/[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 341 by ph10, Sat Apr 19 16:41:04 2008 UTC revision 446 by ph10, Tue Sep 15 10:49:50 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) 1997-2008 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 363  never be called in byte mode. To make su Line 395  never be called in byte mode. To make su
395  support is omitted, we don't even define it. */  support is omitted, we don't even define it. */
396    
397  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
 #define NEXTCHAR(p) p++;  
398  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
399  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
400  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
# Line 373  support is omitted, we don't even define Line 404  support is omitted, we don't even define
404    
405  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
406    
 /* Advance a character pointer one byte in non-UTF-8 mode and by one character  
 in UTF-8 mode. */  
   
 #define NEXTCHAR(p) \  
   p++; \  
   if (utf8) { while((*p & 0xc0) == 0x80) p++; }  
   
407  /* Get the next UTF-8 character, not advancing the pointer. This is called when  /* Get the next UTF-8 character, not advancing the pointer. This is called when
408  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
409    
# Line 468  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 491  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 510  time, run time, or study time, respectiv Line 556  time, run time, or study time, respectiv
556  #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| \
557                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
558    
559  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
560    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
561     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
562     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 518  time, run time, or study time, respectiv Line 564  time, run time, or study time, respectiv
564     PCRE_JAVASCRIPT_COMPAT)     PCRE_JAVASCRIPT_COMPAT)
565    
566  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
567    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
568     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
569       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
570    
571  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
572    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
573     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
574     PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)     PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
575       PCRE_NO_START_OPTIMIZE)
576    
577  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
578    
# Line 549  variable-length repeat, or a anything ot Line 597  variable-length repeat, or a anything ot
597  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
598  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
599    
600  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
601    environments where these macros are defined elsewhere. Unfortunately, there
602    is no way to do the same for the typedef. */
603    
604  typedef int BOOL;  typedef int BOOL;
605    
606    #ifndef FALSE
607  #define FALSE   0  #define FALSE   0
608  #define TRUE    1  #define TRUE    1
609    #endif
610    
611    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
612    character constants like '*' because the compiler would emit their EBCDIC code,
613    which is different from their ASCII/UTF-8 code. Instead we define macros for
614    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
615    is enabled. When UTF-8 support is not enabled, the definitions use character
616    literals. Both character and string versions of each character are needed, and
617    there are some longer strings as well.
618    
619    This means that, on EBCDIC platforms, the PCRE library can handle either
620    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
621    would need different lookups depending on whether PCRE_UTF8 was set or not.
622    This would make it impossible to use characters in switch/case statements,
623    which would reduce performance. For a theoretical use (which nobody has asked
624    for) in a minority area (EBCDIC platforms), this is not sensible. Any
625    application that did need both could compile two versions of the library, using
626    macros to give the functions distinct names. */
627    
628    #ifndef SUPPORT_UTF8
629    
630    /* UTF-8 support is not enabled; use the platform-dependent character literals
631    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
632    
633    #define CHAR_HT                     '\t'
634    #define CHAR_VT                     '\v'
635    #define CHAR_FF                     '\f'
636    #define CHAR_CR                     '\r'
637    #define CHAR_NL                     '\n'
638    #define CHAR_BS                     '\b'
639    #define CHAR_BEL                    '\a'
640    #ifdef EBCDIC
641    #define CHAR_ESC                    '\047'
642    #define CHAR_DEL                    '\007'
643    #else
644    #define CHAR_ESC                    '\033'
645    #define CHAR_DEL                    '\177'
646    #endif
647    
648    #define CHAR_SPACE                  ' '
649    #define CHAR_EXCLAMATION_MARK       '!'
650    #define CHAR_QUOTATION_MARK         '"'
651    #define CHAR_NUMBER_SIGN            '#'
652    #define CHAR_DOLLAR_SIGN            '$'
653    #define CHAR_PERCENT_SIGN           '%'
654    #define CHAR_AMPERSAND              '&'
655    #define CHAR_APOSTROPHE             '\''
656    #define CHAR_LEFT_PARENTHESIS       '('
657    #define CHAR_RIGHT_PARENTHESIS      ')'
658    #define CHAR_ASTERISK               '*'
659    #define CHAR_PLUS                   '+'
660    #define CHAR_COMMA                  ','
661    #define CHAR_MINUS                  '-'
662    #define CHAR_DOT                    '.'
663    #define CHAR_SLASH                  '/'
664    #define CHAR_0                      '0'
665    #define CHAR_1                      '1'
666    #define CHAR_2                      '2'
667    #define CHAR_3                      '3'
668    #define CHAR_4                      '4'
669    #define CHAR_5                      '5'
670    #define CHAR_6                      '6'
671    #define CHAR_7                      '7'
672    #define CHAR_8                      '8'
673    #define CHAR_9                      '9'
674    #define CHAR_COLON                  ':'
675    #define CHAR_SEMICOLON              ';'
676    #define CHAR_LESS_THAN_SIGN         '<'
677    #define CHAR_EQUALS_SIGN            '='
678    #define CHAR_GREATER_THAN_SIGN      '>'
679    #define CHAR_QUESTION_MARK          '?'
680    #define CHAR_COMMERCIAL_AT          '@'
681    #define CHAR_A                      'A'
682    #define CHAR_B                      'B'
683    #define CHAR_C                      'C'
684    #define CHAR_D                      'D'
685    #define CHAR_E                      'E'
686    #define CHAR_F                      'F'
687    #define CHAR_G                      'G'
688    #define CHAR_H                      'H'
689    #define CHAR_I                      'I'
690    #define CHAR_J                      'J'
691    #define CHAR_K                      'K'
692    #define CHAR_L                      'L'
693    #define CHAR_M                      'M'
694    #define CHAR_N                      'N'
695    #define CHAR_O                      'O'
696    #define CHAR_P                      'P'
697    #define CHAR_Q                      'Q'
698    #define CHAR_R                      'R'
699    #define CHAR_S                      'S'
700    #define CHAR_T                      'T'
701    #define CHAR_U                      'U'
702    #define CHAR_V                      'V'
703    #define CHAR_W                      'W'
704    #define CHAR_X                      'X'
705    #define CHAR_Y                      'Y'
706    #define CHAR_Z                      'Z'
707    #define CHAR_LEFT_SQUARE_BRACKET    '['
708    #define CHAR_BACKSLASH              '\\'
709    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
710    #define CHAR_CIRCUMFLEX_ACCENT      '^'
711    #define CHAR_UNDERSCORE             '_'
712    #define CHAR_GRAVE_ACCENT           '`'
713    #define CHAR_a                      'a'
714    #define CHAR_b                      'b'
715    #define CHAR_c                      'c'
716    #define CHAR_d                      'd'
717    #define CHAR_e                      'e'
718    #define CHAR_f                      'f'
719    #define CHAR_g                      'g'
720    #define CHAR_h                      'h'
721    #define CHAR_i                      'i'
722    #define CHAR_j                      'j'
723    #define CHAR_k                      'k'
724    #define CHAR_l                      'l'
725    #define CHAR_m                      'm'
726    #define CHAR_n                      'n'
727    #define CHAR_o                      'o'
728    #define CHAR_p                      'p'
729    #define CHAR_q                      'q'
730    #define CHAR_r                      'r'
731    #define CHAR_s                      's'
732    #define CHAR_t                      't'
733    #define CHAR_u                      'u'
734    #define CHAR_v                      'v'
735    #define CHAR_w                      'w'
736    #define CHAR_x                      'x'
737    #define CHAR_y                      'y'
738    #define CHAR_z                      'z'
739    #define CHAR_LEFT_CURLY_BRACKET     '{'
740    #define CHAR_VERTICAL_LINE          '|'
741    #define CHAR_RIGHT_CURLY_BRACKET    '}'
742    #define CHAR_TILDE                  '~'
743    
744    #define STR_HT                      "\t"
745    #define STR_VT                      "\v"
746    #define STR_FF                      "\f"
747    #define STR_CR                      "\r"
748    #define STR_NL                      "\n"
749    #define STR_BS                      "\b"
750    #define STR_BEL                     "\a"
751    #ifdef EBCDIC
752    #define STR_ESC                     "\047"
753    #define STR_DEL                     "\007"
754    #else
755    #define STR_ESC                     "\033"
756    #define STR_DEL                     "\177"
757    #endif
758    
759    #define STR_SPACE                   " "
760    #define STR_EXCLAMATION_MARK        "!"
761    #define STR_QUOTATION_MARK          "\""
762    #define STR_NUMBER_SIGN             "#"
763    #define STR_DOLLAR_SIGN             "$"
764    #define STR_PERCENT_SIGN            "%"
765    #define STR_AMPERSAND               "&"
766    #define STR_APOSTROPHE              "'"
767    #define STR_LEFT_PARENTHESIS        "("
768    #define STR_RIGHT_PARENTHESIS       ")"
769    #define STR_ASTERISK                "*"
770    #define STR_PLUS                    "+"
771    #define STR_COMMA                   ","
772    #define STR_MINUS                   "-"
773    #define STR_DOT                     "."
774    #define STR_SLASH                   "/"
775    #define STR_0                       "0"
776    #define STR_1                       "1"
777    #define STR_2                       "2"
778    #define STR_3                       "3"
779    #define STR_4                       "4"
780    #define STR_5                       "5"
781    #define STR_6                       "6"
782    #define STR_7                       "7"
783    #define STR_8                       "8"
784    #define STR_9                       "9"
785    #define STR_COLON                   ":"
786    #define STR_SEMICOLON               ";"
787    #define STR_LESS_THAN_SIGN          "<"
788    #define STR_EQUALS_SIGN             "="
789    #define STR_GREATER_THAN_SIGN       ">"
790    #define STR_QUESTION_MARK           "?"
791    #define STR_COMMERCIAL_AT           "@"
792    #define STR_A                       "A"
793    #define STR_B                       "B"
794    #define STR_C                       "C"
795    #define STR_D                       "D"
796    #define STR_E                       "E"
797    #define STR_F                       "F"
798    #define STR_G                       "G"
799    #define STR_H                       "H"
800    #define STR_I                       "I"
801    #define STR_J                       "J"
802    #define STR_K                       "K"
803    #define STR_L                       "L"
804    #define STR_M                       "M"
805    #define STR_N                       "N"
806    #define STR_O                       "O"
807    #define STR_P                       "P"
808    #define STR_Q                       "Q"
809    #define STR_R                       "R"
810    #define STR_S                       "S"
811    #define STR_T                       "T"
812    #define STR_U                       "U"
813    #define STR_V                       "V"
814    #define STR_W                       "W"
815    #define STR_X                       "X"
816    #define STR_Y                       "Y"
817    #define STR_Z                       "Z"
818    #define STR_LEFT_SQUARE_BRACKET     "["
819    #define STR_BACKSLASH               "\\"
820    #define STR_RIGHT_SQUARE_BRACKET    "]"
821    #define STR_CIRCUMFLEX_ACCENT       "^"
822    #define STR_UNDERSCORE              "_"
823    #define STR_GRAVE_ACCENT            "`"
824    #define STR_a                       "a"
825    #define STR_b                       "b"
826    #define STR_c                       "c"
827    #define STR_d                       "d"
828    #define STR_e                       "e"
829    #define STR_f                       "f"
830    #define STR_g                       "g"
831    #define STR_h                       "h"
832    #define STR_i                       "i"
833    #define STR_j                       "j"
834    #define STR_k                       "k"
835    #define STR_l                       "l"
836    #define STR_m                       "m"
837    #define STR_n                       "n"
838    #define STR_o                       "o"
839    #define STR_p                       "p"
840    #define STR_q                       "q"
841    #define STR_r                       "r"
842    #define STR_s                       "s"
843    #define STR_t                       "t"
844    #define STR_u                       "u"
845    #define STR_v                       "v"
846    #define STR_w                       "w"
847    #define STR_x                       "x"
848    #define STR_y                       "y"
849    #define STR_z                       "z"
850    #define STR_LEFT_CURLY_BRACKET      "{"
851    #define STR_VERTICAL_LINE           "|"
852    #define STR_RIGHT_CURLY_BRACKET     "}"
853    #define STR_TILDE                   "~"
854    
855    #define STRING_ACCEPT0              "ACCEPT\0"
856    #define STRING_COMMIT0              "COMMIT\0"
857    #define STRING_F0                   "F\0"
858    #define STRING_FAIL0                "FAIL\0"
859    #define STRING_PRUNE0               "PRUNE\0"
860    #define STRING_SKIP0                "SKIP\0"
861    #define STRING_THEN                 "THEN"
862    
863    #define STRING_alpha0               "alpha\0"
864    #define STRING_lower0               "lower\0"
865    #define STRING_upper0               "upper\0"
866    #define STRING_alnum0               "alnum\0"
867    #define STRING_ascii0               "ascii\0"
868    #define STRING_blank0               "blank\0"
869    #define STRING_cntrl0               "cntrl\0"
870    #define STRING_digit0               "digit\0"
871    #define STRING_graph0               "graph\0"
872    #define STRING_print0               "print\0"
873    #define STRING_punct0               "punct\0"
874    #define STRING_space0               "space\0"
875    #define STRING_word0                "word\0"
876    #define STRING_xdigit               "xdigit"
877    
878    #define STRING_DEFINE               "DEFINE"
879    
880    #define STRING_CR_RIGHTPAR          "CR)"
881    #define STRING_LF_RIGHTPAR          "LF)"
882    #define STRING_CRLF_RIGHTPAR        "CRLF)"
883    #define STRING_ANY_RIGHTPAR         "ANY)"
884    #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"
885    #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
886    #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
887    #define STRING_UTF8_RIGHTPAR        "UTF8)"
888    
889    #else  /* SUPPORT_UTF8 */
890    
891    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
892    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
893    only. */
894    
895    #define CHAR_HT                     '\011'
896    #define CHAR_VT                     '\013'
897    #define CHAR_FF                     '\014'
898    #define CHAR_CR                     '\015'
899    #define CHAR_NL                     '\012'
900    #define CHAR_BS                     '\010'
901    #define CHAR_BEL                    '\007'
902    #define CHAR_ESC                    '\033'
903    #define CHAR_DEL                    '\177'
904    
905    #define CHAR_SPACE                  '\040'
906    #define CHAR_EXCLAMATION_MARK       '\041'
907    #define CHAR_QUOTATION_MARK         '\042'
908    #define CHAR_NUMBER_SIGN            '\043'
909    #define CHAR_DOLLAR_SIGN            '\044'
910    #define CHAR_PERCENT_SIGN           '\045'
911    #define CHAR_AMPERSAND              '\046'
912    #define CHAR_APOSTROPHE             '\047'
913    #define CHAR_LEFT_PARENTHESIS       '\050'
914    #define CHAR_RIGHT_PARENTHESIS      '\051'
915    #define CHAR_ASTERISK               '\052'
916    #define CHAR_PLUS                   '\053'
917    #define CHAR_COMMA                  '\054'
918    #define CHAR_MINUS                  '\055'
919    #define CHAR_DOT                    '\056'
920    #define CHAR_SLASH                  '\057'
921    #define CHAR_0                      '\060'
922    #define CHAR_1                      '\061'
923    #define CHAR_2                      '\062'
924    #define CHAR_3                      '\063'
925    #define CHAR_4                      '\064'
926    #define CHAR_5                      '\065'
927    #define CHAR_6                      '\066'
928    #define CHAR_7                      '\067'
929    #define CHAR_8                      '\070'
930    #define CHAR_9                      '\071'
931    #define CHAR_COLON                  '\072'
932    #define CHAR_SEMICOLON              '\073'
933    #define CHAR_LESS_THAN_SIGN         '\074'
934    #define CHAR_EQUALS_SIGN            '\075'
935    #define CHAR_GREATER_THAN_SIGN      '\076'
936    #define CHAR_QUESTION_MARK          '\077'
937    #define CHAR_COMMERCIAL_AT          '\100'
938    #define CHAR_A                      '\101'
939    #define CHAR_B                      '\102'
940    #define CHAR_C                      '\103'
941    #define CHAR_D                      '\104'
942    #define CHAR_E                      '\105'
943    #define CHAR_F                      '\106'
944    #define CHAR_G                      '\107'
945    #define CHAR_H                      '\110'
946    #define CHAR_I                      '\111'
947    #define CHAR_J                      '\112'
948    #define CHAR_K                      '\113'
949    #define CHAR_L                      '\114'
950    #define CHAR_M                      '\115'
951    #define CHAR_N                      '\116'
952    #define CHAR_O                      '\117'
953    #define CHAR_P                      '\120'
954    #define CHAR_Q                      '\121'
955    #define CHAR_R                      '\122'
956    #define CHAR_S                      '\123'
957    #define CHAR_T                      '\124'
958    #define CHAR_U                      '\125'
959    #define CHAR_V                      '\126'
960    #define CHAR_W                      '\127'
961    #define CHAR_X                      '\130'
962    #define CHAR_Y                      '\131'
963    #define CHAR_Z                      '\132'
964    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
965    #define CHAR_BACKSLASH              '\134'
966    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
967    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
968    #define CHAR_UNDERSCORE             '\137'
969    #define CHAR_GRAVE_ACCENT           '\140'
970    #define CHAR_a                      '\141'
971    #define CHAR_b                      '\142'
972    #define CHAR_c                      '\143'
973    #define CHAR_d                      '\144'
974    #define CHAR_e                      '\145'
975    #define CHAR_f                      '\146'
976    #define CHAR_g                      '\147'
977    #define CHAR_h                      '\150'
978    #define CHAR_i                      '\151'
979    #define CHAR_j                      '\152'
980    #define CHAR_k                      '\153'
981    #define CHAR_l                      '\154'
982    #define CHAR_m                      '\155'
983    #define CHAR_n                      '\156'
984    #define CHAR_o                      '\157'
985    #define CHAR_p                      '\160'
986    #define CHAR_q                      '\161'
987    #define CHAR_r                      '\162'
988    #define CHAR_s                      '\163'
989    #define CHAR_t                      '\164'
990    #define CHAR_u                      '\165'
991    #define CHAR_v                      '\166'
992    #define CHAR_w                      '\167'
993    #define CHAR_x                      '\170'
994    #define CHAR_y                      '\171'
995    #define CHAR_z                      '\172'
996    #define CHAR_LEFT_CURLY_BRACKET     '\173'
997    #define CHAR_VERTICAL_LINE          '\174'
998    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
999    #define CHAR_TILDE                  '\176'
1000    
1001    #define STR_HT                      "\011"
1002    #define STR_VT                      "\013"
1003    #define STR_FF                      "\014"
1004    #define STR_CR                      "\015"
1005    #define STR_NL                      "\012"
1006    #define STR_BS                      "\010"
1007    #define STR_BEL                     "\007"
1008    #define STR_ESC                     "\033"
1009    #define STR_DEL                     "\177"
1010    
1011    #define STR_SPACE                   "\040"
1012    #define STR_EXCLAMATION_MARK        "\041"
1013    #define STR_QUOTATION_MARK          "\042"
1014    #define STR_NUMBER_SIGN             "\043"
1015    #define STR_DOLLAR_SIGN             "\044"
1016    #define STR_PERCENT_SIGN            "\045"
1017    #define STR_AMPERSAND               "\046"
1018    #define STR_APOSTROPHE              "\047"
1019    #define STR_LEFT_PARENTHESIS        "\050"
1020    #define STR_RIGHT_PARENTHESIS       "\051"
1021    #define STR_ASTERISK                "\052"
1022    #define STR_PLUS                    "\053"
1023    #define STR_COMMA                   "\054"
1024    #define STR_MINUS                   "\055"
1025    #define STR_DOT                     "\056"
1026    #define STR_SLASH                   "\057"
1027    #define STR_0                       "\060"
1028    #define STR_1                       "\061"
1029    #define STR_2                       "\062"
1030    #define STR_3                       "\063"
1031    #define STR_4                       "\064"
1032    #define STR_5                       "\065"
1033    #define STR_6                       "\066"
1034    #define STR_7                       "\067"
1035    #define STR_8                       "\070"
1036    #define STR_9                       "\071"
1037    #define STR_COLON                   "\072"
1038    #define STR_SEMICOLON               "\073"
1039    #define STR_LESS_THAN_SIGN          "\074"
1040    #define STR_EQUALS_SIGN             "\075"
1041    #define STR_GREATER_THAN_SIGN       "\076"
1042    #define STR_QUESTION_MARK           "\077"
1043    #define STR_COMMERCIAL_AT           "\100"
1044    #define STR_A                       "\101"
1045    #define STR_B                       "\102"
1046    #define STR_C                       "\103"
1047    #define STR_D                       "\104"
1048    #define STR_E                       "\105"
1049    #define STR_F                       "\106"
1050    #define STR_G                       "\107"
1051    #define STR_H                       "\110"
1052    #define STR_I                       "\111"
1053    #define STR_J                       "\112"
1054    #define STR_K                       "\113"
1055    #define STR_L                       "\114"
1056    #define STR_M                       "\115"
1057    #define STR_N                       "\116"
1058    #define STR_O                       "\117"
1059    #define STR_P                       "\120"
1060    #define STR_Q                       "\121"
1061    #define STR_R                       "\122"
1062    #define STR_S                       "\123"
1063    #define STR_T                       "\124"
1064    #define STR_U                       "\125"
1065    #define STR_V                       "\126"
1066    #define STR_W                       "\127"
1067    #define STR_X                       "\130"
1068    #define STR_Y                       "\131"
1069    #define STR_Z                       "\132"
1070    #define STR_LEFT_SQUARE_BRACKET     "\133"
1071    #define STR_BACKSLASH               "\134"
1072    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1073    #define STR_CIRCUMFLEX_ACCENT       "\136"
1074    #define STR_UNDERSCORE              "\137"
1075    #define STR_GRAVE_ACCENT            "\140"
1076    #define STR_a                       "\141"
1077    #define STR_b                       "\142"
1078    #define STR_c                       "\143"
1079    #define STR_d                       "\144"
1080    #define STR_e                       "\145"
1081    #define STR_f                       "\146"
1082    #define STR_g                       "\147"
1083    #define STR_h                       "\150"
1084    #define STR_i                       "\151"
1085    #define STR_j                       "\152"
1086    #define STR_k                       "\153"
1087    #define STR_l                       "\154"
1088    #define STR_m                       "\155"
1089    #define STR_n                       "\156"
1090    #define STR_o                       "\157"
1091    #define STR_p                       "\160"
1092    #define STR_q                       "\161"
1093    #define STR_r                       "\162"
1094    #define STR_s                       "\163"
1095    #define STR_t                       "\164"
1096    #define STR_u                       "\165"
1097    #define STR_v                       "\166"
1098    #define STR_w                       "\167"
1099    #define STR_x                       "\170"
1100    #define STR_y                       "\171"
1101    #define STR_z                       "\172"
1102    #define STR_LEFT_CURLY_BRACKET      "\173"
1103    #define STR_VERTICAL_LINE           "\174"
1104    #define STR_RIGHT_CURLY_BRACKET     "\175"
1105    #define STR_TILDE                   "\176"
1106    
1107    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1108    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1109    #define STRING_F0                   STR_F "\0"
1110    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1111    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1112    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1113    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1114    
1115    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1116    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1117    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1118    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1119    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1120    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1121    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1122    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1123    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1124    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1125    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1126    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1127    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1128    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1129    
1130    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1131    
1132    #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS
1133    #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS
1134    #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1135    #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1136    #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1137    #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
1138    #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
1139    #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1140    
1141    #endif  /* SUPPORT_UTF8 */
1142    
1143  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1144    
1145  #ifndef ESC_e  #ifndef ESC_e
1146  #define ESC_e 27  #define ESC_e CHAR_ESC
1147  #endif  #endif
1148    
1149  #ifndef ESC_f  #ifndef ESC_f
1150  #define ESC_f '\f'  #define ESC_f CHAR_FF
1151  #endif  #endif
1152    
1153  #ifndef ESC_n  #ifndef ESC_n
1154  #define ESC_n '\n'  #define ESC_n CHAR_NL
1155  #endif  #endif
1156    
1157  #ifndef ESC_r  #ifndef ESC_r
1158  #define ESC_r '\r'  #define ESC_r CHAR_CR
1159  #endif  #endif
1160    
1161  /* 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
1162  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1163    
1164  #ifndef ESC_tee  #ifndef ESC_tee
1165  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1166  #endif  #endif
1167    
1168  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
# Line 605  contain UTF-8 characters with values gre Line 1189  contain UTF-8 characters with values gre
1189  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
1190  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
1191  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
1192  corresponds to "." rather than an escape sequence, and another for OP_ALLANY  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
1193  (which is used for [^] in JavaScript compatibility mode).  (which is used for [^] in JavaScript compatibility mode).
1194    
1195  The final escape must be ESC_REF as subsequent values are used for  The final escape must be ESC_REF as subsequent values are used for
# Line 616  put in between that don't consume a char Line 1200  put in between that don't consume a char
1200  */  */
1201    
1202  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,
1203         ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,         ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1204         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
1205         ESC_REF };         ESC_REF };
1206    
1207    
# Line 780  enum { Line 1364  enum {
1364    
1365    OP_FAIL,           /* 109 */    OP_FAIL,           /* 109 */
1366    OP_ACCEPT,         /* 110 */    OP_ACCEPT,         /* 110 */
1367    
1368    /* This is used to skip a subpattern with a {0} quantifier */    /* This is used to skip a subpattern with a {0} quantifier */
1369    
1370    OP_SKIPZERO        /* 111 */    OP_SKIPZERO        /* 111 */
1371  };  };
1372    
# Line 809  for debugging. The macro is referenced o Line 1393  for debugging. The macro is referenced o
1393    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1394    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
1395    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \    "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
1396    "Skip zero"    "Skip zero"
1397    
1398    
1399  /* 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 984  typedef struct recursion_info { Line 1568  typedef struct recursion_info {
1568    USPTR save_start;             /* Old value of mstart */    USPTR save_start;             /* Old value of mstart */
1569    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1570    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1571      int offset_top;               /* Current value of offset_top */
1572  } recursion_info;  } recursion_info;
1573    
1574  /* 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 1015  typedef struct match_data { Line 1600  typedef struct match_data {
1600    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1601    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1602    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1603    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */    BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1604    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1605    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1606    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1607    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1608    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */    BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1609    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
# Line 1026  typedef struct match_data { Line 1611  typedef struct match_data {
1611    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1612    USPTR  start_match_ptr;       /* Start of matched string */    USPTR  start_match_ptr;       /* Start of matched string */
1613    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1614      USPTR  start_used_ptr;        /* Earliest consulted character */
1615      int    partial;               /* PARTIAL options */
1616    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1617    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1618    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
# Line 1042  typedef struct dfa_match_data { Line 1629  typedef struct dfa_match_data {
1629    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;     /* Start of the compiled pattern */
1630    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
1631    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;    /* End of subject string */
1632      const uschar *start_used_ptr; /* Earliest consulted character */
1633    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1634      int   start_offset;           /* The start offset value */
1635    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1636    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1637    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
# Line 1126  extern BOOL _pcre_is_newline(con Line 1715  extern BOOL _pcre_is_newline(con
1715  extern int          _pcre_ord2utf8(int, uschar *);  extern int          _pcre_ord2utf8(int, uschar *);
1716  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1717                        const pcre_study_data *, pcre_study_data *);                        const pcre_study_data *, pcre_study_data *);
 extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  
 extern unsigned int _pcre_ucp_othercase(const unsigned int);  
1718  extern int          _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_valid_utf8(const uschar *, int);
1719  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1720                        int *, BOOL);                        int *, BOOL);
1721  extern BOOL         _pcre_xclass(int, const uschar *);  extern BOOL         _pcre_xclass(int, const uschar *);
1722    
1723    
1724    /* Unicode character database (UCD) */
1725    
1726    typedef struct {
1727      uschar script;
1728      uschar chartype;
1729      pcre_int32 other_case;
1730    } ucd_record;
1731    
1732    extern const ucd_record  _pcre_ucd_records[];
1733    extern const uschar      _pcre_ucd_stage1[];
1734    extern const pcre_uint16 _pcre_ucd_stage2[];
1735    extern const int         _pcre_ucp_gentype[];
1736    
1737    
1738    /* UCD access macros */
1739    
1740    #define UCD_BLOCK_SIZE 128
1741    #define GET_UCD(ch) (_pcre_ucd_records + \
1742            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1743            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1744    
1745    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1746    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1747    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1748    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1749    
1750  #endif  #endif
1751    
1752  /* End of pcre_internal.h */  /* End of pcre_internal.h */

Legend:
Removed from v.341  
changed lines
  Added in v.446

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12