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

Diff of /code/trunk/pcre_internal.h

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

revision 384 by ph10, Sun Mar 8 16:27:43 2009 UTC revision 457 by ph10, Sat Oct 3 16:24:08 2009 UTC
# 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 521  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 533  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 540  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| \
# Line 548  time, run time, or study time, respectiv Line 565  time, run time, or study time, respectiv
565     PCRE_JAVASCRIPT_COMPAT)     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 590  typedef int BOOL; Line 609  typedef int BOOL;
609  #define TRUE    1  #define TRUE    1
610  #endif  #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 814  enum { Line 1365  enum {
1365    
1366    OP_FAIL,           /* 109 */    OP_FAIL,           /* 109 */
1367    OP_ACCEPT,         /* 110 */    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 */    /* This is used to skip a subpattern with a {0} quantifier */
1371    
1372    OP_SKIPZERO        /* 111 */    OP_SKIPZERO        /* 112 */
1373  };  };
1374    
1375    
# Line 843  for debugging. The macro is referenced o Line 1395  for debugging. The macro is referenced o
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    "Skip zero"    "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 908  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, 1                        /* FAIL, ACCEPT, SKIPZERO                 */    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 924  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, ERR64 };         ERR60, ERR61, ERR62, ERR63, ERR64, ERR65 };
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 940  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 967  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 983  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 */
# Line 995  typedef struct compile_data { Line 1558  typedef struct compile_data {
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 1018  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 1052  typedef struct match_data { Line 1617  typedef struct match_data {
1617    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    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 1060  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 1076  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 1155  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_valid_utf8(const uschar *, int);                         const pcre_study_data *, pcre_study_data *);
1733  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern int           _pcre_valid_utf8(const uschar *, int);
1734                        int *, BOOL);  extern BOOL          _pcre_was_newline(const uschar *, int, const uschar *,
1735  extern BOOL         _pcre_xclass(int, const uschar *);                         int *, BOOL);
1736    extern BOOL          _pcre_xclass(int, const uschar *);
1737    
1738    
1739  /* Unicode character database (UCD) */  /* Unicode character database (UCD) */

Legend:
Removed from v.384  
changed lines
  Added in v.457

  ViewVC Help
Powered by ViewVC 1.1.5