/[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 145 by ph10, Wed Apr 4 14:06:52 2007 UTC revision 442 by ph10, Fri Sep 11 10:21:02 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 86  setjmp and stdarg are used is when NO_RE Line 96  setjmp and stdarg are used is when NO_RE
96  /* When compiling a DLL for Windows, the exported symbols have to be declared  /* When compiling a DLL for Windows, the exported symbols have to be declared
97  using some MS magic. I found some useful information on this web page:  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  http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
99  information there, using __declspec(dllesport) without "extern" we have a  information there, using __declspec(dllexport) without "extern" we have a
100  definition; with "extern" we have a declaration. The settings here override the  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,  setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
102  which is all that is needed for applications, which import the symbols. We use:  which is all that is needed for applications (they just import the symbols). We
103    use:
104    
105    PCRE_EXP_DECL       for declarations    PCRE_EXP_DECL       for declarations
106    PCRE_EXP_DEFN       for definitions of exported functions    PCRE_EXP_DEFN       for definitions of exported functions
# Line 102  Windows, the two should always be the sa Line 113  Windows, the two should always be the sa
113    
114  The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,  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  which is an application, but needs to import this file in order to "peek" at
116  internals, can #include pcre.h first, can get an application's-eye view.  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,  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  special-purpose environments) might want to stick other stuff in front of
# Line 111  PCRE_EXP_DATA_DEFN only if they are not Line 122  PCRE_EXP_DATA_DEFN only if they are not
122    
123  #ifndef PCRE_EXP_DECL  #ifndef PCRE_EXP_DECL
124  #  ifdef _WIN32  #  ifdef _WIN32
125  #    ifdef DLL_EXPORT  #    ifndef PCRE_STATIC
126  #      define PCRE_EXP_DECL       extern __declspec(dllexport)  #      define PCRE_EXP_DECL       extern __declspec(dllexport)
127  #      define PCRE_EXP_DEFN       __declspec(dllexport)  #      define PCRE_EXP_DEFN       __declspec(dllexport)
128  #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)  #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
# Line 120  PCRE_EXP_DATA_DEFN only if they are not Line 131  PCRE_EXP_DATA_DEFN only if they are not
131  #      define PCRE_EXP_DEFN  #      define PCRE_EXP_DEFN
132  #      define PCRE_EXP_DATA_DEFN  #      define PCRE_EXP_DATA_DEFN
133  #    endif  #    endif
 #  
134  #  else  #  else
135  #    ifdef __cplusplus  #    ifdef __cplusplus
136  #      define PCRE_EXP_DECL       extern "C"  #      define PCRE_EXP_DECL       extern "C"
# Line 136  PCRE_EXP_DATA_DEFN only if they are not Line 146  PCRE_EXP_DATA_DEFN only if they are not
146  #  endif  #  endif
147  #endif  #endif
148    
149    /* When compiling with the MSVC compiler, it is sometimes necessary to include
150    a "calling convention" before exported function names. (This is secondhand
151    information; I know nothing about MSVC myself). For example, something like
152    
153      void __cdecl function(....)
154    
155    might be needed. In order so make this easy, all the exported functions have
156    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
157    set, we ensure here that it has no effect. */
158    
159    #ifndef PCRE_CALL_CONVENTION
160    #define PCRE_CALL_CONVENTION
161    #endif
162    
163  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We
164  cannot determine these outside the compilation (e.g. by running a program as  cannot determine these outside the compilation (e.g. by running a program as
165  part of "configure") because PCRE is often cross-compiled for use on other  part of "configure") because PCRE is often cross-compiled for use on other
# Line 144  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 174  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 201  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 232  must begin with PCRE_. */ Line 261  must begin with PCRE_. */
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 361  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 369  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 461  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 481  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    
# Line 500  bits. */ Line 553  bits. */
553  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
554  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
555    
556  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY)  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
557                               PCRE_NEWLINE_ANYCRLF)
558    
559  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
560    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
561     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
562     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
563     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
564       PCRE_JAVASCRIPT_COMPAT)
565    
566  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
567    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
568     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
569       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
570    
571  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
572    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
573     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
574       PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
575       PCRE_NO_START_OPTIMIZE)
576    
577  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
578    
# Line 539  variable-length repeat, or a anything ot Line 597  variable-length repeat, or a anything ot
597  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
598  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
599    
600  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
601    environments where these macros are defined elsewhere. Unfortunately, there
602    is no way to do the same for the typedef. */
603    
604  typedef int BOOL;  typedef int BOOL;
605    
606    #ifndef FALSE
607  #define FALSE   0  #define FALSE   0
608  #define TRUE    1  #define TRUE    1
609    #endif
610    
611    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
612    character constants like '*' because the compiler would emit their EBCDIC code,
613    which is different from their ASCII/UTF-8 code. Instead we define macros for
614    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
615    is enabled. When UTF-8 support is not enabled, the definitions use character
616    literals. Both character and string versions of each character are needed, and
617    there are some longer strings as well.
618    
619    This means that, on EBCDIC platforms, the PCRE library can handle either
620    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
621    would need different lookups depending on whether PCRE_UTF8 was set or not.
622    This would make it impossible to use characters in switch/case statements,
623    which would reduce performance. For a theoretical use (which nobody has asked
624    for) in a minority area (EBCDIC platforms), this is not sensible. Any
625    application that did need both could compile two versions of the library, using
626    macros to give the functions distinct names. */
627    
628    #ifndef SUPPORT_UTF8
629    
630    /* UTF-8 support is not enabled; use the platform-dependent character literals
631    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
632    
633    #define CHAR_HT                     '\t'
634    #define CHAR_VT                     '\v'
635    #define CHAR_FF                     '\f'
636    #define CHAR_CR                     '\r'
637    #define CHAR_NL                     '\n'
638    #define CHAR_BS                     '\b'
639    #define CHAR_BEL                    '\a'
640    #ifdef EBCDIC
641    #define CHAR_ESC                    '\047'
642    #define CHAR_DEL                    '\007'
643    #else
644    #define CHAR_ESC                    '\033'
645    #define CHAR_DEL                    '\177'
646    #endif
647    
648    #define CHAR_SPACE                  ' '
649    #define CHAR_EXCLAMATION_MARK       '!'
650    #define CHAR_QUOTATION_MARK         '"'
651    #define CHAR_NUMBER_SIGN            '#'
652    #define CHAR_DOLLAR_SIGN            '$'
653    #define CHAR_PERCENT_SIGN           '%'
654    #define CHAR_AMPERSAND              '&'
655    #define CHAR_APOSTROPHE             '\''
656    #define CHAR_LEFT_PARENTHESIS       '('
657    #define CHAR_RIGHT_PARENTHESIS      ')'
658    #define CHAR_ASTERISK               '*'
659    #define CHAR_PLUS                   '+'
660    #define CHAR_COMMA                  ','
661    #define CHAR_MINUS                  '-'
662    #define CHAR_DOT                    '.'
663    #define CHAR_SLASH                  '/'
664    #define CHAR_0                      '0'
665    #define CHAR_1                      '1'
666    #define CHAR_2                      '2'
667    #define CHAR_3                      '3'
668    #define CHAR_4                      '4'
669    #define CHAR_5                      '5'
670    #define CHAR_6                      '6'
671    #define CHAR_7                      '7'
672    #define CHAR_8                      '8'
673    #define CHAR_9                      '9'
674    #define CHAR_COLON                  ':'
675    #define CHAR_SEMICOLON              ';'
676    #define CHAR_LESS_THAN_SIGN         '<'
677    #define CHAR_EQUALS_SIGN            '='
678    #define CHAR_GREATER_THAN_SIGN      '>'
679    #define CHAR_QUESTION_MARK          '?'
680    #define CHAR_COMMERCIAL_AT          '@'
681    #define CHAR_A                      'A'
682    #define CHAR_B                      'B'
683    #define CHAR_C                      'C'
684    #define CHAR_D                      'D'
685    #define CHAR_E                      'E'
686    #define CHAR_F                      'F'
687    #define CHAR_G                      'G'
688    #define CHAR_H                      'H'
689    #define CHAR_I                      'I'
690    #define CHAR_J                      'J'
691    #define CHAR_K                      'K'
692    #define CHAR_L                      'L'
693    #define CHAR_M                      'M'
694    #define CHAR_N                      'N'
695    #define CHAR_O                      'O'
696    #define CHAR_P                      'P'
697    #define CHAR_Q                      'Q'
698    #define CHAR_R                      'R'
699    #define CHAR_S                      'S'
700    #define CHAR_T                      'T'
701    #define CHAR_U                      'U'
702    #define CHAR_V                      'V'
703    #define CHAR_W                      'W'
704    #define CHAR_X                      'X'
705    #define CHAR_Y                      'Y'
706    #define CHAR_Z                      'Z'
707    #define CHAR_LEFT_SQUARE_BRACKET    '['
708    #define CHAR_BACKSLASH              '\\'
709    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
710    #define CHAR_CIRCUMFLEX_ACCENT      '^'
711    #define CHAR_UNDERSCORE             '_'
712    #define CHAR_GRAVE_ACCENT           '`'
713    #define CHAR_a                      'a'
714    #define CHAR_b                      'b'
715    #define CHAR_c                      'c'
716    #define CHAR_d                      'd'
717    #define CHAR_e                      'e'
718    #define CHAR_f                      'f'
719    #define CHAR_g                      'g'
720    #define CHAR_h                      'h'
721    #define CHAR_i                      'i'
722    #define CHAR_j                      'j'
723    #define CHAR_k                      'k'
724    #define CHAR_l                      'l'
725    #define CHAR_m                      'm'
726    #define CHAR_n                      'n'
727    #define CHAR_o                      'o'
728    #define CHAR_p                      'p'
729    #define CHAR_q                      'q'
730    #define CHAR_r                      'r'
731    #define CHAR_s                      's'
732    #define CHAR_t                      't'
733    #define CHAR_u                      'u'
734    #define CHAR_v                      'v'
735    #define CHAR_w                      'w'
736    #define CHAR_x                      'x'
737    #define CHAR_y                      'y'
738    #define CHAR_z                      'z'
739    #define CHAR_LEFT_CURLY_BRACKET     '{'
740    #define CHAR_VERTICAL_LINE          '|'
741    #define CHAR_RIGHT_CURLY_BRACKET    '}'
742    #define CHAR_TILDE                  '~'
743    
744    #define STR_HT                      "\t"
745    #define STR_VT                      "\v"
746    #define STR_FF                      "\f"
747    #define STR_CR                      "\r"
748    #define STR_NL                      "\n"
749    #define STR_BS                      "\b"
750    #define STR_BEL                     "\a"
751    #ifdef EBCDIC
752    #define STR_ESC                     "\047"
753    #define STR_DEL                     "\007"
754    #else
755    #define STR_ESC                     "\033"
756    #define STR_DEL                     "\177"
757    #endif
758    
759    #define STR_SPACE                   " "
760    #define STR_EXCLAMATION_MARK        "!"
761    #define STR_QUOTATION_MARK          "\""
762    #define STR_NUMBER_SIGN             "#"
763    #define STR_DOLLAR_SIGN             "$"
764    #define STR_PERCENT_SIGN            "%"
765    #define STR_AMPERSAND               "&"
766    #define STR_APOSTROPHE              "'"
767    #define STR_LEFT_PARENTHESIS        "("
768    #define STR_RIGHT_PARENTHESIS       ")"
769    #define STR_ASTERISK                "*"
770    #define STR_PLUS                    "+"
771    #define STR_COMMA                   ","
772    #define STR_MINUS                   "-"
773    #define STR_DOT                     "."
774    #define STR_SLASH                   "/"
775    #define STR_0                       "0"
776    #define STR_1                       "1"
777    #define STR_2                       "2"
778    #define STR_3                       "3"
779    #define STR_4                       "4"
780    #define STR_5                       "5"
781    #define STR_6                       "6"
782    #define STR_7                       "7"
783    #define STR_8                       "8"
784    #define STR_9                       "9"
785    #define STR_COLON                   ":"
786    #define STR_SEMICOLON               ";"
787    #define STR_LESS_THAN_SIGN          "<"
788    #define STR_EQUALS_SIGN             "="
789    #define STR_GREATER_THAN_SIGN       ">"
790    #define STR_QUESTION_MARK           "?"
791    #define STR_COMMERCIAL_AT           "@"
792    #define STR_A                       "A"
793    #define STR_B                       "B"
794    #define STR_C                       "C"
795    #define STR_D                       "D"
796    #define STR_E                       "E"
797    #define STR_F                       "F"
798    #define STR_G                       "G"
799    #define STR_H                       "H"
800    #define STR_I                       "I"
801    #define STR_J                       "J"
802    #define STR_K                       "K"
803    #define STR_L                       "L"
804    #define STR_M                       "M"
805    #define STR_N                       "N"
806    #define STR_O                       "O"
807    #define STR_P                       "P"
808    #define STR_Q                       "Q"
809    #define STR_R                       "R"
810    #define STR_S                       "S"
811    #define STR_T                       "T"
812    #define STR_U                       "U"
813    #define STR_V                       "V"
814    #define STR_W                       "W"
815    #define STR_X                       "X"
816    #define STR_Y                       "Y"
817    #define STR_Z                       "Z"
818    #define STR_LEFT_SQUARE_BRACKET     "["
819    #define STR_BACKSLASH               "\\"
820    #define STR_RIGHT_SQUARE_BRACKET    "]"
821    #define STR_CIRCUMFLEX_ACCENT       "^"
822    #define STR_UNDERSCORE              "_"
823    #define STR_GRAVE_ACCENT            "`"
824    #define STR_a                       "a"
825    #define STR_b                       "b"
826    #define STR_c                       "c"
827    #define STR_d                       "d"
828    #define STR_e                       "e"
829    #define STR_f                       "f"
830    #define STR_g                       "g"
831    #define STR_h                       "h"
832    #define STR_i                       "i"
833    #define STR_j                       "j"
834    #define STR_k                       "k"
835    #define STR_l                       "l"
836    #define STR_m                       "m"
837    #define STR_n                       "n"
838    #define STR_o                       "o"
839    #define STR_p                       "p"
840    #define STR_q                       "q"
841    #define STR_r                       "r"
842    #define STR_s                       "s"
843    #define STR_t                       "t"
844    #define STR_u                       "u"
845    #define STR_v                       "v"
846    #define STR_w                       "w"
847    #define STR_x                       "x"
848    #define STR_y                       "y"
849    #define STR_z                       "z"
850    #define STR_LEFT_CURLY_BRACKET      "{"
851    #define STR_VERTICAL_LINE           "|"
852    #define STR_RIGHT_CURLY_BRACKET     "}"
853    #define STR_TILDE                   "~"
854    
855    #define STRING_ACCEPT0              "ACCEPT\0"
856    #define STRING_COMMIT0              "COMMIT\0"
857    #define STRING_F0                   "F\0"
858    #define STRING_FAIL0                "FAIL\0"
859    #define STRING_PRUNE0               "PRUNE\0"
860    #define STRING_SKIP0                "SKIP\0"
861    #define STRING_THEN                 "THEN"
862    
863    #define STRING_alpha0               "alpha\0"
864    #define STRING_lower0               "lower\0"
865    #define STRING_upper0               "upper\0"
866    #define STRING_alnum0               "alnum\0"
867    #define STRING_ascii0               "ascii\0"
868    #define STRING_blank0               "blank\0"
869    #define STRING_cntrl0               "cntrl\0"
870    #define STRING_digit0               "digit\0"
871    #define STRING_graph0               "graph\0"
872    #define STRING_print0               "print\0"
873    #define STRING_punct0               "punct\0"
874    #define STRING_space0               "space\0"
875    #define STRING_word0                "word\0"
876    #define STRING_xdigit               "xdigit"
877    
878    #define STRING_DEFINE               "DEFINE"
879    
880    #define STRING_CR_RIGHTPAR          "CR)"
881    #define STRING_LF_RIGHTPAR          "LF)"
882    #define STRING_CRLF_RIGHTPAR        "CRLF)"
883    #define STRING_ANY_RIGHTPAR         "ANY)"
884    #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"
885    #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
886    #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
887    #define STRING_UTF8_RIGHTPAR        "UTF8)"
888    
889    #else  /* SUPPORT_UTF8 */
890    
891    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
892    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
893    only. */
894    
895    #define CHAR_HT                     '\011'
896    #define CHAR_VT                     '\013'
897    #define CHAR_FF                     '\014'
898    #define CHAR_CR                     '\015'
899    #define CHAR_NL                     '\012'
900    #define CHAR_BS                     '\010'
901    #define CHAR_BEL                    '\007'
902    #define CHAR_ESC                    '\033'
903    #define CHAR_DEL                    '\177'
904    
905    #define CHAR_SPACE                  '\040'
906    #define CHAR_EXCLAMATION_MARK       '\041'
907    #define CHAR_QUOTATION_MARK         '\042'
908    #define CHAR_NUMBER_SIGN            '\043'
909    #define CHAR_DOLLAR_SIGN            '\044'
910    #define CHAR_PERCENT_SIGN           '\045'
911    #define CHAR_AMPERSAND              '\046'
912    #define CHAR_APOSTROPHE             '\047'
913    #define CHAR_LEFT_PARENTHESIS       '\050'
914    #define CHAR_RIGHT_PARENTHESIS      '\051'
915    #define CHAR_ASTERISK               '\052'
916    #define CHAR_PLUS                   '\053'
917    #define CHAR_COMMA                  '\054'
918    #define CHAR_MINUS                  '\055'
919    #define CHAR_DOT                    '\056'
920    #define CHAR_SLASH                  '\057'
921    #define CHAR_0                      '\060'
922    #define CHAR_1                      '\061'
923    #define CHAR_2                      '\062'
924    #define CHAR_3                      '\063'
925    #define CHAR_4                      '\064'
926    #define CHAR_5                      '\065'
927    #define CHAR_6                      '\066'
928    #define CHAR_7                      '\067'
929    #define CHAR_8                      '\070'
930    #define CHAR_9                      '\071'
931    #define CHAR_COLON                  '\072'
932    #define CHAR_SEMICOLON              '\073'
933    #define CHAR_LESS_THAN_SIGN         '\074'
934    #define CHAR_EQUALS_SIGN            '\075'
935    #define CHAR_GREATER_THAN_SIGN      '\076'
936    #define CHAR_QUESTION_MARK          '\077'
937    #define CHAR_COMMERCIAL_AT          '\100'
938    #define CHAR_A                      '\101'
939    #define CHAR_B                      '\102'
940    #define CHAR_C                      '\103'
941    #define CHAR_D                      '\104'
942    #define CHAR_E                      '\105'
943    #define CHAR_F                      '\106'
944    #define CHAR_G                      '\107'
945    #define CHAR_H                      '\110'
946    #define CHAR_I                      '\111'
947    #define CHAR_J                      '\112'
948    #define CHAR_K                      '\113'
949    #define CHAR_L                      '\114'
950    #define CHAR_M                      '\115'
951    #define CHAR_N                      '\116'
952    #define CHAR_O                      '\117'
953    #define CHAR_P                      '\120'
954    #define CHAR_Q                      '\121'
955    #define CHAR_R                      '\122'
956    #define CHAR_S                      '\123'
957    #define CHAR_T                      '\124'
958    #define CHAR_U                      '\125'
959    #define CHAR_V                      '\126'
960    #define CHAR_W                      '\127'
961    #define CHAR_X                      '\130'
962    #define CHAR_Y                      '\131'
963    #define CHAR_Z                      '\132'
964    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
965    #define CHAR_BACKSLASH              '\134'
966    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
967    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
968    #define CHAR_UNDERSCORE             '\137'
969    #define CHAR_GRAVE_ACCENT           '\140'
970    #define CHAR_a                      '\141'
971    #define CHAR_b                      '\142'
972    #define CHAR_c                      '\143'
973    #define CHAR_d                      '\144'
974    #define CHAR_e                      '\145'
975    #define CHAR_f                      '\146'
976    #define CHAR_g                      '\147'
977    #define CHAR_h                      '\150'
978    #define CHAR_i                      '\151'
979    #define CHAR_j                      '\152'
980    #define CHAR_k                      '\153'
981    #define CHAR_l                      '\154'
982    #define CHAR_m                      '\155'
983    #define CHAR_n                      '\156'
984    #define CHAR_o                      '\157'
985    #define CHAR_p                      '\160'
986    #define CHAR_q                      '\161'
987    #define CHAR_r                      '\162'
988    #define CHAR_s                      '\163'
989    #define CHAR_t                      '\164'
990    #define CHAR_u                      '\165'
991    #define CHAR_v                      '\166'
992    #define CHAR_w                      '\167'
993    #define CHAR_x                      '\170'
994    #define CHAR_y                      '\171'
995    #define CHAR_z                      '\172'
996    #define CHAR_LEFT_CURLY_BRACKET     '\173'
997    #define CHAR_VERTICAL_LINE          '\174'
998    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
999    #define CHAR_TILDE                  '\176'
1000    
1001    #define STR_HT                      "\011"
1002    #define STR_VT                      "\013"
1003    #define STR_FF                      "\014"
1004    #define STR_CR                      "\015"
1005    #define STR_NL                      "\012"
1006    #define STR_BS                      "\010"
1007    #define STR_BEL                     "\007"
1008    #define STR_ESC                     "\033"
1009    #define STR_DEL                     "\177"
1010    
1011    #define STR_SPACE                   "\040"
1012    #define STR_EXCLAMATION_MARK        "\041"
1013    #define STR_QUOTATION_MARK          "\042"
1014    #define STR_NUMBER_SIGN             "\043"
1015    #define STR_DOLLAR_SIGN             "\044"
1016    #define STR_PERCENT_SIGN            "\045"
1017    #define STR_AMPERSAND               "\046"
1018    #define STR_APOSTROPHE              "\047"
1019    #define STR_LEFT_PARENTHESIS        "\050"
1020    #define STR_RIGHT_PARENTHESIS       "\051"
1021    #define STR_ASTERISK                "\052"
1022    #define STR_PLUS                    "\053"
1023    #define STR_COMMA                   "\054"
1024    #define STR_MINUS                   "\055"
1025    #define STR_DOT                     "\056"
1026    #define STR_SLASH                   "\057"
1027    #define STR_0                       "\060"
1028    #define STR_1                       "\061"
1029    #define STR_2                       "\062"
1030    #define STR_3                       "\063"
1031    #define STR_4                       "\064"
1032    #define STR_5                       "\065"
1033    #define STR_6                       "\066"
1034    #define STR_7                       "\067"
1035    #define STR_8                       "\070"
1036    #define STR_9                       "\071"
1037    #define STR_COLON                   "\072"
1038    #define STR_SEMICOLON               "\073"
1039    #define STR_LESS_THAN_SIGN          "\074"
1040    #define STR_EQUALS_SIGN             "\075"
1041    #define STR_GREATER_THAN_SIGN       "\076"
1042    #define STR_QUESTION_MARK           "\077"
1043    #define STR_COMMERCIAL_AT           "\100"
1044    #define STR_A                       "\101"
1045    #define STR_B                       "\102"
1046    #define STR_C                       "\103"
1047    #define STR_D                       "\104"
1048    #define STR_E                       "\105"
1049    #define STR_F                       "\106"
1050    #define STR_G                       "\107"
1051    #define STR_H                       "\110"
1052    #define STR_I                       "\111"
1053    #define STR_J                       "\112"
1054    #define STR_K                       "\113"
1055    #define STR_L                       "\114"
1056    #define STR_M                       "\115"
1057    #define STR_N                       "\116"
1058    #define STR_O                       "\117"
1059    #define STR_P                       "\120"
1060    #define STR_Q                       "\121"
1061    #define STR_R                       "\122"
1062    #define STR_S                       "\123"
1063    #define STR_T                       "\124"
1064    #define STR_U                       "\125"
1065    #define STR_V                       "\126"
1066    #define STR_W                       "\127"
1067    #define STR_X                       "\130"
1068    #define STR_Y                       "\131"
1069    #define STR_Z                       "\132"
1070    #define STR_LEFT_SQUARE_BRACKET     "\133"
1071    #define STR_BACKSLASH               "\134"
1072    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1073    #define STR_CIRCUMFLEX_ACCENT       "\136"
1074    #define STR_UNDERSCORE              "\137"
1075    #define STR_GRAVE_ACCENT            "\140"
1076    #define STR_a                       "\141"
1077    #define STR_b                       "\142"
1078    #define STR_c                       "\143"
1079    #define STR_d                       "\144"
1080    #define STR_e                       "\145"
1081    #define STR_f                       "\146"
1082    #define STR_g                       "\147"
1083    #define STR_h                       "\150"
1084    #define STR_i                       "\151"
1085    #define STR_j                       "\152"
1086    #define STR_k                       "\153"
1087    #define STR_l                       "\154"
1088    #define STR_m                       "\155"
1089    #define STR_n                       "\156"
1090    #define STR_o                       "\157"
1091    #define STR_p                       "\160"
1092    #define STR_q                       "\161"
1093    #define STR_r                       "\162"
1094    #define STR_s                       "\163"
1095    #define STR_t                       "\164"
1096    #define STR_u                       "\165"
1097    #define STR_v                       "\166"
1098    #define STR_w                       "\167"
1099    #define STR_x                       "\170"
1100    #define STR_y                       "\171"
1101    #define STR_z                       "\172"
1102    #define STR_LEFT_CURLY_BRACKET      "\173"
1103    #define STR_VERTICAL_LINE           "\174"
1104    #define STR_RIGHT_CURLY_BRACKET     "\175"
1105    #define STR_TILDE                   "\176"
1106    
1107    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1108    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1109    #define STRING_F0                   STR_F "\0"
1110    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1111    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1112    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1113    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1114    
1115    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1116    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1117    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1118    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1119    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1120    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1121    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1122    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1123    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1124    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1125    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1126    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1127    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1128    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1129    
1130    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1131    
1132    #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS
1133    #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS
1134    #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1135    #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1136    #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1137    #define STRING_BSR_ANYCRLF_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1138    #define STRING_BSR_UNICODE_RIGHTPAR STR_B STR_S STR_R STR_UNDERSCORE STR_U STR_N STR_I STR_C STR_O STR_D STR_E STR_RIGHT_PARENTHESIS
1139    #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1140    
1141    #endif  /* SUPPORT_UTF8 */
1142    
1143  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1144    
1145  #ifndef ESC_e  #ifndef ESC_e
1146  #define ESC_e 27  #define ESC_e CHAR_ESC
1147  #endif  #endif
1148    
1149  #ifndef ESC_f  #ifndef ESC_f
1150  #define ESC_f '\f'  #define ESC_f CHAR_FF
1151  #endif  #endif
1152    
1153  #ifndef ESC_n  #ifndef ESC_n
1154  #define ESC_n '\n'  #define ESC_n CHAR_NL
1155  #endif  #endif
1156    
1157  #ifndef ESC_r  #ifndef ESC_r
1158  #define ESC_r '\r'  #define ESC_r CHAR_CR
1159  #endif  #endif
1160    
1161  /* We can't officially use ESC_t because it is a POSIX reserved identifier  /* We can't officially use ESC_t because it is a POSIX reserved identifier
1162  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1163    
1164  #ifndef ESC_tee  #ifndef ESC_tee
1165  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1166  #endif  #endif
1167    
1168  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
# Line 595  contain UTF-8 characters with values gre Line 1189  contain UTF-8 characters with values gre
1189  value such as \n. They must have non-zero values, as check_escape() returns  value such as \n. They must have non-zero values, as check_escape() returns
1190  their negation. Also, they must appear in the same order as in the opcode  their negation. Also, they must appear in the same order as in the opcode
1191  definitions below, up to ESC_z. There's a dummy for OP_ANY because it  definitions below, up to ESC_z. There's a dummy for OP_ANY because it
1192  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
1193  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  (which is used for [^] in JavaScript compatibility mode).
 There are two tests in the code for an escape greater than ESC_b and less than  
 ESC_Z to detect the types that may be repeated. These are the types that  
 consume characters. If any new escapes are put in between that don't consume a  
 character, that code will have to change. */  
   
 enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,  
        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 };  
1194    
1195    The final escape must be ESC_REF as subsequent values are used for
1196    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1197    greater than ESC_b and less than ESC_Z to detect the types that may be
1198    repeated. These are the types that consume characters. If any new escapes are
1199    put in between that don't consume a character, that code will have to change.
1200    */
1201    
1202  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
1203  that extract substrings. Starting from 1 (i.e. after OP_END), the values up to         ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1204  OP_EOD must correspond in order to the list of escapes immediately above.         ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
1205           ESC_REF };
1206    
1207  To keep stored, compiled patterns compatible, new opcodes should be added  
1208  immediately before OP_BRA, where (since release 7.0) a gap is left for this  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
1209  purpose.  OP_EOD must correspond in order to the list of escapes immediately above.
1210    
1211  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1212  that follow must also be updated to match. There is also a table called  that follow must also be updated to match. There is also a table called
# Line 626  enum { Line 1219  enum {
1219    
1220    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
1221    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
1222    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
1223    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
1224    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
1225    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
1226    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
1227    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
1228    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
1229    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
1230    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
1231    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
1232    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
1233    OP_PROP,           /* 14 \p (Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1234    OP_ANYNL,          /* 15 \R (any newline sequence) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1235    OP_EXTUNI,         /* 16 \X (extended Unicode sequence */    OP_PROP,           /* 16 \p (Unicode property) */
1236    OP_EODN,           /* 17 End of data or \n at end of data: \Z. */    OP_ANYNL,          /* 17 \R (any newline sequence) */
1237    OP_EOD,            /* 18 End of data: \z */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1238      OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1239    OP_OPT,            /* 19 Set runtime options */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1240    OP_CIRC,           /* 20 Start of line - varies with multiline switch */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1241    OP_DOLL,           /* 21 End of line - varies with multiline switch */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1242    OP_CHAR,           /* 22 Match one character, casefully */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1243    OP_CHARNC,         /* 23 Match one character, caselessly */    OP_EOD,            /* 24 End of data: \z */
1244    OP_NOT,            /* 24 Match one character, not the following one */  
1245      OP_OPT,            /* 25 Set runtime options */
1246    OP_STAR,           /* 25 The maximizing and minimizing versions of */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
1247    OP_MINSTAR,        /* 26 these six opcodes must come in pairs, with */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
1248    OP_PLUS,           /* 27 the minimizing one second. */    OP_CHAR,           /* 28 Match one character, casefully */
1249    OP_MINPLUS,        /* 28 This first set applies to single characters.*/    OP_CHARNC,         /* 29 Match one character, caselessly */
1250    OP_QUERY,          /* 29 */    OP_NOT,            /* 30 Match one character, not the following one */
1251    OP_MINQUERY,       /* 30 */  
1252      OP_STAR,           /* 31 The maximizing and minimizing versions of */
1253    OP_UPTO,           /* 31 From 0 to n matches */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
1254    OP_MINUPTO,        /* 32 */    OP_PLUS,           /* 33 the minimizing one second. */
1255    OP_EXACT,          /* 33 Exactly n matches */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
1256      OP_QUERY,          /* 35 */
1257    OP_POSSTAR,        /* 34 Possessified star */    OP_MINQUERY,       /* 36 */
1258    OP_POSPLUS,        /* 35 Possessified plus */  
1259    OP_POSQUERY,       /* 36 Posesssified query */    OP_UPTO,           /* 37 From 0 to n matches */
1260    OP_POSUPTO,        /* 37 Possessified upto */    OP_MINUPTO,        /* 38 */
1261      OP_EXACT,          /* 39 Exactly n matches */
1262    OP_NOTSTAR,        /* 38 The maximizing and minimizing versions of */  
1263    OP_NOTMINSTAR,     /* 39 these six opcodes must come in pairs, with */    OP_POSSTAR,        /* 40 Possessified star */
1264    OP_NOTPLUS,        /* 40 the minimizing one second. They must be in */    OP_POSPLUS,        /* 41 Possessified plus */
1265    OP_NOTMINPLUS,     /* 41 exactly the same order as those above. */    OP_POSQUERY,       /* 42 Posesssified query */
1266    OP_NOTQUERY,       /* 42 This set applies to "not" single characters. */    OP_POSUPTO,        /* 43 Possessified upto */
1267    OP_NOTMINQUERY,    /* 43 */  
1268      OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
1269    OP_NOTUPTO,        /* 44 From 0 to n matches */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
1270    OP_NOTMINUPTO,     /* 45 */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
1271    OP_NOTEXACT,       /* 46 Exactly n matches */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
1272      OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
1273    OP_NOTPOSSTAR,     /* 47 Possessified versions */    OP_NOTMINQUERY,    /* 49 */
1274    OP_NOTPOSPLUS,     /* 48 */  
1275    OP_NOTPOSQUERY,    /* 49 */    OP_NOTUPTO,        /* 50 From 0 to n matches */
1276    OP_NOTPOSUPTO,     /* 50 */    OP_NOTMINUPTO,     /* 51 */
1277      OP_NOTEXACT,       /* 52 Exactly n matches */
1278    OP_TYPESTAR,       /* 51 The maximizing and minimizing versions of */  
1279    OP_TYPEMINSTAR,    /* 52 these six opcodes must come in pairs, with */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
1280    OP_TYPEPLUS,       /* 53 the minimizing one second. These codes must */    OP_NOTPOSPLUS,     /* 54 */
1281    OP_TYPEMINPLUS,    /* 54 be in exactly the same order as those above. */    OP_NOTPOSQUERY,    /* 55 */
1282    OP_TYPEQUERY,      /* 55 This set applies to character types such as \d */    OP_NOTPOSUPTO,     /* 56 */
1283    OP_TYPEMINQUERY,   /* 56 */  
1284      OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
1285    OP_TYPEUPTO,       /* 57 From 0 to n matches */    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
1286    OP_TYPEMINUPTO,    /* 58 */    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
1287    OP_TYPEEXACT,      /* 59 Exactly n matches */    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
1288      OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
1289    OP_TYPEPOSSTAR,    /* 60 Possessified versions */    OP_TYPEMINQUERY,   /* 62 */
1290    OP_TYPEPOSPLUS,    /* 61 */  
1291    OP_TYPEPOSQUERY,   /* 62 */    OP_TYPEUPTO,       /* 63 From 0 to n matches */
1292    OP_TYPEPOSUPTO,    /* 63 */    OP_TYPEMINUPTO,    /* 64 */
1293      OP_TYPEEXACT,      /* 65 Exactly n matches */
1294    OP_CRSTAR,         /* 64 The maximizing and minimizing versions of */  
1295    OP_CRMINSTAR,      /* 65 all these opcodes must come in pairs, with */    OP_TYPEPOSSTAR,    /* 66 Possessified versions */
1296    OP_CRPLUS,         /* 66 the minimizing one second. These codes must */    OP_TYPEPOSPLUS,    /* 67 */
1297    OP_CRMINPLUS,      /* 67 be in exactly the same order as those above. */    OP_TYPEPOSQUERY,   /* 68 */
1298    OP_CRQUERY,        /* 68 These are for character classes and back refs */    OP_TYPEPOSUPTO,    /* 69 */
1299    OP_CRMINQUERY,     /* 69 */  
1300    OP_CRRANGE,        /* 70 These are different to the three sets above. */    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
1301    OP_CRMINRANGE,     /* 71 */    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
1302      OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
1303      OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
1304      OP_CRQUERY,        /* 74 These are for character classes and back refs */
1305      OP_CRMINQUERY,     /* 75 */
1306      OP_CRRANGE,        /* 76 These are different to the three sets above. */
1307      OP_CRMINRANGE,     /* 77 */
1308    
1309    OP_CLASS,          /* 72 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
1310    OP_NCLASS,         /* 73 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
1311                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
1312                             character > 255 is encountered. */                             character > 255 is encountered. */
1313    
1314    OP_XCLASS,         /* 74 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
1315                             class. This does both positive and negative. */                             class. This does both positive and negative. */
1316    
1317    OP_REF,            /* 75 Match a back reference */    OP_REF,            /* 81 Match a back reference */
1318    OP_RECURSE,        /* 76 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
1319    OP_CALLOUT,        /* 77 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
1320    
1321    OP_ALT,            /* 78 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
1322    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 */
1323    OP_KETRMAX,        /* 80 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
1324    OP_KETRMIN,        /* 81 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
1325    
1326    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
1327    
1328    OP_ASSERT,         /* 82 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
1329    OP_ASSERT_NOT,     /* 83 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
1330    OP_ASSERTBACK,     /* 84 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
1331    OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
1332    OP_REVERSE,        /* 86 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
1333    
1334    /* 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,
1335    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. */
1336    
1337    OP_ONCE,           /* 87 Atomic group */    OP_ONCE,           /* 93 Atomic group */
1338    OP_BRA,            /* 88 Start of non-capturing bracket */    OP_BRA,            /* 94 Start of non-capturing bracket */
1339    OP_CBRA,           /* 89 Start of capturing bracket */    OP_CBRA,           /* 95 Start of capturing bracket */
1340    OP_COND,           /* 90 Conditional group */    OP_COND,           /* 96 Conditional group */
1341    
1342    /* 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
1343    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1344    
1345    OP_SBRA,           /* 91 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
1346    OP_SCBRA,          /* 92 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1347    OP_SCOND,          /* 93 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
   
   OP_CREF,           /* 94 Used to hold a capture number as condition */  
   OP_RREF,           /* 95 Used to hold a recursion number as condition */  
   OP_DEF,            /* 96 The DEFINE condition */  
1348    
1349    OP_BRAZERO,        /* 97 These two must remain together and in this */    OP_CREF,           /* 100 Used to hold a capture number as condition */
1350    OP_BRAMINZERO      /* 98 order. */    OP_RREF,           /* 101 Used to hold a recursion number as condition */
1351      OP_DEF,            /* 102 The DEFINE condition */
1352    
1353      OP_BRAZERO,        /* 103 These two must remain together and in this */
1354      OP_BRAMINZERO,     /* 104 order. */
1355    
1356      /* These are backtracking control verbs */
1357    
1358      OP_PRUNE,          /* 105 */
1359      OP_SKIP,           /* 106 */
1360      OP_THEN,           /* 107 */
1361      OP_COMMIT,         /* 108 */
1362    
1363      /* These are forced failure and success verbs */
1364    
1365      OP_FAIL,           /* 109 */
1366      OP_ACCEPT,         /* 110 */
1367    
1368      /* This is used to skip a subpattern with a {0} quantifier */
1369    
1370      OP_SKIPZERO        /* 111 */
1371  };  };
1372    
1373    
# Line 760  enum { Line 1375  enum {
1375  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
1376    
1377  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1378    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1379    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1380    "notprop", "prop", "anynl", "extuni",                           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1381    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
1382    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
1383    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1384    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
# Line 775  for debugging. The macro is referenced o Line 1390  for debugging. The macro is referenced o
1390    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
1391    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1392    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1393    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1394    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
1395      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
1396      "Skip zero"
1397    
1398    
1399  /* This macro defines the length of fixed length operations in the compiled  /* This macro defines the length of fixed length operations in the compiled
# Line 790  in UTF-8 mode. The code that uses this t Line 1407  in UTF-8 mode. The code that uses this t
1407    
1408  #define OP_LENGTHS \  #define OP_LENGTHS \
1409    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1410    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                     */ \
1411    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1412    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1413      3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
1414      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1415    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
1416    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1417    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 838  in UTF-8 mode. The code that uses this t Line 1457  in UTF-8 mode. The code that uses this t
1457    3,                             /* RREF                                   */ \    3,                             /* RREF                                   */ \
1458    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1459    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1460      1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
1461      1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
1462    
1463    
1464  /* A magic value for OP_RREF to indicate the "any recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
# Line 852  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 1473  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
1473         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
1474         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1475         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1476         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1477           ERR60, ERR61, ERR62, ERR63, ERR64 };
1478    
1479  /* 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
1480  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 874  NOTE NOTE NOTE: Line 1496  NOTE NOTE NOTE:
1496  typedef struct real_pcre {  typedef struct real_pcre {
1497    pcre_uint32 magic_number;    pcre_uint32 magic_number;
1498    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1499    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
1500    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
1501      pcre_uint16 dummy1;             /* For future use */
1502    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
1503    pcre_uint16 top_backref;    pcre_uint16 top_backref;
1504    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 915  typedef struct compile_data { Line 1537  typedef struct compile_data {
1537    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1538    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
1539    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
1540    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
1541      int  final_bracount;          /* Saved value after first pass */
1542    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
1543    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
1544    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
1545      int  external_flags;          /* External flag bits to be set */
1546    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1547    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
1548    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1549    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1550    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 941  typedef struct recursion_info { Line 1565  typedef struct recursion_info {
1565    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1566    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
1567    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 */
1568    USPTR save_start;             /* Old value of md->start_match */    USPTR save_start;             /* Old value of mstart */
1569    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1570    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1571  } recursion_info;  } recursion_info;
1572    
 /* 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;  
   
1573  /* 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
1574  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
1575  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops. */
# Line 985  typedef struct match_data { Line 1599  typedef struct match_data {
1599    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1600    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1601    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1602      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1603    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1604    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1605    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1606    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1607      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1608    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1609    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1610    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1611    USPTR  start_match;           /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
1612    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1613      USPTR  start_used_ptr;        /* Earliest consulted character */
1614      int    partial;               /* PARTIAL options */
1615    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1616    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1617    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
# Line 1001  typedef struct match_data { Line 1619  typedef struct match_data {
1619    int    eptrn;                 /* Next free eptrblock */    int    eptrn;                 /* Next free eptrblock */
1620    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1621    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 */  
1622  } match_data;  } match_data;
1623    
1624  /* 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 1011  typedef struct dfa_match_data { Line 1628  typedef struct dfa_match_data {
1628    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;     /* Start of the compiled pattern */
1629    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
1630    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;    /* End of subject string */
1631      const uschar *start_used_ptr; /* Earliest consulted character */
1632    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1633      int   start_offset;           /* The start offset value */
1634    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1635    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1636    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
# Line 1026  typedef struct dfa_match_data { Line 1645  typedef struct dfa_match_data {
1645  #define ctype_letter  0x02  #define ctype_letter  0x02
1646  #define ctype_digit   0x04  #define ctype_digit   0x04
1647  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1648  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1649  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1650    
1651  /* 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 1054  total length. */ Line 1673  total length. */
1673  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1674    
1675  /* 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
1676  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1677    relocations in shared libraries, it now has an offset into a single string
1678    instead. */
1679    
1680  typedef struct {  typedef struct {
1681    const char *name;    pcre_uint16 name_offset;
1682    pcre_uint16 type;    pcre_uint16 type;
1683    pcre_uint16 value;    pcre_uint16 value;
1684  } ucp_type_table;  } ucp_type_table;
# Line 1075  extern const uschar _pcre_utf8_table4[]; Line 1696  extern const uschar _pcre_utf8_table4[];
1696    
1697  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1698    
1699    extern const char   _pcre_utt_names[];
1700  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1701  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1702    
# Line 1087  extern const uschar _pcre_OP_lengths[]; Line 1709  extern const uschar _pcre_OP_lengths[];
1709  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
1710  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1711    
1712  extern BOOL         _pcre_is_newline(const uschar *, const uschar *, int *,  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,
1713                        BOOL);                        int *, BOOL);
1714  extern int          _pcre_ord2utf8(int, uschar *);  extern int          _pcre_ord2utf8(int, uschar *);
1715  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1716                        const pcre_study_data *, pcre_study_data *);                        const pcre_study_data *, pcre_study_data *);
 extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  
 extern unsigned int _pcre_ucp_othercase(const unsigned int);  
1717  extern int          _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_valid_utf8(const uschar *, int);
1718  extern BOOL         _pcre_was_newline(const uschar *, const uschar *, int *,  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1719                        BOOL);                        int *, BOOL);
1720  extern BOOL         _pcre_xclass(int, const uschar *);  extern BOOL         _pcre_xclass(int, const uschar *);
1721    
1722    
1723    /* Unicode character database (UCD) */
1724    
1725    typedef struct {
1726      uschar script;
1727      uschar chartype;
1728      pcre_int32 other_case;
1729    } ucd_record;
1730    
1731    extern const ucd_record  _pcre_ucd_records[];
1732    extern const uschar      _pcre_ucd_stage1[];
1733    extern const pcre_uint16 _pcre_ucd_stage2[];
1734    extern const int         _pcre_ucp_gentype[];
1735    
1736    
1737    /* UCD access macros */
1738    
1739    #define UCD_BLOCK_SIZE 128
1740    #define GET_UCD(ch) (_pcre_ucd_records + \
1741            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1742            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1743    
1744    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1745    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1746    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1747    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1748    
1749  #endif  #endif
1750    
1751  /* End of pcre_internal.h */  /* End of pcre_internal.h */

Legend:
Removed from v.145  
changed lines
  Added in v.442

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12