/[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 230 by ph10, Mon Sep 10 13:23:56 2007 UTC revision 393 by ph10, Tue Mar 17 21:47:54 2009 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) 1997-2009 University of Cambridge
11    
12  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
13  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 51  functions whose names all begin with "_p Line 51  functions whose names all begin with "_p
51  #define DEBUG  #define DEBUG
52  #endif  #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
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
69  inline, and there are *still* stupid compilers about that don't like indented  inline, and there are *still* stupid compilers about that don't like indented
70  pre-processor statements, or at least there were when I first wrote this. After  pre-processor statements, or at least there were when I first wrote this. After
# Line 132  PCRE_EXP_DATA_DEFN only if they are not Line 146  PCRE_EXP_DATA_DEFN only if they are not
146  #  endif  #  endif
147  #endif  #endif
148    
149    /* When compiling with the MSVC compiler, it is sometimes necessary to include
150    a "calling convention" before exported function names. (This is secondhand
151    information; I know nothing about MSVC myself). For example, something like
152    
153      void __cdecl function(....)
154    
155    might be needed. In order so make this easy, all the exported functions have
156    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
157    set, we ensure here that it has no effect. */
158    
159    #ifndef PCRE_CALL_CONVENTION
160    #define PCRE_CALL_CONVENTION
161    #endif
162    
163  /* 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
164  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
165  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 168  preprocessor time in standard C environm
168    
169  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
170    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
171      typedef short pcre_int16;
172  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
173    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
174      typedef int pcre_int16;
175  #else  #else
176    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
177  #endif  #endif
178    
179  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
180    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
181      typedef int pcre_int32;
182  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
183    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
184      typedef long int pcre_int32;
185  #else  #else
186    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
187  #endif  #endif
# Line 460  if there are extra bytes. This is called Line 492  if there are extra bytes. This is called
492      len += gcaa; \      len += gcaa; \
493      }      }
494    
495    /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
496    pointer, incrementing length if there are extra bytes. This is called when we
497    know we are in UTF-8 mode. */
498    
499    #define GETCHARLENTEST(c, eptr, len) \
500      c = *eptr; \
501      if (utf8 && c >= 0xc0) \
502        { \
503        int gcii; \
504        int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
505        int gcss = 6*gcaa; \
506        c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
507        for (gcii = 1; gcii <= gcaa; gcii++) \
508          { \
509          gcss -= 6; \
510          c |= (eptr[gcii] & 0x3f) << gcss; \
511          } \
512        len += gcaa; \
513        }
514    
515  /* 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
516  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
517  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 481  Standard C system should have one. */ Line 533  Standard C system should have one. */
533    
534  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
535    
536  /* Private flags containing information about the compiled regex. They used to  /* Private flags containing information about the compiled regex. They used to
537  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
538  are in a 16-bit flags word. */  are in a 16-bit flags word. */
539    
540  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
# Line 502  time, run time, or study time, respectiv Line 554  time, run time, or study time, respectiv
554  #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| \
555                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
556    
557  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
558    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
559     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
560     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| \
561     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
562       PCRE_JAVASCRIPT_COMPAT)
563    
564  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
565    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
566     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
567       PCRE_NO_START_OPTIMIZE)
568    
569  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
570    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
571     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
572       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
573    
574  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
575    
# Line 539  variable-length repeat, or a anything ot Line 594  variable-length repeat, or a anything ot
594  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
595  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
596    
597  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
598    environments where these macros are defined elsewhere. Unfortunately, there
599    is no way to do the same for the typedef. */
600    
601  typedef int BOOL;  typedef int BOOL;
602    
603    #ifndef FALSE
604  #define FALSE   0  #define FALSE   0
605  #define TRUE    1  #define TRUE    1
606    #endif
607    
608    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
609    character constants like '*' because the compiler would emit their EBCDIC code,
610    which is different from their ASCII/UTF-8 code. Instead we define macros for
611    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
612    is enabled. When UTF-8 support is not enabled, the definitions use character
613    literals. Both character and string versions of each character are needed, and
614    there are some longer strings as well.
615    
616    This means that, on EBCDIC platforms, the PCRE library can handle either
617    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
618    would need different lookups depending on whether PCRE_UTF8 was set or not.
619    This would make it impossible to use characters in switch/case statements,
620    which would reduce performance. For a theoretical use (which nobody has asked
621    for) in a minority area (EBCDIC platforms), this is not sensible. Any
622    application that did need both could compile two versions of the library, using
623    macros to give the functions distinct names. */
624    
625    #ifndef SUPPORT_UTF8
626    
627    /* UTF-8 support is not enabled; use the platform-dependent character literals
628    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
629    
630    #define CHAR_HT                     '\t'
631    #define CHAR_VT                     '\v'
632    #define CHAR_FF                     '\f'
633    #define CHAR_CR                     '\r'
634    #define CHAR_NL                     '\n'
635    #define CHAR_BS                     '\b'
636    #define CHAR_BEL                    '\a'
637    #ifdef EBCDIC
638    #define CHAR_ESC                    '\047'
639    #define CHAR_DEL                    '\007'
640    #else
641    #define CHAR_ESC                    '\033'
642    #define CHAR_DEL                    '\177'
643    #endif
644    
645    #define CHAR_SPACE                  ' '
646    #define CHAR_EXCLAMATION_MARK       '!'
647    #define CHAR_QUOTATION_MARK         '"'
648    #define CHAR_NUMBER_SIGN            '#'
649    #define CHAR_DOLLAR_SIGN            '$'
650    #define CHAR_PERCENT_SIGN           '%'
651    #define CHAR_AMPERSAND              '&'
652    #define CHAR_APOSTROPHE             '\''
653    #define CHAR_LEFT_PARENTHESIS       '('
654    #define CHAR_RIGHT_PARENTHESIS      ')'
655    #define CHAR_ASTERISK               '*'
656    #define CHAR_PLUS                   '+'
657    #define CHAR_COMMA                  ','
658    #define CHAR_MINUS                  '-'
659    #define CHAR_DOT                    '.'
660    #define CHAR_SLASH                  '/'
661    #define CHAR_0                      '0'
662    #define CHAR_1                      '1'
663    #define CHAR_2                      '2'
664    #define CHAR_3                      '3'
665    #define CHAR_4                      '4'
666    #define CHAR_5                      '5'
667    #define CHAR_6                      '6'
668    #define CHAR_7                      '7'
669    #define CHAR_8                      '8'
670    #define CHAR_9                      '9'
671    #define CHAR_COLON                  ':'
672    #define CHAR_SEMICOLON              ';'
673    #define CHAR_LESS_THAN_SIGN         '<'
674    #define CHAR_EQUALS_SIGN            '='
675    #define CHAR_GREATER_THAN_SIGN      '>'
676    #define CHAR_QUESTION_MARK          '?'
677    #define CHAR_COMMERCIAL_AT          '@'
678    #define CHAR_A                      'A'
679    #define CHAR_B                      'B'
680    #define CHAR_C                      'C'
681    #define CHAR_D                      'D'
682    #define CHAR_E                      'E'
683    #define CHAR_F                      'F'
684    #define CHAR_G                      'G'
685    #define CHAR_H                      'H'
686    #define CHAR_I                      'I'
687    #define CHAR_J                      'J'
688    #define CHAR_K                      'K'
689    #define CHAR_L                      'L'
690    #define CHAR_M                      'M'
691    #define CHAR_N                      'N'
692    #define CHAR_O                      'O'
693    #define CHAR_P                      'P'
694    #define CHAR_Q                      'Q'
695    #define CHAR_R                      'R'
696    #define CHAR_S                      'S'
697    #define CHAR_T                      'T'
698    #define CHAR_U                      'U'
699    #define CHAR_V                      'V'
700    #define CHAR_W                      'W'
701    #define CHAR_X                      'X'
702    #define CHAR_Y                      'Y'
703    #define CHAR_Z                      'Z'
704    #define CHAR_LEFT_SQUARE_BRACKET    '['
705    #define CHAR_BACKSLASH              '\\'
706    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
707    #define CHAR_CIRCUMFLEX_ACCENT      '^'
708    #define CHAR_UNDERSCORE             '_'
709    #define CHAR_GRAVE_ACCENT           '`'
710    #define CHAR_a                      'a'
711    #define CHAR_b                      'b'
712    #define CHAR_c                      'c'
713    #define CHAR_d                      'd'
714    #define CHAR_e                      'e'
715    #define CHAR_f                      'f'
716    #define CHAR_g                      'g'
717    #define CHAR_h                      'h'
718    #define CHAR_i                      'i'
719    #define CHAR_j                      'j'
720    #define CHAR_k                      'k'
721    #define CHAR_l                      'l'
722    #define CHAR_m                      'm'
723    #define CHAR_n                      'n'
724    #define CHAR_o                      'o'
725    #define CHAR_p                      'p'
726    #define CHAR_q                      'q'
727    #define CHAR_r                      'r'
728    #define CHAR_s                      's'
729    #define CHAR_t                      't'
730    #define CHAR_u                      'u'
731    #define CHAR_v                      'v'
732    #define CHAR_w                      'w'
733    #define CHAR_x                      'x'
734    #define CHAR_y                      'y'
735    #define CHAR_z                      'z'
736    #define CHAR_LEFT_CURLY_BRACKET     '{'
737    #define CHAR_VERTICAL_LINE          '|'
738    #define CHAR_RIGHT_CURLY_BRACKET    '}'
739    #define CHAR_TILDE                  '~'
740    
741    #define STR_HT                      "\t"
742    #define STR_VT                      "\v"
743    #define STR_FF                      "\f"
744    #define STR_CR                      "\r"
745    #define STR_NL                      "\n"
746    #define STR_BS                      "\b"
747    #define STR_BEL                     "\a"
748    #ifdef EBCDIC
749    #define STR_ESC                     "\047"
750    #define STR_DEL                     "\007"
751    #else
752    #define STR_ESC                     "\033"
753    #define STR_DEL                     "\177"
754    #endif
755    
756    #define STR_SPACE                   " "
757    #define STR_EXCLAMATION_MARK        "!"
758    #define STR_QUOTATION_MARK          "\""
759    #define STR_NUMBER_SIGN             "#"
760    #define STR_DOLLAR_SIGN             "$"
761    #define STR_PERCENT_SIGN            "%"
762    #define STR_AMPERSAND               "&"
763    #define STR_APOSTROPHE              "'"
764    #define STR_LEFT_PARENTHESIS        "("
765    #define STR_RIGHT_PARENTHESIS       ")"
766    #define STR_ASTERISK                "*"
767    #define STR_PLUS                    "+"
768    #define STR_COMMA                   ","
769    #define STR_MINUS                   "-"
770    #define STR_DOT                     "."
771    #define STR_SLASH                   "/"
772    #define STR_0                       "0"
773    #define STR_1                       "1"
774    #define STR_2                       "2"
775    #define STR_3                       "3"
776    #define STR_4                       "4"
777    #define STR_5                       "5"
778    #define STR_6                       "6"
779    #define STR_7                       "7"
780    #define STR_8                       "8"
781    #define STR_9                       "9"
782    #define STR_COLON                   ":"
783    #define STR_SEMICOLON               ";"
784    #define STR_LESS_THAN_SIGN          "<"
785    #define STR_EQUALS_SIGN             "="
786    #define STR_GREATER_THAN_SIGN       ">"
787    #define STR_QUESTION_MARK           "?"
788    #define STR_COMMERCIAL_AT           "@"
789    #define STR_A                       "A"
790    #define STR_B                       "B"
791    #define STR_C                       "C"
792    #define STR_D                       "D"
793    #define STR_E                       "E"
794    #define STR_F                       "F"
795    #define STR_G                       "G"
796    #define STR_H                       "H"
797    #define STR_I                       "I"
798    #define STR_J                       "J"
799    #define STR_K                       "K"
800    #define STR_L                       "L"
801    #define STR_M                       "M"
802    #define STR_N                       "N"
803    #define STR_O                       "O"
804    #define STR_P                       "P"
805    #define STR_Q                       "Q"
806    #define STR_R                       "R"
807    #define STR_S                       "S"
808    #define STR_T                       "T"
809    #define STR_U                       "U"
810    #define STR_V                       "V"
811    #define STR_W                       "W"
812    #define STR_X                       "X"
813    #define STR_Y                       "Y"
814    #define STR_Z                       "Z"
815    #define STR_LEFT_SQUARE_BRACKET     "["
816    #define STR_BACKSLASH               "\\"
817    #define STR_RIGHT_SQUARE_BRACKET    "]"
818    #define STR_CIRCUMFLEX_ACCENT       "^"
819    #define STR_UNDERSCORE              "_"
820    #define STR_GRAVE_ACCENT            "`"
821    #define STR_a                       "a"
822    #define STR_b                       "b"
823    #define STR_c                       "c"
824    #define STR_d                       "d"
825    #define STR_e                       "e"
826    #define STR_f                       "f"
827    #define STR_g                       "g"
828    #define STR_h                       "h"
829    #define STR_i                       "i"
830    #define STR_j                       "j"
831    #define STR_k                       "k"
832    #define STR_l                       "l"
833    #define STR_m                       "m"
834    #define STR_n                       "n"
835    #define STR_o                       "o"
836    #define STR_p                       "p"
837    #define STR_q                       "q"
838    #define STR_r                       "r"
839    #define STR_s                       "s"
840    #define STR_t                       "t"
841    #define STR_u                       "u"
842    #define STR_v                       "v"
843    #define STR_w                       "w"
844    #define STR_x                       "x"
845    #define STR_y                       "y"
846    #define STR_z                       "z"
847    #define STR_LEFT_CURLY_BRACKET      "{"
848    #define STR_VERTICAL_LINE           "|"
849    #define STR_RIGHT_CURLY_BRACKET     "}"
850    #define STR_TILDE                   "~"
851    
852    #define STRING_ACCEPT0              "ACCEPT\0"
853    #define STRING_COMMIT0              "COMMIT\0"
854    #define STRING_F0                   "F\0"
855    #define STRING_FAIL0                "FAIL\0"
856    #define STRING_PRUNE0               "PRUNE\0"
857    #define STRING_SKIP0                "SKIP\0"
858    #define STRING_THEN                 "THEN"
859    
860    #define STRING_alpha0               "alpha\0"
861    #define STRING_lower0               "lower\0"
862    #define STRING_upper0               "upper\0"
863    #define STRING_alnum0               "alnum\0"
864    #define STRING_ascii0               "ascii\0"
865    #define STRING_blank0               "blank\0"
866    #define STRING_cntrl0               "cntrl\0"
867    #define STRING_digit0               "digit\0"
868    #define STRING_graph0               "graph\0"
869    #define STRING_print0               "print\0"
870    #define STRING_punct0               "punct\0"
871    #define STRING_space0               "space\0"
872    #define STRING_word0                "word\0"
873    #define STRING_xdigit               "xdigit"
874    
875    #define STRING_DEFINE               "DEFINE"
876    
877    #define STRING_CR_RIGHTPAR          "CR)"
878    #define STRING_LF_RIGHTPAR          "LF)"
879    #define STRING_CRLF_RIGHTPAR        "CRLF)"
880    #define STRING_ANY_RIGHTPAR         "ANY)"
881    #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"
882    #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
883    #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
884    
885    #else  /* SUPPORT_UTF8 */
886    
887    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
888    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
889    only. */
890    
891    #define CHAR_HT                     '\011'
892    #define CHAR_VT                     '\013'
893    #define CHAR_FF                     '\014'
894    #define CHAR_CR                     '\015'
895    #define CHAR_NL                     '\012'
896    #define CHAR_BS                     '\010'
897    #define CHAR_BEL                    '\007'
898    #define CHAR_ESC                    '\033'
899    #define CHAR_DEL                    '\177'
900    
901    #define CHAR_SPACE                  '\040'
902    #define CHAR_EXCLAMATION_MARK       '\041'
903    #define CHAR_QUOTATION_MARK         '\042'
904    #define CHAR_NUMBER_SIGN            '\043'
905    #define CHAR_DOLLAR_SIGN            '\044'
906    #define CHAR_PERCENT_SIGN           '\045'
907    #define CHAR_AMPERSAND              '\046'
908    #define CHAR_APOSTROPHE             '\047'
909    #define CHAR_LEFT_PARENTHESIS       '\050'
910    #define CHAR_RIGHT_PARENTHESIS      '\051'
911    #define CHAR_ASTERISK               '\052'
912    #define CHAR_PLUS                   '\053'
913    #define CHAR_COMMA                  '\054'
914    #define CHAR_MINUS                  '\055'
915    #define CHAR_DOT                    '\056'
916    #define CHAR_SLASH                  '\057'
917    #define CHAR_0                      '\060'
918    #define CHAR_1                      '\061'
919    #define CHAR_2                      '\062'
920    #define CHAR_3                      '\063'
921    #define CHAR_4                      '\064'
922    #define CHAR_5                      '\065'
923    #define CHAR_6                      '\066'
924    #define CHAR_7                      '\067'
925    #define CHAR_8                      '\070'
926    #define CHAR_9                      '\071'
927    #define CHAR_COLON                  '\072'
928    #define CHAR_SEMICOLON              '\073'
929    #define CHAR_LESS_THAN_SIGN         '\074'
930    #define CHAR_EQUALS_SIGN            '\075'
931    #define CHAR_GREATER_THAN_SIGN      '\076'
932    #define CHAR_QUESTION_MARK          '\077'
933    #define CHAR_COMMERCIAL_AT          '\100'
934    #define CHAR_A                      '\101'
935    #define CHAR_B                      '\102'
936    #define CHAR_C                      '\103'
937    #define CHAR_D                      '\104'
938    #define CHAR_E                      '\105'
939    #define CHAR_F                      '\106'
940    #define CHAR_G                      '\107'
941    #define CHAR_H                      '\110'
942    #define CHAR_I                      '\111'
943    #define CHAR_J                      '\112'
944    #define CHAR_K                      '\113'
945    #define CHAR_L                      '\114'
946    #define CHAR_M                      '\115'
947    #define CHAR_N                      '\116'
948    #define CHAR_O                      '\117'
949    #define CHAR_P                      '\120'
950    #define CHAR_Q                      '\121'
951    #define CHAR_R                      '\122'
952    #define CHAR_S                      '\123'
953    #define CHAR_T                      '\124'
954    #define CHAR_U                      '\125'
955    #define CHAR_V                      '\126'
956    #define CHAR_W                      '\127'
957    #define CHAR_X                      '\130'
958    #define CHAR_Y                      '\131'
959    #define CHAR_Z                      '\132'
960    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
961    #define CHAR_BACKSLASH              '\134'
962    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
963    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
964    #define CHAR_UNDERSCORE             '\137'
965    #define CHAR_GRAVE_ACCENT           '\140'
966    #define CHAR_a                      '\141'
967    #define CHAR_b                      '\142'
968    #define CHAR_c                      '\143'
969    #define CHAR_d                      '\144'
970    #define CHAR_e                      '\145'
971    #define CHAR_f                      '\146'
972    #define CHAR_g                      '\147'
973    #define CHAR_h                      '\150'
974    #define CHAR_i                      '\151'
975    #define CHAR_j                      '\152'
976    #define CHAR_k                      '\153'
977    #define CHAR_l                      '\154'
978    #define CHAR_m                      '\155'
979    #define CHAR_n                      '\156'
980    #define CHAR_o                      '\157'
981    #define CHAR_p                      '\160'
982    #define CHAR_q                      '\161'
983    #define CHAR_r                      '\162'
984    #define CHAR_s                      '\163'
985    #define CHAR_t                      '\164'
986    #define CHAR_u                      '\165'
987    #define CHAR_v                      '\166'
988    #define CHAR_w                      '\167'
989    #define CHAR_x                      '\170'
990    #define CHAR_y                      '\171'
991    #define CHAR_z                      '\172'
992    #define CHAR_LEFT_CURLY_BRACKET     '\173'
993    #define CHAR_VERTICAL_LINE          '\174'
994    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
995    #define CHAR_TILDE                  '\176'
996    
997    #define STR_HT                      "\011"
998    #define STR_VT                      "\013"
999    #define STR_FF                      "\014"
1000    #define STR_CR                      "\015"
1001    #define STR_NL                      "\012"
1002    #define STR_BS                      "\010"
1003    #define STR_BEL                     "\007"
1004    #define STR_ESC                     "\033"
1005    #define STR_DEL                     "\177"
1006    
1007    #define STR_SPACE                   "\040"
1008    #define STR_EXCLAMATION_MARK        "\041"
1009    #define STR_QUOTATION_MARK          "\042"
1010    #define STR_NUMBER_SIGN             "\043"
1011    #define STR_DOLLAR_SIGN             "\044"
1012    #define STR_PERCENT_SIGN            "\045"
1013    #define STR_AMPERSAND               "\046"
1014    #define STR_APOSTROPHE              "\047"
1015    #define STR_LEFT_PARENTHESIS        "\050"
1016    #define STR_RIGHT_PARENTHESIS       "\051"
1017    #define STR_ASTERISK                "\052"
1018    #define STR_PLUS                    "\053"
1019    #define STR_COMMA                   "\054"
1020    #define STR_MINUS                   "\055"
1021    #define STR_DOT                     "\056"
1022    #define STR_SLASH                   "\057"
1023    #define STR_0                       "\060"
1024    #define STR_1                       "\061"
1025    #define STR_2                       "\062"
1026    #define STR_3                       "\063"
1027    #define STR_4                       "\064"
1028    #define STR_5                       "\065"
1029    #define STR_6                       "\066"
1030    #define STR_7                       "\067"
1031    #define STR_8                       "\070"
1032    #define STR_9                       "\071"
1033    #define STR_COLON                   "\072"
1034    #define STR_SEMICOLON               "\073"
1035    #define STR_LESS_THAN_SIGN          "\074"
1036    #define STR_EQUALS_SIGN             "\075"
1037    #define STR_GREATER_THAN_SIGN       "\076"
1038    #define STR_QUESTION_MARK           "\077"
1039    #define STR_COMMERCIAL_AT           "\100"
1040    #define STR_A                       "\101"
1041    #define STR_B                       "\102"
1042    #define STR_C                       "\103"
1043    #define STR_D                       "\104"
1044    #define STR_E                       "\105"
1045    #define STR_F                       "\106"
1046    #define STR_G                       "\107"
1047    #define STR_H                       "\110"
1048    #define STR_I                       "\111"
1049    #define STR_J                       "\112"
1050    #define STR_K                       "\113"
1051    #define STR_L                       "\114"
1052    #define STR_M                       "\115"
1053    #define STR_N                       "\116"
1054    #define STR_O                       "\117"
1055    #define STR_P                       "\120"
1056    #define STR_Q                       "\121"
1057    #define STR_R                       "\122"
1058    #define STR_S                       "\123"
1059    #define STR_T                       "\124"
1060    #define STR_U                       "\125"
1061    #define STR_V                       "\126"
1062    #define STR_W                       "\127"
1063    #define STR_X                       "\130"
1064    #define STR_Y                       "\131"
1065    #define STR_Z                       "\132"
1066    #define STR_LEFT_SQUARE_BRACKET     "\133"
1067    #define STR_BACKSLASH               "\134"
1068    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1069    #define STR_CIRCUMFLEX_ACCENT       "\136"
1070    #define STR_UNDERSCORE              "\137"
1071    #define STR_GRAVE_ACCENT            "\140"
1072    #define STR_a                       "\141"
1073    #define STR_b                       "\142"
1074    #define STR_c                       "\143"
1075    #define STR_d                       "\144"
1076    #define STR_e                       "\145"
1077    #define STR_f                       "\146"
1078    #define STR_g                       "\147"
1079    #define STR_h                       "\150"
1080    #define STR_i                       "\151"
1081    #define STR_j                       "\152"
1082    #define STR_k                       "\153"
1083    #define STR_l                       "\154"
1084    #define STR_m                       "\155"
1085    #define STR_n                       "\156"
1086    #define STR_o                       "\157"
1087    #define STR_p                       "\160"
1088    #define STR_q                       "\161"
1089    #define STR_r                       "\162"
1090    #define STR_s                       "\163"
1091    #define STR_t                       "\164"
1092    #define STR_u                       "\165"
1093    #define STR_v                       "\166"
1094    #define STR_w                       "\167"
1095    #define STR_x                       "\170"
1096    #define STR_y                       "\171"
1097    #define STR_z                       "\172"
1098    #define STR_LEFT_CURLY_BRACKET      "\173"
1099    #define STR_VERTICAL_LINE           "\174"
1100    #define STR_RIGHT_CURLY_BRACKET     "\175"
1101    #define STR_TILDE                   "\176"
1102    
1103    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1104    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1105    #define STRING_F0                   STR_F "\0"
1106    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1107    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1108    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1109    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1110    
1111    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1112    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1113    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1114    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1115    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1116    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1117    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1118    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1119    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1120    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1121    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1122    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1123    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1124    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1125    
1126    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1127    
1128    #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS
1129    #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS
1130    #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1131    #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1132    #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1133    #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
1134    #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
1135    
1136    #endif  /* SUPPORT_UTF8 */
1137    
1138  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1139    
1140  #ifndef ESC_e  #ifndef ESC_e
1141  #define ESC_e 27  #define ESC_e CHAR_ESC
1142  #endif  #endif
1143    
1144  #ifndef ESC_f  #ifndef ESC_f
1145  #define ESC_f '\f'  #define ESC_f CHAR_FF
1146  #endif  #endif
1147    
1148  #ifndef ESC_n  #ifndef ESC_n
1149  #define ESC_n '\n'  #define ESC_n CHAR_NL
1150  #endif  #endif
1151    
1152  #ifndef ESC_r  #ifndef ESC_r
1153  #define ESC_r '\r'  #define ESC_r CHAR_CR
1154  #endif  #endif
1155    
1156  /* 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
1157  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1158    
1159  #ifndef ESC_tee  #ifndef ESC_tee
1160  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1161  #endif  #endif
1162    
1163  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
# Line 595  contain UTF-8 characters with values gre Line 1184  contain UTF-8 characters with values gre
1184  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
1185  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
1186  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
1187  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
1188  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  (which is used for [^] in JavaScript compatibility mode).
1189  There are two tests in the code for an escape greater than ESC_b and less than  
1190  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
1191  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
1192  character, that code will have to change. */  greater than ESC_b and less than ESC_Z to detect the types that may be
1193    repeated. These are the types that consume characters. If any new escapes are
1194    put in between that don't consume a character, that code will have to change.
1195    */
1196    
1197  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,
1198         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,
1199         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,
1200           ESC_REF };
1201    
1202    
1203  /* 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
# Line 630  enum { Line 1223  enum {
1223    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1224    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1225    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1226    OP_ANY,            /* 12 Match any character */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
1227    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
1228    OP_NOTPROP,        /* 14 \P (not Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1229    OP_PROP,           /* 15 \p (Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1230    OP_ANYNL,          /* 16 \R (any newline sequence) */    OP_PROP,           /* 16 \p (Unicode property) */
1231    OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */    OP_ANYNL,          /* 17 \R (any newline sequence) */
1232    OP_HSPACE,         /* 18 \h (horizontal whitespace) */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1233    OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1234    OP_VSPACE,         /* 20 \v (vertical whitespace) */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1235    OP_EXTUNI,         /* 21 \X (extended Unicode sequence */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1236    OP_EODN,           /* 22 End of data or \n at end of data: \Z. */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1237    OP_EOD,            /* 23 End of data: \z */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1238      OP_EOD,            /* 24 End of data: \z */
1239    OP_OPT,            /* 24 Set runtime options */  
1240    OP_CIRC,           /* 25 Start of line - varies with multiline switch */    OP_OPT,            /* 25 Set runtime options */
1241    OP_DOLL,           /* 26 End of line - varies with multiline switch */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
1242    OP_CHAR,           /* 27 Match one character, casefully */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
1243    OP_CHARNC,         /* 28 Match one character, caselessly */    OP_CHAR,           /* 28 Match one character, casefully */
1244    OP_NOT,            /* 29 Match one character, not the following one */    OP_CHARNC,         /* 29 Match one character, caselessly */
1245      OP_NOT,            /* 30 Match one character, not the following one */
1246    OP_STAR,           /* 30 The maximizing and minimizing versions of */  
1247    OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */    OP_STAR,           /* 31 The maximizing and minimizing versions of */
1248    OP_PLUS,           /* 32 the minimizing one second. */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
1249    OP_MINPLUS,        /* 33 This first set applies to single characters.*/    OP_PLUS,           /* 33 the minimizing one second. */
1250    OP_QUERY,          /* 34 */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
1251    OP_MINQUERY,       /* 35 */    OP_QUERY,          /* 35 */
1252      OP_MINQUERY,       /* 36 */
1253    OP_UPTO,           /* 36 From 0 to n matches */  
1254    OP_MINUPTO,        /* 37 */    OP_UPTO,           /* 37 From 0 to n matches */
1255    OP_EXACT,          /* 38 Exactly n matches */    OP_MINUPTO,        /* 38 */
1256      OP_EXACT,          /* 39 Exactly n matches */
1257    OP_POSSTAR,        /* 39 Possessified star */  
1258    OP_POSPLUS,        /* 40 Possessified plus */    OP_POSSTAR,        /* 40 Possessified star */
1259    OP_POSQUERY,       /* 41 Posesssified query */    OP_POSPLUS,        /* 41 Possessified plus */
1260    OP_POSUPTO,        /* 42 Possessified upto */    OP_POSQUERY,       /* 42 Posesssified query */
1261      OP_POSUPTO,        /* 43 Possessified upto */
1262    OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */  
1263    OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
1264    OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
1265    OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
1266    OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
1267    OP_NOTMINQUERY,    /* 48 */    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
1268      OP_NOTMINQUERY,    /* 49 */
1269    OP_NOTUPTO,        /* 49 From 0 to n matches */  
1270    OP_NOTMINUPTO,     /* 50 */    OP_NOTUPTO,        /* 50 From 0 to n matches */
1271    OP_NOTEXACT,       /* 51 Exactly n matches */    OP_NOTMINUPTO,     /* 51 */
1272      OP_NOTEXACT,       /* 52 Exactly n matches */
1273    OP_NOTPOSSTAR,     /* 52 Possessified versions */  
1274    OP_NOTPOSPLUS,     /* 53 */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
1275    OP_NOTPOSQUERY,    /* 54 */    OP_NOTPOSPLUS,     /* 54 */
1276    OP_NOTPOSUPTO,     /* 55 */    OP_NOTPOSQUERY,    /* 55 */
1277      OP_NOTPOSUPTO,     /* 56 */
1278    OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */  
1279    OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
1280    OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
1281    OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
1282    OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
1283    OP_TYPEMINQUERY,   /* 61 */    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
1284      OP_TYPEMINQUERY,   /* 62 */
1285    OP_TYPEUPTO,       /* 62 From 0 to n matches */  
1286    OP_TYPEMINUPTO,    /* 63 */    OP_TYPEUPTO,       /* 63 From 0 to n matches */
1287    OP_TYPEEXACT,      /* 64 Exactly n matches */    OP_TYPEMINUPTO,    /* 64 */
1288      OP_TYPEEXACT,      /* 65 Exactly n matches */
1289    OP_TYPEPOSSTAR,    /* 65 Possessified versions */  
1290    OP_TYPEPOSPLUS,    /* 66 */    OP_TYPEPOSSTAR,    /* 66 Possessified versions */
1291    OP_TYPEPOSQUERY,   /* 67 */    OP_TYPEPOSPLUS,    /* 67 */
1292    OP_TYPEPOSUPTO,    /* 68 */    OP_TYPEPOSQUERY,   /* 68 */
1293      OP_TYPEPOSUPTO,    /* 69 */
1294    OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */  
1295    OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
1296    OP_CRPLUS,         /* 71 the minimizing one second. These codes must */    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
1297    OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
1298    OP_CRQUERY,        /* 73 These are for character classes and back refs */    OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
1299    OP_CRMINQUERY,     /* 74 */    OP_CRQUERY,        /* 74 These are for character classes and back refs */
1300    OP_CRRANGE,        /* 75 These are different to the three sets above. */    OP_CRMINQUERY,     /* 75 */
1301    OP_CRMINRANGE,     /* 76 */    OP_CRRANGE,        /* 76 These are different to the three sets above. */
1302      OP_CRMINRANGE,     /* 77 */
1303    
1304    OP_CLASS,          /* 77 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
1305    OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
1306                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
1307                             character > 255 is encountered. */                             character > 255 is encountered. */
1308    
1309    OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
1310                             class. This does both positive and negative. */                             class. This does both positive and negative. */
1311    
1312    OP_REF,            /* 80 Match a back reference */    OP_REF,            /* 81 Match a back reference */
1313    OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
1314    OP_CALLOUT,        /* 82 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
1315    
1316    OP_ALT,            /* 83 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
1317    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 */
1318    OP_KETRMAX,        /* 85 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
1319    OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
1320    
1321    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
1322    
1323    OP_ASSERT,         /* 87 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
1324    OP_ASSERT_NOT,     /* 88 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
1325    OP_ASSERTBACK,     /* 89 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
1326    OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
1327    OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
1328    
1329    /* 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,
1330    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. */
1331    
1332    OP_ONCE,           /* 92 Atomic group */    OP_ONCE,           /* 93 Atomic group */
1333    OP_BRA,            /* 93 Start of non-capturing bracket */    OP_BRA,            /* 94 Start of non-capturing bracket */
1334    OP_CBRA,           /* 94 Start of capturing bracket */    OP_CBRA,           /* 95 Start of capturing bracket */
1335    OP_COND,           /* 95 Conditional group */    OP_COND,           /* 96 Conditional group */
1336    
1337    /* 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
1338    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1339    
1340    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
1341    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1342    OP_SCOND,          /* 98 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
1343    
1344    OP_CREF,           /* 99 Used to hold a capture number as condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
1345    OP_RREF,           /* 100 Used to hold a recursion number as condition */    OP_RREF,           /* 101 Used to hold a recursion number as condition */
1346    OP_DEF,            /* 101 The DEFINE condition */    OP_DEF,            /* 102 The DEFINE condition */
1347    
1348    OP_BRAZERO,        /* 102 These two must remain together and in this */    OP_BRAZERO,        /* 103 These two must remain together and in this */
1349    OP_BRAMINZERO,     /* 103 order. */    OP_BRAMINZERO,     /* 104 order. */
1350    
1351    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1352    
1353    OP_PRUNE,          /* 104 */    OP_PRUNE,          /* 105 */
1354    OP_SKIP,           /* 105 */    OP_SKIP,           /* 106 */
1355    OP_THEN,           /* 106 */    OP_THEN,           /* 107 */
1356    OP_COMMIT,         /* 107 */    OP_COMMIT,         /* 108 */
1357    
1358    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1359    
1360    OP_FAIL,           /* 108 */    OP_FAIL,           /* 109 */
1361    OP_ACCEPT          /* 109 */    OP_ACCEPT,         /* 110 */
1362    
1363      /* This is used to skip a subpattern with a {0} quantifier */
1364    
1365      OP_SKIPZERO        /* 111 */
1366  };  };
1367    
1368    
# Line 773  for debugging. The macro is referenced o Line 1371  for debugging. The macro is referenced o
1371    
1372  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1373    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1374    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1375    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1376    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1377    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
# Line 789  for debugging. The macro is referenced o Line 1387  for debugging. The macro is referenced o
1387    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1388    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1389    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
1390    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT"    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
1391      "Skip zero"
1392    
1393    
1394  /* 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 805  in UTF-8 mode. The code that uses this t Line 1404  in UTF-8 mode. The code that uses this t
1404    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1405    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1406    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1407    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1408    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
1409    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1410    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
# Line 854  in UTF-8 mode. The code that uses this t Line 1453  in UTF-8 mode. The code that uses this t
1453    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1454    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1455    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
1456    1, 1                           /* FAIL, ACCEPT                           */    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
1457    
1458    
1459  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
# Line 870  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 1469  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
1469         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1470         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1471         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1472         ERR60, ERR61 };         ERR60, ERR61, ERR62, ERR63, ERR64 };
1473    
1474  /* 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
1475  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 933  typedef struct compile_data { Line 1532  typedef struct compile_data {
1532    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1533    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
1534    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
1535    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
1536      int  final_bracount;          /* Saved value after first pass */
1537    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
1538    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
1539    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
1540    int  external_flags;          /* External flag bits to be set */    int  external_flags;          /* External flag bits to be set */
1541    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1542    BOOL had_accept;              /* (*ACCEPT) encountered */    BOOL had_accept;              /* (*ACCEPT) encountered */
1543    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
# Line 994  typedef struct match_data { Line 1594  typedef struct match_data {
1594    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1595    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1596    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1597      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1598    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1599    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1600    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
1601    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1602      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1603    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1604    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1605    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
# Line 1034  typedef struct dfa_match_data { Line 1636  typedef struct dfa_match_data {
1636  #define ctype_letter  0x02  #define ctype_letter  0x02
1637  #define ctype_digit   0x04  #define ctype_digit   0x04
1638  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1639  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1640  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1641    
1642  /* 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 1062  total length. */ Line 1664  total length. */
1664  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1665    
1666  /* 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
1667  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1668    relocations in shared libraries, it now has an offset into a single string
1669    instead. */
1670    
1671  typedef struct {  typedef struct {
1672    const char *name;    pcre_uint16 name_offset;
1673    pcre_uint16 type;    pcre_uint16 type;
1674    pcre_uint16 value;    pcre_uint16 value;
1675  } ucp_type_table;  } ucp_type_table;
# Line 1083  extern const uschar _pcre_utf8_table4[]; Line 1687  extern const uschar _pcre_utf8_table4[];
1687    
1688  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1689    
1690    extern const char   _pcre_utt_names[];
1691  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1692  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1693    
# Line 1100  extern BOOL _pcre_is_newline(con Line 1705  extern BOOL _pcre_is_newline(con
1705  extern int          _pcre_ord2utf8(int, uschar *);  extern int          _pcre_ord2utf8(int, uschar *);
1706  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1707                        const pcre_study_data *, pcre_study_data *);                        const pcre_study_data *, pcre_study_data *);
 extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  
 extern unsigned int _pcre_ucp_othercase(const unsigned int);  
1708  extern int          _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_valid_utf8(const uschar *, int);
1709  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1710                        int *, BOOL);                        int *, BOOL);
1711  extern BOOL         _pcre_xclass(int, const uschar *);  extern BOOL         _pcre_xclass(int, const uschar *);
1712    
1713    
1714    /* Unicode character database (UCD) */
1715    
1716    typedef struct {
1717      uschar script;
1718      uschar chartype;
1719      pcre_int32 other_case;
1720    } ucd_record;
1721    
1722    extern const ucd_record  _pcre_ucd_records[];
1723    extern const uschar      _pcre_ucd_stage1[];
1724    extern const pcre_uint16 _pcre_ucd_stage2[];
1725    extern const int         _pcre_ucp_gentype[];
1726    
1727    
1728    /* UCD access macros */
1729    
1730    #define UCD_BLOCK_SIZE 128
1731    #define GET_UCD(ch) (_pcre_ucd_records + \
1732            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1733            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1734    
1735    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1736    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1737    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1738    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1739    
1740  #endif  #endif
1741    
1742  /* End of pcre_internal.h */  /* End of pcre_internal.h */

Legend:
Removed from v.230  
changed lines
  Added in v.393

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12