/[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 240 by ph10, Tue Sep 11 15:47:20 2007 UTC revision 475 by ph10, Sat Jan 2 18:21:30 2010 UTC
# Line 7  Line 7 
7  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2007 University of Cambridge             Copyright (c) 1997-2010 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 132  PCRE_EXP_DATA_DEFN only if they are not Line 144  PCRE_EXP_DATA_DEFN only if they are not
144  #  endif  #  endif
145  #endif  #endif
146    
147    /* When compiling with the MSVC compiler, it is sometimes necessary to include
148    a "calling convention" before exported function names. (This is secondhand
149    information; I know nothing about MSVC myself). For example, something like
150    
151      void __cdecl function(....)
152    
153    might be needed. In order so make this easy, all the exported functions have
154    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
155    set, we ensure here that it has no effect. */
156    
157    #ifndef PCRE_CALL_CONVENTION
158    #define PCRE_CALL_CONVENTION
159    #endif
160    
161  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We
162  cannot determine these outside the compilation (e.g. by running a program as  cannot determine these outside the compilation (e.g. by running a program as
163  part of "configure") because PCRE is often cross-compiled for use on other  part of "configure") because PCRE is often cross-compiled for use on other
# Line 140  preprocessor time in standard C environm Line 166  preprocessor time in standard C environm
166    
167  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
168    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
169      typedef short pcre_int16;
170  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
171    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
172      typedef int pcre_int16;
173  #else  #else
174    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
175  #endif  #endif
176    
177  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
178    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
179      typedef int pcre_int32;
180  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
181    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
182      typedef long int pcre_int32;
183  #else  #else
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. The macro int64_t
192    may be set by "configure". */
193    
194    #if HAVE_STDINT_H
195    #include <stdint.h>
196    #endif
197    
198    #if defined INT64_MAX || defined int64_t
199    #define INT64_OR_DOUBLE int64_t
200    #else
201    #define INT64_OR_DOUBLE double
202    #endif
203    
204  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
205  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
206  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 460  if there are extra bytes. This is called Line 507  if there are extra bytes. This is called
507      len += gcaa; \      len += gcaa; \
508      }      }
509    
510    /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
511    pointer, incrementing length if there are extra bytes. This is called when we
512    know we are in UTF-8 mode. */
513    
514    #define GETCHARLENTEST(c, eptr, len) \
515      c = *eptr; \
516      if (utf8 && c >= 0xc0) \
517        { \
518        int gcii; \
519        int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
520        int gcss = 6*gcaa; \
521        c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
522        for (gcii = 1; gcii <= gcaa; gcii++) \
523          { \
524          gcss -= 6; \
525          c |= (eptr[gcii] & 0x3f) << gcss; \
526          } \
527        len += gcaa; \
528        }
529    
530  /* 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
531  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
532  because almost all calls are already within a block of UTF-8 only code. */  because almost all calls are already within a block of UTF-8 only code. */
# Line 483  Standard C system should have one. */ Line 550  Standard C system should have one. */
550    
551  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
552  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
553  are in a 16-bit flags word. */  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
554    the restrictions on partial matching have been lifted. It remains for backwards
555    compatibility. */
556    
557  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
558  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
# Line 495  are in a 16-bit flags word. */ Line 564  are in a 16-bit flags word. */
564  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
565    
566  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
567    #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */
568    
569  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
570  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 502  time, run time, or study time, respectiv Line 572  time, run time, or study time, respectiv
572  #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| \
573                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
574    
575  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
576    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
577     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
578     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| \
579     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
580       PCRE_JAVASCRIPT_COMPAT)
581    
582  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
583    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
584     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
585       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
586    
587  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
588    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
589     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
590     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
591       PCRE_NO_START_OPTIMIZE)
592    
593  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
594    
# Line 540  variable-length repeat, or a anything ot Line 613  variable-length repeat, or a anything ot
613  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
614  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
615    
616  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
617    environments where these macros are defined elsewhere. Unfortunately, there
618    is no way to do the same for the typedef. */
619    
620  typedef int BOOL;  typedef int BOOL;
621    
622    #ifndef FALSE
623  #define FALSE   0  #define FALSE   0
624  #define TRUE    1  #define TRUE    1
625    #endif
626    
627    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
628    character constants like '*' because the compiler would emit their EBCDIC code,
629    which is different from their ASCII/UTF-8 code. Instead we define macros for
630    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
631    is enabled. When UTF-8 support is not enabled, the definitions use character
632    literals. Both character and string versions of each character are needed, and
633    there are some longer strings as well.
634    
635    This means that, on EBCDIC platforms, the PCRE library can handle either
636    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
637    would need different lookups depending on whether PCRE_UTF8 was set or not.
638    This would make it impossible to use characters in switch/case statements,
639    which would reduce performance. For a theoretical use (which nobody has asked
640    for) in a minority area (EBCDIC platforms), this is not sensible. Any
641    application that did need both could compile two versions of the library, using
642    macros to give the functions distinct names. */
643    
644    #ifndef SUPPORT_UTF8
645    
646    /* UTF-8 support is not enabled; use the platform-dependent character literals
647    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
648    
649    #define CHAR_HT                     '\t'
650    #define CHAR_VT                     '\v'
651    #define CHAR_FF                     '\f'
652    #define CHAR_CR                     '\r'
653    #define CHAR_NL                     '\n'
654    #define CHAR_BS                     '\b'
655    #define CHAR_BEL                    '\a'
656    #ifdef EBCDIC
657    #define CHAR_ESC                    '\047'
658    #define CHAR_DEL                    '\007'
659    #else
660    #define CHAR_ESC                    '\033'
661    #define CHAR_DEL                    '\177'
662    #endif
663    
664    #define CHAR_SPACE                  ' '
665    #define CHAR_EXCLAMATION_MARK       '!'
666    #define CHAR_QUOTATION_MARK         '"'
667    #define CHAR_NUMBER_SIGN            '#'
668    #define CHAR_DOLLAR_SIGN            '$'
669    #define CHAR_PERCENT_SIGN           '%'
670    #define CHAR_AMPERSAND              '&'
671    #define CHAR_APOSTROPHE             '\''
672    #define CHAR_LEFT_PARENTHESIS       '('
673    #define CHAR_RIGHT_PARENTHESIS      ')'
674    #define CHAR_ASTERISK               '*'
675    #define CHAR_PLUS                   '+'
676    #define CHAR_COMMA                  ','
677    #define CHAR_MINUS                  '-'
678    #define CHAR_DOT                    '.'
679    #define CHAR_SLASH                  '/'
680    #define CHAR_0                      '0'
681    #define CHAR_1                      '1'
682    #define CHAR_2                      '2'
683    #define CHAR_3                      '3'
684    #define CHAR_4                      '4'
685    #define CHAR_5                      '5'
686    #define CHAR_6                      '6'
687    #define CHAR_7                      '7'
688    #define CHAR_8                      '8'
689    #define CHAR_9                      '9'
690    #define CHAR_COLON                  ':'
691    #define CHAR_SEMICOLON              ';'
692    #define CHAR_LESS_THAN_SIGN         '<'
693    #define CHAR_EQUALS_SIGN            '='
694    #define CHAR_GREATER_THAN_SIGN      '>'
695    #define CHAR_QUESTION_MARK          '?'
696    #define CHAR_COMMERCIAL_AT          '@'
697    #define CHAR_A                      'A'
698    #define CHAR_B                      'B'
699    #define CHAR_C                      'C'
700    #define CHAR_D                      'D'
701    #define CHAR_E                      'E'
702    #define CHAR_F                      'F'
703    #define CHAR_G                      'G'
704    #define CHAR_H                      'H'
705    #define CHAR_I                      'I'
706    #define CHAR_J                      'J'
707    #define CHAR_K                      'K'
708    #define CHAR_L                      'L'
709    #define CHAR_M                      'M'
710    #define CHAR_N                      'N'
711    #define CHAR_O                      'O'
712    #define CHAR_P                      'P'
713    #define CHAR_Q                      'Q'
714    #define CHAR_R                      'R'
715    #define CHAR_S                      'S'
716    #define CHAR_T                      'T'
717    #define CHAR_U                      'U'
718    #define CHAR_V                      'V'
719    #define CHAR_W                      'W'
720    #define CHAR_X                      'X'
721    #define CHAR_Y                      'Y'
722    #define CHAR_Z                      'Z'
723    #define CHAR_LEFT_SQUARE_BRACKET    '['
724    #define CHAR_BACKSLASH              '\\'
725    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
726    #define CHAR_CIRCUMFLEX_ACCENT      '^'
727    #define CHAR_UNDERSCORE             '_'
728    #define CHAR_GRAVE_ACCENT           '`'
729    #define CHAR_a                      'a'
730    #define CHAR_b                      'b'
731    #define CHAR_c                      'c'
732    #define CHAR_d                      'd'
733    #define CHAR_e                      'e'
734    #define CHAR_f                      'f'
735    #define CHAR_g                      'g'
736    #define CHAR_h                      'h'
737    #define CHAR_i                      'i'
738    #define CHAR_j                      'j'
739    #define CHAR_k                      'k'
740    #define CHAR_l                      'l'
741    #define CHAR_m                      'm'
742    #define CHAR_n                      'n'
743    #define CHAR_o                      'o'
744    #define CHAR_p                      'p'
745    #define CHAR_q                      'q'
746    #define CHAR_r                      'r'
747    #define CHAR_s                      's'
748    #define CHAR_t                      't'
749    #define CHAR_u                      'u'
750    #define CHAR_v                      'v'
751    #define CHAR_w                      'w'
752    #define CHAR_x                      'x'
753    #define CHAR_y                      'y'
754    #define CHAR_z                      'z'
755    #define CHAR_LEFT_CURLY_BRACKET     '{'
756    #define CHAR_VERTICAL_LINE          '|'
757    #define CHAR_RIGHT_CURLY_BRACKET    '}'
758    #define CHAR_TILDE                  '~'
759    
760    #define STR_HT                      "\t"
761    #define STR_VT                      "\v"
762    #define STR_FF                      "\f"
763    #define STR_CR                      "\r"
764    #define STR_NL                      "\n"
765    #define STR_BS                      "\b"
766    #define STR_BEL                     "\a"
767    #ifdef EBCDIC
768    #define STR_ESC                     "\047"
769    #define STR_DEL                     "\007"
770    #else
771    #define STR_ESC                     "\033"
772    #define STR_DEL                     "\177"
773    #endif
774    
775    #define STR_SPACE                   " "
776    #define STR_EXCLAMATION_MARK        "!"
777    #define STR_QUOTATION_MARK          "\""
778    #define STR_NUMBER_SIGN             "#"
779    #define STR_DOLLAR_SIGN             "$"
780    #define STR_PERCENT_SIGN            "%"
781    #define STR_AMPERSAND               "&"
782    #define STR_APOSTROPHE              "'"
783    #define STR_LEFT_PARENTHESIS        "("
784    #define STR_RIGHT_PARENTHESIS       ")"
785    #define STR_ASTERISK                "*"
786    #define STR_PLUS                    "+"
787    #define STR_COMMA                   ","
788    #define STR_MINUS                   "-"
789    #define STR_DOT                     "."
790    #define STR_SLASH                   "/"
791    #define STR_0                       "0"
792    #define STR_1                       "1"
793    #define STR_2                       "2"
794    #define STR_3                       "3"
795    #define STR_4                       "4"
796    #define STR_5                       "5"
797    #define STR_6                       "6"
798    #define STR_7                       "7"
799    #define STR_8                       "8"
800    #define STR_9                       "9"
801    #define STR_COLON                   ":"
802    #define STR_SEMICOLON               ";"
803    #define STR_LESS_THAN_SIGN          "<"
804    #define STR_EQUALS_SIGN             "="
805    #define STR_GREATER_THAN_SIGN       ">"
806    #define STR_QUESTION_MARK           "?"
807    #define STR_COMMERCIAL_AT           "@"
808    #define STR_A                       "A"
809    #define STR_B                       "B"
810    #define STR_C                       "C"
811    #define STR_D                       "D"
812    #define STR_E                       "E"
813    #define STR_F                       "F"
814    #define STR_G                       "G"
815    #define STR_H                       "H"
816    #define STR_I                       "I"
817    #define STR_J                       "J"
818    #define STR_K                       "K"
819    #define STR_L                       "L"
820    #define STR_M                       "M"
821    #define STR_N                       "N"
822    #define STR_O                       "O"
823    #define STR_P                       "P"
824    #define STR_Q                       "Q"
825    #define STR_R                       "R"
826    #define STR_S                       "S"
827    #define STR_T                       "T"
828    #define STR_U                       "U"
829    #define STR_V                       "V"
830    #define STR_W                       "W"
831    #define STR_X                       "X"
832    #define STR_Y                       "Y"
833    #define STR_Z                       "Z"
834    #define STR_LEFT_SQUARE_BRACKET     "["
835    #define STR_BACKSLASH               "\\"
836    #define STR_RIGHT_SQUARE_BRACKET    "]"
837    #define STR_CIRCUMFLEX_ACCENT       "^"
838    #define STR_UNDERSCORE              "_"
839    #define STR_GRAVE_ACCENT            "`"
840    #define STR_a                       "a"
841    #define STR_b                       "b"
842    #define STR_c                       "c"
843    #define STR_d                       "d"
844    #define STR_e                       "e"
845    #define STR_f                       "f"
846    #define STR_g                       "g"
847    #define STR_h                       "h"
848    #define STR_i                       "i"
849    #define STR_j                       "j"
850    #define STR_k                       "k"
851    #define STR_l                       "l"
852    #define STR_m                       "m"
853    #define STR_n                       "n"
854    #define STR_o                       "o"
855    #define STR_p                       "p"
856    #define STR_q                       "q"
857    #define STR_r                       "r"
858    #define STR_s                       "s"
859    #define STR_t                       "t"
860    #define STR_u                       "u"
861    #define STR_v                       "v"
862    #define STR_w                       "w"
863    #define STR_x                       "x"
864    #define STR_y                       "y"
865    #define STR_z                       "z"
866    #define STR_LEFT_CURLY_BRACKET      "{"
867    #define STR_VERTICAL_LINE           "|"
868    #define STR_RIGHT_CURLY_BRACKET     "}"
869    #define STR_TILDE                   "~"
870    
871    #define STRING_ACCEPT0              "ACCEPT\0"
872    #define STRING_COMMIT0              "COMMIT\0"
873    #define STRING_F0                   "F\0"
874    #define STRING_FAIL0                "FAIL\0"
875    #define STRING_PRUNE0               "PRUNE\0"
876    #define STRING_SKIP0                "SKIP\0"
877    #define STRING_THEN                 "THEN"
878    
879    #define STRING_alpha0               "alpha\0"
880    #define STRING_lower0               "lower\0"
881    #define STRING_upper0               "upper\0"
882    #define STRING_alnum0               "alnum\0"
883    #define STRING_ascii0               "ascii\0"
884    #define STRING_blank0               "blank\0"
885    #define STRING_cntrl0               "cntrl\0"
886    #define STRING_digit0               "digit\0"
887    #define STRING_graph0               "graph\0"
888    #define STRING_print0               "print\0"
889    #define STRING_punct0               "punct\0"
890    #define STRING_space0               "space\0"
891    #define STRING_word0                "word\0"
892    #define STRING_xdigit               "xdigit"
893    
894    #define STRING_DEFINE               "DEFINE"
895    
896    #define STRING_CR_RIGHTPAR          "CR)"
897    #define STRING_LF_RIGHTPAR          "LF)"
898    #define STRING_CRLF_RIGHTPAR        "CRLF)"
899    #define STRING_ANY_RIGHTPAR         "ANY)"
900    #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"
901    #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
902    #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
903    #define STRING_UTF8_RIGHTPAR        "UTF8)"
904    
905    #else  /* SUPPORT_UTF8 */
906    
907    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
908    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
909    only. */
910    
911    #define CHAR_HT                     '\011'
912    #define CHAR_VT                     '\013'
913    #define CHAR_FF                     '\014'
914    #define CHAR_CR                     '\015'
915    #define CHAR_NL                     '\012'
916    #define CHAR_BS                     '\010'
917    #define CHAR_BEL                    '\007'
918    #define CHAR_ESC                    '\033'
919    #define CHAR_DEL                    '\177'
920    
921    #define CHAR_SPACE                  '\040'
922    #define CHAR_EXCLAMATION_MARK       '\041'
923    #define CHAR_QUOTATION_MARK         '\042'
924    #define CHAR_NUMBER_SIGN            '\043'
925    #define CHAR_DOLLAR_SIGN            '\044'
926    #define CHAR_PERCENT_SIGN           '\045'
927    #define CHAR_AMPERSAND              '\046'
928    #define CHAR_APOSTROPHE             '\047'
929    #define CHAR_LEFT_PARENTHESIS       '\050'
930    #define CHAR_RIGHT_PARENTHESIS      '\051'
931    #define CHAR_ASTERISK               '\052'
932    #define CHAR_PLUS                   '\053'
933    #define CHAR_COMMA                  '\054'
934    #define CHAR_MINUS                  '\055'
935    #define CHAR_DOT                    '\056'
936    #define CHAR_SLASH                  '\057'
937    #define CHAR_0                      '\060'
938    #define CHAR_1                      '\061'
939    #define CHAR_2                      '\062'
940    #define CHAR_3                      '\063'
941    #define CHAR_4                      '\064'
942    #define CHAR_5                      '\065'
943    #define CHAR_6                      '\066'
944    #define CHAR_7                      '\067'
945    #define CHAR_8                      '\070'
946    #define CHAR_9                      '\071'
947    #define CHAR_COLON                  '\072'
948    #define CHAR_SEMICOLON              '\073'
949    #define CHAR_LESS_THAN_SIGN         '\074'
950    #define CHAR_EQUALS_SIGN            '\075'
951    #define CHAR_GREATER_THAN_SIGN      '\076'
952    #define CHAR_QUESTION_MARK          '\077'
953    #define CHAR_COMMERCIAL_AT          '\100'
954    #define CHAR_A                      '\101'
955    #define CHAR_B                      '\102'
956    #define CHAR_C                      '\103'
957    #define CHAR_D                      '\104'
958    #define CHAR_E                      '\105'
959    #define CHAR_F                      '\106'
960    #define CHAR_G                      '\107'
961    #define CHAR_H                      '\110'
962    #define CHAR_I                      '\111'
963    #define CHAR_J                      '\112'
964    #define CHAR_K                      '\113'
965    #define CHAR_L                      '\114'
966    #define CHAR_M                      '\115'
967    #define CHAR_N                      '\116'
968    #define CHAR_O                      '\117'
969    #define CHAR_P                      '\120'
970    #define CHAR_Q                      '\121'
971    #define CHAR_R                      '\122'
972    #define CHAR_S                      '\123'
973    #define CHAR_T                      '\124'
974    #define CHAR_U                      '\125'
975    #define CHAR_V                      '\126'
976    #define CHAR_W                      '\127'
977    #define CHAR_X                      '\130'
978    #define CHAR_Y                      '\131'
979    #define CHAR_Z                      '\132'
980    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
981    #define CHAR_BACKSLASH              '\134'
982    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
983    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
984    #define CHAR_UNDERSCORE             '\137'
985    #define CHAR_GRAVE_ACCENT           '\140'
986    #define CHAR_a                      '\141'
987    #define CHAR_b                      '\142'
988    #define CHAR_c                      '\143'
989    #define CHAR_d                      '\144'
990    #define CHAR_e                      '\145'
991    #define CHAR_f                      '\146'
992    #define CHAR_g                      '\147'
993    #define CHAR_h                      '\150'
994    #define CHAR_i                      '\151'
995    #define CHAR_j                      '\152'
996    #define CHAR_k                      '\153'
997    #define CHAR_l                      '\154'
998    #define CHAR_m                      '\155'
999    #define CHAR_n                      '\156'
1000    #define CHAR_o                      '\157'
1001    #define CHAR_p                      '\160'
1002    #define CHAR_q                      '\161'
1003    #define CHAR_r                      '\162'
1004    #define CHAR_s                      '\163'
1005    #define CHAR_t                      '\164'
1006    #define CHAR_u                      '\165'
1007    #define CHAR_v                      '\166'
1008    #define CHAR_w                      '\167'
1009    #define CHAR_x                      '\170'
1010    #define CHAR_y                      '\171'
1011    #define CHAR_z                      '\172'
1012    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1013    #define CHAR_VERTICAL_LINE          '\174'
1014    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1015    #define CHAR_TILDE                  '\176'
1016    
1017    #define STR_HT                      "\011"
1018    #define STR_VT                      "\013"
1019    #define STR_FF                      "\014"
1020    #define STR_CR                      "\015"
1021    #define STR_NL                      "\012"
1022    #define STR_BS                      "\010"
1023    #define STR_BEL                     "\007"
1024    #define STR_ESC                     "\033"
1025    #define STR_DEL                     "\177"
1026    
1027    #define STR_SPACE                   "\040"
1028    #define STR_EXCLAMATION_MARK        "\041"
1029    #define STR_QUOTATION_MARK          "\042"
1030    #define STR_NUMBER_SIGN             "\043"
1031    #define STR_DOLLAR_SIGN             "\044"
1032    #define STR_PERCENT_SIGN            "\045"
1033    #define STR_AMPERSAND               "\046"
1034    #define STR_APOSTROPHE              "\047"
1035    #define STR_LEFT_PARENTHESIS        "\050"
1036    #define STR_RIGHT_PARENTHESIS       "\051"
1037    #define STR_ASTERISK                "\052"
1038    #define STR_PLUS                    "\053"
1039    #define STR_COMMA                   "\054"
1040    #define STR_MINUS                   "\055"
1041    #define STR_DOT                     "\056"
1042    #define STR_SLASH                   "\057"
1043    #define STR_0                       "\060"
1044    #define STR_1                       "\061"
1045    #define STR_2                       "\062"
1046    #define STR_3                       "\063"
1047    #define STR_4                       "\064"
1048    #define STR_5                       "\065"
1049    #define STR_6                       "\066"
1050    #define STR_7                       "\067"
1051    #define STR_8                       "\070"
1052    #define STR_9                       "\071"
1053    #define STR_COLON                   "\072"
1054    #define STR_SEMICOLON               "\073"
1055    #define STR_LESS_THAN_SIGN          "\074"
1056    #define STR_EQUALS_SIGN             "\075"
1057    #define STR_GREATER_THAN_SIGN       "\076"
1058    #define STR_QUESTION_MARK           "\077"
1059    #define STR_COMMERCIAL_AT           "\100"
1060    #define STR_A                       "\101"
1061    #define STR_B                       "\102"
1062    #define STR_C                       "\103"
1063    #define STR_D                       "\104"
1064    #define STR_E                       "\105"
1065    #define STR_F                       "\106"
1066    #define STR_G                       "\107"
1067    #define STR_H                       "\110"
1068    #define STR_I                       "\111"
1069    #define STR_J                       "\112"
1070    #define STR_K                       "\113"
1071    #define STR_L                       "\114"
1072    #define STR_M                       "\115"
1073    #define STR_N                       "\116"
1074    #define STR_O                       "\117"
1075    #define STR_P                       "\120"
1076    #define STR_Q                       "\121"
1077    #define STR_R                       "\122"
1078    #define STR_S                       "\123"
1079    #define STR_T                       "\124"
1080    #define STR_U                       "\125"
1081    #define STR_V                       "\126"
1082    #define STR_W                       "\127"
1083    #define STR_X                       "\130"
1084    #define STR_Y                       "\131"
1085    #define STR_Z                       "\132"
1086    #define STR_LEFT_SQUARE_BRACKET     "\133"
1087    #define STR_BACKSLASH               "\134"
1088    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1089    #define STR_CIRCUMFLEX_ACCENT       "\136"
1090    #define STR_UNDERSCORE              "\137"
1091    #define STR_GRAVE_ACCENT            "\140"
1092    #define STR_a                       "\141"
1093    #define STR_b                       "\142"
1094    #define STR_c                       "\143"
1095    #define STR_d                       "\144"
1096    #define STR_e                       "\145"
1097    #define STR_f                       "\146"
1098    #define STR_g                       "\147"
1099    #define STR_h                       "\150"
1100    #define STR_i                       "\151"
1101    #define STR_j                       "\152"
1102    #define STR_k                       "\153"
1103    #define STR_l                       "\154"
1104    #define STR_m                       "\155"
1105    #define STR_n                       "\156"
1106    #define STR_o                       "\157"
1107    #define STR_p                       "\160"
1108    #define STR_q                       "\161"
1109    #define STR_r                       "\162"
1110    #define STR_s                       "\163"
1111    #define STR_t                       "\164"
1112    #define STR_u                       "\165"
1113    #define STR_v                       "\166"
1114    #define STR_w                       "\167"
1115    #define STR_x                       "\170"
1116    #define STR_y                       "\171"
1117    #define STR_z                       "\172"
1118    #define STR_LEFT_CURLY_BRACKET      "\173"
1119    #define STR_VERTICAL_LINE           "\174"
1120    #define STR_RIGHT_CURLY_BRACKET     "\175"
1121    #define STR_TILDE                   "\176"
1122    
1123    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1124    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1125    #define STRING_F0                   STR_F "\0"
1126    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1127    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1128    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1129    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1130    
1131    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1132    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1133    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1134    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1135    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1136    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1137    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1138    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1139    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1140    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1141    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1142    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1143    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1144    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1145    
1146    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1147    
1148    #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS
1149    #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS
1150    #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1151    #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1152    #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1153    #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
1154    #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
1155    #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1156    
1157    #endif  /* SUPPORT_UTF8 */
1158    
1159  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1160    
1161  #ifndef ESC_e  #ifndef ESC_e
1162  #define ESC_e 27  #define ESC_e CHAR_ESC
1163  #endif  #endif
1164    
1165  #ifndef ESC_f  #ifndef ESC_f
1166  #define ESC_f '\f'  #define ESC_f CHAR_FF
1167  #endif  #endif
1168    
1169  #ifndef ESC_n  #ifndef ESC_n
1170  #define ESC_n '\n'  #define ESC_n CHAR_NL
1171  #endif  #endif
1172    
1173  #ifndef ESC_r  #ifndef ESC_r
1174  #define ESC_r '\r'  #define ESC_r CHAR_CR
1175  #endif  #endif
1176    
1177  /* 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
1178  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1179    
1180  #ifndef ESC_tee  #ifndef ESC_tee
1181  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1182  #endif  #endif
1183    
1184  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
# Line 596  contain UTF-8 characters with values gre Line 1205  contain UTF-8 characters with values gre
1205  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
1206  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
1207  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_ANY because it
1208  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
1209  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  (which is used for [^] in JavaScript compatibility mode).
1210  There are two tests in the code for an escape greater than ESC_b and less than  
1211  ESC_Z to detect the types that may be repeated. These are the types that  The final escape must be ESC_REF as subsequent values are used for
1212  consume characters. If any new escapes are put in between that don't consume a  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1213  character, that code will have to change. */  greater than ESC_b and less than ESC_Z to detect the types that may be
1214    repeated. These are the types that consume characters. If any new escapes are
1215    put in between that don't consume a character, that code will have to change.
1216    */
1217    
1218  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,
1219         ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h,         ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1220         ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_k, ESC_REF };         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
1221           ESC_REF };
1222    
1223    
1224  /* 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
1225  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.
1226    
1227  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1228  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
1229  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1230    
1231  enum {  enum {
1232    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 631  enum { Line 1244  enum {
1244    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1245    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1246    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1247    OP_ANY,            /* 12 Match any character */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
1248    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
1249    OP_NOTPROP,        /* 14 \P (not Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1250    OP_PROP,           /* 15 \p (Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1251    OP_ANYNL,          /* 16 \R (any newline sequence) */    OP_PROP,           /* 16 \p (Unicode property) */
1252    OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */    OP_ANYNL,          /* 17 \R (any newline sequence) */
1253    OP_HSPACE,         /* 18 \h (horizontal whitespace) */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1254    OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1255    OP_VSPACE,         /* 20 \v (vertical whitespace) */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1256    OP_EXTUNI,         /* 21 \X (extended Unicode sequence */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1257    OP_EODN,           /* 22 End of data or \n at end of data: \Z. */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1258    OP_EOD,            /* 23 End of data: \z */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1259      OP_EOD,            /* 24 End of data: \z */
1260    OP_OPT,            /* 24 Set runtime options */  
1261    OP_CIRC,           /* 25 Start of line - varies with multiline switch */    OP_OPT,            /* 25 Set runtime options */
1262    OP_DOLL,           /* 26 End of line - varies with multiline switch */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
1263    OP_CHAR,           /* 27 Match one character, casefully */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
1264    OP_CHARNC,         /* 28 Match one character, caselessly */    OP_CHAR,           /* 28 Match one character, casefully */
1265    OP_NOT,            /* 29 Match one character, not the following one */    OP_CHARNC,         /* 29 Match one character, caselessly */
1266      OP_NOT,            /* 30 Match one character, not the following one */
1267    OP_STAR,           /* 30 The maximizing and minimizing versions of */  
1268    OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */    OP_STAR,           /* 31 The maximizing and minimizing versions of */
1269    OP_PLUS,           /* 32 the minimizing one second. */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
1270    OP_MINPLUS,        /* 33 This first set applies to single characters.*/    OP_PLUS,           /* 33 the minimizing one second. */
1271    OP_QUERY,          /* 34 */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
1272    OP_MINQUERY,       /* 35 */    OP_QUERY,          /* 35 */
1273      OP_MINQUERY,       /* 36 */
1274    OP_UPTO,           /* 36 From 0 to n matches */  
1275    OP_MINUPTO,        /* 37 */    OP_UPTO,           /* 37 From 0 to n matches */
1276    OP_EXACT,          /* 38 Exactly n matches */    OP_MINUPTO,        /* 38 */
1277      OP_EXACT,          /* 39 Exactly n matches */
1278    OP_POSSTAR,        /* 39 Possessified star */  
1279    OP_POSPLUS,        /* 40 Possessified plus */    OP_POSSTAR,        /* 40 Possessified star */
1280    OP_POSQUERY,       /* 41 Posesssified query */    OP_POSPLUS,        /* 41 Possessified plus */
1281    OP_POSUPTO,        /* 42 Possessified upto */    OP_POSQUERY,       /* 42 Posesssified query */
1282      OP_POSUPTO,        /* 43 Possessified upto */
1283    OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */  
1284    OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
1285    OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
1286    OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
1287    OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
1288    OP_NOTMINQUERY,    /* 48 */    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
1289      OP_NOTMINQUERY,    /* 49 */
1290    OP_NOTUPTO,        /* 49 From 0 to n matches */  
1291    OP_NOTMINUPTO,     /* 50 */    OP_NOTUPTO,        /* 50 From 0 to n matches */
1292    OP_NOTEXACT,       /* 51 Exactly n matches */    OP_NOTMINUPTO,     /* 51 */
1293      OP_NOTEXACT,       /* 52 Exactly n matches */
1294    OP_NOTPOSSTAR,     /* 52 Possessified versions */  
1295    OP_NOTPOSPLUS,     /* 53 */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
1296    OP_NOTPOSQUERY,    /* 54 */    OP_NOTPOSPLUS,     /* 54 */
1297    OP_NOTPOSUPTO,     /* 55 */    OP_NOTPOSQUERY,    /* 55 */
1298      OP_NOTPOSUPTO,     /* 56 */
1299    OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */  
1300    OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
1301    OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
1302    OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
1303    OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
1304    OP_TYPEMINQUERY,   /* 61 */    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
1305      OP_TYPEMINQUERY,   /* 62 */
1306    OP_TYPEUPTO,       /* 62 From 0 to n matches */  
1307    OP_TYPEMINUPTO,    /* 63 */    OP_TYPEUPTO,       /* 63 From 0 to n matches */
1308    OP_TYPEEXACT,      /* 64 Exactly n matches */    OP_TYPEMINUPTO,    /* 64 */
1309      OP_TYPEEXACT,      /* 65 Exactly n matches */
1310    OP_TYPEPOSSTAR,    /* 65 Possessified versions */  
1311    OP_TYPEPOSPLUS,    /* 66 */    OP_TYPEPOSSTAR,    /* 66 Possessified versions */
1312    OP_TYPEPOSQUERY,   /* 67 */    OP_TYPEPOSPLUS,    /* 67 */
1313    OP_TYPEPOSUPTO,    /* 68 */    OP_TYPEPOSQUERY,   /* 68 */
1314      OP_TYPEPOSUPTO,    /* 69 */
1315    OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */  
1316    OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
1317    OP_CRPLUS,         /* 71 the minimizing one second. These codes must */    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
1318    OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
1319    OP_CRQUERY,        /* 73 These are for character classes and back refs */    OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
1320    OP_CRMINQUERY,     /* 74 */    OP_CRQUERY,        /* 74 These are for character classes and back refs */
1321    OP_CRRANGE,        /* 75 These are different to the three sets above. */    OP_CRMINQUERY,     /* 75 */
1322    OP_CRMINRANGE,     /* 76 */    OP_CRRANGE,        /* 76 These are different to the three sets above. */
1323      OP_CRMINRANGE,     /* 77 */
1324    
1325    OP_CLASS,          /* 77 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
1326    OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
1327                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
1328                             character > 255 is encountered. */                             character > 255 is encountered. */
1329    
1330    OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
1331                             class. This does both positive and negative. */                             class. This does both positive and negative. */
1332    
1333    OP_REF,            /* 80 Match a back reference */    OP_REF,            /* 81 Match a back reference */
1334    OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
1335    OP_CALLOUT,        /* 82 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
1336    
1337    OP_ALT,            /* 83 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
1338    OP_KET,            /* 84 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */
1339    OP_KETRMAX,        /* 85 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
1340    OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
1341    
1342    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
1343    
1344    OP_ASSERT,         /* 87 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
1345    OP_ASSERT_NOT,     /* 88 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
1346    OP_ASSERTBACK,     /* 89 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
1347    OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
1348    OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
1349    
1350    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
1351    as there's a test for >= ONCE for a subpattern that isn't an assertion. */    as there's a test for >= ONCE for a subpattern that isn't an assertion. */
1352    
1353    OP_ONCE,           /* 92 Atomic group */    OP_ONCE,           /* 93 Atomic group */
1354    OP_BRA,            /* 93 Start of non-capturing bracket */    OP_BRA,            /* 94 Start of non-capturing bracket */
1355    OP_CBRA,           /* 94 Start of capturing bracket */    OP_CBRA,           /* 95 Start of capturing bracket */
1356    OP_COND,           /* 95 Conditional group */    OP_COND,           /* 96 Conditional group */
1357    
1358    /* These three must follow the previous three, in the same order. There's a    /* These three must follow the previous three, in the same order. There's a
1359    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1360    
1361    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
1362    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1363    OP_SCOND,          /* 98 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
1364    
1365    OP_CREF,           /* 99 Used to hold a capture number as condition */    /* The next two pairs must (respectively) be kept together. */
1366    OP_RREF,           /* 100 Used to hold a recursion number as condition */  
1367    OP_DEF,            /* 101 The DEFINE condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
1368      OP_NCREF,          /* 101 Same, but generaged by a name reference*/
1369      OP_RREF,           /* 102 Used to hold a recursion number as condition */
1370      OP_NRREF,          /* 103 Same, but generaged by a name reference*/
1371      OP_DEF,            /* 104 The DEFINE condition */
1372    
1373    OP_BRAZERO,        /* 102 These two must remain together and in this */    OP_BRAZERO,        /* 105 These two must remain together and in this */
1374    OP_BRAMINZERO,     /* 103 order. */    OP_BRAMINZERO,     /* 106 order. */
1375    
1376    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1377    
1378    OP_PRUNE,          /* 104 */    OP_PRUNE,          /* 107 */
1379    OP_SKIP,           /* 105 */    OP_SKIP,           /* 108 */
1380    OP_THEN,           /* 106 */    OP_THEN,           /* 109 */
1381    OP_COMMIT,         /* 107 */    OP_COMMIT,         /* 110 */
1382    
1383    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1384    
1385    OP_FAIL,           /* 108 */    OP_FAIL,           /* 111 */
1386    OP_ACCEPT          /* 109 */    OP_ACCEPT,         /* 112 */
1387      OP_CLOSE,          /* 113 Used before OP_ACCEPT to close open captures */
1388    
1389      /* This is used to skip a subpattern with a {0} quantifier */
1390    
1391      OP_SKIPZERO        /* 114 */
1392  };  };
1393    
1394    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1395    definitions that follow must also be updated to match. There are also tables
1396    called "coptable" cna "poptable" in pcre_dfa_exec.c that must be updated. */
1397    
1398    
1399  /* 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
1400  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
1401    
1402  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1403    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1404    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1405    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1406    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1407    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
# Line 789  for debugging. The macro is referenced o Line 1416  for debugging. The macro is referenced o
1416    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1417    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1418    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1419    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1420    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT"    "Brazero", "Braminzero",                                        \
1421      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
1422      "Close", "Skip zero"
1423    
1424    
1425  /* This macro defines the length of fixed length operations in the compiled  /* This macro defines the length of fixed length operations in the compiled
# Line 806  in UTF-8 mode. The code that uses this t Line 1435  in UTF-8 mode. The code that uses this t
1435    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1436    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1437    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1438    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1439    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
1440    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1441    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
# Line 850  in UTF-8 mode. The code that uses this t Line 1479  in UTF-8 mode. The code that uses this t
1479    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1480    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1481    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1482    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1483    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1484    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1485    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1486    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
1487    1, 1                           /* FAIL, ACCEPT                           */    1, 1, 3, 1                     /* FAIL, ACCEPT, CLOSE, SKIPZERO          */
1488    
1489    
1490  /* 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"
1491    condition. */
1492    
1493  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1494    
# Line 871  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 1501  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
1501         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1502         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1503         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1504         ERR60, ERR61 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65 };
1505    
1506  /* 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
1507  code vector run on as long as necessary after the end. We store an explicit  code vector run on as long as necessary after the end. We store an explicit
# Line 887  Because people can now save and re-use c Line 1517  Because people can now save and re-use c
1517  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
1518  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
1519  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1520  NOTE NOTE NOTE:  NOTE NOTE NOTE
1521  */  */
1522    
1523  typedef struct real_pcre {  typedef struct real_pcre {
# Line 914  remark (see NOTE above) about extending Line 1544  remark (see NOTE above) about extending
1544    
1545  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1546    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1547    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1548    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1549      pcre_uint32 minlength;          /* Minimum subject length */
1550  } pcre_study_data;  } pcre_study_data;
1551    
1552    /* Structure for building a chain of open capturing subpatterns during
1553    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1554    encountered. */
1555    
1556    typedef struct open_capitem {
1557      struct open_capitem *next;    /* Chain link */
1558      pcre_uint16 number;           /* Capture number */
1559    } open_capitem;
1560    
1561  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1562  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1563    
# Line 930  typedef struct compile_data { Line 1570  typedef struct compile_data {
1570    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1571    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1572    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1573      open_capitem *open_caps;      /* Chain of open capture items */
1574    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1575    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1576    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
1577    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
1578    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
1579      int  final_bracount;          /* Saved value after first pass */
1580    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
1581    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
1582    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
1583    int  external_flags;          /* External flag bits to be set */    int  external_flags;          /* External flag bits to be set */
1584    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1585    BOOL had_accept;              /* (*ACCEPT) encountered */    BOOL had_accept;              /* (*ACCEPT) encountered */
1586      BOOL check_lookbehind;        /* Lookbehinds need later checking */
1587    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1588    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1589    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 951  branches, for testing for left recursion Line 1594  branches, for testing for left recursion
1594    
1595  typedef struct branch_chain {  typedef struct branch_chain {
1596    struct branch_chain *outer;    struct branch_chain *outer;
1597    uschar *current;    uschar *current_branch;
1598  } branch_chain;  } branch_chain;
1599    
1600  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
# Line 964  typedef struct recursion_info { Line 1607  typedef struct recursion_info {
1607    USPTR save_start;             /* Old value of mstart */    USPTR save_start;             /* Old value of mstart */
1608    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1609    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1610      int save_offset_top;          /* Current value of offset_top */
1611  } recursion_info;  } recursion_info;
1612    
1613  /* 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
# Line 988  typedef struct match_data { Line 1632  typedef struct match_data {
1632    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1633    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1634    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1635      int    name_count;            /* Number of names in name table */
1636      int    name_entry_size;       /* Size of entry in names table */
1637      uschar *name_table;           /* Table of names */
1638    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1639    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1640    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
# Line 995  typedef struct match_data { Line 1642  typedef struct match_data {
1642    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1643    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1644    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1645      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1646    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1647    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1648    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1649    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1650    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1651    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
# Line 1005  typedef struct match_data { Line 1653  typedef struct match_data {
1653    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1654    USPTR  start_match_ptr;       /* Start of matched string */    USPTR  start_match_ptr;       /* Start of matched string */
1655    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1656      USPTR  start_used_ptr;        /* Earliest consulted character */
1657      int    partial;               /* PARTIAL options */
1658    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1659    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1660    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
# Line 1021  typedef struct dfa_match_data { Line 1671  typedef struct dfa_match_data {
1671    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;     /* Start of the compiled pattern */
1672    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
1673    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;    /* End of subject string */
1674      const uschar *start_used_ptr; /* Earliest consulted character */
1675    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1676      int   start_offset;           /* The start offset value */
1677    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1678    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1679    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
# Line 1036  typedef struct dfa_match_data { Line 1688  typedef struct dfa_match_data {
1688  #define ctype_letter  0x02  #define ctype_letter  0x02
1689  #define ctype_digit   0x04  #define ctype_digit   0x04
1690  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1691  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1692  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1693    
1694  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
# Line 1065  total length. */ Line 1717  total length. */
1717    
1718  /* Layout of the UCP type table that translates property names into types and  /* Layout of the UCP type table that translates property names into types and
1719  codes. Each entry used to point directly to a name, but to reduce the number of  codes. Each entry used to point directly to a name, but to reduce the number of
1720  relocations in shared libraries, it now has an offset into a single string  relocations in shared libraries, it now has an offset into a single string
1721  instead. */  instead. */
1722    
1723  typedef struct {  typedef struct {
1724    pcre_uint16 name_offset;    pcre_uint16 name_offset;
1725    pcre_uint16 type;    pcre_uint16 type;
1726    pcre_uint16 value;    pcre_uint16 value;
1727  } ucp_type_table;  } ucp_type_table;
# Line 1100  extern const uschar _pcre_OP_lengths[]; Line 1752  extern const uschar _pcre_OP_lengths[];
1752  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
1753  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1754    
1755  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1756                        int *, BOOL);  extern BOOL          _pcre_is_newline(const uschar *, int, const uschar *,
1757  extern int          _pcre_ord2utf8(int, uschar *);                         int *, BOOL);
1758  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern int           _pcre_ord2utf8(int, uschar *);
1759                        const pcre_study_data *, pcre_study_data *);  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1760  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);                         const pcre_study_data *, pcre_study_data *);
1761  extern unsigned int _pcre_ucp_othercase(const unsigned int);  extern int           _pcre_valid_utf8(const uschar *, int);
1762  extern int          _pcre_valid_utf8(const uschar *, int);  extern BOOL          _pcre_was_newline(const uschar *, int, const uschar *,
1763  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,                         int *, BOOL);
1764                        int *, BOOL);  extern BOOL          _pcre_xclass(int, const uschar *);
1765  extern BOOL         _pcre_xclass(int, const uschar *);  
1766    
1767    /* Unicode character database (UCD) */
1768    
1769    typedef struct {
1770      uschar script;
1771      uschar chartype;
1772      pcre_int32 other_case;
1773    } ucd_record;
1774    
1775    extern const ucd_record  _pcre_ucd_records[];
1776    extern const uschar      _pcre_ucd_stage1[];
1777    extern const pcre_uint16 _pcre_ucd_stage2[];
1778    extern const int         _pcre_ucp_gentype[];
1779    
1780    
1781    /* UCD access macros */
1782    
1783    #define UCD_BLOCK_SIZE 128
1784    #define GET_UCD(ch) (_pcre_ucd_records + \
1785            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1786            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1787    
1788    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1789    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1790    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1791    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1792    
1793  #endif  #endif
1794    

Legend:
Removed from v.240  
changed lines
  Added in v.475

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12