/[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 376 by ph10, Sun Mar 1 12:00:59 2009 UTC revision 667 by ph10, Mon Aug 22 14:57:32 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          c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
483              ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
484          eptr += 3; \
485          } \
486        else if ((c & 0x04) == 0) \
487        { \        { \
488        gcss -= 6; \        c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
489        c |= (eptr[gcii] & 0x3f) << gcss; \            ((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) \
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     while (gcaa-- > 0) \  
522        { \        { \
523        gcss -= 6; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
524        c |= (*eptr++ & 0x3f) << gcss; \        len++; \
525          } \
526        else if ((c & 0x10)  == 0) \
527          { \
528          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
529          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);
559      { \  
560      int gcii; \  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
561      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  pointer, incrementing length if there are extra bytes. This is called when we
562      int gcss = 6*gcaa; \  do not know if we are in UTF-8 mode. */
563      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
564      for (gcii = 1; gcii <= gcaa; gcii++) \  #define GETCHARLENTEST(c, eptr, len) \
565        { \    c = *eptr; \
566        gcss -= 6; \    if (utf8 && c >= 0xc0) GETUTF8LEN(c, eptr, len);
       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 484  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 495  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 510  are in a 16-bit flags word. */ Line 595  are in a 16-bit flags word. */
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    
598  /* Options for the "extra" block produced by pcre_study(). */  /* Flags for the "extra" block produced by pcre_study(). */
599    
600  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
601    #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
602    
603  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
604  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 520  time, run time, or study time, respectiv Line 606  time, run time, or study time, respectiv
606  #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| \
607                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
608    
609  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
610    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
611     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
612     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| \
613     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
614     PCRE_JAVASCRIPT_COMPAT)     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
615    
616  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
617    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
618     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
619       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
620    
621  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
622    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
623     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
624     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
625       PCRE_NO_START_OPTIMIZE)
626    
627  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS \
628       PCRE_STUDY_JIT_COMPILE
629    
630  /* 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
631  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 570  typedef int BOOL; Line 659  typedef int BOOL;
659  #define TRUE    1  #define TRUE    1
660  #endif  #endif
661    
662    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
663    character constants like '*' because the compiler would emit their EBCDIC code,
664    which is different from their ASCII/UTF-8 code. Instead we define macros for
665    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
666    is enabled. When UTF-8 support is not enabled, the definitions use character
667    literals. Both character and string versions of each character are needed, and
668    there are some longer strings as well.
669    
670    This means that, on EBCDIC platforms, the PCRE library can handle either
671    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
672    would need different lookups depending on whether PCRE_UTF8 was set or not.
673    This would make it impossible to use characters in switch/case statements,
674    which would reduce performance. For a theoretical use (which nobody has asked
675    for) in a minority area (EBCDIC platforms), this is not sensible. Any
676    application that did need both could compile two versions of the library, using
677    macros to give the functions distinct names. */
678    
679    #ifndef SUPPORT_UTF8
680    
681    /* UTF-8 support is not enabled; use the platform-dependent character literals
682    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
683    
684    #define CHAR_HT                     '\t'
685    #define CHAR_VT                     '\v'
686    #define CHAR_FF                     '\f'
687    #define CHAR_CR                     '\r'
688    #define CHAR_NL                     '\n'
689    #define CHAR_BS                     '\b'
690    #define CHAR_BEL                    '\a'
691    #ifdef EBCDIC
692    #define CHAR_ESC                    '\047'
693    #define CHAR_DEL                    '\007'
694    #else
695    #define CHAR_ESC                    '\033'
696    #define CHAR_DEL                    '\177'
697    #endif
698    
699    #define CHAR_SPACE                  ' '
700    #define CHAR_EXCLAMATION_MARK       '!'
701    #define CHAR_QUOTATION_MARK         '"'
702    #define CHAR_NUMBER_SIGN            '#'
703    #define CHAR_DOLLAR_SIGN            '$'
704    #define CHAR_PERCENT_SIGN           '%'
705    #define CHAR_AMPERSAND              '&'
706    #define CHAR_APOSTROPHE             '\''
707    #define CHAR_LEFT_PARENTHESIS       '('
708    #define CHAR_RIGHT_PARENTHESIS      ')'
709    #define CHAR_ASTERISK               '*'
710    #define CHAR_PLUS                   '+'
711    #define CHAR_COMMA                  ','
712    #define CHAR_MINUS                  '-'
713    #define CHAR_DOT                    '.'
714    #define CHAR_SLASH                  '/'
715    #define CHAR_0                      '0'
716    #define CHAR_1                      '1'
717    #define CHAR_2                      '2'
718    #define CHAR_3                      '3'
719    #define CHAR_4                      '4'
720    #define CHAR_5                      '5'
721    #define CHAR_6                      '6'
722    #define CHAR_7                      '7'
723    #define CHAR_8                      '8'
724    #define CHAR_9                      '9'
725    #define CHAR_COLON                  ':'
726    #define CHAR_SEMICOLON              ';'
727    #define CHAR_LESS_THAN_SIGN         '<'
728    #define CHAR_EQUALS_SIGN            '='
729    #define CHAR_GREATER_THAN_SIGN      '>'
730    #define CHAR_QUESTION_MARK          '?'
731    #define CHAR_COMMERCIAL_AT          '@'
732    #define CHAR_A                      'A'
733    #define CHAR_B                      'B'
734    #define CHAR_C                      'C'
735    #define CHAR_D                      'D'
736    #define CHAR_E                      'E'
737    #define CHAR_F                      'F'
738    #define CHAR_G                      'G'
739    #define CHAR_H                      'H'
740    #define CHAR_I                      'I'
741    #define CHAR_J                      'J'
742    #define CHAR_K                      'K'
743    #define CHAR_L                      'L'
744    #define CHAR_M                      'M'
745    #define CHAR_N                      'N'
746    #define CHAR_O                      'O'
747    #define CHAR_P                      'P'
748    #define CHAR_Q                      'Q'
749    #define CHAR_R                      'R'
750    #define CHAR_S                      'S'
751    #define CHAR_T                      'T'
752    #define CHAR_U                      'U'
753    #define CHAR_V                      'V'
754    #define CHAR_W                      'W'
755    #define CHAR_X                      'X'
756    #define CHAR_Y                      'Y'
757    #define CHAR_Z                      'Z'
758    #define CHAR_LEFT_SQUARE_BRACKET    '['
759    #define CHAR_BACKSLASH              '\\'
760    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
761    #define CHAR_CIRCUMFLEX_ACCENT      '^'
762    #define CHAR_UNDERSCORE             '_'
763    #define CHAR_GRAVE_ACCENT           '`'
764    #define CHAR_a                      'a'
765    #define CHAR_b                      'b'
766    #define CHAR_c                      'c'
767    #define CHAR_d                      'd'
768    #define CHAR_e                      'e'
769    #define CHAR_f                      'f'
770    #define CHAR_g                      'g'
771    #define CHAR_h                      'h'
772    #define CHAR_i                      'i'
773    #define CHAR_j                      'j'
774    #define CHAR_k                      'k'
775    #define CHAR_l                      'l'
776    #define CHAR_m                      'm'
777    #define CHAR_n                      'n'
778    #define CHAR_o                      'o'
779    #define CHAR_p                      'p'
780    #define CHAR_q                      'q'
781    #define CHAR_r                      'r'
782    #define CHAR_s                      's'
783    #define CHAR_t                      't'
784    #define CHAR_u                      'u'
785    #define CHAR_v                      'v'
786    #define CHAR_w                      'w'
787    #define CHAR_x                      'x'
788    #define CHAR_y                      'y'
789    #define CHAR_z                      'z'
790    #define CHAR_LEFT_CURLY_BRACKET     '{'
791    #define CHAR_VERTICAL_LINE          '|'
792    #define CHAR_RIGHT_CURLY_BRACKET    '}'
793    #define CHAR_TILDE                  '~'
794    
795    #define STR_HT                      "\t"
796    #define STR_VT                      "\v"
797    #define STR_FF                      "\f"
798    #define STR_CR                      "\r"
799    #define STR_NL                      "\n"
800    #define STR_BS                      "\b"
801    #define STR_BEL                     "\a"
802    #ifdef EBCDIC
803    #define STR_ESC                     "\047"
804    #define STR_DEL                     "\007"
805    #else
806    #define STR_ESC                     "\033"
807    #define STR_DEL                     "\177"
808    #endif
809    
810    #define STR_SPACE                   " "
811    #define STR_EXCLAMATION_MARK        "!"
812    #define STR_QUOTATION_MARK          "\""
813    #define STR_NUMBER_SIGN             "#"
814    #define STR_DOLLAR_SIGN             "$"
815    #define STR_PERCENT_SIGN            "%"
816    #define STR_AMPERSAND               "&"
817    #define STR_APOSTROPHE              "'"
818    #define STR_LEFT_PARENTHESIS        "("
819    #define STR_RIGHT_PARENTHESIS       ")"
820    #define STR_ASTERISK                "*"
821    #define STR_PLUS                    "+"
822    #define STR_COMMA                   ","
823    #define STR_MINUS                   "-"
824    #define STR_DOT                     "."
825    #define STR_SLASH                   "/"
826    #define STR_0                       "0"
827    #define STR_1                       "1"
828    #define STR_2                       "2"
829    #define STR_3                       "3"
830    #define STR_4                       "4"
831    #define STR_5                       "5"
832    #define STR_6                       "6"
833    #define STR_7                       "7"
834    #define STR_8                       "8"
835    #define STR_9                       "9"
836    #define STR_COLON                   ":"
837    #define STR_SEMICOLON               ";"
838    #define STR_LESS_THAN_SIGN          "<"
839    #define STR_EQUALS_SIGN             "="
840    #define STR_GREATER_THAN_SIGN       ">"
841    #define STR_QUESTION_MARK           "?"
842    #define STR_COMMERCIAL_AT           "@"
843    #define STR_A                       "A"
844    #define STR_B                       "B"
845    #define STR_C                       "C"
846    #define STR_D                       "D"
847    #define STR_E                       "E"
848    #define STR_F                       "F"
849    #define STR_G                       "G"
850    #define STR_H                       "H"
851    #define STR_I                       "I"
852    #define STR_J                       "J"
853    #define STR_K                       "K"
854    #define STR_L                       "L"
855    #define STR_M                       "M"
856    #define STR_N                       "N"
857    #define STR_O                       "O"
858    #define STR_P                       "P"
859    #define STR_Q                       "Q"
860    #define STR_R                       "R"
861    #define STR_S                       "S"
862    #define STR_T                       "T"
863    #define STR_U                       "U"
864    #define STR_V                       "V"
865    #define STR_W                       "W"
866    #define STR_X                       "X"
867    #define STR_Y                       "Y"
868    #define STR_Z                       "Z"
869    #define STR_LEFT_SQUARE_BRACKET     "["
870    #define STR_BACKSLASH               "\\"
871    #define STR_RIGHT_SQUARE_BRACKET    "]"
872    #define STR_CIRCUMFLEX_ACCENT       "^"
873    #define STR_UNDERSCORE              "_"
874    #define STR_GRAVE_ACCENT            "`"
875    #define STR_a                       "a"
876    #define STR_b                       "b"
877    #define STR_c                       "c"
878    #define STR_d                       "d"
879    #define STR_e                       "e"
880    #define STR_f                       "f"
881    #define STR_g                       "g"
882    #define STR_h                       "h"
883    #define STR_i                       "i"
884    #define STR_j                       "j"
885    #define STR_k                       "k"
886    #define STR_l                       "l"
887    #define STR_m                       "m"
888    #define STR_n                       "n"
889    #define STR_o                       "o"
890    #define STR_p                       "p"
891    #define STR_q                       "q"
892    #define STR_r                       "r"
893    #define STR_s                       "s"
894    #define STR_t                       "t"
895    #define STR_u                       "u"
896    #define STR_v                       "v"
897    #define STR_w                       "w"
898    #define STR_x                       "x"
899    #define STR_y                       "y"
900    #define STR_z                       "z"
901    #define STR_LEFT_CURLY_BRACKET      "{"
902    #define STR_VERTICAL_LINE           "|"
903    #define STR_RIGHT_CURLY_BRACKET     "}"
904    #define STR_TILDE                   "~"
905    
906    #define STRING_ACCEPT0              "ACCEPT\0"
907    #define STRING_COMMIT0              "COMMIT\0"
908    #define STRING_F0                   "F\0"
909    #define STRING_FAIL0                "FAIL\0"
910    #define STRING_MARK0                "MARK\0"
911    #define STRING_PRUNE0               "PRUNE\0"
912    #define STRING_SKIP0                "SKIP\0"
913    #define STRING_THEN                 "THEN"
914    
915    #define STRING_alpha0               "alpha\0"
916    #define STRING_lower0               "lower\0"
917    #define STRING_upper0               "upper\0"
918    #define STRING_alnum0               "alnum\0"
919    #define STRING_ascii0               "ascii\0"
920    #define STRING_blank0               "blank\0"
921    #define STRING_cntrl0               "cntrl\0"
922    #define STRING_digit0               "digit\0"
923    #define STRING_graph0               "graph\0"
924    #define STRING_print0               "print\0"
925    #define STRING_punct0               "punct\0"
926    #define STRING_space0               "space\0"
927    #define STRING_word0                "word\0"
928    #define STRING_xdigit               "xdigit"
929    
930    #define STRING_DEFINE               "DEFINE"
931    
932    #define STRING_CR_RIGHTPAR             "CR)"
933    #define STRING_LF_RIGHTPAR             "LF)"
934    #define STRING_CRLF_RIGHTPAR           "CRLF)"
935    #define STRING_ANY_RIGHTPAR            "ANY)"
936    #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
937    #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
938    #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
939    #define STRING_UTF8_RIGHTPAR           "UTF8)"
940    #define STRING_UCP_RIGHTPAR            "UCP)"
941    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
942    
943    #else  /* SUPPORT_UTF8 */
944    
945    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
946    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
947    only. */
948    
949    #define CHAR_HT                     '\011'
950    #define CHAR_VT                     '\013'
951    #define CHAR_FF                     '\014'
952    #define CHAR_CR                     '\015'
953    #define CHAR_NL                     '\012'
954    #define CHAR_BS                     '\010'
955    #define CHAR_BEL                    '\007'
956    #define CHAR_ESC                    '\033'
957    #define CHAR_DEL                    '\177'
958    
959    #define CHAR_SPACE                  '\040'
960    #define CHAR_EXCLAMATION_MARK       '\041'
961    #define CHAR_QUOTATION_MARK         '\042'
962    #define CHAR_NUMBER_SIGN            '\043'
963    #define CHAR_DOLLAR_SIGN            '\044'
964    #define CHAR_PERCENT_SIGN           '\045'
965    #define CHAR_AMPERSAND              '\046'
966    #define CHAR_APOSTROPHE             '\047'
967    #define CHAR_LEFT_PARENTHESIS       '\050'
968    #define CHAR_RIGHT_PARENTHESIS      '\051'
969    #define CHAR_ASTERISK               '\052'
970    #define CHAR_PLUS                   '\053'
971    #define CHAR_COMMA                  '\054'
972    #define CHAR_MINUS                  '\055'
973    #define CHAR_DOT                    '\056'
974    #define CHAR_SLASH                  '\057'
975    #define CHAR_0                      '\060'
976    #define CHAR_1                      '\061'
977    #define CHAR_2                      '\062'
978    #define CHAR_3                      '\063'
979    #define CHAR_4                      '\064'
980    #define CHAR_5                      '\065'
981    #define CHAR_6                      '\066'
982    #define CHAR_7                      '\067'
983    #define CHAR_8                      '\070'
984    #define CHAR_9                      '\071'
985    #define CHAR_COLON                  '\072'
986    #define CHAR_SEMICOLON              '\073'
987    #define CHAR_LESS_THAN_SIGN         '\074'
988    #define CHAR_EQUALS_SIGN            '\075'
989    #define CHAR_GREATER_THAN_SIGN      '\076'
990    #define CHAR_QUESTION_MARK          '\077'
991    #define CHAR_COMMERCIAL_AT          '\100'
992    #define CHAR_A                      '\101'
993    #define CHAR_B                      '\102'
994    #define CHAR_C                      '\103'
995    #define CHAR_D                      '\104'
996    #define CHAR_E                      '\105'
997    #define CHAR_F                      '\106'
998    #define CHAR_G                      '\107'
999    #define CHAR_H                      '\110'
1000    #define CHAR_I                      '\111'
1001    #define CHAR_J                      '\112'
1002    #define CHAR_K                      '\113'
1003    #define CHAR_L                      '\114'
1004    #define CHAR_M                      '\115'
1005    #define CHAR_N                      '\116'
1006    #define CHAR_O                      '\117'
1007    #define CHAR_P                      '\120'
1008    #define CHAR_Q                      '\121'
1009    #define CHAR_R                      '\122'
1010    #define CHAR_S                      '\123'
1011    #define CHAR_T                      '\124'
1012    #define CHAR_U                      '\125'
1013    #define CHAR_V                      '\126'
1014    #define CHAR_W                      '\127'
1015    #define CHAR_X                      '\130'
1016    #define CHAR_Y                      '\131'
1017    #define CHAR_Z                      '\132'
1018    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
1019    #define CHAR_BACKSLASH              '\134'
1020    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
1021    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
1022    #define CHAR_UNDERSCORE             '\137'
1023    #define CHAR_GRAVE_ACCENT           '\140'
1024    #define CHAR_a                      '\141'
1025    #define CHAR_b                      '\142'
1026    #define CHAR_c                      '\143'
1027    #define CHAR_d                      '\144'
1028    #define CHAR_e                      '\145'
1029    #define CHAR_f                      '\146'
1030    #define CHAR_g                      '\147'
1031    #define CHAR_h                      '\150'
1032    #define CHAR_i                      '\151'
1033    #define CHAR_j                      '\152'
1034    #define CHAR_k                      '\153'
1035    #define CHAR_l                      '\154'
1036    #define CHAR_m                      '\155'
1037    #define CHAR_n                      '\156'
1038    #define CHAR_o                      '\157'
1039    #define CHAR_p                      '\160'
1040    #define CHAR_q                      '\161'
1041    #define CHAR_r                      '\162'
1042    #define CHAR_s                      '\163'
1043    #define CHAR_t                      '\164'
1044    #define CHAR_u                      '\165'
1045    #define CHAR_v                      '\166'
1046    #define CHAR_w                      '\167'
1047    #define CHAR_x                      '\170'
1048    #define CHAR_y                      '\171'
1049    #define CHAR_z                      '\172'
1050    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1051    #define CHAR_VERTICAL_LINE          '\174'
1052    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1053    #define CHAR_TILDE                  '\176'
1054    
1055    #define STR_HT                      "\011"
1056    #define STR_VT                      "\013"
1057    #define STR_FF                      "\014"
1058    #define STR_CR                      "\015"
1059    #define STR_NL                      "\012"
1060    #define STR_BS                      "\010"
1061    #define STR_BEL                     "\007"
1062    #define STR_ESC                     "\033"
1063    #define STR_DEL                     "\177"
1064    
1065    #define STR_SPACE                   "\040"
1066    #define STR_EXCLAMATION_MARK        "\041"
1067    #define STR_QUOTATION_MARK          "\042"
1068    #define STR_NUMBER_SIGN             "\043"
1069    #define STR_DOLLAR_SIGN             "\044"
1070    #define STR_PERCENT_SIGN            "\045"
1071    #define STR_AMPERSAND               "\046"
1072    #define STR_APOSTROPHE              "\047"
1073    #define STR_LEFT_PARENTHESIS        "\050"
1074    #define STR_RIGHT_PARENTHESIS       "\051"
1075    #define STR_ASTERISK                "\052"
1076    #define STR_PLUS                    "\053"
1077    #define STR_COMMA                   "\054"
1078    #define STR_MINUS                   "\055"
1079    #define STR_DOT                     "\056"
1080    #define STR_SLASH                   "\057"
1081    #define STR_0                       "\060"
1082    #define STR_1                       "\061"
1083    #define STR_2                       "\062"
1084    #define STR_3                       "\063"
1085    #define STR_4                       "\064"
1086    #define STR_5                       "\065"
1087    #define STR_6                       "\066"
1088    #define STR_7                       "\067"
1089    #define STR_8                       "\070"
1090    #define STR_9                       "\071"
1091    #define STR_COLON                   "\072"
1092    #define STR_SEMICOLON               "\073"
1093    #define STR_LESS_THAN_SIGN          "\074"
1094    #define STR_EQUALS_SIGN             "\075"
1095    #define STR_GREATER_THAN_SIGN       "\076"
1096    #define STR_QUESTION_MARK           "\077"
1097    #define STR_COMMERCIAL_AT           "\100"
1098    #define STR_A                       "\101"
1099    #define STR_B                       "\102"
1100    #define STR_C                       "\103"
1101    #define STR_D                       "\104"
1102    #define STR_E                       "\105"
1103    #define STR_F                       "\106"
1104    #define STR_G                       "\107"
1105    #define STR_H                       "\110"
1106    #define STR_I                       "\111"
1107    #define STR_J                       "\112"
1108    #define STR_K                       "\113"
1109    #define STR_L                       "\114"
1110    #define STR_M                       "\115"
1111    #define STR_N                       "\116"
1112    #define STR_O                       "\117"
1113    #define STR_P                       "\120"
1114    #define STR_Q                       "\121"
1115    #define STR_R                       "\122"
1116    #define STR_S                       "\123"
1117    #define STR_T                       "\124"
1118    #define STR_U                       "\125"
1119    #define STR_V                       "\126"
1120    #define STR_W                       "\127"
1121    #define STR_X                       "\130"
1122    #define STR_Y                       "\131"
1123    #define STR_Z                       "\132"
1124    #define STR_LEFT_SQUARE_BRACKET     "\133"
1125    #define STR_BACKSLASH               "\134"
1126    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1127    #define STR_CIRCUMFLEX_ACCENT       "\136"
1128    #define STR_UNDERSCORE              "\137"
1129    #define STR_GRAVE_ACCENT            "\140"
1130    #define STR_a                       "\141"
1131    #define STR_b                       "\142"
1132    #define STR_c                       "\143"
1133    #define STR_d                       "\144"
1134    #define STR_e                       "\145"
1135    #define STR_f                       "\146"
1136    #define STR_g                       "\147"
1137    #define STR_h                       "\150"
1138    #define STR_i                       "\151"
1139    #define STR_j                       "\152"
1140    #define STR_k                       "\153"
1141    #define STR_l                       "\154"
1142    #define STR_m                       "\155"
1143    #define STR_n                       "\156"
1144    #define STR_o                       "\157"
1145    #define STR_p                       "\160"
1146    #define STR_q                       "\161"
1147    #define STR_r                       "\162"
1148    #define STR_s                       "\163"
1149    #define STR_t                       "\164"
1150    #define STR_u                       "\165"
1151    #define STR_v                       "\166"
1152    #define STR_w                       "\167"
1153    #define STR_x                       "\170"
1154    #define STR_y                       "\171"
1155    #define STR_z                       "\172"
1156    #define STR_LEFT_CURLY_BRACKET      "\173"
1157    #define STR_VERTICAL_LINE           "\174"
1158    #define STR_RIGHT_CURLY_BRACKET     "\175"
1159    #define STR_TILDE                   "\176"
1160    
1161    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1162    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1163    #define STRING_F0                   STR_F "\0"
1164    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1165    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1166    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1167    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1168    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1169    
1170    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1171    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1172    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1173    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1174    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1175    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1176    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1177    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1178    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1179    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1180    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1181    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1182    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1183    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1184    
1185    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1186    
1187    #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1188    #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1189    #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1190    #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1191    #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1192    #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
1193    #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
1194    #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1195    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1196    #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
1197    
1198    #endif  /* SUPPORT_UTF8 */
1199    
1200  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1201    
1202  #ifndef ESC_e  #ifndef ESC_e
1203  #define ESC_e 27  #define ESC_e CHAR_ESC
1204  #endif  #endif
1205    
1206  #ifndef ESC_f  #ifndef ESC_f
1207  #define ESC_f '\f'  #define ESC_f CHAR_FF
1208  #endif  #endif
1209    
1210  #ifndef ESC_n  #ifndef ESC_n
1211  #define ESC_n '\n'  #define ESC_n CHAR_NL
1212  #endif  #endif
1213    
1214  #ifndef ESC_r  #ifndef ESC_r
1215  #define ESC_r '\r'  #define ESC_r CHAR_CR
1216  #endif  #endif
1217    
1218  /* 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
1219  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1220    
1221  #ifndef ESC_tee  #ifndef ESC_tee
1222  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1223  #endif  #endif
1224    
1225  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
1226    
1227  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1228  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1229  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1230  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1231  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1232    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1233    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1234    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1235    #define PT_WORD       8    /* Word - L plus N plus underscore */
1236    
1237  /* 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
1238  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 618  contain UTF-8 characters with values gre Line 1249  contain UTF-8 characters with values gre
1249  /* 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
1250  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
1251  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
1252  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
1253  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
1254  (which is used for [^] in JavaScript compatibility mode).  used for [^] in JavaScript compatibility mode. In non-DOTALL mode, "." behaves
1255    like \N.
1256    
1257    The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1258    when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1259    They must be contiguous, and remain in order so that the replacements can be
1260    looked up from a table.
1261    
1262  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
1263  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 630  put in between that don't consume a char Line 1267  put in between that don't consume a char
1267  */  */
1268    
1269  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,
1270         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,
1271         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,
1272           ESC_E, ESC_Q, ESC_g, ESC_k,
1273           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1274         ESC_REF };         ESC_REF };
1275    
   
1276  /* 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
1277  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.
1278    
1279  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1280  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
1281  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1282    
1283  enum {  enum {
1284    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 658  enum { Line 1296  enum {
1296    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1297    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1298    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1299    OP_ANY,            /* 12 Match any character (subject to DOTALL) */  
1300    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */    OP_ANY,            /* 12 Match any character except newline */
1301      OP_ALLANY,         /* 13 Match any character */
1302    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 */
1303    OP_NOTPROP,        /* 15 \P (not Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1304    OP_PROP,           /* 16 \p (Unicode property) */    OP_PROP,           /* 16 \p (Unicode property) */
# Line 672  enum { Line 1311  enum {
1311    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. */
1312    OP_EOD,            /* 24 End of data: \z */    OP_EOD,            /* 24 End of data: \z */
1313    
1314    OP_OPT,            /* 25 Set runtime options */    OP_CIRC,           /* 25 Start of line - not multiline */
1315    OP_CIRC,           /* 26 Start of line - varies with multiline switch */    OP_CIRCM,          /* 26 Start of line - multiline */
1316    OP_DOLL,           /* 27 End of line - varies with multiline switch */    OP_DOLL,           /* 27 End of line - not multiline */
1317    OP_CHAR,           /* 28 Match one character, casefully */    OP_DOLLM,          /* 28 End of line - multiline */
1318    OP_CHARNC,         /* 29 Match one character, caselessly */    OP_CHAR,           /* 29 Match one character, casefully */
1319    OP_NOT,            /* 30 Match one character, not the following one */    OP_CHARI,          /* 30 Match one character, caselessly */
1320      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1321    OP_STAR,           /* 31 The maximizing and minimizing versions of */    OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1322    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */  
1323    OP_PLUS,           /* 33 the minimizing one second. */    /* The following sets of 13 opcodes must always be kept in step because
1324    OP_MINPLUS,        /* 34 This first set applies to single characters.*/    the offset from the first one is used to generate the others. */
1325    OP_QUERY,          /* 35 */  
1326    OP_MINQUERY,       /* 36 */    /**** Single characters, caseful, must precede the caseless ones ****/
1327    
1328    OP_UPTO,           /* 37 From 0 to n matches */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1329    OP_MINUPTO,        /* 38 */    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1330    OP_EXACT,          /* 39 Exactly n matches */    OP_PLUS,           /* 35 the minimizing one second. */
1331      OP_MINPLUS,        /* 36 */
1332    OP_POSSTAR,        /* 40 Possessified star */    OP_QUERY,          /* 37 */
1333    OP_POSPLUS,        /* 41 Possessified plus */    OP_MINQUERY,       /* 38 */
1334    OP_POSQUERY,       /* 42 Posesssified query */  
1335    OP_POSUPTO,        /* 43 Possessified upto */    OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1336      OP_MINUPTO,        /* 40 */
1337    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */    OP_EXACT,          /* 41 Exactly n matches */
1338    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */  
1339    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */    OP_POSSTAR,        /* 42 Possessified star, caseful */
1340    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */    OP_POSPLUS,        /* 43 Possessified plus, caseful */
1341    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
1342    OP_NOTMINQUERY,    /* 49 */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
1343    
1344    OP_NOTUPTO,        /* 50 From 0 to n matches */    /**** Single characters, caseless, must follow the caseful ones */
1345    OP_NOTMINUPTO,     /* 51 */  
1346    OP_NOTEXACT,       /* 52 Exactly n matches */    OP_STARI,          /* 46 */
1347      OP_MINSTARI,       /* 47 */
1348    OP_NOTPOSSTAR,     /* 53 Possessified versions */    OP_PLUSI,          /* 48 */
1349    OP_NOTPOSPLUS,     /* 54 */    OP_MINPLUSI,       /* 49 */
1350    OP_NOTPOSQUERY,    /* 55 */    OP_QUERYI,         /* 50 */
1351    OP_NOTPOSUPTO,     /* 56 */    OP_MINQUERYI,      /* 51 */
1352    
1353    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */    OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1354    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */    OP_MINUPTOI,       /* 53 */
1355    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */    OP_EXACTI,         /* 54 */
1356    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */  
1357    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */    OP_POSSTARI,       /* 55 Possessified star, caseless */
1358    OP_TYPEMINQUERY,   /* 62 */    OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1359      OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1360    OP_TYPEUPTO,       /* 63 From 0 to n matches */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1361    OP_TYPEMINUPTO,    /* 64 */  
1362    OP_TYPEEXACT,      /* 65 Exactly n matches */    /**** The negated ones must follow the non-negated ones, and match them ****/
1363      /**** Negated single character, caseful; must precede the caseless ones ****/
1364    OP_TYPEPOSSTAR,    /* 66 Possessified versions */  
1365    OP_TYPEPOSPLUS,    /* 67 */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1366    OP_TYPEPOSQUERY,   /* 68 */    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1367    OP_TYPEPOSUPTO,    /* 69 */    OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1368      OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1369    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */    OP_NOTQUERY,       /* 63 */
1370    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */    OP_NOTMINQUERY,    /* 64 */
1371    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */  
1372    OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */    OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1373    OP_CRQUERY,        /* 74 These are for character classes and back refs */    OP_NOTMINUPTO,     /* 66 */
1374    OP_CRMINQUERY,     /* 75 */    OP_NOTEXACT,       /* 67 Exactly n matches */
1375    OP_CRRANGE,        /* 76 These are different to the three sets above. */  
1376    OP_CRMINRANGE,     /* 77 */    OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1377      OP_NOTPOSPLUS,     /* 69 */
1378    OP_CLASS,          /* 78 Match a character class, chars < 256 only */    OP_NOTPOSQUERY,    /* 70 */
1379    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative    OP_NOTPOSUPTO,     /* 71 */
1380                             class - the difference is relevant only when a UTF-8  
1381                             character > 255 is encountered. */    /**** Negated single character, caseless; must follow the caseful ones ****/
1382    
1383    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the    OP_NOTSTARI,       /* 72 */
1384                             class. This does both positive and negative. */    OP_NOTMINSTARI,    /* 73 */
1385      OP_NOTPLUSI,       /* 74 */
1386    OP_REF,            /* 81 Match a back reference */    OP_NOTMINPLUSI,    /* 75 */
1387    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */    OP_NOTQUERYI,      /* 76 */
1388    OP_CALLOUT,        /* 83 Call out to external function if provided */    OP_NOTMINQUERYI,   /* 77 */
1389    
1390    OP_ALT,            /* 84 Start of alternation */    OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1391    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */    OP_NOTMINUPTOI,    /* 79 */
1392    OP_KETRMAX,        /* 86 These two must remain together and in this */    OP_NOTEXACTI,      /* 80 Exactly n matches */
1393    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */  
1394      OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1395    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    OP_NOTPOSPLUSI,    /* 82 */
1396      OP_NOTPOSQUERYI,   /* 83 */
1397    OP_ASSERT,         /* 88 Positive lookahead */    OP_NOTPOSUPTOI,    /* 84 */
1398    OP_ASSERT_NOT,     /* 89 Negative lookahead */  
1399    OP_ASSERTBACK,     /* 90 Positive lookbehind */    /**** Character types ****/
1400    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */  
1401    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1402      OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1403    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1404    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. */
1405      OP_TYPEQUERY,      /* 89 */
1406    OP_ONCE,           /* 93 Atomic group */    OP_TYPEMINQUERY,   /* 90 */
1407    OP_BRA,            /* 94 Start of non-capturing bracket */  
1408    OP_CBRA,           /* 95 Start of capturing bracket */    OP_TYPEUPTO,       /* 91 From 0 to n matches */
1409    OP_COND,           /* 96 Conditional group */    OP_TYPEMINUPTO,    /* 92 */
1410      OP_TYPEEXACT,      /* 93 Exactly n matches */
1411    
1412      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1413      OP_TYPEPOSPLUS,    /* 95 */
1414      OP_TYPEPOSQUERY,   /* 96 */
1415      OP_TYPEPOSUPTO,    /* 97 */
1416    
1417      /* These are used for character classes and back references; only the
1418      first six are the same as the sets above. */
1419    
1420      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1421      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1422      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1423      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1424      OP_CRQUERY,        /* 102 */
1425      OP_CRMINQUERY,     /* 103 */
1426    
1427      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1428      OP_CRMINRANGE,     /* 105 */
1429    
1430      /* End of quantifier opcodes */
1431    
1432      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1433      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1434                                  class - the difference is relevant only when a
1435                                  UTF-8 character > 255 is encountered. */
1436      OP_XCLASS,         /* 108 Extended class for handling UTF-8 chars within the
1437                                  class. This does both positive and negative. */
1438      OP_REF,            /* 109 Match a back reference, casefully */
1439      OP_REFI,           /* 110 Match a back reference, caselessly */
1440      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1441      OP_CALLOUT,        /* 112 Call out to external function if provided */
1442    
1443      OP_ALT,            /* 113 Start of alternation */
1444      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1445      OP_KETRMAX,        /* 115 These two must remain together and in this */
1446      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1447      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1448    
1449      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1450      asserts must remain in order. */
1451    
1452      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1453      OP_ASSERT,         /* 119 Positive lookahead */
1454      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1455      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1456      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1457    
1458      /* ONCE, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately after the
1459      assertions, with ONCE first, as there's a test for >= ONCE for a subpattern
1460      that isn't an assertion. The POS versions must immediately follow the non-POS
1461      versions in each case. */
1462    
1463      OP_ONCE,           /* 123 Atomic group */
1464      OP_BRA,            /* 124 Start of non-capturing bracket */
1465      OP_BRAPOS,         /* 125 Ditto, with unlimited, possessive repeat */
1466      OP_CBRA,           /* 126 Start of capturing bracket */
1467      OP_CBRAPOS,        /* 127 Ditto, with unlimited, possessive repeat */
1468      OP_COND,           /* 128 Conditional group */
1469    
1470    /* 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
1471    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1472    
1473    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 129 Start of non-capturing bracket, check empty  */
1474    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */    OP_SBRAPOS,        /* 130 Ditto, with unlimited, possessive repeat */
1475    OP_SCOND,          /* 99 Conditional group, check empty */    OP_SCBRA,          /* 131 Start of capturing bracket, check empty */
1476      OP_SCBRAPOS,       /* 132 Ditto, with unlimited, possessive repeat */
1477    OP_CREF,           /* 100 Used to hold a capture number as condition */    OP_SCOND,          /* 133 Conditional group, check empty */
1478    OP_RREF,           /* 101 Used to hold a recursion number as condition */  
1479    OP_DEF,            /* 102 The DEFINE condition */    /* The next two pairs must (respectively) be kept together. */
1480    
1481    OP_BRAZERO,        /* 103 These two must remain together and in this */    OP_CREF,           /* 134 Used to hold a capture number as condition */
1482    OP_BRAMINZERO,     /* 104 order. */    OP_NCREF,          /* 135 Same, but generated by a name reference*/
1483      OP_RREF,           /* 136 Used to hold a recursion number as condition */
1484      OP_NRREF,          /* 137 Same, but generated by a name reference*/
1485      OP_DEF,            /* 138 The DEFINE condition */
1486    
1487      OP_BRAZERO,        /* 139 These two must remain together and in this */
1488      OP_BRAMINZERO,     /* 140 order. */
1489      OP_BRAPOSZERO,     /* 141 */
1490    
1491    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1492    
1493    OP_PRUNE,          /* 105 */    OP_MARK,           /* 142 always has an argument */
1494    OP_SKIP,           /* 106 */    OP_PRUNE,          /* 143 */
1495    OP_THEN,           /* 107 */    OP_PRUNE_ARG,      /* 144 same, but with argument */
1496    OP_COMMIT,         /* 108 */    OP_SKIP,           /* 145 */
1497      OP_SKIP_ARG,       /* 146 same, but with argument */
1498      OP_THEN,           /* 147 */
1499      OP_THEN_ARG,       /* 148 same, but with argument */
1500      OP_COMMIT,         /* 149 */
1501    
1502    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1503    
1504    OP_FAIL,           /* 109 */    OP_FAIL,           /* 150 */
1505    OP_ACCEPT,         /* 110 */    OP_ACCEPT,         /* 151 */
1506      OP_ASSERT_ACCEPT,  /* 152 Used inside assertions */
1507      OP_CLOSE,          /* 153 Used before OP_ACCEPT to close open captures */
1508    
1509    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1510    
1511    OP_SKIPZERO        /* 111 */    OP_SKIPZERO,       /* 154 */
1512    
1513      /* This is not an opcode, but is used to check that tables indexed by opcode
1514      are the correct length, in order to catch updating errors - there have been
1515      some in the past. */
1516    
1517      OP_TABLE_LENGTH
1518  };  };
1519    
1520    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1521    definitions that follow must also be updated to match. There are also tables
1522    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1523    
1524    
1525  /* 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
1526  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
1527    only in pcre_printint.c, which fills out the full names in many cases (and in
1528    some cases doesn't actually use these names at all). */
1529    
1530  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1531    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1532    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1533    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1534    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1535    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1536    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1537      "{", "{", "{",                                                  \
1538    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1539    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1540      "{", "{", "{",                                                  \
1541      "*+","++", "?+", "{",                                           \
1542      "*", "*?", "+", "+?", "?", "??",                                \
1543      "{", "{", "{",                                                  \
1544      "*+","++", "?+", "{",                                           \
1545      "*", "*?", "+", "+?", "?", "??",                                \
1546      "{", "{", "{",                                                  \
1547    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1548    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1549    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1550    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1551    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1552    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1553    "AssertB", "AssertB not", "Reverse",                            \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1554    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1555    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Once",                                                         \
1556    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "Bra", "BraPos", "CBra", "CBraPos",                             \
1557    "Skip zero"    "Cond",                                                         \
1558      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1559      "SCond",                                                        \
1560      "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1561      "Brazero", "Braminzero", "Braposzero",                          \
1562      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1563      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1564      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1565      "Close", "Skip zero"
1566    
1567    
1568  /* 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 840  in UTF-8 mode. The code that uses this t Line 1579  in UTF-8 mode. The code that uses this t
1579    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1580    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1581    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1582    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1583    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1584    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1,                             /* \X                                     */ \
1585      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1586    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1587    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1588    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1589    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1590    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1591    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1592      4, 4, 4,                       /* upto, minupto, exact      ** mode      */ \
1593    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
1594      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1595      4, 4, 4,                       /* upto I, minupto I, exact I             */ \
1596      2, 2, 2, 4,                    /* *+I, ++I, ?+I, upto+I                  */ \
1597    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1598    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1599    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
1600    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2, 2, 2, 4,                    /* Possessive NOT *, +, ?, upto           */ \
1601      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1602      4, 4, 4,                       /* NOT upto I, minupto I, exact I         */ \
1603      2, 2, 2, 4,                    /* Possessive NOT *I, +I, ?I, upto I      */ \
1604    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1605    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1606    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
# Line 865  in UTF-8 mode. The code that uses this t Line 1612  in UTF-8 mode. The code that uses this t
1612   33,                             /* NCLASS                                 */ \   33,                             /* NCLASS                                 */ \
1613    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1614    3,                             /* REF                                    */ \    3,                             /* REF                                    */ \
1615      3,                             /* REFI                                   */ \
1616    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1617    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1618    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
1619    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
1620    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
1621    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
1622      1+LINK_SIZE,                   /* KetRpos                                */ \
1623      1+LINK_SIZE,                   /* Reverse                                */ \
1624    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
1625    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
1626    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1627    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
   1+LINK_SIZE,                   /* Reverse                                */ \  
1628    1+LINK_SIZE,                   /* ONCE                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1629    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
1630      1+LINK_SIZE,                   /* BRAPOS                                 */ \
1631    3+LINK_SIZE,                   /* CBRA                                   */ \    3+LINK_SIZE,                   /* CBRA                                   */ \
1632      3+LINK_SIZE,                   /* CBRAPOS                                */ \
1633    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1634    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1635      1+LINK_SIZE,                   /* SBRAPOS                                */ \
1636    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1637      3+LINK_SIZE,                   /* SCBRAPOS                               */ \
1638    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1639    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1640    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1641    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1642    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
1643    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1644    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */    1, 3,                          /* SKIP, SKIP_ARG                         */ \
1645      1+LINK_SIZE, 3+LINK_SIZE,      /* THEN, THEN_ARG                         */ \
1646      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
1647      3, 1                           /* CLOSE, SKIPZERO  */
1648    
1649  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1650    condition. */
1651    
1652  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1653    
1654  /* 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
1655  tracked. */  easily be tracked. When a new number is added, the table called eint in
1656    pcreposix.c must be updated. */
1657    
1658  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1659         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
# Line 904  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 1661  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
1661         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1662         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1663         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1664         ERR60, ERR61, ERR62, ERR63, ERR64 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
1665           ERRCOUNT };
1666    
1667  /* 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
1668  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 920  Because people can now save and re-use c Line 1678  Because people can now save and re-use c
1678  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
1679  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
1680  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1681  NOTE NOTE NOTE:  NOTE NOTE NOTE
1682  */  */
1683    
1684  typedef struct real_pcre {  typedef struct real_pcre {
# Line 947  remark (see NOTE above) about extending Line 1705  remark (see NOTE above) about extending
1705    
1706  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1707    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1708    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1709    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1710      pcre_uint32 minlength;          /* Minimum subject length */
1711  } pcre_study_data;  } pcre_study_data;
1712    
1713    /* Structure for building a chain of open capturing subpatterns during
1714    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1715    encountered. This is also used to identify subpatterns that contain recursive
1716    back references to themselves, so that they can be made atomic. */
1717    
1718    typedef struct open_capitem {
1719      struct open_capitem *next;    /* Chain link */
1720      pcre_uint16 number;           /* Capture number */
1721      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1722    } open_capitem;
1723    
1724  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1725  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1726    
# Line 963  typedef struct compile_data { Line 1733  typedef struct compile_data {
1733    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1734    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1735    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1736      open_capitem *open_caps;      /* Chain of open capture items */
1737    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1738    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1739    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
# Line 971  typedef struct compile_data { Line 1742  typedef struct compile_data {
1742    int  final_bracount;          /* Saved value after first pass */    int  final_bracount;          /* Saved value after first pass */
1743    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
1744    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
1745      int  assert_depth;            /* Depth of nested assertions */
1746    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
1747    int  external_flags;          /* External flag bits to be set */    int  external_flags;          /* External flag bits to be set */
1748    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1749    BOOL had_accept;              /* (*ACCEPT) encountered */    BOOL had_accept;              /* (*ACCEPT) encountered */
1750      BOOL check_lookbehind;        /* Lookbehinds need later checking */
1751    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1752    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1753    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
1754  } compile_data;  } compile_data;
1755    
1756  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
1757  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
1758    
1759  typedef struct branch_chain {  typedef struct branch_chain {
1760    struct branch_chain *outer;    struct branch_chain *outer;
1761    uschar *current;    uschar *current_branch;
1762  } branch_chain;  } branch_chain;
1763    
1764  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
1765  call within the pattern. */  call within the pattern; used by pcre_exec(). */
1766    
1767  typedef struct recursion_info {  typedef struct recursion_info {
1768    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1769    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
1770    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
1771    USPTR save_start;             /* Old value of mstart */    int saved_max;                  /* Number of saved offsets */
1772    int *offset_save;             /* Pointer to start of saved offsets */    USPTR subject_position;         /* Position at start of recursion */
   int saved_max;                /* Number of saved offsets */  
1773  } recursion_info;  } recursion_info;
1774    
1775    /* A similar structure for pcre_dfa_exec(). */
1776    
1777    typedef struct dfa_recursion_info {
1778      struct dfa_recursion_info *prevrec;
1779      int group_num;
1780      USPTR subject_position;
1781    } dfa_recursion_info;
1782    
1783  /* 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
1784  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
1785  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops; used by
1786    pcre_exec(). */
1787    
1788  typedef struct eptrblock {  typedef struct eptrblock {
1789    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
# Line 1022  typedef struct match_data { Line 1803  typedef struct match_data {
1803    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1804    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1805    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1806      int    name_count;            /* Number of names in name table */
1807      int    name_entry_size;       /* Size of entry in names table */
1808      uschar *name_table;           /* Table of names */
1809    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1810    const uschar *lcc;            /* Points to lower casing table */    const  uschar *lcc;           /* Points to lower casing table */
1811    const uschar *ctypes;         /* Points to table of type maps */    const  uschar *ctypes;        /* Points to table of type maps */
1812    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
1813    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1814    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1815    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1816    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1817      BOOL   use_ucp;               /* PCRE_UCP flag */
1818    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1819    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1820    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1821    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1822    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1823    const uschar *start_code;     /* For use when recursing */    const  uschar *start_code;    /* For use when recursing */
1824    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1825    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1826    USPTR  start_match_ptr;       /* Start of matched string */    USPTR  start_match_ptr;       /* Start of matched string */
1827    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1828      USPTR  start_used_ptr;        /* Earliest consulted character */
1829      int    partial;               /* PARTIAL options */
1830    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1831    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1832    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
1833      int    match_function_type;   /* Set for certain special calls of MATCH() */
1834    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
1835    int    eptrn;                 /* Next free eptrblock */    int    eptrn;                 /* Next free eptrblock */
1836    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1837    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1838      const  uschar *mark;          /* Mark pointer to pass back */
1839      const  uschar *once_target;   /* Where to back up to for atomic groups */
1840  } match_data;  } match_data;
1841    
1842  /* 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
1843  functions. */  functions. */
1844    
1845  typedef struct dfa_match_data {  typedef struct dfa_match_data {
1846    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;      /* Start of the compiled pattern */
1847    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;   /* Start of the subject string */
1848    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;     /* End of subject string */
1849    const uschar *tables;         /* Character tables */    const uschar *start_used_ptr;  /* Earliest consulted character */
1850    int   moptions;               /* Match options */    const uschar *tables;          /* Character tables */
1851    int   poptions;               /* Pattern options */    int   start_offset;            /* The start offset value */
1852    int    nltype;                /* Newline type */    int   moptions;                /* Match options */
1853    int    nllen;                 /* Newline string length */    int   poptions;                /* Pattern options */
1854    uschar nl[4];                 /* Newline string when fixed */    int    nltype;                 /* Newline type */
1855    void  *callout_data;          /* To pass back to callouts */    int    nllen;                  /* Newline string length */
1856      uschar nl[4];                  /* Newline string when fixed */
1857      void  *callout_data;           /* To pass back to callouts */
1858      dfa_recursion_info *recursive; /* Linked list of recursion data */
1859  } dfa_match_data;  } dfa_match_data;
1860    
1861  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 1120  extern const int _pcre_utf8_table2[]; Line 1913  extern const int _pcre_utf8_table2[];
1913  extern const int    _pcre_utf8_table3[];  extern const int    _pcre_utf8_table3[];
1914  extern const uschar _pcre_utf8_table4[];  extern const uschar _pcre_utf8_table4[];
1915    
1916    #ifdef SUPPORT_JIT
1917    extern const uschar _pcre_utf8_char_sizes[];
1918    #endif
1919    
1920  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1921    
1922  extern const char   _pcre_utt_names[];  extern const char   _pcre_utt_names[];
# Line 1135  extern const uschar _pcre_OP_lengths[]; Line 1932  extern const uschar _pcre_OP_lengths[];
1932  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
1933  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1934    
1935  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1936                        int *, BOOL);  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);
1937  extern int          _pcre_ord2utf8(int, uschar *);  extern int           _pcre_ord2utf8(int, uschar *);
1938  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1939                        const pcre_study_data *, pcre_study_data *);                         const pcre_study_data *, pcre_study_data *);
1940  extern int          _pcre_valid_utf8(const uschar *, int);  extern int           _pcre_valid_utf8(USPTR, int, int *);
1941  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);
1942                        int *, BOOL);  extern BOOL          _pcre_xclass(int, const uschar *);
1943  extern BOOL         _pcre_xclass(int, const uschar *);  
1944    #ifdef SUPPORT_JIT
1945    extern void          _pcre_jit_compile(const real_pcre *, pcre_extra *);
1946    extern int           _pcre_jit_exec(const real_pcre *, void *, PCRE_SPTR,
1947                            int, int, int, int *, int);
1948    extern void          _pcre_jit_free(void *);
1949    #endif
1950    
1951  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */
1952    
# Line 1158  extern const ucd_record _pcre_ucd_recor Line 1960  extern const ucd_record _pcre_ucd_recor
1960  extern const uschar      _pcre_ucd_stage1[];  extern const uschar      _pcre_ucd_stage1[];
1961  extern const pcre_uint16 _pcre_ucd_stage2[];  extern const pcre_uint16 _pcre_ucd_stage2[];
1962  extern const int         _pcre_ucp_gentype[];  extern const int         _pcre_ucp_gentype[];
1963    #ifdef SUPPORT_JIT
1964    extern const int         _pcre_ucp_typerange[];
1965    #endif
1966    
1967  /* UCD access macros */  /* UCD access macros */
1968    
1969  #define UCD_BLOCK_SIZE 128  #define UCD_BLOCK_SIZE 128
1970  #define GET_UCD(ch) (_pcre_ucd_records + \  #define GET_UCD(ch) (_pcre_ucd_records + \
1971          _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \          _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1972          UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])          UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
1973    
1974  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype  #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1975  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script  #define UCD_SCRIPT(ch)    GET_UCD(ch)->script

Legend:
Removed from v.376  
changed lines
  Added in v.667

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12