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

Diff of /code/trunk/pcre_internal.h

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 137 by ph10, Thu Mar 29 13:56:00 2007 UTC revision 964 by ph10, Fri May 4 13:03:39 2012 UTC
# Line 7  Line 7 
7  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2007 University of Cambridge             Copyright (c) 1997-2012 University of Cambridge
11    
12  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
13  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 40  POSSIBILITY OF SUCH DAMAGE. Line 40  POSSIBILITY OF SUCH DAMAGE.
40    
41  /* This header contains definitions that are shared between the different  /* This header contains definitions that are shared between the different
42  modules, but which are not relevant to the exported API. This includes some  modules, but which are not relevant to the exported API. This includes some
43  functions whose names all begin with "_pcre_". */  functions whose names all begin with "_pcre_" or "_pcre16_" depending on
44    the PRIV macro. */
45    
46  #ifndef PCRE_INTERNAL_H  #ifndef PCRE_INTERNAL_H
47  #define PCRE_INTERNAL_H  #define PCRE_INTERNAL_H
48    
49  /* Define DEBUG to get debugging output on stdout. */  /* Define PCRE_DEBUG to get debugging output on stdout. */
50    
51  #if 0  #if 0
52  #define DEBUG  #define PCRE_DEBUG
53    #endif
54    
55    /* PCRE is compiled as an 8 bit library if it is not requested otherwise. */
56    #ifndef COMPILE_PCRE16
57    #define COMPILE_PCRE8
58    #endif
59    
60    /* If SUPPORT_UCP is defined, SUPPORT_UTF must also be defined. The
61    "configure" script ensures this, but not everybody uses "configure". */
62    
63    #if defined SUPPORT_UCP && !(defined SUPPORT_UTF)
64    #define SUPPORT_UTF 1
65    #endif
66    
67    /* We define SUPPORT_UTF if SUPPORT_UTF8 is enabled for compatibility
68    reasons with existing code. */
69    
70    #if defined SUPPORT_UTF8 && !(defined SUPPORT_UTF)
71    #define SUPPORT_UTF 1
72    #endif
73    
74    /* Fixme: SUPPORT_UTF8 should be eventually disappear from the code.
75    Until then we define it if SUPPORT_UTF is defined. */
76    
77    #if defined SUPPORT_UTF && !(defined SUPPORT_UTF8)
78    #define SUPPORT_UTF8 1
79    #endif
80    
81    /* We do not support both EBCDIC and UTF-8/16 at the same time. The "configure"
82    script prevents both being selected, but not everybody uses "configure". */
83    
84    #if defined EBCDIC && defined SUPPORT_UTF
85    #error The use of both EBCDIC and SUPPORT_UTF8/16 is not supported.
86  #endif  #endif
87    
88  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef
# Line 60  It turns out that the Mac Debugging.h he Line 94  It turns out that the Mac Debugging.h he
94  be absolutely sure we get our version. */  be absolutely sure we get our version. */
95    
96  #undef DPRINTF  #undef DPRINTF
97  #ifdef DEBUG  #ifdef PCRE_DEBUG
98  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
99  #else  #else
100  #define DPRINTF(p) /* Nothing */  #define DPRINTF(p) /* Nothing */
101  #endif  #endif
102    
103    
 /* Get the definitions provided by running "configure" */  
   
 #include "config.h"  
   
104  /* Standard C headers plus the external interface definition. The only time  /* Standard C headers plus the external interface definition. The only time
105  setjmp and stdarg are used is when NO_RECURSE is set. */  setjmp and stdarg are used is when NO_RECURSE is set. */
106    
107  #include <ctype.h>  #include <ctype.h>
108  #include <limits.h>  #include <limits.h>
 #include <setjmp.h>  
 #include <stdarg.h>  
109  #include <stddef.h>  #include <stddef.h>
110  #include <stdio.h>  #include <stdio.h>
111  #include <stdlib.h>  #include <stdlib.h>
112  #include <string.h>  #include <string.h>
113    
114  #ifndef PCRE_SPY  /* When compiling a DLL for Windows, the exported symbols have to be declared
115  #define PCRE_DEFINITION       /* Win32 __declspec(export) trigger for .dll */  using some MS magic. I found some useful information on this web page:
116    http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
117    information there, using __declspec(dllexport) without "extern" we have a
118    definition; with "extern" we have a declaration. The settings here override the
119    setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
120    which is all that is needed for applications (they just import the symbols). We
121    use:
122    
123      PCRE_EXP_DECL       for declarations
124      PCRE_EXP_DEFN       for definitions of exported functions
125      PCRE_EXP_DATA_DEFN  for definitions of exported variables
126    
127    The reason for the two DEFN macros is that in non-Windows environments, one
128    does not want to have "extern" before variable definitions because it leads to
129    compiler warnings. So we distinguish between functions and variables. In
130    Windows, the two should always be the same.
131    
132    The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
133    which is an application, but needs to import this file in order to "peek" at
134    internals, can #include pcre.h first to get an application's-eye view.
135    
136    In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
137    special-purpose environments) might want to stick other stuff in front of
138    exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
139    PCRE_EXP_DATA_DEFN only if they are not already set. */
140    
141    #ifndef PCRE_EXP_DECL
142    #  ifdef _WIN32
143    #    ifndef PCRE_STATIC
144    #      define PCRE_EXP_DECL       extern __declspec(dllexport)
145    #      define PCRE_EXP_DEFN       __declspec(dllexport)
146    #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
147    #    else
148    #      define PCRE_EXP_DECL       extern
149    #      define PCRE_EXP_DEFN
150    #      define PCRE_EXP_DATA_DEFN
151    #    endif
152    #  else
153    #    ifdef __cplusplus
154    #      define PCRE_EXP_DECL       extern "C"
155    #    else
156    #      define PCRE_EXP_DECL       extern
157    #    endif
158    #    ifndef PCRE_EXP_DEFN
159    #      define PCRE_EXP_DEFN       PCRE_EXP_DECL
160    #    endif
161    #    ifndef PCRE_EXP_DATA_DEFN
162    #      define PCRE_EXP_DATA_DEFN
163    #    endif
164    #  endif
165    #endif
166    
167    /* When compiling with the MSVC compiler, it is sometimes necessary to include
168    a "calling convention" before exported function names. (This is secondhand
169    information; I know nothing about MSVC myself). For example, something like
170    
171      void __cdecl function(....)
172    
173    might be needed. In order so make this easy, all the exported functions have
174    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
175    set, we ensure here that it has no effect. */
176    
177    #ifndef PCRE_CALL_CONVENTION
178    #define PCRE_CALL_CONVENTION
179  #endif  #endif
180    
181  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 8, 16 and 32-bit integers. We
182  cannot determine these outside the compilation (e.g. by running a program as  cannot determine these outside the compilation (e.g. by running a program as
183  part of "configure") because PCRE is often cross-compiled for use on other  part of "configure") because PCRE is often cross-compiled for use on other
184  systems. Instead we make use of the maximum sizes that are available at  systems. Instead we make use of the maximum sizes that are available at
185  preprocessor time in standard C environments. */  preprocessor time in standard C environments. */
186    
187    typedef unsigned char pcre_uint8;
188    
189  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
190    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
191      typedef short pcre_int16;
192  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
193    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
194      typedef int pcre_int16;
195  #else  #else
196    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
197  #endif  #endif
198    
199  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
200    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
201      typedef int pcre_int32;
202  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
203    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
204      typedef long int pcre_int32;
205  #else  #else
206    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
207  #endif  #endif
208    
209    /* When checking for integer overflow in pcre_compile(), we need to handle
210    large integers. If a 64-bit integer type is available, we can use that.
211    Otherwise we have to cast to double, which of course requires floating point
212    arithmetic. Handle this by defining a macro for the appropriate type. If
213    stdint.h is available, include it; it may define INT64_MAX. Systems that do not
214    have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
215    by "configure". */
216    
217    #if HAVE_STDINT_H
218    #include <stdint.h>
219    #elif HAVE_INTTYPES_H
220    #include <inttypes.h>
221    #endif
222    
223    #if defined INT64_MAX || defined int64_t
224    #define INT64_OR_DOUBLE int64_t
225    #else
226    #define INT64_OR_DOUBLE double
227    #endif
228    
229  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
230  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
231  However, we leave the interface to the outside world as char *, because that  However, we leave the interface to the outside world as char * or short *,
232  should make things easier for callers. We define a short type for unsigned char  because that should make things easier for callers. This character type is
233  to save lots of typing. I tried "uchar", but it causes problems on Digital  called pcre_uchar.
234  Unix, where it is defined in sys/types, so use "uschar" instead. */  
235    The IN_UCHARS macro multiply its argument with the byte size of the current
236    pcre_uchar type. Useful for memcpy and such operations, whose require the
237    byte size of their input/output buffers.
238    
239    The MAX_255 macro checks whether its pcre_uchar input is less than 256.
240    
241    The TABLE_GET macro is designed for accessing elements of tables whose contain
242    exactly 256 items. When the character is able to contain more than 256
243    items, some check is needed before accessing these tables.
244    */
245    
246  typedef unsigned char uschar;  #ifdef COMPILE_PCRE8
247    
248    typedef unsigned char pcre_uchar;
249    #define IN_UCHARS(x) (x)
250    #define MAX_255(c) 1
251    #define TABLE_GET(c, table, default) ((table)[c])
252    
253    #else
254    
255    #ifdef COMPILE_PCRE16
256    #if USHRT_MAX != 65535
257    /* This is a warning message. Change PCRE_UCHAR16 to a 16 bit data type in
258    pcre.h(.in) and disable (comment out) this message. */
259    #error Warning: PCRE_UCHAR16 is not a 16 bit data type.
260    #endif
261    
262    typedef pcre_uint16 pcre_uchar;
263    #define IN_UCHARS(x) ((x) << 1)
264    #define MAX_255(c) ((c) <= 255u)
265    #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
266    
267    #else
268    #error Unsupported compiling mode
269    #endif /* COMPILE_PCRE16 */
270    
271    #endif /* COMPILE_PCRE8 */
272    
273  /* This is an unsigned int value that no character can ever have. UTF-8  /* This is an unsigned int value that no character can ever have. UTF-8
274  characters only go up to 0x7fffffff (though Unicode doesn't go beyond  characters only go up to 0x7fffffff (though Unicode doesn't go beyond
# Line 125  characters only go up to 0x7fffffff (tho Line 277  characters only go up to 0x7fffffff (tho
277  #define NOTACHAR 0xffffffff  #define NOTACHAR 0xffffffff
278    
279  /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,  /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
280  and "all" at present). The following macros are used to package up testing for  "any" and "anycrlf" at present). The following macros are used to package up
281  newlines. NLBLOCK, PSSTART, and PSEND are defined in the various modules to  testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
282  indicate in which datablock the parameters exist, and what the start/end of  modules to indicate in which datablock the parameters exist, and what the
283  string field names are. */  start/end of string field names are. */
284    
285  #define NLTYPE_FIXED   0     /* Newline is a fixed length string */  #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
286  #define NLTYPE_ANY     1     /* Newline is any Unicode line ending */  #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
287    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
288    
289  /* This macro checks for a newline at the given position */  /* This macro checks for a newline at the given position */
290    
291  #define IS_NEWLINE(p) \  #define IS_NEWLINE(p) \
292    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
293      ((p) < NLBLOCK->PSEND && \      ((p) < NLBLOCK->PSEND && \
294       _pcre_is_newline((p), NLBLOCK->PSEND, &(NLBLOCK->nllen), utf8) \       PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, \
295      ) \         &(NLBLOCK->nllen), utf)) \
296      : \      : \
297      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
298       (p)[0] == NLBLOCK->nl[0] && \       (p)[0] == NLBLOCK->nl[0] && \
# Line 152  string field names are. */ Line 305  string field names are. */
305  #define WAS_NEWLINE(p) \  #define WAS_NEWLINE(p) \
306    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
307      ((p) > NLBLOCK->PSSTART && \      ((p) > NLBLOCK->PSSTART && \
308       _pcre_was_newline((p), NLBLOCK->PSSTART, &(NLBLOCK->nllen), utf8) \       PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
309      ) \         &(NLBLOCK->nllen), utf)) \
310      : \      : \
311      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
312       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
# Line 171  used for the external interface and appe Line 324  used for the external interface and appe
324  must begin with PCRE_. */  must begin with PCRE_. */
325    
326  #ifdef CUSTOM_SUBJECT_PTR  #ifdef CUSTOM_SUBJECT_PTR
327  #define PCRE_SPTR CUSTOM_SUBJECT_PTR  #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
 #define USPTR CUSTOM_SUBJECT_PTR  
328  #else  #else
329  #define PCRE_SPTR const char *  #define PCRE_PUCHAR const pcre_uchar *
 #define USPTR const unsigned char *  
330  #endif  #endif
331    
332  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
333  values. */  values. */
334    
335  #include <pcre.h>  #include "pcre.h"
336  #include "ucp.h"  #include "ucp.h"
337    
338  /* When compiling for use with the Virtual Pascal compiler, these functions  /* When compiling for use with the Virtual Pascal compiler, these functions
# Line 245  The macros are controlled by the value o Line 396  The macros are controlled by the value o
396  the config.h file, but can be overridden by using -D on the command line. This  the config.h file, but can be overridden by using -D on the command line. This
397  is automated on Unix systems via the "configure" command. */  is automated on Unix systems via the "configure" command. */
398    
399    #ifdef COMPILE_PCRE8
400    
401  #if LINK_SIZE == 2  #if LINK_SIZE == 2
402    
403  #define PUT(a,n,d)   \  #define PUT(a,n,d)   \
# Line 281  is automated on Unix systems via the "co Line 434  is automated on Unix systems via the "co
434  #define GET(a,n) \  #define GET(a,n) \
435    (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])    (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
436    
437  #define MAX_PATTERN_SIZE (1 << 30)   /* Keep it positive */  /* Keep it positive */
438    #define MAX_PATTERN_SIZE (1 << 30)
439    
440    #else
441    #error LINK_SIZE must be either 2, 3, or 4
442    #endif
443    
444    #else /* COMPILE_PCRE8 */
445    
446    #ifdef COMPILE_PCRE16
447    
448    #if LINK_SIZE == 2
449    
450    #undef LINK_SIZE
451    #define LINK_SIZE 1
452    
453    #define PUT(a,n,d)   \
454      (a[n] = (d))
455    
456    #define GET(a,n) \
457      (a[n])
458    
459    #define MAX_PATTERN_SIZE (1 << 16)
460    
461    #elif LINK_SIZE == 3 || LINK_SIZE == 4
462    
463    #undef LINK_SIZE
464    #define LINK_SIZE 2
465    
466    #define PUT(a,n,d)   \
467      (a[n] = (d) >> 16), \
468      (a[(n)+1] = (d) & 65535)
469    
470    #define GET(a,n) \
471      (((a)[n] << 16) | (a)[(n)+1])
472    
473    /* Keep it positive */
474    #define MAX_PATTERN_SIZE (1 << 30)
475    
476  #else  #else
477  #error LINK_SIZE must be either 2, 3, or 4  #error LINK_SIZE must be either 2, 3, or 4
478  #endif  #endif
479    
480    #else
481    #error Unsupported compiling mode
482    #endif /* COMPILE_PCRE16 */
483    
484    #endif /* COMPILE_PCRE8 */
485    
486  /* Convenience macro defined in terms of the others */  /* Convenience macro defined in terms of the others */
487    
# Line 298  is automated on Unix systems via the "co Line 492  is automated on Unix systems via the "co
492  offsets changes. There are used for repeat counts and for other things such as  offsets changes. There are used for repeat counts and for other things such as
493  capturing parenthesis numbers in back references. */  capturing parenthesis numbers in back references. */
494    
495    #ifdef COMPILE_PCRE8
496    
497    #define IMM2_SIZE 2
498    
499  #define PUT2(a,n,d)   \  #define PUT2(a,n,d)   \
500    a[n] = (d) >> 8; \    a[n] = (d) >> 8; \
501    a[(n)+1] = (d) & 255    a[(n)+1] = (d) & 255
# Line 305  capturing parenthesis numbers in back re Line 503  capturing parenthesis numbers in back re
503  #define GET2(a,n) \  #define GET2(a,n) \
504    (((a)[n] << 8) | (a)[(n)+1])    (((a)[n] << 8) | (a)[(n)+1])
505    
506  #define PUT2INC(a,n,d)  PUT2(a,n,d), a += 2  #else /* COMPILE_PCRE8 */
507    
508    #ifdef COMPILE_PCRE16
509    
510    #define IMM2_SIZE 1
511    
512    #define PUT2(a,n,d)   \
513       a[n] = d
514    
515    #define GET2(a,n) \
516       a[n]
517    
518    #else
519    #error Unsupported compiling mode
520    #endif /* COMPILE_PCRE16 */
521    
522    #endif /* COMPILE_PCRE8 */
523    
524  /* When UTF-8 encoding is being used, a character is no longer just a single  #define PUT2INC(a,n,d)  PUT2(a,n,d), a += IMM2_SIZE
 byte. The macros for character handling generate simple sequences when used in  
 byte-mode, and more complicated ones for UTF-8 characters. */  
525    
526  #ifndef SUPPORT_UTF8  /* The maximum length of a MARK name is currently one data unit; it may be
527    changed in future to be a fixed number of bytes or to depend on LINK_SIZE. */
528    
529    #define MAX_MARK ((1 << (sizeof(pcre_uchar)*8)) - 1)
530    
531    /* When UTF encoding is being used, a character is no longer just a single
532    character. The macros for character handling generate simple sequences when
533    used in character-mode, and more complicated ones for UTF characters.
534    GETCHARLENTEST and other macros are not used when UTF is not supported,
535    so they are not defined. To make sure they can never even appear when
536    UTF support is omitted, we don't even define them. */
537    
538    #ifndef SUPPORT_UTF
539    
540    /* #define MAX_VALUE_FOR_SINGLE_CHAR */
541    /* #define HAS_EXTRALEN(c) */
542    /* #define GET_EXTRALEN(c) */
543    /* #define NOT_FIRSTCHAR(c) */
544  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
545  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
546  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
547  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
548  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
549  #define BACKCHAR(eptr)  /* #define GETCHARLENTEST(c, eptr, len) */
550    /* #define BACKCHAR(eptr) */
551    /* #define FORWARDCHAR(eptr) */
552    /* #define ACROSSCHAR(condition, eptr, action) */
553    
554    #else   /* SUPPORT_UTF */
555    
556    #ifdef COMPILE_PCRE8
557    
558    /* These macros were originally written in the form of loops that used data
559    from the tables whose names start with PRIV(utf8_table). They were rewritten by
560    a user so as not to use loops, because in some environments this gives a
561    significant performance advantage, and it seems never to do any harm. */
562    
563    /* Tells the biggest code point which can be encoded as a single character. */
564    
565    #define MAX_VALUE_FOR_SINGLE_CHAR 127
566    
567    /* Tests whether the code point needs extra characters to decode. */
568    
569    #define HAS_EXTRALEN(c) ((c) >= 0xc0)
570    
571    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
572    Otherwise it has an undefined behaviour. */
573    
574    #define GET_EXTRALEN(c) (PRIV(utf8_table4)[(c) & 0x3f])
575    
576    /* Returns TRUE, if the given character is not the first character
577    of a UTF sequence. */
578    
579  #else   /* SUPPORT_UTF8 */  #define NOT_FIRSTCHAR(c) (((c) & 0xc0) == 0x80)
580    
581    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
582    advancing the pointer. */
583    
584    #define GETUTF8(c, eptr) \
585        { \
586        if ((c & 0x20) == 0) \
587          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
588        else if ((c & 0x10) == 0) \
589          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
590        else if ((c & 0x08) == 0) \
591          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
592          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
593        else if ((c & 0x04) == 0) \
594          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
595              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
596              (eptr[4] & 0x3f); \
597        else \
598          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
599              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
600              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
601        }
602    
603  /* Get the next UTF-8 character, not advancing the pointer. This is called when  /* Get the next UTF-8 character, not advancing the pointer. This is called when
604  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
605    
606  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
607    c = *eptr; \    c = *eptr; \
608    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8(c, eptr);
     { \  
     int gcii; \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     for (gcii = 1; gcii <= gcaa; gcii++) \  
       { \  
       gcss -= 6; \  
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     }  
609    
610  /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the  /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
611  pointer. */  pointer. */
612    
613  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
614    c = *eptr; \    c = *eptr; \
615    if (utf8 && c >= 0xc0) \    if (utf && c >= 0xc0) GETUTF8(c, eptr);
616    
617    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
618    the pointer. */
619    
620    #define GETUTF8INC(c, eptr) \
621      { \      { \
622      int gcii; \      if ((c & 0x20) == 0) \
623      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
624      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
625      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        { \
626      for (gcii = 1; gcii <= gcaa; gcii++) \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
627          eptr += 2; \
628          } \
629        else if ((c & 0x08) == 0) \
630          { \
631          c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
632              ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
633          eptr += 3; \
634          } \
635        else if ((c & 0x04) == 0) \
636        { \        { \
637        gcss -= 6; \        c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
638        c |= (eptr[gcii] & 0x3f) << gcss; \            ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
639              (eptr[3] & 0x3f); \
640          eptr += 4; \
641          } \
642        else \
643          { \
644          c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
645              ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
646              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
647          eptr += 5; \
648        } \        } \
649      }      }
650    
# Line 363  know we are in UTF-8 mode. */ Line 653  know we are in UTF-8 mode. */
653    
654  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
655    c = *eptr++; \    c = *eptr++; \
656    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8INC(c, eptr);
     { \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     while (gcaa-- > 0) \  
       { \  
       gcss -= 6; \  
       c |= (*eptr++ & 0x3f) << gcss; \  
       } \  
     }  
657    
658  /* Get the next character, testing for UTF-8 mode, and advancing the pointer */  /* Get the next character, testing for UTF-8 mode, and advancing the pointer.
659    This is called when we don't know if we are in UTF-8 mode. */
660    
661  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
662    c = *eptr++; \    c = *eptr++; \
663    if (utf8 && c >= 0xc0) \    if (utf && c >= 0xc0) GETUTF8INC(c, eptr);
664    
665    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
666    advancing the pointer, incrementing the length. */
667    
668    #define GETUTF8LEN(c, eptr, len) \
669      { \      { \
670      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
671      int gcss = 6*gcaa; \        { \
672      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
673      while (gcaa-- > 0) \        len++; \
674          } \
675        else if ((c & 0x10)  == 0) \
676        { \        { \
677        gcss -= 6; \        c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
678        c |= (*eptr++ & 0x3f) << gcss; \        len += 2; \
679          } \
680        else if ((c & 0x08)  == 0) \
681          {\
682          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
683              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
684          len += 3; \
685          } \
686        else if ((c & 0x04)  == 0) \
687          { \
688          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
689              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
690              (eptr[4] & 0x3f); \
691          len += 4; \
692          } \
693        else \
694          {\
695          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
696              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
697              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
698          len += 5; \
699        } \        } \
700      }      }
701    
# Line 396  if there are extra bytes. This is called Line 704  if there are extra bytes. This is called
704    
705  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
706    c = *eptr; \    c = *eptr; \
707    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
708      { \  
709      int gcii; \  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
710      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  pointer, incrementing length if there are extra bytes. This is called when we
711      int gcss = 6*gcaa; \  do not know if we are in UTF-8 mode. */
712      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
713      for (gcii = 1; gcii <= gcaa; gcii++) \  #define GETCHARLENTEST(c, eptr, len) \
714        { \    c = *eptr; \
715        gcss -= 6; \    if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len);
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
716    
717  /* If the pointer is not at the start of a character, move it back until  /* If the pointer is not at the start of a character, move it back until
718  it is. Called only in UTF-8 mode. */  it is. This is called only in UTF-8 mode - we don't put a test within the macro
719    because almost all calls are already within a block of UTF-8 only code. */
720    
721    #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
722    
723    /* Same as above, just in the other direction. */
724    #define FORWARDCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr++
725    
726    /* Same as above, but it allows a fully customizable form. */
727    #define ACROSSCHAR(condition, eptr, action) \
728      while((condition) && ((eptr) & 0xc0) == 0x80) action
729    
730    #else /* COMPILE_PCRE8 */
731    
732    #ifdef COMPILE_PCRE16
733    
734    /* Tells the biggest code point which can be encoded as a single character. */
735    
736    #define MAX_VALUE_FOR_SINGLE_CHAR 65535
737    
738    /* Tests whether the code point needs extra characters to decode. */
739    
740  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define HAS_EXTRALEN(c) (((c) & 0xfc00) == 0xd800)
741    
742    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
743    Otherwise it has an undefined behaviour. */
744    
745    #define GET_EXTRALEN(c) 1
746    
747    /* Returns TRUE, if the given character is not the first character
748    of a UTF sequence. */
749    
750    #define NOT_FIRSTCHAR(c) (((c) & 0xfc00) == 0xdc00)
751    
752    /* Base macro to pick up the low surrogate of a UTF-16 character, not
753    advancing the pointer. */
754    
755    #define GETUTF16(c, eptr) \
756       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; }
757    
758    /* Get the next UTF-16 character, not advancing the pointer. This is called when
759    we know we are in UTF-16 mode. */
760    
761    #define GETCHAR(c, eptr) \
762      c = *eptr; \
763      if ((c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
764    
765    /* Get the next UTF-16 character, testing for UTF-16 mode, and not advancing the
766    pointer. */
767    
768    #define GETCHARTEST(c, eptr) \
769      c = *eptr; \
770      if (utf && (c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
771    
772    /* Base macro to pick up the low surrogate of a UTF-16 character, advancing
773    the pointer. */
774    
775    #define GETUTF16INC(c, eptr) \
776       { c = (((c & 0x3ff) << 10) | (*eptr++ & 0x3ff)) + 0x10000; }
777    
778    /* Get the next UTF-16 character, advancing the pointer. This is called when we
779    know we are in UTF-16 mode. */
780    
781    #define GETCHARINC(c, eptr) \
782      c = *eptr++; \
783      if ((c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
784    
785    /* Get the next character, testing for UTF-16 mode, and advancing the pointer.
786    This is called when we don't know if we are in UTF-16 mode. */
787    
788    #define GETCHARINCTEST(c, eptr) \
789      c = *eptr++; \
790      if (utf && (c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
791    
792    /* Base macro to pick up the low surrogate of a UTF-16 character, not
793    advancing the pointer, incrementing the length. */
794    
795    #define GETUTF16LEN(c, eptr, len) \
796       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; len++; }
797    
798    /* Get the next UTF-16 character, not advancing the pointer, incrementing
799    length if there is a low surrogate. This is called when we know we are in
800    UTF-16 mode. */
801    
802    #define GETCHARLEN(c, eptr, len) \
803      c = *eptr; \
804      if ((c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
805    
806    /* Get the next UTF-816character, testing for UTF-16 mode, not advancing the
807    pointer, incrementing length if there is a low surrogate. This is called when
808    we do not know if we are in UTF-16 mode. */
809    
810    #define GETCHARLENTEST(c, eptr, len) \
811      c = *eptr; \
812      if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
813    
814    /* If the pointer is not at the start of a character, move it back until
815    it is. This is called only in UTF-16 mode - we don't put a test within the
816    macro because almost all calls are already within a block of UTF-16 only
817    code. */
818    
819    #define BACKCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr--
820    
821    /* Same as above, just in the other direction. */
822    #define FORWARDCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr++
823    
824    /* Same as above, but it allows a fully customizable form. */
825    #define ACROSSCHAR(condition, eptr, action) \
826      if ((condition) && ((eptr) & 0xfc00) == 0xdc00) action
827    
828  #endif  #endif
829    
830    #endif /* COMPILE_PCRE8 */
831    
832    #endif  /* SUPPORT_UTF */
833    
834    
835  /* In case there is no definition of offsetof() provided - though any proper  /* In case there is no definition of offsetof() provided - though any proper
836  Standard C system should have one. */  Standard C system should have one. */
# Line 426  Standard C system should have one. */ Line 840  Standard C system should have one. */
840  #endif  #endif
841    
842    
843  /* These are the public options that can change during matching. */  /* Private flags containing information about the compiled regex. They used to
844    live at the top end of the options word, but that got almost full, so now they
845  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
846    the restrictions on partial matching have been lifted. It remains for backwards
847    compatibility. */
848    
849  /* Private options flags start at the most significant end of the four bytes.  #ifdef COMPILE_PCRE8
850  The public options defined in pcre.h start at the least significant end. Make  #define PCRE_MODE          0x0001  /* compiled in 8 bit mode */
851  sure they don't overlap! The bits are getting a bit scarce now -- when we run  #endif
852  out, there is a dummy word in the structure that could be used for the private  #ifdef COMPILE_PCRE16
853  bits. */  #define PCRE_MODE          0x0002  /* compiled in 16 bit mode */
854    #endif
855  #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  #define PCRE_FIRSTSET      0x0010  /* first_char is set */
856  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FCH_CASELESS  0x0020  /* caseless first char */
857  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x0040  /* req_byte is set */
858  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_RCH_CASELESS  0x0080  /* caseless requested char */
859  #define PCRE_JCHANGED      0x08000000  /* j option changes within regex */  #define PCRE_STARTLINE     0x0100  /* start after \n for multiline */
860    #define PCRE_NOPARTIAL     0x0200  /* can't use partial with this regex */
861    #define PCRE_JCHANGED      0x0400  /* j option used in regex */
862    #define PCRE_HASCRORLF     0x0800  /* explicit \r or \n in pattern */
863    #define PCRE_HASTHEN       0x1000  /* pattern contains (*THEN) */
864    
865  /* Options for the "extra" block produced by pcre_study(). */  /* Flags for the "extra" block produced by pcre_study(). */
866    
867  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
868    #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
869    
870  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
871  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
872    
873  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY)  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
874                               PCRE_NEWLINE_ANYCRLF)
875    
876  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
877    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
878     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
879     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
880     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
881       PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
882    
883  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
884    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
885     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
886       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
887    
888  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
889    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
890     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
891       PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
892       PCRE_NO_START_OPTIMIZE)
893    
894    #define PUBLIC_STUDY_OPTIONS \
895       (PCRE_STUDY_JIT_COMPILE|PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE| \
896        PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE)
897    
898  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  /* Magic number to provide a small check against being handed junk. */
   
 /* Magic number to provide a small check against being handed junk. Also used  
 to detect whether a pattern was compiled on a host of different endianness. */  
899    
900  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */
901    
902    /* This variable is used to detect a loaded regular expression
903    in different endianness. */
904    
905    #define REVERSED_MAGIC_NUMBER  0x45524350UL   /* 'ERCP' */
906    
907  /* Negative values for the firstchar and reqchar variables */  /* Negative values for the firstchar and reqchar variables */
908    
909  #define REQ_UNSET (-2)  #define REQ_UNSET (-2)
# Line 482  req_byte match. */ Line 914  req_byte match. */
914    
915  #define REQ_BYTE_MAX 1000  #define REQ_BYTE_MAX 1000
916    
917  /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
918  variable-length repeat, or a anything other than literal characters. */  environments where these macros are defined elsewhere. Unfortunately, there
919    is no way to do the same for the typedef. */
 #define REQ_CASELESS 0x0100    /* indicates caselessness */  
 #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  
   
 /* Miscellaneous definitions */  
920    
921  typedef int BOOL;  typedef int BOOL;
922    
923    #ifndef FALSE
924  #define FALSE   0  #define FALSE   0
925  #define TRUE    1  #define TRUE    1
926    #endif
927    
928    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
929    character constants like '*' because the compiler would emit their EBCDIC code,
930    which is different from their ASCII/UTF-8 code. Instead we define macros for
931    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
932    is enabled. When UTF-8 support is not enabled, the definitions use character
933    literals. Both character and string versions of each character are needed, and
934    there are some longer strings as well.
935    
936    This means that, on EBCDIC platforms, the PCRE library can handle either
937    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
938    would need different lookups depending on whether PCRE_UTF8 was set or not.
939    This would make it impossible to use characters in switch/case statements,
940    which would reduce performance. For a theoretical use (which nobody has asked
941    for) in a minority area (EBCDIC platforms), this is not sensible. Any
942    application that did need both could compile two versions of the library, using
943    macros to give the functions distinct names. */
944    
945    #ifndef SUPPORT_UTF
946    
947    /* UTF-8 support is not enabled; use the platform-dependent character literals
948    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
949    
950    #define CHAR_HT                     '\t'
951    #define CHAR_VT                     '\v'
952    #define CHAR_FF                     '\f'
953    #define CHAR_CR                     '\r'
954    #define CHAR_NL                     '\n'
955    #define CHAR_BS                     '\b'
956    #define CHAR_BEL                    '\a'
957    #ifdef EBCDIC
958    #define CHAR_ESC                    '\047'
959    #define CHAR_DEL                    '\007'
960    #else
961    #define CHAR_ESC                    '\033'
962    #define CHAR_DEL                    '\177'
963    #endif
964    
965    #define CHAR_SPACE                  ' '
966    #define CHAR_EXCLAMATION_MARK       '!'
967    #define CHAR_QUOTATION_MARK         '"'
968    #define CHAR_NUMBER_SIGN            '#'
969    #define CHAR_DOLLAR_SIGN            '$'
970    #define CHAR_PERCENT_SIGN           '%'
971    #define CHAR_AMPERSAND              '&'
972    #define CHAR_APOSTROPHE             '\''
973    #define CHAR_LEFT_PARENTHESIS       '('
974    #define CHAR_RIGHT_PARENTHESIS      ')'
975    #define CHAR_ASTERISK               '*'
976    #define CHAR_PLUS                   '+'
977    #define CHAR_COMMA                  ','
978    #define CHAR_MINUS                  '-'
979    #define CHAR_DOT                    '.'
980    #define CHAR_SLASH                  '/'
981    #define CHAR_0                      '0'
982    #define CHAR_1                      '1'
983    #define CHAR_2                      '2'
984    #define CHAR_3                      '3'
985    #define CHAR_4                      '4'
986    #define CHAR_5                      '5'
987    #define CHAR_6                      '6'
988    #define CHAR_7                      '7'
989    #define CHAR_8                      '8'
990    #define CHAR_9                      '9'
991    #define CHAR_COLON                  ':'
992    #define CHAR_SEMICOLON              ';'
993    #define CHAR_LESS_THAN_SIGN         '<'
994    #define CHAR_EQUALS_SIGN            '='
995    #define CHAR_GREATER_THAN_SIGN      '>'
996    #define CHAR_QUESTION_MARK          '?'
997    #define CHAR_COMMERCIAL_AT          '@'
998    #define CHAR_A                      'A'
999    #define CHAR_B                      'B'
1000    #define CHAR_C                      'C'
1001    #define CHAR_D                      'D'
1002    #define CHAR_E                      'E'
1003    #define CHAR_F                      'F'
1004    #define CHAR_G                      'G'
1005    #define CHAR_H                      'H'
1006    #define CHAR_I                      'I'
1007    #define CHAR_J                      'J'
1008    #define CHAR_K                      'K'
1009    #define CHAR_L                      'L'
1010    #define CHAR_M                      'M'
1011    #define CHAR_N                      'N'
1012    #define CHAR_O                      'O'
1013    #define CHAR_P                      'P'
1014    #define CHAR_Q                      'Q'
1015    #define CHAR_R                      'R'
1016    #define CHAR_S                      'S'
1017    #define CHAR_T                      'T'
1018    #define CHAR_U                      'U'
1019    #define CHAR_V                      'V'
1020    #define CHAR_W                      'W'
1021    #define CHAR_X                      'X'
1022    #define CHAR_Y                      'Y'
1023    #define CHAR_Z                      'Z'
1024    #define CHAR_LEFT_SQUARE_BRACKET    '['
1025    #define CHAR_BACKSLASH              '\\'
1026    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
1027    #define CHAR_CIRCUMFLEX_ACCENT      '^'
1028    #define CHAR_UNDERSCORE             '_'
1029    #define CHAR_GRAVE_ACCENT           '`'
1030    #define CHAR_a                      'a'
1031    #define CHAR_b                      'b'
1032    #define CHAR_c                      'c'
1033    #define CHAR_d                      'd'
1034    #define CHAR_e                      'e'
1035    #define CHAR_f                      'f'
1036    #define CHAR_g                      'g'
1037    #define CHAR_h                      'h'
1038    #define CHAR_i                      'i'
1039    #define CHAR_j                      'j'
1040    #define CHAR_k                      'k'
1041    #define CHAR_l                      'l'
1042    #define CHAR_m                      'm'
1043    #define CHAR_n                      'n'
1044    #define CHAR_o                      'o'
1045    #define CHAR_p                      'p'
1046    #define CHAR_q                      'q'
1047    #define CHAR_r                      'r'
1048    #define CHAR_s                      's'
1049    #define CHAR_t                      't'
1050    #define CHAR_u                      'u'
1051    #define CHAR_v                      'v'
1052    #define CHAR_w                      'w'
1053    #define CHAR_x                      'x'
1054    #define CHAR_y                      'y'
1055    #define CHAR_z                      'z'
1056    #define CHAR_LEFT_CURLY_BRACKET     '{'
1057    #define CHAR_VERTICAL_LINE          '|'
1058    #define CHAR_RIGHT_CURLY_BRACKET    '}'
1059    #define CHAR_TILDE                  '~'
1060    
1061    #define STR_HT                      "\t"
1062    #define STR_VT                      "\v"
1063    #define STR_FF                      "\f"
1064    #define STR_CR                      "\r"
1065    #define STR_NL                      "\n"
1066    #define STR_BS                      "\b"
1067    #define STR_BEL                     "\a"
1068    #ifdef EBCDIC
1069    #define STR_ESC                     "\047"
1070    #define STR_DEL                     "\007"
1071    #else
1072    #define STR_ESC                     "\033"
1073    #define STR_DEL                     "\177"
1074    #endif
1075    
1076    #define STR_SPACE                   " "
1077    #define STR_EXCLAMATION_MARK        "!"
1078    #define STR_QUOTATION_MARK          "\""
1079    #define STR_NUMBER_SIGN             "#"
1080    #define STR_DOLLAR_SIGN             "$"
1081    #define STR_PERCENT_SIGN            "%"
1082    #define STR_AMPERSAND               "&"
1083    #define STR_APOSTROPHE              "'"
1084    #define STR_LEFT_PARENTHESIS        "("
1085    #define STR_RIGHT_PARENTHESIS       ")"
1086    #define STR_ASTERISK                "*"
1087    #define STR_PLUS                    "+"
1088    #define STR_COMMA                   ","
1089    #define STR_MINUS                   "-"
1090    #define STR_DOT                     "."
1091    #define STR_SLASH                   "/"
1092    #define STR_0                       "0"
1093    #define STR_1                       "1"
1094    #define STR_2                       "2"
1095    #define STR_3                       "3"
1096    #define STR_4                       "4"
1097    #define STR_5                       "5"
1098    #define STR_6                       "6"
1099    #define STR_7                       "7"
1100    #define STR_8                       "8"
1101    #define STR_9                       "9"
1102    #define STR_COLON                   ":"
1103    #define STR_SEMICOLON               ";"
1104    #define STR_LESS_THAN_SIGN          "<"
1105    #define STR_EQUALS_SIGN             "="
1106    #define STR_GREATER_THAN_SIGN       ">"
1107    #define STR_QUESTION_MARK           "?"
1108    #define STR_COMMERCIAL_AT           "@"
1109    #define STR_A                       "A"
1110    #define STR_B                       "B"
1111    #define STR_C                       "C"
1112    #define STR_D                       "D"
1113    #define STR_E                       "E"
1114    #define STR_F                       "F"
1115    #define STR_G                       "G"
1116    #define STR_H                       "H"
1117    #define STR_I                       "I"
1118    #define STR_J                       "J"
1119    #define STR_K                       "K"
1120    #define STR_L                       "L"
1121    #define STR_M                       "M"
1122    #define STR_N                       "N"
1123    #define STR_O                       "O"
1124    #define STR_P                       "P"
1125    #define STR_Q                       "Q"
1126    #define STR_R                       "R"
1127    #define STR_S                       "S"
1128    #define STR_T                       "T"
1129    #define STR_U                       "U"
1130    #define STR_V                       "V"
1131    #define STR_W                       "W"
1132    #define STR_X                       "X"
1133    #define STR_Y                       "Y"
1134    #define STR_Z                       "Z"
1135    #define STR_LEFT_SQUARE_BRACKET     "["
1136    #define STR_BACKSLASH               "\\"
1137    #define STR_RIGHT_SQUARE_BRACKET    "]"
1138    #define STR_CIRCUMFLEX_ACCENT       "^"
1139    #define STR_UNDERSCORE              "_"
1140    #define STR_GRAVE_ACCENT            "`"
1141    #define STR_a                       "a"
1142    #define STR_b                       "b"
1143    #define STR_c                       "c"
1144    #define STR_d                       "d"
1145    #define STR_e                       "e"
1146    #define STR_f                       "f"
1147    #define STR_g                       "g"
1148    #define STR_h                       "h"
1149    #define STR_i                       "i"
1150    #define STR_j                       "j"
1151    #define STR_k                       "k"
1152    #define STR_l                       "l"
1153    #define STR_m                       "m"
1154    #define STR_n                       "n"
1155    #define STR_o                       "o"
1156    #define STR_p                       "p"
1157    #define STR_q                       "q"
1158    #define STR_r                       "r"
1159    #define STR_s                       "s"
1160    #define STR_t                       "t"
1161    #define STR_u                       "u"
1162    #define STR_v                       "v"
1163    #define STR_w                       "w"
1164    #define STR_x                       "x"
1165    #define STR_y                       "y"
1166    #define STR_z                       "z"
1167    #define STR_LEFT_CURLY_BRACKET      "{"
1168    #define STR_VERTICAL_LINE           "|"
1169    #define STR_RIGHT_CURLY_BRACKET     "}"
1170    #define STR_TILDE                   "~"
1171    
1172    #define STRING_ACCEPT0              "ACCEPT\0"
1173    #define STRING_COMMIT0              "COMMIT\0"
1174    #define STRING_F0                   "F\0"
1175    #define STRING_FAIL0                "FAIL\0"
1176    #define STRING_MARK0                "MARK\0"
1177    #define STRING_PRUNE0               "PRUNE\0"
1178    #define STRING_SKIP0                "SKIP\0"
1179    #define STRING_THEN                 "THEN"
1180    
1181    #define STRING_alpha0               "alpha\0"
1182    #define STRING_lower0               "lower\0"
1183    #define STRING_upper0               "upper\0"
1184    #define STRING_alnum0               "alnum\0"
1185    #define STRING_ascii0               "ascii\0"
1186    #define STRING_blank0               "blank\0"
1187    #define STRING_cntrl0               "cntrl\0"
1188    #define STRING_digit0               "digit\0"
1189    #define STRING_graph0               "graph\0"
1190    #define STRING_print0               "print\0"
1191    #define STRING_punct0               "punct\0"
1192    #define STRING_space0               "space\0"
1193    #define STRING_word0                "word\0"
1194    #define STRING_xdigit               "xdigit"
1195    
1196    #define STRING_DEFINE               "DEFINE"
1197    
1198    #define STRING_CR_RIGHTPAR             "CR)"
1199    #define STRING_LF_RIGHTPAR             "LF)"
1200    #define STRING_CRLF_RIGHTPAR           "CRLF)"
1201    #define STRING_ANY_RIGHTPAR            "ANY)"
1202    #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
1203    #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
1204    #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
1205    #ifdef COMPILE_PCRE8
1206    #define STRING_UTF_RIGHTPAR            "UTF8)"
1207    #endif
1208    #ifdef COMPILE_PCRE16
1209    #define STRING_UTF_RIGHTPAR            "UTF16)"
1210    #endif
1211    #define STRING_UCP_RIGHTPAR            "UCP)"
1212    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
1213    
1214    #else  /* SUPPORT_UTF */
1215    
1216    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
1217    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
1218    only. */
1219    
1220    #define CHAR_HT                     '\011'
1221    #define CHAR_VT                     '\013'
1222    #define CHAR_FF                     '\014'
1223    #define CHAR_CR                     '\015'
1224    #define CHAR_NL                     '\012'
1225    #define CHAR_BS                     '\010'
1226    #define CHAR_BEL                    '\007'
1227    #define CHAR_ESC                    '\033'
1228    #define CHAR_DEL                    '\177'
1229    
1230    #define CHAR_SPACE                  '\040'
1231    #define CHAR_EXCLAMATION_MARK       '\041'
1232    #define CHAR_QUOTATION_MARK         '\042'
1233    #define CHAR_NUMBER_SIGN            '\043'
1234    #define CHAR_DOLLAR_SIGN            '\044'
1235    #define CHAR_PERCENT_SIGN           '\045'
1236    #define CHAR_AMPERSAND              '\046'
1237    #define CHAR_APOSTROPHE             '\047'
1238    #define CHAR_LEFT_PARENTHESIS       '\050'
1239    #define CHAR_RIGHT_PARENTHESIS      '\051'
1240    #define CHAR_ASTERISK               '\052'
1241    #define CHAR_PLUS                   '\053'
1242    #define CHAR_COMMA                  '\054'
1243    #define CHAR_MINUS                  '\055'
1244    #define CHAR_DOT                    '\056'
1245    #define CHAR_SLASH                  '\057'
1246    #define CHAR_0                      '\060'
1247    #define CHAR_1                      '\061'
1248    #define CHAR_2                      '\062'
1249    #define CHAR_3                      '\063'
1250    #define CHAR_4                      '\064'
1251    #define CHAR_5                      '\065'
1252    #define CHAR_6                      '\066'
1253    #define CHAR_7                      '\067'
1254    #define CHAR_8                      '\070'
1255    #define CHAR_9                      '\071'
1256    #define CHAR_COLON                  '\072'
1257    #define CHAR_SEMICOLON              '\073'
1258    #define CHAR_LESS_THAN_SIGN         '\074'
1259    #define CHAR_EQUALS_SIGN            '\075'
1260    #define CHAR_GREATER_THAN_SIGN      '\076'
1261    #define CHAR_QUESTION_MARK          '\077'
1262    #define CHAR_COMMERCIAL_AT          '\100'
1263    #define CHAR_A                      '\101'
1264    #define CHAR_B                      '\102'
1265    #define CHAR_C                      '\103'
1266    #define CHAR_D                      '\104'
1267    #define CHAR_E                      '\105'
1268    #define CHAR_F                      '\106'
1269    #define CHAR_G                      '\107'
1270    #define CHAR_H                      '\110'
1271    #define CHAR_I                      '\111'
1272    #define CHAR_J                      '\112'
1273    #define CHAR_K                      '\113'
1274    #define CHAR_L                      '\114'
1275    #define CHAR_M                      '\115'
1276    #define CHAR_N                      '\116'
1277    #define CHAR_O                      '\117'
1278    #define CHAR_P                      '\120'
1279    #define CHAR_Q                      '\121'
1280    #define CHAR_R                      '\122'
1281    #define CHAR_S                      '\123'
1282    #define CHAR_T                      '\124'
1283    #define CHAR_U                      '\125'
1284    #define CHAR_V                      '\126'
1285    #define CHAR_W                      '\127'
1286    #define CHAR_X                      '\130'
1287    #define CHAR_Y                      '\131'
1288    #define CHAR_Z                      '\132'
1289    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
1290    #define CHAR_BACKSLASH              '\134'
1291    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
1292    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
1293    #define CHAR_UNDERSCORE             '\137'
1294    #define CHAR_GRAVE_ACCENT           '\140'
1295    #define CHAR_a                      '\141'
1296    #define CHAR_b                      '\142'
1297    #define CHAR_c                      '\143'
1298    #define CHAR_d                      '\144'
1299    #define CHAR_e                      '\145'
1300    #define CHAR_f                      '\146'
1301    #define CHAR_g                      '\147'
1302    #define CHAR_h                      '\150'
1303    #define CHAR_i                      '\151'
1304    #define CHAR_j                      '\152'
1305    #define CHAR_k                      '\153'
1306    #define CHAR_l                      '\154'
1307    #define CHAR_m                      '\155'
1308    #define CHAR_n                      '\156'
1309    #define CHAR_o                      '\157'
1310    #define CHAR_p                      '\160'
1311    #define CHAR_q                      '\161'
1312    #define CHAR_r                      '\162'
1313    #define CHAR_s                      '\163'
1314    #define CHAR_t                      '\164'
1315    #define CHAR_u                      '\165'
1316    #define CHAR_v                      '\166'
1317    #define CHAR_w                      '\167'
1318    #define CHAR_x                      '\170'
1319    #define CHAR_y                      '\171'
1320    #define CHAR_z                      '\172'
1321    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1322    #define CHAR_VERTICAL_LINE          '\174'
1323    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1324    #define CHAR_TILDE                  '\176'
1325    
1326    #define STR_HT                      "\011"
1327    #define STR_VT                      "\013"
1328    #define STR_FF                      "\014"
1329    #define STR_CR                      "\015"
1330    #define STR_NL                      "\012"
1331    #define STR_BS                      "\010"
1332    #define STR_BEL                     "\007"
1333    #define STR_ESC                     "\033"
1334    #define STR_DEL                     "\177"
1335    
1336    #define STR_SPACE                   "\040"
1337    #define STR_EXCLAMATION_MARK        "\041"
1338    #define STR_QUOTATION_MARK          "\042"
1339    #define STR_NUMBER_SIGN             "\043"
1340    #define STR_DOLLAR_SIGN             "\044"
1341    #define STR_PERCENT_SIGN            "\045"
1342    #define STR_AMPERSAND               "\046"
1343    #define STR_APOSTROPHE              "\047"
1344    #define STR_LEFT_PARENTHESIS        "\050"
1345    #define STR_RIGHT_PARENTHESIS       "\051"
1346    #define STR_ASTERISK                "\052"
1347    #define STR_PLUS                    "\053"
1348    #define STR_COMMA                   "\054"
1349    #define STR_MINUS                   "\055"
1350    #define STR_DOT                     "\056"
1351    #define STR_SLASH                   "\057"
1352    #define STR_0                       "\060"
1353    #define STR_1                       "\061"
1354    #define STR_2                       "\062"
1355    #define STR_3                       "\063"
1356    #define STR_4                       "\064"
1357    #define STR_5                       "\065"
1358    #define STR_6                       "\066"
1359    #define STR_7                       "\067"
1360    #define STR_8                       "\070"
1361    #define STR_9                       "\071"
1362    #define STR_COLON                   "\072"
1363    #define STR_SEMICOLON               "\073"
1364    #define STR_LESS_THAN_SIGN          "\074"
1365    #define STR_EQUALS_SIGN             "\075"
1366    #define STR_GREATER_THAN_SIGN       "\076"
1367    #define STR_QUESTION_MARK           "\077"
1368    #define STR_COMMERCIAL_AT           "\100"
1369    #define STR_A                       "\101"
1370    #define STR_B                       "\102"
1371    #define STR_C                       "\103"
1372    #define STR_D                       "\104"
1373    #define STR_E                       "\105"
1374    #define STR_F                       "\106"
1375    #define STR_G                       "\107"
1376    #define STR_H                       "\110"
1377    #define STR_I                       "\111"
1378    #define STR_J                       "\112"
1379    #define STR_K                       "\113"
1380    #define STR_L                       "\114"
1381    #define STR_M                       "\115"
1382    #define STR_N                       "\116"
1383    #define STR_O                       "\117"
1384    #define STR_P                       "\120"
1385    #define STR_Q                       "\121"
1386    #define STR_R                       "\122"
1387    #define STR_S                       "\123"
1388    #define STR_T                       "\124"
1389    #define STR_U                       "\125"
1390    #define STR_V                       "\126"
1391    #define STR_W                       "\127"
1392    #define STR_X                       "\130"
1393    #define STR_Y                       "\131"
1394    #define STR_Z                       "\132"
1395    #define STR_LEFT_SQUARE_BRACKET     "\133"
1396    #define STR_BACKSLASH               "\134"
1397    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1398    #define STR_CIRCUMFLEX_ACCENT       "\136"
1399    #define STR_UNDERSCORE              "\137"
1400    #define STR_GRAVE_ACCENT            "\140"
1401    #define STR_a                       "\141"
1402    #define STR_b                       "\142"
1403    #define STR_c                       "\143"
1404    #define STR_d                       "\144"
1405    #define STR_e                       "\145"
1406    #define STR_f                       "\146"
1407    #define STR_g                       "\147"
1408    #define STR_h                       "\150"
1409    #define STR_i                       "\151"
1410    #define STR_j                       "\152"
1411    #define STR_k                       "\153"
1412    #define STR_l                       "\154"
1413    #define STR_m                       "\155"
1414    #define STR_n                       "\156"
1415    #define STR_o                       "\157"
1416    #define STR_p                       "\160"
1417    #define STR_q                       "\161"
1418    #define STR_r                       "\162"
1419    #define STR_s                       "\163"
1420    #define STR_t                       "\164"
1421    #define STR_u                       "\165"
1422    #define STR_v                       "\166"
1423    #define STR_w                       "\167"
1424    #define STR_x                       "\170"
1425    #define STR_y                       "\171"
1426    #define STR_z                       "\172"
1427    #define STR_LEFT_CURLY_BRACKET      "\173"
1428    #define STR_VERTICAL_LINE           "\174"
1429    #define STR_RIGHT_CURLY_BRACKET     "\175"
1430    #define STR_TILDE                   "\176"
1431    
1432    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1433    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1434    #define STRING_F0                   STR_F "\0"
1435    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1436    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1437    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1438    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1439    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1440    
1441    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1442    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1443    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1444    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1445    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1446    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1447    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1448    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1449    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1450    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1451    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1452    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1453    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1454    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1455    
1456    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1457    
1458    #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1459    #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1460    #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1461    #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1462    #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1463    #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
1464    #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
1465    #ifdef COMPILE_PCRE8
1466    #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1467    #endif
1468    #ifdef COMPILE_PCRE16
1469    #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS
1470    #endif
1471    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1472    #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
1473    
1474    #endif  /* SUPPORT_UTF */
1475    
1476  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1477    
1478  #ifndef ESC_e  #ifndef ESC_e
1479  #define ESC_e 27  #define ESC_e CHAR_ESC
1480  #endif  #endif
1481    
1482  #ifndef ESC_f  #ifndef ESC_f
1483  #define ESC_f '\f'  #define ESC_f CHAR_FF
1484  #endif  #endif
1485    
1486  #ifndef ESC_n  #ifndef ESC_n
1487  #define ESC_n '\n'  #define ESC_n CHAR_NL
1488  #endif  #endif
1489    
1490  #ifndef ESC_r  #ifndef ESC_r
1491  #define ESC_r '\r'  #define ESC_r CHAR_CR
1492  #endif  #endif
1493    
1494  /* We can't officially use ESC_t because it is a POSIX reserved identifier  /* We can't officially use ESC_t because it is a POSIX reserved identifier
1495  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1496    
1497  #ifndef ESC_tee  #ifndef ESC_tee
1498  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1499  #endif  #endif
1500    
1501  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
1502    
1503  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1504  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1505  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1506  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1507  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1508    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1509    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1510    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1511    #define PT_WORD       8    /* Word - L plus N plus underscore */
1512    
1513  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
1514  contain UTF-8 characters with values greater than 255. */  contain characters with values greater than 255. */
1515    
1516  #define XCL_NOT    0x01    /* Flag: this is a negative class */  #define XCL_NOT    0x01    /* Flag: this is a negative class */
1517  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */
# Line 543  contain UTF-8 characters with values gre Line 1525  contain UTF-8 characters with values gre
1525  /* These are escaped items that aren't just an encoding of a particular data  /* These are escaped items that aren't just an encoding of a particular data
1526  value such as \n. They must have non-zero values, as check_escape() returns  value such as \n. They must have non-zero values, as check_escape() returns
1527  their negation. Also, they must appear in the same order as in the opcode  their negation. Also, they must appear in the same order as in the opcode
1528  definitions below, up to ESC_z. There's a dummy for OP_ANY because it  definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it
1529  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1530  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In
1531  There are two tests in the code for an escape greater than ESC_b and less than  non-DOTALL mode, "." behaves like \N.
1532  ESC_Z to detect the types that may be repeated. These are the types that  
1533  consume characters. If any new escapes are put in between that don't consume a  The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1534  character, that code will have to change. */  when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1535    They must be contiguous, and remain in order so that the replacements can be
1536  enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,  looked up from a table.
1537         ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_X, ESC_Z, ESC_z,  
1538         ESC_E, ESC_Q, ESC_k, ESC_REF };  The final escape must be ESC_REF as subsequent values are used for
1539    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1540    greater than ESC_b and less than ESC_Z to detect the types that may be
1541    repeated. These are the types that consume characters. If any new escapes are
1542    put in between that don't consume a character, that code will have to change.
1543    */
1544    
1545    enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
1546           ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1547           ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1548           ESC_E, ESC_Q, ESC_g, ESC_k,
1549           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1550           ESC_REF };
1551    
1552  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to  
1553  OP_EOD must correspond in order to the list of escapes immediately above.  OP_EOD must correspond in order to the list of escapes immediately above.
1554    
 To keep stored, compiled patterns compatible, new opcodes should be added  
 immediately before OP_BRA, where (since release 7.0) a gap is left for this  
 purpose.  
   
1555  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1556  that follow must also be updated to match. There is also a table called  that follow must also be updated to match. There are also tables called
1557  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1558    
1559  enum {  enum {
1560    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 575  enum { Line 1563  enum {
1563    
1564    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
1565    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
1566    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
1567    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
1568    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
1569    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
1570    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
1571    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
1572    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
1573    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
1574    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
1575    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */  
1576    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ANY,            /* 12 Match any character except newline */
1577    OP_PROP,           /* 14 \p (Unicode property) */    OP_ALLANY,         /* 13 Match any character */
1578    OP_ANYNL,          /* 15 \R (any newline sequence) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1579    OP_EXTUNI,         /* 16 \X (extended Unicode sequence */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1580    OP_EODN,           /* 17 End of data or \n at end of data: \Z. */    OP_PROP,           /* 16 \p (Unicode property) */
1581    OP_EOD,            /* 18 End of data: \z */    OP_ANYNL,          /* 17 \R (any newline sequence) */
1582      OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1583    OP_OPT,            /* 19 Set runtime options */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1584    OP_CIRC,           /* 20 Start of line - varies with multiline switch */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1585    OP_DOLL,           /* 21 End of line - varies with multiline switch */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1586    OP_CHAR,           /* 22 Match one character, casefully */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1587    OP_CHARNC,         /* 23 Match one character, caselessly */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1588    OP_NOT,            /* 24 Match one character, not the following one */    OP_EOD,            /* 24 End of data: \z */
1589    
1590    OP_STAR,           /* 25 The maximizing and minimizing versions of */    OP_CIRC,           /* 25 Start of line - not multiline */
1591    OP_MINSTAR,        /* 26 these six opcodes must come in pairs, with */    OP_CIRCM,          /* 26 Start of line - multiline */
1592    OP_PLUS,           /* 27 the minimizing one second. */    OP_DOLL,           /* 27 End of line - not multiline */
1593    OP_MINPLUS,        /* 28 This first set applies to single characters.*/    OP_DOLLM,          /* 28 End of line - multiline */
1594    OP_QUERY,          /* 29 */    OP_CHAR,           /* 29 Match one character, casefully */
1595    OP_MINQUERY,       /* 30 */    OP_CHARI,          /* 30 Match one character, caselessly */
1596      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1597    OP_UPTO,           /* 31 From 0 to n matches */    OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1598    OP_MINUPTO,        /* 32 */  
1599    OP_EXACT,          /* 33 Exactly n matches */    /* The following sets of 13 opcodes must always be kept in step because
1600      the offset from the first one is used to generate the others. */
1601    OP_POSSTAR,        /* 34 Possessified star */  
1602    OP_POSPLUS,        /* 35 Possessified plus */    /**** Single characters, caseful, must precede the caseless ones ****/
1603    OP_POSQUERY,       /* 36 Posesssified query */  
1604    OP_POSUPTO,        /* 37 Possessified upto */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1605      OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1606    OP_NOTSTAR,        /* 38 The maximizing and minimizing versions of */    OP_PLUS,           /* 35 the minimizing one second. */
1607    OP_NOTMINSTAR,     /* 39 these six opcodes must come in pairs, with */    OP_MINPLUS,        /* 36 */
1608    OP_NOTPLUS,        /* 40 the minimizing one second. They must be in */    OP_QUERY,          /* 37 */
1609    OP_NOTMINPLUS,     /* 41 exactly the same order as those above. */    OP_MINQUERY,       /* 38 */
1610    OP_NOTQUERY,       /* 42 This set applies to "not" single characters. */  
1611    OP_NOTMINQUERY,    /* 43 */    OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1612      OP_MINUPTO,        /* 40 */
1613    OP_NOTUPTO,        /* 44 From 0 to n matches */    OP_EXACT,          /* 41 Exactly n matches */
1614    OP_NOTMINUPTO,     /* 45 */  
1615    OP_NOTEXACT,       /* 46 Exactly n matches */    OP_POSSTAR,        /* 42 Possessified star, caseful */
1616      OP_POSPLUS,        /* 43 Possessified plus, caseful */
1617    OP_NOTPOSSTAR,     /* 47 Possessified versions */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
1618    OP_NOTPOSPLUS,     /* 48 */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
1619    OP_NOTPOSQUERY,    /* 49 */  
1620    OP_NOTPOSUPTO,     /* 50 */    /**** Single characters, caseless, must follow the caseful ones */
1621    
1622    OP_TYPESTAR,       /* 51 The maximizing and minimizing versions of */    OP_STARI,          /* 46 */
1623    OP_TYPEMINSTAR,    /* 52 these six opcodes must come in pairs, with */    OP_MINSTARI,       /* 47 */
1624    OP_TYPEPLUS,       /* 53 the minimizing one second. These codes must */    OP_PLUSI,          /* 48 */
1625    OP_TYPEMINPLUS,    /* 54 be in exactly the same order as those above. */    OP_MINPLUSI,       /* 49 */
1626    OP_TYPEQUERY,      /* 55 This set applies to character types such as \d */    OP_QUERYI,         /* 50 */
1627    OP_TYPEMINQUERY,   /* 56 */    OP_MINQUERYI,      /* 51 */
1628    
1629    OP_TYPEUPTO,       /* 57 From 0 to n matches */    OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1630    OP_TYPEMINUPTO,    /* 58 */    OP_MINUPTOI,       /* 53 */
1631    OP_TYPEEXACT,      /* 59 Exactly n matches */    OP_EXACTI,         /* 54 */
1632    
1633    OP_TYPEPOSSTAR,    /* 60 Possessified versions */    OP_POSSTARI,       /* 55 Possessified star, caseless */
1634    OP_TYPEPOSPLUS,    /* 61 */    OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1635    OP_TYPEPOSQUERY,   /* 62 */    OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1636    OP_TYPEPOSUPTO,    /* 63 */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1637    
1638    OP_CRSTAR,         /* 64 The maximizing and minimizing versions of */    /**** The negated ones must follow the non-negated ones, and match them ****/
1639    OP_CRMINSTAR,      /* 65 all these opcodes must come in pairs, with */    /**** Negated single character, caseful; must precede the caseless ones ****/
1640    OP_CRPLUS,         /* 66 the minimizing one second. These codes must */  
1641    OP_CRMINPLUS,      /* 67 be in exactly the same order as those above. */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1642    OP_CRQUERY,        /* 68 These are for character classes and back refs */    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1643    OP_CRMINQUERY,     /* 69 */    OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1644    OP_CRRANGE,        /* 70 These are different to the three sets above. */    OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1645    OP_CRMINRANGE,     /* 71 */    OP_NOTQUERY,       /* 63 */
1646      OP_NOTMINQUERY,    /* 64 */
1647    OP_CLASS,          /* 72 Match a character class, chars < 256 only */  
1648    OP_NCLASS,         /* 73 Same, but the bitmap was created from a negative    OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1649                             class - the difference is relevant only when a UTF-8    OP_NOTMINUPTO,     /* 66 */
1650                             character > 255 is encountered. */    OP_NOTEXACT,       /* 67 Exactly n matches */
1651    
1652    OP_XCLASS,         /* 74 Extended class for handling UTF-8 chars within the    OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1653                             class. This does both positive and negative. */    OP_NOTPOSPLUS,     /* 69 */
1654      OP_NOTPOSQUERY,    /* 70 */
1655    OP_REF,            /* 75 Match a back reference */    OP_NOTPOSUPTO,     /* 71 */
1656    OP_RECURSE,        /* 76 Match a numbered subpattern (possibly recursive) */  
1657    OP_CALLOUT,        /* 77 Call out to external function if provided */    /**** Negated single character, caseless; must follow the caseful ones ****/
1658    
1659    OP_ALT,            /* 78 Start of alternation */    OP_NOTSTARI,       /* 72 */
1660    OP_KET,            /* 79 End of group that doesn't have an unbounded repeat */    OP_NOTMINSTARI,    /* 73 */
1661    OP_KETRMAX,        /* 80 These two must remain together and in this */    OP_NOTPLUSI,       /* 74 */
1662    OP_KETRMIN,        /* 81 order. They are for groups the repeat for ever. */    OP_NOTMINPLUSI,    /* 75 */
1663      OP_NOTQUERYI,      /* 76 */
1664    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    OP_NOTMINQUERYI,   /* 77 */
1665    
1666    OP_ASSERT,         /* 82 Positive lookahead */    OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1667    OP_ASSERT_NOT,     /* 83 Negative lookahead */    OP_NOTMINUPTOI,    /* 79 */
1668    OP_ASSERTBACK,     /* 84 Positive lookbehind */    OP_NOTEXACTI,      /* 80 Exactly n matches */
1669    OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */  
1670    OP_REVERSE,        /* 86 Move pointer back - used in lookbehind assertions */    OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1671      OP_NOTPOSPLUSI,    /* 82 */
1672    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    OP_NOTPOSQUERYI,   /* 83 */
1673    as there's a test for >= ONCE for a subpattern that isn't an assertion. */    OP_NOTPOSUPTOI,    /* 84 */
1674    
1675    OP_ONCE,           /* 87 Atomic group */    /**** Character types ****/
1676    OP_BRA,            /* 88 Start of non-capturing bracket */  
1677    OP_CBRA,           /* 89 Start of capturing bracket */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1678    OP_COND,           /* 90 Conditional group */    OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1679      OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1680      OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
1681      OP_TYPEQUERY,      /* 89 */
1682      OP_TYPEMINQUERY,   /* 90 */
1683    
1684      OP_TYPEUPTO,       /* 91 From 0 to n matches */
1685      OP_TYPEMINUPTO,    /* 92 */
1686      OP_TYPEEXACT,      /* 93 Exactly n matches */
1687    
1688      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1689      OP_TYPEPOSPLUS,    /* 95 */
1690      OP_TYPEPOSQUERY,   /* 96 */
1691      OP_TYPEPOSUPTO,    /* 97 */
1692    
1693      /* These are used for character classes and back references; only the
1694      first six are the same as the sets above. */
1695    
1696      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1697      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1698      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1699      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1700      OP_CRQUERY,        /* 102 */
1701      OP_CRMINQUERY,     /* 103 */
1702    
1703      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1704      OP_CRMINRANGE,     /* 105 */
1705    
1706      /* End of quantifier opcodes */
1707    
1708      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1709      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1710                                  class - the difference is relevant only when a
1711                                  character > 255 is encountered. */
1712      OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the
1713                                  class. This does both positive and negative. */
1714      OP_REF,            /* 109 Match a back reference, casefully */
1715      OP_REFI,           /* 110 Match a back reference, caselessly */
1716      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1717      OP_CALLOUT,        /* 112 Call out to external function if provided */
1718    
1719      OP_ALT,            /* 113 Start of alternation */
1720      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1721      OP_KETRMAX,        /* 115 These two must remain together and in this */
1722      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1723      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1724    
1725      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1726      asserts must remain in order. */
1727    
1728      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1729      OP_ASSERT,         /* 119 Positive lookahead */
1730      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1731      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1732      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1733    
1734      /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
1735      after the assertions, with ONCE first, as there's a test for >= ONCE for a
1736      subpattern that isn't an assertion. The POS versions must immediately follow
1737      the non-POS versions in each case. */
1738    
1739      OP_ONCE,           /* 123 Atomic group, contains captures */
1740      OP_ONCE_NC,        /* 124 Atomic group containing no captures */
1741      OP_BRA,            /* 125 Start of non-capturing bracket */
1742      OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
1743      OP_CBRA,           /* 127 Start of capturing bracket */
1744      OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
1745      OP_COND,           /* 129 Conditional group */
1746    
1747    /* These three must follow the previous three, in the same order. There's a    /* These five must follow the previous five, in the same order. There's a
1748    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1749    
1750    OP_SBRA,           /* 91 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
1751    OP_SCBRA,          /* 92 Start of capturing bracket, check empty */    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
1752    OP_SCOND,          /* 93 Conditional group, check empty */    OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
1753      OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
1754    OP_CREF,           /* 94 Used to hold a capture number as condition */    OP_SCOND,          /* 134 Conditional group, check empty */
1755    OP_RREF,           /* 95 Used to hold a recursion number as condition */  
1756    OP_DEF,            /* 96 The DEFINE condition */    /* The next two pairs must (respectively) be kept together. */
1757    
1758      OP_CREF,           /* 135 Used to hold a capture number as condition */
1759      OP_NCREF,          /* 136 Same, but generated by a name reference*/
1760      OP_RREF,           /* 137 Used to hold a recursion number as condition */
1761      OP_NRREF,          /* 138 Same, but generated by a name reference*/
1762      OP_DEF,            /* 139 The DEFINE condition */
1763    
1764      OP_BRAZERO,        /* 140 These two must remain together and in this */
1765      OP_BRAMINZERO,     /* 141 order. */
1766      OP_BRAPOSZERO,     /* 142 */
1767    
1768      /* These are backtracking control verbs */
1769    
1770      OP_MARK,           /* 143 always has an argument */
1771      OP_PRUNE,          /* 144 */
1772      OP_PRUNE_ARG,      /* 145 same, but with argument */
1773      OP_SKIP,           /* 146 */
1774      OP_SKIP_ARG,       /* 147 same, but with argument */
1775      OP_THEN,           /* 148 */
1776      OP_THEN_ARG,       /* 149 same, but with argument */
1777      OP_COMMIT,         /* 150 */
1778    
1779      /* These are forced failure and success verbs */
1780    
1781      OP_FAIL,           /* 151 */
1782      OP_ACCEPT,         /* 152 */
1783      OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
1784      OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
1785    
1786      /* This is used to skip a subpattern with a {0} quantifier */
1787    
1788      OP_SKIPZERO,       /* 155 */
1789    
1790      /* This is not an opcode, but is used to check that tables indexed by opcode
1791      are the correct length, in order to catch updating errors - there have been
1792      some in the past. */
1793    
1794    OP_BRAZERO,        /* 97 These two must remain together and in this */    OP_TABLE_LENGTH
   OP_BRAMINZERO      /* 98 order. */  
1795  };  };
1796    
1797    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1798    definitions that follow must also be updated to match. There are also tables
1799    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1800    
1801    
1802  /* This macro defines textual names for all the opcodes. These are used only  /* This macro defines textual names for all the opcodes. These are used only
1803  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging, and some of them are only partial names. The macro is referenced
1804    only in pcre_printint.c, which fills out the full names in many cases (and in
1805    some cases doesn't actually use these names at all). */
1806    
1807  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1808    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1809    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1810    "notprop", "prop", "anynl", "extuni",                           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1811    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
1812    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1813    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1814      "{", "{", "{",                                                  \
1815    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1816    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1817      "{", "{", "{",                                                  \
1818      "*+","++", "?+", "{",                                           \
1819      "*", "*?", "+", "+?", "?", "??",                                \
1820      "{", "{", "{",                                                  \
1821      "*+","++", "?+", "{",                                           \
1822      "*", "*?", "+", "+?", "?", "??",                                \
1823      "{", "{", "{",                                                  \
1824    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1825    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1826    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1827    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1828    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1829    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1830    "AssertB", "AssertB not", "Reverse",                            \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1831    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1832    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Once", "Once_NC",                                              \
1833      "Bra", "BraPos", "CBra", "CBraPos",                             \
1834      "Cond",                                                         \
1835      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1836      "SCond",                                                        \
1837      "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1838      "Brazero", "Braminzero", "Braposzero",                          \
1839      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1840      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1841      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1842      "Close", "Skip zero"
1843    
1844    
1845  /* This macro defines the length of fixed length operations in the compiled  /* This macro defines the length of fixed length operations in the compiled
# Line 739  in UTF-8 mode. The code that uses this t Line 1853  in UTF-8 mode. The code that uses this t
1853    
1854  #define OP_LENGTHS \  #define OP_LENGTHS \
1855    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1856    1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1857    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1858    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1859    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    3, 3,                          /* \P, \p                                 */ \
1860      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1861      1,                             /* \X                                     */ \
1862      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1863    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1864    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1865    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1866    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1867    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1868    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1869    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto, minupto             ** mode      */ \
1870      2+IMM2_SIZE,                   /* exact                                  */ \
1871      2, 2, 2, 2+IMM2_SIZE,          /* *+, ++, ?+, upto+                      */ \
1872      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1873      2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto I, minupto I                      */ \
1874      2+IMM2_SIZE,                   /* exact I                                */ \
1875      2, 2, 2, 2+IMM2_SIZE,          /* *+I, ++I, ?+I, upto+I                  */ \
1876    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1877    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1878    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto, minupto                      */ \
1879    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2+IMM2_SIZE,                   /* NOT exact                              */ \
1880      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *, +, ?, upto           */ \
1881      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1882      2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto I, minupto I                  */ \
1883      2+IMM2_SIZE,                   /* NOT exact I                            */ \
1884      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *I, +I, ?I, upto I      */ \
1885    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1886    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1887    4, 4, 4,                       /* Type upto, minupto, exact              */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* Type upto, minupto                     */ \
1888    2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \    2+IMM2_SIZE,                   /* Type exact                             */ \
1889      2, 2, 2, 2+IMM2_SIZE,          /* Possessive *+, ++, ?+, upto+           */ \
1890    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
1891    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
1892    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
1893   33,                             /* CLASS                                  */ \    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
1894   33,                             /* NCLASS                                 */ \    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
1895    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1896    3,                             /* REF                                    */ \    1+IMM2_SIZE,                   /* REF                                    */ \
1897      1+IMM2_SIZE,                   /* REFI                                   */ \
1898    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1899    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1900    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
1901    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
1902    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
1903    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
1904      1+LINK_SIZE,                   /* KetRpos                                */ \
1905      1+LINK_SIZE,                   /* Reverse                                */ \
1906    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
1907    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
1908    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1909    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
   1+LINK_SIZE,                   /* Reverse                                */ \  
1910    1+LINK_SIZE,                   /* ONCE                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1911      1+LINK_SIZE,                   /* ONCE_NC                                */ \
1912    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
1913    3+LINK_SIZE,                   /* CBRA                                   */ \    1+LINK_SIZE,                   /* BRAPOS                                 */ \
1914      1+LINK_SIZE+IMM2_SIZE,         /* CBRA                                   */ \
1915      1+LINK_SIZE+IMM2_SIZE,         /* CBRAPOS                                */ \
1916    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1917    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1918    3+LINK_SIZE,                   /* SCBRA                                  */ \    1+LINK_SIZE,                   /* SBRAPOS                                */ \
1919      1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
1920      1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
1921    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1922    3,                             /* CREF                                   */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \
1923    3,                             /* RREF                                   */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \
1924    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1925    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
1926      3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1927      1, 3,                          /* SKIP, SKIP_ARG                         */ \
1928      1, 3,                          /* THEN, THEN_ARG                         */ \
1929      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
1930      1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
1931    
1932    /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1933  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  condition. */
1934    
1935  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1936    
1937  /* Error code numbers. They are given names so that they can more easily be  /* Compile time error code numbers. They are given names so that they can more
1938  tracked. */  easily be tracked. When a new number is added, the table called eint in
1939    pcreposix.c must be updated. */
1940    
1941  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1942         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
1943         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
1944         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1945         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1946         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1947           ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
1948           ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERRCOUNT };
1949    
1950    /* JIT compiling modes. The function list is indexed by them. */
1951    enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
1952           JIT_NUMBER_OF_COMPILE_MODES };
1953    
1954  /* The real format of the start of the pcre block; the index of names and the  /* The real format of the start of the pcre block; the index of names and the
1955  code vector run on as long as necessary after the end. We store an explicit  code vector run on as long as necessary after the end. We store an explicit
# Line 817  Because people can now save and re-use c Line 1965  Because people can now save and re-use c
1965  structure should be made at the end, and something earlier (e.g. a new  structure should be made at the end, and something earlier (e.g. a new
1966  flag in the options or one of the dummy fields) should indicate that the new  flag in the options or one of the dummy fields) should indicate that the new
1967  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1968  NOTE NOTE NOTE:  NOTE NOTE NOTE
1969  */  */
1970    
1971  typedef struct real_pcre {  #ifdef COMPILE_PCRE8
1972    #define REAL_PCRE real_pcre
1973    #else
1974    #define REAL_PCRE real_pcre16
1975    #endif
1976    
1977    typedef struct REAL_PCRE {
1978    pcre_uint32 magic_number;    pcre_uint32 magic_number;
1979    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1980    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
1981    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
1982      pcre_uint16 max_lookbehind;     /* Longest lookbehind (characters) */
1983    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;        /* Highest numbered group */
1984    pcre_uint16 top_backref;    pcre_uint16 top_backref;        /* Highest numbered back reference */
1985    pcre_uint16 first_byte;    pcre_uint16 first_char;         /* Starting character */
1986    pcre_uint16 req_byte;    pcre_uint16 req_char;           /* This character must be seen */
1987    pcre_uint16 name_table_offset;  /* Offset to name table that follows */    pcre_uint16 name_table_offset;  /* Offset to name table that follows */
1988    pcre_uint16 name_entry_size;    /* Size of any name items */    pcre_uint16 name_entry_size;    /* Size of any name items */
1989    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
1990    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
1991      const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
1992    const unsigned char *tables;    /* Pointer to tables or NULL for std */    const pcre_uint8 *nullpad;      /* NULL padding */
1993    const unsigned char *nullpad;   /* NULL padding */  } REAL_PCRE;
 } real_pcre;  
1994    
1995  /* The format of the block used to store data from pcre_study(). The same  /* The format of the block used to store data from pcre_study(). The same
1996  remark (see NOTE above) about extending this structure applies. */  remark (see NOTE above) about extending this structure applies. */
1997    
1998  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1999    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
2000    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
2001    uschar start_bits[32];    pcre_uint8 start_bits[32];      /* Starting char bits */
2002      pcre_uint32 minlength;          /* Minimum subject length */
2003  } pcre_study_data;  } pcre_study_data;
2004    
2005    /* Structure for building a chain of open capturing subpatterns during
2006    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
2007    encountered. This is also used to identify subpatterns that contain recursive
2008    back references to themselves, so that they can be made atomic. */
2009    
2010    typedef struct open_capitem {
2011      struct open_capitem *next;    /* Chain link */
2012      pcre_uint16 number;           /* Capture number */
2013      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
2014    } open_capitem;
2015    
2016  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
2017  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
2018    
2019  typedef struct compile_data {  typedef struct compile_data {
2020    const uschar *lcc;            /* Points to lower casing table */    const pcre_uint8 *lcc;            /* Points to lower casing table */
2021    const uschar *fcc;            /* Points to case-flipping table */    const pcre_uint8 *fcc;            /* Points to case-flipping table */
2022    const uschar *cbits;          /* Points to character type table */    const pcre_uint8 *cbits;          /* Points to character type table */
2023    const uschar *ctypes;         /* Points to table of type maps */    const pcre_uint8 *ctypes;         /* Points to table of type maps */
2024    const uschar *start_workspace;/* The start of working space */    const pcre_uchar *start_workspace;/* The start of working space */
2025    const uschar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
2026    const uschar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
2027    const uschar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
2028    uschar *hwm;                  /* High watermark of workspace */    open_capitem *open_caps;          /* Chain of open capture items */
2029    uschar *name_table;           /* The name/number table */    pcre_uchar *hwm;                  /* High watermark of workspace */
2030    int  names_found;             /* Number of entries so far */    pcre_uchar *name_table;           /* The name/number table */
2031    int  name_entry_size;         /* Size of each entry */    int  names_found;                 /* Number of entries so far */
2032    int  bracount;                /* Count of capturing parens */    int  name_entry_size;             /* Size of each entry */
2033    int  top_backref;             /* Maximum back reference */    int  workspace_size;              /* Size of workspace */
2034    unsigned int backref_map;     /* Bitmap of low back refs */    int  bracount;                    /* Count of capturing parens as we compile */
2035    int  external_options;        /* External (initial) options */    int  final_bracount;              /* Saved value after first pass */
2036    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  max_lookbehind;              /* Maximum lookbehind (characters) */
2037    BOOL nopartial;               /* Set TRUE if partial won't work */    int  top_backref;                 /* Maximum back reference */
2038    int  nltype;                  /* Newline type */    unsigned int backref_map;         /* Bitmap of low back refs */
2039    int  nllen;                   /* Newline string length */    int  assert_depth;                /* Depth of nested assertions */
2040    uschar nl[4];                 /* Newline string when fixed length */    int  external_options;            /* External (initial) options */
2041      int  external_flags;              /* External flag bits to be set */
2042      int  req_varyopt;                 /* "After variable item" flag for reqbyte */
2043      BOOL had_accept;                  /* (*ACCEPT) encountered */
2044      BOOL check_lookbehind;            /* Lookbehinds need later checking */
2045      int  nltype;                      /* Newline type */
2046      int  nllen;                       /* Newline string length */
2047      pcre_uchar nl[4];                 /* Newline string when fixed length */
2048  } compile_data;  } compile_data;
2049    
2050  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
2051  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
2052    
2053  typedef struct branch_chain {  typedef struct branch_chain {
2054    struct branch_chain *outer;    struct branch_chain *outer;
2055    uschar *current;    pcre_uchar *current_branch;
2056  } branch_chain;  } branch_chain;
2057    
2058  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
2059  call within the pattern. */  call within the pattern; used by pcre_exec(). */
2060    
2061  typedef struct recursion_info {  typedef struct recursion_info {
2062    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
2063    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
2064    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
2065    USPTR save_start;             /* Old value of md->start_match */    int saved_max;                  /* Number of saved offsets */
2066    int *offset_save;             /* Pointer to start of saved offsets */    PCRE_PUCHAR subject_position;   /* Position at start of recursion */
   int saved_max;                /* Number of saved offsets */  
2067  } recursion_info;  } recursion_info;
2068    
2069  /* When compiling in a mode that doesn't use recursive calls to match(),  /* A similar structure for pcre_dfa_exec(). */
 a structure is used to remember local variables on the heap. It is defined in  
 pcre_exec.c, close to the match() function, so that it is easy to keep it in  
 step with any changes of local variable. However, the pointer to the current  
 frame must be saved in some "static" place over a longjmp(). We declare the  
 structure here so that we can put a pointer in the match_data structure. NOTE:  
 This isn't used for a "normal" compilation of pcre. */  
2070    
2071  struct heapframe;  typedef struct dfa_recursion_info {
2072      struct dfa_recursion_info *prevrec;
2073      int group_num;
2074      PCRE_PUCHAR subject_position;
2075    } dfa_recursion_info;
2076    
2077  /* Structure for building a chain of data for holding the values of the subject  /* Structure for building a chain of data for holding the values of the subject
2078  pointer at the start of each subpattern, so as to detect when an empty string  pointer at the start of each subpattern, so as to detect when an empty string
2079  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops; used by
2080    pcre_exec(). */
2081    
2082  typedef struct eptrblock {  typedef struct eptrblock {
2083    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
2084    USPTR epb_saved_eptr;    PCRE_PUCHAR epb_saved_eptr;
2085  } eptrblock;  } eptrblock;
2086    
2087    
# Line 922  typedef struct match_data { Line 2092  typedef struct match_data {
2092    unsigned long int match_call_count;      /* As it says */    unsigned long int match_call_count;      /* As it says */
2093    unsigned long int match_limit;           /* As it says */    unsigned long int match_limit;           /* As it says */
2094    unsigned long int match_limit_recursion; /* As it says */    unsigned long int match_limit_recursion; /* As it says */
2095    int   *offset_vector;         /* Offset vector */    int   *offset_vector;           /* Offset vector */
2096    int    offset_end;            /* One past the end */    int    offset_end;              /* One past the end */
2097    int    offset_max;            /* The maximum usable for return data */    int    offset_max;              /* The maximum usable for return data */
2098    int    nltype;                /* Newline type */    int    nltype;                  /* Newline type */
2099    int    nllen;                 /* Newline string length */    int    nllen;                   /* Newline string length */
2100    uschar nl[4];                 /* Newline string when fixed */    int    name_count;              /* Number of names in name table */
2101    const uschar *lcc;            /* Points to lower casing table */    int    name_entry_size;         /* Size of entry in names table */
2102    const uschar *ctypes;         /* Points to table of type maps */    pcre_uchar *name_table;         /* Table of names */
2103    BOOL   offset_overflow;       /* Set if too many extractions */    pcre_uchar nl[4];               /* Newline string when fixed */
2104    BOOL   notbol;                /* NOTBOL flag */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
2105    BOOL   noteol;                /* NOTEOL flag */    const  pcre_uint8 *fcc;         /* Points to case-flipping table */
2106    BOOL   utf8;                  /* UTF8 flag */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
2107    BOOL   endonly;               /* Dollar not before final \n */    BOOL   offset_overflow;         /* Set if too many extractions */
2108    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notbol;                  /* NOTBOL flag */
2109    BOOL   partial;               /* PARTIAL flag */    BOOL   noteol;                  /* NOTEOL flag */
2110    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   utf;                     /* UTF-8 / UTF-16 flag */
2111    const uschar *start_code;     /* For use when recursing */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
2112    USPTR  start_subject;         /* Start of the subject string */    BOOL   use_ucp;                 /* PCRE_UCP flag */
2113    USPTR  end_subject;           /* End of the subject string */    BOOL   endonly;                 /* Dollar not before final \n */
2114    USPTR  start_match;           /* Start of this match attempt */    BOOL   notempty;                /* Empty string match not wanted */
2115    USPTR  end_match_ptr;         /* Subject position at end match */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
2116    int    end_offset_top;        /* Highwater mark at end of match */    BOOL   hitend;                  /* Hit the end of the subject at some point */
2117    int    capture_last;          /* Most recent capture number */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
2118    int    start_offset;          /* The start offset value */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
2119    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    BOOL   ignore_skip_arg;         /* For re-run when SKIP name not found */
2120    int    eptrn;                 /* Next free eptrblock */    const  pcre_uchar *start_code;  /* For use when recursing */
2121    recursion_info *recursive;    /* Linked list of recursion data */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
2122    void  *callout_data;          /* To pass back to callouts */    PCRE_PUCHAR end_subject;        /* End of the subject string */
2123    struct heapframe *thisframe;  /* Used only when compiling for no recursion */    PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
2124      PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
2125      PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
2126      int    partial;                 /* PARTIAL options */
2127      int    end_offset_top;          /* Highwater mark at end of match */
2128      int    capture_last;            /* Most recent capture number */
2129      int    start_offset;            /* The start offset value */
2130      int    match_function_type;     /* Set for certain special calls of MATCH() */
2131      eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
2132      int    eptrn;                   /* Next free eptrblock */
2133      recursion_info *recursive;      /* Linked list of recursion data */
2134      void  *callout_data;            /* To pass back to callouts */
2135      const  pcre_uchar *mark;        /* Mark pointer to pass back on success */
2136      const  pcre_uchar *nomatch_mark;/* Mark pointer to pass back on failure */
2137      const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
2138    #ifdef NO_RECURSE
2139      void  *match_frames_base;       /* For remembering malloc'd frames */
2140    #endif
2141  } match_data;  } match_data;
2142    
2143  /* A similar structure is used for the same purpose by the DFA matching  /* A similar structure is used for the same purpose by the DFA matching
2144  functions. */  functions. */
2145    
2146  typedef struct dfa_match_data {  typedef struct dfa_match_data {
2147    const uschar *start_code;     /* Start of the compiled pattern */    const pcre_uchar *start_code;     /* Start of the compiled pattern */
2148    const uschar *start_subject;  /* Start of the subject string */    const pcre_uchar *start_subject ; /* Start of the subject string */
2149    const uschar *end_subject;    /* End of subject string */    const pcre_uchar *end_subject;    /* End of subject string */
2150    const uschar *tables;         /* Character tables */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
2151    int   moptions;               /* Match options */    const pcre_uint8 *tables;         /* Character tables */
2152    int   poptions;               /* Pattern options */    int   start_offset;               /* The start offset value */
2153    int    nltype;                /* Newline type */    int   moptions;                   /* Match options */
2154    int    nllen;                 /* Newline string length */    int   poptions;                   /* Pattern options */
2155    uschar nl[4];                 /* Newline string when fixed */    int   nltype;                     /* Newline type */
2156    void  *callout_data;          /* To pass back to callouts */    int   nllen;                      /* Newline string length */
2157      pcre_uchar nl[4];                 /* Newline string when fixed */
2158      void *callout_data;               /* To pass back to callouts */
2159      dfa_recursion_info *recursive;    /* Linked list of recursion data */
2160  } dfa_match_data;  } dfa_match_data;
2161    
2162  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 975  typedef struct dfa_match_data { Line 2165  typedef struct dfa_match_data {
2165  #define ctype_letter  0x02  #define ctype_letter  0x02
2166  #define ctype_digit   0x04  #define ctype_digit   0x04
2167  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
2168  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
2169  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
2170    
2171  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
# Line 1002  total length. */ Line 2192  total length. */
2192  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
2193  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
2194    
2195    /* Internal function and data prefixes. */
2196    
2197    #ifdef COMPILE_PCRE8
2198    #ifndef PUBL
2199    #define PUBL(name) pcre_##name
2200    #endif
2201    #ifndef PRIV
2202    #define PRIV(name) _pcre_##name
2203    #endif
2204    #else /* COMPILE_PCRE8 */
2205    #ifdef COMPILE_PCRE16
2206    #ifndef PUBL
2207    #define PUBL(name) pcre16_##name
2208    #endif
2209    #ifndef PRIV
2210    #define PRIV(name) _pcre16_##name
2211    #endif
2212    #else
2213    #error Unsupported compiling mode
2214    #endif /* COMPILE_PCRE16 */
2215    #endif /* COMPILE_PCRE8 */
2216    
2217  /* Layout of the UCP type table that translates property names into types and  /* Layout of the UCP type table that translates property names into types and
2218  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
2219    relocations in shared libraries, it now has an offset into a single string
2220    instead. */
2221    
2222  typedef struct {  typedef struct {
2223    const char *name;    pcre_uint16 name_offset;
2224    pcre_uint16 type;    pcre_uint16 type;
2225    pcre_uint16 value;    pcre_uint16 value;
2226  } ucp_type_table;  } ucp_type_table;
# Line 1017  of the exported public functions. They h Line 2231  of the exported public functions. They h
2231  but are not part of the PCRE public API. The data for these tables is in the  but are not part of the PCRE public API. The data for these tables is in the
2232  pcre_tables.c module. */  pcre_tables.c module. */
2233    
2234  extern const int    _pcre_utf8_table1[];  #ifdef COMPILE_PCRE8
2235  extern const int    _pcre_utf8_table2[];  
2236  extern const int    _pcre_utf8_table3[];  extern const int            PRIV(utf8_table1)[];
2237  extern const uschar _pcre_utf8_table4[];  extern const int            PRIV(utf8_table1_size);
2238    extern const int            PRIV(utf8_table2)[];
2239    extern const int            PRIV(utf8_table3)[];
2240    extern const pcre_uint8     PRIV(utf8_table4)[];
2241    
2242  extern const int    _pcre_utf8_table1_size;  #endif /* COMPILE_PCRE8 */
2243    
2244  extern const ucp_type_table _pcre_utt[];  extern const char           PRIV(utt_names)[];
2245  extern const int _pcre_utt_size;  extern const ucp_type_table PRIV(utt)[];
2246    extern const int            PRIV(utt_size);
2247    
2248  extern const uschar _pcre_default_tables[];  extern const pcre_uint8     PRIV(default_tables)[];
2249    
2250  extern const uschar _pcre_OP_lengths[];  extern const pcre_uint8     PRIV(OP_lengths)[];
2251    
2252    
2253  /* Internal shared functions. These are functions that are used by more than  /* Internal shared functions. These are functions that are used by more than
2254  one of the exported public functions. They have to be "external" in the C  one of the exported public functions. They have to be "external" in the C
2255  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
2256    
2257  extern BOOL         _pcre_is_newline(const uschar *, const uschar *, int *,  /* String comparison functions. */
2258                        BOOL);  #ifdef COMPILE_PCRE8
2259  extern int          _pcre_ord2utf8(int, uschar *);  
2260  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  #define STRCMP_UC_UC(str1, str2) \
2261                        const pcre_study_data *, pcre_study_data *);    strcmp((char *)(str1), (char *)(str2))
2262  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  #define STRCMP_UC_C8(str1, str2) \
2263  extern unsigned int _pcre_ucp_othercase(const unsigned int);    strcmp((char *)(str1), (str2))
2264  extern int          _pcre_valid_utf8(const uschar *, int);  #define STRNCMP_UC_UC(str1, str2, num) \
2265  extern BOOL         _pcre_was_newline(const uschar *, const uschar *, int *,    strncmp((char *)(str1), (char *)(str2), (num))
2266                        BOOL);  #define STRNCMP_UC_C8(str1, str2, num) \
2267  extern BOOL         _pcre_xclass(int, const uschar *);    strncmp((char *)(str1), (str2), (num))
2268    #define STRLEN_UC(str) strlen((const char *)str)
2269    
2270    #else
2271    
2272    extern int               PRIV(strcmp_uc_uc)(const pcre_uchar *,
2273                               const pcre_uchar *);
2274    extern int               PRIV(strcmp_uc_c8)(const pcre_uchar *,
2275                               const char *);
2276    extern int               PRIV(strncmp_uc_uc)(const pcre_uchar *,
2277                               const pcre_uchar *, unsigned int num);
2278    extern int               PRIV(strncmp_uc_c8)(const pcre_uchar *,
2279                               const char *, unsigned int num);
2280    extern unsigned int      PRIV(strlen_uc)(const pcre_uchar *str);
2281    
2282    #define STRCMP_UC_UC(str1, str2) \
2283      PRIV(strcmp_uc_uc)((str1), (str2))
2284    #define STRCMP_UC_C8(str1, str2) \
2285      PRIV(strcmp_uc_c8)((str1), (str2))
2286    #define STRNCMP_UC_UC(str1, str2, num) \
2287      PRIV(strncmp_uc_uc)((str1), (str2), (num))
2288    #define STRNCMP_UC_C8(str1, str2, num) \
2289      PRIV(strncmp_uc_c8)((str1), (str2), (num))
2290    #define STRLEN_UC(str) PRIV(strlen_uc)(str)
2291    
2292    #endif /* COMPILE_PCRE8 */
2293    
2294    extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2295    extern BOOL              PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2296                               int *, BOOL);
2297    extern int               PRIV(ord2utf)(pcre_uint32, pcre_uchar *);
2298    extern int               PRIV(valid_utf)(PCRE_PUCHAR, int, int *);
2299    extern BOOL              PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2300                               int *, BOOL);
2301    extern BOOL              PRIV(xclass)(int, const pcre_uchar *, BOOL);
2302    
2303    #ifdef SUPPORT_JIT
2304    extern void              PRIV(jit_compile)(const REAL_PCRE *,
2305                               PUBL(extra) *, int);
2306    extern int               PRIV(jit_exec)(const REAL_PCRE *, const PUBL(extra) *,
2307                               const pcre_uchar *, int, int, int, int *, int);
2308    extern void              PRIV(jit_free)(void *);
2309    extern int               PRIV(jit_get_size)(void *);
2310    extern const char*       PRIV(jit_get_target)(void);
2311    #endif
2312    
2313    /* Unicode character database (UCD) */
2314    
2315    typedef struct {
2316      pcre_uint8 script;
2317      pcre_uint8 chartype;
2318      pcre_int32 other_case;
2319    } ucd_record;
2320    
2321    extern const ucd_record  PRIV(ucd_records)[];
2322    extern const pcre_uint8  PRIV(ucd_stage1)[];
2323    extern const pcre_uint16 PRIV(ucd_stage2)[];
2324    extern const int         PRIV(ucp_gentype)[];
2325    #ifdef SUPPORT_JIT
2326    extern const int         PRIV(ucp_typerange)[];
2327    #endif
2328    
2329    #ifdef SUPPORT_UCP
2330    /* UCD access macros */
2331    
2332    #define UCD_BLOCK_SIZE 128
2333    #define GET_UCD(ch) (PRIV(ucd_records) + \
2334            PRIV(ucd_stage2)[PRIV(ucd_stage1)[(ch) / UCD_BLOCK_SIZE] * \
2335            UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
2336    
2337    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
2338    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
2339    #define UCD_CATEGORY(ch)  PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2340    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
2341    
2342    #endif /* SUPPORT_UCP */
2343    
2344  #endif  #endif
2345    

Legend:
Removed from v.137  
changed lines
  Added in v.964

  ViewVC Help
Powered by ViewVC 1.1.5