/[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 384 by ph10, Sun Mar 8 16:27:43 2009 UTC revision 788 by ph10, Tue Dec 6 15:38:01 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) 1997-2009 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 45  functions whose names all begin with "_p Line 45  functions whose names all begin with "_p
45  #ifndef PCRE_INTERNAL_H  #ifndef PCRE_INTERNAL_H
46  #define PCRE_INTERNAL_H  #define PCRE_INTERNAL_H
47    
48  /* Define DEBUG to get debugging output on stdout. */  /* Define PCRE_DEBUG to get debugging output on stdout. */
49    
50  #if 0  #if 0
51  #define DEBUG  #define PCRE_DEBUG
52    #endif
53    
54    /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"
55    script prevents both being selected, but not everybody uses "configure". */
56    
57    #if defined EBCDIC && defined SUPPORT_UTF8
58    #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.
59    #endif
60    
61    /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
62    "configure" script ensures this, but not everybody uses "configure". */
63    
64    #if defined SUPPORT_UCP && !defined SUPPORT_UTF8
65    #define SUPPORT_UTF8 1
66  #endif  #endif
67    
68  /* 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 74  It turns out that the Mac Debugging.h he
74  be absolutely sure we get our version. */  be absolutely sure we get our version. */
75    
76  #undef DPRINTF  #undef DPRINTF
77  #ifdef DEBUG  #ifdef PCRE_DEBUG
78  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
79  #else  #else
80  #define DPRINTF(p) /* Nothing */  #define DPRINTF(p) /* Nothing */
# Line 72  setjmp and stdarg are used is when NO_RE Line 86  setjmp and stdarg are used is when NO_RE
86    
87  #include <ctype.h>  #include <ctype.h>
88  #include <limits.h>  #include <limits.h>
 #include <setjmp.h>  
 #include <stdarg.h>  
89  #include <stddef.h>  #include <stddef.h>
90  #include <stdio.h>  #include <stdio.h>
91  #include <stdlib.h>  #include <stdlib.h>
# Line 172  preprocessor time in standard C environm Line 184  preprocessor time in standard C environm
184    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
185  #endif  #endif
186    
187    /* When checking for integer overflow in pcre_compile(), we need to handle
188    large integers. If a 64-bit integer type is available, we can use that.
189    Otherwise we have to cast to double, which of course requires floating point
190    arithmetic. Handle this by defining a macro for the appropriate type. If
191    stdint.h is available, include it; it may define INT64_MAX. Systems that do not
192    have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
193    by "configure". */
194    
195    #if HAVE_STDINT_H
196    #include <stdint.h>
197    #elif HAVE_INTTYPES_H
198    #include <inttypes.h>
199    #endif
200    
201    #if defined INT64_MAX || defined int64_t
202    #define INT64_OR_DOUBLE int64_t
203    #else
204    #define INT64_OR_DOUBLE double
205    #endif
206    
207  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
208  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
209  However, we leave the interface to the outside world as char *, because that  However, we leave the interface to the outside world as char *, because that
# Line 376  capturing parenthesis numbers in back re Line 408  capturing parenthesis numbers in back re
408    
409  /* 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
410  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
411  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should  byte-mode, and more complicated ones for UTF-8 characters. GETCHARLENTEST is
412  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
413  support is omitted, we don't even define it. */  never be called in byte mode. To make sure they can never even appear when
414    UTF-8 support is omitted, we don't even define them. */
415    
416  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
417  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 386  support is omitted, we don't even define Line 419  support is omitted, we don't even define
419  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
420  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
421  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
422    /* #define GETCHARLENTEST(c, eptr, len) */
423  /* #define BACKCHAR(eptr) */  /* #define BACKCHAR(eptr) */
424    
425  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
426    
427    /* These macros were originally written in the form of loops that used data
428    from the tables whose names start with _pcre_utf8_table. They were rewritten by
429    a user so as not to use loops, because in some environments this gives a
430    significant performance advantage, and it seems never to do any harm. */
431    
432    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
433    advancing the pointer. */
434    
435    #define GETUTF8(c, eptr) \
436        { \
437        if ((c & 0x20) == 0) \
438          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
439        else if ((c & 0x10) == 0) \
440          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
441        else if ((c & 0x08) == 0) \
442          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
443          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
444        else if ((c & 0x04) == 0) \
445          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
446              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
447              (eptr[4] & 0x3f); \
448        else \
449          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
450              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
451              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
452        }
453    
454  /* 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
455  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
456    
457  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
458    c = *eptr; \    c = *eptr; \
459    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; \  
       } \  
     }  
460    
461  /* 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
462  pointer. */  pointer. */
463    
464  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
465    c = *eptr; \    c = *eptr; \
466    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8(c, eptr);
467    
468    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
469    the pointer. */
470    
471    #define GETUTF8INC(c, eptr) \
472      { \      { \
473      int gcii; \      if ((c & 0x20) == 0) \
474      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
475      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
476      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        { \
477      for (gcii = 1; gcii <= gcaa; gcii++) \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
478          eptr += 2; \
479          } \
480        else if ((c & 0x08) == 0) \
481        { \        { \
482        gcss -= 6; \        c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
483        c |= (eptr[gcii] & 0x3f) << gcss; \            ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
484          eptr += 3; \
485          } \
486        else if ((c & 0x04) == 0) \
487          { \
488          c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
489              ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
490              (eptr[3] & 0x3f); \
491          eptr += 4; \
492          } \
493        else \
494          { \
495          c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
496              ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
497              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
498          eptr += 5; \
499        } \        } \
500      }      }
501    
# Line 431  know we are in UTF-8 mode. */ Line 504  know we are in UTF-8 mode. */
504    
505  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
506    c = *eptr++; \    c = *eptr++; \
507    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; \  
       } \  
     }  
508    
509  /* 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.
510    This is called when we don't know if we are in UTF-8 mode. */
511    
512  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
513    c = *eptr++; \    c = *eptr++; \
514    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8INC(c, eptr);
515    
516    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
517    advancing the pointer, incrementing the length. */
518    
519    #define GETUTF8LEN(c, eptr, len) \
520      { \      { \
521      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
522      int gcss = 6*gcaa; \        { \
523      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
524      while (gcaa-- > 0) \        len++; \
525          } \
526        else if ((c & 0x10)  == 0) \
527        { \        { \
528        gcss -= 6; \        c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
529        c |= (*eptr++ & 0x3f) << gcss; \        len += 2; \
530          } \
531        else if ((c & 0x08)  == 0) \
532          {\
533          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
534              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
535          len += 3; \
536          } \
537        else if ((c & 0x04)  == 0) \
538          { \
539          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
540              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
541              (eptr[4] & 0x3f); \
542          len += 4; \
543          } \
544        else \
545          {\
546          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
547              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
548              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
549          len += 5; \
550        } \        } \
551      }      }
552    
# Line 464  if there are extra bytes. This is called Line 555  if there are extra bytes. This is called
555    
556  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
557    c = *eptr; \    c = *eptr; \
558    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; \  
     }  
559    
560  /* 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
561  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
562  know we are in UTF-8 mode. */  do not know if we are in UTF-8 mode. */
563    
564  #define GETCHARLENTEST(c, eptr, len) \  #define GETCHARLENTEST(c, eptr, len) \
565    c = *eptr; \    c = *eptr; \
566    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; \  
     }  
567    
568  /* 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
569  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 504  because almost all calls are already wit Line 571  because almost all calls are already wit
571    
572  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
573    
574  #endif  #endif  /* SUPPORT_UTF8 */
575    
576    
577  /* 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 515  Standard C system should have one. */ Line 582  Standard C system should have one. */
582  #endif  #endif
583    
584    
 /* These are the public options that can change during matching. */  
   
 #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  
   
585  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
586  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
587  are in a 16-bit flags word. */  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
588    the restrictions on partial matching have been lifted. It remains for backwards
589    compatibility. */
590    
591  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
592  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
# Line 529  are in a 16-bit flags word. */ Line 594  are in a 16-bit flags word. */
594  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
595  #define PCRE_JCHANGED      0x0010  /* j option used in regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
596  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
597    #define PCRE_HASTHEN       0x0040  /* pattern contains (*THEN) */
598    
599  /* Options for the "extra" block produced by pcre_study(). */  /* Flags for the "extra" block produced by pcre_study(). */
600    
601  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
602    #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
603    
604  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
605  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 540  time, run time, or study time, respectiv Line 607  time, run time, or study time, respectiv
607  #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| \
608                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
609    
610  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
611    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
612     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
613     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| \
614     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
615     PCRE_JAVASCRIPT_COMPAT)     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
616    
617  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
618    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
619     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
620       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
621    
622  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
623    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
624     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
625     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
626       PCRE_NO_START_OPTIMIZE)
627    
628  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS \
629       PCRE_STUDY_JIT_COMPILE
630    
631  /* 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
632  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 590  typedef int BOOL; Line 660  typedef int BOOL;
660  #define TRUE    1  #define TRUE    1
661  #endif  #endif
662    
663    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
664    character constants like '*' because the compiler would emit their EBCDIC code,
665    which is different from their ASCII/UTF-8 code. Instead we define macros for
666    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
667    is enabled. When UTF-8 support is not enabled, the definitions use character
668    literals. Both character and string versions of each character are needed, and
669    there are some longer strings as well.
670    
671    This means that, on EBCDIC platforms, the PCRE library can handle either
672    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
673    would need different lookups depending on whether PCRE_UTF8 was set or not.
674    This would make it impossible to use characters in switch/case statements,
675    which would reduce performance. For a theoretical use (which nobody has asked
676    for) in a minority area (EBCDIC platforms), this is not sensible. Any
677    application that did need both could compile two versions of the library, using
678    macros to give the functions distinct names. */
679    
680    #ifndef SUPPORT_UTF8
681    
682    /* UTF-8 support is not enabled; use the platform-dependent character literals
683    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
684    
685    #define CHAR_HT                     '\t'
686    #define CHAR_VT                     '\v'
687    #define CHAR_FF                     '\f'
688    #define CHAR_CR                     '\r'
689    #define CHAR_NL                     '\n'
690    #define CHAR_BS                     '\b'
691    #define CHAR_BEL                    '\a'
692    #ifdef EBCDIC
693    #define CHAR_ESC                    '\047'
694    #define CHAR_DEL                    '\007'
695    #else
696    #define CHAR_ESC                    '\033'
697    #define CHAR_DEL                    '\177'
698    #endif
699    
700    #define CHAR_SPACE                  ' '
701    #define CHAR_EXCLAMATION_MARK       '!'
702    #define CHAR_QUOTATION_MARK         '"'
703    #define CHAR_NUMBER_SIGN            '#'
704    #define CHAR_DOLLAR_SIGN            '$'
705    #define CHAR_PERCENT_SIGN           '%'
706    #define CHAR_AMPERSAND              '&'
707    #define CHAR_APOSTROPHE             '\''
708    #define CHAR_LEFT_PARENTHESIS       '('
709    #define CHAR_RIGHT_PARENTHESIS      ')'
710    #define CHAR_ASTERISK               '*'
711    #define CHAR_PLUS                   '+'
712    #define CHAR_COMMA                  ','
713    #define CHAR_MINUS                  '-'
714    #define CHAR_DOT                    '.'
715    #define CHAR_SLASH                  '/'
716    #define CHAR_0                      '0'
717    #define CHAR_1                      '1'
718    #define CHAR_2                      '2'
719    #define CHAR_3                      '3'
720    #define CHAR_4                      '4'
721    #define CHAR_5                      '5'
722    #define CHAR_6                      '6'
723    #define CHAR_7                      '7'
724    #define CHAR_8                      '8'
725    #define CHAR_9                      '9'
726    #define CHAR_COLON                  ':'
727    #define CHAR_SEMICOLON              ';'
728    #define CHAR_LESS_THAN_SIGN         '<'
729    #define CHAR_EQUALS_SIGN            '='
730    #define CHAR_GREATER_THAN_SIGN      '>'
731    #define CHAR_QUESTION_MARK          '?'
732    #define CHAR_COMMERCIAL_AT          '@'
733    #define CHAR_A                      'A'
734    #define CHAR_B                      'B'
735    #define CHAR_C                      'C'
736    #define CHAR_D                      'D'
737    #define CHAR_E                      'E'
738    #define CHAR_F                      'F'
739    #define CHAR_G                      'G'
740    #define CHAR_H                      'H'
741    #define CHAR_I                      'I'
742    #define CHAR_J                      'J'
743    #define CHAR_K                      'K'
744    #define CHAR_L                      'L'
745    #define CHAR_M                      'M'
746    #define CHAR_N                      'N'
747    #define CHAR_O                      'O'
748    #define CHAR_P                      'P'
749    #define CHAR_Q                      'Q'
750    #define CHAR_R                      'R'
751    #define CHAR_S                      'S'
752    #define CHAR_T                      'T'
753    #define CHAR_U                      'U'
754    #define CHAR_V                      'V'
755    #define CHAR_W                      'W'
756    #define CHAR_X                      'X'
757    #define CHAR_Y                      'Y'
758    #define CHAR_Z                      'Z'
759    #define CHAR_LEFT_SQUARE_BRACKET    '['
760    #define CHAR_BACKSLASH              '\\'
761    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
762    #define CHAR_CIRCUMFLEX_ACCENT      '^'
763    #define CHAR_UNDERSCORE             '_'
764    #define CHAR_GRAVE_ACCENT           '`'
765    #define CHAR_a                      'a'
766    #define CHAR_b                      'b'
767    #define CHAR_c                      'c'
768    #define CHAR_d                      'd'
769    #define CHAR_e                      'e'
770    #define CHAR_f                      'f'
771    #define CHAR_g                      'g'
772    #define CHAR_h                      'h'
773    #define CHAR_i                      'i'
774    #define CHAR_j                      'j'
775    #define CHAR_k                      'k'
776    #define CHAR_l                      'l'
777    #define CHAR_m                      'm'
778    #define CHAR_n                      'n'
779    #define CHAR_o                      'o'
780    #define CHAR_p                      'p'
781    #define CHAR_q                      'q'
782    #define CHAR_r                      'r'
783    #define CHAR_s                      's'
784    #define CHAR_t                      't'
785    #define CHAR_u                      'u'
786    #define CHAR_v                      'v'
787    #define CHAR_w                      'w'
788    #define CHAR_x                      'x'
789    #define CHAR_y                      'y'
790    #define CHAR_z                      'z'
791    #define CHAR_LEFT_CURLY_BRACKET     '{'
792    #define CHAR_VERTICAL_LINE          '|'
793    #define CHAR_RIGHT_CURLY_BRACKET    '}'
794    #define CHAR_TILDE                  '~'
795    
796    #define STR_HT                      "\t"
797    #define STR_VT                      "\v"
798    #define STR_FF                      "\f"
799    #define STR_CR                      "\r"
800    #define STR_NL                      "\n"
801    #define STR_BS                      "\b"
802    #define STR_BEL                     "\a"
803    #ifdef EBCDIC
804    #define STR_ESC                     "\047"
805    #define STR_DEL                     "\007"
806    #else
807    #define STR_ESC                     "\033"
808    #define STR_DEL                     "\177"
809    #endif
810    
811    #define STR_SPACE                   " "
812    #define STR_EXCLAMATION_MARK        "!"
813    #define STR_QUOTATION_MARK          "\""
814    #define STR_NUMBER_SIGN             "#"
815    #define STR_DOLLAR_SIGN             "$"
816    #define STR_PERCENT_SIGN            "%"
817    #define STR_AMPERSAND               "&"
818    #define STR_APOSTROPHE              "'"
819    #define STR_LEFT_PARENTHESIS        "("
820    #define STR_RIGHT_PARENTHESIS       ")"
821    #define STR_ASTERISK                "*"
822    #define STR_PLUS                    "+"
823    #define STR_COMMA                   ","
824    #define STR_MINUS                   "-"
825    #define STR_DOT                     "."
826    #define STR_SLASH                   "/"
827    #define STR_0                       "0"
828    #define STR_1                       "1"
829    #define STR_2                       "2"
830    #define STR_3                       "3"
831    #define STR_4                       "4"
832    #define STR_5                       "5"
833    #define STR_6                       "6"
834    #define STR_7                       "7"
835    #define STR_8                       "8"
836    #define STR_9                       "9"
837    #define STR_COLON                   ":"
838    #define STR_SEMICOLON               ";"
839    #define STR_LESS_THAN_SIGN          "<"
840    #define STR_EQUALS_SIGN             "="
841    #define STR_GREATER_THAN_SIGN       ">"
842    #define STR_QUESTION_MARK           "?"
843    #define STR_COMMERCIAL_AT           "@"
844    #define STR_A                       "A"
845    #define STR_B                       "B"
846    #define STR_C                       "C"
847    #define STR_D                       "D"
848    #define STR_E                       "E"
849    #define STR_F                       "F"
850    #define STR_G                       "G"
851    #define STR_H                       "H"
852    #define STR_I                       "I"
853    #define STR_J                       "J"
854    #define STR_K                       "K"
855    #define STR_L                       "L"
856    #define STR_M                       "M"
857    #define STR_N                       "N"
858    #define STR_O                       "O"
859    #define STR_P                       "P"
860    #define STR_Q                       "Q"
861    #define STR_R                       "R"
862    #define STR_S                       "S"
863    #define STR_T                       "T"
864    #define STR_U                       "U"
865    #define STR_V                       "V"
866    #define STR_W                       "W"
867    #define STR_X                       "X"
868    #define STR_Y                       "Y"
869    #define STR_Z                       "Z"
870    #define STR_LEFT_SQUARE_BRACKET     "["
871    #define STR_BACKSLASH               "\\"
872    #define STR_RIGHT_SQUARE_BRACKET    "]"
873    #define STR_CIRCUMFLEX_ACCENT       "^"
874    #define STR_UNDERSCORE              "_"
875    #define STR_GRAVE_ACCENT            "`"
876    #define STR_a                       "a"
877    #define STR_b                       "b"
878    #define STR_c                       "c"
879    #define STR_d                       "d"
880    #define STR_e                       "e"
881    #define STR_f                       "f"
882    #define STR_g                       "g"
883    #define STR_h                       "h"
884    #define STR_i                       "i"
885    #define STR_j                       "j"
886    #define STR_k                       "k"
887    #define STR_l                       "l"
888    #define STR_m                       "m"
889    #define STR_n                       "n"
890    #define STR_o                       "o"
891    #define STR_p                       "p"
892    #define STR_q                       "q"
893    #define STR_r                       "r"
894    #define STR_s                       "s"
895    #define STR_t                       "t"
896    #define STR_u                       "u"
897    #define STR_v                       "v"
898    #define STR_w                       "w"
899    #define STR_x                       "x"
900    #define STR_y                       "y"
901    #define STR_z                       "z"
902    #define STR_LEFT_CURLY_BRACKET      "{"
903    #define STR_VERTICAL_LINE           "|"
904    #define STR_RIGHT_CURLY_BRACKET     "}"
905    #define STR_TILDE                   "~"
906    
907    #define STRING_ACCEPT0              "ACCEPT\0"
908    #define STRING_COMMIT0              "COMMIT\0"
909    #define STRING_F0                   "F\0"
910    #define STRING_FAIL0                "FAIL\0"
911    #define STRING_MARK0                "MARK\0"
912    #define STRING_PRUNE0               "PRUNE\0"
913    #define STRING_SKIP0                "SKIP\0"
914    #define STRING_THEN                 "THEN"
915    
916    #define STRING_alpha0               "alpha\0"
917    #define STRING_lower0               "lower\0"
918    #define STRING_upper0               "upper\0"
919    #define STRING_alnum0               "alnum\0"
920    #define STRING_ascii0               "ascii\0"
921    #define STRING_blank0               "blank\0"
922    #define STRING_cntrl0               "cntrl\0"
923    #define STRING_digit0               "digit\0"
924    #define STRING_graph0               "graph\0"
925    #define STRING_print0               "print\0"
926    #define STRING_punct0               "punct\0"
927    #define STRING_space0               "space\0"
928    #define STRING_word0                "word\0"
929    #define STRING_xdigit               "xdigit"
930    
931    #define STRING_DEFINE               "DEFINE"
932    
933    #define STRING_CR_RIGHTPAR             "CR)"
934    #define STRING_LF_RIGHTPAR             "LF)"
935    #define STRING_CRLF_RIGHTPAR           "CRLF)"
936    #define STRING_ANY_RIGHTPAR            "ANY)"
937    #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
938    #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
939    #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
940    #define STRING_UTF8_RIGHTPAR           "UTF8)"
941    #define STRING_UCP_RIGHTPAR            "UCP)"
942    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
943    
944    #else  /* SUPPORT_UTF8 */
945    
946    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
947    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
948    only. */
949    
950    #define CHAR_HT                     '\011'
951    #define CHAR_VT                     '\013'
952    #define CHAR_FF                     '\014'
953    #define CHAR_CR                     '\015'
954    #define CHAR_NL                     '\012'
955    #define CHAR_BS                     '\010'
956    #define CHAR_BEL                    '\007'
957    #define CHAR_ESC                    '\033'
958    #define CHAR_DEL                    '\177'
959    
960    #define CHAR_SPACE                  '\040'
961    #define CHAR_EXCLAMATION_MARK       '\041'
962    #define CHAR_QUOTATION_MARK         '\042'
963    #define CHAR_NUMBER_SIGN            '\043'
964    #define CHAR_DOLLAR_SIGN            '\044'
965    #define CHAR_PERCENT_SIGN           '\045'
966    #define CHAR_AMPERSAND              '\046'
967    #define CHAR_APOSTROPHE             '\047'
968    #define CHAR_LEFT_PARENTHESIS       '\050'
969    #define CHAR_RIGHT_PARENTHESIS      '\051'
970    #define CHAR_ASTERISK               '\052'
971    #define CHAR_PLUS                   '\053'
972    #define CHAR_COMMA                  '\054'
973    #define CHAR_MINUS                  '\055'
974    #define CHAR_DOT                    '\056'
975    #define CHAR_SLASH                  '\057'
976    #define CHAR_0                      '\060'
977    #define CHAR_1                      '\061'
978    #define CHAR_2                      '\062'
979    #define CHAR_3                      '\063'
980    #define CHAR_4                      '\064'
981    #define CHAR_5                      '\065'
982    #define CHAR_6                      '\066'
983    #define CHAR_7                      '\067'
984    #define CHAR_8                      '\070'
985    #define CHAR_9                      '\071'
986    #define CHAR_COLON                  '\072'
987    #define CHAR_SEMICOLON              '\073'
988    #define CHAR_LESS_THAN_SIGN         '\074'
989    #define CHAR_EQUALS_SIGN            '\075'
990    #define CHAR_GREATER_THAN_SIGN      '\076'
991    #define CHAR_QUESTION_MARK          '\077'
992    #define CHAR_COMMERCIAL_AT          '\100'
993    #define CHAR_A                      '\101'
994    #define CHAR_B                      '\102'
995    #define CHAR_C                      '\103'
996    #define CHAR_D                      '\104'
997    #define CHAR_E                      '\105'
998    #define CHAR_F                      '\106'
999    #define CHAR_G                      '\107'
1000    #define CHAR_H                      '\110'
1001    #define CHAR_I                      '\111'
1002    #define CHAR_J                      '\112'
1003    #define CHAR_K                      '\113'
1004    #define CHAR_L                      '\114'
1005    #define CHAR_M                      '\115'
1006    #define CHAR_N                      '\116'
1007    #define CHAR_O                      '\117'
1008    #define CHAR_P                      '\120'
1009    #define CHAR_Q                      '\121'
1010    #define CHAR_R                      '\122'
1011    #define CHAR_S                      '\123'
1012    #define CHAR_T                      '\124'
1013    #define CHAR_U                      '\125'
1014    #define CHAR_V                      '\126'
1015    #define CHAR_W                      '\127'
1016    #define CHAR_X                      '\130'
1017    #define CHAR_Y                      '\131'
1018    #define CHAR_Z                      '\132'
1019    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
1020    #define CHAR_BACKSLASH              '\134'
1021    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
1022    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
1023    #define CHAR_UNDERSCORE             '\137'
1024    #define CHAR_GRAVE_ACCENT           '\140'
1025    #define CHAR_a                      '\141'
1026    #define CHAR_b                      '\142'
1027    #define CHAR_c                      '\143'
1028    #define CHAR_d                      '\144'
1029    #define CHAR_e                      '\145'
1030    #define CHAR_f                      '\146'
1031    #define CHAR_g                      '\147'
1032    #define CHAR_h                      '\150'
1033    #define CHAR_i                      '\151'
1034    #define CHAR_j                      '\152'
1035    #define CHAR_k                      '\153'
1036    #define CHAR_l                      '\154'
1037    #define CHAR_m                      '\155'
1038    #define CHAR_n                      '\156'
1039    #define CHAR_o                      '\157'
1040    #define CHAR_p                      '\160'
1041    #define CHAR_q                      '\161'
1042    #define CHAR_r                      '\162'
1043    #define CHAR_s                      '\163'
1044    #define CHAR_t                      '\164'
1045    #define CHAR_u                      '\165'
1046    #define CHAR_v                      '\166'
1047    #define CHAR_w                      '\167'
1048    #define CHAR_x                      '\170'
1049    #define CHAR_y                      '\171'
1050    #define CHAR_z                      '\172'
1051    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1052    #define CHAR_VERTICAL_LINE          '\174'
1053    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1054    #define CHAR_TILDE                  '\176'
1055    
1056    #define STR_HT                      "\011"
1057    #define STR_VT                      "\013"
1058    #define STR_FF                      "\014"
1059    #define STR_CR                      "\015"
1060    #define STR_NL                      "\012"
1061    #define STR_BS                      "\010"
1062    #define STR_BEL                     "\007"
1063    #define STR_ESC                     "\033"
1064    #define STR_DEL                     "\177"
1065    
1066    #define STR_SPACE                   "\040"
1067    #define STR_EXCLAMATION_MARK        "\041"
1068    #define STR_QUOTATION_MARK          "\042"
1069    #define STR_NUMBER_SIGN             "\043"
1070    #define STR_DOLLAR_SIGN             "\044"
1071    #define STR_PERCENT_SIGN            "\045"
1072    #define STR_AMPERSAND               "\046"
1073    #define STR_APOSTROPHE              "\047"
1074    #define STR_LEFT_PARENTHESIS        "\050"
1075    #define STR_RIGHT_PARENTHESIS       "\051"
1076    #define STR_ASTERISK                "\052"
1077    #define STR_PLUS                    "\053"
1078    #define STR_COMMA                   "\054"
1079    #define STR_MINUS                   "\055"
1080    #define STR_DOT                     "\056"
1081    #define STR_SLASH                   "\057"
1082    #define STR_0                       "\060"
1083    #define STR_1                       "\061"
1084    #define STR_2                       "\062"
1085    #define STR_3                       "\063"
1086    #define STR_4                       "\064"
1087    #define STR_5                       "\065"
1088    #define STR_6                       "\066"
1089    #define STR_7                       "\067"
1090    #define STR_8                       "\070"
1091    #define STR_9                       "\071"
1092    #define STR_COLON                   "\072"
1093    #define STR_SEMICOLON               "\073"
1094    #define STR_LESS_THAN_SIGN          "\074"
1095    #define STR_EQUALS_SIGN             "\075"
1096    #define STR_GREATER_THAN_SIGN       "\076"
1097    #define STR_QUESTION_MARK           "\077"
1098    #define STR_COMMERCIAL_AT           "\100"
1099    #define STR_A                       "\101"
1100    #define STR_B                       "\102"
1101    #define STR_C                       "\103"
1102    #define STR_D                       "\104"
1103    #define STR_E                       "\105"
1104    #define STR_F                       "\106"
1105    #define STR_G                       "\107"
1106    #define STR_H                       "\110"
1107    #define STR_I                       "\111"
1108    #define STR_J                       "\112"
1109    #define STR_K                       "\113"
1110    #define STR_L                       "\114"
1111    #define STR_M                       "\115"
1112    #define STR_N                       "\116"
1113    #define STR_O                       "\117"
1114    #define STR_P                       "\120"
1115    #define STR_Q                       "\121"
1116    #define STR_R                       "\122"
1117    #define STR_S                       "\123"
1118    #define STR_T                       "\124"
1119    #define STR_U                       "\125"
1120    #define STR_V                       "\126"
1121    #define STR_W                       "\127"
1122    #define STR_X                       "\130"
1123    #define STR_Y                       "\131"
1124    #define STR_Z                       "\132"
1125    #define STR_LEFT_SQUARE_BRACKET     "\133"
1126    #define STR_BACKSLASH               "\134"
1127    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1128    #define STR_CIRCUMFLEX_ACCENT       "\136"
1129    #define STR_UNDERSCORE              "\137"
1130    #define STR_GRAVE_ACCENT            "\140"
1131    #define STR_a                       "\141"
1132    #define STR_b                       "\142"
1133    #define STR_c                       "\143"
1134    #define STR_d                       "\144"
1135    #define STR_e                       "\145"
1136    #define STR_f                       "\146"
1137    #define STR_g                       "\147"
1138    #define STR_h                       "\150"
1139    #define STR_i                       "\151"
1140    #define STR_j                       "\152"
1141    #define STR_k                       "\153"
1142    #define STR_l                       "\154"
1143    #define STR_m                       "\155"
1144    #define STR_n                       "\156"
1145    #define STR_o                       "\157"
1146    #define STR_p                       "\160"
1147    #define STR_q                       "\161"
1148    #define STR_r                       "\162"
1149    #define STR_s                       "\163"
1150    #define STR_t                       "\164"
1151    #define STR_u                       "\165"
1152    #define STR_v                       "\166"
1153    #define STR_w                       "\167"
1154    #define STR_x                       "\170"
1155    #define STR_y                       "\171"
1156    #define STR_z                       "\172"
1157    #define STR_LEFT_CURLY_BRACKET      "\173"
1158    #define STR_VERTICAL_LINE           "\174"
1159    #define STR_RIGHT_CURLY_BRACKET     "\175"
1160    #define STR_TILDE                   "\176"
1161    
1162    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1163    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1164    #define STRING_F0                   STR_F "\0"
1165    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1166    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1167    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1168    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1169    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1170    
1171    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1172    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1173    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1174    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1175    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1176    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1177    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1178    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1179    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1180    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1181    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1182    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1183    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1184    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1185    
1186    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1187    
1188    #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1189    #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1190    #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1191    #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1192    #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1193    #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
1194    #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
1195    #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1196    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1197    #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
1198    
1199    #endif  /* SUPPORT_UTF8 */
1200    
1201  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1202    
1203  #ifndef ESC_e  #ifndef ESC_e
1204  #define ESC_e 27  #define ESC_e CHAR_ESC
1205  #endif  #endif
1206    
1207  #ifndef ESC_f  #ifndef ESC_f
1208  #define ESC_f '\f'  #define ESC_f CHAR_FF
1209  #endif  #endif
1210    
1211  #ifndef ESC_n  #ifndef ESC_n
1212  #define ESC_n '\n'  #define ESC_n CHAR_NL
1213  #endif  #endif
1214    
1215  #ifndef ESC_r  #ifndef ESC_r
1216  #define ESC_r '\r'  #define ESC_r CHAR_CR
1217  #endif  #endif
1218    
1219  /* 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
1220  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1221    
1222  #ifndef ESC_tee  #ifndef ESC_tee
1223  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1224  #endif  #endif
1225    
1226  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
1227    
1228  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1229  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1230  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1231  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1232  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1233    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1234    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1235    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1236    #define PT_WORD       8    /* Word - L plus N plus underscore */
1237    
1238  /* 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
1239  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 638  contain UTF-8 characters with values gre Line 1250  contain UTF-8 characters with values gre
1250  /* 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
1251  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
1252  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
1253  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
1254  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
1255  (which is used for [^] in JavaScript compatibility mode).  used for [^] in JavaScript compatibility mode, and for \C in non-utf8 mode. In
1256    non-DOTALL mode, "." behaves like \N.
1257    
1258    The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1259    when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1260    They must be contiguous, and remain in order so that the replacements can be
1261    looked up from a table.
1262    
1263  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
1264  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 650  put in between that don't consume a char Line 1268  put in between that don't consume a char
1268  */  */
1269    
1270  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,
1271         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,
1272         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,
1273           ESC_E, ESC_Q, ESC_g, ESC_k,
1274           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1275         ESC_REF };         ESC_REF };
1276    
   
1277  /* 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
1278  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.
1279    
1280  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1281  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
1282  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1283    
1284  enum {  enum {
1285    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 678  enum { Line 1297  enum {
1297    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1298    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1299    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1300    OP_ANY,            /* 12 Match any character (subject to DOTALL) */  
1301    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */    OP_ANY,            /* 12 Match any character except newline */
1302      OP_ALLANY,         /* 13 Match any character */
1303    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 */
1304    OP_NOTPROP,        /* 15 \P (not Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1305    OP_PROP,           /* 16 \p (Unicode property) */    OP_PROP,           /* 16 \p (Unicode property) */
# Line 692  enum { Line 1312  enum {
1312    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. */
1313    OP_EOD,            /* 24 End of data: \z */    OP_EOD,            /* 24 End of data: \z */
1314    
1315    OP_OPT,            /* 25 Set runtime options */    OP_CIRC,           /* 25 Start of line - not multiline */
1316    OP_CIRC,           /* 26 Start of line - varies with multiline switch */    OP_CIRCM,          /* 26 Start of line - multiline */
1317    OP_DOLL,           /* 27 End of line - varies with multiline switch */    OP_DOLL,           /* 27 End of line - not multiline */
1318    OP_CHAR,           /* 28 Match one character, casefully */    OP_DOLLM,          /* 28 End of line - multiline */
1319    OP_CHARNC,         /* 29 Match one character, caselessly */    OP_CHAR,           /* 29 Match one character, casefully */
1320    OP_NOT,            /* 30 Match one character, not the following one */    OP_CHARI,          /* 30 Match one character, caselessly */
1321      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1322    OP_STAR,           /* 31 The maximizing and minimizing versions of */    OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1323    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */  
1324    OP_PLUS,           /* 33 the minimizing one second. */    /* The following sets of 13 opcodes must always be kept in step because
1325    OP_MINPLUS,        /* 34 This first set applies to single characters.*/    the offset from the first one is used to generate the others. */
1326    OP_QUERY,          /* 35 */  
1327    OP_MINQUERY,       /* 36 */    /**** Single characters, caseful, must precede the caseless ones ****/
1328    
1329    OP_UPTO,           /* 37 From 0 to n matches */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1330    OP_MINUPTO,        /* 38 */    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1331    OP_EXACT,          /* 39 Exactly n matches */    OP_PLUS,           /* 35 the minimizing one second. */
1332      OP_MINPLUS,        /* 36 */
1333    OP_POSSTAR,        /* 40 Possessified star */    OP_QUERY,          /* 37 */
1334    OP_POSPLUS,        /* 41 Possessified plus */    OP_MINQUERY,       /* 38 */
1335    OP_POSQUERY,       /* 42 Posesssified query */  
1336    OP_POSUPTO,        /* 43 Possessified upto */    OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1337      OP_MINUPTO,        /* 40 */
1338    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */    OP_EXACT,          /* 41 Exactly n matches */
1339    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */  
1340    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */    OP_POSSTAR,        /* 42 Possessified star, caseful */
1341    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */    OP_POSPLUS,        /* 43 Possessified plus, caseful */
1342    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
1343    OP_NOTMINQUERY,    /* 49 */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
1344    
1345    OP_NOTUPTO,        /* 50 From 0 to n matches */    /**** Single characters, caseless, must follow the caseful ones */
1346    OP_NOTMINUPTO,     /* 51 */  
1347    OP_NOTEXACT,       /* 52 Exactly n matches */    OP_STARI,          /* 46 */
1348      OP_MINSTARI,       /* 47 */
1349    OP_NOTPOSSTAR,     /* 53 Possessified versions */    OP_PLUSI,          /* 48 */
1350    OP_NOTPOSPLUS,     /* 54 */    OP_MINPLUSI,       /* 49 */
1351    OP_NOTPOSQUERY,    /* 55 */    OP_QUERYI,         /* 50 */
1352    OP_NOTPOSUPTO,     /* 56 */    OP_MINQUERYI,      /* 51 */
1353    
1354    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */    OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1355    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */    OP_MINUPTOI,       /* 53 */
1356    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */    OP_EXACTI,         /* 54 */
1357    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */  
1358    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */    OP_POSSTARI,       /* 55 Possessified star, caseless */
1359    OP_TYPEMINQUERY,   /* 62 */    OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1360      OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1361    OP_TYPEUPTO,       /* 63 From 0 to n matches */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1362    OP_TYPEMINUPTO,    /* 64 */  
1363    OP_TYPEEXACT,      /* 65 Exactly n matches */    /**** The negated ones must follow the non-negated ones, and match them ****/
1364      /**** Negated single character, caseful; must precede the caseless ones ****/
1365    OP_TYPEPOSSTAR,    /* 66 Possessified versions */  
1366    OP_TYPEPOSPLUS,    /* 67 */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1367    OP_TYPEPOSQUERY,   /* 68 */    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1368    OP_TYPEPOSUPTO,    /* 69 */    OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1369      OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1370    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */    OP_NOTQUERY,       /* 63 */
1371    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */    OP_NOTMINQUERY,    /* 64 */
1372    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */  
1373    OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */    OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1374    OP_CRQUERY,        /* 74 These are for character classes and back refs */    OP_NOTMINUPTO,     /* 66 */
1375    OP_CRMINQUERY,     /* 75 */    OP_NOTEXACT,       /* 67 Exactly n matches */
1376    OP_CRRANGE,        /* 76 These are different to the three sets above. */  
1377    OP_CRMINRANGE,     /* 77 */    OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1378      OP_NOTPOSPLUS,     /* 69 */
1379    OP_CLASS,          /* 78 Match a character class, chars < 256 only */    OP_NOTPOSQUERY,    /* 70 */
1380    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative    OP_NOTPOSUPTO,     /* 71 */
1381                             class - the difference is relevant only when a UTF-8  
1382                             character > 255 is encountered. */    /**** Negated single character, caseless; must follow the caseful ones ****/
1383    
1384    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the    OP_NOTSTARI,       /* 72 */
1385                             class. This does both positive and negative. */    OP_NOTMINSTARI,    /* 73 */
1386      OP_NOTPLUSI,       /* 74 */
1387    OP_REF,            /* 81 Match a back reference */    OP_NOTMINPLUSI,    /* 75 */
1388    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */    OP_NOTQUERYI,      /* 76 */
1389    OP_CALLOUT,        /* 83 Call out to external function if provided */    OP_NOTMINQUERYI,   /* 77 */
1390    
1391    OP_ALT,            /* 84 Start of alternation */    OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1392    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */    OP_NOTMINUPTOI,    /* 79 */
1393    OP_KETRMAX,        /* 86 These two must remain together and in this */    OP_NOTEXACTI,      /* 80 Exactly n matches */
1394    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */  
1395      OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1396    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    OP_NOTPOSPLUSI,    /* 82 */
1397      OP_NOTPOSQUERYI,   /* 83 */
1398    OP_ASSERT,         /* 88 Positive lookahead */    OP_NOTPOSUPTOI,    /* 84 */
1399    OP_ASSERT_NOT,     /* 89 Negative lookahead */  
1400    OP_ASSERTBACK,     /* 90 Positive lookbehind */    /**** Character types ****/
1401    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */  
1402    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1403      OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1404    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1405    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. */
1406      OP_TYPEQUERY,      /* 89 */
1407    OP_ONCE,           /* 93 Atomic group */    OP_TYPEMINQUERY,   /* 90 */
1408    OP_BRA,            /* 94 Start of non-capturing bracket */  
1409    OP_CBRA,           /* 95 Start of capturing bracket */    OP_TYPEUPTO,       /* 91 From 0 to n matches */
1410    OP_COND,           /* 96 Conditional group */    OP_TYPEMINUPTO,    /* 92 */
1411      OP_TYPEEXACT,      /* 93 Exactly n matches */
1412    
1413      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1414      OP_TYPEPOSPLUS,    /* 95 */
1415      OP_TYPEPOSQUERY,   /* 96 */
1416      OP_TYPEPOSUPTO,    /* 97 */
1417    
1418      /* These are used for character classes and back references; only the
1419      first six are the same as the sets above. */
1420    
1421      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1422      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1423      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1424      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1425      OP_CRQUERY,        /* 102 */
1426      OP_CRMINQUERY,     /* 103 */
1427    
1428      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1429      OP_CRMINRANGE,     /* 105 */
1430    
1431      /* End of quantifier opcodes */
1432    
1433      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1434      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1435                                  class - the difference is relevant only when a
1436                                  UTF-8 character > 255 is encountered. */
1437      OP_XCLASS,         /* 108 Extended class for handling UTF-8 chars within the
1438                                  class. This does both positive and negative. */
1439      OP_REF,            /* 109 Match a back reference, casefully */
1440      OP_REFI,           /* 110 Match a back reference, caselessly */
1441      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1442      OP_CALLOUT,        /* 112 Call out to external function if provided */
1443    
1444      OP_ALT,            /* 113 Start of alternation */
1445      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1446      OP_KETRMAX,        /* 115 These two must remain together and in this */
1447      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1448      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1449    
1450      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1451      asserts must remain in order. */
1452    
1453      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1454      OP_ASSERT,         /* 119 Positive lookahead */
1455      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1456      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1457      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1458    
1459      /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
1460      after the assertions, with ONCE first, as there's a test for >= ONCE for a
1461      subpattern that isn't an assertion. The POS versions must immediately follow
1462      the non-POS versions in each case. */
1463    
1464      OP_ONCE,           /* 123 Atomic group, contains captures */
1465      OP_ONCE_NC,        /* 124 Atomic group containing no captures */
1466      OP_BRA,            /* 125 Start of non-capturing bracket */
1467      OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
1468      OP_CBRA,           /* 127 Start of capturing bracket */
1469      OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
1470      OP_COND,           /* 129 Conditional group */
1471    
1472    /* 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
1473    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1474    
1475    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
1476    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
1477    OP_SCOND,          /* 99 Conditional group, check empty */    OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
1478      OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
1479    OP_CREF,           /* 100 Used to hold a capture number as condition */    OP_SCOND,          /* 134 Conditional group, check empty */
1480    OP_RREF,           /* 101 Used to hold a recursion number as condition */  
1481    OP_DEF,            /* 102 The DEFINE condition */    /* The next two pairs must (respectively) be kept together. */
1482    
1483    OP_BRAZERO,        /* 103 These two must remain together and in this */    OP_CREF,           /* 135 Used to hold a capture number as condition */
1484    OP_BRAMINZERO,     /* 104 order. */    OP_NCREF,          /* 136 Same, but generated by a name reference*/
1485      OP_RREF,           /* 137 Used to hold a recursion number as condition */
1486      OP_NRREF,          /* 138 Same, but generated by a name reference*/
1487      OP_DEF,            /* 139 The DEFINE condition */
1488    
1489      OP_BRAZERO,        /* 140 These two must remain together and in this */
1490      OP_BRAMINZERO,     /* 141 order. */
1491      OP_BRAPOSZERO,     /* 142 */
1492    
1493    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1494    
1495    OP_PRUNE,          /* 105 */    OP_MARK,           /* 143 always has an argument */
1496    OP_SKIP,           /* 106 */    OP_PRUNE,          /* 144 */
1497    OP_THEN,           /* 107 */    OP_PRUNE_ARG,      /* 145 same, but with argument */
1498    OP_COMMIT,         /* 108 */    OP_SKIP,           /* 146 */
1499      OP_SKIP_ARG,       /* 147 same, but with argument */
1500      OP_THEN,           /* 148 */
1501      OP_THEN_ARG,       /* 149 same, but with argument */
1502      OP_COMMIT,         /* 150 */
1503    
1504    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1505    
1506    OP_FAIL,           /* 109 */    OP_FAIL,           /* 151 */
1507    OP_ACCEPT,         /* 110 */    OP_ACCEPT,         /* 152 */
1508      OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
1509      OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
1510    
1511    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1512    
1513    OP_SKIPZERO        /* 111 */    OP_SKIPZERO,       /* 155 */
1514    
1515      /* This is not an opcode, but is used to check that tables indexed by opcode
1516      are the correct length, in order to catch updating errors - there have been
1517      some in the past. */
1518    
1519      OP_TABLE_LENGTH
1520  };  };
1521    
1522    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1523    definitions that follow must also be updated to match. There are also tables
1524    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1525    
1526    
1527  /* 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
1528  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
1529    only in pcre_printint.c, which fills out the full names in many cases (and in
1530    some cases doesn't actually use these names at all). */
1531    
1532  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1533    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1534    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1535    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1536    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1537    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1538    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1539      "{", "{", "{",                                                  \
1540    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1541    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1542      "{", "{", "{",                                                  \
1543      "*+","++", "?+", "{",                                           \
1544      "*", "*?", "+", "+?", "?", "??",                                \
1545      "{", "{", "{",                                                  \
1546      "*+","++", "?+", "{",                                           \
1547      "*", "*?", "+", "+?", "?", "??",                                \
1548      "{", "{", "{",                                                  \
1549    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1550    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1551    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1552    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1553    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1554    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1555    "AssertB", "AssertB not", "Reverse",                            \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1556    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1557    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Once", "Once_NC",                                              \
1558    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "Bra", "BraPos", "CBra", "CBraPos",                             \
1559    "Skip zero"    "Cond",                                                         \
1560      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1561      "SCond",                                                        \
1562      "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1563      "Brazero", "Braminzero", "Braposzero",                          \
1564      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1565      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1566      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1567      "Close", "Skip zero"
1568    
1569    
1570  /* 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 860  in UTF-8 mode. The code that uses this t Line 1581  in UTF-8 mode. The code that uses this t
1581    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1582    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1583    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1584    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1585    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1586    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1,                             /* \X                                     */ \
1587      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1588    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1589    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1590    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1591    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1592    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1593    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1594      4, 4, 4,                       /* upto, minupto, exact      ** mode      */ \
1595    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
1596      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1597      4, 4, 4,                       /* upto I, minupto I, exact I             */ \
1598      2, 2, 2, 4,                    /* *+I, ++I, ?+I, upto+I                  */ \
1599    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1600    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1601    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
1602    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2, 2, 2, 4,                    /* Possessive NOT *, +, ?, upto           */ \
1603      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1604      4, 4, 4,                       /* NOT upto I, minupto I, exact I         */ \
1605      2, 2, 2, 4,                    /* Possessive NOT *I, +I, ?I, upto I      */ \
1606    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1607    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1608    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
# Line 885  in UTF-8 mode. The code that uses this t Line 1614  in UTF-8 mode. The code that uses this t
1614   33,                             /* NCLASS                                 */ \   33,                             /* NCLASS                                 */ \
1615    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1616    3,                             /* REF                                    */ \    3,                             /* REF                                    */ \
1617      3,                             /* REFI                                   */ \
1618    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1619    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1620    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
1621    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
1622    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
1623    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
1624      1+LINK_SIZE,                   /* KetRpos                                */ \
1625      1+LINK_SIZE,                   /* Reverse                                */ \
1626    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
1627    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
1628    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1629    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
   1+LINK_SIZE,                   /* Reverse                                */ \  
1630    1+LINK_SIZE,                   /* ONCE                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1631      1+LINK_SIZE,                   /* ONCE_NC                                */ \
1632    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
1633      1+LINK_SIZE,                   /* BRAPOS                                 */ \
1634    3+LINK_SIZE,                   /* CBRA                                   */ \    3+LINK_SIZE,                   /* CBRA                                   */ \
1635      3+LINK_SIZE,                   /* CBRAPOS                                */ \
1636    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1637    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1638      1+LINK_SIZE,                   /* SBRAPOS                                */ \
1639    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1640      3+LINK_SIZE,                   /* SCBRAPOS                               */ \
1641    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1642    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1643    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1644    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1645    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
1646    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1647    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */    1, 3,                          /* SKIP, SKIP_ARG                         */ \
1648      1, 3,                          /* THEN, THEN_ARG                         */ \
1649      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
1650      3, 1                           /* CLOSE, SKIPZERO  */
1651    
1652    /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1653  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  condition. */
1654    
1655  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1656    
1657  /* 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
1658  tracked. */  easily be tracked. When a new number is added, the table called eint in
1659    pcreposix.c must be updated. */
1660    
1661  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1662         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
# Line 924  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 1664  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1664         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1665         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1666         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1667         ERR60, ERR61, ERR62, ERR63, ERR64 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
1668           ERR70, ERR71, ERR72, ERRCOUNT };
1669    
1670  /* 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
1671  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 940  Because people can now save and re-use c Line 1681  Because people can now save and re-use c
1681  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
1682  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
1683  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1684  NOTE NOTE NOTE:  NOTE NOTE NOTE
1685  */  */
1686    
1687  typedef struct real_pcre {  typedef struct real_pcre {
# Line 967  remark (see NOTE above) about extending Line 1708  remark (see NOTE above) about extending
1708    
1709  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1710    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1711    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1712    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1713      pcre_uint32 minlength;          /* Minimum subject length */
1714  } pcre_study_data;  } pcre_study_data;
1715    
1716    /* Structure for building a chain of open capturing subpatterns during
1717    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1718    encountered. This is also used to identify subpatterns that contain recursive
1719    back references to themselves, so that they can be made atomic. */
1720    
1721    typedef struct open_capitem {
1722      struct open_capitem *next;    /* Chain link */
1723      pcre_uint16 number;           /* Capture number */
1724      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1725    } open_capitem;
1726    
1727  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1728  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1729    
# Line 983  typedef struct compile_data { Line 1736  typedef struct compile_data {
1736    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1737    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1738    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1739      open_capitem *open_caps;      /* Chain of open capture items */
1740    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1741    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1742    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
1743    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
1744      int  workspace_size;          /* Size of workspace */
1745    int  bracount;                /* Count of capturing parens as we compile */    int  bracount;                /* Count of capturing parens as we compile */
1746    int  final_bracount;          /* Saved value after first pass */    int  final_bracount;          /* Saved value after first pass */
1747    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
1748    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
1749      int  assert_depth;            /* Depth of nested assertions */
1750    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
1751    int  external_flags;          /* External flag bits to be set */    int  external_flags;          /* External flag bits to be set */
1752    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1753    BOOL had_accept;              /* (*ACCEPT) encountered */    BOOL had_accept;              /* (*ACCEPT) encountered */
1754      BOOL check_lookbehind;        /* Lookbehinds need later checking */
1755    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1756    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1757    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
1758  } compile_data;  } compile_data;
1759    
1760  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
1761  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
1762    
1763  typedef struct branch_chain {  typedef struct branch_chain {
1764    struct branch_chain *outer;    struct branch_chain *outer;
1765    uschar *current;    uschar *current_branch;
1766  } branch_chain;  } branch_chain;
1767    
1768  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
1769  call within the pattern. */  call within the pattern; used by pcre_exec(). */
1770    
1771  typedef struct recursion_info {  typedef struct recursion_info {
1772    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1773    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
1774    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
1775    USPTR save_start;             /* Old value of mstart */    int saved_max;                  /* Number of saved offsets */
1776    int *offset_save;             /* Pointer to start of saved offsets */    USPTR subject_position;         /* Position at start of recursion */
   int saved_max;                /* Number of saved offsets */  
1777  } recursion_info;  } recursion_info;
1778    
1779    /* A similar structure for pcre_dfa_exec(). */
1780    
1781    typedef struct dfa_recursion_info {
1782      struct dfa_recursion_info *prevrec;
1783      int group_num;
1784      USPTR subject_position;
1785    } dfa_recursion_info;
1786    
1787  /* 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
1788  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
1789  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops; used by
1790    pcre_exec(). */
1791    
1792  typedef struct eptrblock {  typedef struct eptrblock {
1793    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
# Line 1042  typedef struct match_data { Line 1807  typedef struct match_data {
1807    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1808    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1809    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1810      int    name_count;            /* Number of names in name table */
1811      int    name_entry_size;       /* Size of entry in names table */
1812      uschar *name_table;           /* Table of names */
1813    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1814    const uschar *lcc;            /* Points to lower casing table */    const  uschar *lcc;           /* Points to lower casing table */
1815    const uschar *ctypes;         /* Points to table of type maps */    const  uschar *ctypes;        /* Points to table of type maps */
1816    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
1817    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1818    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1819    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1820    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1821      BOOL   use_ucp;               /* PCRE_UCP flag */
1822    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1823    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1824    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1825    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1826    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1827    const uschar *start_code;     /* For use when recursing */    BOOL   hasthen;               /* Pattern contains (*THEN) */
1828      BOOL   ignore_skip_arg;       /* For re-run when SKIP name not found */
1829      const  uschar *start_code;    /* For use when recursing */
1830    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1831    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1832    USPTR  start_match_ptr;       /* Start of matched string */    USPTR  start_match_ptr;       /* Start of matched string */
1833    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1834      USPTR  start_used_ptr;        /* Earliest consulted character */
1835      int    partial;               /* PARTIAL options */
1836    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1837    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1838    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
1839      int    match_function_type;   /* Set for certain special calls of MATCH() */
1840    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
1841    int    eptrn;                 /* Next free eptrblock */    int    eptrn;                 /* Next free eptrblock */
1842    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1843    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1844      const  uschar *mark;          /* Mark pointer to pass back on success */
1845      const  uschar *nomatch_mark;  /* Mark pointer to pass back on failure */
1846      const  uschar *once_target;   /* Where to back up to for atomic groups */
1847  } match_data;  } match_data;
1848    
1849  /* 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
1850  functions. */  functions. */
1851    
1852  typedef struct dfa_match_data {  typedef struct dfa_match_data {
1853    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;      /* Start of the compiled pattern */
1854    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;   /* Start of the subject string */
1855    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;     /* End of subject string */
1856    const uschar *tables;         /* Character tables */    const uschar *start_used_ptr;  /* Earliest consulted character */
1857    int   moptions;               /* Match options */    const uschar *tables;          /* Character tables */
1858    int   poptions;               /* Pattern options */    int   start_offset;            /* The start offset value */
1859    int    nltype;                /* Newline type */    int   moptions;                /* Match options */
1860    int    nllen;                 /* Newline string length */    int   poptions;                /* Pattern options */
1861    uschar nl[4];                 /* Newline string when fixed */    int    nltype;                 /* Newline type */
1862    void  *callout_data;          /* To pass back to callouts */    int    nllen;                  /* Newline string length */
1863      uschar nl[4];                  /* Newline string when fixed */
1864      void  *callout_data;           /* To pass back to callouts */
1865      dfa_recursion_info *recursive; /* Linked list of recursion data */
1866  } dfa_match_data;  } dfa_match_data;
1867    
1868  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 1140  extern const int    _pcre_utf8_table2[]; Line 1920  extern const int    _pcre_utf8_table2[];
1920  extern const int    _pcre_utf8_table3[];  extern const int    _pcre_utf8_table3[];
1921  extern const uschar _pcre_utf8_table4[];  extern const uschar _pcre_utf8_table4[];
1922    
1923    #ifdef SUPPORT_JIT
1924    extern const uschar _pcre_utf8_char_sizes[];
1925    #endif
1926    
1927  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1928    
1929  extern const char   _pcre_utt_names[];  extern const char   _pcre_utt_names[];
# Line 1155  extern const uschar _pcre_OP_lengths[]; Line 1939  extern const uschar _pcre_OP_lengths[];
1939  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
1940  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1941    
1942  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1943                        int *, BOOL);  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);
1944  extern int          _pcre_ord2utf8(int, uschar *);  extern int           _pcre_ord2utf8(int, uschar *);
1945  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1946                        const pcre_study_data *, pcre_study_data *);                         const pcre_study_data *, pcre_study_data *);
1947  extern int          _pcre_valid_utf8(const uschar *, int);  extern int           _pcre_valid_utf8(USPTR, int, int *);
1948  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);
1949                        int *, BOOL);  extern BOOL          _pcre_xclass(int, const uschar *);
1950  extern BOOL         _pcre_xclass(int, const uschar *);  
1951    #ifdef SUPPORT_JIT
1952    extern void          _pcre_jit_compile(const real_pcre *, pcre_extra *);
1953    extern int           _pcre_jit_exec(const real_pcre *, void *, PCRE_SPTR,
1954                            int, int, int, int, int *, int);
1955    extern void          _pcre_jit_free(void *);
1956    extern int           _pcre_jit_get_size(void *);
1957    #endif
1958    
1959  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */
1960    
# Line 1178  extern const ucd_record  _pcre_ucd_recor Line 1968  extern const ucd_record  _pcre_ucd_recor
1968  extern const uschar      _pcre_ucd_stage1[];  extern const uschar      _pcre_ucd_stage1[];
1969  extern const pcre_uint16 _pcre_ucd_stage2[];  extern const pcre_uint16 _pcre_ucd_stage2[];
1970  extern const int         _pcre_ucp_gentype[];  extern const int         _pcre_ucp_gentype[];
1971    #ifdef SUPPORT_JIT
1972    extern const int         _pcre_ucp_typerange[];
1973    #endif
1974    
1975  /* UCD access macros */  /* UCD access macros */
1976    
1977  #define UCD_BLOCK_SIZE 128  #define UCD_BLOCK_SIZE 128
1978  #define GET_UCD(ch) (_pcre_ucd_records + \  #define GET_UCD(ch) (_pcre_ucd_records + \
1979          _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \          _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1980          UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])          UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
1981    
1982  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1983  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script

Legend:
Removed from v.384  
changed lines
  Added in v.788

  ViewVC Help
Powered by ViewVC 1.1.5