1 /*************************************************
2 * Perl-Compatible Regular Expressions *
3 *************************************************/
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.
9 Written by Philip Hazel
10 Copyright (c) 1997-2014 University of Cambridge
12 -----------------------------------------------------------------------------
13 Redistribution and use in source and binary forms, with or without
14 modification, are permitted provided that the following conditions are met:
16 * Redistributions of source code must retain the above copyright notice,
17 this list of conditions and the following disclaimer.
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.
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.
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 -----------------------------------------------------------------------------
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. */
46 #ifndef PCRE_INTERNAL_H
47 #define PCRE_INTERNAL_H
49 /* Define PCRE_DEBUG to get debugging output on stdout. */
55 /* PCRE is compiled as an 8 bit library if it is not requested otherwise. */
57 #if !defined COMPILE_PCRE16 && !defined COMPILE_PCRE32
61 /* If SUPPORT_UCP is defined, SUPPORT_UTF must also be defined. The
62 "configure" script ensures this, but not everybody uses "configure". */
64 #if defined SUPPORT_UCP && !(defined SUPPORT_UTF)
68 /* We define SUPPORT_UTF if SUPPORT_UTF8 is enabled for compatibility
69 reasons with existing code. */
71 #if defined SUPPORT_UTF8 && !(defined SUPPORT_UTF)
75 /* Fixme: SUPPORT_UTF8 should be eventually disappear from the code.
76 Until then we define it if SUPPORT_UTF is defined. */
78 #if defined SUPPORT_UTF && !(defined SUPPORT_UTF8)
79 #define SUPPORT_UTF8 1
82 /* We do not support both EBCDIC and UTF-8/16/32 at the same time. The "configure"
83 script prevents both being selected, but not everybody uses "configure". */
85 #if defined EBCDIC && defined SUPPORT_UTF
86 #error The use of both EBCDIC and SUPPORT_UTF is not supported.
89 /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef
90 inline, and there are *still* stupid compilers about that don't like indented
91 pre-processor statements, or at least there were when I first wrote this. After
92 all, it had only been about 10 years then...
94 It turns out that the Mac Debugging.h header also defines the macro DPRINTF, so
95 be absolutely sure we get our version. */
99 #define DPRINTF(p) printf p
101 #define DPRINTF(p) /* Nothing */
105 /* Standard C headers plus the external interface definition. The only time
106 setjmp and stdarg are used is when NO_RECURSE is set. */
115 /* Valgrind (memcheck) support */
117 #ifdef SUPPORT_VALGRIND
118 #include <valgrind/memcheck.h>
121 /* When compiling a DLL for Windows, the exported symbols have to be declared
122 using some MS magic. I found some useful information on this web page:
123 http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
124 information there, using __declspec(dllexport) without "extern" we have a
125 definition; with "extern" we have a declaration. The settings here override the
126 setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
127 which is all that is needed for applications (they just import the symbols). We
130 PCRE_EXP_DECL for declarations
131 PCRE_EXP_DEFN for definitions of exported functions
132 PCRE_EXP_DATA_DEFN for definitions of exported variables
134 The reason for the two DEFN macros is that in non-Windows environments, one
135 does not want to have "extern" before variable definitions because it leads to
136 compiler warnings. So we distinguish between functions and variables. In
137 Windows, the two should always be the same.
139 The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
140 which is an application, but needs to import this file in order to "peek" at
141 internals, can #include pcre.h first to get an application's-eye view.
143 In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
144 special-purpose environments) might want to stick other stuff in front of
145 exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
146 PCRE_EXP_DATA_DEFN only if they are not already set. */
148 #ifndef PCRE_EXP_DECL
151 # define PCRE_EXP_DECL extern __declspec(dllexport)
152 # define PCRE_EXP_DEFN __declspec(dllexport)
153 # define PCRE_EXP_DATA_DEFN __declspec(dllexport)
155 # define PCRE_EXP_DECL extern
156 # define PCRE_EXP_DEFN
157 # define PCRE_EXP_DATA_DEFN
161 # define PCRE_EXP_DECL extern "C"
163 # define PCRE_EXP_DECL extern
165 # ifndef PCRE_EXP_DEFN
166 # define PCRE_EXP_DEFN PCRE_EXP_DECL
168 # ifndef PCRE_EXP_DATA_DEFN
169 # define PCRE_EXP_DATA_DEFN
174 /* When compiling with the MSVC compiler, it is sometimes necessary to include
175 a "calling convention" before exported function names. (This is secondhand
176 information; I know nothing about MSVC myself). For example, something like
178 void __cdecl function(....)
180 might be needed. In order so make this easy, all the exported functions have
181 PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
182 set, we ensure here that it has no effect. */
184 #ifndef PCRE_CALL_CONVENTION
185 #define PCRE_CALL_CONVENTION
188 /* We need to have types that specify unsigned 8, 16 and 32-bit integers. We
189 cannot determine these outside the compilation (e.g. by running a program as
190 part of "configure") because PCRE is often cross-compiled for use on other
191 systems. Instead we make use of the maximum sizes that are available at
192 preprocessor time in standard C environments. */
194 typedef unsigned char pcre_uint8;
196 #if USHRT_MAX == 65535
197 typedef unsigned short pcre_uint16;
198 typedef short pcre_int16;
199 #define PCRE_UINT16_MAX USHRT_MAX
200 #define PCRE_INT16_MAX SHRT_MAX
201 #elif UINT_MAX == 65535
202 typedef unsigned int pcre_uint16;
203 typedef int pcre_int16;
204 #define PCRE_UINT16_MAX UINT_MAX
205 #define PCRE_INT16_MAX INT_MAX
207 #error Cannot determine a type for 16-bit integers
210 #if UINT_MAX == 4294967295U
211 typedef unsigned int pcre_uint32;
212 typedef int pcre_int32;
213 #define PCRE_UINT32_MAX UINT_MAX
214 #define PCRE_INT32_MAX INT_MAX
215 #elif ULONG_MAX == 4294967295UL
216 typedef unsigned long int pcre_uint32;
217 typedef long int pcre_int32;
218 #define PCRE_UINT32_MAX ULONG_MAX
219 #define PCRE_INT32_MAX LONG_MAX
221 #error Cannot determine a type for 32-bit integers
224 /* When checking for integer overflow in pcre_compile(), we need to handle
225 large integers. If a 64-bit integer type is available, we can use that.
226 Otherwise we have to cast to double, which of course requires floating point
227 arithmetic. Handle this by defining a macro for the appropriate type. If
228 stdint.h is available, include it; it may define INT64_MAX. Systems that do not
229 have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
232 #if defined HAVE_STDINT_H
234 #elif defined HAVE_INTTYPES_H
235 #include <inttypes.h>
238 #if defined INT64_MAX || defined int64_t
239 #define INT64_OR_DOUBLE int64_t
241 #define INT64_OR_DOUBLE double
244 /* All character handling must be done as unsigned characters. Otherwise there
245 are problems with top-bit-set characters and functions such as isspace().
246 However, we leave the interface to the outside world as char * or short *,
247 because that should make things easier for callers. This character type is
250 The IN_UCHARS macro multiply its argument with the byte size of the current
251 pcre_uchar type. Useful for memcpy and such operations, whose require the
252 byte size of their input/output buffers.
254 The MAX_255 macro checks whether its pcre_uchar input is less than 256.
256 The TABLE_GET macro is designed for accessing elements of tables whose contain
257 exactly 256 items. When the character is able to contain more than 256
258 items, some check is needed before accessing these tables.
261 #if defined COMPILE_PCRE8
263 typedef unsigned char pcre_uchar;
264 #define IN_UCHARS(x) (x)
266 #define TABLE_GET(c, table, default) ((table)[c])
268 #elif defined COMPILE_PCRE16
270 #if USHRT_MAX != 65535
271 /* This is a warning message. Change PCRE_UCHAR16 to a 16 bit data type in
272 pcre.h(.in) and disable (comment out) this message. */
273 #error Warning: PCRE_UCHAR16 is not a 16 bit data type.
276 typedef pcre_uint16 pcre_uchar;
277 #define UCHAR_SHIFT (1)
278 #define IN_UCHARS(x) ((x) << UCHAR_SHIFT)
279 #define MAX_255(c) ((c) <= 255u)
280 #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
282 #elif defined COMPILE_PCRE32
284 typedef pcre_uint32 pcre_uchar;
285 #define UCHAR_SHIFT (2)
286 #define IN_UCHARS(x) ((x) << UCHAR_SHIFT)
287 #define MAX_255(c) ((c) <= 255u)
288 #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
291 #error Unsupported compiling mode
292 #endif /* COMPILE_PCRE[8|16|32] */
294 /* This is an unsigned int value that no character can ever have. UTF-8
295 characters only go up to 0x7fffffff (though Unicode doesn't go beyond
298 #define NOTACHAR 0xffffffff
300 /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
301 "any" and "anycrlf" at present). The following macros are used to package up
302 testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
303 modules to indicate in which datablock the parameters exist, and what the
304 start/end of string field names are. */
306 #define NLTYPE_FIXED 0 /* Newline is a fixed length string */
307 #define NLTYPE_ANY 1 /* Newline is any Unicode line ending */
308 #define NLTYPE_ANYCRLF 2 /* Newline is CR, LF, or CRLF */
310 /* This macro checks for a newline at the given position */
312 #define IS_NEWLINE(p) \
313 ((NLBLOCK->nltype != NLTYPE_FIXED)? \
314 ((p) < NLBLOCK->PSEND && \
315 PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, \
316 &(NLBLOCK->nllen), utf)) \
318 ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
319 UCHAR21TEST(p) == NLBLOCK->nl[0] && \
320 (NLBLOCK->nllen == 1 || UCHAR21TEST(p+1) == NLBLOCK->nl[1]) \
324 /* This macro checks for a newline immediately preceding the given position */
326 #define WAS_NEWLINE(p) \
327 ((NLBLOCK->nltype != NLTYPE_FIXED)? \
328 ((p) > NLBLOCK->PSSTART && \
329 PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
330 &(NLBLOCK->nllen), utf)) \
332 ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
333 UCHAR21TEST(p - NLBLOCK->nllen) == NLBLOCK->nl[0] && \
334 (NLBLOCK->nllen == 1 || UCHAR21TEST(p - NLBLOCK->nllen + 1) == NLBLOCK->nl[1]) \
338 /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
339 with a custom type. This makes it possible, for example, to allow pcre_exec()
340 to process subject strings that are discontinuous by using a smart pointer
341 class. It must always be possible to inspect all of the subject string in
342 pcre_exec() because of the way it backtracks. Two macros are required in the
343 normal case, for sign-unspecified and unsigned char pointers. The former is
344 used for the external interface and appears in pcre.h, which is why its name
345 must begin with PCRE_. */
347 #ifdef CUSTOM_SUBJECT_PTR
348 #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
350 #define PCRE_PUCHAR const pcre_uchar *
353 /* Include the public PCRE header and the definitions of UCP character property
359 #ifdef COMPILE_PCRE32
360 /* Assert that the public PCRE_UCHAR32 is a 32-bit type */
361 typedef int __assert_pcre_uchar32_size[sizeof(PCRE_UCHAR32) == 4 ? 1 : -1];
364 /* When compiling for use with the Virtual Pascal compiler, these functions
365 need to have their names changed. PCRE must be compiled with the -DVPCOMPAT
366 option on the command line. */
369 #define strlen(s) _strlen(s)
370 #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
371 #define memcmp(s,c,n) _memcmp(s,c,n)
372 #define memcpy(d,s,n) _memcpy(d,s,n)
373 #define memmove(d,s,n) _memmove(d,s,n)
374 #define memset(s,c,n) _memset(s,c,n)
377 /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
378 define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
379 is set. Otherwise, include an emulating function for those systems that have
380 neither (there some non-Unix environments where this is the case). */
383 #undef memmove /* some systems may have a macro */
385 #define memmove(a, b, c) bcopy(b, a, c)
386 #else /* HAVE_BCOPY */
388 pcre_memmove(void *d, const void *s, size_t n)
391 unsigned char *dest = (unsigned char *)d;
392 const unsigned char *src = (const unsigned char *)s;
397 for (i = 0; i < n; ++i) *(--dest) = *(--src);
402 for (i = 0; i < n; ++i) *dest++ = *src++;
403 return (void *)(dest - n);
406 #define memmove(a, b, c) pcre_memmove(a, b, c)
407 #endif /* not HAVE_BCOPY */
408 #endif /* not HAVE_MEMMOVE */
409 #endif /* not VPCOMPAT */
412 /* PCRE keeps offsets in its compiled code as 2-byte quantities (always stored
413 in big-endian order) by default. These are used, for example, to link from the
414 start of a subpattern to its alternatives and its end. The use of 2 bytes per
415 offset limits the size of the compiled regex to around 64K, which is big enough
416 for almost everybody. However, I received a request for an even bigger limit.
417 For this reason, and also to make the code easier to maintain, the storing and
418 loading of offsets from the byte string is now handled by the macros that are
421 The macros are controlled by the value of LINK_SIZE. This defaults to 2 in
422 the config.h file, but can be overridden by using -D on the command line. This
423 is automated on Unix systems via the "configure" command. */
425 #if defined COMPILE_PCRE8
431 (a[(n)+1] = (d) & 255)
434 (((a)[n] << 8) | (a)[(n)+1])
436 #define MAX_PATTERN_SIZE (1 << 16)
442 (a[n] = (d) >> 16), \
443 (a[(n)+1] = (d) >> 8), \
444 (a[(n)+2] = (d) & 255)
447 (((a)[n] << 16) | ((a)[(n)+1] << 8) | (a)[(n)+2])
449 #define MAX_PATTERN_SIZE (1 << 24)
455 (a[n] = (d) >> 24), \
456 (a[(n)+1] = (d) >> 16), \
457 (a[(n)+2] = (d) >> 8), \
458 (a[(n)+3] = (d) & 255)
461 (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
463 /* Keep it positive */
464 #define MAX_PATTERN_SIZE (1 << 30)
467 #error LINK_SIZE must be either 2, 3, or 4
470 #elif defined COMPILE_PCRE16
474 /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
484 #define MAX_PATTERN_SIZE (1 << 16)
486 #elif LINK_SIZE == 3 || LINK_SIZE == 4
488 /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
493 (a[n] = (d) >> 16), \
494 (a[(n)+1] = (d) & 65535)
497 (((a)[n] << 16) | (a)[(n)+1])
499 /* Keep it positive */
500 #define MAX_PATTERN_SIZE (1 << 30)
503 #error LINK_SIZE must be either 2, 3, or 4
506 #elif defined COMPILE_PCRE32
508 /* Only supported LINK_SIZE is 4 */
509 /* Redefine LINK_SIZE as a multiple of sizeof(pcre_uchar) */
519 /* Keep it positive */
520 #define MAX_PATTERN_SIZE (1 << 30)
523 #error Unsupported compiling mode
524 #endif /* COMPILE_PCRE[8|16|32] */
526 /* Convenience macro defined in terms of the others */
528 #define PUTINC(a,n,d) PUT(a,n,d), a += LINK_SIZE
531 /* PCRE uses some other 2-byte quantities that do not change when the size of
532 offsets changes. There are used for repeat counts and for other things such as
533 capturing parenthesis numbers in back references. */
535 #if defined COMPILE_PCRE8
539 #define PUT2(a,n,d) \
543 /* For reasons that I do not understand, the expression in this GET2 macro is
544 treated by gcc as a signed expression, even when a is declared as unsigned. It
545 seems that any kind of arithmetic results in a signed value. */
548 (unsigned int)(((a)[n] << 8) | (a)[(n)+1])
550 #elif defined COMPILE_PCRE16
554 #define PUT2(a,n,d) \
560 #elif defined COMPILE_PCRE32
564 #define PUT2(a,n,d) \
571 #error Unsupported compiling mode
572 #endif /* COMPILE_PCRE[8|16|32] */
574 #define PUT2INC(a,n,d) PUT2(a,n,d), a += IMM2_SIZE
576 /* The maximum length of a MARK name is currently one data unit; it may be
577 changed in future to be a fixed number of bytes or to depend on LINK_SIZE. */
579 #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
580 #define MAX_MARK ((1u << 16) - 1)
582 #define MAX_MARK ((1u << 8) - 1)
585 /* There is a proposed future special "UTF-21" mode, in which only the lowest
586 21 bits of a 32-bit character are interpreted as UTF, with the remaining 11
587 high-order bits available to the application for other uses. In preparation for
588 the future implementation of this mode, there are macros that load a data item
589 and, if in this special mode, mask it to 21 bits. These macros all have names
590 starting with UCHAR21. In all other modes, including the normal 32-bit
591 library, the macros all have the same simple definitions. When the new mode is
592 implemented, it is expected that these definitions will be varied appropriately
593 using #ifdef when compiling the library that supports the special mode. */
595 #define UCHAR21(eptr) (*(eptr))
596 #define UCHAR21TEST(eptr) (*(eptr))
597 #define UCHAR21INC(eptr) (*(eptr)++)
598 #define UCHAR21INCTEST(eptr) (*(eptr)++)
600 /* When UTF encoding is being used, a character is no longer just a single
601 byte in 8-bit mode or a single short in 16-bit mode. The macros for character
602 handling generate simple sequences when used in the basic mode, and more
603 complicated ones for UTF characters. GETCHARLENTEST and other macros are not
604 used when UTF is not supported. To make sure they can never even appear when
605 UTF support is omitted, we don't even define them. */
609 /* #define MAX_VALUE_FOR_SINGLE_CHAR */
610 /* #define HAS_EXTRALEN(c) */
611 /* #define GET_EXTRALEN(c) */
612 /* #define NOT_FIRSTCHAR(c) */
613 #define GETCHAR(c, eptr) c = *eptr;
614 #define GETCHARTEST(c, eptr) c = *eptr;
615 #define GETCHARINC(c, eptr) c = *eptr++;
616 #define GETCHARINCTEST(c, eptr) c = *eptr++;
617 #define GETCHARLEN(c, eptr, len) c = *eptr;
618 /* #define GETCHARLENTEST(c, eptr, len) */
619 /* #define BACKCHAR(eptr) */
620 /* #define FORWARDCHAR(eptr) */
621 /* #define ACROSSCHAR(condition, eptr, action) */
623 #else /* SUPPORT_UTF */
625 /* Tests whether the code point needs extra characters to decode. */
627 #define HASUTF8EXTRALEN(c) ((c) >= 0xc0)
629 /* Base macro to pick up the remaining bytes of a UTF-8 character, not
630 advancing the pointer. */
632 #define GETUTF8(c, eptr) \
634 if ((c & 0x20) == 0) \
635 c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
636 else if ((c & 0x10) == 0) \
637 c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
638 else if ((c & 0x08) == 0) \
639 c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
640 ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
641 else if ((c & 0x04) == 0) \
642 c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
643 ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
646 c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
647 ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
648 ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
651 /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
654 #define GETUTF8INC(c, eptr) \
656 if ((c & 0x20) == 0) \
657 c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
658 else if ((c & 0x10) == 0) \
660 c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
663 else if ((c & 0x08) == 0) \
665 c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
666 ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
669 else if ((c & 0x04) == 0) \
671 c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
672 ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
678 c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
679 ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
680 ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
685 #if defined COMPILE_PCRE8
687 /* These macros were originally written in the form of loops that used data
688 from the tables whose names start with PRIV(utf8_table). They were rewritten by
689 a user so as not to use loops, because in some environments this gives a
690 significant performance advantage, and it seems never to do any harm. */
692 /* Tells the biggest code point which can be encoded as a single character. */
694 #define MAX_VALUE_FOR_SINGLE_CHAR 127
696 /* Tests whether the code point needs extra characters to decode. */
698 #define HAS_EXTRALEN(c) ((c) >= 0xc0)
700 /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
701 Otherwise it has an undefined behaviour. */
703 #define GET_EXTRALEN(c) (PRIV(utf8_table4)[(c) & 0x3f])
705 /* Returns TRUE, if the given character is not the first character
706 of a UTF sequence. */
708 #define NOT_FIRSTCHAR(c) (((c) & 0xc0) == 0x80)
710 /* Get the next UTF-8 character, not advancing the pointer. This is called when
711 we know we are in UTF-8 mode. */
713 #define GETCHAR(c, eptr) \
715 if (c >= 0xc0) GETUTF8(c, eptr);
717 /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
720 #define GETCHARTEST(c, eptr) \
722 if (utf && c >= 0xc0) GETUTF8(c, eptr);
724 /* Get the next UTF-8 character, advancing the pointer. This is called when we
725 know we are in UTF-8 mode. */
727 #define GETCHARINC(c, eptr) \
729 if (c >= 0xc0) GETUTF8INC(c, eptr);
731 /* Get the next character, testing for UTF-8 mode, and advancing the pointer.
732 This is called when we don't know if we are in UTF-8 mode. */
734 #define GETCHARINCTEST(c, eptr) \
736 if (utf && c >= 0xc0) GETUTF8INC(c, eptr);
738 /* Base macro to pick up the remaining bytes of a UTF-8 character, not
739 advancing the pointer, incrementing the length. */
741 #define GETUTF8LEN(c, eptr, len) \
743 if ((c & 0x20) == 0) \
745 c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
748 else if ((c & 0x10) == 0) \
750 c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
753 else if ((c & 0x08) == 0) \
755 c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
756 ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
759 else if ((c & 0x04) == 0) \
761 c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
762 ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
768 c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
769 ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
770 ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
775 /* Get the next UTF-8 character, not advancing the pointer, incrementing length
776 if there are extra bytes. This is called when we know we are in UTF-8 mode. */
778 #define GETCHARLEN(c, eptr, len) \
780 if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
782 /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
783 pointer, incrementing length if there are extra bytes. This is called when we
784 do not know if we are in UTF-8 mode. */
786 #define GETCHARLENTEST(c, eptr, len) \
788 if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len);
790 /* If the pointer is not at the start of a character, move it back until
791 it is. This is called only in UTF-8 mode - we don't put a test within the macro
792 because almost all calls are already within a block of UTF-8 only code. */
794 #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
796 /* Same as above, just in the other direction. */
797 #define FORWARDCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr++
799 /* Same as above, but it allows a fully customizable form. */
800 #define ACROSSCHAR(condition, eptr, action) \
801 while((condition) && ((eptr) & 0xc0) == 0x80) action
803 #elif defined COMPILE_PCRE16
805 /* Tells the biggest code point which can be encoded as a single character. */
807 #define MAX_VALUE_FOR_SINGLE_CHAR 65535
809 /* Tests whether the code point needs extra characters to decode. */
811 #define HAS_EXTRALEN(c) (((c) & 0xfc00) == 0xd800)
813 /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
814 Otherwise it has an undefined behaviour. */
816 #define GET_EXTRALEN(c) 1
818 /* Returns TRUE, if the given character is not the first character
819 of a UTF sequence. */
821 #define NOT_FIRSTCHAR(c) (((c) & 0xfc00) == 0xdc00)
823 /* Base macro to pick up the low surrogate of a UTF-16 character, not
824 advancing the pointer. */
826 #define GETUTF16(c, eptr) \
827 { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; }
829 /* Get the next UTF-16 character, not advancing the pointer. This is called when
830 we know we are in UTF-16 mode. */
832 #define GETCHAR(c, eptr) \
834 if ((c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
836 /* Get the next UTF-16 character, testing for UTF-16 mode, and not advancing the
839 #define GETCHARTEST(c, eptr) \
841 if (utf && (c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
843 /* Base macro to pick up the low surrogate of a UTF-16 character, advancing
846 #define GETUTF16INC(c, eptr) \
847 { c = (((c & 0x3ff) << 10) | (*eptr++ & 0x3ff)) + 0x10000; }
849 /* Get the next UTF-16 character, advancing the pointer. This is called when we
850 know we are in UTF-16 mode. */
852 #define GETCHARINC(c, eptr) \
854 if ((c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
856 /* Get the next character, testing for UTF-16 mode, and advancing the pointer.
857 This is called when we don't know if we are in UTF-16 mode. */
859 #define GETCHARINCTEST(c, eptr) \
861 if (utf && (c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
863 /* Base macro to pick up the low surrogate of a UTF-16 character, not
864 advancing the pointer, incrementing the length. */
866 #define GETUTF16LEN(c, eptr, len) \
867 { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; len++; }
869 /* Get the next UTF-16 character, not advancing the pointer, incrementing
870 length if there is a low surrogate. This is called when we know we are in
873 #define GETCHARLEN(c, eptr, len) \
875 if ((c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
877 /* Get the next UTF-816character, testing for UTF-16 mode, not advancing the
878 pointer, incrementing length if there is a low surrogate. This is called when
879 we do not know if we are in UTF-16 mode. */
881 #define GETCHARLENTEST(c, eptr, len) \
883 if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
885 /* If the pointer is not at the start of a character, move it back until
886 it is. This is called only in UTF-16 mode - we don't put a test within the
887 macro because almost all calls are already within a block of UTF-16 only
890 #define BACKCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr--
892 /* Same as above, just in the other direction. */
893 #define FORWARDCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr++
895 /* Same as above, but it allows a fully customizable form. */
896 #define ACROSSCHAR(condition, eptr, action) \
897 if ((condition) && ((eptr) & 0xfc00) == 0xdc00) action
899 #elif defined COMPILE_PCRE32
901 /* These are trivial for the 32-bit library, since all UTF-32 characters fit
902 into one pcre_uchar unit. */
903 #define MAX_VALUE_FOR_SINGLE_CHAR (0x10ffffu)
904 #define HAS_EXTRALEN(c) (0)
905 #define GET_EXTRALEN(c) (0)
906 #define NOT_FIRSTCHAR(c) (0)
908 /* Get the next UTF-32 character, not advancing the pointer. This is called when
909 we know we are in UTF-32 mode. */
911 #define GETCHAR(c, eptr) \
914 /* Get the next UTF-32 character, testing for UTF-32 mode, and not advancing the
917 #define GETCHARTEST(c, eptr) \
920 /* Get the next UTF-32 character, advancing the pointer. This is called when we
921 know we are in UTF-32 mode. */
923 #define GETCHARINC(c, eptr) \
926 /* Get the next character, testing for UTF-32 mode, and advancing the pointer.
927 This is called when we don't know if we are in UTF-32 mode. */
929 #define GETCHARINCTEST(c, eptr) \
932 /* Get the next UTF-32 character, not advancing the pointer, not incrementing
933 length (since all UTF-32 is of length 1). This is called when we know we are in
936 #define GETCHARLEN(c, eptr, len) \
939 /* Get the next UTF-32character, testing for UTF-32 mode, not advancing the
940 pointer, not incrementing the length (since all UTF-32 is of length 1).
941 This is called when we do not know if we are in UTF-32 mode. */
943 #define GETCHARLENTEST(c, eptr, len) \
946 /* If the pointer is not at the start of a character, move it back until
947 it is. This is called only in UTF-32 mode - we don't put a test within the
948 macro because almost all calls are already within a block of UTF-32 only
950 These are all no-ops since all UTF-32 characters fit into one pcre_uchar. */
952 #define BACKCHAR(eptr) do { } while (0)
954 /* Same as above, just in the other direction. */
955 #define FORWARDCHAR(eptr) do { } while (0)
957 /* Same as above, but it allows a fully customizable form. */
958 #define ACROSSCHAR(condition, eptr, action) do { } while (0)
961 #error Unsupported compiling mode
962 #endif /* COMPILE_PCRE[8|16|32] */
964 #endif /* SUPPORT_UTF */
966 /* Tests for Unicode horizontal and vertical whitespace characters must check a
967 number of different values. Using a switch statement for this generates the
968 fastest code (no loop, no memory access), and there are several places in the
969 interpreter code where this happens. In order to ensure that all the case lists
970 remain in step, we use macros so that there is only one place where the lists
973 These values are also required as lists in pcre_compile.c when processing \h,
974 \H, \v and \V in a character class. The lists are defined in pcre_tables.c, but
975 macros that define the values are here so that all the definitions are
976 together. The lists must be in ascending character order, terminated by
977 NOTACHAR (which is 0xffffffff).
979 Any changes should ensure that the various macros are kept in step with each
980 other. NOTE: The values also appear in pcre_jit_compile.c. */
982 /* ------ ASCII/Unicode environments ------ */
986 #define HSPACE_LIST \
987 CHAR_HT, CHAR_SPACE, 0xa0, \
988 0x1680, 0x180e, 0x2000, 0x2001, 0x2002, 0x2003, 0x2004, 0x2005, \
989 0x2006, 0x2007, 0x2008, 0x2009, 0x200A, 0x202f, 0x205f, 0x3000, \
992 #define HSPACE_MULTIBYTE_CASES \
993 case 0x1680: /* OGHAM SPACE MARK */ \
994 case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ \
995 case 0x2000: /* EN QUAD */ \
996 case 0x2001: /* EM QUAD */ \
997 case 0x2002: /* EN SPACE */ \
998 case 0x2003: /* EM SPACE */ \
999 case 0x2004: /* THREE-PER-EM SPACE */ \
1000 case 0x2005: /* FOUR-PER-EM SPACE */ \
1001 case 0x2006: /* SIX-PER-EM SPACE */ \
1002 case 0x2007: /* FIGURE SPACE */ \
1003 case 0x2008: /* PUNCTUATION SPACE */ \
1004 case 0x2009: /* THIN SPACE */ \
1005 case 0x200A: /* HAIR SPACE */ \
1006 case 0x202f: /* NARROW NO-BREAK SPACE */ \
1007 case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ \
1008 case 0x3000 /* IDEOGRAPHIC SPACE */
1010 #define HSPACE_BYTE_CASES \
1013 case 0xa0 /* NBSP */
1015 #define HSPACE_CASES \
1016 HSPACE_BYTE_CASES: \
1017 HSPACE_MULTIBYTE_CASES
1019 #define VSPACE_LIST \
1020 CHAR_LF, CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, 0x2028, 0x2029, NOTACHAR
1022 #define VSPACE_MULTIBYTE_CASES \
1023 case 0x2028: /* LINE SEPARATOR */ \
1024 case 0x2029 /* PARAGRAPH SEPARATOR */
1026 #define VSPACE_BYTE_CASES \
1033 #define VSPACE_CASES \
1034 VSPACE_BYTE_CASES: \
1035 VSPACE_MULTIBYTE_CASES
1037 /* ------ EBCDIC environments ------ */
1040 #define HSPACE_LIST CHAR_HT, CHAR_SPACE
1042 #define HSPACE_BYTE_CASES \
1046 #define HSPACE_CASES HSPACE_BYTE_CASES
1049 #define VSPACE_LIST \
1050 CHAR_VT, CHAR_FF, CHAR_CR, CHAR_NEL, CHAR_LF, NOTACHAR
1052 #define VSPACE_LIST \
1053 CHAR_VT, CHAR_FF, CHAR_CR, CHAR_LF, CHAR_NEL, NOTACHAR
1056 #define VSPACE_BYTE_CASES \
1063 #define VSPACE_CASES VSPACE_BYTE_CASES
1066 /* ------ End of whitespace macros ------ */
1070 /* Private flags containing information about the compiled regex. They used to
1071 live at the top end of the options word, but that got almost full, so they were
1072 moved to a 16-bit flags word - which got almost full, so now they are in a
1073 32-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as the
1074 restrictions on partial matching have been lifted. It remains for backwards
1077 #define PCRE_MODE8 0x00000001 /* compiled in 8 bit mode */
1078 #define PCRE_MODE16 0x00000002 /* compiled in 16 bit mode */
1079 #define PCRE_MODE32 0x00000004 /* compiled in 32 bit mode */
1080 #define PCRE_FIRSTSET 0x00000010 /* first_char is set */
1081 #define PCRE_FCH_CASELESS 0x00000020 /* caseless first char */
1082 #define PCRE_REQCHSET 0x00000040 /* req_byte is set */
1083 #define PCRE_RCH_CASELESS 0x00000080 /* caseless requested char */
1084 #define PCRE_STARTLINE 0x00000100 /* start after \n for multiline */
1085 #define PCRE_NOPARTIAL 0x00000200 /* can't use partial with this regex */
1086 #define PCRE_JCHANGED 0x00000400 /* j option used in regex */
1087 #define PCRE_HASCRORLF 0x00000800 /* explicit \r or \n in pattern */
1088 #define PCRE_HASTHEN 0x00001000 /* pattern contains (*THEN) */
1089 #define PCRE_MLSET 0x00002000 /* match limit set by regex */
1090 #define PCRE_RLSET 0x00004000 /* recursion limit set by regex */
1091 #define PCRE_MATCH_EMPTY 0x00008000 /* pattern can match empty string */
1093 #if defined COMPILE_PCRE8
1094 #define PCRE_MODE PCRE_MODE8
1095 #elif defined COMPILE_PCRE16
1096 #define PCRE_MODE PCRE_MODE16
1097 #elif defined COMPILE_PCRE32
1098 #define PCRE_MODE PCRE_MODE32
1100 #define PCRE_MODE_MASK (PCRE_MODE8 | PCRE_MODE16 | PCRE_MODE32)
1102 /* Flags for the "extra" block produced by pcre_study(). */
1104 #define PCRE_STUDY_MAPPED 0x0001 /* a map of starting chars exists */
1105 #define PCRE_STUDY_MINLEN 0x0002 /* a minimum length field exists */
1107 /* Masks for identifying the public options that are permitted at compile
1108 time, run time, or study time, respectively. */
1110 #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
1111 PCRE_NEWLINE_ANYCRLF)
1113 #define PUBLIC_COMPILE_OPTIONS \
1114 (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
1115 PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
1116 PCRE_NO_AUTO_CAPTURE|PCRE_NO_AUTO_POSSESS| \
1117 PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
1118 PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
1119 PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE|PCRE_NEVER_UTF)
1121 #define PUBLIC_EXEC_OPTIONS \
1122 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
1123 PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
1124 PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
1126 #define PUBLIC_DFA_EXEC_OPTIONS \
1127 (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
1128 PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
1129 PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
1130 PCRE_NO_START_OPTIMIZE)
1132 #define PUBLIC_STUDY_OPTIONS \
1133 (PCRE_STUDY_JIT_COMPILE|PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE| \
1134 PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE|PCRE_STUDY_EXTRA_NEEDED)
1136 #define PUBLIC_JIT_EXEC_OPTIONS \
1137 (PCRE_NO_UTF8_CHECK|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|\
1138 PCRE_NOTEMPTY_ATSTART|PCRE_PARTIAL_SOFT|PCRE_PARTIAL_HARD)
1140 /* Magic number to provide a small check against being handed junk. */
1142 #define MAGIC_NUMBER 0x50435245UL /* 'PCRE' */
1144 /* This variable is used to detect a loaded regular expression
1145 in different endianness. */
1147 #define REVERSED_MAGIC_NUMBER 0x45524350UL /* 'ERCP' */
1149 /* The maximum remaining length of subject we are prepared to search for a
1152 #define REQ_BYTE_MAX 1000
1154 /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
1155 environments where these macros are defined elsewhere. Unfortunately, there
1156 is no way to do the same for the typedef. */
1165 /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
1166 character constants like '*' because the compiler would emit their EBCDIC code,
1167 which is different from their ASCII/UTF-8 code. Instead we define macros for
1168 the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
1169 is enabled. When UTF-8 support is not enabled, the definitions use character
1170 literals. Both character and string versions of each character are needed, and
1171 there are some longer strings as well.
1173 This means that, on EBCDIC platforms, the PCRE library can handle either
1174 EBCDIC, or UTF-8, but not both. To support both in the same compiled library
1175 would need different lookups depending on whether PCRE_UTF8 was set or not.
1176 This would make it impossible to use characters in switch/case statements,
1177 which would reduce performance. For a theoretical use (which nobody has asked
1178 for) in a minority area (EBCDIC platforms), this is not sensible. Any
1179 application that did need both could compile two versions of the library, using
1180 macros to give the functions distinct names. */
1184 /* UTF-8 support is not enabled; use the platform-dependent character literals
1185 so that PCRE works in both ASCII and EBCDIC environments, but only in non-UTF
1186 mode. Newline characters are problematic in EBCDIC. Though it has CR and LF
1187 characters, a common practice has been to use its NL (0x15) character as the
1188 line terminator in C-like processing environments. However, sometimes the LF
1189 (0x25) character is used instead, according to this Unicode document:
1191 http://unicode.org/standard/reports/tr13/tr13-5.html
1193 PCRE defaults EBCDIC NL to 0x15, but has a build-time option to select 0x25
1194 instead. Whichever is *not* chosen is defined as NEL.
1196 In both ASCII and EBCDIC environments, CHAR_NL and CHAR_LF are synonyms for the
1202 #define CHAR_NL '\x15'
1203 #define CHAR_NEL '\x25'
1204 #define STR_NL "\x15"
1205 #define STR_NEL "\x25"
1207 #define CHAR_NL '\x25'
1208 #define CHAR_NEL '\x15'
1209 #define STR_NL "\x25"
1210 #define STR_NEL "\x15"
1213 #define CHAR_LF CHAR_NL
1214 #define STR_LF STR_NL
1216 #define CHAR_ESC '\047'
1217 #define CHAR_DEL '\007'
1218 #define STR_ESC "\047"
1219 #define STR_DEL "\007"
1221 #else /* Not EBCDIC */
1223 /* In ASCII/Unicode, linefeed is '\n' and we equate this to NL for
1224 compatibility. NEL is the Unicode newline character; make sure it is
1225 a positive value. */
1227 #define CHAR_LF '\n'
1228 #define CHAR_NL CHAR_LF
1229 #define CHAR_NEL ((unsigned char)'\x85')
1230 #define CHAR_ESC '\033'
1231 #define CHAR_DEL '\177'
1234 #define STR_NL STR_LF
1235 #define STR_NEL "\x85"
1236 #define STR_ESC "\033"
1237 #define STR_DEL "\177"
1241 /* The remaining definitions work in both environments. */
1243 #define CHAR_NULL '\0'
1244 #define CHAR_HT '\t'
1245 #define CHAR_VT '\v'
1246 #define CHAR_FF '\f'
1247 #define CHAR_CR '\r'
1248 #define CHAR_BS '\b'
1249 #define CHAR_BEL '\a'
1251 #define CHAR_SPACE ' '
1252 #define CHAR_EXCLAMATION_MARK '!'
1253 #define CHAR_QUOTATION_MARK '"'
1254 #define CHAR_NUMBER_SIGN '#'
1255 #define CHAR_DOLLAR_SIGN '$'
1256 #define CHAR_PERCENT_SIGN '%'
1257 #define CHAR_AMPERSAND '&'
1258 #define CHAR_APOSTROPHE '\''
1259 #define CHAR_LEFT_PARENTHESIS '('
1260 #define CHAR_RIGHT_PARENTHESIS ')'
1261 #define CHAR_ASTERISK '*'
1262 #define CHAR_PLUS '+'
1263 #define CHAR_COMMA ','
1264 #define CHAR_MINUS '-'
1265 #define CHAR_DOT '.'
1266 #define CHAR_SLASH '/'
1277 #define CHAR_COLON ':'
1278 #define CHAR_SEMICOLON ';'
1279 #define CHAR_LESS_THAN_SIGN '<'
1280 #define CHAR_EQUALS_SIGN '='
1281 #define CHAR_GREATER_THAN_SIGN '>'
1282 #define CHAR_QUESTION_MARK '?'
1283 #define CHAR_COMMERCIAL_AT '@'
1310 #define CHAR_LEFT_SQUARE_BRACKET '['
1311 #define CHAR_BACKSLASH '\\'
1312 #define CHAR_RIGHT_SQUARE_BRACKET ']'
1313 #define CHAR_CIRCUMFLEX_ACCENT '^'
1314 #define CHAR_UNDERSCORE '_'
1315 #define CHAR_GRAVE_ACCENT '`'
1342 #define CHAR_LEFT_CURLY_BRACKET '{'
1343 #define CHAR_VERTICAL_LINE '|'
1344 #define CHAR_RIGHT_CURLY_BRACKET '}'
1345 #define CHAR_TILDE '~'
1352 #define STR_BEL "\a"
1354 #define STR_SPACE " "
1355 #define STR_EXCLAMATION_MARK "!"
1356 #define STR_QUOTATION_MARK "\""
1357 #define STR_NUMBER_SIGN "#"
1358 #define STR_DOLLAR_SIGN "$"
1359 #define STR_PERCENT_SIGN "%"
1360 #define STR_AMPERSAND "&"
1361 #define STR_APOSTROPHE "'"
1362 #define STR_LEFT_PARENTHESIS "("
1363 #define STR_RIGHT_PARENTHESIS ")"
1364 #define STR_ASTERISK "*"
1365 #define STR_PLUS "+"
1366 #define STR_COMMA ","
1367 #define STR_MINUS "-"
1369 #define STR_SLASH "/"
1380 #define STR_COLON ":"
1381 #define STR_SEMICOLON ";"
1382 #define STR_LESS_THAN_SIGN "<"
1383 #define STR_EQUALS_SIGN "="
1384 #define STR_GREATER_THAN_SIGN ">"
1385 #define STR_QUESTION_MARK "?"
1386 #define STR_COMMERCIAL_AT "@"
1413 #define STR_LEFT_SQUARE_BRACKET "["
1414 #define STR_BACKSLASH "\\"
1415 #define STR_RIGHT_SQUARE_BRACKET "]"
1416 #define STR_CIRCUMFLEX_ACCENT "^"
1417 #define STR_UNDERSCORE "_"
1418 #define STR_GRAVE_ACCENT "`"
1445 #define STR_LEFT_CURLY_BRACKET "{"
1446 #define STR_VERTICAL_LINE "|"
1447 #define STR_RIGHT_CURLY_BRACKET "}"
1448 #define STR_TILDE "~"
1450 #define STRING_ACCEPT0 "ACCEPT\0"
1451 #define STRING_COMMIT0 "COMMIT\0"
1452 #define STRING_F0 "F\0"
1453 #define STRING_FAIL0 "FAIL\0"
1454 #define STRING_MARK0 "MARK\0"
1455 #define STRING_PRUNE0 "PRUNE\0"
1456 #define STRING_SKIP0 "SKIP\0"
1457 #define STRING_THEN "THEN"
1459 #define STRING_alpha0 "alpha\0"
1460 #define STRING_lower0 "lower\0"
1461 #define STRING_upper0 "upper\0"
1462 #define STRING_alnum0 "alnum\0"
1463 #define STRING_ascii0 "ascii\0"
1464 #define STRING_blank0 "blank\0"
1465 #define STRING_cntrl0 "cntrl\0"
1466 #define STRING_digit0 "digit\0"
1467 #define STRING_graph0 "graph\0"
1468 #define STRING_print0 "print\0"
1469 #define STRING_punct0 "punct\0"
1470 #define STRING_space0 "space\0"
1471 #define STRING_word0 "word\0"
1472 #define STRING_xdigit "xdigit"
1474 #define STRING_DEFINE "DEFINE"
1475 #define STRING_WEIRD_STARTWORD "[:<:]]"
1476 #define STRING_WEIRD_ENDWORD "[:>:]]"
1478 #define STRING_CR_RIGHTPAR "CR)"
1479 #define STRING_LF_RIGHTPAR "LF)"
1480 #define STRING_CRLF_RIGHTPAR "CRLF)"
1481 #define STRING_ANY_RIGHTPAR "ANY)"
1482 #define STRING_ANYCRLF_RIGHTPAR "ANYCRLF)"
1483 #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
1484 #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
1485 #define STRING_UTF8_RIGHTPAR "UTF8)"
1486 #define STRING_UTF16_RIGHTPAR "UTF16)"
1487 #define STRING_UTF32_RIGHTPAR "UTF32)"
1488 #define STRING_UTF_RIGHTPAR "UTF)"
1489 #define STRING_UCP_RIGHTPAR "UCP)"
1490 #define STRING_NO_AUTO_POSSESS_RIGHTPAR "NO_AUTO_POSSESS)"
1491 #define STRING_NO_START_OPT_RIGHTPAR "NO_START_OPT)"
1492 #define STRING_LIMIT_MATCH_EQ "LIMIT_MATCH="
1493 #define STRING_LIMIT_RECURSION_EQ "LIMIT_RECURSION="
1495 #else /* SUPPORT_UTF */
1497 /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
1498 works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
1501 #define CHAR_HT '\011'
1502 #define CHAR_VT '\013'
1503 #define CHAR_FF '\014'
1504 #define CHAR_CR '\015'
1505 #define CHAR_LF '\012'
1506 #define CHAR_NL CHAR_LF
1507 #define CHAR_NEL ((unsigned char)'\x85')
1508 #define CHAR_BS '\010'
1509 #define CHAR_BEL '\007'
1510 #define CHAR_ESC '\033'
1511 #define CHAR_DEL '\177'
1513 #define CHAR_NULL '\0'
1514 #define CHAR_SPACE '\040'
1515 #define CHAR_EXCLAMATION_MARK '\041'
1516 #define CHAR_QUOTATION_MARK '\042'
1517 #define CHAR_NUMBER_SIGN '\043'
1518 #define CHAR_DOLLAR_SIGN '\044'
1519 #define CHAR_PERCENT_SIGN '\045'
1520 #define CHAR_AMPERSAND '\046'
1521 #define CHAR_APOSTROPHE '\047'
1522 #define CHAR_LEFT_PARENTHESIS '\050'
1523 #define CHAR_RIGHT_PARENTHESIS '\051'
1524 #define CHAR_ASTERISK '\052'
1525 #define CHAR_PLUS '\053'
1526 #define CHAR_COMMA '\054'
1527 #define CHAR_MINUS '\055'
1528 #define CHAR_DOT '\056'
1529 #define CHAR_SLASH '\057'
1530 #define CHAR_0 '\060'
1531 #define CHAR_1 '\061'
1532 #define CHAR_2 '\062'
1533 #define CHAR_3 '\063'
1534 #define CHAR_4 '\064'
1535 #define CHAR_5 '\065'
1536 #define CHAR_6 '\066'
1537 #define CHAR_7 '\067'
1538 #define CHAR_8 '\070'
1539 #define CHAR_9 '\071'
1540 #define CHAR_COLON '\072'
1541 #define CHAR_SEMICOLON '\073'
1542 #define CHAR_LESS_THAN_SIGN '\074'
1543 #define CHAR_EQUALS_SIGN '\075'
1544 #define CHAR_GREATER_THAN_SIGN '\076'
1545 #define CHAR_QUESTION_MARK '\077'
1546 #define CHAR_COMMERCIAL_AT '\100'
1547 #define CHAR_A '\101'
1548 #define CHAR_B '\102'
1549 #define CHAR_C '\103'
1550 #define CHAR_D '\104'
1551 #define CHAR_E '\105'
1552 #define CHAR_F '\106'
1553 #define CHAR_G '\107'
1554 #define CHAR_H '\110'
1555 #define CHAR_I '\111'
1556 #define CHAR_J '\112'
1557 #define CHAR_K '\113'
1558 #define CHAR_L '\114'
1559 #define CHAR_M '\115'
1560 #define CHAR_N '\116'
1561 #define CHAR_O '\117'
1562 #define CHAR_P '\120'
1563 #define CHAR_Q '\121'
1564 #define CHAR_R '\122'
1565 #define CHAR_S '\123'
1566 #define CHAR_T '\124'
1567 #define CHAR_U '\125'
1568 #define CHAR_V '\126'
1569 #define CHAR_W '\127'
1570 #define CHAR_X '\130'
1571 #define CHAR_Y '\131'
1572 #define CHAR_Z '\132'
1573 #define CHAR_LEFT_SQUARE_BRACKET '\133'
1574 #define CHAR_BACKSLASH '\134'
1575 #define CHAR_RIGHT_SQUARE_BRACKET '\135'
1576 #define CHAR_CIRCUMFLEX_ACCENT '\136'
1577 #define CHAR_UNDERSCORE '\137'
1578 #define CHAR_GRAVE_ACCENT '\140'
1579 #define CHAR_a '\141'
1580 #define CHAR_b '\142'
1581 #define CHAR_c '\143'
1582 #define CHAR_d '\144'
1583 #define CHAR_e '\145'
1584 #define CHAR_f '\146'
1585 #define CHAR_g '\147'
1586 #define CHAR_h '\150'
1587 #define CHAR_i '\151'
1588 #define CHAR_j '\152'
1589 #define CHAR_k '\153'
1590 #define CHAR_l '\154'
1591 #define CHAR_m '\155'
1592 #define CHAR_n '\156'
1593 #define CHAR_o '\157'
1594 #define CHAR_p '\160'
1595 #define CHAR_q '\161'
1596 #define CHAR_r '\162'
1597 #define CHAR_s '\163'
1598 #define CHAR_t '\164'
1599 #define CHAR_u '\165'
1600 #define CHAR_v '\166'
1601 #define CHAR_w '\167'
1602 #define CHAR_x '\170'
1603 #define CHAR_y '\171'
1604 #define CHAR_z '\172'
1605 #define CHAR_LEFT_CURLY_BRACKET '\173'
1606 #define CHAR_VERTICAL_LINE '\174'
1607 #define CHAR_RIGHT_CURLY_BRACKET '\175'
1608 #define CHAR_TILDE '\176'
1610 #define STR_HT "\011"
1611 #define STR_VT "\013"
1612 #define STR_FF "\014"
1613 #define STR_CR "\015"
1614 #define STR_NL "\012"
1615 #define STR_BS "\010"
1616 #define STR_BEL "\007"
1617 #define STR_ESC "\033"
1618 #define STR_DEL "\177"
1620 #define STR_SPACE "\040"
1621 #define STR_EXCLAMATION_MARK "\041"
1622 #define STR_QUOTATION_MARK "\042"
1623 #define STR_NUMBER_SIGN "\043"
1624 #define STR_DOLLAR_SIGN "\044"
1625 #define STR_PERCENT_SIGN "\045"
1626 #define STR_AMPERSAND "\046"
1627 #define STR_APOSTROPHE "\047"
1628 #define STR_LEFT_PARENTHESIS "\050"
1629 #define STR_RIGHT_PARENTHESIS "\051"
1630 #define STR_ASTERISK "\052"
1631 #define STR_PLUS "\053"
1632 #define STR_COMMA "\054"
1633 #define STR_MINUS "\055"
1634 #define STR_DOT "\056"
1635 #define STR_SLASH "\057"
1636 #define STR_0 "\060"
1637 #define STR_1 "\061"
1638 #define STR_2 "\062"
1639 #define STR_3 "\063"
1640 #define STR_4 "\064"
1641 #define STR_5 "\065"
1642 #define STR_6 "\066"
1643 #define STR_7 "\067"
1644 #define STR_8 "\070"
1645 #define STR_9 "\071"
1646 #define STR_COLON "\072"
1647 #define STR_SEMICOLON "\073"
1648 #define STR_LESS_THAN_SIGN "\074"
1649 #define STR_EQUALS_SIGN "\075"
1650 #define STR_GREATER_THAN_SIGN "\076"
1651 #define STR_QUESTION_MARK "\077"
1652 #define STR_COMMERCIAL_AT "\100"
1653 #define STR_A "\101"
1654 #define STR_B "\102"
1655 #define STR_C "\103"
1656 #define STR_D "\104"
1657 #define STR_E "\105"
1658 #define STR_F "\106"
1659 #define STR_G "\107"
1660 #define STR_H "\110"
1661 #define STR_I "\111"
1662 #define STR_J "\112"
1663 #define STR_K "\113"
1664 #define STR_L "\114"
1665 #define STR_M "\115"
1666 #define STR_N "\116"
1667 #define STR_O "\117"
1668 #define STR_P "\120"
1669 #define STR_Q "\121"
1670 #define STR_R "\122"
1671 #define STR_S "\123"
1672 #define STR_T "\124"
1673 #define STR_U "\125"
1674 #define STR_V "\126"
1675 #define STR_W "\127"
1676 #define STR_X "\130"
1677 #define STR_Y "\131"
1678 #define STR_Z "\132"
1679 #define STR_LEFT_SQUARE_BRACKET "\133"
1680 #define STR_BACKSLASH "\134"
1681 #define STR_RIGHT_SQUARE_BRACKET "\135"
1682 #define STR_CIRCUMFLEX_ACCENT "\136"
1683 #define STR_UNDERSCORE "\137"
1684 #define STR_GRAVE_ACCENT "\140"
1685 #define STR_a "\141"
1686 #define STR_b "\142"
1687 #define STR_c "\143"
1688 #define STR_d "\144"
1689 #define STR_e "\145"
1690 #define STR_f "\146"
1691 #define STR_g "\147"
1692 #define STR_h "\150"
1693 #define STR_i "\151"
1694 #define STR_j "\152"
1695 #define STR_k "\153"
1696 #define STR_l "\154"
1697 #define STR_m "\155"
1698 #define STR_n "\156"
1699 #define STR_o "\157"
1700 #define STR_p "\160"
1701 #define STR_q "\161"
1702 #define STR_r "\162"
1703 #define STR_s "\163"
1704 #define STR_t "\164"
1705 #define STR_u "\165"
1706 #define STR_v "\166"
1707 #define STR_w "\167"
1708 #define STR_x "\170"
1709 #define STR_y "\171"
1710 #define STR_z "\172"
1711 #define STR_LEFT_CURLY_BRACKET "\173"
1712 #define STR_VERTICAL_LINE "\174"
1713 #define STR_RIGHT_CURLY_BRACKET "\175"
1714 #define STR_TILDE "\176"
1716 #define STRING_ACCEPT0 STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1717 #define STRING_COMMIT0 STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1718 #define STRING_F0 STR_F "\0"
1719 #define STRING_FAIL0 STR_F STR_A STR_I STR_L "\0"
1720 #define STRING_MARK0 STR_M STR_A STR_R STR_K "\0"
1721 #define STRING_PRUNE0 STR_P STR_R STR_U STR_N STR_E "\0"
1722 #define STRING_SKIP0 STR_S STR_K STR_I STR_P "\0"
1723 #define STRING_THEN STR_T STR_H STR_E STR_N
1725 #define STRING_alpha0 STR_a STR_l STR_p STR_h STR_a "\0"
1726 #define STRING_lower0 STR_l STR_o STR_w STR_e STR_r "\0"
1727 #define STRING_upper0 STR_u STR_p STR_p STR_e STR_r "\0"
1728 #define STRING_alnum0 STR_a STR_l STR_n STR_u STR_m "\0"
1729 #define STRING_ascii0 STR_a STR_s STR_c STR_i STR_i "\0"
1730 #define STRING_blank0 STR_b STR_l STR_a STR_n STR_k "\0"
1731 #define STRING_cntrl0 STR_c STR_n STR_t STR_r STR_l "\0"
1732 #define STRING_digit0 STR_d STR_i STR_g STR_i STR_t "\0"
1733 #define STRING_graph0 STR_g STR_r STR_a STR_p STR_h "\0"
1734 #define STRING_print0 STR_p STR_r STR_i STR_n STR_t "\0"
1735 #define STRING_punct0 STR_p STR_u STR_n STR_c STR_t "\0"
1736 #define STRING_space0 STR_s STR_p STR_a STR_c STR_e "\0"
1737 #define STRING_word0 STR_w STR_o STR_r STR_d "\0"
1738 #define STRING_xdigit STR_x STR_d STR_i STR_g STR_i STR_t
1740 #define STRING_DEFINE STR_D STR_E STR_F STR_I STR_N STR_E
1741 #define STRING_WEIRD_STARTWORD STR_LEFT_SQUARE_BRACKET STR_COLON STR_LESS_THAN_SIGN STR_COLON STR_RIGHT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET
1742 #define STRING_WEIRD_ENDWORD STR_LEFT_SQUARE_BRACKET STR_COLON STR_GREATER_THAN_SIGN STR_COLON STR_RIGHT_SQUARE_BRACKET STR_RIGHT_SQUARE_BRACKET
1744 #define STRING_CR_RIGHTPAR STR_C STR_R STR_RIGHT_PARENTHESIS
1745 #define STRING_LF_RIGHTPAR STR_L STR_F STR_RIGHT_PARENTHESIS
1746 #define STRING_CRLF_RIGHTPAR STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1747 #define STRING_ANY_RIGHTPAR STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1748 #define STRING_ANYCRLF_RIGHTPAR STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1749 #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
1750 #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
1751 #define STRING_UTF8_RIGHTPAR STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1752 #define STRING_UTF16_RIGHTPAR STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS
1753 #define STRING_UTF32_RIGHTPAR STR_U STR_T STR_F STR_3 STR_2 STR_RIGHT_PARENTHESIS
1754 #define STRING_UTF_RIGHTPAR STR_U STR_T STR_F STR_RIGHT_PARENTHESIS
1755 #define STRING_UCP_RIGHTPAR STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1756 #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
1757 #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
1758 #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
1759 #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
1761 #endif /* SUPPORT_UTF */
1763 /* Escape items that are just an encoding of a particular data value. */
1766 #define ESC_e CHAR_ESC
1770 #define ESC_f CHAR_FF
1774 #define ESC_n CHAR_LF
1778 #define ESC_r CHAR_CR
1781 /* We can't officially use ESC_t because it is a POSIX reserved identifier
1782 (presumably because of all the others like size_t). */
1785 #define ESC_tee CHAR_HT
1788 /* Codes for different types of Unicode property */
1790 #define PT_ANY 0 /* Any property - matches all chars */
1791 #define PT_LAMP 1 /* L& - the union of Lu, Ll, Lt */
1792 #define PT_GC 2 /* Specified general characteristic (e.g. L) */
1793 #define PT_PC 3 /* Specified particular characteristic (e.g. Lu) */
1794 #define PT_SC 4 /* Script (e.g. Han) */
1795 #define PT_ALNUM 5 /* Alphanumeric - the union of L and N */
1796 #define PT_SPACE 6 /* Perl space - Z plus 9,10,12,13 */
1797 #define PT_PXSPACE 7 /* POSIX space - Z plus 9,10,11,12,13 */
1798 #define PT_WORD 8 /* Word - L plus N plus underscore */
1799 #define PT_CLIST 9 /* Pseudo-property: match character list */
1800 #define PT_UCNC 10 /* Universal Character nameable character */
1801 #define PT_TABSIZE 11 /* Size of square table for autopossessify tests */
1803 /* The following special properties are used only in XCLASS items, when POSIX
1804 classes are specified and PCRE_UCP is set - in other words, for Unicode
1805 handling of these classes. They are not available via the \p or \P escapes like
1806 those in the above list, and so they do not take part in the autopossessifying
1809 #define PT_PXGRAPH 11 /* [:graph:] - characters that mark the paper */
1810 #define PT_PXPRINT 12 /* [:print:] - [:graph:] plus non-control spaces */
1811 #define PT_PXPUNCT 13 /* [:punct:] - punctuation characters */
1813 /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
1814 contain characters with values greater than 255. */
1816 #define XCL_NOT 0x01 /* Flag: this is a negative class */
1817 #define XCL_MAP 0x02 /* Flag: a 32-byte map is present */
1818 #define XCL_HASPROP 0x04 /* Flag: property checks are present. */
1820 #define XCL_END 0 /* Marks end of individual items */
1821 #define XCL_SINGLE 1 /* Single item (one multibyte char) follows */
1822 #define XCL_RANGE 2 /* A range (two multibyte chars) follows */
1823 #define XCL_PROP 3 /* Unicode property (2-byte property code follows) */
1824 #define XCL_NOTPROP 4 /* Unicode inverted property (ditto) */
1826 /* These are escaped items that aren't just an encoding of a particular data
1827 value such as \n. They must have non-zero values, as check_escape() returns 0
1828 for a data character. Also, they must appear in the same order as in the
1829 opcode definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it
1830 corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1831 used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In
1832 non-DOTALL mode, "." behaves like \N.
1834 The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1835 when PCRE_UCP is set and replacement of \d etc by \p sequences is required.
1836 They must be contiguous, and remain in order so that the replacements can be
1837 looked up from a table.
1839 Negative numbers are used to encode a backreference (\1, \2, \3, etc.) in
1840 check_escape(). There are two tests in the code for an escape
1841 greater than ESC_b and less than ESC_Z to detect the types that may be
1842 repeated. These are the types that consume characters. If any new escapes are
1843 put in between that don't consume a character, that code will have to change.
1846 enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
1847 ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1848 ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1849 ESC_E, ESC_Q, ESC_g, ESC_k,
1850 ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu };
1853 /********************** Opcode definitions ******************/
1855 /****** NOTE NOTE NOTE ******
1857 Starting from 1 (i.e. after OP_END), the values up to OP_EOD must correspond in
1858 order to the list of escapes immediately above. Furthermore, values up to
1859 OP_DOLLM must not be changed without adjusting the table called autoposstab in
1862 Whenever this list is updated, the two macro definitions that follow must be
1863 updated to match. The possessification table called "opcode_possessify" in
1864 pcre_compile.c must also be updated, and also the tables called "coptable"
1865 and "poptable" in pcre_dfa_exec.c.
1867 ****** NOTE NOTE NOTE ******/
1870 /* The values between FIRST_AUTOTAB_OP and LAST_AUTOTAB_RIGHT_OP, inclusive,
1871 are used in a table for deciding whether a repeated character type can be
1872 auto-possessified. */
1874 #define FIRST_AUTOTAB_OP OP_NOT_DIGIT
1875 #define LAST_AUTOTAB_LEFT_OP OP_EXTUNI
1876 #define LAST_AUTOTAB_RIGHT_OP OP_DOLLM
1879 OP_END, /* 0 End of pattern */
1881 /* Values corresponding to backslashed metacharacters */
1883 OP_SOD, /* 1 Start of data: \A */
1884 OP_SOM, /* 2 Start of match (subject + offset): \G */
1885 OP_SET_SOM, /* 3 Set start of match (\K) */
1886 OP_NOT_WORD_BOUNDARY, /* 4 \B */
1887 OP_WORD_BOUNDARY, /* 5 \b */
1888 OP_NOT_DIGIT, /* 6 \D */
1889 OP_DIGIT, /* 7 \d */
1890 OP_NOT_WHITESPACE, /* 8 \S */
1891 OP_WHITESPACE, /* 9 \s */
1892 OP_NOT_WORDCHAR, /* 10 \W */
1893 OP_WORDCHAR, /* 11 \w */
1895 OP_ANY, /* 12 Match any character except newline (\N) */
1896 OP_ALLANY, /* 13 Match any character */
1897 OP_ANYBYTE, /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1898 OP_NOTPROP, /* 15 \P (not Unicode property) */
1899 OP_PROP, /* 16 \p (Unicode property) */
1900 OP_ANYNL, /* 17 \R (any newline sequence) */
1901 OP_NOT_HSPACE, /* 18 \H (not horizontal whitespace) */
1902 OP_HSPACE, /* 19 \h (horizontal whitespace) */
1903 OP_NOT_VSPACE, /* 20 \V (not vertical whitespace) */
1904 OP_VSPACE, /* 21 \v (vertical whitespace) */
1905 OP_EXTUNI, /* 22 \X (extended Unicode sequence */
1906 OP_EODN, /* 23 End of data or \n at end of data (\Z) */
1907 OP_EOD, /* 24 End of data (\z) */
1909 /* Line end assertions */
1911 OP_DOLL, /* 25 End of line - not multiline */
1912 OP_DOLLM, /* 26 End of line - multiline */
1913 OP_CIRC, /* 27 Start of line - not multiline */
1914 OP_CIRCM, /* 28 Start of line - multiline */
1916 /* Single characters; caseful must precede the caseless ones */
1918 OP_CHAR, /* 29 Match one character, casefully */
1919 OP_CHARI, /* 30 Match one character, caselessly */
1920 OP_NOT, /* 31 Match one character, not the given one, casefully */
1921 OP_NOTI, /* 32 Match one character, not the given one, caselessly */
1923 /* The following sets of 13 opcodes must always be kept in step because
1924 the offset from the first one is used to generate the others. */
1926 /* Repeated characters; caseful must precede the caseless ones */
1928 OP_STAR, /* 33 The maximizing and minimizing versions of */
1929 OP_MINSTAR, /* 34 these six opcodes must come in pairs, with */
1930 OP_PLUS, /* 35 the minimizing one second. */
1931 OP_MINPLUS, /* 36 */
1933 OP_MINQUERY, /* 38 */
1935 OP_UPTO, /* 39 From 0 to n matches of one character, caseful*/
1936 OP_MINUPTO, /* 40 */
1937 OP_EXACT, /* 41 Exactly n matches */
1939 OP_POSSTAR, /* 42 Possessified star, caseful */
1940 OP_POSPLUS, /* 43 Possessified plus, caseful */
1941 OP_POSQUERY, /* 44 Posesssified query, caseful */
1942 OP_POSUPTO, /* 45 Possessified upto, caseful */
1944 /* Repeated characters; caseless must follow the caseful ones */
1947 OP_MINSTARI, /* 47 */
1949 OP_MINPLUSI, /* 49 */
1951 OP_MINQUERYI, /* 51 */
1953 OP_UPTOI, /* 52 From 0 to n matches of one character, caseless */
1954 OP_MINUPTOI, /* 53 */
1957 OP_POSSTARI, /* 55 Possessified star, caseless */
1958 OP_POSPLUSI, /* 56 Possessified plus, caseless */
1959 OP_POSQUERYI, /* 57 Posesssified query, caseless */
1960 OP_POSUPTOI, /* 58 Possessified upto, caseless */
1962 /* The negated ones must follow the non-negated ones, and match them */
1963 /* Negated repeated character, caseful; must precede the caseless ones */
1965 OP_NOTSTAR, /* 59 The maximizing and minimizing versions of */
1966 OP_NOTMINSTAR, /* 60 these six opcodes must come in pairs, with */
1967 OP_NOTPLUS, /* 61 the minimizing one second. They must be in */
1968 OP_NOTMINPLUS, /* 62 exactly the same order as those above. */
1969 OP_NOTQUERY, /* 63 */
1970 OP_NOTMINQUERY, /* 64 */
1972 OP_NOTUPTO, /* 65 From 0 to n matches, caseful */
1973 OP_NOTMINUPTO, /* 66 */
1974 OP_NOTEXACT, /* 67 Exactly n matches */
1976 OP_NOTPOSSTAR, /* 68 Possessified versions, caseful */
1977 OP_NOTPOSPLUS, /* 69 */
1978 OP_NOTPOSQUERY, /* 70 */
1979 OP_NOTPOSUPTO, /* 71 */
1981 /* Negated repeated character, caseless; must follow the caseful ones */
1983 OP_NOTSTARI, /* 72 */
1984 OP_NOTMINSTARI, /* 73 */
1985 OP_NOTPLUSI, /* 74 */
1986 OP_NOTMINPLUSI, /* 75 */
1987 OP_NOTQUERYI, /* 76 */
1988 OP_NOTMINQUERYI, /* 77 */
1990 OP_NOTUPTOI, /* 78 From 0 to n matches, caseless */
1991 OP_NOTMINUPTOI, /* 79 */
1992 OP_NOTEXACTI, /* 80 Exactly n matches */
1994 OP_NOTPOSSTARI, /* 81 Possessified versions, caseless */
1995 OP_NOTPOSPLUSI, /* 82 */
1996 OP_NOTPOSQUERYI, /* 83 */
1997 OP_NOTPOSUPTOI, /* 84 */
1999 /* Character types */
2001 OP_TYPESTAR, /* 85 The maximizing and minimizing versions of */
2002 OP_TYPEMINSTAR, /* 86 these six opcodes must come in pairs, with */
2003 OP_TYPEPLUS, /* 87 the minimizing one second. These codes must */
2004 OP_TYPEMINPLUS, /* 88 be in exactly the same order as those above. */
2005 OP_TYPEQUERY, /* 89 */
2006 OP_TYPEMINQUERY, /* 90 */
2008 OP_TYPEUPTO, /* 91 From 0 to n matches */
2009 OP_TYPEMINUPTO, /* 92 */
2010 OP_TYPEEXACT, /* 93 Exactly n matches */
2012 OP_TYPEPOSSTAR, /* 94 Possessified versions */
2013 OP_TYPEPOSPLUS, /* 95 */
2014 OP_TYPEPOSQUERY, /* 96 */
2015 OP_TYPEPOSUPTO, /* 97 */
2017 /* These are used for character classes and back references; only the
2018 first six are the same as the sets above. */
2020 OP_CRSTAR, /* 98 The maximizing and minimizing versions of */
2021 OP_CRMINSTAR, /* 99 all these opcodes must come in pairs, with */
2022 OP_CRPLUS, /* 100 the minimizing one second. These codes must */
2023 OP_CRMINPLUS, /* 101 be in exactly the same order as those above. */
2024 OP_CRQUERY, /* 102 */
2025 OP_CRMINQUERY, /* 103 */
2027 OP_CRRANGE, /* 104 These are different to the three sets above. */
2028 OP_CRMINRANGE, /* 105 */
2030 OP_CRPOSSTAR, /* 106 Possessified versions */
2031 OP_CRPOSPLUS, /* 107 */
2032 OP_CRPOSQUERY, /* 108 */
2033 OP_CRPOSRANGE, /* 109 */
2035 /* End of quantifier opcodes */
2037 OP_CLASS, /* 110 Match a character class, chars < 256 only */
2038 OP_NCLASS, /* 111 Same, but the bitmap was created from a negative
2039 class - the difference is relevant only when a
2040 character > 255 is encountered. */
2041 OP_XCLASS, /* 112 Extended class for handling > 255 chars within the
2042 class. This does both positive and negative. */
2043 OP_REF, /* 113 Match a back reference, casefully */
2044 OP_REFI, /* 114 Match a back reference, caselessly */
2045 OP_DNREF, /* 115 Match a duplicate name backref, casefully */
2046 OP_DNREFI, /* 116 Match a duplicate name backref, caselessly */
2047 OP_RECURSE, /* 117 Match a numbered subpattern (possibly recursive) */
2048 OP_CALLOUT, /* 118 Call out to external function if provided */
2050 OP_ALT, /* 119 Start of alternation */
2051 OP_KET, /* 120 End of group that doesn't have an unbounded repeat */
2052 OP_KETRMAX, /* 121 These two must remain together and in this */
2053 OP_KETRMIN, /* 122 order. They are for groups the repeat for ever. */
2054 OP_KETRPOS, /* 123 Possessive unlimited repeat. */
2056 /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
2057 asserts must remain in order. */
2059 OP_REVERSE, /* 124 Move pointer back - used in lookbehind assertions */
2060 OP_ASSERT, /* 125 Positive lookahead */
2061 OP_ASSERT_NOT, /* 126 Negative lookahead */
2062 OP_ASSERTBACK, /* 127 Positive lookbehind */
2063 OP_ASSERTBACK_NOT, /* 128 Negative lookbehind */
2065 /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
2066 after the assertions, with ONCE first, as there's a test for >= ONCE for a
2067 subpattern that isn't an assertion. The POS versions must immediately follow
2068 the non-POS versions in each case. */
2070 OP_ONCE, /* 129 Atomic group, contains captures */
2071 OP_ONCE_NC, /* 130 Atomic group containing no captures */
2072 OP_BRA, /* 131 Start of non-capturing bracket */
2073 OP_BRAPOS, /* 132 Ditto, with unlimited, possessive repeat */
2074 OP_CBRA, /* 133 Start of capturing bracket */
2075 OP_CBRAPOS, /* 134 Ditto, with unlimited, possessive repeat */
2076 OP_COND, /* 135 Conditional group */
2078 /* These five must follow the previous five, in the same order. There's a
2079 check for >= SBRA to distinguish the two sets. */
2081 OP_SBRA, /* 136 Start of non-capturing bracket, check empty */
2082 OP_SBRAPOS, /* 137 Ditto, with unlimited, possessive repeat */
2083 OP_SCBRA, /* 138 Start of capturing bracket, check empty */
2084 OP_SCBRAPOS, /* 139 Ditto, with unlimited, possessive repeat */
2085 OP_SCOND, /* 140 Conditional group, check empty */
2087 /* The next two pairs must (respectively) be kept together. */
2089 OP_CREF, /* 141 Used to hold a capture number as condition */
2090 OP_DNCREF, /* 142 Used to point to duplicate names as a condition */
2091 OP_RREF, /* 143 Used to hold a recursion number as condition */
2092 OP_DNRREF, /* 144 Used to point to duplicate names as a condition */
2093 OP_DEF, /* 145 The DEFINE condition */
2095 OP_BRAZERO, /* 146 These two must remain together and in this */
2096 OP_BRAMINZERO, /* 147 order. */
2097 OP_BRAPOSZERO, /* 148 */
2099 /* These are backtracking control verbs */
2101 OP_MARK, /* 149 always has an argument */
2103 OP_PRUNE_ARG, /* 151 same, but with argument */
2105 OP_SKIP_ARG, /* 153 same, but with argument */
2107 OP_THEN_ARG, /* 155 same, but with argument */
2108 OP_COMMIT, /* 156 */
2110 /* These are forced failure and success verbs */
2113 OP_ACCEPT, /* 158 */
2114 OP_ASSERT_ACCEPT, /* 159 Used inside assertions */
2115 OP_CLOSE, /* 160 Used before OP_ACCEPT to close open captures */
2117 /* This is used to skip a subpattern with a {0} quantifier */
2119 OP_SKIPZERO, /* 161 */
2121 /* This is not an opcode, but is used to check that tables indexed by opcode
2122 are the correct length, in order to catch updating errors - there have been
2123 some in the past. */
2128 /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
2129 definitions that follow must also be updated to match. There are also tables
2130 called "opcode_possessify" in pcre_compile.c and "coptable" and "poptable" in
2131 pcre_dfa_exec.c that must be updated. */
2134 /* This macro defines textual names for all the opcodes. These are used only
2135 for debugging, and some of them are only partial names. The macro is referenced
2136 only in pcre_printint.c, which fills out the full names in many cases (and in
2137 some cases doesn't actually use these names at all). */
2139 #define OP_NAME_LIST \
2140 "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d", \
2141 "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte", \
2142 "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v", \
2143 "extuni", "\\Z", "\\z", \
2144 "$", "$", "^", "^", "char", "chari", "not", "noti", \
2145 "*", "*?", "+", "+?", "?", "??", \
2147 "*+","++", "?+", "{", \
2148 "*", "*?", "+", "+?", "?", "??", \
2150 "*+","++", "?+", "{", \
2151 "*", "*?", "+", "+?", "?", "??", \
2153 "*+","++", "?+", "{", \
2154 "*", "*?", "+", "+?", "?", "??", \
2156 "*+","++", "?+", "{", \
2157 "*", "*?", "+", "+?", "?", "??", "{", "{", "{", \
2158 "*+","++", "?+", "{", \
2159 "*", "*?", "+", "+?", "?", "??", "{", "{", \
2160 "*+","++", "?+", "{", \
2161 "class", "nclass", "xclass", "Ref", "Refi", "DnRef", "DnRefi", \
2162 "Recurse", "Callout", \
2163 "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos", \
2164 "Reverse", "Assert", "Assert not", "AssertB", "AssertB not", \
2165 "Once", "Once_NC", \
2166 "Bra", "BraPos", "CBra", "CBraPos", \
2168 "SBra", "SBraPos", "SCBra", "SCBraPos", \
2170 "Cond ref", "Cond dnref", "Cond rec", "Cond dnrec", "Cond def", \
2171 "Brazero", "Braminzero", "Braposzero", \
2172 "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP", \
2173 "*THEN", "*THEN", "*COMMIT", "*FAIL", \
2174 "*ACCEPT", "*ASSERT_ACCEPT", \
2175 "Close", "Skip zero"
2178 /* This macro defines the length of fixed length operations in the compiled
2179 regex. The lengths are used when searching for specific things, and also in the
2180 debugging printing of a compiled regex. We use a macro so that it can be
2181 defined close to the definitions of the opcodes themselves.
2183 As things have been extended, some of these are no longer fixed lenths, but are
2184 minima instead. For example, the length of a single-character repeat may vary
2185 in UTF-8 mode. The code that uses this table must know about such things. */
2187 #define OP_LENGTHS \
2189 1, 1, 1, 1, 1, /* \A, \G, \K, \B, \b */ \
2190 1, 1, 1, 1, 1, 1, /* \D, \d, \S, \s, \W, \w */ \
2191 1, 1, 1, /* Any, AllAny, Anybyte */ \
2192 3, 3, /* \P, \p */ \
2193 1, 1, 1, 1, 1, /* \R, \H, \h, \V, \v */ \
2195 1, 1, 1, 1, 1, 1, /* \Z, \z, $, $M ^, ^M */ \
2196 2, /* Char - the minimum length */ \
2197 2, /* Chari - the minimum length */ \
2200 /* Positive single-char repeats ** These are */ \
2201 2, 2, 2, 2, 2, 2, /* *, *?, +, +?, ?, ?? ** minima in */ \
2202 2+IMM2_SIZE, 2+IMM2_SIZE, /* upto, minupto ** mode */ \
2203 2+IMM2_SIZE, /* exact */ \
2204 2, 2, 2, 2+IMM2_SIZE, /* *+, ++, ?+, upto+ */ \
2205 2, 2, 2, 2, 2, 2, /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8 */ \
2206 2+IMM2_SIZE, 2+IMM2_SIZE, /* upto I, minupto I */ \
2207 2+IMM2_SIZE, /* exact I */ \
2208 2, 2, 2, 2+IMM2_SIZE, /* *+I, ++I, ?+I, upto+I */ \
2209 /* Negative single-char repeats - only for chars < 256 */ \
2210 2, 2, 2, 2, 2, 2, /* NOT *, *?, +, +?, ?, ?? */ \
2211 2+IMM2_SIZE, 2+IMM2_SIZE, /* NOT upto, minupto */ \
2212 2+IMM2_SIZE, /* NOT exact */ \
2213 2, 2, 2, 2+IMM2_SIZE, /* Possessive NOT *, +, ?, upto */ \
2214 2, 2, 2, 2, 2, 2, /* NOT *I, *?I, +I, +?I, ?I, ??I */ \
2215 2+IMM2_SIZE, 2+IMM2_SIZE, /* NOT upto I, minupto I */ \
2216 2+IMM2_SIZE, /* NOT exact I */ \
2217 2, 2, 2, 2+IMM2_SIZE, /* Possessive NOT *I, +I, ?I, upto I */ \
2218 /* Positive type repeats */ \
2219 2, 2, 2, 2, 2, 2, /* Type *, *?, +, +?, ?, ?? */ \
2220 2+IMM2_SIZE, 2+IMM2_SIZE, /* Type upto, minupto */ \
2221 2+IMM2_SIZE, /* Type exact */ \
2222 2, 2, 2, 2+IMM2_SIZE, /* Possessive *+, ++, ?+, upto+ */ \
2223 /* Character class & ref repeats */ \
2224 1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ \
2225 1+2*IMM2_SIZE, 1+2*IMM2_SIZE, /* CRRANGE, CRMINRANGE */ \
2226 1, 1, 1, 1+2*IMM2_SIZE, /* Possessive *+, ++, ?+, CRPOSRANGE */ \
2227 1+(32/sizeof(pcre_uchar)), /* CLASS */ \
2228 1+(32/sizeof(pcre_uchar)), /* NCLASS */ \
2229 0, /* XCLASS - variable length */ \
2230 1+IMM2_SIZE, /* REF */ \
2231 1+IMM2_SIZE, /* REFI */ \
2232 1+2*IMM2_SIZE, /* DNREF */ \
2233 1+2*IMM2_SIZE, /* DNREFI */ \
2234 1+LINK_SIZE, /* RECURSE */ \
2235 2+2*LINK_SIZE, /* CALLOUT */ \
2236 1+LINK_SIZE, /* Alt */ \
2237 1+LINK_SIZE, /* Ket */ \
2238 1+LINK_SIZE, /* KetRmax */ \
2239 1+LINK_SIZE, /* KetRmin */ \
2240 1+LINK_SIZE, /* KetRpos */ \
2241 1+LINK_SIZE, /* Reverse */ \
2242 1+LINK_SIZE, /* Assert */ \
2243 1+LINK_SIZE, /* Assert not */ \
2244 1+LINK_SIZE, /* Assert behind */ \
2245 1+LINK_SIZE, /* Assert behind not */ \
2246 1+LINK_SIZE, /* ONCE */ \
2247 1+LINK_SIZE, /* ONCE_NC */ \
2248 1+LINK_SIZE, /* BRA */ \
2249 1+LINK_SIZE, /* BRAPOS */ \
2250 1+LINK_SIZE+IMM2_SIZE, /* CBRA */ \
2251 1+LINK_SIZE+IMM2_SIZE, /* CBRAPOS */ \
2252 1+LINK_SIZE, /* COND */ \
2253 1+LINK_SIZE, /* SBRA */ \
2254 1+LINK_SIZE, /* SBRAPOS */ \
2255 1+LINK_SIZE+IMM2_SIZE, /* SCBRA */ \
2256 1+LINK_SIZE+IMM2_SIZE, /* SCBRAPOS */ \
2257 1+LINK_SIZE, /* SCOND */ \
2258 1+IMM2_SIZE, 1+2*IMM2_SIZE, /* CREF, DNCREF */ \
2259 1+IMM2_SIZE, 1+2*IMM2_SIZE, /* RREF, DNRREF */ \
2261 1, 1, 1, /* BRAZERO, BRAMINZERO, BRAPOSZERO */ \
2262 3, 1, 3, /* MARK, PRUNE, PRUNE_ARG */ \
2263 1, 3, /* SKIP, SKIP_ARG */ \
2264 1, 3, /* THEN, THEN_ARG */ \
2265 1, 1, 1, 1, /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT */ \
2266 1+IMM2_SIZE, 1 /* CLOSE, SKIPZERO */
2268 /* A magic value for OP_RREF to indicate the "any recursion" condition. */
2270 #define RREF_ANY 0xffff
2272 /* Compile time error code numbers. They are given names so that they can more
2273 easily be tracked. When a new number is added, the table called eint in
2274 pcreposix.c must be updated. */
2276 enum { ERR0, ERR1, ERR2, ERR3, ERR4, ERR5, ERR6, ERR7, ERR8, ERR9,
2277 ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
2278 ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
2279 ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
2280 ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
2281 ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
2282 ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
2283 ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERR77, ERR78, ERR79,
2284 ERR80, ERR81, ERR82, ERR83, ERR84, ERR85, ERRCOUNT };
2286 /* JIT compiling modes. The function list is indexed by them. */
2288 enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
2289 JIT_NUMBER_OF_COMPILE_MODES };
2291 /* The real format of the start of the pcre block; the index of names and the
2292 code vector run on as long as necessary after the end. We store an explicit
2293 offset to the name table so that if a regex is compiled on one host, saved, and
2294 then run on another where the size of pointers is different, all might still
2297 The size of the structure must be a multiple of 8 bytes. For the case of
2298 compiled-on-4 and run-on-8, we include an extra pointer that is always NULL so
2299 that there are an even number of pointers which therefore are a multiple of 8
2302 It is necessary to fork the struct for the 32 bit library, since it needs to
2303 use pcre_uint32 for first_char and req_char. We can't put an ifdef inside the
2304 typedef because pcretest needs access to the struct of the 8-, 16- and 32-bit
2308 When new fields are added to these structures, remember to adjust the code in
2309 pcre_byte_order.c that is concerned with swapping the byte order of the fields
2310 when a compiled regex is reloaded on a host with different endianness.
2312 There is also similar byte-flipping code in pcretest.c, which is used for
2313 testing the byte-flipping features. It must also be kept in step.
2317 typedef struct real_pcre8_or_16 {
2318 pcre_uint32 magic_number;
2319 pcre_uint32 size; /* Total that was malloced */
2320 pcre_uint32 options; /* Public options */
2321 pcre_uint32 flags; /* Private flags */
2322 pcre_uint32 limit_match; /* Limit set from regex */
2323 pcre_uint32 limit_recursion; /* Limit set from regex */
2324 pcre_uint16 first_char; /* Starting character */
2325 pcre_uint16 req_char; /* This character must be seen */
2326 pcre_uint16 max_lookbehind; /* Longest lookbehind (characters) */
2327 pcre_uint16 top_bracket; /* Highest numbered group */
2328 pcre_uint16 top_backref; /* Highest numbered back reference */
2329 pcre_uint16 name_table_offset; /* Offset to name table that follows */
2330 pcre_uint16 name_entry_size; /* Size of any name items */
2331 pcre_uint16 name_count; /* Number of name items */
2332 pcre_uint16 ref_count; /* Reference count */
2333 pcre_uint16 dummy1; /* To ensure size is a multiple of 8 */
2334 pcre_uint16 dummy2; /* To ensure size is a multiple of 8 */
2335 pcre_uint16 dummy3; /* To ensure size is a multiple of 8 */
2336 const pcre_uint8 *tables; /* Pointer to tables or NULL for std */
2337 void *nullpad; /* NULL padding */
2340 typedef struct real_pcre8_or_16 real_pcre;
2341 typedef struct real_pcre8_or_16 real_pcre16;
2343 typedef struct real_pcre32 {
2344 pcre_uint32 magic_number;
2345 pcre_uint32 size; /* Total that was malloced */
2346 pcre_uint32 options; /* Public options */
2347 pcre_uint32 flags; /* Private flags */
2348 pcre_uint32 limit_match; /* Limit set from regex */
2349 pcre_uint32 limit_recursion; /* Limit set from regex */
2350 pcre_uint32 first_char; /* Starting character */
2351 pcre_uint32 req_char; /* This character must be seen */
2352 pcre_uint16 max_lookbehind; /* Longest lookbehind (characters) */
2353 pcre_uint16 top_bracket; /* Highest numbered group */
2354 pcre_uint16 top_backref; /* Highest numbered back reference */
2355 pcre_uint16 name_table_offset; /* Offset to name table that follows */
2356 pcre_uint16 name_entry_size; /* Size of any name items */
2357 pcre_uint16 name_count; /* Number of name items */
2358 pcre_uint16 ref_count; /* Reference count */
2359 pcre_uint16 dummy; /* To ensure size is a multiple of 8 */
2360 const pcre_uint8 *tables; /* Pointer to tables or NULL for std */
2361 void *nullpad; /* NULL padding */
2364 #if defined COMPILE_PCRE8
2365 #define REAL_PCRE real_pcre
2366 #elif defined COMPILE_PCRE16
2367 #define REAL_PCRE real_pcre16
2368 #elif defined COMPILE_PCRE32
2369 #define REAL_PCRE real_pcre32
2372 /* Assert that the size of REAL_PCRE is divisible by 8 */
2373 typedef int __assert_real_pcre_size_divisible_8[(sizeof(REAL_PCRE) % 8) == 0 ? 1 : -1];
2375 /* Needed in pcretest to access some fields in the real_pcre* structures
2376 * directly. They're unified for 8/16/32 bits since the structs only differ
2377 * after these fields; if that ever changes, need to fork those defines into
2378 * 8/16 and 32 bit versions. */
2379 #define REAL_PCRE_MAGIC(re) (((REAL_PCRE*)re)->magic_number)
2380 #define REAL_PCRE_SIZE(re) (((REAL_PCRE*)re)->size)
2381 #define REAL_PCRE_OPTIONS(re) (((REAL_PCRE*)re)->options)
2382 #define REAL_PCRE_FLAGS(re) (((REAL_PCRE*)re)->flags)
2384 /* The format of the block used to store data from pcre_study(). The same
2385 remark (see NOTE above) about extending this structure applies. */
2387 typedef struct pcre_study_data {
2388 pcre_uint32 size; /* Total that was malloced */
2389 pcre_uint32 flags; /* Private flags */
2390 pcre_uint8 start_bits[32]; /* Starting char bits */
2391 pcre_uint32 minlength; /* Minimum subject length */
2394 /* Structure for building a chain of open capturing subpatterns during
2395 compiling, so that instructions to close them can be compiled when (*ACCEPT) is
2396 encountered. This is also used to identify subpatterns that contain recursive
2397 back references to themselves, so that they can be made atomic. */
2399 typedef struct open_capitem {
2400 struct open_capitem *next; /* Chain link */
2401 pcre_uint16 number; /* Capture number */
2402 pcre_uint16 flag; /* Set TRUE if recursive back ref */
2405 /* Structure for building a list of named groups during the first pass of
2408 typedef struct named_group {
2409 const pcre_uchar *name; /* Points to the name in the pattern */
2410 int length; /* Length of the name */
2411 pcre_uint32 number; /* Group number */
2414 /* Structure for passing "static" information around between the functions
2415 doing the compiling, so that they are thread-safe. */
2417 typedef struct compile_data {
2418 const pcre_uint8 *lcc; /* Points to lower casing table */
2419 const pcre_uint8 *fcc; /* Points to case-flipping table */
2420 const pcre_uint8 *cbits; /* Points to character type table */
2421 const pcre_uint8 *ctypes; /* Points to table of type maps */
2422 const pcre_uchar *start_workspace;/* The start of working space */
2423 const pcre_uchar *start_code; /* The start of the compiled code */
2424 const pcre_uchar *start_pattern; /* The start of the pattern */
2425 const pcre_uchar *end_pattern; /* The end of the pattern */
2426 pcre_uchar *hwm; /* High watermark of workspace */
2427 open_capitem *open_caps; /* Chain of open capture items */
2428 named_group *named_groups; /* Points to vector in pre-compile */
2429 pcre_uchar *name_table; /* The name/number table */
2430 int names_found; /* Number of entries so far */
2431 int name_entry_size; /* Size of each entry */
2432 int named_group_list_size; /* Number of entries in the list */
2433 int workspace_size; /* Size of workspace */
2434 unsigned int bracount; /* Count of capturing parens as we compile */
2435 int final_bracount; /* Saved value after first pass */
2436 int max_lookbehind; /* Maximum lookbehind (characters) */
2437 int top_backref; /* Maximum back reference */
2438 unsigned int backref_map; /* Bitmap of low back refs */
2439 unsigned int namedrefcount; /* Number of backreferences by name */
2440 int parens_depth; /* Depth of nested parentheses */
2441 int assert_depth; /* Depth of nested assertions */
2442 pcre_uint32 external_options; /* External (initial) options */
2443 pcre_uint32 external_flags; /* External flag bits to be set */
2444 int req_varyopt; /* "After variable item" flag for reqbyte */
2445 BOOL had_accept; /* (*ACCEPT) encountered */
2446 BOOL had_pruneorskip; /* (*PRUNE) or (*SKIP) encountered */
2447 BOOL check_lookbehind; /* Lookbehinds need later checking */
2448 BOOL dupnames; /* Duplicate names exist */
2449 int nltype; /* Newline type */
2450 int nllen; /* Newline string length */
2451 pcre_uchar nl[4]; /* Newline string when fixed length */
2454 /* Structure for maintaining a chain of pointers to the currently incomplete
2455 branches, for testing for left recursion while compiling. */
2457 typedef struct branch_chain {
2458 struct branch_chain *outer;
2459 pcre_uchar *current_branch;
2462 /* Structure for items in a linked list that represents an explicit recursive
2463 call within the pattern; used by pcre_exec(). */
2465 typedef struct recursion_info {
2466 struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
2467 unsigned int group_num; /* Number of group that was called */
2468 int *offset_save; /* Pointer to start of saved offsets */
2469 int saved_max; /* Number of saved offsets */
2470 int saved_capture_last; /* Last capture number */
2471 PCRE_PUCHAR subject_position; /* Position at start of recursion */
2474 /* A similar structure for pcre_dfa_exec(). */
2476 typedef struct dfa_recursion_info {
2477 struct dfa_recursion_info *prevrec;
2479 PCRE_PUCHAR subject_position;
2480 } dfa_recursion_info;
2482 /* Structure for building a chain of data for holding the values of the subject
2483 pointer at the start of each subpattern, so as to detect when an empty string
2484 has been matched by a subpattern - to break infinite loops; used by
2487 typedef struct eptrblock {
2488 struct eptrblock *epb_prev;
2489 PCRE_PUCHAR epb_saved_eptr;
2493 /* Structure for passing "static" information around between the functions
2494 doing traditional NFA matching, so that they are thread-safe. */
2496 typedef struct match_data {
2497 unsigned long int match_call_count; /* As it says */
2498 unsigned long int match_limit; /* As it says */
2499 unsigned long int match_limit_recursion; /* As it says */
2500 int *offset_vector; /* Offset vector */
2501 int offset_end; /* One past the end */
2502 int offset_max; /* The maximum usable for return data */
2503 int nltype; /* Newline type */
2504 int nllen; /* Newline string length */
2505 int name_count; /* Number of names in name table */
2506 int name_entry_size; /* Size of entry in names table */
2507 unsigned int skip_arg_count; /* For counting SKIP_ARGs */
2508 unsigned int ignore_skip_arg; /* For re-run when SKIP arg name not found */
2509 pcre_uchar *name_table; /* Table of names */
2510 pcre_uchar nl[4]; /* Newline string when fixed */
2511 const pcre_uint8 *lcc; /* Points to lower casing table */
2512 const pcre_uint8 *fcc; /* Points to case-flipping table */
2513 const pcre_uint8 *ctypes; /* Points to table of type maps */
2514 BOOL notbol; /* NOTBOL flag */
2515 BOOL noteol; /* NOTEOL flag */
2516 BOOL utf; /* UTF-8 / UTF-16 flag */
2517 BOOL jscript_compat; /* JAVASCRIPT_COMPAT flag */
2518 BOOL use_ucp; /* PCRE_UCP flag */
2519 BOOL endonly; /* Dollar not before final \n */
2520 BOOL notempty; /* Empty string match not wanted */
2521 BOOL notempty_atstart; /* Empty string match at start not wanted */
2522 BOOL hitend; /* Hit the end of the subject at some point */
2523 BOOL bsr_anycrlf; /* \R is just any CRLF, not full Unicode */
2524 BOOL hasthen; /* Pattern contains (*THEN) */
2525 const pcre_uchar *start_code; /* For use when recursing */
2526 PCRE_PUCHAR start_subject; /* Start of the subject string */
2527 PCRE_PUCHAR end_subject; /* End of the subject string */
2528 PCRE_PUCHAR start_match_ptr; /* Start of matched string */
2529 PCRE_PUCHAR end_match_ptr; /* Subject position at end match */
2530 PCRE_PUCHAR start_used_ptr; /* Earliest consulted character */
2531 int partial; /* PARTIAL options */
2532 int end_offset_top; /* Highwater mark at end of match */
2533 pcre_int32 capture_last; /* Most recent capture number + overflow flag */
2534 int start_offset; /* The start offset value */
2535 int match_function_type; /* Set for certain special calls of MATCH() */
2536 eptrblock *eptrchain; /* Chain of eptrblocks for tail recursions */
2537 int eptrn; /* Next free eptrblock */
2538 recursion_info *recursive; /* Linked list of recursion data */
2539 void *callout_data; /* To pass back to callouts */
2540 const pcre_uchar *mark; /* Mark pointer to pass back on success */
2541 const pcre_uchar *nomatch_mark;/* Mark pointer to pass back on failure */
2542 const pcre_uchar *once_target; /* Where to back up to for atomic groups */
2544 void *match_frames_base; /* For remembering malloc'd frames */
2548 /* A similar structure is used for the same purpose by the DFA matching
2551 typedef struct dfa_match_data {
2552 const pcre_uchar *start_code; /* Start of the compiled pattern */
2553 const pcre_uchar *start_subject ; /* Start of the subject string */
2554 const pcre_uchar *end_subject; /* End of subject string */
2555 const pcre_uchar *start_used_ptr; /* Earliest consulted character */
2556 const pcre_uint8 *tables; /* Character tables */
2557 int start_offset; /* The start offset value */
2558 int moptions; /* Match options */
2559 int poptions; /* Pattern options */
2560 int nltype; /* Newline type */
2561 int nllen; /* Newline string length */
2562 pcre_uchar nl[4]; /* Newline string when fixed */
2563 void *callout_data; /* To pass back to callouts */
2564 dfa_recursion_info *recursive; /* Linked list of recursion data */
2567 /* Bit definitions for entries in the pcre_ctypes table. */
2569 #define ctype_space 0x01
2570 #define ctype_letter 0x02
2571 #define ctype_digit 0x04
2572 #define ctype_xdigit 0x08
2573 #define ctype_word 0x10 /* alphanumeric or '_' */
2574 #define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
2576 /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
2577 of bits for a class map. Some classes are built by combining these tables. */
2579 #define cbit_space 0 /* [:space:] or \s */
2580 #define cbit_xdigit 32 /* [:xdigit:] */
2581 #define cbit_digit 64 /* [:digit:] or \d */
2582 #define cbit_upper 96 /* [:upper:] */
2583 #define cbit_lower 128 /* [:lower:] */
2584 #define cbit_word 160 /* [:word:] or \w */
2585 #define cbit_graph 192 /* [:graph:] */
2586 #define cbit_print 224 /* [:print:] */
2587 #define cbit_punct 256 /* [:punct:] */
2588 #define cbit_cntrl 288 /* [:cntrl:] */
2589 #define cbit_length 320 /* Length of the cbits table */
2591 /* Offsets of the various tables from the base tables pointer, and
2594 #define lcc_offset 0
2595 #define fcc_offset 256
2596 #define cbits_offset 512
2597 #define ctypes_offset (cbits_offset + cbit_length)
2598 #define tables_length (ctypes_offset + 256)
2600 /* Internal function and data prefixes. */
2602 #if defined COMPILE_PCRE8
2604 #define PUBL(name) pcre_##name
2607 #define PRIV(name) _pcre_##name
2609 #elif defined COMPILE_PCRE16
2611 #define PUBL(name) pcre16_##name
2614 #define PRIV(name) _pcre16_##name
2616 #elif defined COMPILE_PCRE32
2618 #define PUBL(name) pcre32_##name
2621 #define PRIV(name) _pcre32_##name
2624 #error Unsupported compiling mode
2625 #endif /* COMPILE_PCRE[8|16|32] */
2627 /* Layout of the UCP type table that translates property names into types and
2628 codes. Each entry used to point directly to a name, but to reduce the number of
2629 relocations in shared libraries, it now has an offset into a single string
2633 pcre_uint16 name_offset;
2639 /* Internal shared data tables. These are tables that are used by more than one
2640 of the exported public functions. They have to be "external" in the C sense,
2641 but are not part of the PCRE public API. The data for these tables is in the
2642 pcre_tables.c module. */
2644 #ifdef COMPILE_PCRE8
2645 extern const int PRIV(utf8_table1)[];
2646 extern const int PRIV(utf8_table1_size);
2647 extern const int PRIV(utf8_table2)[];
2648 extern const int PRIV(utf8_table3)[];
2649 extern const pcre_uint8 PRIV(utf8_table4)[];
2650 #endif /* COMPILE_PCRE8 */
2652 extern const char PRIV(utt_names)[];
2653 extern const ucp_type_table PRIV(utt)[];
2654 extern const int PRIV(utt_size);
2656 extern const pcre_uint8 PRIV(OP_lengths)[];
2657 extern const pcre_uint8 PRIV(default_tables)[];
2659 extern const pcre_uint32 PRIV(hspace_list)[];
2660 extern const pcre_uint32 PRIV(vspace_list)[];
2663 /* Internal shared functions. These are functions that are used by more than
2664 one of the exported public functions. They have to be "external" in the C
2665 sense, but are not part of the PCRE public API. */
2667 /* String comparison functions. */
2668 #if defined COMPILE_PCRE8
2670 #define STRCMP_UC_UC(str1, str2) \
2671 strcmp((char *)(str1), (char *)(str2))
2672 #define STRCMP_UC_C8(str1, str2) \
2673 strcmp((char *)(str1), (str2))
2674 #define STRNCMP_UC_UC(str1, str2, num) \
2675 strncmp((char *)(str1), (char *)(str2), (num))
2676 #define STRNCMP_UC_C8(str1, str2, num) \
2677 strncmp((char *)(str1), (str2), (num))
2678 #define STRLEN_UC(str) strlen((const char *)str)
2680 #elif defined COMPILE_PCRE16 || defined COMPILE_PCRE32
2682 extern int PRIV(strcmp_uc_uc)(const pcre_uchar *,
2683 const pcre_uchar *);
2684 extern int PRIV(strcmp_uc_c8)(const pcre_uchar *,
2686 extern int PRIV(strncmp_uc_uc)(const pcre_uchar *,
2687 const pcre_uchar *, unsigned int num);
2688 extern int PRIV(strncmp_uc_c8)(const pcre_uchar *,
2689 const char *, unsigned int num);
2690 extern unsigned int PRIV(strlen_uc)(const pcre_uchar *str);
2692 #define STRCMP_UC_UC(str1, str2) \
2693 PRIV(strcmp_uc_uc)((str1), (str2))
2694 #define STRCMP_UC_C8(str1, str2) \
2695 PRIV(strcmp_uc_c8)((str1), (str2))
2696 #define STRNCMP_UC_UC(str1, str2, num) \
2697 PRIV(strncmp_uc_uc)((str1), (str2), (num))
2698 #define STRNCMP_UC_C8(str1, str2, num) \
2699 PRIV(strncmp_uc_c8)((str1), (str2), (num))
2700 #define STRLEN_UC(str) PRIV(strlen_uc)(str)
2702 #endif /* COMPILE_PCRE[8|16|32] */
2704 #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
2706 #define STRCMP_UC_UC_TEST(str1, str2) STRCMP_UC_UC(str1, str2)
2707 #define STRCMP_UC_C8_TEST(str1, str2) STRCMP_UC_C8(str1, str2)
2709 #elif defined COMPILE_PCRE32
2711 extern int PRIV(strcmp_uc_uc_utf)(const pcre_uchar *,
2712 const pcre_uchar *);
2713 extern int PRIV(strcmp_uc_c8_utf)(const pcre_uchar *,
2716 #define STRCMP_UC_UC_TEST(str1, str2) \
2717 (utf ? PRIV(strcmp_uc_uc_utf)((str1), (str2)) : PRIV(strcmp_uc_uc)((str1), (str2)))
2718 #define STRCMP_UC_C8_TEST(str1, str2) \
2719 (utf ? PRIV(strcmp_uc_c8_utf)((str1), (str2)) : PRIV(strcmp_uc_c8)((str1), (str2)))
2721 #endif /* COMPILE_PCRE[8|16|32] */
2723 extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2724 extern BOOL PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2726 extern unsigned int PRIV(ord2utf)(pcre_uint32, pcre_uchar *);
2727 extern int PRIV(valid_utf)(PCRE_PUCHAR, int, int *);
2728 extern BOOL PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2730 extern BOOL PRIV(xclass)(pcre_uint32, const pcre_uchar *, BOOL);
2733 extern void PRIV(jit_compile)(const REAL_PCRE *,
2734 PUBL(extra) *, int);
2735 extern int PRIV(jit_exec)(const PUBL(extra) *,
2736 const pcre_uchar *, int, int, int, int *, int);
2737 extern void PRIV(jit_free)(void *);
2738 extern int PRIV(jit_get_size)(void *);
2739 extern const char* PRIV(jit_get_target)(void);
2742 /* Unicode character database (UCD) */
2745 pcre_uint8 script; /* ucp_Arabic, etc. */
2746 pcre_uint8 chartype; /* ucp_Cc, etc. (general categories) */
2747 pcre_uint8 gbprop; /* ucp_gbControl, etc. (grapheme break property) */
2748 pcre_uint8 caseset; /* offset to multichar other cases or zero */
2749 pcre_int32 other_case; /* offset to other case, or zero if none */
2752 extern const pcre_uint32 PRIV(ucd_caseless_sets)[];
2753 extern const ucd_record PRIV(ucd_records)[];
2754 extern const pcre_uint8 PRIV(ucd_stage1)[];
2755 extern const pcre_uint16 PRIV(ucd_stage2)[];
2756 extern const pcre_uint32 PRIV(ucp_gentype)[];
2757 extern const pcre_uint32 PRIV(ucp_gbtable)[];
2759 extern const int PRIV(ucp_typerange)[];
2763 /* UCD access macros */
2765 #define UCD_BLOCK_SIZE 128
2766 #define GET_UCD(ch) (PRIV(ucd_records) + \
2767 PRIV(ucd_stage2)[PRIV(ucd_stage1)[(int)(ch) / UCD_BLOCK_SIZE] * \
2768 UCD_BLOCK_SIZE + (int)(ch) % UCD_BLOCK_SIZE])
2770 #define UCD_CHARTYPE(ch) GET_UCD(ch)->chartype
2771 #define UCD_SCRIPT(ch) GET_UCD(ch)->script
2772 #define UCD_CATEGORY(ch) PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2773 #define UCD_GRAPHBREAK(ch) GET_UCD(ch)->gbprop
2774 #define UCD_CASESET(ch) GET_UCD(ch)->caseset
2775 #define UCD_OTHERCASE(ch) ((pcre_uint32)((int)ch + (int)(GET_UCD(ch)->other_case)))
2777 #endif /* SUPPORT_UCP */
2781 /* End of pcre_internal.h */
~ian / pcre3.git / blob