/[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 270 by ph10, Fri Nov 16 19:55:16 2007 UTC revision 1028 by ph10, Thu Sep 6 16:55:38 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 pcre_uchar;
249    #define IN_UCHARS(x) (x)
250    #define MAX_255(c) 1
251    #define TABLE_GET(c, table, default) ((table)[c])
252    
253    #else
254    
255    #ifdef COMPILE_PCRE16
256    #if USHRT_MAX != 65535
257    /* This is a warning message. Change PCRE_UCHAR16 to a 16 bit data type in
258    pcre.h(.in) and disable (comment out) this message. */
259    #error Warning: PCRE_UCHAR16 is not a 16 bit data type.
260    #endif
261    
262    typedef pcre_uint16 pcre_uchar;
263    #define IN_UCHARS(x) ((x) << 1)
264    #define MAX_255(c) ((c) <= 255u)
265    #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
266    
267    #else
268    #error Unsupported compiling mode
269    #endif /* COMPILE_PCRE16 */
270    
271  typedef unsigned char uschar;  #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    #define PUT2INC(a,n,d)  PUT2(a,n,d), a += IMM2_SIZE
525    
526    /* The maximum length of a MARK name is currently one data unit; it may be
527    changed in future to be a fixed number of bytes or to depend on LINK_SIZE. */
528    
529    #define MAX_MARK ((1 << (sizeof(pcre_uchar)*8)) - 1)
530    
531    /* When UTF encoding is being used, a character is no longer just a single
532    character. The macros for character handling generate simple sequences when
533    used in character-mode, and more complicated ones for UTF characters.
534    GETCHARLENTEST and other macros are not used when UTF is not supported,
535    so they are not defined. To make sure they can never even appear when
536    UTF support is omitted, we don't even define them. */
537    
538  /* When UTF-8 encoding is being used, a character is no longer just a single  #ifndef SUPPORT_UTF
 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. */  
539    
540  #ifndef SUPPORT_UTF8  /* #define MAX_VALUE_FOR_SINGLE_CHAR */
541    /* #define HAS_EXTRALEN(c) */
542    /* #define GET_EXTRALEN(c) */
543    /* #define NOT_FIRSTCHAR(c) */
544  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
545  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
546  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
547  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
548  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
549    /* #define GETCHARLENTEST(c, eptr, len) */
550  /* #define BACKCHAR(eptr) */  /* #define BACKCHAR(eptr) */
551    /* #define FORWARDCHAR(eptr) */
552    /* #define ACROSSCHAR(condition, eptr, action) */
553    
554  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF */
555    
556    #ifdef COMPILE_PCRE8
557    
558    /* These macros were originally written in the form of loops that used data
559    from the tables whose names start with PRIV(utf8_table). They were rewritten by
560    a user so as not to use loops, because in some environments this gives a
561    significant performance advantage, and it seems never to do any harm. */
562    
563    /* Tells the biggest code point which can be encoded as a single character. */
564    
565    #define MAX_VALUE_FOR_SINGLE_CHAR 127
566    
567    /* Tests whether the code point needs extra characters to decode. */
568    
569    #define HAS_EXTRALEN(c) ((c) >= 0xc0)
570    
571    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
572    Otherwise it has an undefined behaviour. */
573    
574    #define GET_EXTRALEN(c) (PRIV(utf8_table4)[(c) & 0x3f])
575    
576    /* Returns TRUE, if the given character is not the first character
577    of a UTF sequence. */
578    
579    #define NOT_FIRSTCHAR(c) (((c) & 0xc0) == 0x80)
580    
581    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
582    advancing the pointer. */
583    
584    #define GETUTF8(c, eptr) \
585        { \
586        if ((c & 0x20) == 0) \
587          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
588        else if ((c & 0x10) == 0) \
589          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
590        else if ((c & 0x08) == 0) \
591          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
592          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
593        else if ((c & 0x04) == 0) \
594          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
595              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
596              (eptr[4] & 0x3f); \
597        else \
598          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
599              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
600              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
601        }
602    
603  /* Get the next UTF-8 character, not advancing the pointer. This is called when  /* Get the next UTF-8 character, not advancing the pointer. This is called when
604  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
605    
606  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
607    c = *eptr; \    c = *eptr; \
608    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8(c, eptr);
     { \  
     int gcii; \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     for (gcii = 1; gcii <= gcaa; gcii++) \  
       { \  
       gcss -= 6; \  
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     }  
609    
610  /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the  /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
611  pointer. */  pointer. */
612    
613  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
614    c = *eptr; \    c = *eptr; \
615    if (utf8 && c >= 0xc0) \    if (utf && c >= 0xc0) GETUTF8(c, eptr);
616    
617    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
618    the pointer. */
619    
620    #define GETUTF8INC(c, eptr) \
621      { \      { \
622      int gcii; \      if ((c & 0x20) == 0) \
623      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
624      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
625      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        { \
626      for (gcii = 1; gcii <= gcaa; gcii++) \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
627          eptr += 2; \
628          } \
629        else if ((c & 0x08) == 0) \
630        { \        { \
631        gcss -= 6; \        c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
632        c |= (eptr[gcii] & 0x3f) << gcss; \            ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
633          eptr += 3; \
634          } \
635        else if ((c & 0x04) == 0) \
636          { \
637          c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
638              ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
639              (eptr[3] & 0x3f); \
640          eptr += 4; \
641          } \
642        else \
643          { \
644          c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
645              ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
646              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
647          eptr += 5; \
648        } \        } \
649      }      }
650    
# Line 413  know we are in UTF-8 mode. */ Line 653  know we are in UTF-8 mode. */
653    
654  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
655    c = *eptr++; \    c = *eptr++; \
656    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8INC(c, eptr);
     { \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     while (gcaa-- > 0) \  
       { \  
       gcss -= 6; \  
       c |= (*eptr++ & 0x3f) << gcss; \  
       } \  
     }  
657    
658  /* Get the next character, testing for UTF-8 mode, and advancing the pointer */  /* Get the next character, testing for UTF-8 mode, and advancing the pointer.
659    This is called when we don't know if we are in UTF-8 mode. */
660    
661  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
662    c = *eptr++; \    c = *eptr++; \
663    if (utf8 && c >= 0xc0) \    if (utf && c >= 0xc0) GETUTF8INC(c, eptr);
664    
665    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
666    advancing the pointer, incrementing the length. */
667    
668    #define GETUTF8LEN(c, eptr, len) \
669      { \      { \
670      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
671      int gcss = 6*gcaa; \        { \
672      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
673      while (gcaa-- > 0) \        len++; \
674          } \
675        else if ((c & 0x10)  == 0) \
676          { \
677          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
678          len += 2; \
679          } \
680        else if ((c & 0x08)  == 0) \
681          {\
682          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
683              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
684          len += 3; \
685          } \
686        else if ((c & 0x04)  == 0) \
687        { \        { \
688        gcss -= 6; \        c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
689        c |= (*eptr++ & 0x3f) << gcss; \            ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
690              (eptr[4] & 0x3f); \
691          len += 4; \
692          } \
693        else \
694          {\
695          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
696              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
697              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
698          len += 5; \
699        } \        } \
700      }      }
701    
# Line 446  if there are extra bytes. This is called Line 704  if there are extra bytes. This is called
704    
705  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
706    c = *eptr; \    c = *eptr; \
707    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
708      { \  
709      int gcii; \  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
710      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  pointer, incrementing length if there are extra bytes. This is called when we
711      int gcss = 6*gcaa; \  do not know if we are in UTF-8 mode. */
712      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
713      for (gcii = 1; gcii <= gcaa; gcii++) \  #define GETCHARLENTEST(c, eptr, len) \
714        { \    c = *eptr; \
715        gcss -= 6; \    if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len);
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
716    
717  /* If the pointer is not at the start of a character, move it back until  /* If the pointer is not at the start of a character, move it back until
718  it is. 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
# Line 466  because almost all calls are already wit Line 720  because almost all calls are already wit
720    
721  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
722    
723    /* Same as above, just in the other direction. */
724    #define FORWARDCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr++
725    
726    /* Same as above, but it allows a fully customizable form. */
727    #define ACROSSCHAR(condition, eptr, action) \
728      while((condition) && ((eptr) & 0xc0) == 0x80) action
729    
730    #else /* COMPILE_PCRE8 */
731    
732    #ifdef COMPILE_PCRE16
733    
734    /* Tells the biggest code point which can be encoded as a single character. */
735    
736    #define MAX_VALUE_FOR_SINGLE_CHAR 65535
737    
738    /* Tests whether the code point needs extra characters to decode. */
739    
740    #define HAS_EXTRALEN(c) (((c) & 0xfc00) == 0xd800)
741    
742    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
743    Otherwise it has an undefined behaviour. */
744    
745    #define GET_EXTRALEN(c) 1
746    
747    /* Returns TRUE, if the given character is not the first character
748    of a UTF sequence. */
749    
750    #define NOT_FIRSTCHAR(c) (((c) & 0xfc00) == 0xdc00)
751    
752    /* Base macro to pick up the low surrogate of a UTF-16 character, not
753    advancing the pointer. */
754    
755    #define GETUTF16(c, eptr) \
756       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; }
757    
758    /* Get the next UTF-16 character, not advancing the pointer. This is called when
759    we know we are in UTF-16 mode. */
760    
761    #define GETCHAR(c, eptr) \
762      c = *eptr; \
763      if ((c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
764    
765    /* Get the next UTF-16 character, testing for UTF-16 mode, and not advancing the
766    pointer. */
767    
768    #define GETCHARTEST(c, eptr) \
769      c = *eptr; \
770      if (utf && (c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
771    
772    /* Base macro to pick up the low surrogate of a UTF-16 character, advancing
773    the pointer. */
774    
775    #define GETUTF16INC(c, eptr) \
776       { c = (((c & 0x3ff) << 10) | (*eptr++ & 0x3ff)) + 0x10000; }
777    
778    /* Get the next UTF-16 character, advancing the pointer. This is called when we
779    know we are in UTF-16 mode. */
780    
781    #define GETCHARINC(c, eptr) \
782      c = *eptr++; \
783      if ((c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
784    
785    /* Get the next character, testing for UTF-16 mode, and advancing the pointer.
786    This is called when we don't know if we are in UTF-16 mode. */
787    
788    #define GETCHARINCTEST(c, eptr) \
789      c = *eptr++; \
790      if (utf && (c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
791    
792    /* Base macro to pick up the low surrogate of a UTF-16 character, not
793    advancing the pointer, incrementing the length. */
794    
795    #define GETUTF16LEN(c, eptr, len) \
796       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; len++; }
797    
798    /* Get the next UTF-16 character, not advancing the pointer, incrementing
799    length if there is a low surrogate. This is called when we know we are in
800    UTF-16 mode. */
801    
802    #define GETCHARLEN(c, eptr, len) \
803      c = *eptr; \
804      if ((c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
805    
806    /* Get the next UTF-816character, testing for UTF-16 mode, not advancing the
807    pointer, incrementing length if there is a low surrogate. This is called when
808    we do not know if we are in UTF-16 mode. */
809    
810    #define GETCHARLENTEST(c, eptr, len) \
811      c = *eptr; \
812      if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
813    
814    /* If the pointer is not at the start of a character, move it back until
815    it is. This is called only in UTF-16 mode - we don't put a test within the
816    macro because almost all calls are already within a block of UTF-16 only
817    code. */
818    
819    #define BACKCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr--
820    
821    /* Same as above, just in the other direction. */
822    #define FORWARDCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr++
823    
824    /* Same as above, but it allows a fully customizable form. */
825    #define ACROSSCHAR(condition, eptr, action) \
826      if ((condition) && ((eptr) & 0xfc00) == 0xdc00) action
827    
828  #endif  #endif
829    
830    #endif /* COMPILE_PCRE8 */
831    
832    #endif  /* SUPPORT_UTF */
833    
834    
835  /* In case there is no definition of offsetof() provided - though any proper  /* In case there is no definition of offsetof() provided - though any proper
836  Standard C system should have one. */  Standard C system should have one. */
# Line 477  Standard C system should have one. */ Line 840  Standard C system should have one. */
840  #endif  #endif
841    
842    
 /* These are the public options that can change during matching. */  
   
 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  
   
843  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
844  live at the top end of the options word, but that got almost full, so now they  live at the top end of the options word, but that got almost full, so now they
845  are in a 16-bit flags word. */  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
846    the restrictions on partial matching have been lifted. It remains for backwards
847    compatibility. */
848    
849  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  #ifdef COMPILE_PCRE8
850  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */  #define PCRE_MODE          0x0001  /* compiled in 8 bit mode */
851  #define PCRE_REQCHSET      0x0004  /* req_byte is set */  #endif
852  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */  #ifdef COMPILE_PCRE16
853  #define PCRE_JCHANGED      0x0010  /* j option used in regex */  #define PCRE_MODE          0x0002  /* compiled in 16 bit mode */
854  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */  #endif
855    #define PCRE_FIRSTSET      0x0010  /* first_char is set */
856    #define PCRE_FCH_CASELESS  0x0020  /* caseless first char */
857    #define PCRE_REQCHSET      0x0040  /* req_byte is set */
858    #define PCRE_RCH_CASELESS  0x0080  /* caseless requested char */
859    #define PCRE_STARTLINE     0x0100  /* start after \n for multiline */
860    #define PCRE_NOPARTIAL     0x0200  /* can't use partial with this regex */
861    #define PCRE_JCHANGED      0x0400  /* j option used in regex */
862    #define PCRE_HASCRORLF     0x0800  /* explicit \r or \n in pattern */
863    #define PCRE_HASTHEN       0x1000  /* pattern contains (*THEN) */
864    
865  /* Options for the "extra" block produced by pcre_study(). */  /* Flags for the "extra" block produced by pcre_study(). */
866    
867  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
868    #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
869    
870  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
871  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 502  time, run time, or study time, respectiv Line 873  time, run time, or study time, respectiv
873  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
874                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
875    
876  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
877    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
878     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
879     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
880     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
881       PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
882    
883  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
884    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
885     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
886       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
887    
888  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
889    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
890     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
891     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
892       PCRE_NO_START_OPTIMIZE)
893  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  
894    #define PUBLIC_STUDY_OPTIONS \
895       (PCRE_STUDY_JIT_COMPILE|PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE| \
896        PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE|PCRE_STUDY_EXTRA_NEEDED)
897    
898  /* 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. */  
899    
900  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */
901    
902    /* This variable is used to detect a loaded regular expression
903    in different endianness. */
904    
905    #define REVERSED_MAGIC_NUMBER  0x45524350UL   /* 'ERCP' */
906    
907  /* Negative values for the firstchar and reqchar variables */  /* Negative values for the firstchar and reqchar variables */
908    
909  #define REQ_UNSET (-2)  #define REQ_UNSET (-2)
# Line 534  req_byte match. */ Line 914  req_byte match. */
914    
915  #define REQ_BYTE_MAX 1000  #define REQ_BYTE_MAX 1000
916    
917  /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
918  variable-length repeat, or a anything other than literal characters. */  environments where these macros are defined elsewhere. Unfortunately, there
919    is no way to do the same for the typedef. */
 #define REQ_CASELESS 0x0100    /* indicates caselessness */  
 #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  
   
 /* Miscellaneous definitions */  
920    
921  typedef int BOOL;  typedef int BOOL;
922    
923    #ifndef FALSE
924  #define FALSE   0  #define FALSE   0
925  #define TRUE    1  #define TRUE    1
926    #endif
927    
928    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
929    character constants like '*' because the compiler would emit their EBCDIC code,
930    which is different from their ASCII/UTF-8 code. Instead we define macros for
931    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
932    is enabled. When UTF-8 support is not enabled, the definitions use character
933    literals. Both character and string versions of each character are needed, and
934    there are some longer strings as well.
935    
936    This means that, on EBCDIC platforms, the PCRE library can handle either
937    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
938    would need different lookups depending on whether PCRE_UTF8 was set or not.
939    This would make it impossible to use characters in switch/case statements,
940    which would reduce performance. For a theoretical use (which nobody has asked
941    for) in a minority area (EBCDIC platforms), this is not sensible. Any
942    application that did need both could compile two versions of the library, using
943    macros to give the functions distinct names. */
944    
945    #ifndef SUPPORT_UTF
946    
947    /* UTF-8 support is not enabled; use the platform-dependent character literals
948    so that PCRE works in both ASCII and EBCDIC environments, but only in non-UTF
949    mode. Newline characters are problematic in EBCDIC. Though it has CR and LF
950    characters, a common practice has been to use its NL (0x15) character as the
951    line terminator in C-like processing environments. However, sometimes the LF
952    (0x25) character is used instead, according to this Unicode document:
953    
954    http://unicode.org/standard/reports/tr13/tr13-5.html
955    
956    PCRE defaults EBCDIC NL to 0x15, but has a build-time option to select 0x25
957    instead. Whichever is *not* chosen is defined as NEL.
958    
959    In both ASCII and EBCDIC environments, CHAR_NL and CHAR_LF are synonyms for the
960    same code point. */
961    
962    #ifdef EBCDIC
963    
964    #ifndef EBCDIC_NL25
965    #define CHAR_NL                     '\x15'
966    #define CHAR_NEL                    '\x25'
967    #define STR_NL                      "\x15"
968    #define STR_NEL                     "\x25"
969    #else
970    #define CHAR_NL                     '\x25'
971    #define CHAR_NEL                    '\x15'
972    #define STR_NL                      "\x25"
973    #define STR_NEL                     "\x15"
974    #endif
975    
976    #define CHAR_LF                     CHAR_NL
977    #define STR_LF                      STR_NL
978    
979    #define CHAR_ESC                    '\047'
980    #define CHAR_DEL                    '\007'
981    #define STR_ESC                     "\047"
982    #define STR_DEL                     "\007"
983    
984    #else  /* Not EBCDIC */
985    
986    /* In ASCII/Unicode, linefeed is '\n' and we equate this to NL for
987    compatibility. NEL is the Unicode newline character. */
988    
989    #define CHAR_LF                     '\n'
990    #define CHAR_NL                     CHAR_LF
991    #define CHAR_NEL                    '\x85'
992    #define CHAR_ESC                    '\033'
993    #define CHAR_DEL                    '\177'
994    
995    #define STR_LF                      "\n"
996    #define STR_NL                      STR_LF
997    #define STR_NEL                     "\x85"
998    #define STR_ESC                     "\033"
999    #define STR_DEL                     "\177"
1000    
1001    #endif  /* EBCDIC */
1002    
1003    /* The remaining definitions work in both environments. */
1004    
1005    #define CHAR_HT                     '\t'
1006    #define CHAR_VT                     '\v'
1007    #define CHAR_FF                     '\f'
1008    #define CHAR_CR                     '\r'
1009    #define CHAR_BS                     '\b'
1010    #define CHAR_BEL                    '\a'
1011    
1012    #define CHAR_SPACE                  ' '
1013    #define CHAR_EXCLAMATION_MARK       '!'
1014    #define CHAR_QUOTATION_MARK         '"'
1015    #define CHAR_NUMBER_SIGN            '#'
1016    #define CHAR_DOLLAR_SIGN            '$'
1017    #define CHAR_PERCENT_SIGN           '%'
1018    #define CHAR_AMPERSAND              '&'
1019    #define CHAR_APOSTROPHE             '\''
1020    #define CHAR_LEFT_PARENTHESIS       '('
1021    #define CHAR_RIGHT_PARENTHESIS      ')'
1022    #define CHAR_ASTERISK               '*'
1023    #define CHAR_PLUS                   '+'
1024    #define CHAR_COMMA                  ','
1025    #define CHAR_MINUS                  '-'
1026    #define CHAR_DOT                    '.'
1027    #define CHAR_SLASH                  '/'
1028    #define CHAR_0                      '0'
1029    #define CHAR_1                      '1'
1030    #define CHAR_2                      '2'
1031    #define CHAR_3                      '3'
1032    #define CHAR_4                      '4'
1033    #define CHAR_5                      '5'
1034    #define CHAR_6                      '6'
1035    #define CHAR_7                      '7'
1036    #define CHAR_8                      '8'
1037    #define CHAR_9                      '9'
1038    #define CHAR_COLON                  ':'
1039    #define CHAR_SEMICOLON              ';'
1040    #define CHAR_LESS_THAN_SIGN         '<'
1041    #define CHAR_EQUALS_SIGN            '='
1042    #define CHAR_GREATER_THAN_SIGN      '>'
1043    #define CHAR_QUESTION_MARK          '?'
1044    #define CHAR_COMMERCIAL_AT          '@'
1045    #define CHAR_A                      'A'
1046    #define CHAR_B                      'B'
1047    #define CHAR_C                      'C'
1048    #define CHAR_D                      'D'
1049    #define CHAR_E                      'E'
1050    #define CHAR_F                      'F'
1051    #define CHAR_G                      'G'
1052    #define CHAR_H                      'H'
1053    #define CHAR_I                      'I'
1054    #define CHAR_J                      'J'
1055    #define CHAR_K                      'K'
1056    #define CHAR_L                      'L'
1057    #define CHAR_M                      'M'
1058    #define CHAR_N                      'N'
1059    #define CHAR_O                      'O'
1060    #define CHAR_P                      'P'
1061    #define CHAR_Q                      'Q'
1062    #define CHAR_R                      'R'
1063    #define CHAR_S                      'S'
1064    #define CHAR_T                      'T'
1065    #define CHAR_U                      'U'
1066    #define CHAR_V                      'V'
1067    #define CHAR_W                      'W'
1068    #define CHAR_X                      'X'
1069    #define CHAR_Y                      'Y'
1070    #define CHAR_Z                      'Z'
1071    #define CHAR_LEFT_SQUARE_BRACKET    '['
1072    #define CHAR_BACKSLASH              '\\'
1073    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
1074    #define CHAR_CIRCUMFLEX_ACCENT      '^'
1075    #define CHAR_UNDERSCORE             '_'
1076    #define CHAR_GRAVE_ACCENT           '`'
1077    #define CHAR_a                      'a'
1078    #define CHAR_b                      'b'
1079    #define CHAR_c                      'c'
1080    #define CHAR_d                      'd'
1081    #define CHAR_e                      'e'
1082    #define CHAR_f                      'f'
1083    #define CHAR_g                      'g'
1084    #define CHAR_h                      'h'
1085    #define CHAR_i                      'i'
1086    #define CHAR_j                      'j'
1087    #define CHAR_k                      'k'
1088    #define CHAR_l                      'l'
1089    #define CHAR_m                      'm'
1090    #define CHAR_n                      'n'
1091    #define CHAR_o                      'o'
1092    #define CHAR_p                      'p'
1093    #define CHAR_q                      'q'
1094    #define CHAR_r                      'r'
1095    #define CHAR_s                      's'
1096    #define CHAR_t                      't'
1097    #define CHAR_u                      'u'
1098    #define CHAR_v                      'v'
1099    #define CHAR_w                      'w'
1100    #define CHAR_x                      'x'
1101    #define CHAR_y                      'y'
1102    #define CHAR_z                      'z'
1103    #define CHAR_LEFT_CURLY_BRACKET     '{'
1104    #define CHAR_VERTICAL_LINE          '|'
1105    #define CHAR_RIGHT_CURLY_BRACKET    '}'
1106    #define CHAR_TILDE                  '~'
1107    
1108    #define STR_HT                      "\t"
1109    #define STR_VT                      "\v"
1110    #define STR_FF                      "\f"
1111    #define STR_CR                      "\r"
1112    #define STR_BS                      "\b"
1113    #define STR_BEL                     "\a"
1114    
1115    #define STR_SPACE                   " "
1116    #define STR_EXCLAMATION_MARK        "!"
1117    #define STR_QUOTATION_MARK          "\""
1118    #define STR_NUMBER_SIGN             "#"
1119    #define STR_DOLLAR_SIGN             "$"
1120    #define STR_PERCENT_SIGN            "%"
1121    #define STR_AMPERSAND               "&"
1122    #define STR_APOSTROPHE              "'"
1123    #define STR_LEFT_PARENTHESIS        "("
1124    #define STR_RIGHT_PARENTHESIS       ")"
1125    #define STR_ASTERISK                "*"
1126    #define STR_PLUS                    "+"
1127    #define STR_COMMA                   ","
1128    #define STR_MINUS                   "-"
1129    #define STR_DOT                     "."
1130    #define STR_SLASH                   "/"
1131    #define STR_0                       "0"
1132    #define STR_1                       "1"
1133    #define STR_2                       "2"
1134    #define STR_3                       "3"
1135    #define STR_4                       "4"
1136    #define STR_5                       "5"
1137    #define STR_6                       "6"
1138    #define STR_7                       "7"
1139    #define STR_8                       "8"
1140    #define STR_9                       "9"
1141    #define STR_COLON                   ":"
1142    #define STR_SEMICOLON               ";"
1143    #define STR_LESS_THAN_SIGN          "<"
1144    #define STR_EQUALS_SIGN             "="
1145    #define STR_GREATER_THAN_SIGN       ">"
1146    #define STR_QUESTION_MARK           "?"
1147    #define STR_COMMERCIAL_AT           "@"
1148    #define STR_A                       "A"
1149    #define STR_B                       "B"
1150    #define STR_C                       "C"
1151    #define STR_D                       "D"
1152    #define STR_E                       "E"
1153    #define STR_F                       "F"
1154    #define STR_G                       "G"
1155    #define STR_H                       "H"
1156    #define STR_I                       "I"
1157    #define STR_J                       "J"
1158    #define STR_K                       "K"
1159    #define STR_L                       "L"
1160    #define STR_M                       "M"
1161    #define STR_N                       "N"
1162    #define STR_O                       "O"
1163    #define STR_P                       "P"
1164    #define STR_Q                       "Q"
1165    #define STR_R                       "R"
1166    #define STR_S                       "S"
1167    #define STR_T                       "T"
1168    #define STR_U                       "U"
1169    #define STR_V                       "V"
1170    #define STR_W                       "W"
1171    #define STR_X                       "X"
1172    #define STR_Y                       "Y"
1173    #define STR_Z                       "Z"
1174    #define STR_LEFT_SQUARE_BRACKET     "["
1175    #define STR_BACKSLASH               "\\"
1176    #define STR_RIGHT_SQUARE_BRACKET    "]"
1177    #define STR_CIRCUMFLEX_ACCENT       "^"
1178    #define STR_UNDERSCORE              "_"
1179    #define STR_GRAVE_ACCENT            "`"
1180    #define STR_a                       "a"
1181    #define STR_b                       "b"
1182    #define STR_c                       "c"
1183    #define STR_d                       "d"
1184    #define STR_e                       "e"
1185    #define STR_f                       "f"
1186    #define STR_g                       "g"
1187    #define STR_h                       "h"
1188    #define STR_i                       "i"
1189    #define STR_j                       "j"
1190    #define STR_k                       "k"
1191    #define STR_l                       "l"
1192    #define STR_m                       "m"
1193    #define STR_n                       "n"
1194    #define STR_o                       "o"
1195    #define STR_p                       "p"
1196    #define STR_q                       "q"
1197    #define STR_r                       "r"
1198    #define STR_s                       "s"
1199    #define STR_t                       "t"
1200    #define STR_u                       "u"
1201    #define STR_v                       "v"
1202    #define STR_w                       "w"
1203    #define STR_x                       "x"
1204    #define STR_y                       "y"
1205    #define STR_z                       "z"
1206    #define STR_LEFT_CURLY_BRACKET      "{"
1207    #define STR_VERTICAL_LINE           "|"
1208    #define STR_RIGHT_CURLY_BRACKET     "}"
1209    #define STR_TILDE                   "~"
1210    
1211    #define STRING_ACCEPT0              "ACCEPT\0"
1212    #define STRING_COMMIT0              "COMMIT\0"
1213    #define STRING_F0                   "F\0"
1214    #define STRING_FAIL0                "FAIL\0"
1215    #define STRING_MARK0                "MARK\0"
1216    #define STRING_PRUNE0               "PRUNE\0"
1217    #define STRING_SKIP0                "SKIP\0"
1218    #define STRING_THEN                 "THEN"
1219    
1220    #define STRING_alpha0               "alpha\0"
1221    #define STRING_lower0               "lower\0"
1222    #define STRING_upper0               "upper\0"
1223    #define STRING_alnum0               "alnum\0"
1224    #define STRING_ascii0               "ascii\0"
1225    #define STRING_blank0               "blank\0"
1226    #define STRING_cntrl0               "cntrl\0"
1227    #define STRING_digit0               "digit\0"
1228    #define STRING_graph0               "graph\0"
1229    #define STRING_print0               "print\0"
1230    #define STRING_punct0               "punct\0"
1231    #define STRING_space0               "space\0"
1232    #define STRING_word0                "word\0"
1233    #define STRING_xdigit               "xdigit"
1234    
1235    #define STRING_DEFINE               "DEFINE"
1236    
1237    #define STRING_CR_RIGHTPAR             "CR)"
1238    #define STRING_LF_RIGHTPAR             "LF)"
1239    #define STRING_CRLF_RIGHTPAR           "CRLF)"
1240    #define STRING_ANY_RIGHTPAR            "ANY)"
1241    #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
1242    #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
1243    #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
1244    #ifdef COMPILE_PCRE8
1245    #define STRING_UTF_RIGHTPAR            "UTF8)"
1246    #endif
1247    #ifdef COMPILE_PCRE16
1248    #define STRING_UTF_RIGHTPAR            "UTF16)"
1249    #endif
1250    #define STRING_UCP_RIGHTPAR            "UCP)"
1251    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
1252    
1253    #else  /* SUPPORT_UTF */
1254    
1255    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
1256    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
1257    only. */
1258    
1259    #define CHAR_HT                     '\011'
1260    #define CHAR_VT                     '\013'
1261    #define CHAR_FF                     '\014'
1262    #define CHAR_CR                     '\015'
1263    #define CHAR_LF                     '\012'
1264    #define CHAR_NL                     CHAR_LF
1265    #define CHAR_NEL                    '\x85'
1266    #define CHAR_BS                     '\010'
1267    #define CHAR_BEL                    '\007'
1268    #define CHAR_ESC                    '\033'
1269    #define CHAR_DEL                    '\177'
1270    
1271    #define CHAR_SPACE                  '\040'
1272    #define CHAR_EXCLAMATION_MARK       '\041'
1273    #define CHAR_QUOTATION_MARK         '\042'
1274    #define CHAR_NUMBER_SIGN            '\043'
1275    #define CHAR_DOLLAR_SIGN            '\044'
1276    #define CHAR_PERCENT_SIGN           '\045'
1277    #define CHAR_AMPERSAND              '\046'
1278    #define CHAR_APOSTROPHE             '\047'
1279    #define CHAR_LEFT_PARENTHESIS       '\050'
1280    #define CHAR_RIGHT_PARENTHESIS      '\051'
1281    #define CHAR_ASTERISK               '\052'
1282    #define CHAR_PLUS                   '\053'
1283    #define CHAR_COMMA                  '\054'
1284    #define CHAR_MINUS                  '\055'
1285    #define CHAR_DOT                    '\056'
1286    #define CHAR_SLASH                  '\057'
1287    #define CHAR_0                      '\060'
1288    #define CHAR_1                      '\061'
1289    #define CHAR_2                      '\062'
1290    #define CHAR_3                      '\063'
1291    #define CHAR_4                      '\064'
1292    #define CHAR_5                      '\065'
1293    #define CHAR_6                      '\066'
1294    #define CHAR_7                      '\067'
1295    #define CHAR_8                      '\070'
1296    #define CHAR_9                      '\071'
1297    #define CHAR_COLON                  '\072'
1298    #define CHAR_SEMICOLON              '\073'
1299    #define CHAR_LESS_THAN_SIGN         '\074'
1300    #define CHAR_EQUALS_SIGN            '\075'
1301    #define CHAR_GREATER_THAN_SIGN      '\076'
1302    #define CHAR_QUESTION_MARK          '\077'
1303    #define CHAR_COMMERCIAL_AT          '\100'
1304    #define CHAR_A                      '\101'
1305    #define CHAR_B                      '\102'
1306    #define CHAR_C                      '\103'
1307    #define CHAR_D                      '\104'
1308    #define CHAR_E                      '\105'
1309    #define CHAR_F                      '\106'
1310    #define CHAR_G                      '\107'
1311    #define CHAR_H                      '\110'
1312    #define CHAR_I                      '\111'
1313    #define CHAR_J                      '\112'
1314    #define CHAR_K                      '\113'
1315    #define CHAR_L                      '\114'
1316    #define CHAR_M                      '\115'
1317    #define CHAR_N                      '\116'
1318    #define CHAR_O                      '\117'
1319    #define CHAR_P                      '\120'
1320    #define CHAR_Q                      '\121'
1321    #define CHAR_R                      '\122'
1322    #define CHAR_S                      '\123'
1323    #define CHAR_T                      '\124'
1324    #define CHAR_U                      '\125'
1325    #define CHAR_V                      '\126'
1326    #define CHAR_W                      '\127'
1327    #define CHAR_X                      '\130'
1328    #define CHAR_Y                      '\131'
1329    #define CHAR_Z                      '\132'
1330    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
1331    #define CHAR_BACKSLASH              '\134'
1332    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
1333    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
1334    #define CHAR_UNDERSCORE             '\137'
1335    #define CHAR_GRAVE_ACCENT           '\140'
1336    #define CHAR_a                      '\141'
1337    #define CHAR_b                      '\142'
1338    #define CHAR_c                      '\143'
1339    #define CHAR_d                      '\144'
1340    #define CHAR_e                      '\145'
1341    #define CHAR_f                      '\146'
1342    #define CHAR_g                      '\147'
1343    #define CHAR_h                      '\150'
1344    #define CHAR_i                      '\151'
1345    #define CHAR_j                      '\152'
1346    #define CHAR_k                      '\153'
1347    #define CHAR_l                      '\154'
1348    #define CHAR_m                      '\155'
1349    #define CHAR_n                      '\156'
1350    #define CHAR_o                      '\157'
1351    #define CHAR_p                      '\160'
1352    #define CHAR_q                      '\161'
1353    #define CHAR_r                      '\162'
1354    #define CHAR_s                      '\163'
1355    #define CHAR_t                      '\164'
1356    #define CHAR_u                      '\165'
1357    #define CHAR_v                      '\166'
1358    #define CHAR_w                      '\167'
1359    #define CHAR_x                      '\170'
1360    #define CHAR_y                      '\171'
1361    #define CHAR_z                      '\172'
1362    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1363    #define CHAR_VERTICAL_LINE          '\174'
1364    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1365    #define CHAR_TILDE                  '\176'
1366    
1367    #define STR_HT                      "\011"
1368    #define STR_VT                      "\013"
1369    #define STR_FF                      "\014"
1370    #define STR_CR                      "\015"
1371    #define STR_NL                      "\012"
1372    #define STR_BS                      "\010"
1373    #define STR_BEL                     "\007"
1374    #define STR_ESC                     "\033"
1375    #define STR_DEL                     "\177"
1376    
1377    #define STR_SPACE                   "\040"
1378    #define STR_EXCLAMATION_MARK        "\041"
1379    #define STR_QUOTATION_MARK          "\042"
1380    #define STR_NUMBER_SIGN             "\043"
1381    #define STR_DOLLAR_SIGN             "\044"
1382    #define STR_PERCENT_SIGN            "\045"
1383    #define STR_AMPERSAND               "\046"
1384    #define STR_APOSTROPHE              "\047"
1385    #define STR_LEFT_PARENTHESIS        "\050"
1386    #define STR_RIGHT_PARENTHESIS       "\051"
1387    #define STR_ASTERISK                "\052"
1388    #define STR_PLUS                    "\053"
1389    #define STR_COMMA                   "\054"
1390    #define STR_MINUS                   "\055"
1391    #define STR_DOT                     "\056"
1392    #define STR_SLASH                   "\057"
1393    #define STR_0                       "\060"
1394    #define STR_1                       "\061"
1395    #define STR_2                       "\062"
1396    #define STR_3                       "\063"
1397    #define STR_4                       "\064"
1398    #define STR_5                       "\065"
1399    #define STR_6                       "\066"
1400    #define STR_7                       "\067"
1401    #define STR_8                       "\070"
1402    #define STR_9                       "\071"
1403    #define STR_COLON                   "\072"
1404    #define STR_SEMICOLON               "\073"
1405    #define STR_LESS_THAN_SIGN          "\074"
1406    #define STR_EQUALS_SIGN             "\075"
1407    #define STR_GREATER_THAN_SIGN       "\076"
1408    #define STR_QUESTION_MARK           "\077"
1409    #define STR_COMMERCIAL_AT           "\100"
1410    #define STR_A                       "\101"
1411    #define STR_B                       "\102"
1412    #define STR_C                       "\103"
1413    #define STR_D                       "\104"
1414    #define STR_E                       "\105"
1415    #define STR_F                       "\106"
1416    #define STR_G                       "\107"
1417    #define STR_H                       "\110"
1418    #define STR_I                       "\111"
1419    #define STR_J                       "\112"
1420    #define STR_K                       "\113"
1421    #define STR_L                       "\114"
1422    #define STR_M                       "\115"
1423    #define STR_N                       "\116"
1424    #define STR_O                       "\117"
1425    #define STR_P                       "\120"
1426    #define STR_Q                       "\121"
1427    #define STR_R                       "\122"
1428    #define STR_S                       "\123"
1429    #define STR_T                       "\124"
1430    #define STR_U                       "\125"
1431    #define STR_V                       "\126"
1432    #define STR_W                       "\127"
1433    #define STR_X                       "\130"
1434    #define STR_Y                       "\131"
1435    #define STR_Z                       "\132"
1436    #define STR_LEFT_SQUARE_BRACKET     "\133"
1437    #define STR_BACKSLASH               "\134"
1438    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1439    #define STR_CIRCUMFLEX_ACCENT       "\136"
1440    #define STR_UNDERSCORE              "\137"
1441    #define STR_GRAVE_ACCENT            "\140"
1442    #define STR_a                       "\141"
1443    #define STR_b                       "\142"
1444    #define STR_c                       "\143"
1445    #define STR_d                       "\144"
1446    #define STR_e                       "\145"
1447    #define STR_f                       "\146"
1448    #define STR_g                       "\147"
1449    #define STR_h                       "\150"
1450    #define STR_i                       "\151"
1451    #define STR_j                       "\152"
1452    #define STR_k                       "\153"
1453    #define STR_l                       "\154"
1454    #define STR_m                       "\155"
1455    #define STR_n                       "\156"
1456    #define STR_o                       "\157"
1457    #define STR_p                       "\160"
1458    #define STR_q                       "\161"
1459    #define STR_r                       "\162"
1460    #define STR_s                       "\163"
1461    #define STR_t                       "\164"
1462    #define STR_u                       "\165"
1463    #define STR_v                       "\166"
1464    #define STR_w                       "\167"
1465    #define STR_x                       "\170"
1466    #define STR_y                       "\171"
1467    #define STR_z                       "\172"
1468    #define STR_LEFT_CURLY_BRACKET      "\173"
1469    #define STR_VERTICAL_LINE           "\174"
1470    #define STR_RIGHT_CURLY_BRACKET     "\175"
1471    #define STR_TILDE                   "\176"
1472    
1473    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1474    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1475    #define STRING_F0                   STR_F "\0"
1476    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1477    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1478    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1479    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1480    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1481    
1482    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1483    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1484    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1485    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1486    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1487    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1488    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1489    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1490    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1491    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1492    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1493    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1494    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1495    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1496    
1497    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1498    
1499    #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1500    #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1501    #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1502    #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1503    #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1504    #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
1505    #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
1506    #ifdef COMPILE_PCRE8
1507    #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1508    #endif
1509    #ifdef COMPILE_PCRE16
1510    #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS
1511    #endif
1512    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1513    #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
1514    
1515    #endif  /* SUPPORT_UTF */
1516    
1517  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1518    
1519  #ifndef ESC_e  #ifndef ESC_e
1520  #define ESC_e 27  #define ESC_e CHAR_ESC
1521  #endif  #endif
1522    
1523  #ifndef ESC_f  #ifndef ESC_f
1524  #define ESC_f '\f'  #define ESC_f CHAR_FF
1525  #endif  #endif
1526    
1527  #ifndef ESC_n  #ifndef ESC_n
1528  #define ESC_n '\n'  #define ESC_n CHAR_LF
1529  #endif  #endif
1530    
1531  #ifndef ESC_r  #ifndef ESC_r
1532  #define ESC_r '\r'  #define ESC_r CHAR_CR
1533  #endif  #endif
1534    
1535  /* 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
1536  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1537    
1538  #ifndef ESC_tee  #ifndef ESC_tee
1539  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1540  #endif  #endif
1541    
1542  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
1543    
1544  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1545  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1546  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1547  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1548  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1549    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1550    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1551    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1552    #define PT_WORD       8    /* Word - L plus N plus underscore */
1553    
1554  /* 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
1555  contain UTF-8 characters with values greater than 255. */  contain characters with values greater than 255. */
1556    
1557  #define XCL_NOT    0x01    /* Flag: this is a negative class */  #define XCL_NOT    0x01    /* Flag: this is a negative class */
1558  #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 1566  contain UTF-8 characters with values gre
1566  /* 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
1567  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
1568  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
1569  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
1570  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
1571  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
1572  There are two tests in the code for an escape greater than ESC_b and less than  non-DOTALL mode, "." behaves like \N.
1573  ESC_Z to detect the types that may be repeated. These are the types that  
1574  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.
1575  character, that code will have to change. */  when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1576    They must be contiguous, and remain in order so that the replacements can be
1577    looked up from a table.
1578    
1579    The final escape must be ESC_REF as subsequent values are used for
1580    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1581    greater than ESC_b and less than ESC_Z to detect the types that may be
1582    repeated. These are the types that consume characters. If any new escapes are
1583    put in between that don't consume a character, that code will have to change.
1584    */
1585    
1586  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,
1587         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,
1588         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,
1589           ESC_E, ESC_Q, ESC_g, ESC_k,
1590           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1591           ESC_REF };
1592    
1593  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
1594  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.
1595    
1596  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1597  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
1598  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1599    
1600  enum {  enum {
1601    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 631  enum { Line 1613  enum {
1613    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1614    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1615    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 */  
1616    
1617    /* These three must follow the previous three, in the same order. There's a    OP_ANY,            /* 12 Match any character except newline */
1618    check for >= SBRA to distinguish the two sets. */    OP_ALLANY,         /* 13 Match any character */
1619      OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1620      OP_NOTPROP,        /* 15 \P (not Unicode property) */
1621      OP_PROP,           /* 16 \p (Unicode property) */
1622      OP_ANYNL,          /* 17 \R (any newline sequence) */
1623      OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1624      OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1625      OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1626      OP_VSPACE,         /* 21 \v (vertical whitespace) */
1627      OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1628      OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1629      OP_EOD,            /* 24 End of data: \z */
1630    
1631      OP_CIRC,           /* 25 Start of line - not multiline */
1632      OP_CIRCM,          /* 26 Start of line - multiline */
1633      OP_DOLL,           /* 27 End of line - not multiline */
1634      OP_DOLLM,          /* 28 End of line - multiline */
1635      OP_CHAR,           /* 29 Match one character, casefully */
1636      OP_CHARI,          /* 30 Match one character, caselessly */
1637      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1638      OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1639    
1640      /* The following sets of 13 opcodes must always be kept in step because
1641      the offset from the first one is used to generate the others. */
1642    
1643      /**** Single characters, caseful, must precede the caseless ones ****/
1644    
1645      OP_STAR,           /* 33 The maximizing and minimizing versions of */
1646      OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1647      OP_PLUS,           /* 35 the minimizing one second. */
1648      OP_MINPLUS,        /* 36 */
1649      OP_QUERY,          /* 37 */
1650      OP_MINQUERY,       /* 38 */
1651    
1652      OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1653      OP_MINUPTO,        /* 40 */
1654      OP_EXACT,          /* 41 Exactly n matches */
1655    
1656      OP_POSSTAR,        /* 42 Possessified star, caseful */
1657      OP_POSPLUS,        /* 43 Possessified plus, caseful */
1658      OP_POSQUERY,       /* 44 Posesssified query, caseful */
1659      OP_POSUPTO,        /* 45 Possessified upto, caseful */
1660    
1661      /**** Single characters, caseless, must follow the caseful ones */
1662    
1663      OP_STARI,          /* 46 */
1664      OP_MINSTARI,       /* 47 */
1665      OP_PLUSI,          /* 48 */
1666      OP_MINPLUSI,       /* 49 */
1667      OP_QUERYI,         /* 50 */
1668      OP_MINQUERYI,      /* 51 */
1669    
1670      OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1671      OP_MINUPTOI,       /* 53 */
1672      OP_EXACTI,         /* 54 */
1673    
1674      OP_POSSTARI,       /* 55 Possessified star, caseless */
1675      OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1676      OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1677      OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1678    
1679      /**** The negated ones must follow the non-negated ones, and match them ****/
1680      /**** Negated single character, caseful; must precede the caseless ones ****/
1681    
1682      OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1683      OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1684      OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1685      OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1686      OP_NOTQUERY,       /* 63 */
1687      OP_NOTMINQUERY,    /* 64 */
1688    
1689      OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1690      OP_NOTMINUPTO,     /* 66 */
1691      OP_NOTEXACT,       /* 67 Exactly n matches */
1692    
1693      OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1694      OP_NOTPOSPLUS,     /* 69 */
1695      OP_NOTPOSQUERY,    /* 70 */
1696      OP_NOTPOSUPTO,     /* 71 */
1697    
1698      /**** Negated single character, caseless; must follow the caseful ones ****/
1699    
1700      OP_NOTSTARI,       /* 72 */
1701      OP_NOTMINSTARI,    /* 73 */
1702      OP_NOTPLUSI,       /* 74 */
1703      OP_NOTMINPLUSI,    /* 75 */
1704      OP_NOTQUERYI,      /* 76 */
1705      OP_NOTMINQUERYI,   /* 77 */
1706    
1707      OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1708      OP_NOTMINUPTOI,    /* 79 */
1709      OP_NOTEXACTI,      /* 80 Exactly n matches */
1710    
1711      OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1712      OP_NOTPOSPLUSI,    /* 82 */
1713      OP_NOTPOSQUERYI,   /* 83 */
1714      OP_NOTPOSUPTOI,    /* 84 */
1715    
1716      /**** Character types ****/
1717    
1718      OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1719      OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1720      OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1721      OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
1722      OP_TYPEQUERY,      /* 89 */
1723      OP_TYPEMINQUERY,   /* 90 */
1724    
1725      OP_TYPEUPTO,       /* 91 From 0 to n matches */
1726      OP_TYPEMINUPTO,    /* 92 */
1727      OP_TYPEEXACT,      /* 93 Exactly n matches */
1728    
1729      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1730      OP_TYPEPOSPLUS,    /* 95 */
1731      OP_TYPEPOSQUERY,   /* 96 */
1732      OP_TYPEPOSUPTO,    /* 97 */
1733    
1734      /* These are used for character classes and back references; only the
1735      first six are the same as the sets above. */
1736    
1737      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1738      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1739      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1740      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1741      OP_CRQUERY,        /* 102 */
1742      OP_CRMINQUERY,     /* 103 */
1743    
1744      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1745      OP_CRMINRANGE,     /* 105 */
1746    
1747      /* End of quantifier opcodes */
1748    
1749      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1750      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1751                                  class - the difference is relevant only when a
1752                                  character > 255 is encountered. */
1753      OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the
1754                                  class. This does both positive and negative. */
1755      OP_REF,            /* 109 Match a back reference, casefully */
1756      OP_REFI,           /* 110 Match a back reference, caselessly */
1757      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1758      OP_CALLOUT,        /* 112 Call out to external function if provided */
1759    
1760      OP_ALT,            /* 113 Start of alternation */
1761      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1762      OP_KETRMAX,        /* 115 These two must remain together and in this */
1763      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1764      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1765    
1766      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1767      asserts must remain in order. */
1768    
1769      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1770      OP_ASSERT,         /* 119 Positive lookahead */
1771      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1772      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1773      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1774    
1775      /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
1776      after the assertions, with ONCE first, as there's a test for >= ONCE for a
1777      subpattern that isn't an assertion. The POS versions must immediately follow
1778      the non-POS versions in each case. */
1779    
1780      OP_ONCE,           /* 123 Atomic group, contains captures */
1781      OP_ONCE_NC,        /* 124 Atomic group containing no captures */
1782      OP_BRA,            /* 125 Start of non-capturing bracket */
1783      OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
1784      OP_CBRA,           /* 127 Start of capturing bracket */
1785      OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
1786      OP_COND,           /* 129 Conditional group */
1787    
1788    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */    /* These five must follow the previous five, in the same order. There's a
1789    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */    check for >= SBRA to distinguish the two sets. */
   OP_SCOND,          /* 98 Conditional group, check empty */  
   
   OP_CREF,           /* 99 Used to hold a capture number as condition */  
   OP_RREF,           /* 100 Used to hold a recursion number as condition */  
   OP_DEF,            /* 101 The DEFINE condition */  
1790    
1791    OP_BRAZERO,        /* 102 These two must remain together and in this */    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
1792    OP_BRAMINZERO,     /* 103 order. */    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
1793      OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
1794      OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
1795      OP_SCOND,          /* 134 Conditional group, check empty */
1796    
1797      /* The next two pairs must (respectively) be kept together. */
1798    
1799      OP_CREF,           /* 135 Used to hold a capture number as condition */
1800      OP_NCREF,          /* 136 Same, but generated by a name reference*/
1801      OP_RREF,           /* 137 Used to hold a recursion number as condition */
1802      OP_NRREF,          /* 138 Same, but generated by a name reference*/
1803      OP_DEF,            /* 139 The DEFINE condition */
1804    
1805      OP_BRAZERO,        /* 140 These two must remain together and in this */
1806      OP_BRAMINZERO,     /* 141 order. */
1807      OP_BRAPOSZERO,     /* 142 */
1808    
1809    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1810    
1811    OP_PRUNE,          /* 104 */    OP_MARK,           /* 143 always has an argument */
1812    OP_SKIP,           /* 105 */    OP_PRUNE,          /* 144 */
1813    OP_THEN,           /* 106 */    OP_PRUNE_ARG,      /* 145 same, but with argument */
1814    OP_COMMIT,         /* 107 */    OP_SKIP,           /* 146 */
1815      OP_SKIP_ARG,       /* 147 same, but with argument */
1816      OP_THEN,           /* 148 */
1817      OP_THEN_ARG,       /* 149 same, but with argument */
1818      OP_COMMIT,         /* 150 */
1819    
1820    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1821    
1822    OP_FAIL,           /* 108 */    OP_FAIL,           /* 151 */
1823    OP_ACCEPT          /* 109 */    OP_ACCEPT,         /* 152 */
1824      OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
1825      OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
1826    
1827      /* This is used to skip a subpattern with a {0} quantifier */
1828    
1829      OP_SKIPZERO,       /* 155 */
1830    
1831      /* This is not an opcode, but is used to check that tables indexed by opcode
1832      are the correct length, in order to catch updating errors - there have been
1833      some in the past. */
1834    
1835      OP_TABLE_LENGTH
1836  };  };
1837    
1838    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1839    definitions that follow must also be updated to match. There are also tables
1840    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1841    
1842    
1843  /* 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
1844  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
1845    only in pcre_printint.c, which fills out the full names in many cases (and in
1846    some cases doesn't actually use these names at all). */
1847    
1848  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1849    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1850    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1851    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1852    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1853    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1854    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1855      "{", "{", "{",                                                  \
1856    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1857    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1858      "{", "{", "{",                                                  \
1859      "*+","++", "?+", "{",                                           \
1860      "*", "*?", "+", "+?", "?", "??",                                \
1861      "{", "{", "{",                                                  \
1862      "*+","++", "?+", "{",                                           \
1863      "*", "*?", "+", "+?", "?", "??",                                \
1864      "{", "{", "{",                                                  \
1865    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1866    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1867    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1868    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1869    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1870    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1871    "AssertB", "AssertB not", "Reverse",                            \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1872    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1873    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Once", "Once_NC",                                              \
1874    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT"    "Bra", "BraPos", "CBra", "CBraPos",                             \
1875      "Cond",                                                         \
1876      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1877      "SCond",                                                        \
1878      "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1879      "Brazero", "Braminzero", "Braposzero",                          \
1880      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1881      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1882      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1883      "Close", "Skip zero"
1884    
1885    
1886  /* 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 806  in UTF-8 mode. The code that uses this t Line 1896  in UTF-8 mode. The code that uses this t
1896    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1897    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1898    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1899    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1900    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1901    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1902    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1,                             /* \X                                     */ \
1903      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1904    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1905    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1906    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1907    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1908    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1909    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1910    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto, minupto             ** mode      */ \
1911      2+IMM2_SIZE,                   /* exact                                  */ \
1912      2, 2, 2, 2+IMM2_SIZE,          /* *+, ++, ?+, upto+                      */ \
1913      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1914      2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto I, minupto I                      */ \
1915      2+IMM2_SIZE,                   /* exact I                                */ \
1916      2, 2, 2, 2+IMM2_SIZE,          /* *+I, ++I, ?+I, upto+I                  */ \
1917    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1918    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1919    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto, minupto                      */ \
1920    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2+IMM2_SIZE,                   /* NOT exact                              */ \
1921      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *, +, ?, upto           */ \
1922      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1923      2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto I, minupto I                  */ \
1924      2+IMM2_SIZE,                   /* NOT exact I                            */ \
1925      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *I, +I, ?I, upto I      */ \
1926    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1927    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1928    4, 4, 4,                       /* Type upto, minupto, exact              */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* Type upto, minupto                     */ \
1929    2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \    2+IMM2_SIZE,                   /* Type exact                             */ \
1930      2, 2, 2, 2+IMM2_SIZE,          /* Possessive *+, ++, ?+, upto+           */ \
1931    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
1932    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
1933    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
1934   33,                             /* CLASS                                  */ \    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
1935   33,                             /* NCLASS                                 */ \    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
1936    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1937    3,                             /* REF                                    */ \    1+IMM2_SIZE,                   /* REF                                    */ \
1938      1+IMM2_SIZE,                   /* REFI                                   */ \
1939    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1940    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1941    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
1942    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
1943    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
1944    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
1945      1+LINK_SIZE,                   /* KetRpos                                */ \
1946      1+LINK_SIZE,                   /* Reverse                                */ \
1947    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
1948    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
1949    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1950    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
   1+LINK_SIZE,                   /* Reverse                                */ \  
1951    1+LINK_SIZE,                   /* ONCE                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1952      1+LINK_SIZE,                   /* ONCE_NC                                */ \
1953    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
1954    3+LINK_SIZE,                   /* CBRA                                   */ \    1+LINK_SIZE,                   /* BRAPOS                                 */ \
1955      1+LINK_SIZE+IMM2_SIZE,         /* CBRA                                   */ \
1956      1+LINK_SIZE+IMM2_SIZE,         /* CBRAPOS                                */ \
1957    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1958    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1959    3+LINK_SIZE,                   /* SCBRA                                  */ \    1+LINK_SIZE,                   /* SBRAPOS                                */ \
1960      1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
1961      1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
1962    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1963    3,                             /* CREF                                   */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \
1964    3,                             /* RREF                                   */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \
1965    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1966    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
1967    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1968    1, 1                           /* FAIL, ACCEPT                           */    1, 3,                          /* SKIP, SKIP_ARG                         */ \
1969      1, 3,                          /* THEN, THEN_ARG                         */ \
1970      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
1971      1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
1972    
1973  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1974    condition. */
1975    
1976  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1977    
1978  /* 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
1979  tracked. */  easily be tracked. When a new number is added, the table called eint in
1980    pcreposix.c must be updated. */
1981    
1982  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1983         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
# Line 871  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 1985  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1985         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1986         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1987         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1988         ERR60, ERR61, ERR62, ERR63 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
1989           ERR70, ERR71, ERR72, ERR73, ERR74, ERR75, ERR76, ERRCOUNT };
1990    
1991    /* JIT compiling modes. The function list is indexed by them. */
1992    enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
1993           JIT_NUMBER_OF_COMPILE_MODES };
1994    
1995  /* 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
1996  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 887  Because people can now save and re-use c Line 2006  Because people can now save and re-use c
2006  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
2007  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
2008  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
2009  NOTE NOTE NOTE:  NOTE NOTE NOTE
2010  */  */
2011    
2012  typedef struct real_pcre {  #ifdef COMPILE_PCRE8
2013    #define REAL_PCRE real_pcre
2014    #else
2015    #define REAL_PCRE real_pcre16
2016    #endif
2017    
2018    typedef struct REAL_PCRE {
2019    pcre_uint32 magic_number;    pcre_uint32 magic_number;
2020    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
2021    pcre_uint32 options;            /* Public options */    pcre_uint32 options;            /* Public options */
2022    pcre_uint16 flags;              /* Private flags */    pcre_uint16 flags;              /* Private flags */
2023    pcre_uint16 dummy1;             /* For future use */    pcre_uint16 max_lookbehind;     /* Longest lookbehind (characters) */
2024    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;        /* Highest numbered group */
2025    pcre_uint16 top_backref;    pcre_uint16 top_backref;        /* Highest numbered back reference */
2026    pcre_uint16 first_byte;    pcre_uint16 first_char;         /* Starting character */
2027    pcre_uint16 req_byte;    pcre_uint16 req_char;           /* This character must be seen */
2028    pcre_uint16 name_table_offset;  /* Offset to name table that follows */    pcre_uint16 name_table_offset;  /* Offset to name table that follows */
2029    pcre_uint16 name_entry_size;    /* Size of any name items */    pcre_uint16 name_entry_size;    /* Size of any name items */
2030    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
2031    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
2032      const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
2033    const unsigned char *tables;    /* Pointer to tables or NULL for std */    const pcre_uint8 *nullpad;      /* NULL padding */
2034    const unsigned char *nullpad;   /* NULL padding */  } REAL_PCRE;
 } real_pcre;  
2035    
2036  /* 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
2037  remark (see NOTE above) about extending this structure applies. */  remark (see NOTE above) about extending this structure applies. */
2038    
2039  typedef struct pcre_study_data {  typedef struct pcre_study_data {
2040    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
2041    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
2042    uschar start_bits[32];    pcre_uint8 start_bits[32];      /* Starting char bits */
2043      pcre_uint32 minlength;          /* Minimum subject length */
2044  } pcre_study_data;  } pcre_study_data;
2045    
2046    /* Structure for building a chain of open capturing subpatterns during
2047    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
2048    encountered. This is also used to identify subpatterns that contain recursive
2049    back references to themselves, so that they can be made atomic. */
2050    
2051    typedef struct open_capitem {
2052      struct open_capitem *next;    /* Chain link */
2053      pcre_uint16 number;           /* Capture number */
2054      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
2055    } open_capitem;
2056    
2057  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
2058  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
2059    
2060  typedef struct compile_data {  typedef struct compile_data {
2061    const uschar *lcc;            /* Points to lower casing table */    const pcre_uint8 *lcc;            /* Points to lower casing table */
2062    const uschar *fcc;            /* Points to case-flipping table */    const pcre_uint8 *fcc;            /* Points to case-flipping table */
2063    const uschar *cbits;          /* Points to character type table */    const pcre_uint8 *cbits;          /* Points to character type table */
2064    const uschar *ctypes;         /* Points to table of type maps */    const pcre_uint8 *ctypes;         /* Points to table of type maps */
2065    const uschar *start_workspace;/* The start of working space */    const pcre_uchar *start_workspace;/* The start of working space */
2066    const uschar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
2067    const uschar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
2068    const uschar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
2069    uschar *hwm;                  /* High watermark of workspace */    open_capitem *open_caps;          /* Chain of open capture items */
2070    uschar *name_table;           /* The name/number table */    pcre_uchar *hwm;                  /* High watermark of workspace */
2071    int  names_found;             /* Number of entries so far */    pcre_uchar *name_table;           /* The name/number table */
2072    int  name_entry_size;         /* Size of each entry */    int  names_found;                 /* Number of entries so far */
2073    int  bracount;                /* Count of capturing parens as we compile */    int  name_entry_size;             /* Size of each entry */
2074    int  final_bracount;          /* Saved value after first pass */    int  workspace_size;              /* Size of workspace */
2075    int  top_backref;             /* Maximum back reference */    int  bracount;                    /* Count of capturing parens as we compile */
2076    unsigned int backref_map;     /* Bitmap of low back refs */    int  final_bracount;              /* Saved value after first pass */
2077    int  external_options;        /* External (initial) options */    int  max_lookbehind;              /* Maximum lookbehind (characters) */
2078    int  external_flags;          /* External flag bits to be set */    int  top_backref;                 /* Maximum back reference */
2079    int  req_varyopt;             /* "After variable item" flag for reqbyte */    unsigned int backref_map;         /* Bitmap of low back refs */
2080    BOOL had_accept;              /* (*ACCEPT) encountered */    int  assert_depth;                /* Depth of nested assertions */
2081    int  nltype;                  /* Newline type */    int  external_options;            /* External (initial) options */
2082    int  nllen;                   /* Newline string length */    int  external_flags;              /* External flag bits to be set */
2083    uschar nl[4];                 /* Newline string when fixed length */    int  req_varyopt;                 /* "After variable item" flag for reqbyte */
2084      BOOL had_accept;                  /* (*ACCEPT) encountered */
2085      BOOL had_pruneorskip;             /* (*PRUNE) or (*SKIP) encountered */
2086      BOOL check_lookbehind;            /* Lookbehinds need later checking */
2087      int  nltype;                      /* Newline type */
2088      int  nllen;                       /* Newline string length */
2089      pcre_uchar nl[4];                 /* Newline string when fixed length */
2090  } compile_data;  } compile_data;
2091    
2092  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
2093  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
2094    
2095  typedef struct branch_chain {  typedef struct branch_chain {
2096    struct branch_chain *outer;    struct branch_chain *outer;
2097    uschar *current;    pcre_uchar *current_branch;
2098  } branch_chain;  } branch_chain;
2099    
2100  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
2101  call within the pattern. */  call within the pattern; used by pcre_exec(). */
2102    
2103  typedef struct recursion_info {  typedef struct recursion_info {
2104    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
2105    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
2106    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
2107    USPTR save_start;             /* Old value of mstart */    int saved_max;                  /* Number of saved offsets */
2108    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 */  
2109  } recursion_info;  } recursion_info;
2110    
2111    /* A similar structure for pcre_dfa_exec(). */
2112    
2113    typedef struct dfa_recursion_info {
2114      struct dfa_recursion_info *prevrec;
2115      int group_num;
2116      PCRE_PUCHAR subject_position;
2117    } dfa_recursion_info;
2118    
2119  /* 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
2120  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
2121  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops; used by
2122    pcre_exec(). */
2123    
2124  typedef struct eptrblock {  typedef struct eptrblock {
2125    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
2126    USPTR epb_saved_eptr;    PCRE_PUCHAR epb_saved_eptr;
2127  } eptrblock;  } eptrblock;
2128    
2129    
# Line 984  typedef struct match_data { Line 2134  typedef struct match_data {
2134    unsigned long int match_call_count;      /* As it says */    unsigned long int match_call_count;      /* As it says */
2135    unsigned long int match_limit;           /* As it says */    unsigned long int match_limit;           /* As it says */
2136    unsigned long int match_limit_recursion; /* As it says */    unsigned long int match_limit_recursion; /* As it says */
2137    int   *offset_vector;         /* Offset vector */    int   *offset_vector;           /* Offset vector */
2138    int    offset_end;            /* One past the end */    int    offset_end;              /* One past the end */
2139    int    offset_max;            /* The maximum usable for return data */    int    offset_max;              /* The maximum usable for return data */
2140    int    nltype;                /* Newline type */    int    nltype;                  /* Newline type */
2141    int    nllen;                 /* Newline string length */    int    nllen;                   /* Newline string length */
2142    uschar nl[4];                 /* Newline string when fixed */    int    name_count;              /* Number of names in name table */
2143    const uschar *lcc;            /* Points to lower casing table */    int    name_entry_size;         /* Size of entry in names table */
2144    const uschar *ctypes;         /* Points to table of type maps */    pcre_uchar *name_table;         /* Table of names */
2145    BOOL   offset_overflow;       /* Set if too many extractions */    pcre_uchar nl[4];               /* Newline string when fixed */
2146    BOOL   notbol;                /* NOTBOL flag */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
2147    BOOL   noteol;                /* NOTEOL flag */    const  pcre_uint8 *fcc;         /* Points to case-flipping table */
2148    BOOL   utf8;                  /* UTF8 flag */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
2149    BOOL   endonly;               /* Dollar not before final \n */    BOOL   offset_overflow;         /* Set if too many extractions */
2150    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notbol;                  /* NOTBOL flag */
2151    BOOL   partial;               /* PARTIAL flag */    BOOL   noteol;                  /* NOTEOL flag */
2152    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   utf;                     /* UTF-8 / UTF-16 flag */
2153    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
2154    const uschar *start_code;     /* For use when recursing */    BOOL   use_ucp;                 /* PCRE_UCP flag */
2155    USPTR  start_subject;         /* Start of the subject string */    BOOL   endonly;                 /* Dollar not before final \n */
2156    USPTR  end_subject;           /* End of the subject string */    BOOL   notempty;                /* Empty string match not wanted */
2157    USPTR  start_match_ptr;       /* Start of matched string */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
2158    USPTR  end_match_ptr;         /* Subject position at end match */    BOOL   hitend;                  /* Hit the end of the subject at some point */
2159    int    end_offset_top;        /* Highwater mark at end of match */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
2160    int    capture_last;          /* Most recent capture number */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
2161    int    start_offset;          /* The start offset value */    BOOL   ignore_skip_arg;         /* For re-run when SKIP name not found */
2162    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    const  pcre_uchar *start_code;  /* For use when recursing */
2163    int    eptrn;                 /* Next free eptrblock */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
2164    recursion_info *recursive;    /* Linked list of recursion data */    PCRE_PUCHAR end_subject;        /* End of the subject string */
2165    void  *callout_data;          /* To pass back to callouts */    PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
2166      PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
2167      PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
2168      int    partial;                 /* PARTIAL options */
2169      int    end_offset_top;          /* Highwater mark at end of match */
2170      int    capture_last;            /* Most recent capture number */
2171      int    start_offset;            /* The start offset value */
2172      int    match_function_type;     /* Set for certain special calls of MATCH() */
2173      eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
2174      int    eptrn;                   /* Next free eptrblock */
2175      recursion_info *recursive;      /* Linked list of recursion data */
2176      void  *callout_data;            /* To pass back to callouts */
2177      const  pcre_uchar *mark;        /* Mark pointer to pass back on success */
2178      const  pcre_uchar *nomatch_mark;/* Mark pointer to pass back on failure */
2179      const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
2180    #ifdef NO_RECURSE
2181      void  *match_frames_base;       /* For remembering malloc'd frames */
2182    #endif
2183  } match_data;  } match_data;
2184    
2185  /* 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
2186  functions. */  functions. */
2187    
2188  typedef struct dfa_match_data {  typedef struct dfa_match_data {
2189    const uschar *start_code;     /* Start of the compiled pattern */    const pcre_uchar *start_code;     /* Start of the compiled pattern */
2190    const uschar *start_subject;  /* Start of the subject string */    const pcre_uchar *start_subject ; /* Start of the subject string */
2191    const uschar *end_subject;    /* End of subject string */    const pcre_uchar *end_subject;    /* End of subject string */
2192    const uschar *tables;         /* Character tables */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
2193    int   moptions;               /* Match options */    const pcre_uint8 *tables;         /* Character tables */
2194    int   poptions;               /* Pattern options */    int   start_offset;               /* The start offset value */
2195    int    nltype;                /* Newline type */    int   moptions;                   /* Match options */
2196    int    nllen;                 /* Newline string length */    int   poptions;                   /* Pattern options */
2197    uschar nl[4];                 /* Newline string when fixed */    int   nltype;                     /* Newline type */
2198    void  *callout_data;          /* To pass back to callouts */    int   nllen;                      /* Newline string length */
2199      pcre_uchar nl[4];                 /* Newline string when fixed */
2200      void *callout_data;               /* To pass back to callouts */
2201      dfa_recursion_info *recursive;    /* Linked list of recursion data */
2202  } dfa_match_data;  } dfa_match_data;
2203    
2204  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 1037  typedef struct dfa_match_data { Line 2207  typedef struct dfa_match_data {
2207  #define ctype_letter  0x02  #define ctype_letter  0x02
2208  #define ctype_digit   0x04  #define ctype_digit   0x04
2209  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
2210  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
2211  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
2212    
2213  /* 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 1064  total length. */ Line 2234  total length. */
2234  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
2235  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
2236    
2237    /* Internal function and data prefixes. */
2238    
2239    #ifdef COMPILE_PCRE8
2240    #ifndef PUBL
2241    #define PUBL(name) pcre_##name
2242    #endif
2243    #ifndef PRIV
2244    #define PRIV(name) _pcre_##name
2245    #endif
2246    #else /* COMPILE_PCRE8 */
2247    #ifdef COMPILE_PCRE16
2248    #ifndef PUBL
2249    #define PUBL(name) pcre16_##name
2250    #endif
2251    #ifndef PRIV
2252    #define PRIV(name) _pcre16_##name
2253    #endif
2254    #else
2255    #error Unsupported compiling mode
2256    #endif /* COMPILE_PCRE16 */
2257    #endif /* COMPILE_PCRE8 */
2258    
2259  /* 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
2260  codes. Each entry used to point directly to a name, but to reduce the number of  codes. Each entry used to point directly to a name, but to reduce the number of
2261  relocations in shared libraries, it now has an offset into a single string  relocations in shared libraries, it now has an offset into a single string
# Line 1081  of the exported public functions. They h Line 2273  of the exported public functions. They h
2273  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
2274  pcre_tables.c module. */  pcre_tables.c module. */
2275    
2276  extern const int    _pcre_utf8_table1[];  #ifdef COMPILE_PCRE8
2277  extern const int    _pcre_utf8_table2[];  
2278  extern const int    _pcre_utf8_table3[];  extern const int            PRIV(utf8_table1)[];
2279  extern const uschar _pcre_utf8_table4[];  extern const int            PRIV(utf8_table1_size);
2280    extern const int            PRIV(utf8_table2)[];
2281    extern const int            PRIV(utf8_table3)[];
2282    extern const pcre_uint8     PRIV(utf8_table4)[];
2283    
2284  extern const int    _pcre_utf8_table1_size;  #endif /* COMPILE_PCRE8 */
2285    
2286  extern const char   _pcre_utt_names[];  extern const char           PRIV(utt_names)[];
2287  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table PRIV(utt)[];
2288  extern const int _pcre_utt_size;  extern const int            PRIV(utt_size);
2289    
2290  extern const uschar _pcre_default_tables[];  extern const pcre_uint8     PRIV(default_tables)[];
2291    
2292  extern const uschar _pcre_OP_lengths[];  extern const pcre_uint8     PRIV(OP_lengths)[];
2293    
2294    
2295  /* Internal shared functions. These are functions that are used by more than  /* Internal shared functions. These are functions that are used by more than
2296  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
2297  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
2298    
2299  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  /* String comparison functions. */
2300                        int *, BOOL);  #ifdef COMPILE_PCRE8
2301  extern int          _pcre_ord2utf8(int, uschar *);  
2302  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  #define STRCMP_UC_UC(str1, str2) \
2303                        const pcre_study_data *, pcre_study_data *);    strcmp((char *)(str1), (char *)(str2))
2304  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  #define STRCMP_UC_C8(str1, str2) \
2305  extern unsigned int _pcre_ucp_othercase(const unsigned int);    strcmp((char *)(str1), (str2))
2306  extern int          _pcre_valid_utf8(const uschar *, int);  #define STRNCMP_UC_UC(str1, str2, num) \
2307  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,    strncmp((char *)(str1), (char *)(str2), (num))
2308                        int *, BOOL);  #define STRNCMP_UC_C8(str1, str2, num) \
2309  extern BOOL         _pcre_xclass(int, const uschar *);    strncmp((char *)(str1), (str2), (num))
2310    #define STRLEN_UC(str) strlen((const char *)str)
2311    
2312    #else
2313    
2314    extern int               PRIV(strcmp_uc_uc)(const pcre_uchar *,
2315                               const pcre_uchar *);
2316    extern int               PRIV(strcmp_uc_c8)(const pcre_uchar *,
2317                               const char *);
2318    extern int               PRIV(strncmp_uc_uc)(const pcre_uchar *,
2319                               const pcre_uchar *, unsigned int num);
2320    extern int               PRIV(strncmp_uc_c8)(const pcre_uchar *,
2321                               const char *, unsigned int num);
2322    extern unsigned int      PRIV(strlen_uc)(const pcre_uchar *str);
2323    
2324    #define STRCMP_UC_UC(str1, str2) \
2325      PRIV(strcmp_uc_uc)((str1), (str2))
2326    #define STRCMP_UC_C8(str1, str2) \
2327      PRIV(strcmp_uc_c8)((str1), (str2))
2328    #define STRNCMP_UC_UC(str1, str2, num) \
2329      PRIV(strncmp_uc_uc)((str1), (str2), (num))
2330    #define STRNCMP_UC_C8(str1, str2, num) \
2331      PRIV(strncmp_uc_c8)((str1), (str2), (num))
2332    #define STRLEN_UC(str) PRIV(strlen_uc)(str)
2333    
2334    #endif /* COMPILE_PCRE8 */
2335    
2336    extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2337    extern BOOL              PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2338                               int *, BOOL);
2339    extern int               PRIV(ord2utf)(pcre_uint32, pcre_uchar *);
2340    extern int               PRIV(valid_utf)(PCRE_PUCHAR, int, int *);
2341    extern BOOL              PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2342                               int *, BOOL);
2343    extern BOOL              PRIV(xclass)(int, const pcre_uchar *, BOOL);
2344    
2345    #ifdef SUPPORT_JIT
2346    extern void              PRIV(jit_compile)(const REAL_PCRE *,
2347                               PUBL(extra) *, int);
2348    extern int               PRIV(jit_exec)(const REAL_PCRE *, const PUBL(extra) *,
2349                               const pcre_uchar *, int, int, int, int *, int);
2350    extern void              PRIV(jit_free)(void *);
2351    extern int               PRIV(jit_get_size)(void *);
2352    extern const char*       PRIV(jit_get_target)(void);
2353    #endif
2354    
2355    /* Unicode character database (UCD) */
2356    
2357    typedef struct {
2358      pcre_uint8 script;     /* ucp_Arabic, etc. */
2359      pcre_uint8 chartype;   /* ucp_Cc, etc. (general categories) */
2360      pcre_uint8 gbprop;     /* ucp_gbControl, etc. (grapheme break property) */
2361      pcre_int32 other_case; /* offset to other case, or zero if none */
2362    } ucd_record;
2363    
2364    extern const ucd_record  PRIV(ucd_records)[];
2365    extern const pcre_uint8  PRIV(ucd_stage1)[];
2366    extern const pcre_uint16 PRIV(ucd_stage2)[];
2367    extern const int         PRIV(ucp_gentype)[];
2368    extern const pcre_uint32 PRIV(ucp_gbtable)[];
2369    #ifdef SUPPORT_JIT
2370    extern const int         PRIV(ucp_typerange)[];
2371    #endif
2372    
2373    #ifdef SUPPORT_UCP
2374    /* UCD access macros */
2375    
2376    #define UCD_BLOCK_SIZE 128
2377    #define GET_UCD(ch) (PRIV(ucd_records) + \
2378            PRIV(ucd_stage2)[PRIV(ucd_stage1)[(ch) / UCD_BLOCK_SIZE] * \
2379            UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
2380    
2381    #define UCD_CHARTYPE(ch)    GET_UCD(ch)->chartype
2382    #define UCD_SCRIPT(ch)      GET_UCD(ch)->script
2383    #define UCD_CATEGORY(ch)    PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2384    #define UCD_GRAPHBREAK(ch)  GET_UCD(ch)->gbprop
2385    #define UCD_OTHERCASE(ch)   (ch + GET_UCD(ch)->other_case)
2386    
2387    #endif /* SUPPORT_UCP */
2388    
2389  #endif  #endif
2390    

Legend:
Removed from v.270  
changed lines
  Added in v.1028

  ViewVC Help
Powered by ViewVC 1.1.5