2622 #define fcc_offset 256
2623 #define cbits_offset 512
2624 #define ctypes_offset (cbits_offset + cbit_length)
2625 #define tables_length (ctypes_offset + 256)
2626
2627 /* Internal function and data prefixes. */
2628
2629 #if defined COMPILE_PCRE8
2630 #if defined(ERLANG_INTEGRATION)
2631 #ifndef PUBL
2632 #define PUBL(name) erts_pcre_##name
2633 #endif
2634 #ifndef PRIV
2635 #define PRIV(name) _erts_pcre_##name
2636 #endif
2637 #else
2638 #ifndef PUBL
2639 #define PUBL(name) pcre_##name
2640 #endif
2641 #ifndef PRIV
2642 #define PRIV(name) _pcre_##name
2643 #endif
2644 #endif
2645 #elif defined COMPILE_PCRE16
2646 #ifndef PUBL
2647 #define PUBL(name) pcre16_##name
2648 #endif
2649 #ifndef PRIV
2650 #define PRIV(name) _pcre16_##name
2651 #endif
2652 #elif defined COMPILE_PCRE32
2653 #ifndef PUBL
2654 #define PUBL(name) pcre32_##name
2655 #endif
2656 #ifndef PRIV
2657 #define PRIV(name) _pcre32_##name
2658 #endif
2659 #else
2660 #error Unsupported compiling mode
2661 #endif /* COMPILE_PCRE[8|16|32] */
2662
2663 /* Layout of the UCP type table that translates property names into types and
2664 codes. Each entry used to point directly to a name, but to reduce the number of
2665 relocations in shared libraries, it now has an offset into a single string
2666 instead. */
2667
2668 typedef struct {
2669 pcre_uint16 name_offset;
2670 pcre_uint16 type;
2671 pcre_uint16 value;
2672 } ucp_type_table;
2673
2674
2675 /* Internal shared data tables. These are tables that are used by more than one
2676 of the exported public functions. They have to be "external" in the C sense,
2677 but are not part of the PCRE public API. The data for these tables is in the
2678 pcre_tables.c module. */
2679
2680 #ifdef COMPILE_PCRE8
2681 extern const int PRIV(utf8_table1)[];
2682 extern const int PRIV(utf8_table1_size);
2683 extern const int PRIV(utf8_table2)[];
2684 extern const int PRIV(utf8_table3)[];
2685 extern const pcre_uint8 PRIV(utf8_table4)[];
2686 #endif /* COMPILE_PCRE8 */
2687
2688 extern const char PRIV(utt_names)[];
2689 extern const ucp_type_table PRIV(utt)[];
2690 extern const int PRIV(utt_size);
2691
2692 extern const pcre_uint8 PRIV(OP_lengths)[];
2693 extern const pcre_uint8 PRIV(default_tables)[];
2694
2695 extern const pcre_uint32 PRIV(hspace_list)[];
2696 extern const pcre_uint32 PRIV(vspace_list)[];
2697
2698
2699 /* Internal shared functions. These are functions that are used by more than
2700 one of the exported public functions. They have to be "external" in the C
2701 sense, but are not part of the PCRE public API. */
2702
2703 /* String comparison functions. */
2704 #if defined COMPILE_PCRE8
2705
2706 #define STRCMP_UC_UC(str1, str2) \
2707 strcmp((char *)(str1), (char *)(str2))
2708 #define STRCMP_UC_C8(str1, str2) \
2709 strcmp((char *)(str1), (str2))
2710 #define STRNCMP_UC_UC(str1, str2, num) \
2711 strncmp((char *)(str1), (char *)(str2), (num))
2712 #define STRNCMP_UC_C8(str1, str2, num) \
2713 strncmp((char *)(str1), (str2), (num))
2714 #define STRLEN_UC(str) strlen((const char *)str)
2715
2716 #elif defined COMPILE_PCRE16 || defined COMPILE_PCRE32
2717
2718 extern int PRIV(strcmp_uc_uc)(const pcre_uchar *,
2719 const pcre_uchar *);
2720 extern int PRIV(strcmp_uc_c8)(const pcre_uchar *,
2721 const char *);
2722 extern int PRIV(strncmp_uc_uc)(const pcre_uchar *,
2723 const pcre_uchar *, unsigned int num);
2724 extern int PRIV(strncmp_uc_c8)(const pcre_uchar *,
2725 const char *, unsigned int num);
2726 extern unsigned int PRIV(strlen_uc)(const pcre_uchar *str);
2727
2728 #define STRCMP_UC_UC(str1, str2) \
2729 PRIV(strcmp_uc_uc)((str1), (str2))
2730 #define STRCMP_UC_C8(str1, str2) \
2731 PRIV(strcmp_uc_c8)((str1), (str2))
2732 #define STRNCMP_UC_UC(str1, str2, num) \
2733 PRIV(strncmp_uc_uc)((str1), (str2), (num))
2734 #define STRNCMP_UC_C8(str1, str2, num) \
2735 PRIV(strncmp_uc_c8)((str1), (str2), (num))
2736 #define STRLEN_UC(str) PRIV(strlen_uc)(str)
2737
2738 #endif /* COMPILE_PCRE[8|16|32] */
2739
2740 #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
2741
2742 #define STRCMP_UC_UC_TEST(str1, str2) STRCMP_UC_UC(str1, str2)
2743 #define STRCMP_UC_C8_TEST(str1, str2) STRCMP_UC_C8(str1, str2)
2744
2745 #elif defined COMPILE_PCRE32
2746
2747 extern int PRIV(strcmp_uc_uc_utf)(const pcre_uchar *,
2748 const pcre_uchar *);
2749 extern int PRIV(strcmp_uc_c8_utf)(const pcre_uchar *,
2750 const char *);
2751
2752 #define STRCMP_UC_UC_TEST(str1, str2) \
2753 (utf ? PRIV(strcmp_uc_uc_utf)((str1), (str2)) : PRIV(strcmp_uc_uc)((str1), (str2)))
2754 #define STRCMP_UC_C8_TEST(str1, str2) \
2755 (utf ? PRIV(strcmp_uc_c8_utf)((str1), (str2)) : PRIV(strcmp_uc_c8)((str1), (str2)))
2756
2757 #endif /* COMPILE_PCRE[8|16|32] */
2758
2759 #if defined(ERLANG_INTEGRATION)
PRIV(valid_utf_ystate)2760 struct PRIV(valid_utf_ystate) {
2761 unsigned int cnt;
2762 int length;
2763 int yielded;
2764 PCRE_PUCHAR p;
2765 };
2766 extern int PRIV(yielding_valid_utf)(PCRE_PUCHAR, int, int *,
2767 struct PRIV(valid_utf_ystate) *);
2768 #endif
2769
2770 extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2771 extern BOOL PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2772 int *, BOOL);
2773 extern unsigned int PRIV(ord2utf)(pcre_uint32, pcre_uchar *);
2774 extern int PRIV(valid_utf)(PCRE_PUCHAR, int, int *);
2775 extern BOOL PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2776 int *, BOOL);
2777 extern BOOL PRIV(xclass)(pcre_uint32, const pcre_uchar *, BOOL);
2778
2779 #ifdef SUPPORT_JIT
2780 extern void PRIV(jit_compile)(const REAL_PCRE *,
2781 PUBL(extra) *, int);
2782 extern int PRIV(jit_exec)(const PUBL(extra) *,
2783 const pcre_uchar *, int, int, int, int *, int);
2784 extern void PRIV(jit_free)(void *);
2785 extern int PRIV(jit_get_size)(void *);
2786 extern const char* PRIV(jit_get_target)(void);
2787 #endif
2788
2789 /* Unicode character database (UCD) */
2790
2791 typedef struct {
2792 pcre_uint8 script; /* ucp_Arabic, etc. */
2793 pcre_uint8 chartype; /* ucp_Cc, etc. (general categories) */
2794 pcre_uint8 gbprop; /* ucp_gbControl, etc. (grapheme break property) */
2795 pcre_uint8 caseset; /* offset to multichar other cases or zero */
2796 pcre_int32 other_case; /* offset to other case, or zero if none */
2797 } ucd_record;
2798
2799 extern const pcre_uint32 PRIV(ucd_caseless_sets)[];
2800 extern const ucd_record PRIV(ucd_records)[];
2801 extern const pcre_uint8 PRIV(ucd_stage1)[];
2802 extern const pcre_uint16 PRIV(ucd_stage2)[];
2803 extern const pcre_uint32 PRIV(ucp_gentype)[];
2804 extern const pcre_uint32 PRIV(ucp_gbtable)[];
2805 #ifdef COMPILE_PCRE32
2806 extern const ucd_record PRIV(dummy_ucd_record)[];
2807 #endif
2808 #ifdef SUPPORT_JIT
2809 extern const int PRIV(ucp_typerange)[];
2810 #endif
2811
2812 #ifdef SUPPORT_UCP
2813 /* UCD access macros */
2814
2815 #define UCD_BLOCK_SIZE 128
2816 #define REAL_GET_UCD(ch) (PRIV(ucd_records) + \
2817 PRIV(ucd_stage2)[PRIV(ucd_stage1)[(int)(ch) / UCD_BLOCK_SIZE] * \
2818 UCD_BLOCK_SIZE + (int)(ch) % UCD_BLOCK_SIZE])
2819
2820 #ifdef COMPILE_PCRE32
2821 #define GET_UCD(ch) ((ch > 0x10ffff)? PRIV(dummy_ucd_record) : REAL_GET_UCD(ch))
2822 #else
2823 #define GET_UCD(ch) REAL_GET_UCD(ch)
2824 #endif
2825
2826 #define UCD_CHARTYPE(ch) GET_UCD(ch)->chartype
2827 #define UCD_SCRIPT(ch) GET_UCD(ch)->script
2828 #define UCD_CATEGORY(ch) PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2829 #define UCD_GRAPHBREAK(ch) GET_UCD(ch)->gbprop
2830 #define UCD_CASESET(ch) GET_UCD(ch)->caseset
2831 #define UCD_OTHERCASE(ch) ((pcre_uint32)((int)ch + (int)(GET_UCD(ch)->other_case)))
2832
2833 #endif /* SUPPORT_UCP */
2834
2835 #endif
2836
2837 /* End of pcre_internal.h */
2838