/[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 270 by ph10, Fri Nov 16 19:55:16 2007 UTC revision 455 by ph10, Sat Sep 26 19:12:32 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 483  Standard C system should have one. */ Line 535  Standard C system should have one. */
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. From release 8.00, PCRE_NOPARTIAL is unused, as
539    the restrictions on partial matching have been lifted. It remains for backwards
540    compatibility. */
541    
542  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
543  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
# Line 495  are in a 16-bit flags word. */ Line 549  are in a 16-bit flags word. */
549  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
550    
551  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
552    #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */
553    
554  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
555  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 502  time, run time, or study time, respectiv Line 557  time, run time, or study time, respectiv
557  #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| \
558                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
559    
560  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
561    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
562     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
563     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| \
564     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
565       PCRE_JAVASCRIPT_COMPAT)
566    
567  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
568    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
569     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
570       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
571    
572  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
573    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
574     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
575     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
576       PCRE_NO_START_OPTIMIZE)
577    
578  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
579    
# Line 540  variable-length repeat, or a anything ot Line 598  variable-length repeat, or a anything ot
598  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
599  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
600    
601  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
602    environments where these macros are defined elsewhere. Unfortunately, there
603    is no way to do the same for the typedef. */
604    
605  typedef int BOOL;  typedef int BOOL;
606    
607    #ifndef FALSE
608  #define FALSE   0  #define FALSE   0
609  #define TRUE    1  #define TRUE    1
610    #endif
611    
612    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
613    character constants like '*' because the compiler would emit their EBCDIC code,
614    which is different from their ASCII/UTF-8 code. Instead we define macros for
615    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
616    is enabled. When UTF-8 support is not enabled, the definitions use character
617    literals. Both character and string versions of each character are needed, and
618    there are some longer strings as well.
619    
620    This means that, on EBCDIC platforms, the PCRE library can handle either
621    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
622    would need different lookups depending on whether PCRE_UTF8 was set or not.
623    This would make it impossible to use characters in switch/case statements,
624    which would reduce performance. For a theoretical use (which nobody has asked
625    for) in a minority area (EBCDIC platforms), this is not sensible. Any
626    application that did need both could compile two versions of the library, using
627    macros to give the functions distinct names. */
628    
629    #ifndef SUPPORT_UTF8
630    
631    /* UTF-8 support is not enabled; use the platform-dependent character literals
632    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
633    
634    #define CHAR_HT                     '\t'
635    #define CHAR_VT                     '\v'
636    #define CHAR_FF                     '\f'
637    #define CHAR_CR                     '\r'
638    #define CHAR_NL                     '\n'
639    #define CHAR_BS                     '\b'
640    #define CHAR_BEL                    '\a'
641    #ifdef EBCDIC
642    #define CHAR_ESC                    '\047'
643    #define CHAR_DEL                    '\007'
644    #else
645    #define CHAR_ESC                    '\033'
646    #define CHAR_DEL                    '\177'
647    #endif
648    
649    #define CHAR_SPACE                  ' '
650    #define CHAR_EXCLAMATION_MARK       '!'
651    #define CHAR_QUOTATION_MARK         '"'
652    #define CHAR_NUMBER_SIGN            '#'
653    #define CHAR_DOLLAR_SIGN            '$'
654    #define CHAR_PERCENT_SIGN           '%'
655    #define CHAR_AMPERSAND              '&'
656    #define CHAR_APOSTROPHE             '\''
657    #define CHAR_LEFT_PARENTHESIS       '('
658    #define CHAR_RIGHT_PARENTHESIS      ')'
659    #define CHAR_ASTERISK               '*'
660    #define CHAR_PLUS                   '+'
661    #define CHAR_COMMA                  ','
662    #define CHAR_MINUS                  '-'
663    #define CHAR_DOT                    '.'
664    #define CHAR_SLASH                  '/'
665    #define CHAR_0                      '0'
666    #define CHAR_1                      '1'
667    #define CHAR_2                      '2'
668    #define CHAR_3                      '3'
669    #define CHAR_4                      '4'
670    #define CHAR_5                      '5'
671    #define CHAR_6                      '6'
672    #define CHAR_7                      '7'
673    #define CHAR_8                      '8'
674    #define CHAR_9                      '9'
675    #define CHAR_COLON                  ':'
676    #define CHAR_SEMICOLON              ';'
677    #define CHAR_LESS_THAN_SIGN         '<'
678    #define CHAR_EQUALS_SIGN            '='
679    #define CHAR_GREATER_THAN_SIGN      '>'
680    #define CHAR_QUESTION_MARK          '?'
681    #define CHAR_COMMERCIAL_AT          '@'
682    #define CHAR_A                      'A'
683    #define CHAR_B                      'B'
684    #define CHAR_C                      'C'
685    #define CHAR_D                      'D'
686    #define CHAR_E                      'E'
687    #define CHAR_F                      'F'
688    #define CHAR_G                      'G'
689    #define CHAR_H                      'H'
690    #define CHAR_I                      'I'
691    #define CHAR_J                      'J'
692    #define CHAR_K                      'K'
693    #define CHAR_L                      'L'
694    #define CHAR_M                      'M'
695    #define CHAR_N                      'N'
696    #define CHAR_O                      'O'
697    #define CHAR_P                      'P'
698    #define CHAR_Q                      'Q'
699    #define CHAR_R                      'R'
700    #define CHAR_S                      'S'
701    #define CHAR_T                      'T'
702    #define CHAR_U                      'U'
703    #define CHAR_V                      'V'
704    #define CHAR_W                      'W'
705    #define CHAR_X                      'X'
706    #define CHAR_Y                      'Y'
707    #define CHAR_Z                      'Z'
708    #define CHAR_LEFT_SQUARE_BRACKET    '['
709    #define CHAR_BACKSLASH              '\\'
710    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
711    #define CHAR_CIRCUMFLEX_ACCENT      '^'
712    #define CHAR_UNDERSCORE             '_'
713    #define CHAR_GRAVE_ACCENT           '`'
714    #define CHAR_a                      'a'
715    #define CHAR_b                      'b'
716    #define CHAR_c                      'c'
717    #define CHAR_d                      'd'
718    #define CHAR_e                      'e'
719    #define CHAR_f                      'f'
720    #define CHAR_g                      'g'
721    #define CHAR_h                      'h'
722    #define CHAR_i                      'i'
723    #define CHAR_j                      'j'
724    #define CHAR_k                      'k'
725    #define CHAR_l                      'l'
726    #define CHAR_m                      'm'
727    #define CHAR_n                      'n'
728    #define CHAR_o                      'o'
729    #define CHAR_p                      'p'
730    #define CHAR_q                      'q'
731    #define CHAR_r                      'r'
732    #define CHAR_s                      's'
733    #define CHAR_t                      't'
734    #define CHAR_u                      'u'
735    #define CHAR_v                      'v'
736    #define CHAR_w                      'w'
737    #define CHAR_x                      'x'
738    #define CHAR_y                      'y'
739    #define CHAR_z                      'z'
740    #define CHAR_LEFT_CURLY_BRACKET     '{'
741    #define CHAR_VERTICAL_LINE          '|'
742    #define CHAR_RIGHT_CURLY_BRACKET    '}'
743    #define CHAR_TILDE                  '~'
744    
745    #define STR_HT                      "\t"
746    #define STR_VT                      "\v"
747    #define STR_FF                      "\f"
748    #define STR_CR                      "\r"
749    #define STR_NL                      "\n"
750    #define STR_BS                      "\b"
751    #define STR_BEL                     "\a"
752    #ifdef EBCDIC
753    #define STR_ESC                     "\047"
754    #define STR_DEL                     "\007"
755    #else
756    #define STR_ESC                     "\033"
757    #define STR_DEL                     "\177"
758    #endif
759    
760    #define STR_SPACE                   " "
761    #define STR_EXCLAMATION_MARK        "!"
762    #define STR_QUOTATION_MARK          "\""
763    #define STR_NUMBER_SIGN             "#"
764    #define STR_DOLLAR_SIGN             "$"
765    #define STR_PERCENT_SIGN            "%"
766    #define STR_AMPERSAND               "&"
767    #define STR_APOSTROPHE              "'"
768    #define STR_LEFT_PARENTHESIS        "("
769    #define STR_RIGHT_PARENTHESIS       ")"
770    #define STR_ASTERISK                "*"
771    #define STR_PLUS                    "+"
772    #define STR_COMMA                   ","
773    #define STR_MINUS                   "-"
774    #define STR_DOT                     "."
775    #define STR_SLASH                   "/"
776    #define STR_0                       "0"
777    #define STR_1                       "1"
778    #define STR_2                       "2"
779    #define STR_3                       "3"
780    #define STR_4                       "4"
781    #define STR_5                       "5"
782    #define STR_6                       "6"
783    #define STR_7                       "7"
784    #define STR_8                       "8"
785    #define STR_9                       "9"
786    #define STR_COLON                   ":"
787    #define STR_SEMICOLON               ";"
788    #define STR_LESS_THAN_SIGN          "<"
789    #define STR_EQUALS_SIGN             "="
790    #define STR_GREATER_THAN_SIGN       ">"
791    #define STR_QUESTION_MARK           "?"
792    #define STR_COMMERCIAL_AT           "@"
793    #define STR_A                       "A"
794    #define STR_B                       "B"
795    #define STR_C                       "C"
796    #define STR_D                       "D"
797    #define STR_E                       "E"
798    #define STR_F                       "F"
799    #define STR_G                       "G"
800    #define STR_H                       "H"
801    #define STR_I                       "I"
802    #define STR_J                       "J"
803    #define STR_K                       "K"
804    #define STR_L                       "L"
805    #define STR_M                       "M"
806    #define STR_N                       "N"
807    #define STR_O                       "O"
808    #define STR_P                       "P"
809    #define STR_Q                       "Q"
810    #define STR_R                       "R"
811    #define STR_S                       "S"
812    #define STR_T                       "T"
813    #define STR_U                       "U"
814    #define STR_V                       "V"
815    #define STR_W                       "W"
816    #define STR_X                       "X"
817    #define STR_Y                       "Y"
818    #define STR_Z                       "Z"
819    #define STR_LEFT_SQUARE_BRACKET     "["
820    #define STR_BACKSLASH               "\\"
821    #define STR_RIGHT_SQUARE_BRACKET    "]"
822    #define STR_CIRCUMFLEX_ACCENT       "^"
823    #define STR_UNDERSCORE              "_"
824    #define STR_GRAVE_ACCENT            "`"
825    #define STR_a                       "a"
826    #define STR_b                       "b"
827    #define STR_c                       "c"
828    #define STR_d                       "d"
829    #define STR_e                       "e"
830    #define STR_f                       "f"
831    #define STR_g                       "g"
832    #define STR_h                       "h"
833    #define STR_i                       "i"
834    #define STR_j                       "j"
835    #define STR_k                       "k"
836    #define STR_l                       "l"
837    #define STR_m                       "m"
838    #define STR_n                       "n"
839    #define STR_o                       "o"
840    #define STR_p                       "p"
841    #define STR_q                       "q"
842    #define STR_r                       "r"
843    #define STR_s                       "s"
844    #define STR_t                       "t"
845    #define STR_u                       "u"
846    #define STR_v                       "v"
847    #define STR_w                       "w"
848    #define STR_x                       "x"
849    #define STR_y                       "y"
850    #define STR_z                       "z"
851    #define STR_LEFT_CURLY_BRACKET      "{"
852    #define STR_VERTICAL_LINE           "|"
853    #define STR_RIGHT_CURLY_BRACKET     "}"
854    #define STR_TILDE                   "~"
855    
856    #define STRING_ACCEPT0              "ACCEPT\0"
857    #define STRING_COMMIT0              "COMMIT\0"
858    #define STRING_F0                   "F\0"
859    #define STRING_FAIL0                "FAIL\0"
860    #define STRING_PRUNE0               "PRUNE\0"
861    #define STRING_SKIP0                "SKIP\0"
862    #define STRING_THEN                 "THEN"
863    
864    #define STRING_alpha0               "alpha\0"
865    #define STRING_lower0               "lower\0"
866    #define STRING_upper0               "upper\0"
867    #define STRING_alnum0               "alnum\0"
868    #define STRING_ascii0               "ascii\0"
869    #define STRING_blank0               "blank\0"
870    #define STRING_cntrl0               "cntrl\0"
871    #define STRING_digit0               "digit\0"
872    #define STRING_graph0               "graph\0"
873    #define STRING_print0               "print\0"
874    #define STRING_punct0               "punct\0"
875    #define STRING_space0               "space\0"
876    #define STRING_word0                "word\0"
877    #define STRING_xdigit               "xdigit"
878    
879    #define STRING_DEFINE               "DEFINE"
880    
881    #define STRING_CR_RIGHTPAR          "CR)"
882    #define STRING_LF_RIGHTPAR          "LF)"
883    #define STRING_CRLF_RIGHTPAR        "CRLF)"
884    #define STRING_ANY_RIGHTPAR         "ANY)"
885    #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"
886    #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
887    #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
888    #define STRING_UTF8_RIGHTPAR        "UTF8)"
889    
890    #else  /* SUPPORT_UTF8 */
891    
892    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
893    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
894    only. */
895    
896    #define CHAR_HT                     '\011'
897    #define CHAR_VT                     '\013'
898    #define CHAR_FF                     '\014'
899    #define CHAR_CR                     '\015'
900    #define CHAR_NL                     '\012'
901    #define CHAR_BS                     '\010'
902    #define CHAR_BEL                    '\007'
903    #define CHAR_ESC                    '\033'
904    #define CHAR_DEL                    '\177'
905    
906    #define CHAR_SPACE                  '\040'
907    #define CHAR_EXCLAMATION_MARK       '\041'
908    #define CHAR_QUOTATION_MARK         '\042'
909    #define CHAR_NUMBER_SIGN            '\043'
910    #define CHAR_DOLLAR_SIGN            '\044'
911    #define CHAR_PERCENT_SIGN           '\045'
912    #define CHAR_AMPERSAND              '\046'
913    #define CHAR_APOSTROPHE             '\047'
914    #define CHAR_LEFT_PARENTHESIS       '\050'
915    #define CHAR_RIGHT_PARENTHESIS      '\051'
916    #define CHAR_ASTERISK               '\052'
917    #define CHAR_PLUS                   '\053'
918    #define CHAR_COMMA                  '\054'
919    #define CHAR_MINUS                  '\055'
920    #define CHAR_DOT                    '\056'
921    #define CHAR_SLASH                  '\057'
922    #define CHAR_0                      '\060'
923    #define CHAR_1                      '\061'
924    #define CHAR_2                      '\062'
925    #define CHAR_3                      '\063'
926    #define CHAR_4                      '\064'
927    #define CHAR_5                      '\065'
928    #define CHAR_6                      '\066'
929    #define CHAR_7                      '\067'
930    #define CHAR_8                      '\070'
931    #define CHAR_9                      '\071'
932    #define CHAR_COLON                  '\072'
933    #define CHAR_SEMICOLON              '\073'
934    #define CHAR_LESS_THAN_SIGN         '\074'
935    #define CHAR_EQUALS_SIGN            '\075'
936    #define CHAR_GREATER_THAN_SIGN      '\076'
937    #define CHAR_QUESTION_MARK          '\077'
938    #define CHAR_COMMERCIAL_AT          '\100'
939    #define CHAR_A                      '\101'
940    #define CHAR_B                      '\102'
941    #define CHAR_C                      '\103'
942    #define CHAR_D                      '\104'
943    #define CHAR_E                      '\105'
944    #define CHAR_F                      '\106'
945    #define CHAR_G                      '\107'
946    #define CHAR_H                      '\110'
947    #define CHAR_I                      '\111'
948    #define CHAR_J                      '\112'
949    #define CHAR_K                      '\113'
950    #define CHAR_L                      '\114'
951    #define CHAR_M                      '\115'
952    #define CHAR_N                      '\116'
953    #define CHAR_O                      '\117'
954    #define CHAR_P                      '\120'
955    #define CHAR_Q                      '\121'
956    #define CHAR_R                      '\122'
957    #define CHAR_S                      '\123'
958    #define CHAR_T                      '\124'
959    #define CHAR_U                      '\125'
960    #define CHAR_V                      '\126'
961    #define CHAR_W                      '\127'
962    #define CHAR_X                      '\130'
963    #define CHAR_Y                      '\131'
964    #define CHAR_Z                      '\132'
965    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
966    #define CHAR_BACKSLASH              '\134'
967    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
968    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
969    #define CHAR_UNDERSCORE             '\137'
970    #define CHAR_GRAVE_ACCENT           '\140'
971    #define CHAR_a                      '\141'
972    #define CHAR_b                      '\142'
973    #define CHAR_c                      '\143'
974    #define CHAR_d                      '\144'
975    #define CHAR_e                      '\145'
976    #define CHAR_f                      '\146'
977    #define CHAR_g                      '\147'
978    #define CHAR_h                      '\150'
979    #define CHAR_i                      '\151'
980    #define CHAR_j                      '\152'
981    #define CHAR_k                      '\153'
982    #define CHAR_l                      '\154'
983    #define CHAR_m                      '\155'
984    #define CHAR_n                      '\156'
985    #define CHAR_o                      '\157'
986    #define CHAR_p                      '\160'
987    #define CHAR_q                      '\161'
988    #define CHAR_r                      '\162'
989    #define CHAR_s                      '\163'
990    #define CHAR_t                      '\164'
991    #define CHAR_u                      '\165'
992    #define CHAR_v                      '\166'
993    #define CHAR_w                      '\167'
994    #define CHAR_x                      '\170'
995    #define CHAR_y                      '\171'
996    #define CHAR_z                      '\172'
997    #define CHAR_LEFT_CURLY_BRACKET     '\173'
998    #define CHAR_VERTICAL_LINE          '\174'
999    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1000    #define CHAR_TILDE                  '\176'
1001    
1002    #define STR_HT                      "\011"
1003    #define STR_VT                      "\013"
1004    #define STR_FF                      "\014"
1005    #define STR_CR                      "\015"
1006    #define STR_NL                      "\012"
1007    #define STR_BS                      "\010"
1008    #define STR_BEL                     "\007"
1009    #define STR_ESC                     "\033"
1010    #define STR_DEL                     "\177"
1011    
1012    #define STR_SPACE                   "\040"
1013    #define STR_EXCLAMATION_MARK        "\041"
1014    #define STR_QUOTATION_MARK          "\042"
1015    #define STR_NUMBER_SIGN             "\043"
1016    #define STR_DOLLAR_SIGN             "\044"
1017    #define STR_PERCENT_SIGN            "\045"
1018    #define STR_AMPERSAND               "\046"
1019    #define STR_APOSTROPHE              "\047"
1020    #define STR_LEFT_PARENTHESIS        "\050"
1021    #define STR_RIGHT_PARENTHESIS       "\051"
1022    #define STR_ASTERISK                "\052"
1023    #define STR_PLUS                    "\053"
1024    #define STR_COMMA                   "\054"
1025    #define STR_MINUS                   "\055"
1026    #define STR_DOT                     "\056"
1027    #define STR_SLASH                   "\057"
1028    #define STR_0                       "\060"
1029    #define STR_1                       "\061"
1030    #define STR_2                       "\062"
1031    #define STR_3                       "\063"
1032    #define STR_4                       "\064"
1033    #define STR_5                       "\065"
1034    #define STR_6                       "\066"
1035    #define STR_7                       "\067"
1036    #define STR_8                       "\070"
1037    #define STR_9                       "\071"
1038    #define STR_COLON                   "\072"
1039    #define STR_SEMICOLON               "\073"
1040    #define STR_LESS_THAN_SIGN          "\074"
1041    #define STR_EQUALS_SIGN             "\075"
1042    #define STR_GREATER_THAN_SIGN       "\076"
1043    #define STR_QUESTION_MARK           "\077"
1044    #define STR_COMMERCIAL_AT           "\100"
1045    #define STR_A                       "\101"
1046    #define STR_B                       "\102"
1047    #define STR_C                       "\103"
1048    #define STR_D                       "\104"
1049    #define STR_E                       "\105"
1050    #define STR_F                       "\106"
1051    #define STR_G                       "\107"
1052    #define STR_H                       "\110"
1053    #define STR_I                       "\111"
1054    #define STR_J                       "\112"
1055    #define STR_K                       "\113"
1056    #define STR_L                       "\114"
1057    #define STR_M                       "\115"
1058    #define STR_N                       "\116"
1059    #define STR_O                       "\117"
1060    #define STR_P                       "\120"
1061    #define STR_Q                       "\121"
1062    #define STR_R                       "\122"
1063    #define STR_S                       "\123"
1064    #define STR_T                       "\124"
1065    #define STR_U                       "\125"
1066    #define STR_V                       "\126"
1067    #define STR_W                       "\127"
1068    #define STR_X                       "\130"
1069    #define STR_Y                       "\131"
1070    #define STR_Z                       "\132"
1071    #define STR_LEFT_SQUARE_BRACKET     "\133"
1072    #define STR_BACKSLASH               "\134"
1073    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1074    #define STR_CIRCUMFLEX_ACCENT       "\136"
1075    #define STR_UNDERSCORE              "\137"
1076    #define STR_GRAVE_ACCENT            "\140"
1077    #define STR_a                       "\141"
1078    #define STR_b                       "\142"
1079    #define STR_c                       "\143"
1080    #define STR_d                       "\144"
1081    #define STR_e                       "\145"
1082    #define STR_f                       "\146"
1083    #define STR_g                       "\147"
1084    #define STR_h                       "\150"
1085    #define STR_i                       "\151"
1086    #define STR_j                       "\152"
1087    #define STR_k                       "\153"
1088    #define STR_l                       "\154"
1089    #define STR_m                       "\155"
1090    #define STR_n                       "\156"
1091    #define STR_o                       "\157"
1092    #define STR_p                       "\160"
1093    #define STR_q                       "\161"
1094    #define STR_r                       "\162"
1095    #define STR_s                       "\163"
1096    #define STR_t                       "\164"
1097    #define STR_u                       "\165"
1098    #define STR_v                       "\166"
1099    #define STR_w                       "\167"
1100    #define STR_x                       "\170"
1101    #define STR_y                       "\171"
1102    #define STR_z                       "\172"
1103    #define STR_LEFT_CURLY_BRACKET      "\173"
1104    #define STR_VERTICAL_LINE           "\174"
1105    #define STR_RIGHT_CURLY_BRACKET     "\175"
1106    #define STR_TILDE                   "\176"
1107    
1108    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1109    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1110    #define STRING_F0                   STR_F "\0"
1111    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1112    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1113    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1114    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1115    
1116    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1117    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1118    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1119    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1120    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1121    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1122    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1123    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1124    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1125    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1126    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1127    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1128    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1129    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1130    
1131    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1132    
1133    #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS
1134    #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS
1135    #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1136    #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1137    #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1138    #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
1139    #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
1140    #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1141    
1142    #endif  /* SUPPORT_UTF8 */
1143    
1144  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1145    
1146  #ifndef ESC_e  #ifndef ESC_e
1147  #define ESC_e 27  #define ESC_e CHAR_ESC
1148  #endif  #endif
1149    
1150  #ifndef ESC_f  #ifndef ESC_f
1151  #define ESC_f '\f'  #define ESC_f CHAR_FF
1152  #endif  #endif
1153    
1154  #ifndef ESC_n  #ifndef ESC_n
1155  #define ESC_n '\n'  #define ESC_n CHAR_NL
1156  #endif  #endif
1157    
1158  #ifndef ESC_r  #ifndef ESC_r
1159  #define ESC_r '\r'  #define ESC_r CHAR_CR
1160  #endif  #endif
1161    
1162  /* 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
1163  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1164    
1165  #ifndef ESC_tee  #ifndef ESC_tee
1166  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1167  #endif  #endif
1168    
1169  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
# Line 596  contain UTF-8 characters with values gre Line 1190  contain UTF-8 characters with values gre
1190  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
1191  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
1192  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
1193  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
1194  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  (which is used for [^] in JavaScript compatibility mode).
1195  There are two tests in the code for an escape greater than ESC_b and less than  
1196  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
1197  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
1198  character, that code will have to change. */  greater than ESC_b and less than ESC_Z to detect the types that may be
1199    repeated. These are the types that consume characters. If any new escapes are
1200    put in between that don't consume a character, that code will have to change.
1201    */
1202    
1203  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,
1204         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,
1205         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,
1206           ESC_REF };
1207    
1208    
1209  /* 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 631  enum { Line 1229  enum {
1229    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1230    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1231    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1232    OP_ANY,            /* 12 Match any character */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
1233    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
1234    OP_NOTPROP,        /* 14 \P (not Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1235    OP_PROP,           /* 15 \p (Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1236    OP_ANYNL,          /* 16 \R (any newline sequence) */    OP_PROP,           /* 16 \p (Unicode property) */
1237    OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */    OP_ANYNL,          /* 17 \R (any newline sequence) */
1238    OP_HSPACE,         /* 18 \h (horizontal whitespace) */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1239    OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1240    OP_VSPACE,         /* 20 \v (vertical whitespace) */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1241    OP_EXTUNI,         /* 21 \X (extended Unicode sequence */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1242    OP_EODN,           /* 22 End of data or \n at end of data: \Z. */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1243    OP_EOD,            /* 23 End of data: \z */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1244      OP_EOD,            /* 24 End of data: \z */
1245    OP_OPT,            /* 24 Set runtime options */  
1246    OP_CIRC,           /* 25 Start of line - varies with multiline switch */    OP_OPT,            /* 25 Set runtime options */
1247    OP_DOLL,           /* 26 End of line - varies with multiline switch */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
1248    OP_CHAR,           /* 27 Match one character, casefully */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
1249    OP_CHARNC,         /* 28 Match one character, caselessly */    OP_CHAR,           /* 28 Match one character, casefully */
1250    OP_NOT,            /* 29 Match one character, not the following one */    OP_CHARNC,         /* 29 Match one character, caselessly */
1251      OP_NOT,            /* 30 Match one character, not the following one */
1252    OP_STAR,           /* 30 The maximizing and minimizing versions of */  
1253    OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */    OP_STAR,           /* 31 The maximizing and minimizing versions of */
1254    OP_PLUS,           /* 32 the minimizing one second. */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
1255    OP_MINPLUS,        /* 33 This first set applies to single characters.*/    OP_PLUS,           /* 33 the minimizing one second. */
1256    OP_QUERY,          /* 34 */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
1257    OP_MINQUERY,       /* 35 */    OP_QUERY,          /* 35 */
1258      OP_MINQUERY,       /* 36 */
1259    OP_UPTO,           /* 36 From 0 to n matches */  
1260    OP_MINUPTO,        /* 37 */    OP_UPTO,           /* 37 From 0 to n matches */
1261    OP_EXACT,          /* 38 Exactly n matches */    OP_MINUPTO,        /* 38 */
1262      OP_EXACT,          /* 39 Exactly n matches */
1263    OP_POSSTAR,        /* 39 Possessified star */  
1264    OP_POSPLUS,        /* 40 Possessified plus */    OP_POSSTAR,        /* 40 Possessified star */
1265    OP_POSQUERY,       /* 41 Posesssified query */    OP_POSPLUS,        /* 41 Possessified plus */
1266    OP_POSUPTO,        /* 42 Possessified upto */    OP_POSQUERY,       /* 42 Posesssified query */
1267      OP_POSUPTO,        /* 43 Possessified upto */
1268    OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */  
1269    OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
1270    OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
1271    OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
1272    OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
1273    OP_NOTMINQUERY,    /* 48 */    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
1274      OP_NOTMINQUERY,    /* 49 */
1275    OP_NOTUPTO,        /* 49 From 0 to n matches */  
1276    OP_NOTMINUPTO,     /* 50 */    OP_NOTUPTO,        /* 50 From 0 to n matches */
1277    OP_NOTEXACT,       /* 51 Exactly n matches */    OP_NOTMINUPTO,     /* 51 */
1278      OP_NOTEXACT,       /* 52 Exactly n matches */
1279    OP_NOTPOSSTAR,     /* 52 Possessified versions */  
1280    OP_NOTPOSPLUS,     /* 53 */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
1281    OP_NOTPOSQUERY,    /* 54 */    OP_NOTPOSPLUS,     /* 54 */
1282    OP_NOTPOSUPTO,     /* 55 */    OP_NOTPOSQUERY,    /* 55 */
1283      OP_NOTPOSUPTO,     /* 56 */
1284    OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */  
1285    OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
1286    OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
1287    OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
1288    OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
1289    OP_TYPEMINQUERY,   /* 61 */    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
1290      OP_TYPEMINQUERY,   /* 62 */
1291    OP_TYPEUPTO,       /* 62 From 0 to n matches */  
1292    OP_TYPEMINUPTO,    /* 63 */    OP_TYPEUPTO,       /* 63 From 0 to n matches */
1293    OP_TYPEEXACT,      /* 64 Exactly n matches */    OP_TYPEMINUPTO,    /* 64 */
1294      OP_TYPEEXACT,      /* 65 Exactly n matches */
1295    OP_TYPEPOSSTAR,    /* 65 Possessified versions */  
1296    OP_TYPEPOSPLUS,    /* 66 */    OP_TYPEPOSSTAR,    /* 66 Possessified versions */
1297    OP_TYPEPOSQUERY,   /* 67 */    OP_TYPEPOSPLUS,    /* 67 */
1298    OP_TYPEPOSUPTO,    /* 68 */    OP_TYPEPOSQUERY,   /* 68 */
1299      OP_TYPEPOSUPTO,    /* 69 */
1300    OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */  
1301    OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
1302    OP_CRPLUS,         /* 71 the minimizing one second. These codes must */    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
1303    OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
1304    OP_CRQUERY,        /* 73 These are for character classes and back refs */    OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
1305    OP_CRMINQUERY,     /* 74 */    OP_CRQUERY,        /* 74 These are for character classes and back refs */
1306    OP_CRRANGE,        /* 75 These are different to the three sets above. */    OP_CRMINQUERY,     /* 75 */
1307    OP_CRMINRANGE,     /* 76 */    OP_CRRANGE,        /* 76 These are different to the three sets above. */
1308      OP_CRMINRANGE,     /* 77 */
1309    
1310    OP_CLASS,          /* 77 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
1311    OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
1312                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
1313                             character > 255 is encountered. */                             character > 255 is encountered. */
1314    
1315    OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
1316                             class. This does both positive and negative. */                             class. This does both positive and negative. */
1317    
1318    OP_REF,            /* 80 Match a back reference */    OP_REF,            /* 81 Match a back reference */
1319    OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
1320    OP_CALLOUT,        /* 82 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
1321    
1322    OP_ALT,            /* 83 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
1323    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 */
1324    OP_KETRMAX,        /* 85 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
1325    OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
1326    
1327    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
1328    
1329    OP_ASSERT,         /* 87 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
1330    OP_ASSERT_NOT,     /* 88 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
1331    OP_ASSERTBACK,     /* 89 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
1332    OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
1333    OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
1334    
1335    /* 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,
1336    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. */
1337    
1338    OP_ONCE,           /* 92 Atomic group */    OP_ONCE,           /* 93 Atomic group */
1339    OP_BRA,            /* 93 Start of non-capturing bracket */    OP_BRA,            /* 94 Start of non-capturing bracket */
1340    OP_CBRA,           /* 94 Start of capturing bracket */    OP_CBRA,           /* 95 Start of capturing bracket */
1341    OP_COND,           /* 95 Conditional group */    OP_COND,           /* 96 Conditional group */
1342    
1343    /* 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
1344    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1345    
1346    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
1347    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1348    OP_SCOND,          /* 98 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
1349    
1350    OP_CREF,           /* 99 Used to hold a capture number as condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
1351    OP_RREF,           /* 100 Used to hold a recursion number as condition */    OP_RREF,           /* 101 Used to hold a recursion number as condition */
1352    OP_DEF,            /* 101 The DEFINE condition */    OP_DEF,            /* 102 The DEFINE condition */
1353    
1354    OP_BRAZERO,        /* 102 These two must remain together and in this */    OP_BRAZERO,        /* 103 These two must remain together and in this */
1355    OP_BRAMINZERO,     /* 103 order. */    OP_BRAMINZERO,     /* 104 order. */
1356    
1357    /* These are backtracking control verbs */    /* These are backtracking control verbs */
1358    
1359    OP_PRUNE,          /* 104 */    OP_PRUNE,          /* 105 */
1360    OP_SKIP,           /* 105 */    OP_SKIP,           /* 106 */
1361    OP_THEN,           /* 106 */    OP_THEN,           /* 107 */
1362    OP_COMMIT,         /* 107 */    OP_COMMIT,         /* 108 */
1363    
1364    /* These are forced failure and success verbs */    /* These are forced failure and success verbs */
1365    
1366    OP_FAIL,           /* 108 */    OP_FAIL,           /* 109 */
1367    OP_ACCEPT          /* 109 */    OP_ACCEPT,         /* 110 */
1368      OP_CLOSE,          /* 111 Used before OP_ACCEPT to close open captures */
1369    
1370      /* This is used to skip a subpattern with a {0} quantifier */
1371    
1372      OP_SKIPZERO        /* 112 */
1373  };  };
1374    
1375    
# Line 774  for debugging. The macro is referenced o Line 1378  for debugging. The macro is referenced o
1378    
1379  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1380    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1381    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1382    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1383    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1384    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
# Line 790  for debugging. The macro is referenced o Line 1394  for debugging. The macro is referenced o
1394    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1395    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1396    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
1397    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT"    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
1398      "Close", "Skip zero"
1399    
1400    
1401  /* This macro defines the length of fixed length operations in the compiled  /* This macro defines the length of fixed length operations in the compiled
# Line 806  in UTF-8 mode. The code that uses this t Line 1411  in UTF-8 mode. The code that uses this t
1411    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1412    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1413    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1414    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1415    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
1416    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1417    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
# Line 855  in UTF-8 mode. The code that uses this t Line 1460  in UTF-8 mode. The code that uses this t
1460    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1461    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1462    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
1463    1, 1                           /* FAIL, ACCEPT                           */    1, 1, 3, 1                     /* FAIL, ACCEPT, CLOSE, SKIPZERO          */
1464    
1465    
1466  /* 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 871  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 1476  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
1476         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1477         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1478         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1479         ERR60, ERR61, ERR62, ERR63 };         ERR60, ERR61, ERR62, ERR63, ERR64 };
1480    
1481  /* 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
1482  code vector run on as long as necessary after the end. We store an explicit  code vector run on as long as necessary after the end. We store an explicit
# Line 887  Because people can now save and re-use c Line 1492  Because people can now save and re-use c
1492  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
1493  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
1494  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1495  NOTE NOTE NOTE:  NOTE NOTE NOTE
1496  */  */
1497    
1498  typedef struct real_pcre {  typedef struct real_pcre {
# Line 914  remark (see NOTE above) about extending Line 1519  remark (see NOTE above) about extending
1519    
1520  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1521    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1522    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1523    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1524      pcre_uint32 minlength;          /* Minimum subject length */
1525  } pcre_study_data;  } pcre_study_data;
1526    
1527    /* Structure for building a chain of open capturing subpatterns during
1528    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1529    encountered. */
1530    
1531    typedef struct open_capitem {
1532      struct open_capitem *next;    /* Chain link */
1533      pcre_uint16 number;           /* Capture number */
1534    } open_capitem;
1535    
1536  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1537  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1538    
# Line 930  typedef struct compile_data { Line 1545  typedef struct compile_data {
1545    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1546    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1547    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1548      open_capitem *open_caps;      /* Chain of open capture items */
1549    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1550    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1551    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
1552    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
1553    int  bracount;                /* Count of capturing parens as we compile */    int  bracount;                /* Count of capturing parens as we compile */
1554    int  final_bracount;          /* Saved value after first pass */    int  final_bracount;          /* Saved value after first pass */
1555    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
1556    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
1557    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
1558    int  external_flags;          /* External flag bits to be set */    int  external_flags;          /* External flag bits to be set */
1559    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1560    BOOL had_accept;              /* (*ACCEPT) encountered */    BOOL had_accept;              /* (*ACCEPT) encountered */
1561      BOOL check_lookbehind;        /* Lookbehinds need later checking */
1562    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1563    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1564    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 965  typedef struct recursion_info { Line 1582  typedef struct recursion_info {
1582    USPTR save_start;             /* Old value of mstart */    USPTR save_start;             /* Old value of mstart */
1583    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1584    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1585      int offset_top;               /* Current value of offset_top */
1586  } recursion_info;  } recursion_info;
1587    
1588  /* 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 996  typedef struct match_data { Line 1614  typedef struct match_data {
1614    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1615    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1616    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1617      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1618    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1619    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1620    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1621    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1622    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1623    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
# Line 1006  typedef struct match_data { Line 1625  typedef struct match_data {
1625    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1626    USPTR  start_match_ptr;       /* Start of matched string */    USPTR  start_match_ptr;       /* Start of matched string */
1627    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1628      USPTR  start_used_ptr;        /* Earliest consulted character */
1629      int    partial;               /* PARTIAL options */
1630    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1631    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1632    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
# Line 1022  typedef struct dfa_match_data { Line 1643  typedef struct dfa_match_data {
1643    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;     /* Start of the compiled pattern */
1644    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
1645    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;    /* End of subject string */
1646      const uschar *start_used_ptr; /* Earliest consulted character */
1647    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1648      int   start_offset;           /* The start offset value */
1649    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1650    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1651    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
# Line 1037  typedef struct dfa_match_data { Line 1660  typedef struct dfa_match_data {
1660  #define ctype_letter  0x02  #define ctype_letter  0x02
1661  #define ctype_digit   0x04  #define ctype_digit   0x04
1662  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1663  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1664  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1665    
1666  /* 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 1101  extern const uschar _pcre_OP_lengths[]; Line 1724  extern const uschar _pcre_OP_lengths[];
1724  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
1725  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1726    
1727  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1728                        int *, BOOL);  extern BOOL          _pcre_is_newline(const uschar *, int, const uschar *,
1729  extern int          _pcre_ord2utf8(int, uschar *);                         int *, BOOL);
1730  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern int           _pcre_ord2utf8(int, uschar *);
1731                        const pcre_study_data *, pcre_study_data *);  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1732  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);                         const pcre_study_data *, pcre_study_data *);
1733  extern unsigned int _pcre_ucp_othercase(const unsigned int);  extern int           _pcre_valid_utf8(const uschar *, int);
1734  extern int          _pcre_valid_utf8(const uschar *, int);  extern BOOL          _pcre_was_newline(const uschar *, int, const uschar *,
1735  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,                         int *, BOOL);
1736                        int *, BOOL);  extern BOOL          _pcre_xclass(int, const uschar *);
1737  extern BOOL         _pcre_xclass(int, const uschar *);  
1738    
1739    /* Unicode character database (UCD) */
1740    
1741    typedef struct {
1742      uschar script;
1743      uschar chartype;
1744      pcre_int32 other_case;
1745    } ucd_record;
1746    
1747    extern const ucd_record  _pcre_ucd_records[];
1748    extern const uschar      _pcre_ucd_stage1[];
1749    extern const pcre_uint16 _pcre_ucd_stage2[];
1750    extern const int         _pcre_ucp_gentype[];
1751    
1752    
1753    /* UCD access macros */
1754    
1755    #define UCD_BLOCK_SIZE 128
1756    #define GET_UCD(ch) (_pcre_ucd_records + \
1757            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1758            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1759    
1760    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1761    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1762    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1763    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1764    
1765  #endif  #endif
1766    

Legend:
Removed from v.270  
changed lines
  Added in v.455

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12