/[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 137 by ph10, Thu Mar 29 13:56:00 2007 UTC revision 462 by ph10, Sat Oct 17 19:55:02 2009 UTC
# Line 7  Line 7 
7  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2007 University of Cambridge             Copyright (c) 1997-2009 University of Cambridge
11    
12  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
13  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 51  functions whose names all begin with "_p Line 51  functions whose names all begin with "_p
51  #define DEBUG  #define DEBUG
52  #endif  #endif
53    
54    /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"
55    script prevents both being selected, but not everybody uses "configure". */
56    
57    #if defined EBCDIC && defined SUPPORT_UTF8
58    #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.
59    #endif
60    
61    /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
62    "configure" script ensures this, but not everybody uses "configure". */
63    
64    #if defined SUPPORT_UCP && !defined SUPPORT_UTF8
65    #define SUPPORT_UTF8 1
66    #endif
67    
68  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef
69  inline, and there are *still* stupid compilers about that don't like indented  inline, and there are *still* stupid compilers about that don't like indented
70  pre-processor statements, or at least there were when I first wrote this. After  pre-processor statements, or at least there were when I first wrote this. After
# Line 67  be absolutely sure we get our version. * Line 81  be absolutely sure we get our version. *
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    
# Line 83  setjmp and stdarg are used is when NO_RE Line 93  setjmp and stdarg are used is when NO_RE
93  #include <stdlib.h>  #include <stdlib.h>
94  #include <string.h>  #include <string.h>
95    
96  #ifndef PCRE_SPY  /* When compiling a DLL for Windows, the exported symbols have to be declared
97  #define PCRE_DEFINITION       /* Win32 __declspec(export) trigger for .dll */  using some MS magic. I found some useful information on this web page:
98    http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
99    information there, using __declspec(dllexport) without "extern" we have a
100    definition; with "extern" we have a declaration. The settings here override the
101    setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
102    which is all that is needed for applications (they just import the symbols). We
103    use:
104    
105      PCRE_EXP_DECL       for declarations
106      PCRE_EXP_DEFN       for definitions of exported functions
107      PCRE_EXP_DATA_DEFN  for definitions of exported variables
108    
109    The reason for the two DEFN macros is that in non-Windows environments, one
110    does not want to have "extern" before variable definitions because it leads to
111    compiler warnings. So we distinguish between functions and variables. In
112    Windows, the two should always be the same.
113    
114    The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
115    which is an application, but needs to import this file in order to "peek" at
116    internals, can #include pcre.h first to get an application's-eye view.
117    
118    In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
119    special-purpose environments) might want to stick other stuff in front of
120    exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
121    PCRE_EXP_DATA_DEFN only if they are not already set. */
122    
123    #ifndef PCRE_EXP_DECL
124    #  ifdef _WIN32
125    #    ifndef PCRE_STATIC
126    #      define PCRE_EXP_DECL       extern __declspec(dllexport)
127    #      define PCRE_EXP_DEFN       __declspec(dllexport)
128    #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
129    #    else
130    #      define PCRE_EXP_DECL       extern
131    #      define PCRE_EXP_DEFN
132    #      define PCRE_EXP_DATA_DEFN
133    #    endif
134    #  else
135    #    ifdef __cplusplus
136    #      define PCRE_EXP_DECL       extern "C"
137    #    else
138    #      define PCRE_EXP_DECL       extern
139    #    endif
140    #    ifndef PCRE_EXP_DEFN
141    #      define PCRE_EXP_DEFN       PCRE_EXP_DECL
142    #    endif
143    #    ifndef PCRE_EXP_DATA_DEFN
144    #      define PCRE_EXP_DATA_DEFN
145    #    endif
146    #  endif
147    #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  #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
# Line 95  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 125  characters only go up to 0x7fffffff (tho Line 202  characters only go up to 0x7fffffff (tho
202  #define NOTACHAR 0xffffffff  #define NOTACHAR 0xffffffff
203    
204  /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,  /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
205  and "all" at present). The following macros are used to package up testing for  "any" and "anycrlf" at present). The following macros are used to package up
206  newlines. NLBLOCK, PSSTART, and PSEND are defined in the various modules to  testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
207  indicate in which datablock the parameters exist, and what the start/end of  modules to indicate in which datablock the parameters exist, and what the
208  string field names are. */  start/end of string field names are. */
209    
210  #define NLTYPE_FIXED   0     /* Newline is a fixed length string */  #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
211  #define NLTYPE_ANY     1     /* Newline is any Unicode line ending */  #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
212    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
213    
214  /* This macro checks for a newline at the given position */  /* This macro checks for a newline at the given position */
215    
216  #define IS_NEWLINE(p) \  #define IS_NEWLINE(p) \
217    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
218      ((p) < NLBLOCK->PSEND && \      ((p) < NLBLOCK->PSEND && \
219       _pcre_is_newline((p), NLBLOCK->PSEND, &(NLBLOCK->nllen), utf8) \       _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
220      ) \         utf8)) \
221      : \      : \
222      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
223       (p)[0] == NLBLOCK->nl[0] && \       (p)[0] == NLBLOCK->nl[0] && \
# Line 152  string field names are. */ Line 230  string field names are. */
230  #define WAS_NEWLINE(p) \  #define WAS_NEWLINE(p) \
231    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
232      ((p) > NLBLOCK->PSSTART && \      ((p) > NLBLOCK->PSSTART && \
233       _pcre_was_newline((p), NLBLOCK->PSSTART, &(NLBLOCK->nllen), utf8) \       _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
234      ) \         &(NLBLOCK->nllen), utf8)) \
235      : \      : \
236      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
237       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
# Line 178  must begin with PCRE_. */ Line 256  must begin with PCRE_. */
256  #define USPTR const unsigned char *  #define USPTR const unsigned char *
257  #endif  #endif
258    
259    
260    
261  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
262  values. */  values. */
263    
264  #include <pcre.h>  #include "pcre.h"
265  #include "ucp.h"  #include "ucp.h"
266    
267  /* When compiling for use with the Virtual Pascal compiler, these functions  /* When compiling for use with the Virtual Pascal compiler, these functions
# Line 310  capturing parenthesis numbers in back re Line 390  capturing parenthesis numbers in back re
390    
391  /* 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
392  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
393  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
394    never be called in byte mode. To make sure it can never even appear when UTF-8
395    support is omitted, we don't even define it. */
396    
397  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
398  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 318  byte-mode, and more complicated ones for Line 400  byte-mode, and more complicated ones for
400  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
401  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
402  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
403  #define BACKCHAR(eptr)  /* #define BACKCHAR(eptr) */
404    
405  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
406    
# Line 410  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. 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
517    because almost all calls are already within a block of UTF-8 only code. */
518    
519  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
520    
521  #endif  #endif
522    
# Line 430  Standard C system should have one. */ Line 533  Standard C system should have one. */
533    
534  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
535    
536  /* Private options flags start at the most significant end of the four bytes.  /* Private flags containing information about the compiled regex. They used to
537  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
538  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
539  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
540  bits. */  compatibility. */
541    
542  #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
543  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
544  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
545  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
546  #define PCRE_JCHANGED      0x08000000  /* j option changes within regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
547    #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
548    
549  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
550    
551  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
552    #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */
553    
554  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
555  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
556    
557  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY)  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
558                               PCRE_NEWLINE_ANYCRLF)
559    
560  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
561    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
562     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
563     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
564     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
565       PCRE_JAVASCRIPT_COMPAT)
566    
567  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
568    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
569     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
570       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
571    
572  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
573    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
574     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
575       PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
576       PCRE_NO_START_OPTIMIZE)
577    
578  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
579    
# Line 488  variable-length repeat, or a anything ot Line 598  variable-length repeat, or a anything ot
598  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
599  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
600    
601  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
602    environments where these macros are defined elsewhere. Unfortunately, there
603    is no way to do the same for the typedef. */
604    
605  typedef int BOOL;  typedef int BOOL;
606    
607    #ifndef FALSE
608  #define FALSE   0  #define FALSE   0
609  #define TRUE    1  #define TRUE    1
610    #endif
611    
612    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
613    character constants like '*' because the compiler would emit their EBCDIC code,
614    which is different from their ASCII/UTF-8 code. Instead we define macros for
615    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
616    is enabled. When UTF-8 support is not enabled, the definitions use character
617    literals. Both character and string versions of each character are needed, and
618    there are some longer strings as well.
619    
620    This means that, on EBCDIC platforms, the PCRE library can handle either
621    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
622    would need different lookups depending on whether PCRE_UTF8 was set or not.
623    This would make it impossible to use characters in switch/case statements,
624    which would reduce performance. For a theoretical use (which nobody has asked
625    for) in a minority area (EBCDIC platforms), this is not sensible. Any
626    application that did need both could compile two versions of the library, using
627    macros to give the functions distinct names. */
628    
629    #ifndef SUPPORT_UTF8
630    
631    /* UTF-8 support is not enabled; use the platform-dependent character literals
632    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
633    
634    #define CHAR_HT                     '\t'
635    #define CHAR_VT                     '\v'
636    #define CHAR_FF                     '\f'
637    #define CHAR_CR                     '\r'
638    #define CHAR_NL                     '\n'
639    #define CHAR_BS                     '\b'
640    #define CHAR_BEL                    '\a'
641    #ifdef EBCDIC
642    #define CHAR_ESC                    '\047'
643    #define CHAR_DEL                    '\007'
644    #else
645    #define CHAR_ESC                    '\033'
646    #define CHAR_DEL                    '\177'
647    #endif
648    
649    #define CHAR_SPACE                  ' '
650    #define CHAR_EXCLAMATION_MARK       '!'
651    #define CHAR_QUOTATION_MARK         '"'
652    #define CHAR_NUMBER_SIGN            '#'
653    #define CHAR_DOLLAR_SIGN            '$'
654    #define CHAR_PERCENT_SIGN           '%'
655    #define CHAR_AMPERSAND              '&'
656    #define CHAR_APOSTROPHE             '\''
657    #define CHAR_LEFT_PARENTHESIS       '('
658    #define CHAR_RIGHT_PARENTHESIS      ')'
659    #define CHAR_ASTERISK               '*'
660    #define CHAR_PLUS                   '+'
661    #define CHAR_COMMA                  ','
662    #define CHAR_MINUS                  '-'
663    #define CHAR_DOT                    '.'
664    #define CHAR_SLASH                  '/'
665    #define CHAR_0                      '0'
666    #define CHAR_1                      '1'
667    #define CHAR_2                      '2'
668    #define CHAR_3                      '3'
669    #define CHAR_4                      '4'
670    #define CHAR_5                      '5'
671    #define CHAR_6                      '6'
672    #define CHAR_7                      '7'
673    #define CHAR_8                      '8'
674    #define CHAR_9                      '9'
675    #define CHAR_COLON                  ':'
676    #define CHAR_SEMICOLON              ';'
677    #define CHAR_LESS_THAN_SIGN         '<'
678    #define CHAR_EQUALS_SIGN            '='
679    #define CHAR_GREATER_THAN_SIGN      '>'
680    #define CHAR_QUESTION_MARK          '?'
681    #define CHAR_COMMERCIAL_AT          '@'
682    #define CHAR_A                      'A'
683    #define CHAR_B                      'B'
684    #define CHAR_C                      'C'
685    #define CHAR_D                      'D'
686    #define CHAR_E                      'E'
687    #define CHAR_F                      'F'
688    #define CHAR_G                      'G'
689    #define CHAR_H                      'H'
690    #define CHAR_I                      'I'
691    #define CHAR_J                      'J'
692    #define CHAR_K                      'K'
693    #define CHAR_L                      'L'
694    #define CHAR_M                      'M'
695    #define CHAR_N                      'N'
696    #define CHAR_O                      'O'
697    #define CHAR_P                      'P'
698    #define CHAR_Q                      'Q'
699    #define CHAR_R                      'R'
700    #define CHAR_S                      'S'
701    #define CHAR_T                      'T'
702    #define CHAR_U                      'U'
703    #define CHAR_V                      'V'
704    #define CHAR_W                      'W'
705    #define CHAR_X                      'X'
706    #define CHAR_Y                      'Y'
707    #define CHAR_Z                      'Z'
708    #define CHAR_LEFT_SQUARE_BRACKET    '['
709    #define CHAR_BACKSLASH              '\\'
710    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
711    #define CHAR_CIRCUMFLEX_ACCENT      '^'
712    #define CHAR_UNDERSCORE             '_'
713    #define CHAR_GRAVE_ACCENT           '`'
714    #define CHAR_a                      'a'
715    #define CHAR_b                      'b'
716    #define CHAR_c                      'c'
717    #define CHAR_d                      'd'
718    #define CHAR_e                      'e'
719    #define CHAR_f                      'f'
720    #define CHAR_g                      'g'
721    #define CHAR_h                      'h'
722    #define CHAR_i                      'i'
723    #define CHAR_j                      'j'
724    #define CHAR_k                      'k'
725    #define CHAR_l                      'l'
726    #define CHAR_m                      'm'
727    #define CHAR_n                      'n'
728    #define CHAR_o                      'o'
729    #define CHAR_p                      'p'
730    #define CHAR_q                      'q'
731    #define CHAR_r                      'r'
732    #define CHAR_s                      's'
733    #define CHAR_t                      't'
734    #define CHAR_u                      'u'
735    #define CHAR_v                      'v'
736    #define CHAR_w                      'w'
737    #define CHAR_x                      'x'
738    #define CHAR_y                      'y'
739    #define CHAR_z                      'z'
740    #define CHAR_LEFT_CURLY_BRACKET     '{'
741    #define CHAR_VERTICAL_LINE          '|'
742    #define CHAR_RIGHT_CURLY_BRACKET    '}'
743    #define CHAR_TILDE                  '~'
744    
745    #define STR_HT                      "\t"
746    #define STR_VT                      "\v"
747    #define STR_FF                      "\f"
748    #define STR_CR                      "\r"
749    #define STR_NL                      "\n"
750    #define STR_BS                      "\b"
751    #define STR_BEL                     "\a"
752    #ifdef EBCDIC
753    #define STR_ESC                     "\047"
754    #define STR_DEL                     "\007"
755    #else
756    #define STR_ESC                     "\033"
757    #define STR_DEL                     "\177"
758    #endif
759    
760    #define STR_SPACE                   " "
761    #define STR_EXCLAMATION_MARK        "!"
762    #define STR_QUOTATION_MARK          "\""
763    #define STR_NUMBER_SIGN             "#"
764    #define STR_DOLLAR_SIGN             "$"
765    #define STR_PERCENT_SIGN            "%"
766    #define STR_AMPERSAND               "&"
767    #define STR_APOSTROPHE              "'"
768    #define STR_LEFT_PARENTHESIS        "("
769    #define STR_RIGHT_PARENTHESIS       ")"
770    #define STR_ASTERISK                "*"
771    #define STR_PLUS                    "+"
772    #define STR_COMMA                   ","
773    #define STR_MINUS                   "-"
774    #define STR_DOT                     "."
775    #define STR_SLASH                   "/"
776    #define STR_0                       "0"
777    #define STR_1                       "1"
778    #define STR_2                       "2"
779    #define STR_3                       "3"
780    #define STR_4                       "4"
781    #define STR_5                       "5"
782    #define STR_6                       "6"
783    #define STR_7                       "7"
784    #define STR_8                       "8"
785    #define STR_9                       "9"
786    #define STR_COLON                   ":"
787    #define STR_SEMICOLON               ";"
788    #define STR_LESS_THAN_SIGN          "<"
789    #define STR_EQUALS_SIGN             "="
790    #define STR_GREATER_THAN_SIGN       ">"
791    #define STR_QUESTION_MARK           "?"
792    #define STR_COMMERCIAL_AT           "@"
793    #define STR_A                       "A"
794    #define STR_B                       "B"
795    #define STR_C                       "C"
796    #define STR_D                       "D"
797    #define STR_E                       "E"
798    #define STR_F                       "F"
799    #define STR_G                       "G"
800    #define STR_H                       "H"
801    #define STR_I                       "I"
802    #define STR_J                       "J"
803    #define STR_K                       "K"
804    #define STR_L                       "L"
805    #define STR_M                       "M"
806    #define STR_N                       "N"
807    #define STR_O                       "O"
808    #define STR_P                       "P"
809    #define STR_Q                       "Q"
810    #define STR_R                       "R"
811    #define STR_S                       "S"
812    #define STR_T                       "T"
813    #define STR_U                       "U"
814    #define STR_V                       "V"
815    #define STR_W                       "W"
816    #define STR_X                       "X"
817    #define STR_Y                       "Y"
818    #define STR_Z                       "Z"
819    #define STR_LEFT_SQUARE_BRACKET     "["
820    #define STR_BACKSLASH               "\\"
821    #define STR_RIGHT_SQUARE_BRACKET    "]"
822    #define STR_CIRCUMFLEX_ACCENT       "^"
823    #define STR_UNDERSCORE              "_"
824    #define STR_GRAVE_ACCENT            "`"
825    #define STR_a                       "a"
826    #define STR_b                       "b"
827    #define STR_c                       "c"
828    #define STR_d                       "d"
829    #define STR_e                       "e"
830    #define STR_f                       "f"
831    #define STR_g                       "g"
832    #define STR_h                       "h"
833    #define STR_i                       "i"
834    #define STR_j                       "j"
835    #define STR_k                       "k"
836    #define STR_l                       "l"
837    #define STR_m                       "m"
838    #define STR_n                       "n"
839    #define STR_o                       "o"
840    #define STR_p                       "p"
841    #define STR_q                       "q"
842    #define STR_r                       "r"
843    #define STR_s                       "s"
844    #define STR_t                       "t"
845    #define STR_u                       "u"
846    #define STR_v                       "v"
847    #define STR_w                       "w"
848    #define STR_x                       "x"
849    #define STR_y                       "y"
850    #define STR_z                       "z"
851    #define STR_LEFT_CURLY_BRACKET      "{"
852    #define STR_VERTICAL_LINE           "|"
853    #define STR_RIGHT_CURLY_BRACKET     "}"
854    #define STR_TILDE                   "~"
855    
856    #define STRING_ACCEPT0              "ACCEPT\0"
857    #define STRING_COMMIT0              "COMMIT\0"
858    #define STRING_F0                   "F\0"
859    #define STRING_FAIL0                "FAIL\0"
860    #define STRING_PRUNE0               "PRUNE\0"
861    #define STRING_SKIP0                "SKIP\0"
862    #define STRING_THEN                 "THEN"
863    
864    #define STRING_alpha0               "alpha\0"
865    #define STRING_lower0               "lower\0"
866    #define STRING_upper0               "upper\0"
867    #define STRING_alnum0               "alnum\0"
868    #define STRING_ascii0               "ascii\0"
869    #define STRING_blank0               "blank\0"
870    #define STRING_cntrl0               "cntrl\0"
871    #define STRING_digit0               "digit\0"
872    #define STRING_graph0               "graph\0"
873    #define STRING_print0               "print\0"
874    #define STRING_punct0               "punct\0"
875    #define STRING_space0               "space\0"
876    #define STRING_word0                "word\0"
877    #define STRING_xdigit               "xdigit"
878    
879    #define STRING_DEFINE               "DEFINE"
880    
881    #define STRING_CR_RIGHTPAR          "CR)"
882    #define STRING_LF_RIGHTPAR          "LF)"
883    #define STRING_CRLF_RIGHTPAR        "CRLF)"
884    #define STRING_ANY_RIGHTPAR         "ANY)"
885    #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"
886    #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
887    #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
888    #define STRING_UTF8_RIGHTPAR        "UTF8)"
889    
890    #else  /* SUPPORT_UTF8 */
891    
892    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
893    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
894    only. */
895    
896    #define CHAR_HT                     '\011'
897    #define CHAR_VT                     '\013'
898    #define CHAR_FF                     '\014'
899    #define CHAR_CR                     '\015'
900    #define CHAR_NL                     '\012'
901    #define CHAR_BS                     '\010'
902    #define CHAR_BEL                    '\007'
903    #define CHAR_ESC                    '\033'
904    #define CHAR_DEL                    '\177'
905    
906    #define CHAR_SPACE                  '\040'
907    #define CHAR_EXCLAMATION_MARK       '\041'
908    #define CHAR_QUOTATION_MARK         '\042'
909    #define CHAR_NUMBER_SIGN            '\043'
910    #define CHAR_DOLLAR_SIGN            '\044'
911    #define CHAR_PERCENT_SIGN           '\045'
912    #define CHAR_AMPERSAND              '\046'
913    #define CHAR_APOSTROPHE             '\047'
914    #define CHAR_LEFT_PARENTHESIS       '\050'
915    #define CHAR_RIGHT_PARENTHESIS      '\051'
916    #define CHAR_ASTERISK               '\052'
917    #define CHAR_PLUS                   '\053'
918    #define CHAR_COMMA                  '\054'
919    #define CHAR_MINUS                  '\055'
920    #define CHAR_DOT                    '\056'
921    #define CHAR_SLASH                  '\057'
922    #define CHAR_0                      '\060'
923    #define CHAR_1                      '\061'
924    #define CHAR_2                      '\062'
925    #define CHAR_3                      '\063'
926    #define CHAR_4                      '\064'
927    #define CHAR_5                      '\065'
928    #define CHAR_6                      '\066'
929    #define CHAR_7                      '\067'
930    #define CHAR_8                      '\070'
931    #define CHAR_9                      '\071'
932    #define CHAR_COLON                  '\072'
933    #define CHAR_SEMICOLON              '\073'
934    #define CHAR_LESS_THAN_SIGN         '\074'
935    #define CHAR_EQUALS_SIGN            '\075'
936    #define CHAR_GREATER_THAN_SIGN      '\076'
937    #define CHAR_QUESTION_MARK          '\077'
938    #define CHAR_COMMERCIAL_AT          '\100'
939    #define CHAR_A                      '\101'
940    #define CHAR_B                      '\102'
941    #define CHAR_C                      '\103'
942    #define CHAR_D                      '\104'
943    #define CHAR_E                      '\105'
944    #define CHAR_F                      '\106'
945    #define CHAR_G                      '\107'
946    #define CHAR_H                      '\110'
947    #define CHAR_I                      '\111'
948    #define CHAR_J                      '\112'
949    #define CHAR_K                      '\113'
950    #define CHAR_L                      '\114'
951    #define CHAR_M                      '\115'
952    #define CHAR_N                      '\116'
953    #define CHAR_O                      '\117'
954    #define CHAR_P                      '\120'
955    #define CHAR_Q                      '\121'
956    #define CHAR_R                      '\122'
957    #define CHAR_S                      '\123'
958    #define CHAR_T                      '\124'
959    #define CHAR_U                      '\125'
960    #define CHAR_V                      '\126'
961    #define CHAR_W                      '\127'
962    #define CHAR_X                      '\130'
963    #define CHAR_Y                      '\131'
964    #define CHAR_Z                      '\132'
965    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
966    #define CHAR_BACKSLASH              '\134'
967    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
968    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
969    #define CHAR_UNDERSCORE             '\137'
970    #define CHAR_GRAVE_ACCENT           '\140'
971    #define CHAR_a                      '\141'
972    #define CHAR_b                      '\142'
973    #define CHAR_c                      '\143'
974    #define CHAR_d                      '\144'
975    #define CHAR_e                      '\145'
976    #define CHAR_f                      '\146'
977    #define CHAR_g                      '\147'
978    #define CHAR_h                      '\150'
979    #define CHAR_i                      '\151'
980    #define CHAR_j                      '\152'
981    #define CHAR_k                      '\153'
982    #define CHAR_l                      '\154'
983    #define CHAR_m                      '\155'
984    #define CHAR_n                      '\156'
985    #define CHAR_o                      '\157'
986    #define CHAR_p                      '\160'
987    #define CHAR_q                      '\161'
988    #define CHAR_r                      '\162'
989    #define CHAR_s                      '\163'
990    #define CHAR_t                      '\164'
991    #define CHAR_u                      '\165'
992    #define CHAR_v                      '\166'
993    #define CHAR_w                      '\167'
994    #define CHAR_x                      '\170'
995    #define CHAR_y                      '\171'
996    #define CHAR_z                      '\172'
997    #define CHAR_LEFT_CURLY_BRACKET     '\173'
998    #define CHAR_VERTICAL_LINE          '\174'
999    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1000    #define CHAR_TILDE                  '\176'
1001    
1002    #define STR_HT                      "\011"
1003    #define STR_VT                      "\013"
1004    #define STR_FF                      "\014"
1005    #define STR_CR                      "\015"
1006    #define STR_NL                      "\012"
1007    #define STR_BS                      "\010"
1008    #define STR_BEL                     "\007"
1009    #define STR_ESC                     "\033"
1010    #define STR_DEL                     "\177"
1011    
1012    #define STR_SPACE                   "\040"
1013    #define STR_EXCLAMATION_MARK        "\041"
1014    #define STR_QUOTATION_MARK          "\042"
1015    #define STR_NUMBER_SIGN             "\043"
1016    #define STR_DOLLAR_SIGN             "\044"
1017    #define STR_PERCENT_SIGN            "\045"
1018    #define STR_AMPERSAND               "\046"
1019    #define STR_APOSTROPHE              "\047"
1020    #define STR_LEFT_PARENTHESIS        "\050"
1021    #define STR_RIGHT_PARENTHESIS       "\051"
1022    #define STR_ASTERISK                "\052"
1023    #define STR_PLUS                    "\053"
1024    #define STR_COMMA                   "\054"
1025    #define STR_MINUS                   "\055"
1026    #define STR_DOT                     "\056"
1027    #define STR_SLASH                   "\057"
1028    #define STR_0                       "\060"
1029    #define STR_1                       "\061"
1030    #define STR_2                       "\062"
1031    #define STR_3                       "\063"
1032    #define STR_4                       "\064"
1033    #define STR_5                       "\065"
1034    #define STR_6                       "\066"
1035    #define STR_7                       "\067"
1036    #define STR_8                       "\070"
1037    #define STR_9                       "\071"
1038    #define STR_COLON                   "\072"
1039    #define STR_SEMICOLON               "\073"
1040    #define STR_LESS_THAN_SIGN          "\074"
1041    #define STR_EQUALS_SIGN             "\075"
1042    #define STR_GREATER_THAN_SIGN       "\076"
1043    #define STR_QUESTION_MARK           "\077"
1044    #define STR_COMMERCIAL_AT           "\100"
1045    #define STR_A                       "\101"
1046    #define STR_B                       "\102"
1047    #define STR_C                       "\103"
1048    #define STR_D                       "\104"
1049    #define STR_E                       "\105"
1050    #define STR_F                       "\106"
1051    #define STR_G                       "\107"
1052    #define STR_H                       "\110"
1053    #define STR_I                       "\111"
1054    #define STR_J                       "\112"
1055    #define STR_K                       "\113"
1056    #define STR_L                       "\114"
1057    #define STR_M                       "\115"
1058    #define STR_N                       "\116"
1059    #define STR_O                       "\117"
1060    #define STR_P                       "\120"
1061    #define STR_Q                       "\121"
1062    #define STR_R                       "\122"
1063    #define STR_S                       "\123"
1064    #define STR_T                       "\124"
1065    #define STR_U                       "\125"
1066    #define STR_V                       "\126"
1067    #define STR_W                       "\127"
1068    #define STR_X                       "\130"
1069    #define STR_Y                       "\131"
1070    #define STR_Z                       "\132"
1071    #define STR_LEFT_SQUARE_BRACKET     "\133"
1072    #define STR_BACKSLASH               "\134"
1073    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1074    #define STR_CIRCUMFLEX_ACCENT       "\136"
1075    #define STR_UNDERSCORE              "\137"
1076    #define STR_GRAVE_ACCENT            "\140"
1077    #define STR_a                       "\141"
1078    #define STR_b                       "\142"
1079    #define STR_c                       "\143"
1080    #define STR_d                       "\144"
1081    #define STR_e                       "\145"
1082    #define STR_f                       "\146"
1083    #define STR_g                       "\147"
1084    #define STR_h                       "\150"
1085    #define STR_i                       "\151"
1086    #define STR_j                       "\152"
1087    #define STR_k                       "\153"
1088    #define STR_l                       "\154"
1089    #define STR_m                       "\155"
1090    #define STR_n                       "\156"
1091    #define STR_o                       "\157"
1092    #define STR_p                       "\160"
1093    #define STR_q                       "\161"
1094    #define STR_r                       "\162"
1095    #define STR_s                       "\163"
1096    #define STR_t                       "\164"
1097    #define STR_u                       "\165"
1098    #define STR_v                       "\166"
1099    #define STR_w                       "\167"
1100    #define STR_x                       "\170"
1101    #define STR_y                       "\171"
1102    #define STR_z                       "\172"
1103    #define STR_LEFT_CURLY_BRACKET      "\173"
1104    #define STR_VERTICAL_LINE           "\174"
1105    #define STR_RIGHT_CURLY_BRACKET     "\175"
1106    #define STR_TILDE                   "\176"
1107    
1108    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1109    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1110    #define STRING_F0                   STR_F "\0"
1111    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1112    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1113    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1114    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1115    
1116    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1117    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1118    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1119    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1120    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1121    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1122    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1123    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1124    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1125    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1126    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1127    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1128    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1129    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1130    
1131    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1132    
1133    #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS
1134    #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS
1135    #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1136    #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1137    #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1138    #define STRING_BSR_ANYCRLF_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1139    #define STRING_BSR_UNICODE_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS
1140    #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1141    
1142    #endif  /* SUPPORT_UTF8 */
1143    
1144  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1145    
1146  #ifndef ESC_e  #ifndef ESC_e
1147  #define ESC_e 27  #define ESC_e CHAR_ESC
1148  #endif  #endif
1149    
1150  #ifndef ESC_f  #ifndef ESC_f
1151  #define ESC_f '\f'  #define ESC_f CHAR_FF
1152  #endif  #endif
1153    
1154  #ifndef ESC_n  #ifndef ESC_n
1155  #define ESC_n '\n'  #define ESC_n CHAR_NL
1156  #endif  #endif
1157    
1158  #ifndef ESC_r  #ifndef ESC_r
1159  #define ESC_r '\r'  #define ESC_r CHAR_CR
1160  #endif  #endif
1161    
1162  /* We can't officially use ESC_t because it is a POSIX reserved identifier  /* We can't officially use ESC_t because it is a POSIX reserved identifier
1163  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1164    
1165  #ifndef ESC_tee  #ifndef ESC_tee
1166  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1167  #endif  #endif
1168    
1169  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
# Line 544  contain UTF-8 characters with values gre Line 1190  contain UTF-8 characters with values gre
1190  value such as \n. They must have non-zero values, as check_escape() returns  value such as \n. They must have non-zero values, as check_escape() returns
1191  their negation. Also, they must appear in the same order as in the opcode  their negation. Also, they must appear in the same order as in the opcode
1192  definitions below, up to ESC_z. There's a dummy for OP_ANY because it  definitions below, up to ESC_z. There's a dummy for OP_ANY because it
1193  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
1194  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  (which is used for [^] in JavaScript compatibility mode).
1195  There are two tests in the code for an escape greater than ESC_b and less than  
1196  ESC_Z to detect the types that may be repeated. These are the types that  The final escape must be ESC_REF as subsequent values are used for
1197  consume characters. If any new escapes are put in between that don't consume a  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1198  character, that code will have to change. */  greater than ESC_b and less than ESC_Z to detect the types that may be
1199    repeated. These are the types that consume characters. If any new escapes are
1200  enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,  put in between that don't consume a character, that code will have to change.
1201         ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_X, ESC_Z, ESC_z,  */
        ESC_E, ESC_Q, ESC_k, ESC_REF };  
1202    
1203    enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
1204           ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1205           ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
1206           ESC_REF };
1207    
 /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  
 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to  
 OP_EOD must correspond in order to the list of escapes immediately above.  
1208    
1209  To keep stored, compiled patterns compatible, new opcodes should be added  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
1210  immediately before OP_BRA, where (since release 7.0) a gap is left for this  OP_EOD must correspond in order to the list of escapes immediately above.
 purpose.  
1211    
1212  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1213  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
1214  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1215    
1216  enum {  enum {
1217    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 575  enum { Line 1220  enum {
1220    
1221    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
1222    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
1223    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
1224    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
1225    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
1226    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
1227    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
1228    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
1229    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
1230    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
1231    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
1232    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
1233    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
1234    OP_PROP,           /* 14 \p (Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1235    OP_ANYNL,          /* 15 \R (any newline sequence) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1236    OP_EXTUNI,         /* 16 \X (extended Unicode sequence */    OP_PROP,           /* 16 \p (Unicode property) */
1237    OP_EODN,           /* 17 End of data or \n at end of data: \Z. */    OP_ANYNL,          /* 17 \R (any newline sequence) */
1238    OP_EOD,            /* 18 End of data: \z */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1239      OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1240    OP_OPT,            /* 19 Set runtime options */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1241    OP_CIRC,           /* 20 Start of line - varies with multiline switch */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1242    OP_DOLL,           /* 21 End of line - varies with multiline switch */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1243    OP_CHAR,           /* 22 Match one character, casefully */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1244    OP_CHARNC,         /* 23 Match one character, caselessly */    OP_EOD,            /* 24 End of data: \z */
1245    OP_NOT,            /* 24 Match one character, not the following one */  
1246      OP_OPT,            /* 25 Set runtime options */
1247    OP_STAR,           /* 25 The maximizing and minimizing versions of */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
1248    OP_MINSTAR,        /* 26 these six opcodes must come in pairs, with */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
1249    OP_PLUS,           /* 27 the minimizing one second. */    OP_CHAR,           /* 28 Match one character, casefully */
1250    OP_MINPLUS,        /* 28 This first set applies to single characters.*/    OP_CHARNC,         /* 29 Match one character, caselessly */
1251    OP_QUERY,          /* 29 */    OP_NOT,            /* 30 Match one character, not the following one */
1252    OP_MINQUERY,       /* 30 */  
1253      OP_STAR,           /* 31 The maximizing and minimizing versions of */
1254    OP_UPTO,           /* 31 From 0 to n matches */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
1255    OP_MINUPTO,        /* 32 */    OP_PLUS,           /* 33 the minimizing one second. */
1256    OP_EXACT,          /* 33 Exactly n matches */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
1257      OP_QUERY,          /* 35 */
1258    OP_POSSTAR,        /* 34 Possessified star */    OP_MINQUERY,       /* 36 */
1259    OP_POSPLUS,        /* 35 Possessified plus */  
1260    OP_POSQUERY,       /* 36 Posesssified query */    OP_UPTO,           /* 37 From 0 to n matches */
1261    OP_POSUPTO,        /* 37 Possessified upto */    OP_MINUPTO,        /* 38 */
1262      OP_EXACT,          /* 39 Exactly n matches */
1263    OP_NOTSTAR,        /* 38 The maximizing and minimizing versions of */  
1264    OP_NOTMINSTAR,     /* 39 these six opcodes must come in pairs, with */    OP_POSSTAR,        /* 40 Possessified star */
1265    OP_NOTPLUS,        /* 40 the minimizing one second. They must be in */    OP_POSPLUS,        /* 41 Possessified plus */
1266    OP_NOTMINPLUS,     /* 41 exactly the same order as those above. */    OP_POSQUERY,       /* 42 Posesssified query */
1267    OP_NOTQUERY,       /* 42 This set applies to "not" single characters. */    OP_POSUPTO,        /* 43 Possessified upto */
1268    OP_NOTMINQUERY,    /* 43 */  
1269      OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
1270    OP_NOTUPTO,        /* 44 From 0 to n matches */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
1271    OP_NOTMINUPTO,     /* 45 */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
1272    OP_NOTEXACT,       /* 46 Exactly n matches */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
1273      OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
1274    OP_NOTPOSSTAR,     /* 47 Possessified versions */    OP_NOTMINQUERY,    /* 49 */
1275    OP_NOTPOSPLUS,     /* 48 */  
1276    OP_NOTPOSQUERY,    /* 49 */    OP_NOTUPTO,        /* 50 From 0 to n matches */
1277    OP_NOTPOSUPTO,     /* 50 */    OP_NOTMINUPTO,     /* 51 */
1278      OP_NOTEXACT,       /* 52 Exactly n matches */
1279    OP_TYPESTAR,       /* 51 The maximizing and minimizing versions of */  
1280    OP_TYPEMINSTAR,    /* 52 these six opcodes must come in pairs, with */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
1281    OP_TYPEPLUS,       /* 53 the minimizing one second. These codes must */    OP_NOTPOSPLUS,     /* 54 */
1282    OP_TYPEMINPLUS,    /* 54 be in exactly the same order as those above. */    OP_NOTPOSQUERY,    /* 55 */
1283    OP_TYPEQUERY,      /* 55 This set applies to character types such as \d */    OP_NOTPOSUPTO,     /* 56 */
1284    OP_TYPEMINQUERY,   /* 56 */  
1285      OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
1286    OP_TYPEUPTO,       /* 57 From 0 to n matches */    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
1287    OP_TYPEMINUPTO,    /* 58 */    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
1288    OP_TYPEEXACT,      /* 59 Exactly n matches */    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
1289      OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
1290    OP_TYPEPOSSTAR,    /* 60 Possessified versions */    OP_TYPEMINQUERY,   /* 62 */
1291    OP_TYPEPOSPLUS,    /* 61 */  
1292    OP_TYPEPOSQUERY,   /* 62 */    OP_TYPEUPTO,       /* 63 From 0 to n matches */
1293    OP_TYPEPOSUPTO,    /* 63 */    OP_TYPEMINUPTO,    /* 64 */
1294      OP_TYPEEXACT,      /* 65 Exactly n matches */
1295    OP_CRSTAR,         /* 64 The maximizing and minimizing versions of */  
1296    OP_CRMINSTAR,      /* 65 all these opcodes must come in pairs, with */    OP_TYPEPOSSTAR,    /* 66 Possessified versions */
1297    OP_CRPLUS,         /* 66 the minimizing one second. These codes must */    OP_TYPEPOSPLUS,    /* 67 */
1298    OP_CRMINPLUS,      /* 67 be in exactly the same order as those above. */    OP_TYPEPOSQUERY,   /* 68 */
1299    OP_CRQUERY,        /* 68 These are for character classes and back refs */    OP_TYPEPOSUPTO,    /* 69 */
1300    OP_CRMINQUERY,     /* 69 */  
1301    OP_CRRANGE,        /* 70 These are different to the three sets above. */    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
1302    OP_CRMINRANGE,     /* 71 */    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
1303      OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
1304      OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
1305      OP_CRQUERY,        /* 74 These are for character classes and back refs */
1306      OP_CRMINQUERY,     /* 75 */
1307      OP_CRRANGE,        /* 76 These are different to the three sets above. */
1308      OP_CRMINRANGE,     /* 77 */
1309    
1310    OP_CLASS,          /* 72 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
1311    OP_NCLASS,         /* 73 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
1312                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
1313                             character > 255 is encountered. */                             character > 255 is encountered. */
1314    
1315    OP_XCLASS,         /* 74 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
1316                             class. This does both positive and negative. */                             class. This does both positive and negative. */
1317    
1318    OP_REF,            /* 75 Match a back reference */    OP_REF,            /* 81 Match a back reference */
1319    OP_RECURSE,        /* 76 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
1320    OP_CALLOUT,        /* 77 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
1321    
1322    OP_ALT,            /* 78 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
1323    OP_KET,            /* 79 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */
1324    OP_KETRMAX,        /* 80 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
1325    OP_KETRMIN,        /* 81 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
1326    
1327    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
1328    
1329    OP_ASSERT,         /* 82 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
1330    OP_ASSERT_NOT,     /* 83 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
1331    OP_ASSERTBACK,     /* 84 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
1332    OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
1333    OP_REVERSE,        /* 86 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
1334    
1335    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
1336    as there's a test for >= ONCE for a subpattern that isn't an assertion. */    as there's a test for >= ONCE for a subpattern that isn't an assertion. */
1337    
1338    OP_ONCE,           /* 87 Atomic group */    OP_ONCE,           /* 93 Atomic group */
1339    OP_BRA,            /* 88 Start of non-capturing bracket */    OP_BRA,            /* 94 Start of non-capturing bracket */
1340    OP_CBRA,           /* 89 Start of capturing bracket */    OP_CBRA,           /* 95 Start of capturing bracket */
1341    OP_COND,           /* 90 Conditional group */    OP_COND,           /* 96 Conditional group */
1342    
1343    /* These three must follow the previous three, in the same order. There's a    /* These three must follow the previous three, in the same order. There's a
1344    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1345    
1346    OP_SBRA,           /* 91 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
1347    OP_SCBRA,          /* 92 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1348    OP_SCOND,          /* 93 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
1349    
1350    OP_CREF,           /* 94 Used to hold a capture number as condition */    /* The next two pairs must (respectively) be kept together. */
   OP_RREF,           /* 95 Used to hold a recursion number as condition */  
   OP_DEF,            /* 96 The DEFINE condition */  
1351    
1352    OP_BRAZERO,        /* 97 These two must remain together and in this */    OP_CREF,           /* 100 Used to hold a capture number as condition */
1353    OP_BRAMINZERO      /* 98 order. */    OP_NCREF,          /* 101 Same, but generaged by a name reference*/
1354      OP_RREF,           /* 102 Used to hold a recursion number as condition */
1355      OP_NRREF,          /* 103 Same, but generaged by a name reference*/
1356      OP_DEF,            /* 104 The DEFINE condition */
1357    
1358      OP_BRAZERO,        /* 105 These two must remain together and in this */
1359      OP_BRAMINZERO,     /* 106 order. */
1360    
1361      /* These are backtracking control verbs */
1362    
1363      OP_PRUNE,          /* 107 */
1364      OP_SKIP,           /* 108 */
1365      OP_THEN,           /* 109 */
1366      OP_COMMIT,         /* 110 */
1367    
1368      /* These are forced failure and success verbs */
1369    
1370      OP_FAIL,           /* 111 */
1371      OP_ACCEPT,         /* 112 */
1372      OP_CLOSE,          /* 113 Used before OP_ACCEPT to close open captures */
1373    
1374      /* This is used to skip a subpattern with a {0} quantifier */
1375    
1376      OP_SKIPZERO        /* 114 */
1377  };  };
1378    
1379    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1380    definitions that follow must also be updated to match. There are also tables
1381    called "coptable" cna "poptable" in pcre_dfa_exec.c that must be updated. */
1382    
1383    
1384  /* 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
1385  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
1386    
1387  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1388    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1389    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1390    "notprop", "prop", "anynl", "extuni",                           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1391    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
1392    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
1393    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1394    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
# Line 724  for debugging. The macro is referenced o Line 1400  for debugging. The macro is referenced o
1400    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
1401    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1402    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1403    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1404    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1405      "Brazero", "Braminzero",                                        \
1406      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
1407      "Close", "Skip zero"
1408    
1409    
1410  /* 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 739  in UTF-8 mode. The code that uses this t Line 1418  in UTF-8 mode. The code that uses this t
1418    
1419  #define OP_LENGTHS \  #define OP_LENGTHS \
1420    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1421    1, 1, 1, 1, 1, 1, 1, 1, 1, 1,  /* \A, \G, \B, \B, \D, \d, \S, \s, \W, \w */ \    1, 1, 1, 1, 1,                 /* \A, \G, \K, \B, \b                     */ \
1422    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1423    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1424      3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
1425      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1426    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
1427    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1428    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 783  in UTF-8 mode. The code that uses this t Line 1464  in UTF-8 mode. The code that uses this t
1464    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1465    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1466    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1467    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1468    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1469    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1470    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1471      1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
1472      1, 1, 3, 1                     /* FAIL, ACCEPT, CLOSE, SKIPZERO          */
1473    
1474    
1475  /* 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"
1476    condition. */
1477    
1478  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1479    
# Line 801  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 1485  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
1485         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
1486         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1487         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1488         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1489           ERR60, ERR61, ERR62, ERR63, ERR64, ERR65 };
1490    
1491  /* 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
1492  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 817  Because people can now save and re-use c Line 1502  Because people can now save and re-use c
1502  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
1503  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
1504  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1505  NOTE NOTE NOTE:  NOTE NOTE NOTE
1506  */  */
1507    
1508  typedef struct real_pcre {  typedef struct real_pcre {
1509    pcre_uint32 magic_number;    pcre_uint32 magic_number;
1510    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1511    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
1512    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
1513      pcre_uint16 dummy1;             /* For future use */
1514    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
1515    pcre_uint16 top_backref;    pcre_uint16 top_backref;
1516    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 844  remark (see NOTE above) about extending Line 1529  remark (see NOTE above) about extending
1529    
1530  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1531    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1532    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1533    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1534      pcre_uint32 minlength;          /* Minimum subject length */
1535  } pcre_study_data;  } pcre_study_data;
1536    
1537    /* Structure for building a chain of open capturing subpatterns during
1538    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1539    encountered. */
1540    
1541    typedef struct open_capitem {
1542      struct open_capitem *next;    /* Chain link */
1543      pcre_uint16 number;           /* Capture number */
1544    } open_capitem;
1545    
1546  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1547  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1548    
# Line 860  typedef struct compile_data { Line 1555  typedef struct compile_data {
1555    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1556    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1557    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1558      open_capitem *open_caps;      /* Chain of open capture items */
1559    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1560    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1561    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
1562    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
1563    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
1564      int  final_bracount;          /* Saved value after first pass */
1565    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
1566    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
1567    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
1568      int  external_flags;          /* External flag bits to be set */
1569    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1570    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
1571      BOOL check_lookbehind;        /* Lookbehinds need later checking */
1572    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1573    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1574    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 890  typedef struct recursion_info { Line 1589  typedef struct recursion_info {
1589    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1590    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
1591    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 */
1592    USPTR save_start;             /* Old value of md->start_match */    USPTR save_start;             /* Old value of mstart */
1593    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1594    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1595      int save_offset_top;          /* Current value of offset_top */
1596  } recursion_info;  } recursion_info;
1597    
 /* When compiling in a mode that doesn't use recursive calls to match(),  
 a structure is used to remember local variables on the heap. It is defined in  
 pcre_exec.c, close to the match() function, so that it is easy to keep it in  
 step with any changes of local variable. However, the pointer to the current  
 frame must be saved in some "static" place over a longjmp(). We declare the  
 structure here so that we can put a pointer in the match_data structure. NOTE:  
 This isn't used for a "normal" compilation of pcre. */  
   
 struct heapframe;  
   
1598  /* 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
1599  pointer at the start of each subpattern, so as to detect when an empty string  pointer at the start of each subpattern, so as to detect when an empty string
1600  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops. */
# Line 927  typedef struct match_data { Line 1617  typedef struct match_data {
1617    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1618    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1619    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1620      int    name_count;            /* Number of names in name table */
1621      int    name_entry_size;       /* Size of entry in names table */
1622      uschar *name_table;           /* Table of names */
1623    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1624    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1625    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
# Line 934  typedef struct match_data { Line 1627  typedef struct match_data {
1627    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1628    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1629    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1630      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1631    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1632    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1633    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1634    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1635      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1636    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1637    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1638    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1639    USPTR  start_match;           /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
1640    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1641      USPTR  start_used_ptr;        /* Earliest consulted character */
1642      int    partial;               /* PARTIAL options */
1643    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1644    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1645    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
# Line 950  typedef struct match_data { Line 1647  typedef struct match_data {
1647    int    eptrn;                 /* Next free eptrblock */    int    eptrn;                 /* Next free eptrblock */
1648    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1649    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
   struct heapframe *thisframe;  /* Used only when compiling for no recursion */  
1650  } match_data;  } match_data;
1651    
1652  /* 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 960  typedef struct dfa_match_data { Line 1656  typedef struct dfa_match_data {
1656    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;     /* Start of the compiled pattern */
1657    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
1658    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;    /* End of subject string */
1659      const uschar *start_used_ptr; /* Earliest consulted character */
1660    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1661      int   start_offset;           /* The start offset value */
1662    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1663    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1664    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
# Line 975  typedef struct dfa_match_data { Line 1673  typedef struct dfa_match_data {
1673  #define ctype_letter  0x02  #define ctype_letter  0x02
1674  #define ctype_digit   0x04  #define ctype_digit   0x04
1675  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1676  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1677  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1678    
1679  /* 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 1003  total length. */ Line 1701  total length. */
1701  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1702    
1703  /* 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
1704  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1705    relocations in shared libraries, it now has an offset into a single string
1706    instead. */
1707    
1708  typedef struct {  typedef struct {
1709    const char *name;    pcre_uint16 name_offset;
1710    pcre_uint16 type;    pcre_uint16 type;
1711    pcre_uint16 value;    pcre_uint16 value;
1712  } ucp_type_table;  } ucp_type_table;
# Line 1024  extern const uschar _pcre_utf8_table4[]; Line 1724  extern const uschar _pcre_utf8_table4[];
1724    
1725  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1726    
1727    extern const char   _pcre_utt_names[];
1728  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1729  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1730    
# Line 1036  extern const uschar _pcre_OP_lengths[]; Line 1737  extern const uschar _pcre_OP_lengths[];
1737  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
1738  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1739    
1740  extern BOOL         _pcre_is_newline(const uschar *, const uschar *, int *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1741                        BOOL);  extern BOOL          _pcre_is_newline(const uschar *, int, const uschar *,
1742  extern int          _pcre_ord2utf8(int, uschar *);                         int *, BOOL);
1743  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern int           _pcre_ord2utf8(int, uschar *);
1744                        const pcre_study_data *, pcre_study_data *);  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1745  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);                         const pcre_study_data *, pcre_study_data *);
1746  extern unsigned int _pcre_ucp_othercase(const unsigned int);  extern int           _pcre_valid_utf8(const uschar *, int);
1747  extern int          _pcre_valid_utf8(const uschar *, int);  extern BOOL          _pcre_was_newline(const uschar *, int, const uschar *,
1748  extern BOOL         _pcre_was_newline(const uschar *, const uschar *, int *,                         int *, BOOL);
1749                        BOOL);  extern BOOL          _pcre_xclass(int, const uschar *);
1750  extern BOOL         _pcre_xclass(int, const uschar *);  
1751    
1752    /* Unicode character database (UCD) */
1753    
1754    typedef struct {
1755      uschar script;
1756      uschar chartype;
1757      pcre_int32 other_case;
1758    } ucd_record;
1759    
1760    extern const ucd_record  _pcre_ucd_records[];
1761    extern const uschar      _pcre_ucd_stage1[];
1762    extern const pcre_uint16 _pcre_ucd_stage2[];
1763    extern const int         _pcre_ucp_gentype[];
1764    
1765    
1766    /* UCD access macros */
1767    
1768    #define UCD_BLOCK_SIZE 128
1769    #define GET_UCD(ch) (_pcre_ucd_records + \
1770            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1771            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1772    
1773    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1774    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1775    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1776    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1777    
1778  #endif  #endif
1779    

Legend:
Removed from v.137  
changed lines
  Added in v.462

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12