/[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 207 by ph10, Mon Aug 6 09:32:14 2007 UTC revision 848 by ph10, Wed Jan 4 16:34:01 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 */
# Line 72  setjmp and stdarg are used is when NO_RE Line 106  setjmp and stdarg are used is when NO_RE
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>
# Line 132  PCRE_EXP_DATA_DEFN only if they are not Line 164  PCRE_EXP_DATA_DEFN only if they are not
164  #  endif  #  endif
165  #endif  #endif
166    
167  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* 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
180    
181    /* 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    #ifdef COMPILE_PCRE8
247    
248  typedef unsigned char uschar;  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_SCHAR16 to a 16 bit data type in
258    pcre.h(.in) and disable (comment out) this message. */
259    #error Warning: PCRE_SCHAR16 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 184  start/end of string field names are. */ Line 291  start/end of string field names are. */
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->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\       PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, \
295         utf8)) \         &(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 198  start/end of string field names are. */ Line 305  start/end of 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->nltype, NLBLOCK->PSSTART, \       PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
309         &(NLBLOCK->nllen), utf8)) \         &(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 217  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    
# Line 293  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 329  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 346  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 353  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. BACKCHAR should  
 never be called in byte mode. To make sure it can never even appear when UTF-8  
 support is omitted, we don't even define it. */  
525    
526  #ifndef SUPPORT_UTF8  /* When UTF encoding is being used, a character is no longer just a single
527    character. The macros for character handling generate simple sequences when
528    used in character-mode, and more complicated ones for UTF characters.
529    GETCHARLENTEST and other macros are not used when UTF is not supported,
530    so they are not defined. To make sure they can never even appear when
531    UTF support is omitted, we don't even define them. */
532    
533    #ifndef SUPPORT_UTF
534    
535    /* #define MAX_VALUE_FOR_SINGLE_CHAR */
536    /* #define HAS_EXTRALEN(c) */
537    /* #define GET_EXTRALEN(c) */
538    /* #define NOT_FIRSTCHAR(c) */
539  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
540  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
541  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
542  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
543  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
544    /* #define GETCHARLENTEST(c, eptr, len) */
545  /* #define BACKCHAR(eptr) */  /* #define BACKCHAR(eptr) */
546    /* #define FORWARDCHAR(eptr) */
547    /* #define ACROSSCHAR(condition, eptr, action) */
548    
549  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF */
550    
551    #ifdef COMPILE_PCRE8
552    
553    /* These macros were originally written in the form of loops that used data
554    from the tables whose names start with PRIV(utf8_table). They were rewritten by
555    a user so as not to use loops, because in some environments this gives a
556    significant performance advantage, and it seems never to do any harm. */
557    
558    /* Tells the biggest code point which can be encoded as a single character. */
559    
560    #define MAX_VALUE_FOR_SINGLE_CHAR 127
561    
562    /* Tests whether the code point needs extra characters to decode. */
563    
564    #define HAS_EXTRALEN(c) ((c) >= 0xc0)
565    
566    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
567    Otherwise it has an undefined behaviour. */
568    
569    #define GET_EXTRALEN(c) (PRIV(utf8_table4)[(c) & 0x3f])
570    
571    /* Returns TRUE, if the given character is not the first character
572    of a UTF sequence. */
573    
574    #define NOT_FIRSTCHAR(c) (((c) & 0xc0) == 0x80)
575    
576    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
577    advancing the pointer. */
578    
579    #define GETUTF8(c, eptr) \
580        { \
581        if ((c & 0x20) == 0) \
582          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
583        else if ((c & 0x10) == 0) \
584          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
585        else if ((c & 0x08) == 0) \
586          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
587          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
588        else if ((c & 0x04) == 0) \
589          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
590              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
591              (eptr[4] & 0x3f); \
592        else \
593          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
594              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
595              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
596        }
597    
598  /* 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
599  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
600    
601  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
602    c = *eptr; \    c = *eptr; \
603    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; \  
       } \  
     }  
604    
605  /* 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
606  pointer. */  pointer. */
607    
608  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
609    c = *eptr; \    c = *eptr; \
610    if (utf8 && c >= 0xc0) \    if (utf && c >= 0xc0) GETUTF8(c, eptr);
611    
612    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
613    the pointer. */
614    
615    #define GETUTF8INC(c, eptr) \
616      { \      { \
617      int gcii; \      if ((c & 0x20) == 0) \
618      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
619      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     for (gcii = 1; gcii <= gcaa; gcii++) \  
620        { \        { \
621        gcss -= 6; \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
622        c |= (eptr[gcii] & 0x3f) << gcss; \        eptr += 2; \
623          } \
624        else if ((c & 0x08) == 0) \
625          { \
626          c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
627              ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
628          eptr += 3; \
629          } \
630        else if ((c & 0x04) == 0) \
631          { \
632          c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
633              ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
634              (eptr[3] & 0x3f); \
635          eptr += 4; \
636          } \
637        else \
638          { \
639          c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
640              ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
641              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
642          eptr += 5; \
643        } \        } \
644      }      }
645    
# Line 413  know we are in UTF-8 mode. */ Line 648  know we are in UTF-8 mode. */
648    
649  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
650    c = *eptr++; \    c = *eptr++; \
651    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; \  
       } \  
     }  
652    
653  /* 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.
654    This is called when we don't know if we are in UTF-8 mode. */
655    
656  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
657    c = *eptr++; \    c = *eptr++; \
658    if (utf8 && c >= 0xc0) \    if (utf && c >= 0xc0) GETUTF8INC(c, eptr);
659    
660    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
661    advancing the pointer, incrementing the length. */
662    
663    #define GETUTF8LEN(c, eptr, len) \
664      { \      { \
665      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
666      int gcss = 6*gcaa; \        { \
667      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
668      while (gcaa-- > 0) \        len++; \
669          } \
670        else if ((c & 0x10)  == 0) \
671        { \        { \
672        gcss -= 6; \        c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
673        c |= (*eptr++ & 0x3f) << gcss; \        len += 2; \
674          } \
675        else if ((c & 0x08)  == 0) \
676          {\
677          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
678              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
679          len += 3; \
680          } \
681        else if ((c & 0x04)  == 0) \
682          { \
683          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
684              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
685              (eptr[4] & 0x3f); \
686          len += 4; \
687          } \
688        else \
689          {\
690          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
691              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
692              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
693          len += 5; \
694        } \        } \
695      }      }
696    
# Line 446  if there are extra bytes. This is called Line 699  if there are extra bytes. This is called
699    
700  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
701    c = *eptr; \    c = *eptr; \
702    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
703      { \  
704      int gcii; \  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
705      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  pointer, incrementing length if there are extra bytes. This is called when we
706      int gcss = 6*gcaa; \  do not know if we are in UTF-8 mode. */
707      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
708      for (gcii = 1; gcii <= gcaa; gcii++) \  #define GETCHARLENTEST(c, eptr, len) \
709        { \    c = *eptr; \
710        gcss -= 6; \    if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len);
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
711    
712  /* 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
713  it is. This is called only in UTF-8 mode - we don't put a test within the macro  it is. This is called only in UTF-8 mode - we don't put a test within the macro
714  because almost all calls are already within a block of UTF-8 only code. */  because almost all calls are already within a block of UTF-8 only code. */
715    
716  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
717    
718    /* Same as above, just in the other direction. */
719    #define FORWARDCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr++
720    
721    /* Same as above, but it allows a fully customizable form. */
722    #define ACROSSCHAR(condition, eptr, action) \
723      while((condition) && ((eptr) & 0xc0) == 0x80) action
724    
725    #else /* COMPILE_PCRE8 */
726    
727    #ifdef COMPILE_PCRE16
728    
729    /* Tells the biggest code point which can be encoded as a single character. */
730    
731    #define MAX_VALUE_FOR_SINGLE_CHAR 65535
732    
733    /* Tests whether the code point needs extra characters to decode. */
734    
735    #define HAS_EXTRALEN(c) (((c) & 0xfc00) == 0xd800)
736    
737    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
738    Otherwise it has an undefined behaviour. */
739    
740    #define GET_EXTRALEN(c) 1
741    
742    /* Returns TRUE, if the given character is not the first character
743    of a UTF sequence. */
744    
745    #define NOT_FIRSTCHAR(c) (((c) & 0xfc00) == 0xdc00)
746    
747    /* Base macro to pick up the low surrogate of a UTF-16 character, not
748    advancing the pointer. */
749    
750    #define GETUTF16(c, eptr) \
751       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; }
752    
753    /* Get the next UTF-16 character, not advancing the pointer. This is called when
754    we know we are in UTF-16 mode. */
755    
756    #define GETCHAR(c, eptr) \
757      c = *eptr; \
758      if ((c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
759    
760    /* Get the next UTF-16 character, testing for UTF-16 mode, and not advancing the
761    pointer. */
762    
763    #define GETCHARTEST(c, eptr) \
764      c = *eptr; \
765      if (utf && (c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
766    
767    /* Base macro to pick up the low surrogate of a UTF-16 character, advancing
768    the pointer. */
769    
770    #define GETUTF16INC(c, eptr) \
771       { c = (((c & 0x3ff) << 10) | (*eptr++ & 0x3ff)) + 0x10000; }
772    
773    /* Get the next UTF-16 character, advancing the pointer. This is called when we
774    know we are in UTF-16 mode. */
775    
776    #define GETCHARINC(c, eptr) \
777      c = *eptr++; \
778      if ((c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
779    
780    /* Get the next character, testing for UTF-16 mode, and advancing the pointer.
781    This is called when we don't know if we are in UTF-16 mode. */
782    
783    #define GETCHARINCTEST(c, eptr) \
784      c = *eptr++; \
785      if (utf && (c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
786    
787    /* Base macro to pick up the low surrogate of a UTF-16 character, not
788    advancing the pointer, incrementing the length. */
789    
790    #define GETUTF16LEN(c, eptr, len) \
791       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; len++; }
792    
793    /* Get the next UTF-16 character, not advancing the pointer, incrementing
794    length if there is a low surrogate. This is called when we know we are in
795    UTF-16 mode. */
796    
797    #define GETCHARLEN(c, eptr, len) \
798      c = *eptr; \
799      if ((c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
800    
801    /* Get the next UTF-816character, testing for UTF-16 mode, not advancing the
802    pointer, incrementing length if there is a low surrogate. This is called when
803    we do not know if we are in UTF-16 mode. */
804    
805    #define GETCHARLENTEST(c, eptr, len) \
806      c = *eptr; \
807      if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
808    
809    /* If the pointer is not at the start of a character, move it back until
810    it is. This is called only in UTF-16 mode - we don't put a test within the
811    macro because almost all calls are already within a block of UTF-16 only
812    code. */
813    
814    #define BACKCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr--
815    
816    /* Same as above, just in the other direction. */
817    #define FORWARDCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr++
818    
819    /* Same as above, but it allows a fully customizable form. */
820    #define ACROSSCHAR(condition, eptr, action) \
821      if ((condition) && ((eptr) & 0xfc00) == 0xdc00) action
822    
823  #endif  #endif
824    
825    #endif /* COMPILE_PCRE8 */
826    
827    #endif  /* SUPPORT_UTF */
828    
829    
830  /* In case there is no definition of offsetof() provided - though any proper  /* In case there is no definition of offsetof() provided - though any proper
831  Standard C system should have one. */  Standard C system should have one. */
# Line 477  Standard C system should have one. */ Line 835  Standard C system should have one. */
835  #endif  #endif
836    
837    
838  /* These are the public options that can change during matching. */  /* Private flags containing information about the compiled regex. They used to
839    live at the top end of the options word, but that got almost full, so now they
840    are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
841    the restrictions on partial matching have been lifted. It remains for backwards
842    compatibility. */
843    
844  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #ifdef COMPILE_PCRE8
845    #define PCRE_MODE          0x0001  /* compiled in 8 bit mode */
846  /* Private options flags start at the most significant end of the four bytes.  #endif
847  The public options defined in pcre.h start at the least significant end. Make  #ifdef COMPILE_PCRE16
848  sure they don't overlap! The bits are getting a bit scarce now -- when we run  #define PCRE_MODE          0x0002  /* compiled in 16 bit mode */
849  out, there is a dummy word in the structure that could be used for the private  #endif
850  bits. */  #define PCRE_FIRSTSET      0x0010  /* first_char is set */
851    #define PCRE_FCH_CASELESS  0x0020  /* caseless first char */
852  #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  #define PCRE_REQCHSET      0x0040  /* req_byte is set */
853  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_RCH_CASELESS  0x0080  /* caseless requested char */
854  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_STARTLINE     0x0100  /* start after \n for multiline */
855  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_NOPARTIAL     0x0200  /* can't use partial with this regex */
856  #define PCRE_JCHANGED      0x08000000  /* j option changes within regex */  #define PCRE_JCHANGED      0x0400  /* j option used in regex */
857    #define PCRE_HASCRORLF     0x0800  /* explicit \r or \n in pattern */
858    #define PCRE_HASTHEN       0x1000  /* pattern contains (*THEN) */
859    
860  /* Options for the "extra" block produced by pcre_study(). */  /* Flags for the "extra" block produced by pcre_study(). */
861    
862  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
863    #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
864    
865  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
866  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 503  time, run time, or study time, respectiv Line 868  time, run time, or study time, respectiv
868  #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| \
869                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
870    
871  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
872    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
873     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
874     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| \
875     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
876       PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
877    
878  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
879    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
880     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
881       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
882    
883  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_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_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
886       PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
887       PCRE_NO_START_OPTIMIZE)
888    
889  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS \
890       PCRE_STUDY_JIT_COMPILE
891    
892  /* Magic number to provide a small check against being handed junk. Also used  /* Magic number to provide a small check against being handed junk. */
 to detect whether a pattern was compiled on a host of different endianness. */  
893    
894  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */
895    
896    /* This variable is used to detect a loaded regular expression
897    in different endianness. */
898    
899    #define REVERSED_MAGIC_NUMBER  0x45524350UL   /* 'ERCP' */
900    
901  /* Negative values for the firstchar and reqchar variables */  /* Negative values for the firstchar and reqchar variables */
902    
903  #define REQ_UNSET (-2)  #define REQ_UNSET (-2)
# Line 534  req_byte match. */ Line 908  req_byte match. */
908    
909  #define REQ_BYTE_MAX 1000  #define REQ_BYTE_MAX 1000
910    
911  /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
912  variable-length repeat, or a anything other than literal characters. */  environments where these macros are defined elsewhere. Unfortunately, there
913    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 */  
914    
915  typedef int BOOL;  typedef int BOOL;
916    
917    #ifndef FALSE
918  #define FALSE   0  #define FALSE   0
919  #define TRUE    1  #define TRUE    1
920    #endif
921    
922    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
923    character constants like '*' because the compiler would emit their EBCDIC code,
924    which is different from their ASCII/UTF-8 code. Instead we define macros for
925    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
926    is enabled. When UTF-8 support is not enabled, the definitions use character
927    literals. Both character and string versions of each character are needed, and
928    there are some longer strings as well.
929    
930    This means that, on EBCDIC platforms, the PCRE library can handle either
931    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
932    would need different lookups depending on whether PCRE_UTF8 was set or not.
933    This would make it impossible to use characters in switch/case statements,
934    which would reduce performance. For a theoretical use (which nobody has asked
935    for) in a minority area (EBCDIC platforms), this is not sensible. Any
936    application that did need both could compile two versions of the library, using
937    macros to give the functions distinct names. */
938    
939    #ifndef SUPPORT_UTF
940    
941    /* UTF-8 support is not enabled; use the platform-dependent character literals
942    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
943    
944    #define CHAR_HT                     '\t'
945    #define CHAR_VT                     '\v'
946    #define CHAR_FF                     '\f'
947    #define CHAR_CR                     '\r'
948    #define CHAR_NL                     '\n'
949    #define CHAR_BS                     '\b'
950    #define CHAR_BEL                    '\a'
951    #ifdef EBCDIC
952    #define CHAR_ESC                    '\047'
953    #define CHAR_DEL                    '\007'
954    #else
955    #define CHAR_ESC                    '\033'
956    #define CHAR_DEL                    '\177'
957    #endif
958    
959    #define CHAR_SPACE                  ' '
960    #define CHAR_EXCLAMATION_MARK       '!'
961    #define CHAR_QUOTATION_MARK         '"'
962    #define CHAR_NUMBER_SIGN            '#'
963    #define CHAR_DOLLAR_SIGN            '$'
964    #define CHAR_PERCENT_SIGN           '%'
965    #define CHAR_AMPERSAND              '&'
966    #define CHAR_APOSTROPHE             '\''
967    #define CHAR_LEFT_PARENTHESIS       '('
968    #define CHAR_RIGHT_PARENTHESIS      ')'
969    #define CHAR_ASTERISK               '*'
970    #define CHAR_PLUS                   '+'
971    #define CHAR_COMMA                  ','
972    #define CHAR_MINUS                  '-'
973    #define CHAR_DOT                    '.'
974    #define CHAR_SLASH                  '/'
975    #define CHAR_0                      '0'
976    #define CHAR_1                      '1'
977    #define CHAR_2                      '2'
978    #define CHAR_3                      '3'
979    #define CHAR_4                      '4'
980    #define CHAR_5                      '5'
981    #define CHAR_6                      '6'
982    #define CHAR_7                      '7'
983    #define CHAR_8                      '8'
984    #define CHAR_9                      '9'
985    #define CHAR_COLON                  ':'
986    #define CHAR_SEMICOLON              ';'
987    #define CHAR_LESS_THAN_SIGN         '<'
988    #define CHAR_EQUALS_SIGN            '='
989    #define CHAR_GREATER_THAN_SIGN      '>'
990    #define CHAR_QUESTION_MARK          '?'
991    #define CHAR_COMMERCIAL_AT          '@'
992    #define CHAR_A                      'A'
993    #define CHAR_B                      'B'
994    #define CHAR_C                      'C'
995    #define CHAR_D                      'D'
996    #define CHAR_E                      'E'
997    #define CHAR_F                      'F'
998    #define CHAR_G                      'G'
999    #define CHAR_H                      'H'
1000    #define CHAR_I                      'I'
1001    #define CHAR_J                      'J'
1002    #define CHAR_K                      'K'
1003    #define CHAR_L                      'L'
1004    #define CHAR_M                      'M'
1005    #define CHAR_N                      'N'
1006    #define CHAR_O                      'O'
1007    #define CHAR_P                      'P'
1008    #define CHAR_Q                      'Q'
1009    #define CHAR_R                      'R'
1010    #define CHAR_S                      'S'
1011    #define CHAR_T                      'T'
1012    #define CHAR_U                      'U'
1013    #define CHAR_V                      'V'
1014    #define CHAR_W                      'W'
1015    #define CHAR_X                      'X'
1016    #define CHAR_Y                      'Y'
1017    #define CHAR_Z                      'Z'
1018    #define CHAR_LEFT_SQUARE_BRACKET    '['
1019    #define CHAR_BACKSLASH              '\\'
1020    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
1021    #define CHAR_CIRCUMFLEX_ACCENT      '^'
1022    #define CHAR_UNDERSCORE             '_'
1023    #define CHAR_GRAVE_ACCENT           '`'
1024    #define CHAR_a                      'a'
1025    #define CHAR_b                      'b'
1026    #define CHAR_c                      'c'
1027    #define CHAR_d                      'd'
1028    #define CHAR_e                      'e'
1029    #define CHAR_f                      'f'
1030    #define CHAR_g                      'g'
1031    #define CHAR_h                      'h'
1032    #define CHAR_i                      'i'
1033    #define CHAR_j                      'j'
1034    #define CHAR_k                      'k'
1035    #define CHAR_l                      'l'
1036    #define CHAR_m                      'm'
1037    #define CHAR_n                      'n'
1038    #define CHAR_o                      'o'
1039    #define CHAR_p                      'p'
1040    #define CHAR_q                      'q'
1041    #define CHAR_r                      'r'
1042    #define CHAR_s                      's'
1043    #define CHAR_t                      't'
1044    #define CHAR_u                      'u'
1045    #define CHAR_v                      'v'
1046    #define CHAR_w                      'w'
1047    #define CHAR_x                      'x'
1048    #define CHAR_y                      'y'
1049    #define CHAR_z                      'z'
1050    #define CHAR_LEFT_CURLY_BRACKET     '{'
1051    #define CHAR_VERTICAL_LINE          '|'
1052    #define CHAR_RIGHT_CURLY_BRACKET    '}'
1053    #define CHAR_TILDE                  '~'
1054    
1055    #define STR_HT                      "\t"
1056    #define STR_VT                      "\v"
1057    #define STR_FF                      "\f"
1058    #define STR_CR                      "\r"
1059    #define STR_NL                      "\n"
1060    #define STR_BS                      "\b"
1061    #define STR_BEL                     "\a"
1062    #ifdef EBCDIC
1063    #define STR_ESC                     "\047"
1064    #define STR_DEL                     "\007"
1065    #else
1066    #define STR_ESC                     "\033"
1067    #define STR_DEL                     "\177"
1068    #endif
1069    
1070    #define STR_SPACE                   " "
1071    #define STR_EXCLAMATION_MARK        "!"
1072    #define STR_QUOTATION_MARK          "\""
1073    #define STR_NUMBER_SIGN             "#"
1074    #define STR_DOLLAR_SIGN             "$"
1075    #define STR_PERCENT_SIGN            "%"
1076    #define STR_AMPERSAND               "&"
1077    #define STR_APOSTROPHE              "'"
1078    #define STR_LEFT_PARENTHESIS        "("
1079    #define STR_RIGHT_PARENTHESIS       ")"
1080    #define STR_ASTERISK                "*"
1081    #define STR_PLUS                    "+"
1082    #define STR_COMMA                   ","
1083    #define STR_MINUS                   "-"
1084    #define STR_DOT                     "."
1085    #define STR_SLASH                   "/"
1086    #define STR_0                       "0"
1087    #define STR_1                       "1"
1088    #define STR_2                       "2"
1089    #define STR_3                       "3"
1090    #define STR_4                       "4"
1091    #define STR_5                       "5"
1092    #define STR_6                       "6"
1093    #define STR_7                       "7"
1094    #define STR_8                       "8"
1095    #define STR_9                       "9"
1096    #define STR_COLON                   ":"
1097    #define STR_SEMICOLON               ";"
1098    #define STR_LESS_THAN_SIGN          "<"
1099    #define STR_EQUALS_SIGN             "="
1100    #define STR_GREATER_THAN_SIGN       ">"
1101    #define STR_QUESTION_MARK           "?"
1102    #define STR_COMMERCIAL_AT           "@"
1103    #define STR_A                       "A"
1104    #define STR_B                       "B"
1105    #define STR_C                       "C"
1106    #define STR_D                       "D"
1107    #define STR_E                       "E"
1108    #define STR_F                       "F"
1109    #define STR_G                       "G"
1110    #define STR_H                       "H"
1111    #define STR_I                       "I"
1112    #define STR_J                       "J"
1113    #define STR_K                       "K"
1114    #define STR_L                       "L"
1115    #define STR_M                       "M"
1116    #define STR_N                       "N"
1117    #define STR_O                       "O"
1118    #define STR_P                       "P"
1119    #define STR_Q                       "Q"
1120    #define STR_R                       "R"
1121    #define STR_S                       "S"
1122    #define STR_T                       "T"
1123    #define STR_U                       "U"
1124    #define STR_V                       "V"
1125    #define STR_W                       "W"
1126    #define STR_X                       "X"
1127    #define STR_Y                       "Y"
1128    #define STR_Z                       "Z"
1129    #define STR_LEFT_SQUARE_BRACKET     "["
1130    #define STR_BACKSLASH               "\\"
1131    #define STR_RIGHT_SQUARE_BRACKET    "]"
1132    #define STR_CIRCUMFLEX_ACCENT       "^"
1133    #define STR_UNDERSCORE              "_"
1134    #define STR_GRAVE_ACCENT            "`"
1135    #define STR_a                       "a"
1136    #define STR_b                       "b"
1137    #define STR_c                       "c"
1138    #define STR_d                       "d"
1139    #define STR_e                       "e"
1140    #define STR_f                       "f"
1141    #define STR_g                       "g"
1142    #define STR_h                       "h"
1143    #define STR_i                       "i"
1144    #define STR_j                       "j"
1145    #define STR_k                       "k"
1146    #define STR_l                       "l"
1147    #define STR_m                       "m"
1148    #define STR_n                       "n"
1149    #define STR_o                       "o"
1150    #define STR_p                       "p"
1151    #define STR_q                       "q"
1152    #define STR_r                       "r"
1153    #define STR_s                       "s"
1154    #define STR_t                       "t"
1155    #define STR_u                       "u"
1156    #define STR_v                       "v"
1157    #define STR_w                       "w"
1158    #define STR_x                       "x"
1159    #define STR_y                       "y"
1160    #define STR_z                       "z"
1161    #define STR_LEFT_CURLY_BRACKET      "{"
1162    #define STR_VERTICAL_LINE           "|"
1163    #define STR_RIGHT_CURLY_BRACKET     "}"
1164    #define STR_TILDE                   "~"
1165    
1166    #define STRING_ACCEPT0              "ACCEPT\0"
1167    #define STRING_COMMIT0              "COMMIT\0"
1168    #define STRING_F0                   "F\0"
1169    #define STRING_FAIL0                "FAIL\0"
1170    #define STRING_MARK0                "MARK\0"
1171    #define STRING_PRUNE0               "PRUNE\0"
1172    #define STRING_SKIP0                "SKIP\0"
1173    #define STRING_THEN                 "THEN"
1174    
1175    #define STRING_alpha0               "alpha\0"
1176    #define STRING_lower0               "lower\0"
1177    #define STRING_upper0               "upper\0"
1178    #define STRING_alnum0               "alnum\0"
1179    #define STRING_ascii0               "ascii\0"
1180    #define STRING_blank0               "blank\0"
1181    #define STRING_cntrl0               "cntrl\0"
1182    #define STRING_digit0               "digit\0"
1183    #define STRING_graph0               "graph\0"
1184    #define STRING_print0               "print\0"
1185    #define STRING_punct0               "punct\0"
1186    #define STRING_space0               "space\0"
1187    #define STRING_word0                "word\0"
1188    #define STRING_xdigit               "xdigit"
1189    
1190    #define STRING_DEFINE               "DEFINE"
1191    
1192    #define STRING_CR_RIGHTPAR             "CR)"
1193    #define STRING_LF_RIGHTPAR             "LF)"
1194    #define STRING_CRLF_RIGHTPAR           "CRLF)"
1195    #define STRING_ANY_RIGHTPAR            "ANY)"
1196    #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
1197    #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
1198    #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
1199    #ifdef COMPILE_PCRE8
1200    #define STRING_UTF_RIGHTPAR            "UTF8)"
1201    #endif
1202    #ifdef COMPILE_PCRE16
1203    #define STRING_UTF_RIGHTPAR            "UTF16)"
1204    #endif
1205    #define STRING_UCP_RIGHTPAR            "UCP)"
1206    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
1207    
1208    #else  /* SUPPORT_UTF */
1209    
1210    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
1211    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
1212    only. */
1213    
1214    #define CHAR_HT                     '\011'
1215    #define CHAR_VT                     '\013'
1216    #define CHAR_FF                     '\014'
1217    #define CHAR_CR                     '\015'
1218    #define CHAR_NL                     '\012'
1219    #define CHAR_BS                     '\010'
1220    #define CHAR_BEL                    '\007'
1221    #define CHAR_ESC                    '\033'
1222    #define CHAR_DEL                    '\177'
1223    
1224    #define CHAR_SPACE                  '\040'
1225    #define CHAR_EXCLAMATION_MARK       '\041'
1226    #define CHAR_QUOTATION_MARK         '\042'
1227    #define CHAR_NUMBER_SIGN            '\043'
1228    #define CHAR_DOLLAR_SIGN            '\044'
1229    #define CHAR_PERCENT_SIGN           '\045'
1230    #define CHAR_AMPERSAND              '\046'
1231    #define CHAR_APOSTROPHE             '\047'
1232    #define CHAR_LEFT_PARENTHESIS       '\050'
1233    #define CHAR_RIGHT_PARENTHESIS      '\051'
1234    #define CHAR_ASTERISK               '\052'
1235    #define CHAR_PLUS                   '\053'
1236    #define CHAR_COMMA                  '\054'
1237    #define CHAR_MINUS                  '\055'
1238    #define CHAR_DOT                    '\056'
1239    #define CHAR_SLASH                  '\057'
1240    #define CHAR_0                      '\060'
1241    #define CHAR_1                      '\061'
1242    #define CHAR_2                      '\062'
1243    #define CHAR_3                      '\063'
1244    #define CHAR_4                      '\064'
1245    #define CHAR_5                      '\065'
1246    #define CHAR_6                      '\066'
1247    #define CHAR_7                      '\067'
1248    #define CHAR_8                      '\070'
1249    #define CHAR_9                      '\071'
1250    #define CHAR_COLON                  '\072'
1251    #define CHAR_SEMICOLON              '\073'
1252    #define CHAR_LESS_THAN_SIGN         '\074'
1253    #define CHAR_EQUALS_SIGN            '\075'
1254    #define CHAR_GREATER_THAN_SIGN      '\076'
1255    #define CHAR_QUESTION_MARK          '\077'
1256    #define CHAR_COMMERCIAL_AT          '\100'
1257    #define CHAR_A                      '\101'
1258    #define CHAR_B                      '\102'
1259    #define CHAR_C                      '\103'
1260    #define CHAR_D                      '\104'
1261    #define CHAR_E                      '\105'
1262    #define CHAR_F                      '\106'
1263    #define CHAR_G                      '\107'
1264    #define CHAR_H                      '\110'
1265    #define CHAR_I                      '\111'
1266    #define CHAR_J                      '\112'
1267    #define CHAR_K                      '\113'
1268    #define CHAR_L                      '\114'
1269    #define CHAR_M                      '\115'
1270    #define CHAR_N                      '\116'
1271    #define CHAR_O                      '\117'
1272    #define CHAR_P                      '\120'
1273    #define CHAR_Q                      '\121'
1274    #define CHAR_R                      '\122'
1275    #define CHAR_S                      '\123'
1276    #define CHAR_T                      '\124'
1277    #define CHAR_U                      '\125'
1278    #define CHAR_V                      '\126'
1279    #define CHAR_W                      '\127'
1280    #define CHAR_X                      '\130'
1281    #define CHAR_Y                      '\131'
1282    #define CHAR_Z                      '\132'
1283    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
1284    #define CHAR_BACKSLASH              '\134'
1285    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
1286    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
1287    #define CHAR_UNDERSCORE             '\137'
1288    #define CHAR_GRAVE_ACCENT           '\140'
1289    #define CHAR_a                      '\141'
1290    #define CHAR_b                      '\142'
1291    #define CHAR_c                      '\143'
1292    #define CHAR_d                      '\144'
1293    #define CHAR_e                      '\145'
1294    #define CHAR_f                      '\146'
1295    #define CHAR_g                      '\147'
1296    #define CHAR_h                      '\150'
1297    #define CHAR_i                      '\151'
1298    #define CHAR_j                      '\152'
1299    #define CHAR_k                      '\153'
1300    #define CHAR_l                      '\154'
1301    #define CHAR_m                      '\155'
1302    #define CHAR_n                      '\156'
1303    #define CHAR_o                      '\157'
1304    #define CHAR_p                      '\160'
1305    #define CHAR_q                      '\161'
1306    #define CHAR_r                      '\162'
1307    #define CHAR_s                      '\163'
1308    #define CHAR_t                      '\164'
1309    #define CHAR_u                      '\165'
1310    #define CHAR_v                      '\166'
1311    #define CHAR_w                      '\167'
1312    #define CHAR_x                      '\170'
1313    #define CHAR_y                      '\171'
1314    #define CHAR_z                      '\172'
1315    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1316    #define CHAR_VERTICAL_LINE          '\174'
1317    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1318    #define CHAR_TILDE                  '\176'
1319    
1320    #define STR_HT                      "\011"
1321    #define STR_VT                      "\013"
1322    #define STR_FF                      "\014"
1323    #define STR_CR                      "\015"
1324    #define STR_NL                      "\012"
1325    #define STR_BS                      "\010"
1326    #define STR_BEL                     "\007"
1327    #define STR_ESC                     "\033"
1328    #define STR_DEL                     "\177"
1329    
1330    #define STR_SPACE                   "\040"
1331    #define STR_EXCLAMATION_MARK        "\041"
1332    #define STR_QUOTATION_MARK          "\042"
1333    #define STR_NUMBER_SIGN             "\043"
1334    #define STR_DOLLAR_SIGN             "\044"
1335    #define STR_PERCENT_SIGN            "\045"
1336    #define STR_AMPERSAND               "\046"
1337    #define STR_APOSTROPHE              "\047"
1338    #define STR_LEFT_PARENTHESIS        "\050"
1339    #define STR_RIGHT_PARENTHESIS       "\051"
1340    #define STR_ASTERISK                "\052"
1341    #define STR_PLUS                    "\053"
1342    #define STR_COMMA                   "\054"
1343    #define STR_MINUS                   "\055"
1344    #define STR_DOT                     "\056"
1345    #define STR_SLASH                   "\057"
1346    #define STR_0                       "\060"
1347    #define STR_1                       "\061"
1348    #define STR_2                       "\062"
1349    #define STR_3                       "\063"
1350    #define STR_4                       "\064"
1351    #define STR_5                       "\065"
1352    #define STR_6                       "\066"
1353    #define STR_7                       "\067"
1354    #define STR_8                       "\070"
1355    #define STR_9                       "\071"
1356    #define STR_COLON                   "\072"
1357    #define STR_SEMICOLON               "\073"
1358    #define STR_LESS_THAN_SIGN          "\074"
1359    #define STR_EQUALS_SIGN             "\075"
1360    #define STR_GREATER_THAN_SIGN       "\076"
1361    #define STR_QUESTION_MARK           "\077"
1362    #define STR_COMMERCIAL_AT           "\100"
1363    #define STR_A                       "\101"
1364    #define STR_B                       "\102"
1365    #define STR_C                       "\103"
1366    #define STR_D                       "\104"
1367    #define STR_E                       "\105"
1368    #define STR_F                       "\106"
1369    #define STR_G                       "\107"
1370    #define STR_H                       "\110"
1371    #define STR_I                       "\111"
1372    #define STR_J                       "\112"
1373    #define STR_K                       "\113"
1374    #define STR_L                       "\114"
1375    #define STR_M                       "\115"
1376    #define STR_N                       "\116"
1377    #define STR_O                       "\117"
1378    #define STR_P                       "\120"
1379    #define STR_Q                       "\121"
1380    #define STR_R                       "\122"
1381    #define STR_S                       "\123"
1382    #define STR_T                       "\124"
1383    #define STR_U                       "\125"
1384    #define STR_V                       "\126"
1385    #define STR_W                       "\127"
1386    #define STR_X                       "\130"
1387    #define STR_Y                       "\131"
1388    #define STR_Z                       "\132"
1389    #define STR_LEFT_SQUARE_BRACKET     "\133"
1390    #define STR_BACKSLASH               "\134"
1391    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1392    #define STR_CIRCUMFLEX_ACCENT       "\136"
1393    #define STR_UNDERSCORE              "\137"
1394    #define STR_GRAVE_ACCENT            "\140"
1395    #define STR_a                       "\141"
1396    #define STR_b                       "\142"
1397    #define STR_c                       "\143"
1398    #define STR_d                       "\144"
1399    #define STR_e                       "\145"
1400    #define STR_f                       "\146"
1401    #define STR_g                       "\147"
1402    #define STR_h                       "\150"
1403    #define STR_i                       "\151"
1404    #define STR_j                       "\152"
1405    #define STR_k                       "\153"
1406    #define STR_l                       "\154"
1407    #define STR_m                       "\155"
1408    #define STR_n                       "\156"
1409    #define STR_o                       "\157"
1410    #define STR_p                       "\160"
1411    #define STR_q                       "\161"
1412    #define STR_r                       "\162"
1413    #define STR_s                       "\163"
1414    #define STR_t                       "\164"
1415    #define STR_u                       "\165"
1416    #define STR_v                       "\166"
1417    #define STR_w                       "\167"
1418    #define STR_x                       "\170"
1419    #define STR_y                       "\171"
1420    #define STR_z                       "\172"
1421    #define STR_LEFT_CURLY_BRACKET      "\173"
1422    #define STR_VERTICAL_LINE           "\174"
1423    #define STR_RIGHT_CURLY_BRACKET     "\175"
1424    #define STR_TILDE                   "\176"
1425    
1426    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1427    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1428    #define STRING_F0                   STR_F "\0"
1429    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1430    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1431    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1432    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1433    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1434    
1435    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1436    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1437    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1438    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1439    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1440    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1441    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1442    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1443    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1444    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1445    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1446    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1447    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1448    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1449    
1450    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1451    
1452    #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1453    #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1454    #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1455    #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1456    #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1457    #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
1458    #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
1459    #ifdef COMPILE_PCRE8
1460    #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1461    #endif
1462    #ifdef COMPILE_PCRE16
1463    #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS
1464    #endif
1465    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1466    #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
1467    
1468    #endif  /* SUPPORT_UTF */
1469    
1470  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1471    
1472  #ifndef ESC_e  #ifndef ESC_e
1473  #define ESC_e 27  #define ESC_e CHAR_ESC
1474  #endif  #endif
1475    
1476  #ifndef ESC_f  #ifndef ESC_f
1477  #define ESC_f '\f'  #define ESC_f CHAR_FF
1478  #endif  #endif
1479    
1480  #ifndef ESC_n  #ifndef ESC_n
1481  #define ESC_n '\n'  #define ESC_n CHAR_NL
1482  #endif  #endif
1483    
1484  #ifndef ESC_r  #ifndef ESC_r
1485  #define ESC_r '\r'  #define ESC_r CHAR_CR
1486  #endif  #endif
1487    
1488  /* 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
1489  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1490    
1491  #ifndef ESC_tee  #ifndef ESC_tee
1492  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1493  #endif  #endif
1494    
1495  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
1496    
1497  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1498  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1499  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1500  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1501  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1502    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1503    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1504    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1505    #define PT_WORD       8    /* Word - L plus N plus underscore */
1506    
1507  /* 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
1508  contain UTF-8 characters with values greater than 255. */  contain characters with values greater than 255. */
1509    
1510  #define XCL_NOT    0x01    /* Flag: this is a negative class */  #define XCL_NOT    0x01    /* Flag: this is a negative class */
1511  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */
# Line 595  contain UTF-8 characters with values gre Line 1519  contain UTF-8 characters with values gre
1519  /* 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
1520  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
1521  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
1522  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
1523  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
1524  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
1525  There are two tests in the code for an escape greater than ESC_b and less than  non-DOTALL mode, "." behaves like \N.
1526  ESC_Z to detect the types that may be repeated. These are the types that  
1527  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.
1528  character, that code will have to change. */  when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1529    They must be contiguous, and remain in order so that the replacements can be
1530    looked up from a table.
1531    
1532    The final escape must be ESC_REF as subsequent values are used for
1533    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1534    greater than ESC_b and less than ESC_Z to detect the types that may be
1535    repeated. These are the types that consume characters. If any new escapes are
1536    put in between that don't consume a character, that code will have to change.
1537    */
1538    
1539  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
1540         ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h,         ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1541         ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_k, ESC_REF };         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1542           ESC_E, ESC_Q, ESC_g, ESC_k,
1543           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1544           ESC_REF };
1545    
1546    /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
 /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  
 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to  
1547  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.
1548    
 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.  
   
1549  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1550  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
1551  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1552    
1553  enum {  enum {
1554    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 636  enum { Line 1566  enum {
1566    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1567    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1568    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
   OP_ANY,            /* 12 Match any character */  
   OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */  
   OP_NOTPROP,        /* 14 \P (not Unicode property) */  
   OP_PROP,           /* 15 \p (Unicode property) */  
   OP_ANYNL,          /* 16 \R (any newline sequence) */  
   OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */  
   OP_HSPACE,         /* 18 \h (horizontal whitespace) */  
   OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */  
   OP_VSPACE,         /* 20 \v (vertical whitespace) */  
   OP_EXTUNI,         /* 21 \X (extended Unicode sequence */  
   OP_EODN,           /* 22 End of data or \n at end of data: \Z. */  
   OP_EOD,            /* 23 End of data: \z */  
   
   OP_OPT,            /* 24 Set runtime options */  
   OP_CIRC,           /* 25 Start of line - varies with multiline switch */  
   OP_DOLL,           /* 26 End of line - varies with multiline switch */  
   OP_CHAR,           /* 27 Match one character, casefully */  
   OP_CHARNC,         /* 28 Match one character, caselessly */  
   OP_NOT,            /* 29 Match one character, not the following one */  
   
   OP_STAR,           /* 30 The maximizing and minimizing versions of */  
   OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */  
   OP_PLUS,           /* 32 the minimizing one second. */  
   OP_MINPLUS,        /* 33 This first set applies to single characters.*/  
   OP_QUERY,          /* 34 */  
   OP_MINQUERY,       /* 35 */  
   
   OP_UPTO,           /* 36 From 0 to n matches */  
   OP_MINUPTO,        /* 37 */  
   OP_EXACT,          /* 38 Exactly n matches */  
   
   OP_POSSTAR,        /* 39 Possessified star */  
   OP_POSPLUS,        /* 40 Possessified plus */  
   OP_POSQUERY,       /* 41 Posesssified query */  
   OP_POSUPTO,        /* 42 Possessified upto */  
   
   OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */  
   OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */  
   OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */  
   OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */  
   OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */  
   OP_NOTMINQUERY,    /* 48 */  
   
   OP_NOTUPTO,        /* 49 From 0 to n matches */  
   OP_NOTMINUPTO,     /* 50 */  
   OP_NOTEXACT,       /* 51 Exactly n matches */  
   
   OP_NOTPOSSTAR,     /* 52 Possessified versions */  
   OP_NOTPOSPLUS,     /* 53 */  
   OP_NOTPOSQUERY,    /* 54 */  
   OP_NOTPOSUPTO,     /* 55 */  
   
   OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */  
   OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */  
   OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */  
   OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */  
   OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */  
   OP_TYPEMINQUERY,   /* 61 */  
   
   OP_TYPEUPTO,       /* 62 From 0 to n matches */  
   OP_TYPEMINUPTO,    /* 63 */  
   OP_TYPEEXACT,      /* 64 Exactly n matches */  
   
   OP_TYPEPOSSTAR,    /* 65 Possessified versions */  
   OP_TYPEPOSPLUS,    /* 66 */  
   OP_TYPEPOSQUERY,   /* 67 */  
   OP_TYPEPOSUPTO,    /* 68 */  
   
   OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */  
   OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */  
   OP_CRPLUS,         /* 71 the minimizing one second. These codes must */  
   OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */  
   OP_CRQUERY,        /* 73 These are for character classes and back refs */  
   OP_CRMINQUERY,     /* 74 */  
   OP_CRRANGE,        /* 75 These are different to the three sets above. */  
   OP_CRMINRANGE,     /* 76 */  
   
   OP_CLASS,          /* 77 Match a character class, chars < 256 only */  
   OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative  
                            class - the difference is relevant only when a UTF-8  
                            character > 255 is encountered. */  
   
   OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the  
                            class. This does both positive and negative. */  
   
   OP_REF,            /* 80 Match a back reference */  
   OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */  
   OP_CALLOUT,        /* 82 Call out to external function if provided */  
   
   OP_ALT,            /* 83 Start of alternation */  
   OP_KET,            /* 84 End of group that doesn't have an unbounded repeat */  
   OP_KETRMAX,        /* 85 These two must remain together and in this */  
   OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */  
   
   /* The assertions must come before BRA, CBRA, ONCE, and COND.*/  
   
   OP_ASSERT,         /* 87 Positive lookahead */  
   OP_ASSERT_NOT,     /* 88 Negative lookahead */  
   OP_ASSERTBACK,     /* 89 Positive lookbehind */  
   OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */  
   OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */  
   
   /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,  
   as there's a test for >= ONCE for a subpattern that isn't an assertion. */  
   
   OP_ONCE,           /* 92 Atomic group */  
   OP_BRA,            /* 93 Start of non-capturing bracket */  
   OP_CBRA,           /* 94 Start of capturing bracket */  
   OP_COND,           /* 95 Conditional group */  
1569    
1570    /* These three must follow the previous three, in the same order. There's a    OP_ANY,            /* 12 Match any character except newline */
1571      OP_ALLANY,         /* 13 Match any character */
1572      OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1573      OP_NOTPROP,        /* 15 \P (not Unicode property) */
1574      OP_PROP,           /* 16 \p (Unicode property) */
1575      OP_ANYNL,          /* 17 \R (any newline sequence) */
1576      OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1577      OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1578      OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1579      OP_VSPACE,         /* 21 \v (vertical whitespace) */
1580      OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1581      OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1582      OP_EOD,            /* 24 End of data: \z */
1583    
1584      OP_CIRC,           /* 25 Start of line - not multiline */
1585      OP_CIRCM,          /* 26 Start of line - multiline */
1586      OP_DOLL,           /* 27 End of line - not multiline */
1587      OP_DOLLM,          /* 28 End of line - multiline */
1588      OP_CHAR,           /* 29 Match one character, casefully */
1589      OP_CHARI,          /* 30 Match one character, caselessly */
1590      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1591      OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1592    
1593      /* The following sets of 13 opcodes must always be kept in step because
1594      the offset from the first one is used to generate the others. */
1595    
1596      /**** Single characters, caseful, must precede the caseless ones ****/
1597    
1598      OP_STAR,           /* 33 The maximizing and minimizing versions of */
1599      OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1600      OP_PLUS,           /* 35 the minimizing one second. */
1601      OP_MINPLUS,        /* 36 */
1602      OP_QUERY,          /* 37 */
1603      OP_MINQUERY,       /* 38 */
1604    
1605      OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1606      OP_MINUPTO,        /* 40 */
1607      OP_EXACT,          /* 41 Exactly n matches */
1608    
1609      OP_POSSTAR,        /* 42 Possessified star, caseful */
1610      OP_POSPLUS,        /* 43 Possessified plus, caseful */
1611      OP_POSQUERY,       /* 44 Posesssified query, caseful */
1612      OP_POSUPTO,        /* 45 Possessified upto, caseful */
1613    
1614      /**** Single characters, caseless, must follow the caseful ones */
1615    
1616      OP_STARI,          /* 46 */
1617      OP_MINSTARI,       /* 47 */
1618      OP_PLUSI,          /* 48 */
1619      OP_MINPLUSI,       /* 49 */
1620      OP_QUERYI,         /* 50 */
1621      OP_MINQUERYI,      /* 51 */
1622    
1623      OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1624      OP_MINUPTOI,       /* 53 */
1625      OP_EXACTI,         /* 54 */
1626    
1627      OP_POSSTARI,       /* 55 Possessified star, caseless */
1628      OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1629      OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1630      OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1631    
1632      /**** The negated ones must follow the non-negated ones, and match them ****/
1633      /**** Negated single character, caseful; must precede the caseless ones ****/
1634    
1635      OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1636      OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1637      OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1638      OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1639      OP_NOTQUERY,       /* 63 */
1640      OP_NOTMINQUERY,    /* 64 */
1641    
1642      OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1643      OP_NOTMINUPTO,     /* 66 */
1644      OP_NOTEXACT,       /* 67 Exactly n matches */
1645    
1646      OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1647      OP_NOTPOSPLUS,     /* 69 */
1648      OP_NOTPOSQUERY,    /* 70 */
1649      OP_NOTPOSUPTO,     /* 71 */
1650    
1651      /**** Negated single character, caseless; must follow the caseful ones ****/
1652    
1653      OP_NOTSTARI,       /* 72 */
1654      OP_NOTMINSTARI,    /* 73 */
1655      OP_NOTPLUSI,       /* 74 */
1656      OP_NOTMINPLUSI,    /* 75 */
1657      OP_NOTQUERYI,      /* 76 */
1658      OP_NOTMINQUERYI,   /* 77 */
1659    
1660      OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1661      OP_NOTMINUPTOI,    /* 79 */
1662      OP_NOTEXACTI,      /* 80 Exactly n matches */
1663    
1664      OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1665      OP_NOTPOSPLUSI,    /* 82 */
1666      OP_NOTPOSQUERYI,   /* 83 */
1667      OP_NOTPOSUPTOI,    /* 84 */
1668    
1669      /**** Character types ****/
1670    
1671      OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1672      OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1673      OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1674      OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
1675      OP_TYPEQUERY,      /* 89 */
1676      OP_TYPEMINQUERY,   /* 90 */
1677    
1678      OP_TYPEUPTO,       /* 91 From 0 to n matches */
1679      OP_TYPEMINUPTO,    /* 92 */
1680      OP_TYPEEXACT,      /* 93 Exactly n matches */
1681    
1682      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1683      OP_TYPEPOSPLUS,    /* 95 */
1684      OP_TYPEPOSQUERY,   /* 96 */
1685      OP_TYPEPOSUPTO,    /* 97 */
1686    
1687      /* These are used for character classes and back references; only the
1688      first six are the same as the sets above. */
1689    
1690      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1691      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1692      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1693      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1694      OP_CRQUERY,        /* 102 */
1695      OP_CRMINQUERY,     /* 103 */
1696    
1697      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1698      OP_CRMINRANGE,     /* 105 */
1699    
1700      /* End of quantifier opcodes */
1701    
1702      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1703      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1704                                  class - the difference is relevant only when a
1705                                  character > 255 is encountered. */
1706      OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the
1707                                  class. This does both positive and negative. */
1708      OP_REF,            /* 109 Match a back reference, casefully */
1709      OP_REFI,           /* 110 Match a back reference, caselessly */
1710      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1711      OP_CALLOUT,        /* 112 Call out to external function if provided */
1712    
1713      OP_ALT,            /* 113 Start of alternation */
1714      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1715      OP_KETRMAX,        /* 115 These two must remain together and in this */
1716      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1717      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1718    
1719      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1720      asserts must remain in order. */
1721    
1722      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1723      OP_ASSERT,         /* 119 Positive lookahead */
1724      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1725      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1726      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1727    
1728      /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
1729      after the assertions, with ONCE first, as there's a test for >= ONCE for a
1730      subpattern that isn't an assertion. The POS versions must immediately follow
1731      the non-POS versions in each case. */
1732    
1733      OP_ONCE,           /* 123 Atomic group, contains captures */
1734      OP_ONCE_NC,        /* 124 Atomic group containing no captures */
1735      OP_BRA,            /* 125 Start of non-capturing bracket */
1736      OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
1737      OP_CBRA,           /* 127 Start of capturing bracket */
1738      OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
1739      OP_COND,           /* 129 Conditional group */
1740    
1741      /* These five must follow the previous five, in the same order. There's a
1742    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1743    
1744    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
1745    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
1746    OP_SCOND,          /* 98 Conditional group, check empty */    OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
1747      OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
1748    OP_CREF,           /* 99 Used to hold a capture number as condition */    OP_SCOND,          /* 134 Conditional group, check empty */
1749    OP_RREF,           /* 100 Used to hold a recursion number as condition */  
1750    OP_DEF,            /* 101 The DEFINE condition */    /* The next two pairs must (respectively) be kept together. */
1751    
1752      OP_CREF,           /* 135 Used to hold a capture number as condition */
1753      OP_NCREF,          /* 136 Same, but generated by a name reference*/
1754      OP_RREF,           /* 137 Used to hold a recursion number as condition */
1755      OP_NRREF,          /* 138 Same, but generated by a name reference*/
1756      OP_DEF,            /* 139 The DEFINE condition */
1757    
1758      OP_BRAZERO,        /* 140 These two must remain together and in this */
1759      OP_BRAMINZERO,     /* 141 order. */
1760      OP_BRAPOSZERO,     /* 142 */
1761    
1762      /* These are backtracking control verbs */
1763    
1764      OP_MARK,           /* 143 always has an argument */
1765      OP_PRUNE,          /* 144 */
1766      OP_PRUNE_ARG,      /* 145 same, but with argument */
1767      OP_SKIP,           /* 146 */
1768      OP_SKIP_ARG,       /* 147 same, but with argument */
1769      OP_THEN,           /* 148 */
1770      OP_THEN_ARG,       /* 149 same, but with argument */
1771      OP_COMMIT,         /* 150 */
1772    
1773      /* These are forced failure and success verbs */
1774    
1775      OP_FAIL,           /* 151 */
1776      OP_ACCEPT,         /* 152 */
1777      OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
1778      OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
1779    
1780      /* This is used to skip a subpattern with a {0} quantifier */
1781    
1782      OP_SKIPZERO,       /* 155 */
1783    
1784      /* This is not an opcode, but is used to check that tables indexed by opcode
1785      are the correct length, in order to catch updating errors - there have been
1786      some in the past. */
1787    
1788    OP_BRAZERO,        /* 102 These two must remain together and in this */    OP_TABLE_LENGTH
   OP_BRAMINZERO      /* 103 order. */  
1789  };  };
1790    
1791    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1792    definitions that follow must also be updated to match. There are also tables
1793    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1794    
1795    
1796  /* 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
1797  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
1798    only in pcre_printint.c, which fills out the full names in many cases (and in
1799    some cases doesn't actually use these names at all). */
1800    
1801  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1802    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1803    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1804    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1805    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1806    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1807    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1808      "{", "{", "{",                                                  \
1809    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1810    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1811      "{", "{", "{",                                                  \
1812      "*+","++", "?+", "{",                                           \
1813      "*", "*?", "+", "+?", "?", "??",                                \
1814      "{", "{", "{",                                                  \
1815      "*+","++", "?+", "{",                                           \
1816      "*", "*?", "+", "+?", "?", "??",                                \
1817      "{", "{", "{",                                                  \
1818    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1819    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1820    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1821    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1822    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1823    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1824    "AssertB", "AssertB not", "Reverse",                            \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1825    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1826    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Once", "Once_NC",                                              \
1827      "Bra", "BraPos", "CBra", "CBraPos",                             \
1828      "Cond",                                                         \
1829      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1830      "SCond",                                                        \
1831      "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1832      "Brazero", "Braminzero", "Braposzero",                          \
1833      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1834      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1835      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1836      "Close", "Skip zero"
1837    
1838    
1839  /* 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 798  in UTF-8 mode. The code that uses this t Line 1849  in UTF-8 mode. The code that uses this t
1849    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1850    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1851    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1852    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1853    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1854    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1855    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1,                             /* \X                                     */ \
1856      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1857    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1858    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1859    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1860    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1861    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1862    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1863    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto, minupto             ** mode      */ \
1864      2+IMM2_SIZE,                   /* exact                                  */ \
1865      2, 2, 2, 2+IMM2_SIZE,          /* *+, ++, ?+, upto+                      */ \
1866      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1867      2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto I, minupto I                      */ \
1868      2+IMM2_SIZE,                   /* exact I                                */ \
1869      2, 2, 2, 2+IMM2_SIZE,          /* *+I, ++I, ?+I, upto+I                  */ \
1870    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1871    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1872    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto, minupto                      */ \
1873    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2+IMM2_SIZE,                   /* NOT exact                              */ \
1874      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *, +, ?, upto           */ \
1875      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1876      2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto I, minupto I                  */ \
1877      2+IMM2_SIZE,                   /* NOT exact I                            */ \
1878      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *I, +I, ?I, upto I      */ \
1879    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1880    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1881    4, 4, 4,                       /* Type upto, minupto, exact              */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* Type upto, minupto                     */ \
1882    2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \    2+IMM2_SIZE,                   /* Type exact                             */ \
1883      2, 2, 2, 2+IMM2_SIZE,          /* Possessive *+, ++, ?+, upto+           */ \
1884    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
1885    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
1886    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
1887   33,                             /* CLASS                                  */ \    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
1888   33,                             /* NCLASS                                 */ \    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
1889    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1890    3,                             /* REF                                    */ \    1+IMM2_SIZE,                   /* REF                                    */ \
1891      1+IMM2_SIZE,                   /* REFI                                   */ \
1892    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1893    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1894    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
1895    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
1896    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
1897    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
1898      1+LINK_SIZE,                   /* KetRpos                                */ \
1899      1+LINK_SIZE,                   /* Reverse                                */ \
1900    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
1901    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
1902    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1903    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
   1+LINK_SIZE,                   /* Reverse                                */ \  
1904    1+LINK_SIZE,                   /* ONCE                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1905      1+LINK_SIZE,                   /* ONCE_NC                                */ \
1906    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
1907    3+LINK_SIZE,                   /* CBRA                                   */ \    1+LINK_SIZE,                   /* BRAPOS                                 */ \
1908      1+LINK_SIZE+IMM2_SIZE,         /* CBRA                                   */ \
1909      1+LINK_SIZE+IMM2_SIZE,         /* CBRAPOS                                */ \
1910    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1911    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1912    3+LINK_SIZE,                   /* SCBRA                                  */ \    1+LINK_SIZE,                   /* SBRAPOS                                */ \
1913      1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
1914      1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
1915    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1916    3,                             /* CREF                                   */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \
1917    3,                             /* RREF                                   */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \
1918    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1919    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
1920      3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1921      1, 3,                          /* SKIP, SKIP_ARG                         */ \
1922      1, 3,                          /* THEN, THEN_ARG                         */ \
1923      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
1924      1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
1925    
1926    /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1927  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  condition. */
1928    
1929  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1930    
1931  /* 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
1932  tracked. */  easily be tracked. When a new number is added, the table called eint in
1933    pcreposix.c must be updated. */
1934    
1935  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1936         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
1937         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
1938         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1939         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1940         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1941           ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
1942           ERR70, ERR71, ERR72, ERR73, ERR74, ERRCOUNT };
1943    
1944  /* 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
1945  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 876  Because people can now save and re-use c Line 1955  Because people can now save and re-use c
1955  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
1956  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
1957  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1958  NOTE NOTE NOTE:  NOTE NOTE NOTE
1959  */  */
1960    
1961  typedef struct real_pcre {  typedef struct real_pcre {
1962    pcre_uint32 magic_number;    pcre_uint32 magic_number;
1963    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1964    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
1965    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
1966      pcre_uint16 dummy1;             /* For future use */
1967    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
1968    pcre_uint16 top_backref;    pcre_uint16 top_backref;
1969    pcre_uint16 first_byte;    pcre_uint16 first_char;         /* Starting character */
1970    pcre_uint16 req_byte;    pcre_uint16 req_char;           /* This character must be seen */
1971    pcre_uint16 name_table_offset;  /* Offset to name table that follows */    pcre_uint16 name_table_offset;  /* Offset to name table that follows */
1972    pcre_uint16 name_entry_size;    /* Size of any name items */    pcre_uint16 name_entry_size;    /* Size of any name items */
1973    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
1974    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
1975    
1976    const unsigned char *tables;    /* Pointer to tables or NULL for std */    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
1977    const unsigned char *nullpad;   /* NULL padding */    const pcre_uint8 *nullpad;      /* NULL padding */
1978  } real_pcre;  } real_pcre;
1979    
1980  /* 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
# Line 903  remark (see NOTE above) about extending Line 1982  remark (see NOTE above) about extending
1982    
1983  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1984    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1985    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1986    uschar start_bits[32];    pcre_uint8 start_bits[32];      /* Starting char bits */
1987      pcre_uint32 minlength;          /* Minimum subject length */
1988  } pcre_study_data;  } pcre_study_data;
1989    
1990    /* Structure for building a chain of open capturing subpatterns during
1991    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1992    encountered. This is also used to identify subpatterns that contain recursive
1993    back references to themselves, so that they can be made atomic. */
1994    
1995    typedef struct open_capitem {
1996      struct open_capitem *next;    /* Chain link */
1997      pcre_uint16 number;           /* Capture number */
1998      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1999    } open_capitem;
2000    
2001  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
2002  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
2003    
2004  typedef struct compile_data {  typedef struct compile_data {
2005    const uschar *lcc;            /* Points to lower casing table */    const pcre_uint8 *lcc;            /* Points to lower casing table */
2006    const uschar *fcc;            /* Points to case-flipping table */    const pcre_uint8 *fcc;            /* Points to case-flipping table */
2007    const uschar *cbits;          /* Points to character type table */    const pcre_uint8 *cbits;          /* Points to character type table */
2008    const uschar *ctypes;         /* Points to table of type maps */    const pcre_uint8 *ctypes;         /* Points to table of type maps */
2009    const uschar *start_workspace;/* The start of working space */    const pcre_uchar *start_workspace;/* The start of working space */
2010    const uschar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
2011    const uschar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
2012    const uschar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
2013    uschar *hwm;                  /* High watermark of workspace */    open_capitem *open_caps;          /* Chain of open capture items */
2014    uschar *name_table;           /* The name/number table */    pcre_uchar *hwm;                  /* High watermark of workspace */
2015    int  names_found;             /* Number of entries so far */    pcre_uchar *name_table;           /* The name/number table */
2016    int  name_entry_size;         /* Size of each entry */    int  names_found;                 /* Number of entries so far */
2017    int  bracount;                /* Count of capturing parens */    int  name_entry_size;             /* Size of each entry */
2018    int  top_backref;             /* Maximum back reference */    int  workspace_size;              /* Size of workspace */
2019    unsigned int backref_map;     /* Bitmap of low back refs */    int  bracount;                    /* Count of capturing parens as we compile */
2020    int  external_options;        /* External (initial) options */    int  final_bracount;              /* Saved value after first pass */
2021    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  top_backref;                 /* Maximum back reference */
2022    BOOL nopartial;               /* Set TRUE if partial won't work */    unsigned int backref_map;         /* Bitmap of low back refs */
2023    int  nltype;                  /* Newline type */    int  assert_depth;                /* Depth of nested assertions */
2024    int  nllen;                   /* Newline string length */    int  external_options;            /* External (initial) options */
2025    uschar nl[4];                 /* Newline string when fixed length */    int  external_flags;              /* External flag bits to be set */
2026      int  req_varyopt;                 /* "After variable item" flag for reqbyte */
2027      BOOL had_accept;                  /* (*ACCEPT) encountered */
2028      BOOL check_lookbehind;            /* Lookbehinds need later checking */
2029      int  nltype;                      /* Newline type */
2030      int  nllen;                       /* Newline string length */
2031      pcre_uchar nl[4];                 /* Newline string when fixed length */
2032  } compile_data;  } compile_data;
2033    
2034  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
2035  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
2036    
2037  typedef struct branch_chain {  typedef struct branch_chain {
2038    struct branch_chain *outer;    struct branch_chain *outer;
2039    uschar *current;    pcre_uchar *current_branch;
2040  } branch_chain;  } branch_chain;
2041    
2042  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
2043  call within the pattern. */  call within the pattern; used by pcre_exec(). */
2044    
2045  typedef struct recursion_info {  typedef struct recursion_info {
2046    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
2047    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
2048    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
2049    USPTR save_start;             /* Old value of mstart */    int saved_max;                  /* Number of saved offsets */
2050    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 */  
2051  } recursion_info;  } recursion_info;
2052    
2053    /* A similar structure for pcre_dfa_exec(). */
2054    
2055    typedef struct dfa_recursion_info {
2056      struct dfa_recursion_info *prevrec;
2057      int group_num;
2058      PCRE_PUCHAR subject_position;
2059    } dfa_recursion_info;
2060    
2061  /* 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
2062  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
2063  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops; used by
2064    pcre_exec(). */
2065    
2066  typedef struct eptrblock {  typedef struct eptrblock {
2067    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
2068    USPTR epb_saved_eptr;    PCRE_PUCHAR epb_saved_eptr;
2069  } eptrblock;  } eptrblock;
2070    
2071    
# Line 971  typedef struct match_data { Line 2076  typedef struct match_data {
2076    unsigned long int match_call_count;      /* As it says */    unsigned long int match_call_count;      /* As it says */
2077    unsigned long int match_limit;           /* As it says */    unsigned long int match_limit;           /* As it says */
2078    unsigned long int match_limit_recursion; /* As it says */    unsigned long int match_limit_recursion; /* As it says */
2079    int   *offset_vector;         /* Offset vector */    int   *offset_vector;           /* Offset vector */
2080    int    offset_end;            /* One past the end */    int    offset_end;              /* One past the end */
2081    int    offset_max;            /* The maximum usable for return data */    int    offset_max;              /* The maximum usable for return data */
2082    int    nltype;                /* Newline type */    int    nltype;                  /* Newline type */
2083    int    nllen;                 /* Newline string length */    int    nllen;                   /* Newline string length */
2084    uschar nl[4];                 /* Newline string when fixed */    int    name_count;              /* Number of names in name table */
2085    const uschar *lcc;            /* Points to lower casing table */    int    name_entry_size;         /* Size of entry in names table */
2086    const uschar *ctypes;         /* Points to table of type maps */    pcre_uchar *name_table;         /* Table of names */
2087    BOOL   offset_overflow;       /* Set if too many extractions */    pcre_uchar nl[4];               /* Newline string when fixed */
2088    BOOL   notbol;                /* NOTBOL flag */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
2089    BOOL   noteol;                /* NOTEOL flag */    const  pcre_uint8 *fcc;         /* Points to case-flipping table */
2090    BOOL   utf8;                  /* UTF8 flag */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
2091    BOOL   endonly;               /* Dollar not before final \n */    BOOL   offset_overflow;         /* Set if too many extractions */
2092    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notbol;                  /* NOTBOL flag */
2093    BOOL   partial;               /* PARTIAL flag */    BOOL   noteol;                  /* NOTEOL flag */
2094    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   utf;                     /* UTF-8 / UTF-16 flag */
2095    const uschar *start_code;     /* For use when recursing */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
2096    USPTR  start_subject;         /* Start of the subject string */    BOOL   use_ucp;                 /* PCRE_UCP flag */
2097    USPTR  end_subject;           /* End of the subject string */    BOOL   endonly;                 /* Dollar not before final \n */
2098    USPTR  start_match_ptr;       /* Start of matched string */    BOOL   notempty;                /* Empty string match not wanted */
2099    USPTR  end_match_ptr;         /* Subject position at end match */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
2100    int    end_offset_top;        /* Highwater mark at end of match */    BOOL   hitend;                  /* Hit the end of the subject at some point */
2101    int    capture_last;          /* Most recent capture number */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
2102    int    start_offset;          /* The start offset value */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
2103    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    BOOL   ignore_skip_arg;         /* For re-run when SKIP name not found */
2104    int    eptrn;                 /* Next free eptrblock */    const  pcre_uchar *start_code;  /* For use when recursing */
2105    recursion_info *recursive;    /* Linked list of recursion data */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
2106    void  *callout_data;          /* To pass back to callouts */    PCRE_PUCHAR end_subject;        /* End of the subject string */
2107      PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
2108      PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
2109      PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
2110      int    partial;                 /* PARTIAL options */
2111      int    end_offset_top;          /* Highwater mark at end of match */
2112      int    capture_last;            /* Most recent capture number */
2113      int    start_offset;            /* The start offset value */
2114      int    match_function_type;     /* Set for certain special calls of MATCH() */
2115      eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
2116      int    eptrn;                   /* Next free eptrblock */
2117      recursion_info *recursive;      /* Linked list of recursion data */
2118      void  *callout_data;            /* To pass back to callouts */
2119      const  pcre_uchar *mark;        /* Mark pointer to pass back on success */
2120      const  pcre_uchar *nomatch_mark;/* Mark pointer to pass back on failure */
2121      const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
2122  } match_data;  } match_data;
2123    
2124  /* 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
2125  functions. */  functions. */
2126    
2127  typedef struct dfa_match_data {  typedef struct dfa_match_data {
2128    const uschar *start_code;     /* Start of the compiled pattern */    const pcre_uchar *start_code;     /* Start of the compiled pattern */
2129    const uschar *start_subject;  /* Start of the subject string */    const pcre_uchar *start_subject ; /* Start of the subject string */
2130    const uschar *end_subject;    /* End of subject string */    const pcre_uchar *end_subject;    /* End of subject string */
2131    const uschar *tables;         /* Character tables */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
2132    int   moptions;               /* Match options */    const pcre_uint8 *tables;         /* Character tables */
2133    int   poptions;               /* Pattern options */    int   start_offset;               /* The start offset value */
2134    int    nltype;                /* Newline type */    int   moptions;                   /* Match options */
2135    int    nllen;                 /* Newline string length */    int   poptions;                   /* Pattern options */
2136    uschar nl[4];                 /* Newline string when fixed */    int   nltype;                     /* Newline type */
2137    void  *callout_data;          /* To pass back to callouts */    int   nllen;                      /* Newline string length */
2138      pcre_uchar nl[4];                 /* Newline string when fixed */
2139      void *callout_data;               /* To pass back to callouts */
2140      dfa_recursion_info *recursive;    /* Linked list of recursion data */
2141  } dfa_match_data;  } dfa_match_data;
2142    
2143  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 1023  typedef struct dfa_match_data { Line 2146  typedef struct dfa_match_data {
2146  #define ctype_letter  0x02  #define ctype_letter  0x02
2147  #define ctype_digit   0x04  #define ctype_digit   0x04
2148  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
2149  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
2150  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
2151    
2152  /* 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 1050  total length. */ Line 2173  total length. */
2173  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
2174  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
2175    
2176    /* Internal function prefix */
2177    
2178    #ifdef COMPILE_PCRE8
2179    #define PUBL(name) pcre_##name
2180    #define PRIV(name) _pcre_##name
2181    #else
2182    #ifdef COMPILE_PCRE16
2183    #define PUBL(name) pcre16_##name
2184    #define PRIV(name) _pcre16_##name
2185    #else
2186    #error Unsupported compiling mode
2187    #endif /* COMPILE_PCRE16 */
2188    #endif /* COMPILE_PCRE8 */
2189    
2190  /* 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
2191  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
2192    relocations in shared libraries, it now has an offset into a single string
2193    instead. */
2194    
2195  typedef struct {  typedef struct {
2196    const char *name;    pcre_uint16 name_offset;
2197    pcre_uint16 type;    pcre_uint16 type;
2198    pcre_uint16 value;    pcre_uint16 value;
2199  } ucp_type_table;  } ucp_type_table;
# Line 1065  of the exported public functions. They h Line 2204  of the exported public functions. They h
2204  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
2205  pcre_tables.c module. */  pcre_tables.c module. */
2206    
2207  extern const int    _pcre_utf8_table1[];  #ifdef COMPILE_PCRE8
 extern const int    _pcre_utf8_table2[];  
 extern const int    _pcre_utf8_table3[];  
 extern const uschar _pcre_utf8_table4[];  
2208    
2209  extern const int    _pcre_utf8_table1_size;  extern const int            PRIV(utf8_table1)[];
2210    extern const int            PRIV(utf8_table1_size);
2211    extern const int            PRIV(utf8_table2)[];
2212    extern const int            PRIV(utf8_table3)[];
2213    extern const pcre_uint8     PRIV(utf8_table4)[];
2214    
2215  extern const ucp_type_table _pcre_utt[];  #endif /* COMPILE_PCRE8 */
 extern const int _pcre_utt_size;  
2216    
2217  extern const uschar _pcre_default_tables[];  extern const char           PRIV(utt_names)[];
2218    extern const ucp_type_table PRIV(utt)[];
2219    extern const int            PRIV(utt_size);
2220    
2221  extern const uschar _pcre_OP_lengths[];  extern const pcre_uint8     PRIV(default_tables)[];
2222    
2223    extern const pcre_uint8     PRIV(OP_lengths)[];
2224    
2225    
2226  /* Internal shared functions. These are functions that are used by more than  /* Internal shared functions. These are functions that are used by more than
2227  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
2228  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
2229    
2230  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  /* String comparison functions. */
2231                        int *, BOOL);  #ifdef COMPILE_PCRE8
2232  extern int          _pcre_ord2utf8(int, uschar *);  
2233  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  #define STRCMP_UC_UC(str1, str2) \
2234                        const pcre_study_data *, pcre_study_data *);    strcmp((char *)(str1), (char *)(str2))
2235  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  #define STRCMP_UC_C8(str1, str2) \
2236  extern unsigned int _pcre_ucp_othercase(const unsigned int);    strcmp((char *)(str1), (str2))
2237  extern int          _pcre_valid_utf8(const uschar *, int);  #define STRNCMP_UC_UC(str1, str2, num) \
2238  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,    strncmp((char *)(str1), (char *)(str2), (num))
2239                        int *, BOOL);  #define STRNCMP_UC_C8(str1, str2, num) \
2240  extern BOOL         _pcre_xclass(int, const uschar *);    strncmp((char *)(str1), (str2), (num))
2241    #define STRLEN_UC(str) strlen((const char *)str)
2242    
2243    #else
2244    
2245    extern int               PRIV(strcmp_uc_uc)(const pcre_uchar *,
2246                               const pcre_uchar *);
2247    extern int               PRIV(strcmp_uc_c8)(const pcre_uchar *,
2248                               const char *);
2249    extern int               PRIV(strncmp_uc_uc)(const pcre_uchar *,
2250                               const pcre_uchar *, unsigned int num);
2251    extern int               PRIV(strncmp_uc_c8)(const pcre_uchar *,
2252                               const char *, unsigned int num);
2253    extern unsigned int      PRIV(strlen_uc)(const pcre_uchar *str);
2254    
2255    #define STRCMP_UC_UC(str1, str2) \
2256      PRIV(strcmp_uc_uc)((str1), (str2))
2257    #define STRCMP_UC_C8(str1, str2) \
2258      PRIV(strcmp_uc_c8)((str1), (str2))
2259    #define STRNCMP_UC_UC(str1, str2, num) \
2260      PRIV(strncmp_uc_uc)((str1), (str2), (num))
2261    #define STRNCMP_UC_C8(str1, str2, num) \
2262      PRIV(strncmp_uc_c8)((str1), (str2), (num))
2263    #define STRLEN_UC(str) PRIV(strlen_uc)(str)
2264    
2265    #endif /* COMPILE_PCRE8 */
2266    
2267    extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2268    extern BOOL              PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2269                               int *, BOOL);
2270    extern int               PRIV(ord2utf)(pcre_uint32, pcre_uchar *);
2271    extern int               PRIV(valid_utf)(PCRE_PUCHAR, int, int *);
2272    extern BOOL              PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2273                               int *, BOOL);
2274    extern BOOL              PRIV(xclass)(int, const pcre_uchar *, BOOL);
2275    
2276    #ifdef SUPPORT_JIT
2277    extern void              PRIV(jit_compile)(const real_pcre *, pcre_extra *);
2278    extern int               PRIV(jit_exec)(const real_pcre *, void *,
2279                               const pcre_uchar *, int, int, int, int, int *, int);
2280    extern void              PRIV(jit_free)(void *);
2281    extern int               PRIV(jit_get_size)(void *);
2282    #endif
2283    
2284    /* Unicode character database (UCD) */
2285    
2286    typedef struct {
2287      pcre_uint8 script;
2288      pcre_uint8 chartype;
2289      pcre_int32 other_case;
2290    } ucd_record;
2291    
2292    extern const ucd_record  PRIV(ucd_records)[];
2293    extern const pcre_uint8  PRIV(ucd_stage1)[];
2294    extern const pcre_uint16 PRIV(ucd_stage2)[];
2295    extern const int         PRIV(ucp_gentype)[];
2296    #ifdef SUPPORT_JIT
2297    extern const int         PRIV(ucp_typerange)[];
2298    #endif
2299    
2300    #ifdef SUPPORT_UCP
2301    /* UCD access macros */
2302    
2303    #define UCD_BLOCK_SIZE 128
2304    #define GET_UCD(ch) (PRIV(ucd_records) + \
2305            PRIV(ucd_stage2)[PRIV(ucd_stage1)[(ch) / UCD_BLOCK_SIZE] * \
2306            UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
2307    
2308    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
2309    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
2310    #define UCD_CATEGORY(ch)  PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2311    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
2312    
2313    #endif /* SUPPORT_UCP */
2314    
2315  #endif  #endif
2316    

Legend:
Removed from v.207  
changed lines
  Added in v.848

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12