/[pcre]/code/trunk/pcre_internal.h
ViewVC logotype

Diff of /code/trunk/pcre_internal.h

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

revision 384 by ph10, Sun Mar 8 16:27:43 2009 UTC revision 550 by ph10, Sun Oct 10 16:24:11 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) 1997-2009 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 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 443  know we are in UTF-8 mode. */ Line 475  know we are in UTF-8 mode. */
475        } \        } \
476      }      }
477    
478  /* 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.
479    This is called when we don't know if we are in UTF-8 mode. */
480    
481  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
482    c = *eptr++; \    c = *eptr++; \
# Line 480  if there are extra bytes. This is called Line 513  if there are extra bytes. This is called
513    
514  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
515  pointer, incrementing length if there are extra bytes. This is called when we  pointer, incrementing length if there are extra bytes. This is called when we
516  know we are in UTF-8 mode. */  do not know if we are in UTF-8 mode. */
517    
518  #define GETCHARLENTEST(c, eptr, len) \  #define GETCHARLENTEST(c, eptr, len) \
519    c = *eptr; \    c = *eptr; \
# Line 521  Standard C system should have one. */ Line 554  Standard C system should have one. */
554    
555  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
556  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
557  are in a 16-bit flags word. */  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
558    the restrictions on partial matching have been lifted. It remains for backwards
559    compatibility. */
560    
561  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
562  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
# Line 533  are in a 16-bit flags word. */ Line 568  are in a 16-bit flags word. */
568  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
569    
570  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
571    #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */
572    
573  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
574  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 540  time, run time, or study time, respectiv Line 576  time, run time, or study time, respectiv
576  #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| \
577                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
578    
579  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
580    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
581     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
582     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| \
583     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
584     PCRE_JAVASCRIPT_COMPAT)     PCRE_JAVASCRIPT_COMPAT|PCRE_UCP)
585    
586  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
587    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
588     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
589       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
590    
591  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
592    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
593     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
594     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
595       PCRE_NO_START_OPTIMIZE)
596    
597  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
598    
# Line 590  typedef int BOOL; Line 628  typedef int BOOL;
628  #define TRUE    1  #define TRUE    1
629  #endif  #endif
630    
631    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
632    character constants like '*' because the compiler would emit their EBCDIC code,
633    which is different from their ASCII/UTF-8 code. Instead we define macros for
634    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
635    is enabled. When UTF-8 support is not enabled, the definitions use character
636    literals. Both character and string versions of each character are needed, and
637    there are some longer strings as well.
638    
639    This means that, on EBCDIC platforms, the PCRE library can handle either
640    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
641    would need different lookups depending on whether PCRE_UTF8 was set or not.
642    This would make it impossible to use characters in switch/case statements,
643    which would reduce performance. For a theoretical use (which nobody has asked
644    for) in a minority area (EBCDIC platforms), this is not sensible. Any
645    application that did need both could compile two versions of the library, using
646    macros to give the functions distinct names. */
647    
648    #ifndef SUPPORT_UTF8
649    
650    /* UTF-8 support is not enabled; use the platform-dependent character literals
651    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
652    
653    #define CHAR_HT                     '\t'
654    #define CHAR_VT                     '\v'
655    #define CHAR_FF                     '\f'
656    #define CHAR_CR                     '\r'
657    #define CHAR_NL                     '\n'
658    #define CHAR_BS                     '\b'
659    #define CHAR_BEL                    '\a'
660    #ifdef EBCDIC
661    #define CHAR_ESC                    '\047'
662    #define CHAR_DEL                    '\007'
663    #else
664    #define CHAR_ESC                    '\033'
665    #define CHAR_DEL                    '\177'
666    #endif
667    
668    #define CHAR_SPACE                  ' '
669    #define CHAR_EXCLAMATION_MARK       '!'
670    #define CHAR_QUOTATION_MARK         '"'
671    #define CHAR_NUMBER_SIGN            '#'
672    #define CHAR_DOLLAR_SIGN            '$'
673    #define CHAR_PERCENT_SIGN           '%'
674    #define CHAR_AMPERSAND              '&'
675    #define CHAR_APOSTROPHE             '\''
676    #define CHAR_LEFT_PARENTHESIS       '('
677    #define CHAR_RIGHT_PARENTHESIS      ')'
678    #define CHAR_ASTERISK               '*'
679    #define CHAR_PLUS                   '+'
680    #define CHAR_COMMA                  ','
681    #define CHAR_MINUS                  '-'
682    #define CHAR_DOT                    '.'
683    #define CHAR_SLASH                  '/'
684    #define CHAR_0                      '0'
685    #define CHAR_1                      '1'
686    #define CHAR_2                      '2'
687    #define CHAR_3                      '3'
688    #define CHAR_4                      '4'
689    #define CHAR_5                      '5'
690    #define CHAR_6                      '6'
691    #define CHAR_7                      '7'
692    #define CHAR_8                      '8'
693    #define CHAR_9                      '9'
694    #define CHAR_COLON                  ':'
695    #define CHAR_SEMICOLON              ';'
696    #define CHAR_LESS_THAN_SIGN         '<'
697    #define CHAR_EQUALS_SIGN            '='
698    #define CHAR_GREATER_THAN_SIGN      '>'
699    #define CHAR_QUESTION_MARK          '?'
700    #define CHAR_COMMERCIAL_AT          '@'
701    #define CHAR_A                      'A'
702    #define CHAR_B                      'B'
703    #define CHAR_C                      'C'
704    #define CHAR_D                      'D'
705    #define CHAR_E                      'E'
706    #define CHAR_F                      'F'
707    #define CHAR_G                      'G'
708    #define CHAR_H                      'H'
709    #define CHAR_I                      'I'
710    #define CHAR_J                      'J'
711    #define CHAR_K                      'K'
712    #define CHAR_L                      'L'
713    #define CHAR_M                      'M'
714    #define CHAR_N                      'N'
715    #define CHAR_O                      'O'
716    #define CHAR_P                      'P'
717    #define CHAR_Q                      'Q'
718    #define CHAR_R                      'R'
719    #define CHAR_S                      'S'
720    #define CHAR_T                      'T'
721    #define CHAR_U                      'U'
722    #define CHAR_V                      'V'
723    #define CHAR_W                      'W'
724    #define CHAR_X                      'X'
725    #define CHAR_Y                      'Y'
726    #define CHAR_Z                      'Z'
727    #define CHAR_LEFT_SQUARE_BRACKET    '['
728    #define CHAR_BACKSLASH              '\\'
729    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
730    #define CHAR_CIRCUMFLEX_ACCENT      '^'
731    #define CHAR_UNDERSCORE             '_'
732    #define CHAR_GRAVE_ACCENT           '`'
733    #define CHAR_a                      'a'
734    #define CHAR_b                      'b'
735    #define CHAR_c                      'c'
736    #define CHAR_d                      'd'
737    #define CHAR_e                      'e'
738    #define CHAR_f                      'f'
739    #define CHAR_g                      'g'
740    #define CHAR_h                      'h'
741    #define CHAR_i                      'i'
742    #define CHAR_j                      'j'
743    #define CHAR_k                      'k'
744    #define CHAR_l                      'l'
745    #define CHAR_m                      'm'
746    #define CHAR_n                      'n'
747    #define CHAR_o                      'o'
748    #define CHAR_p                      'p'
749    #define CHAR_q                      'q'
750    #define CHAR_r                      'r'
751    #define CHAR_s                      's'
752    #define CHAR_t                      't'
753    #define CHAR_u                      'u'
754    #define CHAR_v                      'v'
755    #define CHAR_w                      'w'
756    #define CHAR_x                      'x'
757    #define CHAR_y                      'y'
758    #define CHAR_z                      'z'
759    #define CHAR_LEFT_CURLY_BRACKET     '{'
760    #define CHAR_VERTICAL_LINE          '|'
761    #define CHAR_RIGHT_CURLY_BRACKET    '}'
762    #define CHAR_TILDE                  '~'
763    
764    #define STR_HT                      "\t"
765    #define STR_VT                      "\v"
766    #define STR_FF                      "\f"
767    #define STR_CR                      "\r"
768    #define STR_NL                      "\n"
769    #define STR_BS                      "\b"
770    #define STR_BEL                     "\a"
771    #ifdef EBCDIC
772    #define STR_ESC                     "\047"
773    #define STR_DEL                     "\007"
774    #else
775    #define STR_ESC                     "\033"
776    #define STR_DEL                     "\177"
777    #endif
778    
779    #define STR_SPACE                   " "
780    #define STR_EXCLAMATION_MARK        "!"
781    #define STR_QUOTATION_MARK          "\""
782    #define STR_NUMBER_SIGN             "#"
783    #define STR_DOLLAR_SIGN             "$"
784    #define STR_PERCENT_SIGN            "%"
785    #define STR_AMPERSAND               "&"
786    #define STR_APOSTROPHE              "'"
787    #define STR_LEFT_PARENTHESIS        "("
788    #define STR_RIGHT_PARENTHESIS       ")"
789    #define STR_ASTERISK                "*"
790    #define STR_PLUS                    "+"
791    #define STR_COMMA                   ","
792    #define STR_MINUS                   "-"
793    #define STR_DOT                     "."
794    #define STR_SLASH                   "/"
795    #define STR_0                       "0"
796    #define STR_1                       "1"
797    #define STR_2                       "2"
798    #define STR_3                       "3"
799    #define STR_4                       "4"
800    #define STR_5                       "5"
801    #define STR_6                       "6"
802    #define STR_7                       "7"
803    #define STR_8                       "8"
804    #define STR_9                       "9"
805    #define STR_COLON                   ":"
806    #define STR_SEMICOLON               ";"
807    #define STR_LESS_THAN_SIGN          "<"
808    #define STR_EQUALS_SIGN             "="
809    #define STR_GREATER_THAN_SIGN       ">"
810    #define STR_QUESTION_MARK           "?"
811    #define STR_COMMERCIAL_AT           "@"
812    #define STR_A                       "A"
813    #define STR_B                       "B"
814    #define STR_C                       "C"
815    #define STR_D                       "D"
816    #define STR_E                       "E"
817    #define STR_F                       "F"
818    #define STR_G                       "G"
819    #define STR_H                       "H"
820    #define STR_I                       "I"
821    #define STR_J                       "J"
822    #define STR_K                       "K"
823    #define STR_L                       "L"
824    #define STR_M                       "M"
825    #define STR_N                       "N"
826    #define STR_O                       "O"
827    #define STR_P                       "P"
828    #define STR_Q                       "Q"
829    #define STR_R                       "R"
830    #define STR_S                       "S"
831    #define STR_T                       "T"
832    #define STR_U                       "U"
833    #define STR_V                       "V"
834    #define STR_W                       "W"
835    #define STR_X                       "X"
836    #define STR_Y                       "Y"
837    #define STR_Z                       "Z"
838    #define STR_LEFT_SQUARE_BRACKET     "["
839    #define STR_BACKSLASH               "\\"
840    #define STR_RIGHT_SQUARE_BRACKET    "]"
841    #define STR_CIRCUMFLEX_ACCENT       "^"
842    #define STR_UNDERSCORE              "_"
843    #define STR_GRAVE_ACCENT            "`"
844    #define STR_a                       "a"
845    #define STR_b                       "b"
846    #define STR_c                       "c"
847    #define STR_d                       "d"
848    #define STR_e                       "e"
849    #define STR_f                       "f"
850    #define STR_g                       "g"
851    #define STR_h                       "h"
852    #define STR_i                       "i"
853    #define STR_j                       "j"
854    #define STR_k                       "k"
855    #define STR_l                       "l"
856    #define STR_m                       "m"
857    #define STR_n                       "n"
858    #define STR_o                       "o"
859    #define STR_p                       "p"
860    #define STR_q                       "q"
861    #define STR_r                       "r"
862    #define STR_s                       "s"
863    #define STR_t                       "t"
864    #define STR_u                       "u"
865    #define STR_v                       "v"
866    #define STR_w                       "w"
867    #define STR_x                       "x"
868    #define STR_y                       "y"
869    #define STR_z                       "z"
870    #define STR_LEFT_CURLY_BRACKET      "{"
871    #define STR_VERTICAL_LINE           "|"
872    #define STR_RIGHT_CURLY_BRACKET     "}"
873    #define STR_TILDE                   "~"
874    
875    #define STRING_ACCEPT0              "ACCEPT\0"
876    #define STRING_COMMIT0              "COMMIT\0"
877    #define STRING_F0                   "F\0"
878    #define STRING_FAIL0                "FAIL\0"
879    #define STRING_MARK0                "MARK\0"
880    #define STRING_PRUNE0               "PRUNE\0"
881    #define STRING_SKIP0                "SKIP\0"
882    #define STRING_THEN                 "THEN"
883    
884    #define STRING_alpha0               "alpha\0"
885    #define STRING_lower0               "lower\0"
886    #define STRING_upper0               "upper\0"
887    #define STRING_alnum0               "alnum\0"
888    #define STRING_ascii0               "ascii\0"
889    #define STRING_blank0               "blank\0"
890    #define STRING_cntrl0               "cntrl\0"
891    #define STRING_digit0               "digit\0"
892    #define STRING_graph0               "graph\0"
893    #define STRING_print0               "print\0"
894    #define STRING_punct0               "punct\0"
895    #define STRING_space0               "space\0"
896    #define STRING_word0                "word\0"
897    #define STRING_xdigit               "xdigit"
898    
899    #define STRING_DEFINE               "DEFINE"
900    
901    #define STRING_CR_RIGHTPAR          "CR)"
902    #define STRING_LF_RIGHTPAR          "LF)"
903    #define STRING_CRLF_RIGHTPAR        "CRLF)"
904    #define STRING_ANY_RIGHTPAR         "ANY)"
905    #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"
906    #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
907    #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
908    #define STRING_UTF8_RIGHTPAR        "UTF8)"
909    #define STRING_UCP_RIGHTPAR         "UCP)"
910    
911    #else  /* SUPPORT_UTF8 */
912    
913    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
914    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
915    only. */
916    
917    #define CHAR_HT                     '\011'
918    #define CHAR_VT                     '\013'
919    #define CHAR_FF                     '\014'
920    #define CHAR_CR                     '\015'
921    #define CHAR_NL                     '\012'
922    #define CHAR_BS                     '\010'
923    #define CHAR_BEL                    '\007'
924    #define CHAR_ESC                    '\033'
925    #define CHAR_DEL                    '\177'
926    
927    #define CHAR_SPACE                  '\040'
928    #define CHAR_EXCLAMATION_MARK       '\041'
929    #define CHAR_QUOTATION_MARK         '\042'
930    #define CHAR_NUMBER_SIGN            '\043'
931    #define CHAR_DOLLAR_SIGN            '\044'
932    #define CHAR_PERCENT_SIGN           '\045'
933    #define CHAR_AMPERSAND              '\046'
934    #define CHAR_APOSTROPHE             '\047'
935    #define CHAR_LEFT_PARENTHESIS       '\050'
936    #define CHAR_RIGHT_PARENTHESIS      '\051'
937    #define CHAR_ASTERISK               '\052'
938    #define CHAR_PLUS                   '\053'
939    #define CHAR_COMMA                  '\054'
940    #define CHAR_MINUS                  '\055'
941    #define CHAR_DOT                    '\056'
942    #define CHAR_SLASH                  '\057'
943    #define CHAR_0                      '\060'
944    #define CHAR_1                      '\061'
945    #define CHAR_2                      '\062'
946    #define CHAR_3                      '\063'
947    #define CHAR_4                      '\064'
948    #define CHAR_5                      '\065'
949    #define CHAR_6                      '\066'
950    #define CHAR_7                      '\067'
951    #define CHAR_8                      '\070'
952    #define CHAR_9                      '\071'
953    #define CHAR_COLON                  '\072'
954    #define CHAR_SEMICOLON              '\073'
955    #define CHAR_LESS_THAN_SIGN         '\074'
956    #define CHAR_EQUALS_SIGN            '\075'
957    #define CHAR_GREATER_THAN_SIGN      '\076'
958    #define CHAR_QUESTION_MARK          '\077'
959    #define CHAR_COMMERCIAL_AT          '\100'
960    #define CHAR_A                      '\101'
961    #define CHAR_B                      '\102'
962    #define CHAR_C                      '\103'
963    #define CHAR_D                      '\104'
964    #define CHAR_E                      '\105'
965    #define CHAR_F                      '\106'
966    #define CHAR_G                      '\107'
967    #define CHAR_H                      '\110'
968    #define CHAR_I                      '\111'
969    #define CHAR_J                      '\112'
970    #define CHAR_K                      '\113'
971    #define CHAR_L                      '\114'
972    #define CHAR_M                      '\115'
973    #define CHAR_N                      '\116'
974    #define CHAR_O                      '\117'
975    #define CHAR_P                      '\120'
976    #define CHAR_Q                      '\121'
977    #define CHAR_R                      '\122'
978    #define CHAR_S                      '\123'
979    #define CHAR_T                      '\124'
980    #define CHAR_U                      '\125'
981    #define CHAR_V                      '\126'
982    #define CHAR_W                      '\127'
983    #define CHAR_X                      '\130'
984    #define CHAR_Y                      '\131'
985    #define CHAR_Z                      '\132'
986    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
987    #define CHAR_BACKSLASH              '\134'
988    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
989    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
990    #define CHAR_UNDERSCORE             '\137'
991    #define CHAR_GRAVE_ACCENT           '\140'
992    #define CHAR_a                      '\141'
993    #define CHAR_b                      '\142'
994    #define CHAR_c                      '\143'
995    #define CHAR_d                      '\144'
996    #define CHAR_e                      '\145'
997    #define CHAR_f                      '\146'
998    #define CHAR_g                      '\147'
999    #define CHAR_h                      '\150'
1000    #define CHAR_i                      '\151'
1001    #define CHAR_j                      '\152'
1002    #define CHAR_k                      '\153'
1003    #define CHAR_l                      '\154'
1004    #define CHAR_m                      '\155'
1005    #define CHAR_n                      '\156'
1006    #define CHAR_o                      '\157'
1007    #define CHAR_p                      '\160'
1008    #define CHAR_q                      '\161'
1009    #define CHAR_r                      '\162'
1010    #define CHAR_s                      '\163'
1011    #define CHAR_t                      '\164'
1012    #define CHAR_u                      '\165'
1013    #define CHAR_v                      '\166'
1014    #define CHAR_w                      '\167'
1015    #define CHAR_x                      '\170'
1016    #define CHAR_y                      '\171'
1017    #define CHAR_z                      '\172'
1018    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1019    #define CHAR_VERTICAL_LINE          '\174'
1020    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1021    #define CHAR_TILDE                  '\176'
1022    
1023    #define STR_HT                      "\011"
1024    #define STR_VT                      "\013"
1025    #define STR_FF                      "\014"
1026    #define STR_CR                      "\015"
1027    #define STR_NL                      "\012"
1028    #define STR_BS                      "\010"
1029    #define STR_BEL                     "\007"
1030    #define STR_ESC                     "\033"
1031    #define STR_DEL                     "\177"
1032    
1033    #define STR_SPACE                   "\040"
1034    #define STR_EXCLAMATION_MARK        "\041"
1035    #define STR_QUOTATION_MARK          "\042"
1036    #define STR_NUMBER_SIGN             "\043"
1037    #define STR_DOLLAR_SIGN             "\044"
1038    #define STR_PERCENT_SIGN            "\045"
1039    #define STR_AMPERSAND               "\046"
1040    #define STR_APOSTROPHE              "\047"
1041    #define STR_LEFT_PARENTHESIS        "\050"
1042    #define STR_RIGHT_PARENTHESIS       "\051"
1043    #define STR_ASTERISK                "\052"
1044    #define STR_PLUS                    "\053"
1045    #define STR_COMMA                   "\054"
1046    #define STR_MINUS                   "\055"
1047    #define STR_DOT                     "\056"
1048    #define STR_SLASH                   "\057"
1049    #define STR_0                       "\060"
1050    #define STR_1                       "\061"
1051    #define STR_2                       "\062"
1052    #define STR_3                       "\063"
1053    #define STR_4                       "\064"
1054    #define STR_5                       "\065"
1055    #define STR_6                       "\066"
1056    #define STR_7                       "\067"
1057    #define STR_8                       "\070"
1058    #define STR_9                       "\071"
1059    #define STR_COLON                   "\072"
1060    #define STR_SEMICOLON               "\073"
1061    #define STR_LESS_THAN_SIGN          "\074"
1062    #define STR_EQUALS_SIGN             "\075"
1063    #define STR_GREATER_THAN_SIGN       "\076"
1064    #define STR_QUESTION_MARK           "\077"
1065    #define STR_COMMERCIAL_AT           "\100"
1066    #define STR_A                       "\101"
1067    #define STR_B                       "\102"
1068    #define STR_C                       "\103"
1069    #define STR_D                       "\104"
1070    #define STR_E                       "\105"
1071    #define STR_F                       "\106"
1072    #define STR_G                       "\107"
1073    #define STR_H                       "\110"
1074    #define STR_I                       "\111"
1075    #define STR_J                       "\112"
1076    #define STR_K                       "\113"
1077    #define STR_L                       "\114"
1078    #define STR_M                       "\115"
1079    #define STR_N                       "\116"
1080    #define STR_O                       "\117"
1081    #define STR_P                       "\120"
1082    #define STR_Q                       "\121"
1083    #define STR_R                       "\122"
1084    #define STR_S                       "\123"
1085    #define STR_T                       "\124"
1086    #define STR_U                       "\125"
1087    #define STR_V                       "\126"
1088    #define STR_W                       "\127"
1089    #define STR_X                       "\130"
1090    #define STR_Y                       "\131"
1091    #define STR_Z                       "\132"
1092    #define STR_LEFT_SQUARE_BRACKET     "\133"
1093    #define STR_BACKSLASH               "\134"
1094    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1095    #define STR_CIRCUMFLEX_ACCENT       "\136"
1096    #define STR_UNDERSCORE              "\137"
1097    #define STR_GRAVE_ACCENT            "\140"
1098    #define STR_a                       "\141"
1099    #define STR_b                       "\142"
1100    #define STR_c                       "\143"
1101    #define STR_d                       "\144"
1102    #define STR_e                       "\145"
1103    #define STR_f                       "\146"
1104    #define STR_g                       "\147"
1105    #define STR_h                       "\150"
1106    #define STR_i                       "\151"
1107    #define STR_j                       "\152"
1108    #define STR_k                       "\153"
1109    #define STR_l                       "\154"
1110    #define STR_m                       "\155"
1111    #define STR_n                       "\156"
1112    #define STR_o                       "\157"
1113    #define STR_p                       "\160"
1114    #define STR_q                       "\161"
1115    #define STR_r                       "\162"
1116    #define STR_s                       "\163"
1117    #define STR_t                       "\164"
1118    #define STR_u                       "\165"
1119    #define STR_v                       "\166"
1120    #define STR_w                       "\167"
1121    #define STR_x                       "\170"
1122    #define STR_y                       "\171"
1123    #define STR_z                       "\172"
1124    #define STR_LEFT_CURLY_BRACKET      "\173"
1125    #define STR_VERTICAL_LINE           "\174"
1126    #define STR_RIGHT_CURLY_BRACKET     "\175"
1127    #define STR_TILDE                   "\176"
1128    
1129    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1130    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1131    #define STRING_F0                   STR_F "\0"
1132    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1133    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1134    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1135    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1136    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1137    
1138    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1139    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1140    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1141    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1142    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1143    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1144    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1145    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1146    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1147    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1148    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1149    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1150    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1151    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1152    
1153    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1154    
1155    #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS
1156    #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS
1157    #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1158    #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1159    #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1160    #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
1161    #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
1162    #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1163    #define STRING_UCP_RIGHTPAR         STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1164    
1165    #endif  /* SUPPORT_UTF8 */
1166    
1167  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1168    
1169  #ifndef ESC_e  #ifndef ESC_e
1170  #define ESC_e 27  #define ESC_e CHAR_ESC
1171  #endif  #endif
1172    
1173  #ifndef ESC_f  #ifndef ESC_f
1174  #define ESC_f '\f'  #define ESC_f CHAR_FF
1175  #endif  #endif
1176    
1177  #ifndef ESC_n  #ifndef ESC_n
1178  #define ESC_n '\n'  #define ESC_n CHAR_NL
1179  #endif  #endif
1180    
1181  #ifndef ESC_r  #ifndef ESC_r
1182  #define ESC_r '\r'  #define ESC_r CHAR_CR
1183  #endif  #endif
1184    
1185  /* 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
1186  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1187    
1188  #ifndef ESC_tee  #ifndef ESC_tee
1189  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1190  #endif  #endif
1191    
1192  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
1193    
1194  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1195  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1196  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1197  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1198  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1199    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1200    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1201    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1202    #define PT_WORD       8    /* Word - L plus N plus underscore */
1203    
1204  /* 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
1205  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 638  contain UTF-8 characters with values gre Line 1216  contain UTF-8 characters with values gre
1216  /* 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
1217  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
1218  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
1219  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
1220  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
1221  (which is used for [^] in JavaScript compatibility mode).  used for [^] in JavaScript compatibility mode. In non-DOTALL mode, "." behaves
1222    like \N.
1223    
1224    The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1225    when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1226    They must be contiguous, and remain in order so that the replacements can be
1227    looked up from a table.
1228    
1229  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
1230  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
# Line 650  put in between that don't consume a char Line 1234  put in between that don't consume a char
1234  */  */
1235    
1236  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,
1237         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,
1238         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,
1239           ESC_E, ESC_Q, ESC_g, ESC_k,
1240           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1241         ESC_REF };         ESC_REF };
1242    
   
1243  /* 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
1244  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.
1245    
1246  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1247  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
1248  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1249    
1250  enum {  enum {
1251    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 678  enum { Line 1263  enum {
1263    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1264    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1265    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1266    OP_ANY,            /* 12 Match any character (subject to DOTALL) */    OP_ANY,            /* 12 Match any character except newline */
1267    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */    OP_ALLANY,         /* 13 Match any character */
1268    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 */
1269    OP_NOTPROP,        /* 15 \P (not Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1270    OP_PROP,           /* 16 \p (Unicode property) */    OP_PROP,           /* 16 \p (Unicode property) */
# Line 796  enum { Line 1381  enum {
1381    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1382    OP_SCOND,          /* 99 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
1383    
1384      /* The next two pairs must (respectively) be kept together. */
1385    
1386    OP_CREF,           /* 100 Used to hold a capture number as condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
1387    OP_RREF,           /* 101 Used to hold a recursion number as condition */    OP_NCREF,          /* 101 Same, but generaged by a name reference*/
1388    OP_DEF,            /* 102 The DEFINE condition */    OP_RREF,           /* 102 Used to hold a recursion number as condition */
1389      OP_NRREF,          /* 103 Same, but generaged by a name reference*/
1390      OP_DEF,            /* 104 The DEFINE condition */
1391    
1392    OP_BRAZERO,        /* 103 These two must remain together and in this */    OP_BRAZERO,        /* 105 These two must remain together and in this */
1393    OP_BRAMINZERO,     /* 104 order. */    OP_BRAMINZERO,     /* 106 order. */
1394    
1395    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1396    
1397    OP_PRUNE,          /* 105 */    OP_MARK,           /* 107 always has an argument */
1398    OP_SKIP,           /* 106 */    OP_PRUNE,          /* 108 */
1399    OP_THEN,           /* 107 */    OP_PRUNE_ARG,      /* 109 same, but with argument */
1400    OP_COMMIT,         /* 108 */    OP_SKIP,           /* 110 */
1401      OP_SKIP_ARG,       /* 111 same, but with argument */
1402      OP_THEN,           /* 112 */
1403      OP_THEN_ARG,       /* 113 same, but with argument */
1404      OP_COMMIT,         /* 114 */
1405    
1406    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1407    
1408    OP_FAIL,           /* 109 */    OP_FAIL,           /* 115 */
1409    OP_ACCEPT,         /* 110 */    OP_ACCEPT,         /* 116 */
1410      OP_CLOSE,          /* 117 Used before OP_ACCEPT to close open captures */
1411    
1412    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1413    
1414    OP_SKIPZERO        /* 111 */    OP_SKIPZERO,       /* 118 */
1415    
1416      /* This is not an opcode, but is used to check that tables indexed by opcode
1417      are the correct length, in order to catch updating errors - there have been
1418      some in the past. */
1419    
1420      OP_TABLE_LENGTH
1421  };  };
1422    
1423    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1424    definitions that follow must also be updated to match. There are also tables
1425    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1426    
1427    
1428  /* 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
1429  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
# Line 841  for debugging. The macro is referenced o Line 1445  for debugging. The macro is referenced o
1445    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1446    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1447    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1448    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1449    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "Brazero", "Braminzero",                                        \
1450    "Skip zero"    "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1451      "*THEN", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",                \
1452      "Close", "Skip zero"
1453    
1454    
1455  /* This macro defines the length of fixed length operations in the compiled  /* This macro defines the length of fixed length operations in the compiled
# Line 860  in UTF-8 mode. The code that uses this t Line 1466  in UTF-8 mode. The code that uses this t
1466    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1467    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1468    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1469    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1470    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1471      1,                             /* \X                                     */ \
1472    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
1473    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1474    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 903  in UTF-8 mode. The code that uses this t Line 1510  in UTF-8 mode. The code that uses this t
1510    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1511    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1512    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1513    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1514    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1515    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1516    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1517    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1518    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */    1, 3,                          /* SKIP, SKIP_ARG                         */ \
1519      1+LINK_SIZE, 3+LINK_SIZE,      /* THEN, THEN_ARG                         */ \
1520      1, 1, 1, 3, 1                  /* COMMIT, FAIL, ACCEPT, CLOSE, SKIPZERO  */
1521    
1522    
1523  /* 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"
1524    condition. */
1525    
1526  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1527    
1528  /* 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
1529  tracked. */  easily be tracked. When a new number is added, the table called eint in
1530    pcreposix.c must be updated. */
1531    
1532  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1533         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
# Line 924  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 1535  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1535         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1536         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1537         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1538         ERR60, ERR61, ERR62, ERR63, ERR64 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERRCOUNT };
1539    
1540  /* 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
1541  code vector run on as long as necessary after the end. We store an explicit  code vector run on as long as necessary after the end. We store an explicit
# Line 940  Because people can now save and re-use c Line 1551  Because people can now save and re-use c
1551  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
1552  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
1553  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1554  NOTE NOTE NOTE:  NOTE NOTE NOTE
1555  */  */
1556    
1557  typedef struct real_pcre {  typedef struct real_pcre {
# Line 967  remark (see NOTE above) about extending Line 1578  remark (see NOTE above) about extending
1578    
1579  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1580    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1581    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1582    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1583      pcre_uint32 minlength;          /* Minimum subject length */
1584  } pcre_study_data;  } pcre_study_data;
1585    
1586    /* Structure for building a chain of open capturing subpatterns during
1587    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1588    encountered. This is also used to identify subpatterns that contain recursive
1589    back references to themselves, so that they can be made atomic. */
1590    
1591    typedef struct open_capitem {
1592      struct open_capitem *next;    /* Chain link */
1593      pcre_uint16 number;           /* Capture number */
1594      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1595    } open_capitem;
1596    
1597  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1598  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1599    
# Line 983  typedef struct compile_data { Line 1606  typedef struct compile_data {
1606    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1607    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1608    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1609      open_capitem *open_caps;      /* Chain of open capture items */
1610    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1611    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1612    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
# Line 995  typedef struct compile_data { Line 1619  typedef struct compile_data {
1619    int  external_flags;          /* External flag bits to be set */    int  external_flags;          /* External flag bits to be set */
1620    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1621    BOOL had_accept;              /* (*ACCEPT) encountered */    BOOL had_accept;              /* (*ACCEPT) encountered */
1622      BOOL check_lookbehind;        /* Lookbehinds need later checking */
1623    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1624    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1625    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 1005  branches, for testing for left recursion Line 1630  branches, for testing for left recursion
1630    
1631  typedef struct branch_chain {  typedef struct branch_chain {
1632    struct branch_chain *outer;    struct branch_chain *outer;
1633    uschar *current;    uschar *current_branch;
1634  } branch_chain;  } branch_chain;
1635    
1636  /* 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 1015  typedef struct recursion_info { Line 1640  typedef struct recursion_info {
1640    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1641    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
1642    const uschar *after_call;     /* "Return value": points after the call in the expr */    const uschar *after_call;     /* "Return value": points after the call in the expr */
   USPTR save_start;             /* Old value of mstart */  
1643    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1644    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1645      int save_offset_top;          /* Current value of offset_top */
1646  } recursion_info;  } recursion_info;
1647    
1648  /* 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 1042  typedef struct match_data { Line 1667  typedef struct match_data {
1667    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1668    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1669    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1670      int    name_count;            /* Number of names in name table */
1671      int    name_entry_size;       /* Size of entry in names table */
1672      uschar *name_table;           /* Table of names */
1673    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1674    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1675    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
# Line 1050  typedef struct match_data { Line 1678  typedef struct match_data {
1678    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1679    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1680    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1681      BOOL   use_ucp;               /* PCRE_UCP flag */
1682    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1683    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1684    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1685    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1686    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1687    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
# Line 1060  typedef struct match_data { Line 1689  typedef struct match_data {
1689    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1690    USPTR  start_match_ptr;       /* Start of matched string */    USPTR  start_match_ptr;       /* Start of matched string */
1691    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1692      USPTR  start_used_ptr;        /* Earliest consulted character */
1693      int    partial;               /* PARTIAL options */
1694    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1695    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1696    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
# Line 1067  typedef struct match_data { Line 1698  typedef struct match_data {
1698    int    eptrn;                 /* Next free eptrblock */    int    eptrn;                 /* Next free eptrblock */
1699    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1700    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1701      const uschar *mark;           /* Mark pointer to pass back */
1702  } match_data;  } match_data;
1703    
1704  /* 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
# Line 1076  typedef struct dfa_match_data { Line 1708  typedef struct dfa_match_data {
1708    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;     /* Start of the compiled pattern */
1709    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
1710    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;    /* End of subject string */
1711      const uschar *start_used_ptr; /* Earliest consulted character */
1712    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1713      int   start_offset;           /* The start offset value */
1714    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1715    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1716    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
# Line 1155  extern const uschar _pcre_OP_lengths[]; Line 1789  extern const uschar _pcre_OP_lengths[];
1789  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
1790  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1791    
1792  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1793                        int *, BOOL);  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);
1794  extern int          _pcre_ord2utf8(int, uschar *);  extern int           _pcre_ord2utf8(int, uschar *);
1795  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1796                        const pcre_study_data *, pcre_study_data *);                         const pcre_study_data *, pcre_study_data *);
1797  extern int          _pcre_valid_utf8(const uschar *, int);  extern int           _pcre_valid_utf8(USPTR, int);
1798  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);
1799                        int *, BOOL);  extern BOOL          _pcre_xclass(int, const uschar *);
 extern BOOL         _pcre_xclass(int, const uschar *);  
1800    
1801    
1802  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */

Legend:
Removed from v.384  
changed lines
  Added in v.550

  ViewVC Help
Powered by ViewVC 1.1.5