/[pcre]/code/branches/pcre16/pcre_internal.h
ViewVC logotype

Diff of /code/branches/pcre16/pcre_internal.h

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

code/trunk/pcre_internal.h revision 483 by ph10, Mon Jan 4 16:19:17 2010 UTC code/branches/pcre16/pcre_internal.h revision 759 by zherczeg, Mon Nov 21 18:57:34 2011 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) 1997-2010 University of Cambridge             Copyright (c) 1997-2011 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 51  functions whose names all begin with "_p Line 51  functions whose names all begin with "_p
51  #define PCRE_DEBUG  #define PCRE_DEBUG
52  #endif  #endif
53    
54  /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"  /* We do not support both EBCDIC and UTF-8/16 at the same time. The "configure"
55  script prevents both being selected, but not everybody uses "configure". */  script prevents both being selected, but not everybody uses "configure". */
56    
57  #if defined EBCDIC && defined SUPPORT_UTF8  #if defined EBCDIC && (defined SUPPORT_UTF8 || defined SUPPORT_UTF16)
58  #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.  #error The use of both EBCDIC and SUPPORT_UTF8/16 is not supported.
59  #endif  #endif
60    
61  /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The  /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
# Line 158  set, we ensure here that it has no effec Line 158  set, we ensure here that it has no effec
158  #define PCRE_CALL_CONVENTION  #define PCRE_CALL_CONVENTION
159  #endif  #endif
160    
161  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 8, 16 and 32-bit integers. We
162  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
163  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
164  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
165  preprocessor time in standard C environments. */  preprocessor time in standard C environments. */
166    
167    typedef unsigned char pcre_uint8;
168    
169  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
170    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
171    typedef short pcre_int16;    typedef short pcre_int16;
# Line 184  preprocessor time in standard C environm Line 186  preprocessor time in standard C environm
186    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
187  #endif  #endif
188    
189  /* When checking for integer overflow in pcre_compile(), we need to handle  /* When checking for integer overflow in pcre_compile(), we need to handle
190  large integers. If a 64-bit integer type is available, we can use that.  large integers. If a 64-bit integer type is available, we can use that.
191  Otherwise we have to cast to double, which of course requires floating point  Otherwise we have to cast to double, which of course requires floating point
192  arithmetic. Handle this by defining a macro for the appropriate type. If  arithmetic. Handle this by defining a macro for the appropriate type. If
193  stdint.h is available, include it; it may define INT64_MAX. The macro int64_t  stdint.h is available, include it; it may define INT64_MAX. Systems that do not
194  may be set by "configure". */  have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
195    by "configure". */
196    
197  #if HAVE_STDINT_H  #if HAVE_STDINT_H
198  #include <stdint.h>  #include <stdint.h>
199    #elif HAVE_INTTYPES_H
200    #include <inttypes.h>
201  #endif  #endif
202    
203  #if defined INT64_MAX || defined int64_t  #if defined INT64_MAX || defined int64_t
# Line 203  may be set by "configure". */ Line 208  may be set by "configure". */
208    
209  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
210  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
211  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 *,
212  should make things easier for callers. We define a short type for unsigned char  because that should make things easier for callers. We define a short type
213  to save lots of typing. I tried "uchar", but it causes problems on Digital  for the current character representation (either 8 or 16 bit) to save lots
214  Unix, where it is defined in sys/types, so use "uschar" instead. */  of typing. I tried "uchar", but it causes problems on Digital Unix, where
215    it is defined in sys/types, so use "uschar" instead. */
216    
217  typedef unsigned char uschar;  #ifndef COMPILE_PCRE16
218    typedef unsigned char pcre_uchar;
219    #else
220    #if USHRT_MAX != 65535
221    /* This is a warning message. Change PCRE_SCHAR16 to a 16 bit data type in
222    pcre.h(.in) and disable (comment out) this message. */
223    #error Warning: PCRE_SCHAR16 is not a 16 bit data type.
224    #endif
225    typedef pcre_uint16 uschar;
226    #endif
227    
228  /* 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
229  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 264  used for the external interface and appe Line 279  used for the external interface and appe
279  must begin with PCRE_. */  must begin with PCRE_. */
280    
281  #ifdef CUSTOM_SUBJECT_PTR  #ifdef CUSTOM_SUBJECT_PTR
282  #define PCRE_SPTR CUSTOM_SUBJECT_PTR  #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
 #define USPTR CUSTOM_SUBJECT_PTR  
283  #else  #else
284  #define PCRE_SPTR const char *  #define PCRE_PUCHAR const pcre_uchar *
 #define USPTR const unsigned char *  
 #endif  
   
285    
286    /* PCRE_SPTR is defined in pcre.h. */
287    #define USPTR const uschar *
288    #endif
289    
290  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
291  values. */  values. */
# Line 405  capturing parenthesis numbers in back re Line 419  capturing parenthesis numbers in back re
419    
420  /* When UTF-8 encoding is being used, a character is no longer just a single  /* When UTF-8 encoding is being used, a character is no longer just a single
421  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
422  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should  byte-mode, and more complicated ones for UTF-8 characters. GETCHARLENTEST is
423  never be called in byte mode. To make sure it can never even appear when UTF-8  not used when UTF-8 is not supported, so it is not defined, and BACKCHAR should
424  support is omitted, we don't even define it. */  never be called in byte mode. To make sure they can never even appear when
425    UTF-8 support is omitted, we don't even define them. */
426    
427  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
428  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 415  support is omitted, we don't even define Line 430  support is omitted, we don't even define
430  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
431  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
432  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
433    /* #define GETCHARLENTEST(c, eptr, len) */
434  /* #define BACKCHAR(eptr) */  /* #define BACKCHAR(eptr) */
435    
436  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
437    
438    /* These macros were originally written in the form of loops that used data
439    from the tables whose names start with _pcre_utf8_table. They were rewritten by
440    a user so as not to use loops, because in some environments this gives a
441    significant performance advantage, and it seems never to do any harm. */
442    
443    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
444    advancing the pointer. */
445    
446    #define GETUTF8(c, eptr) \
447        { \
448        if ((c & 0x20) == 0) \
449          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
450        else if ((c & 0x10) == 0) \
451          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
452        else if ((c & 0x08) == 0) \
453          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
454          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
455        else if ((c & 0x04) == 0) \
456          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
457              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
458              (eptr[4] & 0x3f); \
459        else \
460          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
461              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
462              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
463        }
464    
465  /* 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
466  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
467    
468  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
469    c = *eptr; \    c = *eptr; \
470    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; \  
       } \  
     }  
471    
472  /* 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
473  pointer. */  pointer. */
474    
475  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
476    c = *eptr; \    c = *eptr; \
477    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8(c, eptr);
478    
479    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
480    the pointer. */
481    
482    #define GETUTF8INC(c, eptr) \
483      { \      { \
484      int gcii; \      if ((c & 0x20) == 0) \
485      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
486      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     for (gcii = 1; gcii <= gcaa; gcii++) \  
487        { \        { \
488        gcss -= 6; \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
489        c |= (eptr[gcii] & 0x3f) << gcss; \        eptr += 2; \
490          } \
491        else if ((c & 0x08) == 0) \
492          { \
493          c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
494              ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
495          eptr += 3; \
496          } \
497        else if ((c & 0x04) == 0) \
498          { \
499          c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
500              ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
501              (eptr[3] & 0x3f); \
502          eptr += 4; \
503          } \
504        else \
505          { \
506          c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
507              ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
508              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
509          eptr += 5; \
510        } \        } \
511      }      }
512    
# Line 460  know we are in UTF-8 mode. */ Line 515  know we are in UTF-8 mode. */
515    
516  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
517    c = *eptr++; \    c = *eptr++; \
518    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; \  
       } \  
     }  
519    
520  /* 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.
521    This is called when we don't know if we are in UTF-8 mode. */
522    
523  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
524    c = *eptr++; \    c = *eptr++; \
525    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8INC(c, eptr);
526    
527    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
528    advancing the pointer, incrementing the length. */
529    
530    #define GETUTF8LEN(c, eptr, len) \
531      { \      { \
532      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
533      int gcss = 6*gcaa; \        { \
534      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
535      while (gcaa-- > 0) \        len++; \
536          } \
537        else if ((c & 0x10)  == 0) \
538          { \
539          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
540          len += 2; \
541          } \
542        else if ((c & 0x08)  == 0) \
543          {\
544          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
545              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
546          len += 3; \
547          } \
548        else if ((c & 0x04)  == 0) \
549        { \        { \
550        gcss -= 6; \        c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
551        c |= (*eptr++ & 0x3f) << gcss; \            ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
552              (eptr[4] & 0x3f); \
553          len += 4; \
554          } \
555        else \
556          {\
557          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
558              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
559              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
560          len += 5; \
561        } \        } \
562      }      }
563    
# Line 493  if there are extra bytes. This is called Line 566  if there are extra bytes. This is called
566    
567  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
568    c = *eptr; \    c = *eptr; \
569    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
     { \  
     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; \  
       } \  
     len += gcaa; \  
     }  
570    
571  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
572  pointer, incrementing length if there are extra bytes. This is called when we  pointer, incrementing length if there are extra bytes. This is called when we
573  know we are in UTF-8 mode. */  do not know if we are in UTF-8 mode. */
574    
575  #define GETCHARLENTEST(c, eptr, len) \  #define GETCHARLENTEST(c, eptr, len) \
576    c = *eptr; \    c = *eptr; \
577    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8LEN(c, eptr, len);
     { \  
     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; \  
       } \  
     len += gcaa; \  
     }  
578    
579  /* 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
580  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 533  because almost all calls are already wit Line 582  because almost all calls are already wit
582    
583  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
584    
585  #endif  #endif  /* SUPPORT_UTF8 */
586    
587    
588  /* In case there is no definition of offsetof() provided - though any proper  /* In case there is no definition of offsetof() provided - though any proper
# Line 544  Standard C system should have one. */ Line 593  Standard C system should have one. */
593  #endif  #endif
594    
595    
 /* These are the public options that can change during matching. */  
   
 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  
   
596  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
597  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
598  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
# Line 560  compatibility. */ Line 605  compatibility. */
605  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
606  #define PCRE_JCHANGED      0x0010  /* j option used in regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
607  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
608    #define PCRE_HASTHEN       0x0040  /* pattern contains (*THEN) */
609    
610  /* Options for the "extra" block produced by pcre_study(). */  /* Flags for the "extra" block produced by pcre_study(). */
611    
612  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
613  #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */  #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
614    
615  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
616  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 577  time, run time, or study time, respectiv Line 623  time, run time, or study time, respectiv
623     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
624     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| \
625     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
626     PCRE_JAVASCRIPT_COMPAT)     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
627    
628  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
629    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
# Line 590  time, run time, or study time, respectiv Line 636  time, run time, or study time, respectiv
636     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
637     PCRE_NO_START_OPTIMIZE)     PCRE_NO_START_OPTIMIZE)
638    
639  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS \
640       PCRE_STUDY_JIT_COMPILE
641    
642  /* Magic number to provide a small check against being handed junk. Also used  /* Magic number to provide a small check against being handed junk. Also used
643  to detect whether a pattern was compiled on a host of different endianness. */  to detect whether a pattern was compiled on a host of different endianness. */
# Line 872  so that PCRE works on both ASCII and EBC Line 919  so that PCRE works on both ASCII and EBC
919  #define STRING_COMMIT0              "COMMIT\0"  #define STRING_COMMIT0              "COMMIT\0"
920  #define STRING_F0                   "F\0"  #define STRING_F0                   "F\0"
921  #define STRING_FAIL0                "FAIL\0"  #define STRING_FAIL0                "FAIL\0"
922    #define STRING_MARK0                "MARK\0"
923  #define STRING_PRUNE0               "PRUNE\0"  #define STRING_PRUNE0               "PRUNE\0"
924  #define STRING_SKIP0                "SKIP\0"  #define STRING_SKIP0                "SKIP\0"
925  #define STRING_THEN                 "THEN"  #define STRING_THEN                 "THEN"
# Line 893  so that PCRE works on both ASCII and EBC Line 941  so that PCRE works on both ASCII and EBC
941    
942  #define STRING_DEFINE               "DEFINE"  #define STRING_DEFINE               "DEFINE"
943    
944  #define STRING_CR_RIGHTPAR          "CR)"  #define STRING_CR_RIGHTPAR             "CR)"
945  #define STRING_LF_RIGHTPAR          "LF)"  #define STRING_LF_RIGHTPAR             "LF)"
946  #define STRING_CRLF_RIGHTPAR        "CRLF)"  #define STRING_CRLF_RIGHTPAR           "CRLF)"
947  #define STRING_ANY_RIGHTPAR         "ANY)"  #define STRING_ANY_RIGHTPAR            "ANY)"
948  #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"  #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
949  #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"  #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
950  #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"  #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
951  #define STRING_UTF8_RIGHTPAR        "UTF8)"  #define STRING_UTF8_RIGHTPAR           "UTF8)"
952    #define STRING_UCP_RIGHTPAR            "UCP)"
953    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
954    
955  #else  /* SUPPORT_UTF8 */  #else  /* SUPPORT_UTF8 */
956    
# Line 1124  only. */ Line 1174  only. */
1174  #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"  #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1175  #define STRING_F0                   STR_F "\0"  #define STRING_F0                   STR_F "\0"
1176  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"  #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1177    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1178  #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"  #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1179  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"  #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1180  #define STRING_THEN                 STR_T STR_H STR_E STR_N  #define STRING_THEN                 STR_T STR_H STR_E STR_N
# Line 1145  only. */ Line 1196  only. */
1196    
1197  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E  #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1198    
1199  #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS  #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1200  #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1201  #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1202  #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS  #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1203  #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS  #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1204  #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  #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
1205  #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  #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
1206  #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS  #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1207    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1208    #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
1209    
1210  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 */
1211    
# Line 1185  only. */ Line 1238  only. */
1238    
1239  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1240  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1241  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1242  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1243  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1244    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1245    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1246    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1247    #define PT_WORD       8    /* Word - L plus N plus underscore */
1248    
1249  /* 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
1250  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 1204  contain UTF-8 characters with values gre Line 1261  contain UTF-8 characters with values gre
1261  /* 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
1262  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
1263  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
1264  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
1265  corresponds to "." rather than an escape sequence, and another for OP_ALLANY  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1266  (which is used for [^] in JavaScript compatibility mode).  used for [^] in JavaScript compatibility mode, and for \C in non-utf8 mode. In
1267    non-DOTALL mode, "." behaves like \N.
1268    
1269    The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1270    when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1271    They must be contiguous, and remain in order so that the replacements can be
1272    looked up from a table.
1273    
1274  The final escape must be ESC_REF as subsequent values are used for  The final escape must be ESC_REF as subsequent values are used for
1275  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
# Line 1216  put in between that don't consume a char Line 1279  put in between that don't consume a char
1279  */  */
1280    
1281  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,
1282         ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,         ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1283         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1284           ESC_E, ESC_Q, ESC_g, ESC_k,
1285           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1286         ESC_REF };         ESC_REF };
1287    
   
1288  /* 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
1289  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.
1290    
# Line 1244  enum { Line 1308  enum {
1308    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1309    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1310    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1311    OP_ANY,            /* 12 Match any character (subject to DOTALL) */  
1312    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */    OP_ANY,            /* 12 Match any character except newline */
1313      OP_ALLANY,         /* 13 Match any character */
1314    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1315    OP_NOTPROP,        /* 15 \P (not Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1316    OP_PROP,           /* 16 \p (Unicode property) */    OP_PROP,           /* 16 \p (Unicode property) */
# Line 1258  enum { Line 1323  enum {
1323    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1324    OP_EOD,            /* 24 End of data: \z */    OP_EOD,            /* 24 End of data: \z */
1325    
1326    OP_OPT,            /* 25 Set runtime options */    OP_CIRC,           /* 25 Start of line - not multiline */
1327    OP_CIRC,           /* 26 Start of line - varies with multiline switch */    OP_CIRCM,          /* 26 Start of line - multiline */
1328    OP_DOLL,           /* 27 End of line - varies with multiline switch */    OP_DOLL,           /* 27 End of line - not multiline */
1329    OP_CHAR,           /* 28 Match one character, casefully */    OP_DOLLM,          /* 28 End of line - multiline */
1330    OP_CHARNC,         /* 29 Match one character, caselessly */    OP_CHAR,           /* 29 Match one character, casefully */
1331    OP_NOT,            /* 30 Match one character, not the following one */    OP_CHARI,          /* 30 Match one character, caselessly */
1332      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1333    OP_STAR,           /* 31 The maximizing and minimizing versions of */    OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1334    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */  
1335    OP_PLUS,           /* 33 the minimizing one second. */    /* The following sets of 13 opcodes must always be kept in step because
1336    OP_MINPLUS,        /* 34 This first set applies to single characters.*/    the offset from the first one is used to generate the others. */
1337    OP_QUERY,          /* 35 */  
1338    OP_MINQUERY,       /* 36 */    /**** Single characters, caseful, must precede the caseless ones ****/
1339    
1340    OP_UPTO,           /* 37 From 0 to n matches */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1341    OP_MINUPTO,        /* 38 */    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1342    OP_EXACT,          /* 39 Exactly n matches */    OP_PLUS,           /* 35 the minimizing one second. */
1343      OP_MINPLUS,        /* 36 */
1344    OP_POSSTAR,        /* 40 Possessified star */    OP_QUERY,          /* 37 */
1345    OP_POSPLUS,        /* 41 Possessified plus */    OP_MINQUERY,       /* 38 */
1346    OP_POSQUERY,       /* 42 Posesssified query */  
1347    OP_POSUPTO,        /* 43 Possessified upto */    OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1348      OP_MINUPTO,        /* 40 */
1349    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */    OP_EXACT,          /* 41 Exactly n matches */
1350    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */  
1351    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */    OP_POSSTAR,        /* 42 Possessified star, caseful */
1352    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */    OP_POSPLUS,        /* 43 Possessified plus, caseful */
1353    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
1354    OP_NOTMINQUERY,    /* 49 */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
1355    
1356    OP_NOTUPTO,        /* 50 From 0 to n matches */    /**** Single characters, caseless, must follow the caseful ones */
1357    OP_NOTMINUPTO,     /* 51 */  
1358    OP_NOTEXACT,       /* 52 Exactly n matches */    OP_STARI,          /* 46 */
1359      OP_MINSTARI,       /* 47 */
1360    OP_NOTPOSSTAR,     /* 53 Possessified versions */    OP_PLUSI,          /* 48 */
1361    OP_NOTPOSPLUS,     /* 54 */    OP_MINPLUSI,       /* 49 */
1362    OP_NOTPOSQUERY,    /* 55 */    OP_QUERYI,         /* 50 */
1363    OP_NOTPOSUPTO,     /* 56 */    OP_MINQUERYI,      /* 51 */
1364    
1365    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */    OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1366    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */    OP_MINUPTOI,       /* 53 */
1367    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */    OP_EXACTI,         /* 54 */
1368    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */  
1369    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */    OP_POSSTARI,       /* 55 Possessified star, caseless */
1370    OP_TYPEMINQUERY,   /* 62 */    OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1371      OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1372    OP_TYPEUPTO,       /* 63 From 0 to n matches */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1373    OP_TYPEMINUPTO,    /* 64 */  
1374    OP_TYPEEXACT,      /* 65 Exactly n matches */    /**** The negated ones must follow the non-negated ones, and match them ****/
1375      /**** Negated single character, caseful; must precede the caseless ones ****/
1376    OP_TYPEPOSSTAR,    /* 66 Possessified versions */  
1377    OP_TYPEPOSPLUS,    /* 67 */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1378    OP_TYPEPOSQUERY,   /* 68 */    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1379    OP_TYPEPOSUPTO,    /* 69 */    OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1380      OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1381    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */    OP_NOTQUERY,       /* 63 */
1382    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */    OP_NOTMINQUERY,    /* 64 */
1383    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */  
1384    OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */    OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1385    OP_CRQUERY,        /* 74 These are for character classes and back refs */    OP_NOTMINUPTO,     /* 66 */
1386    OP_CRMINQUERY,     /* 75 */    OP_NOTEXACT,       /* 67 Exactly n matches */
1387    OP_CRRANGE,        /* 76 These are different to the three sets above. */  
1388    OP_CRMINRANGE,     /* 77 */    OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1389      OP_NOTPOSPLUS,     /* 69 */
1390    OP_CLASS,          /* 78 Match a character class, chars < 256 only */    OP_NOTPOSQUERY,    /* 70 */
1391    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative    OP_NOTPOSUPTO,     /* 71 */
1392                             class - the difference is relevant only when a UTF-8  
1393                             character > 255 is encountered. */    /**** Negated single character, caseless; must follow the caseful ones ****/
1394    
1395    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the    OP_NOTSTARI,       /* 72 */
1396                             class. This does both positive and negative. */    OP_NOTMINSTARI,    /* 73 */
1397      OP_NOTPLUSI,       /* 74 */
1398    OP_REF,            /* 81 Match a back reference */    OP_NOTMINPLUSI,    /* 75 */
1399    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */    OP_NOTQUERYI,      /* 76 */
1400    OP_CALLOUT,        /* 83 Call out to external function if provided */    OP_NOTMINQUERYI,   /* 77 */
1401    
1402    OP_ALT,            /* 84 Start of alternation */    OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1403    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */    OP_NOTMINUPTOI,    /* 79 */
1404    OP_KETRMAX,        /* 86 These two must remain together and in this */    OP_NOTEXACTI,      /* 80 Exactly n matches */
1405    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */  
1406      OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1407    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    OP_NOTPOSPLUSI,    /* 82 */
1408      OP_NOTPOSQUERYI,   /* 83 */
1409    OP_ASSERT,         /* 88 Positive lookahead */    OP_NOTPOSUPTOI,    /* 84 */
1410    OP_ASSERT_NOT,     /* 89 Negative lookahead */  
1411    OP_ASSERTBACK,     /* 90 Positive lookbehind */    /**** Character types ****/
1412    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */  
1413    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1414      OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1415    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1416    as there's a test for >= ONCE for a subpattern that isn't an assertion. */    OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
1417      OP_TYPEQUERY,      /* 89 */
1418    OP_ONCE,           /* 93 Atomic group */    OP_TYPEMINQUERY,   /* 90 */
1419    OP_BRA,            /* 94 Start of non-capturing bracket */  
1420    OP_CBRA,           /* 95 Start of capturing bracket */    OP_TYPEUPTO,       /* 91 From 0 to n matches */
1421    OP_COND,           /* 96 Conditional group */    OP_TYPEMINUPTO,    /* 92 */
1422      OP_TYPEEXACT,      /* 93 Exactly n matches */
1423    
1424      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1425      OP_TYPEPOSPLUS,    /* 95 */
1426      OP_TYPEPOSQUERY,   /* 96 */
1427      OP_TYPEPOSUPTO,    /* 97 */
1428    
1429      /* These are used for character classes and back references; only the
1430      first six are the same as the sets above. */
1431    
1432      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1433      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1434      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1435      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1436      OP_CRQUERY,        /* 102 */
1437      OP_CRMINQUERY,     /* 103 */
1438    
1439      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1440      OP_CRMINRANGE,     /* 105 */
1441    
1442      /* End of quantifier opcodes */
1443    
1444      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1445      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1446                                  class - the difference is relevant only when a
1447                                  UTF-8 character > 255 is encountered. */
1448      OP_XCLASS,         /* 108 Extended class for handling UTF-8 chars within the
1449                                  class. This does both positive and negative. */
1450      OP_REF,            /* 109 Match a back reference, casefully */
1451      OP_REFI,           /* 110 Match a back reference, caselessly */
1452      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1453      OP_CALLOUT,        /* 112 Call out to external function if provided */
1454    
1455      OP_ALT,            /* 113 Start of alternation */
1456      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1457      OP_KETRMAX,        /* 115 These two must remain together and in this */
1458      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1459      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1460    
1461      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1462      asserts must remain in order. */
1463    
1464      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1465      OP_ASSERT,         /* 119 Positive lookahead */
1466      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1467      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1468      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1469    
1470      /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
1471      after the assertions, with ONCE first, as there's a test for >= ONCE for a
1472      subpattern that isn't an assertion. The POS versions must immediately follow
1473      the non-POS versions in each case. */
1474    
1475      OP_ONCE,           /* 123 Atomic group, contains captures */
1476      OP_ONCE_NC,        /* 124 Atomic group containing no captures */
1477      OP_BRA,            /* 125 Start of non-capturing bracket */
1478      OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
1479      OP_CBRA,           /* 127 Start of capturing bracket */
1480      OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
1481      OP_COND,           /* 129 Conditional group */
1482    
1483    /* These three must follow the previous three, in the same order. There's a    /* These five must follow the previous five, in the same order. There's a
1484    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1485    
1486    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
1487    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
1488    OP_SCOND,          /* 99 Conditional group, check empty */    OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
1489      OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
1490      OP_SCOND,          /* 134 Conditional group, check empty */
1491    
1492    /* The next two pairs must (respectively) be kept together. */    /* The next two pairs must (respectively) be kept together. */
1493    
1494    OP_CREF,           /* 100 Used to hold a capture number as condition */    OP_CREF,           /* 135 Used to hold a capture number as condition */
1495    OP_NCREF,          /* 101 Same, but generaged by a name reference*/    OP_NCREF,          /* 136 Same, but generated by a name reference*/
1496    OP_RREF,           /* 102 Used to hold a recursion number as condition */    OP_RREF,           /* 137 Used to hold a recursion number as condition */
1497    OP_NRREF,          /* 103 Same, but generaged by a name reference*/    OP_NRREF,          /* 138 Same, but generated by a name reference*/
1498    OP_DEF,            /* 104 The DEFINE condition */    OP_DEF,            /* 139 The DEFINE condition */
1499    
1500    OP_BRAZERO,        /* 105 These two must remain together and in this */    OP_BRAZERO,        /* 140 These two must remain together and in this */
1501    OP_BRAMINZERO,     /* 106 order. */    OP_BRAMINZERO,     /* 141 order. */
1502      OP_BRAPOSZERO,     /* 142 */
1503    
1504    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1505    
1506    OP_PRUNE,          /* 107 */    OP_MARK,           /* 143 always has an argument */
1507    OP_SKIP,           /* 108 */    OP_PRUNE,          /* 144 */
1508    OP_THEN,           /* 109 */    OP_PRUNE_ARG,      /* 145 same, but with argument */
1509    OP_COMMIT,         /* 110 */    OP_SKIP,           /* 146 */
1510      OP_SKIP_ARG,       /* 147 same, but with argument */
1511      OP_THEN,           /* 148 */
1512      OP_THEN_ARG,       /* 149 same, but with argument */
1513      OP_COMMIT,         /* 150 */
1514    
1515    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1516    
1517    OP_FAIL,           /* 111 */    OP_FAIL,           /* 151 */
1518    OP_ACCEPT,         /* 112 */    OP_ACCEPT,         /* 152 */
1519    OP_CLOSE,          /* 113 Used before OP_ACCEPT to close open captures */    OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
1520      OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
1521    
1522    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1523    
1524    OP_SKIPZERO        /* 114 */    OP_SKIPZERO,       /* 155 */
1525    
1526      /* This is not an opcode, but is used to check that tables indexed by opcode
1527      are the correct length, in order to catch updating errors - there have been
1528      some in the past. */
1529    
1530      OP_TABLE_LENGTH
1531  };  };
1532    
1533  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1534  definitions that follow must also be updated to match. There are also tables  definitions that follow must also be updated to match. There are also tables
1535  called "coptable" cna "poptable" in pcre_dfa_exec.c that must be updated. */  called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1536    
1537    
1538  /* 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
1539  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
1540    only in pcre_printint.c, which fills out the full names in many cases (and in
1541    some cases doesn't actually use these names at all). */
1542    
1543  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1544    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1545    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1546    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1547    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1548    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1549    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1550      "{", "{", "{",                                                  \
1551    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1552    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1553      "{", "{", "{",                                                  \
1554      "*+","++", "?+", "{",                                           \
1555      "*", "*?", "+", "+?", "?", "??",                                \
1556      "{", "{", "{",                                                  \
1557      "*+","++", "?+", "{",                                           \
1558      "*", "*?", "+", "+?", "?", "??",                                \
1559      "{", "{", "{",                                                  \
1560    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1561    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1562    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1563    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1564    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1565    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1566    "AssertB", "AssertB not", "Reverse",                            \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1567    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1568      "Once", "Once_NC",                                              \
1569      "Bra", "BraPos", "CBra", "CBraPos",                             \
1570      "Cond",                                                         \
1571      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1572      "SCond",                                                        \
1573    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1574    "Brazero", "Braminzero",                                        \    "Brazero", "Braminzero", "Braposzero",                          \
1575    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1576      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1577      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1578    "Close", "Skip zero"    "Close", "Skip zero"
1579    
1580    
# Line 1436  in UTF-8 mode. The code that uses this t Line 1592  in UTF-8 mode. The code that uses this t
1592    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1593    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1594    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1595    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1596    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1597    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1,                             /* \X                                     */ \
1598      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1599    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1600    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1601    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1602    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1603    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1604    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1605      4, 4, 4,                       /* upto, minupto, exact      ** mode      */ \
1606    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
1607      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1608      4, 4, 4,                       /* upto I, minupto I, exact I             */ \
1609      2, 2, 2, 4,                    /* *+I, ++I, ?+I, upto+I                  */ \
1610    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1611    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1612    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
1613    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2, 2, 2, 4,                    /* Possessive NOT *, +, ?, upto           */ \
1614      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1615      4, 4, 4,                       /* NOT upto I, minupto I, exact I         */ \
1616      2, 2, 2, 4,                    /* Possessive NOT *I, +I, ?I, upto I      */ \
1617    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1618    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1619    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
# Line 1461  in UTF-8 mode. The code that uses this t Line 1625  in UTF-8 mode. The code that uses this t
1625   33,                             /* NCLASS                                 */ \   33,                             /* NCLASS                                 */ \
1626    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1627    3,                             /* REF                                    */ \    3,                             /* REF                                    */ \
1628      3,                             /* REFI                                   */ \
1629    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1630    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1631    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
1632    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
1633    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
1634    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
1635      1+LINK_SIZE,                   /* KetRpos                                */ \
1636      1+LINK_SIZE,                   /* Reverse                                */ \
1637    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
1638    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
1639    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1640    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
   1+LINK_SIZE,                   /* Reverse                                */ \  
1641    1+LINK_SIZE,                   /* ONCE                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1642      1+LINK_SIZE,                   /* ONCE_NC                                */ \
1643    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
1644      1+LINK_SIZE,                   /* BRAPOS                                 */ \
1645    3+LINK_SIZE,                   /* CBRA                                   */ \    3+LINK_SIZE,                   /* CBRA                                   */ \
1646      3+LINK_SIZE,                   /* CBRAPOS                                */ \
1647    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1648    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1649      1+LINK_SIZE,                   /* SBRAPOS                                */ \
1650    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1651      3+LINK_SIZE,                   /* SCBRAPOS                               */ \
1652    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1653    3, 3,                          /* CREF, NCREF                            */ \    3, 3,                          /* CREF, NCREF                            */ \
1654    3, 3,                          /* RREF, NRREF                            */ \    3, 3,                          /* RREF, NRREF                            */ \
1655    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1656    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
1657    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1658    1, 1, 3, 1                     /* FAIL, ACCEPT, CLOSE, SKIPZERO          */    1, 3,                          /* SKIP, SKIP_ARG                         */ \
1659      1, 3,                          /* THEN, THEN_ARG                         */ \
1660      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
1661      3, 1                           /* CLOSE, SKIPZERO  */
1662    
1663  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1664  condition. */  condition. */
1665    
1666  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1667    
1668  /* 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
1669  tracked. */  easily be tracked. When a new number is added, the table called eint in
1670    pcreposix.c must be updated. */
1671    
1672  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1673         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
# Line 1501  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 1675  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1675         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1676         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1677         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1678         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
1679           ERR70, ERRCOUNT };
1680    
1681  /* 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
1682  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 1535  typedef struct real_pcre { Line 1710  typedef struct real_pcre {
1710    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
1711    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
1712    
1713    const unsigned char *tables;    /* Pointer to tables or NULL for std */    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
1714    const unsigned char *nullpad;   /* NULL padding */    const pcre_uint8 *nullpad;      /* NULL padding */
1715  } real_pcre;  } real_pcre;
1716    
1717  /* The format of the block used to store data from pcre_study(). The same  /* The format of the block used to store data from pcre_study(). The same
# Line 1545  remark (see NOTE above) about extending Line 1720  remark (see NOTE above) about extending
1720  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1721    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1722    pcre_uint32 flags;              /* Private flags */    pcre_uint32 flags;              /* Private flags */
1723    uschar start_bits[32];          /* Starting char bits */    pcre_uint8 start_bits[32];      /* Starting char bits */
1724    pcre_uint32 minlength;          /* Minimum subject length */    pcre_uint32 minlength;          /* Minimum subject length */
1725  } pcre_study_data;  } pcre_study_data;
1726    
1727  /* Structure for building a chain of open capturing subpatterns during  /* Structure for building a chain of open capturing subpatterns during
1728  compiling, so that instructions to close them can be compiled when (*ACCEPT) is  compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1729  encountered. */  encountered. This is also used to identify subpatterns that contain recursive
1730    back references to themselves, so that they can be made atomic. */
1731    
1732  typedef struct open_capitem {  typedef struct open_capitem {
1733    struct open_capitem *next;    /* Chain link */    struct open_capitem *next;    /* Chain link */
1734    pcre_uint16 number;           /* Capture number */    pcre_uint16 number;           /* Capture number */
1735      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1736  } open_capitem;  } open_capitem;
1737    
1738  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1739  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1740    
1741  typedef struct compile_data {  typedef struct compile_data {
1742    const uschar *lcc;            /* Points to lower casing table */    const pcre_uint8 *lcc;            /* Points to lower casing table */
1743    const uschar *fcc;            /* Points to case-flipping table */    const pcre_uint8 *fcc;            /* Points to case-flipping table */
1744    const uschar *cbits;          /* Points to character type table */    const pcre_uint8 *cbits;          /* Points to character type table */
1745    const uschar *ctypes;         /* Points to table of type maps */    const pcre_uint8 *ctypes;         /* Points to table of type maps */
1746    const uschar *start_workspace;/* The start of working space */    const pcre_uchar *start_workspace;/* The start of working space */
1747    const uschar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
1748    const uschar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
1749    const uschar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
1750    open_capitem *open_caps;      /* Chain of open capture items */    open_capitem *open_caps;          /* Chain of open capture items */
1751    uschar *hwm;                  /* High watermark of workspace */    pcre_uchar *hwm;                  /* High watermark of workspace */
1752    uschar *name_table;           /* The name/number table */    pcre_uchar *name_table;           /* The name/number table */
1753    int  names_found;             /* Number of entries so far */    int  names_found;                 /* Number of entries so far */
1754    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;             /* Size of each entry */
1755    int  bracount;                /* Count of capturing parens as we compile */    int  bracount;                    /* Count of capturing parens as we compile */
1756    int  final_bracount;          /* Saved value after first pass */    int  final_bracount;              /* Saved value after first pass */
1757    int  top_backref;             /* Maximum back reference */    int  top_backref;                 /* Maximum back reference */
1758    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;         /* Bitmap of low back refs */
1759    int  external_options;        /* External (initial) options */    int  assert_depth;                /* Depth of nested assertions */
1760    int  external_flags;          /* External flag bits to be set */    int  external_options;            /* External (initial) options */
1761    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  external_flags;              /* External flag bits to be set */
1762    BOOL had_accept;              /* (*ACCEPT) encountered */    int  req_varyopt;                 /* "After variable item" flag for reqbyte */
1763    BOOL check_lookbehind;        /* Lookbehinds need later checking */    BOOL had_accept;                  /* (*ACCEPT) encountered */
1764    int  nltype;                  /* Newline type */    BOOL check_lookbehind;            /* Lookbehinds need later checking */
1765    int  nllen;                   /* Newline string length */    int  nltype;                      /* Newline type */
1766    uschar nl[4];                 /* Newline string when fixed length */    int  nllen;                       /* Newline string length */
1767      pcre_uchar nl[4];                 /* Newline string when fixed length */
1768  } compile_data;  } compile_data;
1769    
1770  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
1771  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
1772    
1773  typedef struct branch_chain {  typedef struct branch_chain {
1774    struct branch_chain *outer;    struct branch_chain *outer;
1775    uschar *current_branch;    pcre_uchar *current_branch;
1776  } branch_chain;  } branch_chain;
1777    
1778  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
1779  call within the pattern. */  call within the pattern; used by pcre_exec(). */
1780    
1781  typedef struct recursion_info {  typedef struct recursion_info {
1782    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1783    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
1784    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
1785    USPTR save_start;             /* Old value of mstart */    int saved_max;                  /* Number of saved offsets */
1786    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 */  
   int save_offset_top;          /* Current value of offset_top */  
1787  } recursion_info;  } recursion_info;
1788    
1789    /* A similar structure for pcre_dfa_exec(). */
1790    
1791    typedef struct dfa_recursion_info {
1792      struct dfa_recursion_info *prevrec;
1793      int group_num;
1794      PCRE_PUCHAR subject_position;
1795    } dfa_recursion_info;
1796    
1797  /* 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
1798  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
1799  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops; used by
1800    pcre_exec(). */
1801    
1802  typedef struct eptrblock {  typedef struct eptrblock {
1803    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
1804    USPTR epb_saved_eptr;    PCRE_PUCHAR epb_saved_eptr;
1805  } eptrblock;  } eptrblock;
1806    
1807    
# Line 1627  typedef struct match_data { Line 1812  typedef struct match_data {
1812    unsigned long int match_call_count;      /* As it says */    unsigned long int match_call_count;      /* As it says */
1813    unsigned long int match_limit;           /* As it says */    unsigned long int match_limit;           /* As it says */
1814    unsigned long int match_limit_recursion; /* As it says */    unsigned long int match_limit_recursion; /* As it says */
1815    int   *offset_vector;         /* Offset vector */    int   *offset_vector;           /* Offset vector */
1816    int    offset_end;            /* One past the end */    int    offset_end;              /* One past the end */
1817    int    offset_max;            /* The maximum usable for return data */    int    offset_max;              /* The maximum usable for return data */
1818    int    nltype;                /* Newline type */    int    nltype;                  /* Newline type */
1819    int    nllen;                 /* Newline string length */    int    nllen;                   /* Newline string length */
1820    int    name_count;            /* Number of names in name table */    int    name_count;              /* Number of names in name table */
1821    int    name_entry_size;       /* Size of entry in names table */    int    name_entry_size;         /* Size of entry in names table */
1822    uschar *name_table;           /* Table of names */    pcre_uchar *name_table;         /* Table of names */
1823    uschar nl[4];                 /* Newline string when fixed */    pcre_uchar nl[4];               /* Newline string when fixed */
1824    const uschar *lcc;            /* Points to lower casing table */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
1825    const uschar *ctypes;         /* Points to table of type maps */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
1826    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;         /* Set if too many extractions */
1827    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                  /* NOTBOL flag */
1828    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                  /* NOTEOL flag */
1829    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                    /* UTF8 flag */
1830    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
1831    BOOL   endonly;               /* Dollar not before final \n */    BOOL   use_ucp;                 /* PCRE_UCP flag */
1832    BOOL   notempty;              /* Empty string match not wanted */    BOOL   endonly;                 /* Dollar not before final \n */
1833    BOOL   notempty_atstart;      /* Empty string match at start not wanted */    BOOL   notempty;                /* Empty string match not wanted */
1834    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
1835    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   hitend;                  /* Hit the end of the subject at some point */
1836    const uschar *start_code;     /* For use when recursing */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
1837    USPTR  start_subject;         /* Start of the subject string */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
1838    USPTR  end_subject;           /* End of the subject string */    const  pcre_uchar *start_code;  /* For use when recursing */
1839    USPTR  start_match_ptr;       /* Start of matched string */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
1840    USPTR  end_match_ptr;         /* Subject position at end match */    PCRE_PUCHAR end_subject;        /* End of the subject string */
1841    USPTR  start_used_ptr;        /* Earliest consulted character */    PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
1842    int    partial;               /* PARTIAL options */    PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
1843    int    end_offset_top;        /* Highwater mark at end of match */    PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
1844    int    capture_last;          /* Most recent capture number */    int    partial;                 /* PARTIAL options */
1845    int    start_offset;          /* The start offset value */    int    end_offset_top;          /* Highwater mark at end of match */
1846    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    int    capture_last;            /* Most recent capture number */
1847    int    eptrn;                 /* Next free eptrblock */    int    start_offset;            /* The start offset value */
1848    recursion_info *recursive;    /* Linked list of recursion data */    int    match_function_type;     /* Set for certain special calls of MATCH() */
1849    void  *callout_data;          /* To pass back to callouts */    eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
1850      int    eptrn;                   /* Next free eptrblock */
1851      recursion_info *recursive;      /* Linked list of recursion data */
1852      void  *callout_data;            /* To pass back to callouts */
1853      const  pcre_uchar *mark;        /* Mark pointer to pass back */
1854      const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
1855  } match_data;  } match_data;
1856    
1857  /* 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
1858  functions. */  functions. */
1859    
1860  typedef struct dfa_match_data {  typedef struct dfa_match_data {
1861    const uschar *start_code;     /* Start of the compiled pattern */    const pcre_uchar *start_code;     /* Start of the compiled pattern */
1862    const uschar *start_subject;  /* Start of the subject string */    const pcre_uchar *start_subject ; /* Start of the subject string */
1863    const uschar *end_subject;    /* End of subject string */    const pcre_uchar *end_subject;    /* End of subject string */
1864    const uschar *start_used_ptr; /* Earliest consulted character */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
1865    const uschar *tables;         /* Character tables */    const pcre_uint8 *tables;         /* Character tables */
1866    int   start_offset;           /* The start offset value */    int   start_offset;               /* The start offset value */
1867    int   moptions;               /* Match options */    int   moptions;                   /* Match options */
1868    int   poptions;               /* Pattern options */    int   poptions;                   /* Pattern options */
1869    int    nltype;                /* Newline type */    int   nltype;                     /* Newline type */
1870    int    nllen;                 /* Newline string length */    int   nllen;                      /* Newline string length */
1871    uschar nl[4];                 /* Newline string when fixed */    pcre_uchar nl[4];                 /* Newline string when fixed */
1872    void  *callout_data;          /* To pass back to callouts */    void *callout_data;               /* To pass back to callouts */
1873      dfa_recursion_info *recursive;    /* Linked list of recursion data */
1874  } dfa_match_data;  } dfa_match_data;
1875    
1876  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 1732  of the exported public functions. They h Line 1923  of the exported public functions. They h
1923  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
1924  pcre_tables.c module. */  pcre_tables.c module. */
1925    
1926  extern const int    _pcre_utf8_table1[];  extern const int            _pcre_utf8_table1[];
1927  extern const int    _pcre_utf8_table2[];  extern const int            _pcre_utf8_table2[];
1928  extern const int    _pcre_utf8_table3[];  extern const int            _pcre_utf8_table3[];
1929  extern const uschar _pcre_utf8_table4[];  extern const pcre_uint8     _pcre_utf8_table4[];
1930    
1931  extern const int    _pcre_utf8_table1_size;  extern const int            _pcre_utf8_table1_size;
1932    
1933  extern const char   _pcre_utt_names[];  extern const char           _pcre_utt_names[];
1934  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1935  extern const int _pcre_utt_size;  extern const int            _pcre_utt_size;
1936    
1937  extern const uschar _pcre_default_tables[];  extern const pcre_uint8     _pcre_default_tables[];
1938    
1939  extern const uschar _pcre_OP_lengths[];  extern const pcre_uint8     _pcre_OP_lengths[];
1940    
1941    
1942  /* Internal shared functions. These are functions that are used by more than  /* Internal shared functions. These are functions that are used by more than
1943  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
1944  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1945    
1946  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);  extern const pcre_uchar *_pcre_find_bracket(const pcre_uchar *, BOOL, int);
1947  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);  extern BOOL              _pcre_is_newline(PCRE_PUCHAR, int, PCRE_PUCHAR,
1948  extern int           _pcre_ord2utf8(int, uschar *);                             int *, BOOL);
1949  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern int               _pcre_ord2utf8(int, pcre_uint8 *);
1950                         const pcre_study_data *, pcre_study_data *);  extern real_pcre        *_pcre_try_flipped(const real_pcre *, real_pcre *,
1951  extern int           _pcre_valid_utf8(USPTR, int);                             const pcre_study_data *, pcre_study_data *);
1952  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);  #ifndef COMPILE_PCRE16
1953  extern BOOL          _pcre_xclass(int, const uschar *);  extern int               _pcre_valid_utf8(PCRE_PUCHAR, int, int *);
1954    #else
1955    extern int               _pcre16_valid_utf16(PCRE_PUCHAR, int, int *);
1956    #endif
1957    extern BOOL              _pcre_was_newline(PCRE_PUCHAR, int, PCRE_PUCHAR,
1958                               int *, BOOL);
1959    extern BOOL              _pcre_xclass(int, const pcre_uchar *);
1960    
1961    #ifdef SUPPORT_JIT
1962    extern void              _pcre_jit_compile(const real_pcre *, pcre_extra *);
1963    extern int               _pcre_jit_exec(const real_pcre *, void *, PCRE_SPTR,
1964                               int, int, int, int, int *, int);
1965    extern void              _pcre_jit_free(void *);
1966    #endif
1967    
1968  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */
1969    
1970  typedef struct {  typedef struct {
1971    uschar script;    pcre_uint8 script;
1972    uschar chartype;    pcre_uint8 chartype;
1973    pcre_int32 other_case;    pcre_int32 other_case;
1974  } ucd_record;  } ucd_record;
1975    
1976  extern const ucd_record  _pcre_ucd_records[];  extern const ucd_record  _pcre_ucd_records[];
1977  extern const uschar      _pcre_ucd_stage1[];  extern const pcre_uint8  _pcre_ucd_stage1[];
1978  extern const pcre_uint16 _pcre_ucd_stage2[];  extern const pcre_uint16 _pcre_ucd_stage2[];
1979  extern const int         _pcre_ucp_gentype[];  extern const int         _pcre_ucp_gentype[];
1980    #ifdef SUPPORT_JIT
1981    extern const int         _pcre_ucp_typerange[];
1982    #endif
1983    
1984  /* UCD access macros */  /* UCD access macros */
1985    
1986  #define UCD_BLOCK_SIZE 128  #define UCD_BLOCK_SIZE 128
1987  #define GET_UCD(ch) (_pcre_ucd_records + \  #define GET_UCD(ch) (_pcre_ucd_records + \
1988          _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \          _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1989          UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])          UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
1990    
1991  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1992  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script

Legend:
Removed from v.483  
changed lines
  Added in v.759

  ViewVC Help
Powered by ViewVC 1.1.5