/[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 178 by ph10, Wed Jun 13 08:44:34 2007 UTC revision 512 by ph10, Tue Mar 30 11:11:52 2010 UTC
# Line 7  Line 7 
7  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2007 University of Cambridge             Copyright (c) 1997-2010 University of Cambridge
11    
12  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
13  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 45  functions whose names all begin with "_p Line 45  functions whose names all begin with "_p
45  #ifndef PCRE_INTERNAL_H  #ifndef PCRE_INTERNAL_H
46  #define PCRE_INTERNAL_H  #define PCRE_INTERNAL_H
47    
48  /* Define DEBUG to get debugging output on stdout. */  /* Define PCRE_DEBUG to get debugging output on stdout. */
49    
50  #if 0  #if 0
51  #define DEBUG  #define PCRE_DEBUG
52    #endif
53    
54    /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"
55    script prevents both being selected, but not everybody uses "configure". */
56    
57    #if defined EBCDIC && defined SUPPORT_UTF8
58    #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.
59    #endif
60    
61    /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
62    "configure" script ensures this, but not everybody uses "configure". */
63    
64    #if defined SUPPORT_UCP && !defined SUPPORT_UTF8
65    #define SUPPORT_UTF8 1
66  #endif  #endif
67    
68  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef
# Line 60  It turns out that the Mac Debugging.h he Line 74  It turns out that the Mac Debugging.h he
74  be absolutely sure we get our version. */  be absolutely sure we get our version. */
75    
76  #undef DPRINTF  #undef DPRINTF
77  #ifdef DEBUG  #ifdef PCRE_DEBUG
78  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
79  #else  #else
80  #define DPRINTF(p) /* Nothing */  #define DPRINTF(p) /* Nothing */
81  #endif  #endif
82    
83    
 /* Get the definitions provided by running "configure" */  
   
 #include "config.h"  
   
84  /* Standard C headers plus the external interface definition. The only time  /* Standard C headers plus the external interface definition. The only time
85  setjmp and stdarg are used is when NO_RECURSE is set. */  setjmp and stdarg are used is when NO_RECURSE is set. */
86    
87  #include <ctype.h>  #include <ctype.h>
88  #include <limits.h>  #include <limits.h>
 #include <setjmp.h>  
 #include <stdarg.h>  
89  #include <stddef.h>  #include <stddef.h>
90  #include <stdio.h>  #include <stdio.h>
91  #include <stdlib.h>  #include <stdlib.h>
# Line 112  PCRE_EXP_DATA_DEFN only if they are not Line 120  PCRE_EXP_DATA_DEFN only if they are not
120    
121  #ifndef PCRE_EXP_DECL  #ifndef PCRE_EXP_DECL
122  #  ifdef _WIN32  #  ifdef _WIN32
123  #    ifdef DLL_EXPORT  #    ifndef PCRE_STATIC
124  #      define PCRE_EXP_DECL       extern __declspec(dllexport)  #      define PCRE_EXP_DECL       extern __declspec(dllexport)
125  #      define PCRE_EXP_DEFN       __declspec(dllexport)  #      define PCRE_EXP_DEFN       __declspec(dllexport)
126  #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)  #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
# Line 121  PCRE_EXP_DATA_DEFN only if they are not Line 129  PCRE_EXP_DATA_DEFN only if they are not
129  #      define PCRE_EXP_DEFN  #      define PCRE_EXP_DEFN
130  #      define PCRE_EXP_DATA_DEFN  #      define PCRE_EXP_DATA_DEFN
131  #    endif  #    endif
 #  
132  #  else  #  else
133  #    ifdef __cplusplus  #    ifdef __cplusplus
134  #      define PCRE_EXP_DECL       extern "C"  #      define PCRE_EXP_DECL       extern "C"
# Line 137  PCRE_EXP_DATA_DEFN only if they are not Line 144  PCRE_EXP_DATA_DEFN only if they are not
144  #  endif  #  endif
145  #endif  #endif
146    
147    /* When compiling with the MSVC compiler, it is sometimes necessary to include
148    a "calling convention" before exported function names. (This is secondhand
149    information; I know nothing about MSVC myself). For example, something like
150    
151      void __cdecl function(....)
152    
153    might be needed. In order so make this easy, all the exported functions have
154    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
155    set, we ensure here that it has no effect. */
156    
157    #ifndef PCRE_CALL_CONVENTION
158    #define PCRE_CALL_CONVENTION
159    #endif
160    
161  /* 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
162  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
163  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 145  preprocessor time in standard C environm Line 166  preprocessor time in standard C environm
166    
167  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
168    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
169      typedef short pcre_int16;
170  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
171    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
172      typedef int pcre_int16;
173  #else  #else
174    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
175  #endif  #endif
176    
177  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
178    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
179      typedef int pcre_int32;
180  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
181    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
182      typedef long int pcre_int32;
183  #else  #else
184    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
185  #endif  #endif
186    
187    /* When checking for integer overflow in pcre_compile(), we need to handle
188    large integers. If a 64-bit integer type is available, we can use that.
189    Otherwise we have to cast to double, which of course requires floating point
190    arithmetic. Handle this by defining a macro for the appropriate type. If
191    stdint.h is available, include it; it may define INT64_MAX. Systems that do not
192    have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
193    by "configure". */
194    
195    #if HAVE_STDINT_H
196    #include <stdint.h>
197    #elif HAVE_INTTYPES_H
198    #include <inttypes.h>
199    #endif
200    
201    #if defined INT64_MAX || defined int64_t
202    #define INT64_OR_DOUBLE int64_t
203    #else
204    #define INT64_OR_DOUBLE double
205    #endif
206    
207  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
208  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
209  However, we leave the interface to the outside world as char *, because that  However, we leave the interface to the outside world as char *, because that
# Line 234  must begin with PCRE_. */ Line 279  must begin with PCRE_. */
279  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
280  values. */  values. */
281    
282  #include <pcre.h>  #include "pcre.h"
283  #include "ucp.h"  #include "ucp.h"
284    
285  /* When compiling for use with the Virtual Pascal compiler, these functions  /* When compiling for use with the Virtual Pascal compiler, these functions
# Line 363  capturing parenthesis numbers in back re Line 408  capturing parenthesis numbers in back re
408    
409  /* When UTF-8 encoding is being used, a character is no longer just a single  /* When UTF-8 encoding is being used, a character is no longer just a single
410  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
411  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
412    never be called in byte mode. To make sure it can never even appear when UTF-8
413    support is omitted, we don't even define it. */
414    
415  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
416  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 371  byte-mode, and more complicated ones for Line 418  byte-mode, and more complicated ones for
418  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
419  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
420  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
421  #define BACKCHAR(eptr)  /* #define BACKCHAR(eptr) */
422    
423  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
424    
# Line 463  if there are extra bytes. This is called Line 510  if there are extra bytes. This is called
510      len += gcaa; \      len += gcaa; \
511      }      }
512    
513    /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
514    pointer, incrementing length if there are extra bytes. This is called when we
515    know we are in UTF-8 mode. */
516    
517    #define GETCHARLENTEST(c, eptr, len) \
518      c = *eptr; \
519      if (utf8 && c >= 0xc0) \
520        { \
521        int gcii; \
522        int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
523        int gcss = 6*gcaa; \
524        c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
525        for (gcii = 1; gcii <= gcaa; gcii++) \
526          { \
527          gcss -= 6; \
528          c |= (eptr[gcii] & 0x3f) << gcss; \
529          } \
530        len += gcaa; \
531        }
532    
533  /* 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
534  it is. Called only in UTF-8 mode. */  it is. This is called only in UTF-8 mode - we don't put a test within the macro
535    because almost all calls are already within a block of UTF-8 only code. */
536    
537  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
538    
539  #endif  #endif
540    
# Line 483  Standard C system should have one. */ Line 551  Standard C system should have one. */
551    
552  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
553    
554  /* Private options flags start at the most significant end of the four bytes.  /* Private flags containing information about the compiled regex. They used to
555  The public options defined in pcre.h start at the least significant end. Make  live at the top end of the options word, but that got almost full, so now they
556  sure they don't overlap! The bits are getting a bit scarce now -- when we run  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
557  out, there is a dummy word in the structure that could be used for the private  the restrictions on partial matching have been lifted. It remains for backwards
558  bits. */  compatibility. */
559    
560  #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
561  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
562  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
563  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
564  #define PCRE_JCHANGED      0x08000000  /* j option changes within regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
565    #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
566    
567  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
568    
569  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
570    #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */
571    
572  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
573  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
# Line 505  time, run time, or study time, respectiv Line 575  time, run time, or study time, respectiv
575  #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| \
576                             PCRE_NEWLINE_ANYCRLF)                             PCRE_NEWLINE_ANYCRLF)
577    
578  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
579    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
580     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
581     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| \
582     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
583       PCRE_JAVASCRIPT_COMPAT)
584    
585  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
586    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
587     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
588       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
589    
590  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
591    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
592     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
593       PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
594       PCRE_NO_START_OPTIMIZE)
595    
596  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
597    
# Line 542  variable-length repeat, or a anything ot Line 616  variable-length repeat, or a anything ot
616  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
617  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
618    
619  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
620    environments where these macros are defined elsewhere. Unfortunately, there
621    is no way to do the same for the typedef. */
622    
623  typedef int BOOL;  typedef int BOOL;
624    
625    #ifndef FALSE
626  #define FALSE   0  #define FALSE   0
627  #define TRUE    1  #define TRUE    1
628    #endif
629    
630    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
631    character constants like '*' because the compiler would emit their EBCDIC code,
632    which is different from their ASCII/UTF-8 code. Instead we define macros for
633    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
634    is enabled. When UTF-8 support is not enabled, the definitions use character
635    literals. Both character and string versions of each character are needed, and
636    there are some longer strings as well.
637    
638    This means that, on EBCDIC platforms, the PCRE library can handle either
639    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
640    would need different lookups depending on whether PCRE_UTF8 was set or not.
641    This would make it impossible to use characters in switch/case statements,
642    which would reduce performance. For a theoretical use (which nobody has asked
643    for) in a minority area (EBCDIC platforms), this is not sensible. Any
644    application that did need both could compile two versions of the library, using
645    macros to give the functions distinct names. */
646    
647    #ifndef SUPPORT_UTF8
648    
649    /* UTF-8 support is not enabled; use the platform-dependent character literals
650    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
651    
652    #define CHAR_HT                     '\t'
653    #define CHAR_VT                     '\v'
654    #define CHAR_FF                     '\f'
655    #define CHAR_CR                     '\r'
656    #define CHAR_NL                     '\n'
657    #define CHAR_BS                     '\b'
658    #define CHAR_BEL                    '\a'
659    #ifdef EBCDIC
660    #define CHAR_ESC                    '\047'
661    #define CHAR_DEL                    '\007'
662    #else
663    #define CHAR_ESC                    '\033'
664    #define CHAR_DEL                    '\177'
665    #endif
666    
667    #define CHAR_SPACE                  ' '
668    #define CHAR_EXCLAMATION_MARK       '!'
669    #define CHAR_QUOTATION_MARK         '"'
670    #define CHAR_NUMBER_SIGN            '#'
671    #define CHAR_DOLLAR_SIGN            '$'
672    #define CHAR_PERCENT_SIGN           '%'
673    #define CHAR_AMPERSAND              '&'
674    #define CHAR_APOSTROPHE             '\''
675    #define CHAR_LEFT_PARENTHESIS       '('
676    #define CHAR_RIGHT_PARENTHESIS      ')'
677    #define CHAR_ASTERISK               '*'
678    #define CHAR_PLUS                   '+'
679    #define CHAR_COMMA                  ','
680    #define CHAR_MINUS                  '-'
681    #define CHAR_DOT                    '.'
682    #define CHAR_SLASH                  '/'
683    #define CHAR_0                      '0'
684    #define CHAR_1                      '1'
685    #define CHAR_2                      '2'
686    #define CHAR_3                      '3'
687    #define CHAR_4                      '4'
688    #define CHAR_5                      '5'
689    #define CHAR_6                      '6'
690    #define CHAR_7                      '7'
691    #define CHAR_8                      '8'
692    #define CHAR_9                      '9'
693    #define CHAR_COLON                  ':'
694    #define CHAR_SEMICOLON              ';'
695    #define CHAR_LESS_THAN_SIGN         '<'
696    #define CHAR_EQUALS_SIGN            '='
697    #define CHAR_GREATER_THAN_SIGN      '>'
698    #define CHAR_QUESTION_MARK          '?'
699    #define CHAR_COMMERCIAL_AT          '@'
700    #define CHAR_A                      'A'
701    #define CHAR_B                      'B'
702    #define CHAR_C                      'C'
703    #define CHAR_D                      'D'
704    #define CHAR_E                      'E'
705    #define CHAR_F                      'F'
706    #define CHAR_G                      'G'
707    #define CHAR_H                      'H'
708    #define CHAR_I                      'I'
709    #define CHAR_J                      'J'
710    #define CHAR_K                      'K'
711    #define CHAR_L                      'L'
712    #define CHAR_M                      'M'
713    #define CHAR_N                      'N'
714    #define CHAR_O                      'O'
715    #define CHAR_P                      'P'
716    #define CHAR_Q                      'Q'
717    #define CHAR_R                      'R'
718    #define CHAR_S                      'S'
719    #define CHAR_T                      'T'
720    #define CHAR_U                      'U'
721    #define CHAR_V                      'V'
722    #define CHAR_W                      'W'
723    #define CHAR_X                      'X'
724    #define CHAR_Y                      'Y'
725    #define CHAR_Z                      'Z'
726    #define CHAR_LEFT_SQUARE_BRACKET    '['
727    #define CHAR_BACKSLASH              '\\'
728    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
729    #define CHAR_CIRCUMFLEX_ACCENT      '^'
730    #define CHAR_UNDERSCORE             '_'
731    #define CHAR_GRAVE_ACCENT           '`'
732    #define CHAR_a                      'a'
733    #define CHAR_b                      'b'
734    #define CHAR_c                      'c'
735    #define CHAR_d                      'd'
736    #define CHAR_e                      'e'
737    #define CHAR_f                      'f'
738    #define CHAR_g                      'g'
739    #define CHAR_h                      'h'
740    #define CHAR_i                      'i'
741    #define CHAR_j                      'j'
742    #define CHAR_k                      'k'
743    #define CHAR_l                      'l'
744    #define CHAR_m                      'm'
745    #define CHAR_n                      'n'
746    #define CHAR_o                      'o'
747    #define CHAR_p                      'p'
748    #define CHAR_q                      'q'
749    #define CHAR_r                      'r'
750    #define CHAR_s                      's'
751    #define CHAR_t                      't'
752    #define CHAR_u                      'u'
753    #define CHAR_v                      'v'
754    #define CHAR_w                      'w'
755    #define CHAR_x                      'x'
756    #define CHAR_y                      'y'
757    #define CHAR_z                      'z'
758    #define CHAR_LEFT_CURLY_BRACKET     '{'
759    #define CHAR_VERTICAL_LINE          '|'
760    #define CHAR_RIGHT_CURLY_BRACKET    '}'
761    #define CHAR_TILDE                  '~'
762    
763    #define STR_HT                      "\t"
764    #define STR_VT                      "\v"
765    #define STR_FF                      "\f"
766    #define STR_CR                      "\r"
767    #define STR_NL                      "\n"
768    #define STR_BS                      "\b"
769    #define STR_BEL                     "\a"
770    #ifdef EBCDIC
771    #define STR_ESC                     "\047"
772    #define STR_DEL                     "\007"
773    #else
774    #define STR_ESC                     "\033"
775    #define STR_DEL                     "\177"
776    #endif
777    
778    #define STR_SPACE                   " "
779    #define STR_EXCLAMATION_MARK        "!"
780    #define STR_QUOTATION_MARK          "\""
781    #define STR_NUMBER_SIGN             "#"
782    #define STR_DOLLAR_SIGN             "$"
783    #define STR_PERCENT_SIGN            "%"
784    #define STR_AMPERSAND               "&"
785    #define STR_APOSTROPHE              "'"
786    #define STR_LEFT_PARENTHESIS        "("
787    #define STR_RIGHT_PARENTHESIS       ")"
788    #define STR_ASTERISK                "*"
789    #define STR_PLUS                    "+"
790    #define STR_COMMA                   ","
791    #define STR_MINUS                   "-"
792    #define STR_DOT                     "."
793    #define STR_SLASH                   "/"
794    #define STR_0                       "0"
795    #define STR_1                       "1"
796    #define STR_2                       "2"
797    #define STR_3                       "3"
798    #define STR_4                       "4"
799    #define STR_5                       "5"
800    #define STR_6                       "6"
801    #define STR_7                       "7"
802    #define STR_8                       "8"
803    #define STR_9                       "9"
804    #define STR_COLON                   ":"
805    #define STR_SEMICOLON               ";"
806    #define STR_LESS_THAN_SIGN          "<"
807    #define STR_EQUALS_SIGN             "="
808    #define STR_GREATER_THAN_SIGN       ">"
809    #define STR_QUESTION_MARK           "?"
810    #define STR_COMMERCIAL_AT           "@"
811    #define STR_A                       "A"
812    #define STR_B                       "B"
813    #define STR_C                       "C"
814    #define STR_D                       "D"
815    #define STR_E                       "E"
816    #define STR_F                       "F"
817    #define STR_G                       "G"
818    #define STR_H                       "H"
819    #define STR_I                       "I"
820    #define STR_J                       "J"
821    #define STR_K                       "K"
822    #define STR_L                       "L"
823    #define STR_M                       "M"
824    #define STR_N                       "N"
825    #define STR_O                       "O"
826    #define STR_P                       "P"
827    #define STR_Q                       "Q"
828    #define STR_R                       "R"
829    #define STR_S                       "S"
830    #define STR_T                       "T"
831    #define STR_U                       "U"
832    #define STR_V                       "V"
833    #define STR_W                       "W"
834    #define STR_X                       "X"
835    #define STR_Y                       "Y"
836    #define STR_Z                       "Z"
837    #define STR_LEFT_SQUARE_BRACKET     "["
838    #define STR_BACKSLASH               "\\"
839    #define STR_RIGHT_SQUARE_BRACKET    "]"
840    #define STR_CIRCUMFLEX_ACCENT       "^"
841    #define STR_UNDERSCORE              "_"
842    #define STR_GRAVE_ACCENT            "`"
843    #define STR_a                       "a"
844    #define STR_b                       "b"
845    #define STR_c                       "c"
846    #define STR_d                       "d"
847    #define STR_e                       "e"
848    #define STR_f                       "f"
849    #define STR_g                       "g"
850    #define STR_h                       "h"
851    #define STR_i                       "i"
852    #define STR_j                       "j"
853    #define STR_k                       "k"
854    #define STR_l                       "l"
855    #define STR_m                       "m"
856    #define STR_n                       "n"
857    #define STR_o                       "o"
858    #define STR_p                       "p"
859    #define STR_q                       "q"
860    #define STR_r                       "r"
861    #define STR_s                       "s"
862    #define STR_t                       "t"
863    #define STR_u                       "u"
864    #define STR_v                       "v"
865    #define STR_w                       "w"
866    #define STR_x                       "x"
867    #define STR_y                       "y"
868    #define STR_z                       "z"
869    #define STR_LEFT_CURLY_BRACKET      "{"
870    #define STR_VERTICAL_LINE           "|"
871    #define STR_RIGHT_CURLY_BRACKET     "}"
872    #define STR_TILDE                   "~"
873    
874    #define STRING_ACCEPT0              "ACCEPT\0"
875    #define STRING_COMMIT0              "COMMIT\0"
876    #define STRING_F0                   "F\0"
877    #define STRING_FAIL0                "FAIL\0"
878    #define STRING_MARK0                "MARK\0"
879    #define STRING_PRUNE0               "PRUNE\0"
880    #define STRING_SKIP0                "SKIP\0"
881    #define STRING_THEN                 "THEN"
882    
883    #define STRING_alpha0               "alpha\0"
884    #define STRING_lower0               "lower\0"
885    #define STRING_upper0               "upper\0"
886    #define STRING_alnum0               "alnum\0"
887    #define STRING_ascii0               "ascii\0"
888    #define STRING_blank0               "blank\0"
889    #define STRING_cntrl0               "cntrl\0"
890    #define STRING_digit0               "digit\0"
891    #define STRING_graph0               "graph\0"
892    #define STRING_print0               "print\0"
893    #define STRING_punct0               "punct\0"
894    #define STRING_space0               "space\0"
895    #define STRING_word0                "word\0"
896    #define STRING_xdigit               "xdigit"
897    
898    #define STRING_DEFINE               "DEFINE"
899    
900    #define STRING_CR_RIGHTPAR          "CR)"
901    #define STRING_LF_RIGHTPAR          "LF)"
902    #define STRING_CRLF_RIGHTPAR        "CRLF)"
903    #define STRING_ANY_RIGHTPAR         "ANY)"
904    #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"
905    #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
906    #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
907    #define STRING_UTF8_RIGHTPAR        "UTF8)"
908    
909    #else  /* SUPPORT_UTF8 */
910    
911    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
912    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
913    only. */
914    
915    #define CHAR_HT                     '\011'
916    #define CHAR_VT                     '\013'
917    #define CHAR_FF                     '\014'
918    #define CHAR_CR                     '\015'
919    #define CHAR_NL                     '\012'
920    #define CHAR_BS                     '\010'
921    #define CHAR_BEL                    '\007'
922    #define CHAR_ESC                    '\033'
923    #define CHAR_DEL                    '\177'
924    
925    #define CHAR_SPACE                  '\040'
926    #define CHAR_EXCLAMATION_MARK       '\041'
927    #define CHAR_QUOTATION_MARK         '\042'
928    #define CHAR_NUMBER_SIGN            '\043'
929    #define CHAR_DOLLAR_SIGN            '\044'
930    #define CHAR_PERCENT_SIGN           '\045'
931    #define CHAR_AMPERSAND              '\046'
932    #define CHAR_APOSTROPHE             '\047'
933    #define CHAR_LEFT_PARENTHESIS       '\050'
934    #define CHAR_RIGHT_PARENTHESIS      '\051'
935    #define CHAR_ASTERISK               '\052'
936    #define CHAR_PLUS                   '\053'
937    #define CHAR_COMMA                  '\054'
938    #define CHAR_MINUS                  '\055'
939    #define CHAR_DOT                    '\056'
940    #define CHAR_SLASH                  '\057'
941    #define CHAR_0                      '\060'
942    #define CHAR_1                      '\061'
943    #define CHAR_2                      '\062'
944    #define CHAR_3                      '\063'
945    #define CHAR_4                      '\064'
946    #define CHAR_5                      '\065'
947    #define CHAR_6                      '\066'
948    #define CHAR_7                      '\067'
949    #define CHAR_8                      '\070'
950    #define CHAR_9                      '\071'
951    #define CHAR_COLON                  '\072'
952    #define CHAR_SEMICOLON              '\073'
953    #define CHAR_LESS_THAN_SIGN         '\074'
954    #define CHAR_EQUALS_SIGN            '\075'
955    #define CHAR_GREATER_THAN_SIGN      '\076'
956    #define CHAR_QUESTION_MARK          '\077'
957    #define CHAR_COMMERCIAL_AT          '\100'
958    #define CHAR_A                      '\101'
959    #define CHAR_B                      '\102'
960    #define CHAR_C                      '\103'
961    #define CHAR_D                      '\104'
962    #define CHAR_E                      '\105'
963    #define CHAR_F                      '\106'
964    #define CHAR_G                      '\107'
965    #define CHAR_H                      '\110'
966    #define CHAR_I                      '\111'
967    #define CHAR_J                      '\112'
968    #define CHAR_K                      '\113'
969    #define CHAR_L                      '\114'
970    #define CHAR_M                      '\115'
971    #define CHAR_N                      '\116'
972    #define CHAR_O                      '\117'
973    #define CHAR_P                      '\120'
974    #define CHAR_Q                      '\121'
975    #define CHAR_R                      '\122'
976    #define CHAR_S                      '\123'
977    #define CHAR_T                      '\124'
978    #define CHAR_U                      '\125'
979    #define CHAR_V                      '\126'
980    #define CHAR_W                      '\127'
981    #define CHAR_X                      '\130'
982    #define CHAR_Y                      '\131'
983    #define CHAR_Z                      '\132'
984    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
985    #define CHAR_BACKSLASH              '\134'
986    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
987    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
988    #define CHAR_UNDERSCORE             '\137'
989    #define CHAR_GRAVE_ACCENT           '\140'
990    #define CHAR_a                      '\141'
991    #define CHAR_b                      '\142'
992    #define CHAR_c                      '\143'
993    #define CHAR_d                      '\144'
994    #define CHAR_e                      '\145'
995    #define CHAR_f                      '\146'
996    #define CHAR_g                      '\147'
997    #define CHAR_h                      '\150'
998    #define CHAR_i                      '\151'
999    #define CHAR_j                      '\152'
1000    #define CHAR_k                      '\153'
1001    #define CHAR_l                      '\154'
1002    #define CHAR_m                      '\155'
1003    #define CHAR_n                      '\156'
1004    #define CHAR_o                      '\157'
1005    #define CHAR_p                      '\160'
1006    #define CHAR_q                      '\161'
1007    #define CHAR_r                      '\162'
1008    #define CHAR_s                      '\163'
1009    #define CHAR_t                      '\164'
1010    #define CHAR_u                      '\165'
1011    #define CHAR_v                      '\166'
1012    #define CHAR_w                      '\167'
1013    #define CHAR_x                      '\170'
1014    #define CHAR_y                      '\171'
1015    #define CHAR_z                      '\172'
1016    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1017    #define CHAR_VERTICAL_LINE          '\174'
1018    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1019    #define CHAR_TILDE                  '\176'
1020    
1021    #define STR_HT                      "\011"
1022    #define STR_VT                      "\013"
1023    #define STR_FF                      "\014"
1024    #define STR_CR                      "\015"
1025    #define STR_NL                      "\012"
1026    #define STR_BS                      "\010"
1027    #define STR_BEL                     "\007"
1028    #define STR_ESC                     "\033"
1029    #define STR_DEL                     "\177"
1030    
1031    #define STR_SPACE                   "\040"
1032    #define STR_EXCLAMATION_MARK        "\041"
1033    #define STR_QUOTATION_MARK          "\042"
1034    #define STR_NUMBER_SIGN             "\043"
1035    #define STR_DOLLAR_SIGN             "\044"
1036    #define STR_PERCENT_SIGN            "\045"
1037    #define STR_AMPERSAND               "\046"
1038    #define STR_APOSTROPHE              "\047"
1039    #define STR_LEFT_PARENTHESIS        "\050"
1040    #define STR_RIGHT_PARENTHESIS       "\051"
1041    #define STR_ASTERISK                "\052"
1042    #define STR_PLUS                    "\053"
1043    #define STR_COMMA                   "\054"
1044    #define STR_MINUS                   "\055"
1045    #define STR_DOT                     "\056"
1046    #define STR_SLASH                   "\057"
1047    #define STR_0                       "\060"
1048    #define STR_1                       "\061"
1049    #define STR_2                       "\062"
1050    #define STR_3                       "\063"
1051    #define STR_4                       "\064"
1052    #define STR_5                       "\065"
1053    #define STR_6                       "\066"
1054    #define STR_7                       "\067"
1055    #define STR_8                       "\070"
1056    #define STR_9                       "\071"
1057    #define STR_COLON                   "\072"
1058    #define STR_SEMICOLON               "\073"
1059    #define STR_LESS_THAN_SIGN          "\074"
1060    #define STR_EQUALS_SIGN             "\075"
1061    #define STR_GREATER_THAN_SIGN       "\076"
1062    #define STR_QUESTION_MARK           "\077"
1063    #define STR_COMMERCIAL_AT           "\100"
1064    #define STR_A                       "\101"
1065    #define STR_B                       "\102"
1066    #define STR_C                       "\103"
1067    #define STR_D                       "\104"
1068    #define STR_E                       "\105"
1069    #define STR_F                       "\106"
1070    #define STR_G                       "\107"
1071    #define STR_H                       "\110"
1072    #define STR_I                       "\111"
1073    #define STR_J                       "\112"
1074    #define STR_K                       "\113"
1075    #define STR_L                       "\114"
1076    #define STR_M                       "\115"
1077    #define STR_N                       "\116"
1078    #define STR_O                       "\117"
1079    #define STR_P                       "\120"
1080    #define STR_Q                       "\121"
1081    #define STR_R                       "\122"
1082    #define STR_S                       "\123"
1083    #define STR_T                       "\124"
1084    #define STR_U                       "\125"
1085    #define STR_V                       "\126"
1086    #define STR_W                       "\127"
1087    #define STR_X                       "\130"
1088    #define STR_Y                       "\131"
1089    #define STR_Z                       "\132"
1090    #define STR_LEFT_SQUARE_BRACKET     "\133"
1091    #define STR_BACKSLASH               "\134"
1092    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1093    #define STR_CIRCUMFLEX_ACCENT       "\136"
1094    #define STR_UNDERSCORE              "\137"
1095    #define STR_GRAVE_ACCENT            "\140"
1096    #define STR_a                       "\141"
1097    #define STR_b                       "\142"
1098    #define STR_c                       "\143"
1099    #define STR_d                       "\144"
1100    #define STR_e                       "\145"
1101    #define STR_f                       "\146"
1102    #define STR_g                       "\147"
1103    #define STR_h                       "\150"
1104    #define STR_i                       "\151"
1105    #define STR_j                       "\152"
1106    #define STR_k                       "\153"
1107    #define STR_l                       "\154"
1108    #define STR_m                       "\155"
1109    #define STR_n                       "\156"
1110    #define STR_o                       "\157"
1111    #define STR_p                       "\160"
1112    #define STR_q                       "\161"
1113    #define STR_r                       "\162"
1114    #define STR_s                       "\163"
1115    #define STR_t                       "\164"
1116    #define STR_u                       "\165"
1117    #define STR_v                       "\166"
1118    #define STR_w                       "\167"
1119    #define STR_x                       "\170"
1120    #define STR_y                       "\171"
1121    #define STR_z                       "\172"
1122    #define STR_LEFT_CURLY_BRACKET      "\173"
1123    #define STR_VERTICAL_LINE           "\174"
1124    #define STR_RIGHT_CURLY_BRACKET     "\175"
1125    #define STR_TILDE                   "\176"
1126    
1127    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1128    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1129    #define STRING_F0                   STR_F "\0"
1130    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1131    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1132    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1133    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1134    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1135    
1136    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1137    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1138    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1139    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1140    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1141    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1142    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1143    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1144    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1145    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1146    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1147    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1148    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1149    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1150    
1151    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1152    
1153    #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS
1154    #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS
1155    #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1156    #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1157    #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1158    #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
1159    #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
1160    #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1161    
1162    #endif  /* SUPPORT_UTF8 */
1163    
1164  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1165    
1166  #ifndef ESC_e  #ifndef ESC_e
1167  #define ESC_e 27  #define ESC_e CHAR_ESC
1168  #endif  #endif
1169    
1170  #ifndef ESC_f  #ifndef ESC_f
1171  #define ESC_f '\f'  #define ESC_f CHAR_FF
1172  #endif  #endif
1173    
1174  #ifndef ESC_n  #ifndef ESC_n
1175  #define ESC_n '\n'  #define ESC_n CHAR_NL
1176  #endif  #endif
1177    
1178  #ifndef ESC_r  #ifndef ESC_r
1179  #define ESC_r '\r'  #define ESC_r CHAR_CR
1180  #endif  #endif
1181    
1182  /* 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
1183  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1184    
1185  #ifndef ESC_tee  #ifndef ESC_tee
1186  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1187  #endif  #endif
1188    
1189  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
# Line 598  contain UTF-8 characters with values gre Line 1210  contain UTF-8 characters with values gre
1210  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
1211  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
1212  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
1213  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
1214  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  (which is used for [^] in JavaScript compatibility mode).
1215  There are two tests in the code for an escape greater than ESC_b and less than  
1216  ESC_Z to detect the types that may be repeated. These are the types that  The final escape must be ESC_REF as subsequent values are used for
1217  consume characters. If any new escapes are put in between that don't consume a  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1218  character, that code will have to change. */  greater than ESC_b and less than ESC_Z to detect the types that may be
1219    repeated. These are the types that consume characters. If any new escapes are
1220    put in between that don't consume a character, that code will have to change.
1221    */
1222    
1223  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,
1224         ESC_W, ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H, ESC_h,         ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1225         ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_k, ESC_REF };         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
1226           ESC_REF };
1227    
1228    
1229  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to  
1230  OP_EOD must correspond in order to the list of escapes immediately above.  OP_EOD must correspond in order to the list of escapes immediately above.
1231    
 To keep stored, compiled patterns compatible, new opcodes should be added  
 immediately before OP_BRA, where (since release 7.0) a gap is left for this  
 purpose.  
   
1232  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1233  that follow must also be updated to match. There is also a table called  that follow must also be updated to match. There are also tables called
1234  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1235    
1236  enum {  enum {
1237    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 638  enum { Line 1249  enum {
1249    OP_WHITESPACE,         /*  9 \s */    OP_WHITESPACE,         /*  9 \s */
1250    OP_NOT_WORDCHAR,       /* 10 \W */    OP_NOT_WORDCHAR,       /* 10 \W */
1251    OP_WORDCHAR,           /* 11 \w */    OP_WORDCHAR,           /* 11 \w */
1252    OP_ANY,            /* 12 Match any character */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
1253    OP_ANYBYTE,        /* 13 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
1254    OP_NOTPROP,        /* 14 \P (not Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1255    OP_PROP,           /* 15 \p (Unicode property) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1256    OP_ANYNL,          /* 16 \R (any newline sequence) */    OP_PROP,           /* 16 \p (Unicode property) */
1257    OP_NOT_HSPACE,     /* 17 \H (not horizontal whitespace) */    OP_ANYNL,          /* 17 \R (any newline sequence) */
1258    OP_HSPACE,         /* 18 \h (horizontal whitespace) */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1259    OP_NOT_VSPACE,     /* 19 \V (not vertical whitespace) */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1260    OP_VSPACE,         /* 20 \v (vertical whitespace) */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1261    OP_EXTUNI,         /* 21 \X (extended Unicode sequence */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1262    OP_EODN,           /* 22 End of data or \n at end of data: \Z. */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1263    OP_EOD,            /* 23 End of data: \z */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1264      OP_EOD,            /* 24 End of data: \z */
1265    OP_OPT,            /* 24 Set runtime options */  
1266    OP_CIRC,           /* 25 Start of line - varies with multiline switch */    OP_OPT,            /* 25 Set runtime options */
1267    OP_DOLL,           /* 26 End of line - varies with multiline switch */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
1268    OP_CHAR,           /* 27 Match one character, casefully */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
1269    OP_CHARNC,         /* 28 Match one character, caselessly */    OP_CHAR,           /* 28 Match one character, casefully */
1270    OP_NOT,            /* 29 Match one character, not the following one */    OP_CHARNC,         /* 29 Match one character, caselessly */
1271      OP_NOT,            /* 30 Match one character, not the following one */
1272    OP_STAR,           /* 30 The maximizing and minimizing versions of */  
1273    OP_MINSTAR,        /* 31 these six opcodes must come in pairs, with */    OP_STAR,           /* 31 The maximizing and minimizing versions of */
1274    OP_PLUS,           /* 32 the minimizing one second. */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
1275    OP_MINPLUS,        /* 33 This first set applies to single characters.*/    OP_PLUS,           /* 33 the minimizing one second. */
1276    OP_QUERY,          /* 34 */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
1277    OP_MINQUERY,       /* 35 */    OP_QUERY,          /* 35 */
1278      OP_MINQUERY,       /* 36 */
1279    OP_UPTO,           /* 36 From 0 to n matches */  
1280    OP_MINUPTO,        /* 37 */    OP_UPTO,           /* 37 From 0 to n matches */
1281    OP_EXACT,          /* 38 Exactly n matches */    OP_MINUPTO,        /* 38 */
1282      OP_EXACT,          /* 39 Exactly n matches */
1283    OP_POSSTAR,        /* 39 Possessified star */  
1284    OP_POSPLUS,        /* 40 Possessified plus */    OP_POSSTAR,        /* 40 Possessified star */
1285    OP_POSQUERY,       /* 41 Posesssified query */    OP_POSPLUS,        /* 41 Possessified plus */
1286    OP_POSUPTO,        /* 42 Possessified upto */    OP_POSQUERY,       /* 42 Posesssified query */
1287      OP_POSUPTO,        /* 43 Possessified upto */
1288    OP_NOTSTAR,        /* 43 The maximizing and minimizing versions of */  
1289    OP_NOTMINSTAR,     /* 44 these six opcodes must come in pairs, with */    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
1290    OP_NOTPLUS,        /* 45 the minimizing one second. They must be in */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
1291    OP_NOTMINPLUS,     /* 46 exactly the same order as those above. */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
1292    OP_NOTQUERY,       /* 47 This set applies to "not" single characters. */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
1293    OP_NOTMINQUERY,    /* 48 */    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
1294      OP_NOTMINQUERY,    /* 49 */
1295    OP_NOTUPTO,        /* 49 From 0 to n matches */  
1296    OP_NOTMINUPTO,     /* 50 */    OP_NOTUPTO,        /* 50 From 0 to n matches */
1297    OP_NOTEXACT,       /* 51 Exactly n matches */    OP_NOTMINUPTO,     /* 51 */
1298      OP_NOTEXACT,       /* 52 Exactly n matches */
1299    OP_NOTPOSSTAR,     /* 52 Possessified versions */  
1300    OP_NOTPOSPLUS,     /* 53 */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
1301    OP_NOTPOSQUERY,    /* 54 */    OP_NOTPOSPLUS,     /* 54 */
1302    OP_NOTPOSUPTO,     /* 55 */    OP_NOTPOSQUERY,    /* 55 */
1303      OP_NOTPOSUPTO,     /* 56 */
1304    OP_TYPESTAR,       /* 56 The maximizing and minimizing versions of */  
1305    OP_TYPEMINSTAR,    /* 57 these six opcodes must come in pairs, with */    OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
1306    OP_TYPEPLUS,       /* 58 the minimizing one second. These codes must */    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
1307    OP_TYPEMINPLUS,    /* 59 be in exactly the same order as those above. */    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
1308    OP_TYPEQUERY,      /* 60 This set applies to character types such as \d */    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
1309    OP_TYPEMINQUERY,   /* 61 */    OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
1310      OP_TYPEMINQUERY,   /* 62 */
1311    OP_TYPEUPTO,       /* 62 From 0 to n matches */  
1312    OP_TYPEMINUPTO,    /* 63 */    OP_TYPEUPTO,       /* 63 From 0 to n matches */
1313    OP_TYPEEXACT,      /* 64 Exactly n matches */    OP_TYPEMINUPTO,    /* 64 */
1314      OP_TYPEEXACT,      /* 65 Exactly n matches */
1315    OP_TYPEPOSSTAR,    /* 65 Possessified versions */  
1316    OP_TYPEPOSPLUS,    /* 66 */    OP_TYPEPOSSTAR,    /* 66 Possessified versions */
1317    OP_TYPEPOSQUERY,   /* 67 */    OP_TYPEPOSPLUS,    /* 67 */
1318    OP_TYPEPOSUPTO,    /* 68 */    OP_TYPEPOSQUERY,   /* 68 */
1319      OP_TYPEPOSUPTO,    /* 69 */
1320    OP_CRSTAR,         /* 69 The maximizing and minimizing versions of */  
1321    OP_CRMINSTAR,      /* 70 all these opcodes must come in pairs, with */    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
1322    OP_CRPLUS,         /* 71 the minimizing one second. These codes must */    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
1323    OP_CRMINPLUS,      /* 72 be in exactly the same order as those above. */    OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
1324    OP_CRQUERY,        /* 73 These are for character classes and back refs */    OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
1325    OP_CRMINQUERY,     /* 74 */    OP_CRQUERY,        /* 74 These are for character classes and back refs */
1326    OP_CRRANGE,        /* 75 These are different to the three sets above. */    OP_CRMINQUERY,     /* 75 */
1327    OP_CRMINRANGE,     /* 76 */    OP_CRRANGE,        /* 76 These are different to the three sets above. */
1328      OP_CRMINRANGE,     /* 77 */
1329    
1330    OP_CLASS,          /* 77 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
1331    OP_NCLASS,         /* 78 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
1332                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
1333                             character > 255 is encountered. */                             character > 255 is encountered. */
1334    
1335    OP_XCLASS,         /* 79 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
1336                             class. This does both positive and negative. */                             class. This does both positive and negative. */
1337    
1338    OP_REF,            /* 80 Match a back reference */    OP_REF,            /* 81 Match a back reference */
1339    OP_RECURSE,        /* 81 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
1340    OP_CALLOUT,        /* 82 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
1341    
1342    OP_ALT,            /* 83 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
1343    OP_KET,            /* 84 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */
1344    OP_KETRMAX,        /* 85 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
1345    OP_KETRMIN,        /* 86 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
1346    
1347    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
1348    
1349    OP_ASSERT,         /* 87 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
1350    OP_ASSERT_NOT,     /* 88 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
1351    OP_ASSERTBACK,     /* 89 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
1352    OP_ASSERTBACK_NOT, /* 90 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
1353    OP_REVERSE,        /* 91 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
1354    
1355    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
1356    as there's a test for >= ONCE for a subpattern that isn't an assertion. */    as there's a test for >= ONCE for a subpattern that isn't an assertion. */
1357    
1358    OP_ONCE,           /* 92 Atomic group */    OP_ONCE,           /* 93 Atomic group */
1359    OP_BRA,            /* 83 Start of non-capturing bracket */    OP_BRA,            /* 94 Start of non-capturing bracket */
1360    OP_CBRA,           /* 94 Start of capturing bracket */    OP_CBRA,           /* 95 Start of capturing bracket */
1361    OP_COND,           /* 95 Conditional group */    OP_COND,           /* 96 Conditional group */
1362    
1363    /* These three must follow the previous three, in the same order. There's a    /* These three must follow the previous three, in the same order. There's a
1364    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1365    
1366    OP_SBRA,           /* 96 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
1367    OP_SCBRA,          /* 97 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1368    OP_SCOND,          /* 98 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
1369    
1370    OP_CREF,           /* 99 Used to hold a capture number as condition */    /* The next two pairs must (respectively) be kept together. */
1371    OP_RREF,           /* 100 Used to hold a recursion number as condition */  
1372    OP_DEF,            /* 101 The DEFINE condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
1373      OP_NCREF,          /* 101 Same, but generaged by a name reference*/
1374      OP_RREF,           /* 102 Used to hold a recursion number as condition */
1375      OP_NRREF,          /* 103 Same, but generaged by a name reference*/
1376      OP_DEF,            /* 104 The DEFINE condition */
1377    
1378    OP_BRAZERO,        /* 102 These two must remain together and in this */    OP_BRAZERO,        /* 105 These two must remain together and in this */
1379    OP_BRAMINZERO      /* 103 order. */    OP_BRAMINZERO,     /* 106 order. */
1380    
1381      /* These are backtracking control verbs */
1382    
1383      OP_MARK,           /* 107 always has an argument */
1384      OP_PRUNE,          /* 108 */
1385      OP_PRUNE_ARG,      /* 109 same, but with argument */
1386      OP_SKIP,           /* 110 */
1387      OP_SKIP_ARG,       /* 111 same, but with argument */
1388      OP_THEN,           /* 112 */
1389      OP_THEN_ARG,       /* 113 same, but with argument */
1390      OP_COMMIT,         /* 114 */
1391    
1392      /* These are forced failure and success verbs */
1393    
1394      OP_FAIL,           /* 115 */
1395      OP_ACCEPT,         /* 116 */
1396      OP_CLOSE,          /* 117 Used before OP_ACCEPT to close open captures */
1397    
1398      /* This is used to skip a subpattern with a {0} quantifier */
1399    
1400      OP_SKIPZERO,       /* 118 */
1401    
1402      /* This is not an opcode, but is used to check that tables indexed by opcode
1403      are the correct length, in order to catch updating errors - there have been
1404      some in the past. */
1405    
1406      OP_TABLE_LENGTH
1407  };  };
1408    
1409    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1410    definitions that follow must also be updated to match. There are also tables
1411    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1412    
1413    
1414  /* This macro defines textual names for all the opcodes. These are used only  /* This macro defines textual names for all the opcodes. These are used only
1415  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
1416    
1417  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1418    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1419    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1420    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1421    "extuni",  "\\Z", "\\z",                                        \    "extuni",  "\\Z", "\\z",                                        \
1422    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
# Line 783  for debugging. The macro is referenced o Line 1430  for debugging. The macro is referenced o
1430    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
1431    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1432    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1433    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1434    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1435      "Brazero", "Braminzero",                                        \
1436      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1437      "*THEN", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",                \
1438      "Close", "Skip zero"
1439    
1440    
1441  /* 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 800  in UTF-8 mode. The code that uses this t Line 1451  in UTF-8 mode. The code that uses this t
1451    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1452    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1453    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1454    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1455    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3,                          /* \P, \p                                 */ \
1456    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \    1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1457      1,                             /* \X                                     */ \
1458    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
1459    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1460    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 844  in UTF-8 mode. The code that uses this t Line 1496  in UTF-8 mode. The code that uses this t
1496    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1497    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1498    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1499    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1500    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1501    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1502    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1503      3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG,                */ \
1504      1, 3, 1, 3,                    /* SKIP, SKIP_ARG, THEN, THEN_ARG,        */ \
1505      1, 1, 1, 3, 1                  /* COMMIT, FAIL, ACCEPT, CLOSE, SKIPZERO  */
1506    
1507    
1508  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1509    condition. */
1510    
1511  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1512    
1513  /* Error code numbers. They are given names so that they can more easily be  /* Compile time error code numbers. They are given names so that they can more
1514  tracked. */  easily be tracked. When a new number is added, the table called eint in
1515    pcreposix.c must be updated. */
1516    
1517  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1518         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
1519         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
1520         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1521         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1522         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1523           ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERRCOUNT };
1524    
1525  /* 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
1526  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 878  Because people can now save and re-use c Line 1536  Because people can now save and re-use c
1536  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
1537  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
1538  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1539  NOTE NOTE NOTE:  NOTE NOTE NOTE
1540  */  */
1541    
1542  typedef struct real_pcre {  typedef struct real_pcre {
1543    pcre_uint32 magic_number;    pcre_uint32 magic_number;
1544    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1545    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
1546    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
1547      pcre_uint16 dummy1;             /* For future use */
1548    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
1549    pcre_uint16 top_backref;    pcre_uint16 top_backref;
1550    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 905  remark (see NOTE above) about extending Line 1563  remark (see NOTE above) about extending
1563    
1564  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1565    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1566    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1567    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1568      pcre_uint32 minlength;          /* Minimum subject length */
1569  } pcre_study_data;  } pcre_study_data;
1570    
1571    /* Structure for building a chain of open capturing subpatterns during
1572    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1573    encountered. This is also used to identify subpatterns that contain recursive
1574    back references to themselves, so that they can be made atomic. */
1575    
1576    typedef struct open_capitem {
1577      struct open_capitem *next;    /* Chain link */
1578      pcre_uint16 number;           /* Capture number */
1579      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1580    } open_capitem;
1581    
1582  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1583  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1584    
# Line 921  typedef struct compile_data { Line 1591  typedef struct compile_data {
1591    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1592    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1593    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1594      open_capitem *open_caps;      /* Chain of open capture items */
1595    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1596    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1597    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
1598    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
1599    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
1600      int  final_bracount;          /* Saved value after first pass */
1601    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
1602    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
1603    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
1604      int  external_flags;          /* External flag bits to be set */
1605    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1606    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
1607      BOOL check_lookbehind;        /* Lookbehinds need later checking */
1608    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1609    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1610    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 941  branches, for testing for left recursion Line 1615  branches, for testing for left recursion
1615    
1616  typedef struct branch_chain {  typedef struct branch_chain {
1617    struct branch_chain *outer;    struct branch_chain *outer;
1618    uschar *current;    uschar *current_branch;
1619  } branch_chain;  } branch_chain;
1620    
1621  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
# Line 951  typedef struct recursion_info { Line 1625  typedef struct recursion_info {
1625    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1626    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
1627    const uschar *after_call;     /* "Return value": points after the call in the expr */    const uschar *after_call;     /* "Return value": points after the call in the expr */
   USPTR save_start;             /* Old value of mstart */  
1628    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1629    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1630      int save_offset_top;          /* Current value of offset_top */
1631  } recursion_info;  } recursion_info;
1632    
1633  /* 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 978  typedef struct match_data { Line 1652  typedef struct match_data {
1652    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1653    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1654    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1655      int    name_count;            /* Number of names in name table */
1656      int    name_entry_size;       /* Size of entry in names table */
1657      uschar *name_table;           /* Table of names */
1658    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1659    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1660    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
# Line 985  typedef struct match_data { Line 1662  typedef struct match_data {
1662    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1663    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1664    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1665      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1666    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1667    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1668    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1669    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1670      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1671    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1672    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1673    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1674    USPTR  start_match_ptr;       /* Start of matched string */    USPTR  start_match_ptr;       /* Start of matched string */
1675    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1676      USPTR  start_used_ptr;        /* Earliest consulted character */
1677      int    partial;               /* PARTIAL options */
1678    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1679    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1680    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
# Line 1001  typedef struct match_data { Line 1682  typedef struct match_data {
1682    int    eptrn;                 /* Next free eptrblock */    int    eptrn;                 /* Next free eptrblock */
1683    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1684    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1685      const uschar *mark;           /* Mark pointer to pass back */
1686  } match_data;  } match_data;
1687    
1688  /* A similar structure is used for the same purpose by the DFA matching  /* A similar structure is used for the same purpose by the DFA matching
# Line 1010  typedef struct dfa_match_data { Line 1692  typedef struct dfa_match_data {
1692    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;     /* Start of the compiled pattern */
1693    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
1694    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;    /* End of subject string */
1695      const uschar *start_used_ptr; /* Earliest consulted character */
1696    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1697      int   start_offset;           /* The start offset value */
1698    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1699    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1700    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
# Line 1025  typedef struct dfa_match_data { Line 1709  typedef struct dfa_match_data {
1709  #define ctype_letter  0x02  #define ctype_letter  0x02
1710  #define ctype_digit   0x04  #define ctype_digit   0x04
1711  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1712  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1713  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1714    
1715  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set  /* Offsets for the bitmap tables in pcre_cbits. Each table contains a set
# Line 1053  total length. */ Line 1737  total length. */
1737  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1738    
1739  /* Layout of the UCP type table that translates property names into types and  /* Layout of the UCP type table that translates property names into types and
1740  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1741    relocations in shared libraries, it now has an offset into a single string
1742    instead. */
1743    
1744  typedef struct {  typedef struct {
1745    const char *name;    pcre_uint16 name_offset;
1746    pcre_uint16 type;    pcre_uint16 type;
1747    pcre_uint16 value;    pcre_uint16 value;
1748  } ucp_type_table;  } ucp_type_table;
# Line 1074  extern const uschar _pcre_utf8_table4[]; Line 1760  extern const uschar _pcre_utf8_table4[];
1760    
1761  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1762    
1763    extern const char   _pcre_utt_names[];
1764  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1765  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1766    
# Line 1086  extern const uschar _pcre_OP_lengths[]; Line 1773  extern const uschar _pcre_OP_lengths[];
1773  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
1774  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1775    
1776  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1777                        int *, BOOL);  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);
1778  extern int          _pcre_ord2utf8(int, uschar *);  extern int           _pcre_ord2utf8(int, uschar *);
1779  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1780                        const pcre_study_data *, pcre_study_data *);                         const pcre_study_data *, pcre_study_data *);
1781  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  extern int           _pcre_valid_utf8(USPTR, int);
1782  extern unsigned int _pcre_ucp_othercase(const unsigned int);  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);
1783  extern int          _pcre_valid_utf8(const uschar *, int);  extern BOOL          _pcre_xclass(int, const uschar *);
1784  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,  
1785                        int *, BOOL);  
1786  extern BOOL         _pcre_xclass(int, const uschar *);  /* Unicode character database (UCD) */
1787    
1788    typedef struct {
1789      uschar script;
1790      uschar chartype;
1791      pcre_int32 other_case;
1792    } ucd_record;
1793    
1794    extern const ucd_record  _pcre_ucd_records[];
1795    extern const uschar      _pcre_ucd_stage1[];
1796    extern const pcre_uint16 _pcre_ucd_stage2[];
1797    extern const int         _pcre_ucp_gentype[];
1798    
1799    
1800    /* UCD access macros */
1801    
1802    #define UCD_BLOCK_SIZE 128
1803    #define GET_UCD(ch) (_pcre_ucd_records + \
1804            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1805            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1806    
1807    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1808    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1809    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1810    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1811    
1812  #endif  #endif
1813    

Legend:
Removed from v.178  
changed lines
  Added in v.512

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12