/[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 108 by ph10, Wed Mar 7 11:11:23 2007 UTC revision 593 by ph10, Sun May 1 08:14:24 2011 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) 1997-2006 University of Cambridge             Copyright (c) 1997-2010 University of Cambridge
11    
12  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
13  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 45  functions whose names all begin with "_p Line 45  functions whose names all begin with "_p
45  #ifndef PCRE_INTERNAL_H  #ifndef PCRE_INTERNAL_H
46  #define PCRE_INTERNAL_H  #define PCRE_INTERNAL_H
47    
48  /* Define DEBUG to get debugging output on stdout. */  /* Define PCRE_DEBUG to get debugging output on stdout. */
49    
50  #if 0  #if 0
51  #define DEBUG  #define PCRE_DEBUG
52    #endif
53    
54    /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"
55    script prevents both being selected, but not everybody uses "configure". */
56    
57    #if defined EBCDIC && defined SUPPORT_UTF8
58    #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.
59    #endif
60    
61    /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
62    "configure" script ensures this, but not everybody uses "configure". */
63    
64    #if defined SUPPORT_UCP && !defined SUPPORT_UTF8
65    #define SUPPORT_UTF8 1
66  #endif  #endif
67    
68  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef
# Line 60  It turns out that the Mac Debugging.h he Line 74  It turns out that the Mac Debugging.h he
74  be absolutely sure we get our version. */  be absolutely sure we get our version. */
75    
76  #undef DPRINTF  #undef DPRINTF
77  #ifdef DEBUG  #ifdef PCRE_DEBUG
78  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
79  #else  #else
80  #define DPRINTF(p) /* Nothing */  #define DPRINTF(p) /* Nothing */
81  #endif  #endif
82    
83    
 /* Get the definitions provided by running "configure" */  
   
 #include "config.h"  
   
84  /* Standard C headers plus the external interface definition. The only time  /* Standard C headers plus the external interface definition. The only time
85  setjmp and stdarg are used is when NO_RECURSE is set. */  setjmp and stdarg are used is when NO_RECURSE is set. */
86    
87  #include <ctype.h>  #include <ctype.h>
88  #include <limits.h>  #include <limits.h>
 #include <setjmp.h>  
 #include <stdarg.h>  
89  #include <stddef.h>  #include <stddef.h>
90  #include <stdio.h>  #include <stdio.h>
91  #include <stdlib.h>  #include <stdlib.h>
92  #include <string.h>  #include <string.h>
93    
94  #ifndef PCRE_SPY  /* When compiling a DLL for Windows, the exported symbols have to be declared
95  #define PCRE_DEFINITION       /* Win32 __declspec(export) trigger for .dll */  using some MS magic. I found some useful information on this web page:
96    http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
97    information there, using __declspec(dllexport) without "extern" we have a
98    definition; with "extern" we have a declaration. The settings here override the
99    setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
100    which is all that is needed for applications (they just import the symbols). We
101    use:
102    
103      PCRE_EXP_DECL       for declarations
104      PCRE_EXP_DEFN       for definitions of exported functions
105      PCRE_EXP_DATA_DEFN  for definitions of exported variables
106    
107    The reason for the two DEFN macros is that in non-Windows environments, one
108    does not want to have "extern" before variable definitions because it leads to
109    compiler warnings. So we distinguish between functions and variables. In
110    Windows, the two should always be the same.
111    
112    The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
113    which is an application, but needs to import this file in order to "peek" at
114    internals, can #include pcre.h first to get an application's-eye view.
115    
116    In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
117    special-purpose environments) might want to stick other stuff in front of
118    exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
119    PCRE_EXP_DATA_DEFN only if they are not already set. */
120    
121    #ifndef PCRE_EXP_DECL
122    #  ifdef _WIN32
123    #    ifndef PCRE_STATIC
124    #      define PCRE_EXP_DECL       extern __declspec(dllexport)
125    #      define PCRE_EXP_DEFN       __declspec(dllexport)
126    #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
127    #    else
128    #      define PCRE_EXP_DECL       extern
129    #      define PCRE_EXP_DEFN
130    #      define PCRE_EXP_DATA_DEFN
131    #    endif
132    #  else
133    #    ifdef __cplusplus
134    #      define PCRE_EXP_DECL       extern "C"
135    #    else
136    #      define PCRE_EXP_DECL       extern
137    #    endif
138    #    ifndef PCRE_EXP_DEFN
139    #      define PCRE_EXP_DEFN       PCRE_EXP_DECL
140    #    endif
141    #    ifndef PCRE_EXP_DATA_DEFN
142    #      define PCRE_EXP_DATA_DEFN
143    #    endif
144    #  endif
145    #endif
146    
147    /* When compiling with the MSVC compiler, it is sometimes necessary to include
148    a "calling convention" before exported function names. (This is secondhand
149    information; I know nothing about MSVC myself). For example, something like
150    
151      void __cdecl function(....)
152    
153    might be needed. In order so make this easy, all the exported functions have
154    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
155    set, we ensure here that it has no effect. */
156    
157    #ifndef PCRE_CALL_CONVENTION
158    #define PCRE_CALL_CONVENTION
159  #endif  #endif
160    
161  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We
# Line 95  preprocessor time in standard C environm Line 166  preprocessor time in standard C environm
166    
167  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
168    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
169      typedef short pcre_int16;
170  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
171    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
172      typedef int pcre_int16;
173  #else  #else
174    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
175  #endif  #endif
176    
177  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
178    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
179      typedef int pcre_int32;
180  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
181    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
182      typedef long int pcre_int32;
183  #else  #else
184    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
185  #endif  #endif
186    
187    /* When checking for integer overflow in pcre_compile(), we need to handle
188    large integers. If a 64-bit integer type is available, we can use that.
189    Otherwise we have to cast to double, which of course requires floating point
190    arithmetic. Handle this by defining a macro for the appropriate type. If
191    stdint.h is available, include it; it may define INT64_MAX. Systems that do not
192    have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
193    by "configure". */
194    
195    #if HAVE_STDINT_H
196    #include <stdint.h>
197    #elif HAVE_INTTYPES_H
198    #include <inttypes.h>
199    #endif
200    
201    #if defined INT64_MAX || defined int64_t
202    #define INT64_OR_DOUBLE int64_t
203    #else
204    #define INT64_OR_DOUBLE double
205    #endif
206    
207  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
208  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
209  However, we leave the interface to the outside world as char *, because that  However, we leave the interface to the outside world as char *, because that
# Line 125  characters only go up to 0x7fffffff (tho Line 220  characters only go up to 0x7fffffff (tho
220  #define NOTACHAR 0xffffffff  #define NOTACHAR 0xffffffff
221    
222  /* 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,
223  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
224  newlines. NLBLOCK, PSSTART, and PSEND are defined in the various modules to  testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
225  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
226  string field names are. */  start/end of string field names are. */
227    
228  #define NLTYPE_FIXED   0     /* Newline is a fixed length string */  #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
229  #define NLTYPE_ANY     1     /* Newline is any Unicode line ending */  #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
230    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
231    
232  /* This macro checks for a newline at the given position */  /* This macro checks for a newline at the given position */
233    
234  #define IS_NEWLINE(p) \  #define IS_NEWLINE(p) \
235    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
236      ((p) < NLBLOCK->PSEND && \      ((p) < NLBLOCK->PSEND && \
237       _pcre_is_newline((p), NLBLOCK->PSEND, &(NLBLOCK->nllen), utf8) \       _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
238      ) \         utf8)) \
239      : \      : \
240      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
241       (p)[0] == NLBLOCK->nl[0] && \       (p)[0] == NLBLOCK->nl[0] && \
# Line 152  string field names are. */ Line 248  string field names are. */
248  #define WAS_NEWLINE(p) \  #define WAS_NEWLINE(p) \
249    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
250      ((p) > NLBLOCK->PSSTART && \      ((p) > NLBLOCK->PSSTART && \
251       _pcre_was_newline((p), NLBLOCK->PSSTART, &(NLBLOCK->nllen), utf8) \       _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
252      ) \         &(NLBLOCK->nllen), utf8)) \
253      : \      : \
254      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
255       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
# Line 178  must begin with PCRE_. */ Line 274  must begin with PCRE_. */
274  #define USPTR const unsigned char *  #define USPTR const unsigned char *
275  #endif  #endif
276    
277    
278    
279  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
280  values. */  values. */
281    
# Line 200  option on the command line. */ Line 298  option on the command line. */
298  /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),  /* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
299  define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY  define a macro for memmove() if HAVE_MEMMOVE is false, provided that HAVE_BCOPY
300  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
301  neither (there some non-Unix environments where this is the case). This assumes  neither (there some non-Unix environments where this is the case). */
 that all calls to memmove are moving strings upwards in store, which is the  
 case in PCRE. */  
302    
303  #if ! HAVE_MEMMOVE  #ifndef HAVE_MEMMOVE
304  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
305  #if HAVE_BCOPY  #ifdef HAVE_BCOPY
306  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
307  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
308  static void *  static void *
# Line 215  pcre_memmove(void *d, const void *s, siz Line 311  pcre_memmove(void *d, const void *s, siz
311  size_t i;  size_t i;
312  unsigned char *dest = (unsigned char *)d;  unsigned char *dest = (unsigned char *)d;
313  const unsigned char *src = (const unsigned char *)s;  const unsigned char *src = (const unsigned char *)s;
314  dest += n;  if (dest > src)
315  src += n;    {
316  for (i = 0; i < n; ++i) *(--dest) = *(--src);    dest += n;
317  return (void *)dest;    src += n;
318      for (i = 0; i < n; ++i) *(--dest) = *(--src);
319      return (void *)dest;
320      }
321    else
322      {
323      for (i = 0; i < n; ++i) *dest++ = *src++;
324      return (void *)(dest - n);
325      }
326  }  }
327  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
328  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 304  capturing parenthesis numbers in back re Line 408  capturing parenthesis numbers in back re
408    
409  /* When UTF-8 encoding is being used, a character is no longer just a single  /* When UTF-8 encoding is being used, a character is no longer just a single
410  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
411  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. GETCHARLENTEST is
412    not used when UTF-8 is not supported, so it is not defined, and BACKCHAR should
413    never be called in byte mode. To make sure they can never even appear when
414    UTF-8 support is omitted, we don't even define them. */
415    
416  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
417  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 312  byte-mode, and more complicated ones for Line 419  byte-mode, and more complicated ones for
419  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
420  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
421  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
422  #define BACKCHAR(eptr)  /* #define GETCHARLENTEST(c, eptr, len) */
423    /* #define BACKCHAR(eptr) */
424    
425  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
426    
427    /* These macros were originally written in the form of loops that used data
428    from the tables whose names start with _pcre_utf8_table. They were rewritten by
429    a user so as not to use loops, because in some environments this gives a
430    significant performance advantage, and it seems never to do any harm. */
431    
432    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
433    advancing the pointer. */
434    
435    #define GETUTF8(c, eptr) \
436        { \
437        if ((c & 0x20) == 0) \
438          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
439        else if ((c & 0x10) == 0) \
440          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
441        else if ((c & 0x08) == 0) \
442          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
443          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
444        else if ((c & 0x04) == 0) \
445          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
446              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
447              (eptr[4] & 0x3f); \
448        else \
449          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
450              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
451              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
452        }
453    
454  /* Get the next UTF-8 character, not advancing the pointer. This is called when  /* Get the next UTF-8 character, not advancing the pointer. This is called when
455  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
456    
457  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
458    c = *eptr; \    c = *eptr; \
459    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8(c, eptr);
     { \  
     int gcii; \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     for (gcii = 1; gcii <= gcaa; gcii++) \  
       { \  
       gcss -= 6; \  
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     }  
460    
461  /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the  /* Get the next UTF-8 character, testing for UTF-8 mode, and not advancing the
462  pointer. */  pointer. */
463    
464  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
465    c = *eptr; \    c = *eptr; \
466    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8(c, eptr);
467    
468    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
469    the pointer. */
470    
471    #define GETUTF8INC(c, eptr) \
472      { \      { \
473      int gcii; \      if ((c & 0x20) == 0) \
474      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
475      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
476      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        { \
477      for (gcii = 1; gcii <= gcaa; gcii++) \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
478          eptr += 2; \
479          } \
480        else if ((c & 0x08) == 0) \
481          { \
482          c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
483              ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
484          eptr += 3; \
485          } \
486        else if ((c & 0x04) == 0) \
487        { \        { \
488        gcss -= 6; \        c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
489        c |= (eptr[gcii] & 0x3f) << gcss; \            ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
490              (eptr[3] & 0x3f); \
491          eptr += 4; \
492          } \
493        else \
494          { \
495          c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
496              ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
497              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
498          eptr += 5; \
499        } \        } \
500      }      }
501    
# Line 357  know we are in UTF-8 mode. */ Line 504  know we are in UTF-8 mode. */
504    
505  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
506    c = *eptr++; \    c = *eptr++; \
507    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8INC(c, eptr);
     { \  
     int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     while (gcaa-- > 0) \  
       { \  
       gcss -= 6; \  
       c |= (*eptr++ & 0x3f) << gcss; \  
       } \  
     }  
508    
509  /* Get the next character, testing for UTF-8 mode, and advancing the pointer */  /* Get the next character, testing for UTF-8 mode, and advancing the pointer.
510    This is called when we don't know if we are in UTF-8 mode. */
511    
512  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
513    c = *eptr++; \    c = *eptr++; \
514    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8INC(c, eptr);
515    
516    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
517    advancing the pointer, incrementing the length. */
518    
519    #define GETUTF8LEN(c, eptr, len) \
520      { \      { \
521      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
     int gcss = 6*gcaa; \  
     c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
     while (gcaa-- > 0) \  
522        { \        { \
523        gcss -= 6; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
524        c |= (*eptr++ & 0x3f) << gcss; \        len++; \
525          } \
526        else if ((c & 0x10)  == 0) \
527          { \
528          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
529          len += 2; \
530          } \
531        else if ((c & 0x08)  == 0) \
532          {\
533          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
534              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
535          len += 3; \
536          } \
537        else if ((c & 0x04)  == 0) \
538          { \
539          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
540              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
541              (eptr[4] & 0x3f); \
542          len += 4; \
543          } \
544        else \
545          {\
546          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
547              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
548              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
549          len += 5; \
550        } \        } \
551      }      }
552    
# Line 390  if there are extra bytes. This is called Line 555  if there are extra bytes. This is called
555    
556  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
557    c = *eptr; \    c = *eptr; \
558    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
559      { \  
560      int gcii; \  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
561      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  pointer, incrementing length if there are extra bytes. This is called when we
562      int gcss = 6*gcaa; \  do not know if we are in UTF-8 mode. */
563      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
564      for (gcii = 1; gcii <= gcaa; gcii++) \  #define GETCHARLENTEST(c, eptr, len) \
565        { \    c = *eptr; \
566        gcss -= 6; \    if (utf8 && c >= 0xc0) GETUTF8LEN(c, eptr, len);
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
567    
568  /* 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
569  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
570    because almost all calls are already within a block of UTF-8 only code. */
571    
572  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
573    
574  #endif  #endif  /* SUPPORT_UTF8 */
575    
576    
577  /* In case there is no definition of offsetof() provided - though any proper  /* In case there is no definition of offsetof() provided - though any proper
# Line 424  Standard C system should have one. */ Line 586  Standard C system should have one. */
586    
587  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
588    
589  /* Private options flags start at the most significant end of the four bytes.  /* Private flags containing information about the compiled regex. They used to
590  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
591  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
592  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
593  bits. */  compatibility. */
594    
595  #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
596  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
597  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
598  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
599  #define PCRE_JCHANGED      0x08000000  /* j option changes within regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
600    #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
601    
602  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
603    
604  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */
605    #define PCRE_STUDY_MINLEN   0x02     /* a minimum length field exists */
606    
607  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
608  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
609    
610  #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| \
611                               PCRE_NEWLINE_ANYCRLF)
612    
613  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
614    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
615     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
616     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| \
617     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
618       PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
619    
620  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
621    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
622     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
623       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
624    
625  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
626    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
627     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
628       PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
629       PCRE_NO_START_OPTIMIZE)
630    
631  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
632    
# Line 482  variable-length repeat, or a anything ot Line 651  variable-length repeat, or a anything ot
651  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
652  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
653    
654  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
655    environments where these macros are defined elsewhere. Unfortunately, there
656    is no way to do the same for the typedef. */
657    
658  typedef int BOOL;  typedef int BOOL;
659    
660    #ifndef FALSE
661  #define FALSE   0  #define FALSE   0
662  #define TRUE    1  #define TRUE    1
663    #endif
664    
665    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
666    character constants like '*' because the compiler would emit their EBCDIC code,
667    which is different from their ASCII/UTF-8 code. Instead we define macros for
668    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
669    is enabled. When UTF-8 support is not enabled, the definitions use character
670    literals. Both character and string versions of each character are needed, and
671    there are some longer strings as well.
672    
673    This means that, on EBCDIC platforms, the PCRE library can handle either
674    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
675    would need different lookups depending on whether PCRE_UTF8 was set or not.
676    This would make it impossible to use characters in switch/case statements,
677    which would reduce performance. For a theoretical use (which nobody has asked
678    for) in a minority area (EBCDIC platforms), this is not sensible. Any
679    application that did need both could compile two versions of the library, using
680    macros to give the functions distinct names. */
681    
682    #ifndef SUPPORT_UTF8
683    
684    /* UTF-8 support is not enabled; use the platform-dependent character literals
685    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
686    
687    #define CHAR_HT                     '\t'
688    #define CHAR_VT                     '\v'
689    #define CHAR_FF                     '\f'
690    #define CHAR_CR                     '\r'
691    #define CHAR_NL                     '\n'
692    #define CHAR_BS                     '\b'
693    #define CHAR_BEL                    '\a'
694    #ifdef EBCDIC
695    #define CHAR_ESC                    '\047'
696    #define CHAR_DEL                    '\007'
697    #else
698    #define CHAR_ESC                    '\033'
699    #define CHAR_DEL                    '\177'
700    #endif
701    
702    #define CHAR_SPACE                  ' '
703    #define CHAR_EXCLAMATION_MARK       '!'
704    #define CHAR_QUOTATION_MARK         '"'
705    #define CHAR_NUMBER_SIGN            '#'
706    #define CHAR_DOLLAR_SIGN            '$'
707    #define CHAR_PERCENT_SIGN           '%'
708    #define CHAR_AMPERSAND              '&'
709    #define CHAR_APOSTROPHE             '\''
710    #define CHAR_LEFT_PARENTHESIS       '('
711    #define CHAR_RIGHT_PARENTHESIS      ')'
712    #define CHAR_ASTERISK               '*'
713    #define CHAR_PLUS                   '+'
714    #define CHAR_COMMA                  ','
715    #define CHAR_MINUS                  '-'
716    #define CHAR_DOT                    '.'
717    #define CHAR_SLASH                  '/'
718    #define CHAR_0                      '0'
719    #define CHAR_1                      '1'
720    #define CHAR_2                      '2'
721    #define CHAR_3                      '3'
722    #define CHAR_4                      '4'
723    #define CHAR_5                      '5'
724    #define CHAR_6                      '6'
725    #define CHAR_7                      '7'
726    #define CHAR_8                      '8'
727    #define CHAR_9                      '9'
728    #define CHAR_COLON                  ':'
729    #define CHAR_SEMICOLON              ';'
730    #define CHAR_LESS_THAN_SIGN         '<'
731    #define CHAR_EQUALS_SIGN            '='
732    #define CHAR_GREATER_THAN_SIGN      '>'
733    #define CHAR_QUESTION_MARK          '?'
734    #define CHAR_COMMERCIAL_AT          '@'
735    #define CHAR_A                      'A'
736    #define CHAR_B                      'B'
737    #define CHAR_C                      'C'
738    #define CHAR_D                      'D'
739    #define CHAR_E                      'E'
740    #define CHAR_F                      'F'
741    #define CHAR_G                      'G'
742    #define CHAR_H                      'H'
743    #define CHAR_I                      'I'
744    #define CHAR_J                      'J'
745    #define CHAR_K                      'K'
746    #define CHAR_L                      'L'
747    #define CHAR_M                      'M'
748    #define CHAR_N                      'N'
749    #define CHAR_O                      'O'
750    #define CHAR_P                      'P'
751    #define CHAR_Q                      'Q'
752    #define CHAR_R                      'R'
753    #define CHAR_S                      'S'
754    #define CHAR_T                      'T'
755    #define CHAR_U                      'U'
756    #define CHAR_V                      'V'
757    #define CHAR_W                      'W'
758    #define CHAR_X                      'X'
759    #define CHAR_Y                      'Y'
760    #define CHAR_Z                      'Z'
761    #define CHAR_LEFT_SQUARE_BRACKET    '['
762    #define CHAR_BACKSLASH              '\\'
763    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
764    #define CHAR_CIRCUMFLEX_ACCENT      '^'
765    #define CHAR_UNDERSCORE             '_'
766    #define CHAR_GRAVE_ACCENT           '`'
767    #define CHAR_a                      'a'
768    #define CHAR_b                      'b'
769    #define CHAR_c                      'c'
770    #define CHAR_d                      'd'
771    #define CHAR_e                      'e'
772    #define CHAR_f                      'f'
773    #define CHAR_g                      'g'
774    #define CHAR_h                      'h'
775    #define CHAR_i                      'i'
776    #define CHAR_j                      'j'
777    #define CHAR_k                      'k'
778    #define CHAR_l                      'l'
779    #define CHAR_m                      'm'
780    #define CHAR_n                      'n'
781    #define CHAR_o                      'o'
782    #define CHAR_p                      'p'
783    #define CHAR_q                      'q'
784    #define CHAR_r                      'r'
785    #define CHAR_s                      's'
786    #define CHAR_t                      't'
787    #define CHAR_u                      'u'
788    #define CHAR_v                      'v'
789    #define CHAR_w                      'w'
790    #define CHAR_x                      'x'
791    #define CHAR_y                      'y'
792    #define CHAR_z                      'z'
793    #define CHAR_LEFT_CURLY_BRACKET     '{'
794    #define CHAR_VERTICAL_LINE          '|'
795    #define CHAR_RIGHT_CURLY_BRACKET    '}'
796    #define CHAR_TILDE                  '~'
797    
798    #define STR_HT                      "\t"
799    #define STR_VT                      "\v"
800    #define STR_FF                      "\f"
801    #define STR_CR                      "\r"
802    #define STR_NL                      "\n"
803    #define STR_BS                      "\b"
804    #define STR_BEL                     "\a"
805    #ifdef EBCDIC
806    #define STR_ESC                     "\047"
807    #define STR_DEL                     "\007"
808    #else
809    #define STR_ESC                     "\033"
810    #define STR_DEL                     "\177"
811    #endif
812    
813    #define STR_SPACE                   " "
814    #define STR_EXCLAMATION_MARK        "!"
815    #define STR_QUOTATION_MARK          "\""
816    #define STR_NUMBER_SIGN             "#"
817    #define STR_DOLLAR_SIGN             "$"
818    #define STR_PERCENT_SIGN            "%"
819    #define STR_AMPERSAND               "&"
820    #define STR_APOSTROPHE              "'"
821    #define STR_LEFT_PARENTHESIS        "("
822    #define STR_RIGHT_PARENTHESIS       ")"
823    #define STR_ASTERISK                "*"
824    #define STR_PLUS                    "+"
825    #define STR_COMMA                   ","
826    #define STR_MINUS                   "-"
827    #define STR_DOT                     "."
828    #define STR_SLASH                   "/"
829    #define STR_0                       "0"
830    #define STR_1                       "1"
831    #define STR_2                       "2"
832    #define STR_3                       "3"
833    #define STR_4                       "4"
834    #define STR_5                       "5"
835    #define STR_6                       "6"
836    #define STR_7                       "7"
837    #define STR_8                       "8"
838    #define STR_9                       "9"
839    #define STR_COLON                   ":"
840    #define STR_SEMICOLON               ";"
841    #define STR_LESS_THAN_SIGN          "<"
842    #define STR_EQUALS_SIGN             "="
843    #define STR_GREATER_THAN_SIGN       ">"
844    #define STR_QUESTION_MARK           "?"
845    #define STR_COMMERCIAL_AT           "@"
846    #define STR_A                       "A"
847    #define STR_B                       "B"
848    #define STR_C                       "C"
849    #define STR_D                       "D"
850    #define STR_E                       "E"
851    #define STR_F                       "F"
852    #define STR_G                       "G"
853    #define STR_H                       "H"
854    #define STR_I                       "I"
855    #define STR_J                       "J"
856    #define STR_K                       "K"
857    #define STR_L                       "L"
858    #define STR_M                       "M"
859    #define STR_N                       "N"
860    #define STR_O                       "O"
861    #define STR_P                       "P"
862    #define STR_Q                       "Q"
863    #define STR_R                       "R"
864    #define STR_S                       "S"
865    #define STR_T                       "T"
866    #define STR_U                       "U"
867    #define STR_V                       "V"
868    #define STR_W                       "W"
869    #define STR_X                       "X"
870    #define STR_Y                       "Y"
871    #define STR_Z                       "Z"
872    #define STR_LEFT_SQUARE_BRACKET     "["
873    #define STR_BACKSLASH               "\\"
874    #define STR_RIGHT_SQUARE_BRACKET    "]"
875    #define STR_CIRCUMFLEX_ACCENT       "^"
876    #define STR_UNDERSCORE              "_"
877    #define STR_GRAVE_ACCENT            "`"
878    #define STR_a                       "a"
879    #define STR_b                       "b"
880    #define STR_c                       "c"
881    #define STR_d                       "d"
882    #define STR_e                       "e"
883    #define STR_f                       "f"
884    #define STR_g                       "g"
885    #define STR_h                       "h"
886    #define STR_i                       "i"
887    #define STR_j                       "j"
888    #define STR_k                       "k"
889    #define STR_l                       "l"
890    #define STR_m                       "m"
891    #define STR_n                       "n"
892    #define STR_o                       "o"
893    #define STR_p                       "p"
894    #define STR_q                       "q"
895    #define STR_r                       "r"
896    #define STR_s                       "s"
897    #define STR_t                       "t"
898    #define STR_u                       "u"
899    #define STR_v                       "v"
900    #define STR_w                       "w"
901    #define STR_x                       "x"
902    #define STR_y                       "y"
903    #define STR_z                       "z"
904    #define STR_LEFT_CURLY_BRACKET      "{"
905    #define STR_VERTICAL_LINE           "|"
906    #define STR_RIGHT_CURLY_BRACKET     "}"
907    #define STR_TILDE                   "~"
908    
909    #define STRING_ACCEPT0              "ACCEPT\0"
910    #define STRING_COMMIT0              "COMMIT\0"
911    #define STRING_F0                   "F\0"
912    #define STRING_FAIL0                "FAIL\0"
913    #define STRING_MARK0                "MARK\0"
914    #define STRING_PRUNE0               "PRUNE\0"
915    #define STRING_SKIP0                "SKIP\0"
916    #define STRING_THEN                 "THEN"
917    
918    #define STRING_alpha0               "alpha\0"
919    #define STRING_lower0               "lower\0"
920    #define STRING_upper0               "upper\0"
921    #define STRING_alnum0               "alnum\0"
922    #define STRING_ascii0               "ascii\0"
923    #define STRING_blank0               "blank\0"
924    #define STRING_cntrl0               "cntrl\0"
925    #define STRING_digit0               "digit\0"
926    #define STRING_graph0               "graph\0"
927    #define STRING_print0               "print\0"
928    #define STRING_punct0               "punct\0"
929    #define STRING_space0               "space\0"
930    #define STRING_word0                "word\0"
931    #define STRING_xdigit               "xdigit"
932    
933    #define STRING_DEFINE               "DEFINE"
934    
935    #define STRING_CR_RIGHTPAR             "CR)"
936    #define STRING_LF_RIGHTPAR             "LF)"
937    #define STRING_CRLF_RIGHTPAR           "CRLF)"
938    #define STRING_ANY_RIGHTPAR            "ANY)"
939    #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
940    #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
941    #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
942    #define STRING_UTF8_RIGHTPAR           "UTF8)"
943    #define STRING_UCP_RIGHTPAR            "UCP)"
944    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
945    
946    #else  /* SUPPORT_UTF8 */
947    
948    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
949    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
950    only. */
951    
952    #define CHAR_HT                     '\011'
953    #define CHAR_VT                     '\013'
954    #define CHAR_FF                     '\014'
955    #define CHAR_CR                     '\015'
956    #define CHAR_NL                     '\012'
957    #define CHAR_BS                     '\010'
958    #define CHAR_BEL                    '\007'
959    #define CHAR_ESC                    '\033'
960    #define CHAR_DEL                    '\177'
961    
962    #define CHAR_SPACE                  '\040'
963    #define CHAR_EXCLAMATION_MARK       '\041'
964    #define CHAR_QUOTATION_MARK         '\042'
965    #define CHAR_NUMBER_SIGN            '\043'
966    #define CHAR_DOLLAR_SIGN            '\044'
967    #define CHAR_PERCENT_SIGN           '\045'
968    #define CHAR_AMPERSAND              '\046'
969    #define CHAR_APOSTROPHE             '\047'
970    #define CHAR_LEFT_PARENTHESIS       '\050'
971    #define CHAR_RIGHT_PARENTHESIS      '\051'
972    #define CHAR_ASTERISK               '\052'
973    #define CHAR_PLUS                   '\053'
974    #define CHAR_COMMA                  '\054'
975    #define CHAR_MINUS                  '\055'
976    #define CHAR_DOT                    '\056'
977    #define CHAR_SLASH                  '\057'
978    #define CHAR_0                      '\060'
979    #define CHAR_1                      '\061'
980    #define CHAR_2                      '\062'
981    #define CHAR_3                      '\063'
982    #define CHAR_4                      '\064'
983    #define CHAR_5                      '\065'
984    #define CHAR_6                      '\066'
985    #define CHAR_7                      '\067'
986    #define CHAR_8                      '\070'
987    #define CHAR_9                      '\071'
988    #define CHAR_COLON                  '\072'
989    #define CHAR_SEMICOLON              '\073'
990    #define CHAR_LESS_THAN_SIGN         '\074'
991    #define CHAR_EQUALS_SIGN            '\075'
992    #define CHAR_GREATER_THAN_SIGN      '\076'
993    #define CHAR_QUESTION_MARK          '\077'
994    #define CHAR_COMMERCIAL_AT          '\100'
995    #define CHAR_A                      '\101'
996    #define CHAR_B                      '\102'
997    #define CHAR_C                      '\103'
998    #define CHAR_D                      '\104'
999    #define CHAR_E                      '\105'
1000    #define CHAR_F                      '\106'
1001    #define CHAR_G                      '\107'
1002    #define CHAR_H                      '\110'
1003    #define CHAR_I                      '\111'
1004    #define CHAR_J                      '\112'
1005    #define CHAR_K                      '\113'
1006    #define CHAR_L                      '\114'
1007    #define CHAR_M                      '\115'
1008    #define CHAR_N                      '\116'
1009    #define CHAR_O                      '\117'
1010    #define CHAR_P                      '\120'
1011    #define CHAR_Q                      '\121'
1012    #define CHAR_R                      '\122'
1013    #define CHAR_S                      '\123'
1014    #define CHAR_T                      '\124'
1015    #define CHAR_U                      '\125'
1016    #define CHAR_V                      '\126'
1017    #define CHAR_W                      '\127'
1018    #define CHAR_X                      '\130'
1019    #define CHAR_Y                      '\131'
1020    #define CHAR_Z                      '\132'
1021    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
1022    #define CHAR_BACKSLASH              '\134'
1023    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
1024    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
1025    #define CHAR_UNDERSCORE             '\137'
1026    #define CHAR_GRAVE_ACCENT           '\140'
1027    #define CHAR_a                      '\141'
1028    #define CHAR_b                      '\142'
1029    #define CHAR_c                      '\143'
1030    #define CHAR_d                      '\144'
1031    #define CHAR_e                      '\145'
1032    #define CHAR_f                      '\146'
1033    #define CHAR_g                      '\147'
1034    #define CHAR_h                      '\150'
1035    #define CHAR_i                      '\151'
1036    #define CHAR_j                      '\152'
1037    #define CHAR_k                      '\153'
1038    #define CHAR_l                      '\154'
1039    #define CHAR_m                      '\155'
1040    #define CHAR_n                      '\156'
1041    #define CHAR_o                      '\157'
1042    #define CHAR_p                      '\160'
1043    #define CHAR_q                      '\161'
1044    #define CHAR_r                      '\162'
1045    #define CHAR_s                      '\163'
1046    #define CHAR_t                      '\164'
1047    #define CHAR_u                      '\165'
1048    #define CHAR_v                      '\166'
1049    #define CHAR_w                      '\167'
1050    #define CHAR_x                      '\170'
1051    #define CHAR_y                      '\171'
1052    #define CHAR_z                      '\172'
1053    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1054    #define CHAR_VERTICAL_LINE          '\174'
1055    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1056    #define CHAR_TILDE                  '\176'
1057    
1058    #define STR_HT                      "\011"
1059    #define STR_VT                      "\013"
1060    #define STR_FF                      "\014"
1061    #define STR_CR                      "\015"
1062    #define STR_NL                      "\012"
1063    #define STR_BS                      "\010"
1064    #define STR_BEL                     "\007"
1065    #define STR_ESC                     "\033"
1066    #define STR_DEL                     "\177"
1067    
1068    #define STR_SPACE                   "\040"
1069    #define STR_EXCLAMATION_MARK        "\041"
1070    #define STR_QUOTATION_MARK          "\042"
1071    #define STR_NUMBER_SIGN             "\043"
1072    #define STR_DOLLAR_SIGN             "\044"
1073    #define STR_PERCENT_SIGN            "\045"
1074    #define STR_AMPERSAND               "\046"
1075    #define STR_APOSTROPHE              "\047"
1076    #define STR_LEFT_PARENTHESIS        "\050"
1077    #define STR_RIGHT_PARENTHESIS       "\051"
1078    #define STR_ASTERISK                "\052"
1079    #define STR_PLUS                    "\053"
1080    #define STR_COMMA                   "\054"
1081    #define STR_MINUS                   "\055"
1082    #define STR_DOT                     "\056"
1083    #define STR_SLASH                   "\057"
1084    #define STR_0                       "\060"
1085    #define STR_1                       "\061"
1086    #define STR_2                       "\062"
1087    #define STR_3                       "\063"
1088    #define STR_4                       "\064"
1089    #define STR_5                       "\065"
1090    #define STR_6                       "\066"
1091    #define STR_7                       "\067"
1092    #define STR_8                       "\070"
1093    #define STR_9                       "\071"
1094    #define STR_COLON                   "\072"
1095    #define STR_SEMICOLON               "\073"
1096    #define STR_LESS_THAN_SIGN          "\074"
1097    #define STR_EQUALS_SIGN             "\075"
1098    #define STR_GREATER_THAN_SIGN       "\076"
1099    #define STR_QUESTION_MARK           "\077"
1100    #define STR_COMMERCIAL_AT           "\100"
1101    #define STR_A                       "\101"
1102    #define STR_B                       "\102"
1103    #define STR_C                       "\103"
1104    #define STR_D                       "\104"
1105    #define STR_E                       "\105"
1106    #define STR_F                       "\106"
1107    #define STR_G                       "\107"
1108    #define STR_H                       "\110"
1109    #define STR_I                       "\111"
1110    #define STR_J                       "\112"
1111    #define STR_K                       "\113"
1112    #define STR_L                       "\114"
1113    #define STR_M                       "\115"
1114    #define STR_N                       "\116"
1115    #define STR_O                       "\117"
1116    #define STR_P                       "\120"
1117    #define STR_Q                       "\121"
1118    #define STR_R                       "\122"
1119    #define STR_S                       "\123"
1120    #define STR_T                       "\124"
1121    #define STR_U                       "\125"
1122    #define STR_V                       "\126"
1123    #define STR_W                       "\127"
1124    #define STR_X                       "\130"
1125    #define STR_Y                       "\131"
1126    #define STR_Z                       "\132"
1127    #define STR_LEFT_SQUARE_BRACKET     "\133"
1128    #define STR_BACKSLASH               "\134"
1129    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1130    #define STR_CIRCUMFLEX_ACCENT       "\136"
1131    #define STR_UNDERSCORE              "\137"
1132    #define STR_GRAVE_ACCENT            "\140"
1133    #define STR_a                       "\141"
1134    #define STR_b                       "\142"
1135    #define STR_c                       "\143"
1136    #define STR_d                       "\144"
1137    #define STR_e                       "\145"
1138    #define STR_f                       "\146"
1139    #define STR_g                       "\147"
1140    #define STR_h                       "\150"
1141    #define STR_i                       "\151"
1142    #define STR_j                       "\152"
1143    #define STR_k                       "\153"
1144    #define STR_l                       "\154"
1145    #define STR_m                       "\155"
1146    #define STR_n                       "\156"
1147    #define STR_o                       "\157"
1148    #define STR_p                       "\160"
1149    #define STR_q                       "\161"
1150    #define STR_r                       "\162"
1151    #define STR_s                       "\163"
1152    #define STR_t                       "\164"
1153    #define STR_u                       "\165"
1154    #define STR_v                       "\166"
1155    #define STR_w                       "\167"
1156    #define STR_x                       "\170"
1157    #define STR_y                       "\171"
1158    #define STR_z                       "\172"
1159    #define STR_LEFT_CURLY_BRACKET      "\173"
1160    #define STR_VERTICAL_LINE           "\174"
1161    #define STR_RIGHT_CURLY_BRACKET     "\175"
1162    #define STR_TILDE                   "\176"
1163    
1164    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1165    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1166    #define STRING_F0                   STR_F "\0"
1167    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1168    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1169    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1170    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1171    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1172    
1173    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1174    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1175    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1176    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1177    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1178    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1179    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1180    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1181    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1182    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1183    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1184    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1185    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1186    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1187    
1188    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1189    
1190    #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1191    #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1192    #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1193    #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1194    #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1195    #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
1196    #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
1197    #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1198    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1199    #define STRING_NO_START_OPT_RIGHTPAR   STR_N STR_O STR_UNDERSCORE STR_S STR_T STR_A STR_R STR_T STR_UNDERSCORE STR_O STR_P STR_T STR_RIGHT_PARENTHESIS
1200    
1201    #endif  /* SUPPORT_UTF8 */
1202    
1203  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1204    
1205  #ifndef ESC_e  #ifndef ESC_e
1206  #define ESC_e 27  #define ESC_e CHAR_ESC
1207  #endif  #endif
1208    
1209  #ifndef ESC_f  #ifndef ESC_f
1210  #define ESC_f '\f'  #define ESC_f CHAR_FF
1211  #endif  #endif
1212    
1213  #ifndef ESC_n  #ifndef ESC_n
1214  #define ESC_n '\n'  #define ESC_n CHAR_NL
1215  #endif  #endif
1216    
1217  #ifndef ESC_r  #ifndef ESC_r
1218  #define ESC_r '\r'  #define ESC_r CHAR_CR
1219  #endif  #endif
1220    
1221  /* 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
1222  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1223    
1224  #ifndef ESC_tee  #ifndef ESC_tee
1225  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1226  #endif  #endif
1227    
1228  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
1229    
1230  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1231  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1232  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1233  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1234  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1235    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1236    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1237    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1238    #define PT_WORD       8    /* Word - L plus N plus underscore */
1239    
1240  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
1241  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 537  contain UTF-8 characters with values gre Line 1252  contain UTF-8 characters with values gre
1252  /* These are escaped items that aren't just an encoding of a particular data  /* These are escaped items that aren't just an encoding of a particular data
1253  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
1254  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
1255  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_ALLANY because it
1256  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1257  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  used for [^] in JavaScript compatibility mode. In non-DOTALL mode, "." behaves
1258  There are two tests in the code for an escape greater than ESC_b and less than  like \N.
1259  ESC_Z to detect the types that may be repeated. These are the types that  
1260  consume characters. If any new escapes are put in between that don't consume a  The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1261  character, that code will have to change. */  when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1262    They must be contiguous, and remain in order so that the replacements can be
1263  enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,  looked up from a table.
1264         ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_X, ESC_Z, ESC_z,  
1265         ESC_E, ESC_Q, ESC_k, ESC_REF };  The final escape must be ESC_REF as subsequent values are used for
1266    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1267    greater than ESC_b and less than ESC_Z to detect the types that may be
1268    repeated. These are the types that consume characters. If any new escapes are
1269    put in between that don't consume a character, that code will have to change.
1270    */
1271    
1272    enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
1273           ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1274           ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1275           ESC_E, ESC_Q, ESC_g, ESC_k,
1276           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1277           ESC_REF };
1278    
1279  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to  
1280  OP_EOD must correspond in order to the list of escapes immediately above.  OP_EOD must correspond in order to the list of escapes immediately above.
1281    
 To keep stored, compiled patterns compatible, new opcodes should be added  
 immediately before OP_BRA, where (since release 7.0) a gap is left for this  
 purpose.  
   
1282  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1283  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
1284  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1285    
1286  enum {  enum {
1287    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 569  enum { Line 1290  enum {
1290    
1291    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
1292    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
1293    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
1294    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
1295    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
1296    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
1297    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
1298    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
1299    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
1300    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
1301    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
1302    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character except newline */
1303    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ALLANY,         /* 13 Match any character */
1304    OP_PROP,           /* 14 \p (Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1305    OP_ANYNL,          /* 15 \R (any newline sequence) */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1306    OP_EXTUNI,         /* 16 \X (extended Unicode sequence */    OP_PROP,           /* 16 \p (Unicode property) */
1307    OP_EODN,           /* 17 End of data or \n at end of data: \Z. */    OP_ANYNL,          /* 17 \R (any newline sequence) */
1308    OP_EOD,            /* 18 End of data: \z */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1309      OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1310    OP_OPT,            /* 19 Set runtime options */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1311    OP_CIRC,           /* 20 Start of line - varies with multiline switch */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1312    OP_DOLL,           /* 21 End of line - varies with multiline switch */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1313    OP_CHAR,           /* 22 Match one character, casefully */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1314    OP_CHARNC,         /* 23 Match one character, caselessly */    OP_EOD,            /* 24 End of data: \z */
1315    OP_NOT,            /* 24 Match one character, not the following one */  
1316      OP_OPT,            /* 25 Set runtime options */
1317    OP_STAR,           /* 25 The maximizing and minimizing versions of */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
1318    OP_MINSTAR,        /* 26 these six opcodes must come in pairs, with */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
1319    OP_PLUS,           /* 27 the minimizing one second. */    OP_CHAR,           /* 28 Match one character, casefully */
1320    OP_MINPLUS,        /* 28 This first set applies to single characters.*/    OP_CHARNC,         /* 29 Match one character, caselessly */
1321    OP_QUERY,          /* 29 */    OP_NOT,            /* 30 Match one character, not the following one */
1322    OP_MINQUERY,       /* 30 */  
1323      OP_STAR,           /* 31 The maximizing and minimizing versions of */
1324    OP_UPTO,           /* 31 From 0 to n matches */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
1325    OP_MINUPTO,        /* 32 */    OP_PLUS,           /* 33 the minimizing one second. */
1326    OP_EXACT,          /* 33 Exactly n matches */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
1327      OP_QUERY,          /* 35 */
1328    OP_POSSTAR,        /* 34 Possessified star */    OP_MINQUERY,       /* 36 */
1329    OP_POSPLUS,        /* 35 Possessified plus */  
1330    OP_POSQUERY,       /* 36 Posesssified query */    OP_UPTO,           /* 37 From 0 to n matches */
1331    OP_POSUPTO,        /* 37 Possessified upto */    OP_MINUPTO,        /* 38 */
1332      OP_EXACT,          /* 39 Exactly n matches */
1333    OP_NOTSTAR,        /* 38 The maximizing and minimizing versions of */  
1334    OP_NOTMINSTAR,     /* 39 these six opcodes must come in pairs, with */    OP_POSSTAR,        /* 40 Possessified star */
1335    OP_NOTPLUS,        /* 40 the minimizing one second. They must be in */    OP_POSPLUS,        /* 41 Possessified plus */
1336    OP_NOTMINPLUS,     /* 41 exactly the same order as those above. */    OP_POSQUERY,       /* 42 Posesssified query */
1337    OP_NOTQUERY,       /* 42 This set applies to "not" single characters. */    OP_POSUPTO,        /* 43 Possessified upto */
1338    OP_NOTMINQUERY,    /* 43 */  
1339      OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
1340    OP_NOTUPTO,        /* 44 From 0 to n matches */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
1341    OP_NOTMINUPTO,     /* 45 */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
1342    OP_NOTEXACT,       /* 46 Exactly n matches */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
1343      OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
1344    OP_NOTPOSSTAR,     /* 47 Possessified versions */    OP_NOTMINQUERY,    /* 49 */
1345    OP_NOTPOSPLUS,     /* 48 */  
1346    OP_NOTPOSQUERY,    /* 49 */    OP_NOTUPTO,        /* 50 From 0 to n matches */
1347    OP_NOTPOSUPTO,     /* 50 */    OP_NOTMINUPTO,     /* 51 */
1348      OP_NOTEXACT,       /* 52 Exactly n matches */
1349    OP_TYPESTAR,       /* 51 The maximizing and minimizing versions of */  
1350    OP_TYPEMINSTAR,    /* 52 these six opcodes must come in pairs, with */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
1351    OP_TYPEPLUS,       /* 53 the minimizing one second. These codes must */    OP_NOTPOSPLUS,     /* 54 */
1352    OP_TYPEMINPLUS,    /* 54 be in exactly the same order as those above. */    OP_NOTPOSQUERY,    /* 55 */
1353    OP_TYPEQUERY,      /* 55 This set applies to character types such as \d */    OP_NOTPOSUPTO,     /* 56 */
1354    OP_TYPEMINQUERY,   /* 56 */  
1355      OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
1356    OP_TYPEUPTO,       /* 57 From 0 to n matches */    OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
1357    OP_TYPEMINUPTO,    /* 58 */    OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
1358    OP_TYPEEXACT,      /* 59 Exactly n matches */    OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
1359      OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
1360    OP_TYPEPOSSTAR,    /* 60 Possessified versions */    OP_TYPEMINQUERY,   /* 62 */
1361    OP_TYPEPOSPLUS,    /* 61 */  
1362    OP_TYPEPOSQUERY,   /* 62 */    OP_TYPEUPTO,       /* 63 From 0 to n matches */
1363    OP_TYPEPOSUPTO,    /* 63 */    OP_TYPEMINUPTO,    /* 64 */
1364      OP_TYPEEXACT,      /* 65 Exactly n matches */
1365    OP_CRSTAR,         /* 64 The maximizing and minimizing versions of */  
1366    OP_CRMINSTAR,      /* 65 all these opcodes must come in pairs, with */    OP_TYPEPOSSTAR,    /* 66 Possessified versions */
1367    OP_CRPLUS,         /* 66 the minimizing one second. These codes must */    OP_TYPEPOSPLUS,    /* 67 */
1368    OP_CRMINPLUS,      /* 67 be in exactly the same order as those above. */    OP_TYPEPOSQUERY,   /* 68 */
1369    OP_CRQUERY,        /* 68 These are for character classes and back refs */    OP_TYPEPOSUPTO,    /* 69 */
1370    OP_CRMINQUERY,     /* 69 */  
1371    OP_CRRANGE,        /* 70 These are different to the three sets above. */    OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
1372    OP_CRMINRANGE,     /* 71 */    OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
1373      OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
1374      OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
1375      OP_CRQUERY,        /* 74 These are for character classes and back refs */
1376      OP_CRMINQUERY,     /* 75 */
1377      OP_CRRANGE,        /* 76 These are different to the three sets above. */
1378      OP_CRMINRANGE,     /* 77 */
1379    
1380    OP_CLASS,          /* 72 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
1381    OP_NCLASS,         /* 73 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
1382                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
1383                             character > 255 is encountered. */                             character > 255 is encountered. */
1384    
1385    OP_XCLASS,         /* 74 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
1386                             class. This does both positive and negative. */                             class. This does both positive and negative. */
1387    
1388    OP_REF,            /* 75 Match a back reference */    OP_REF,            /* 81 Match a back reference */
1389    OP_RECURSE,        /* 76 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
1390    OP_CALLOUT,        /* 77 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
1391    
1392    OP_ALT,            /* 78 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
1393    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 */
1394    OP_KETRMAX,        /* 80 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
1395    OP_KETRMIN,        /* 81 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
1396    
1397    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
1398    
1399    OP_ASSERT,         /* 82 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
1400    OP_ASSERT_NOT,     /* 83 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
1401    OP_ASSERTBACK,     /* 84 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
1402    OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
1403    OP_REVERSE,        /* 86 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
1404    
1405    /* 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,
1406    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. */
1407    
1408    OP_ONCE,           /* 87 Atomic group */    OP_ONCE,           /* 93 Atomic group */
1409    OP_BRA,            /* 88 Start of non-capturing bracket */    OP_BRA,            /* 94 Start of non-capturing bracket */
1410    OP_CBRA,           /* 89 Start of capturing bracket */    OP_CBRA,           /* 95 Start of capturing bracket */
1411    OP_COND,           /* 90 Conditional group */    OP_COND,           /* 96 Conditional group */
1412    
1413    /* 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
1414    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1415    
1416    OP_SBRA,           /* 91 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
1417    OP_SCBRA,          /* 92 Start of capturing bracket, check empty */    OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1418    OP_SCOND,          /* 93 Conditional group, check empty */    OP_SCOND,          /* 99 Conditional group, check empty */
1419    
1420    OP_CREF,           /* 94 Used to hold a capture number as condition */    /* The next two pairs must (respectively) be kept together. */
1421    OP_RREF,           /* 95 Used to hold a recursion number as condition */  
1422    OP_DEF,            /* 96 The DEFINE condition */    OP_CREF,           /* 100 Used to hold a capture number as condition */
1423      OP_NCREF,          /* 101 Same, but generated by a name reference*/
1424      OP_RREF,           /* 102 Used to hold a recursion number as condition */
1425      OP_NRREF,          /* 103 Same, but generated by a name reference*/
1426      OP_DEF,            /* 104 The DEFINE condition */
1427    
1428      OP_BRAZERO,        /* 105 These two must remain together and in this */
1429      OP_BRAMINZERO,     /* 106 order. */
1430    
1431      /* These are backtracking control verbs */
1432    
1433      OP_MARK,           /* 107 always has an argument */
1434      OP_PRUNE,          /* 108 */
1435      OP_PRUNE_ARG,      /* 109 same, but with argument */
1436      OP_SKIP,           /* 110 */
1437      OP_SKIP_ARG,       /* 111 same, but with argument */
1438      OP_THEN,           /* 112 */
1439      OP_THEN_ARG,       /* 113 same, but with argument */
1440      OP_COMMIT,         /* 114 */
1441    
1442      /* These are forced failure and success verbs */
1443    
1444      OP_FAIL,           /* 115 */
1445      OP_ACCEPT,         /* 116 */
1446      OP_CLOSE,          /* 117 Used before OP_ACCEPT to close open captures */
1447    
1448      /* This is used to skip a subpattern with a {0} quantifier */
1449    
1450      OP_SKIPZERO,       /* 118 */
1451    
1452    OP_BRAZERO,        /* 97 These two must remain together and in this */    /* This is not an opcode, but is used to check that tables indexed by opcode
1453    OP_BRAMINZERO      /* 98 order. */    are the correct length, in order to catch updating errors - there have been
1454      some in the past. */
1455    
1456      OP_TABLE_LENGTH
1457  };  };
1458    
1459    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1460    definitions that follow must also be updated to match. There are also tables
1461    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1462    
1463    
1464  /* 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
1465  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
1466    
1467  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1468    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1469    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1470    "notprop", "prop", "anynl", "extuni",                           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1471    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
1472    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
1473    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1474    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
# Line 718  for debugging. The macro is referenced o Line 1480  for debugging. The macro is referenced o
1480    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
1481    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1482    "AssertB", "AssertB not", "Reverse",                            \    "AssertB", "AssertB not", "Reverse",                            \
1483    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1484    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1485      "Brazero", "Braminzero",                                        \
1486      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1487      "*THEN", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",                \
1488      "Close", "Skip zero"
1489    
1490    
1491  /* 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 733  in UTF-8 mode. The code that uses this t Line 1499  in UTF-8 mode. The code that uses this t
1499    
1500  #define OP_LENGTHS \  #define OP_LENGTHS \
1501    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1502    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                     */ \
1503    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1504    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1505      3, 3,                          /* \P, \p                                 */ \
1506      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1507      1,                             /* \X                                     */ \
1508    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
1509    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1510    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 777  in UTF-8 mode. The code that uses this t Line 1546  in UTF-8 mode. The code that uses this t
1546    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1547    3+LINK_SIZE,                   /* SCBRA                                  */ \    3+LINK_SIZE,                   /* SCBRA                                  */ \
1548    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1549    3,                             /* CREF                                   */ \    3, 3,                          /* CREF, NCREF                            */ \
1550    3,                             /* RREF                                   */ \    3, 3,                          /* RREF, NRREF                            */ \
1551    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1552    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1553      3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1554      1, 3,                          /* SKIP, SKIP_ARG                         */ \
1555      1+LINK_SIZE, 3+LINK_SIZE,      /* THEN, THEN_ARG                         */ \
1556      1, 1, 1, 3, 1                  /* COMMIT, FAIL, ACCEPT, CLOSE, SKIPZERO  */
1557    
1558    
1559  /* 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"
1560    condition. */
1561    
1562  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1563    
1564  /* Error code numbers. They are given names so that they can more easily be  /* Compile time error code numbers. They are given names so that they can more
1565  tracked. */  easily be tracked. When a new number is added, the table called eint in
1566    pcreposix.c must be updated. */
1567    
1568  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1569         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
1570         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
1571         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1572         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1573         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1574           ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68,
1575           ERRCOUNT };
1576    
1577  /* 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
1578  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 811  Because people can now save and re-use c Line 1588  Because people can now save and re-use c
1588  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
1589  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
1590  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1591  NOTE NOTE NOTE:  NOTE NOTE NOTE
1592  */  */
1593    
1594  typedef struct real_pcre {  typedef struct real_pcre {
1595    pcre_uint32 magic_number;    pcre_uint32 magic_number;
1596    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1597    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
1598    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
1599      pcre_uint16 dummy1;             /* For future use */
1600    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
1601    pcre_uint16 top_backref;    pcre_uint16 top_backref;
1602    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 838  remark (see NOTE above) about extending Line 1615  remark (see NOTE above) about extending
1615    
1616  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1617    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1618    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1619    uschar start_bits[32];    uschar start_bits[32];          /* Starting char bits */
1620      pcre_uint32 minlength;          /* Minimum subject length */
1621  } pcre_study_data;  } pcre_study_data;
1622    
1623    /* Structure for building a chain of open capturing subpatterns during
1624    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1625    encountered. This is also used to identify subpatterns that contain recursive
1626    back references to themselves, so that they can be made atomic. */
1627    
1628    typedef struct open_capitem {
1629      struct open_capitem *next;    /* Chain link */
1630      pcre_uint16 number;           /* Capture number */
1631      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1632    } open_capitem;
1633    
1634  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1635  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1636    
# Line 854  typedef struct compile_data { Line 1643  typedef struct compile_data {
1643    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1644    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1645    const uschar *end_pattern;    /* The end of the pattern */    const uschar *end_pattern;    /* The end of the pattern */
1646      open_capitem *open_caps;      /* Chain of open capture items */
1647    uschar *hwm;                  /* High watermark of workspace */    uschar *hwm;                  /* High watermark of workspace */
1648    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1649    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
1650    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
1651    int  bracount;                /* Count of capturing parens */    int  bracount;                /* Count of capturing parens as we compile */
1652      int  final_bracount;          /* Saved value after first pass */
1653    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
1654    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
1655    int  external_options;        /* External (initial) options */    int  external_options;        /* External (initial) options */
1656      int  external_flags;          /* External flag bits to be set */
1657    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1658    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
1659      BOOL check_lookbehind;        /* Lookbehinds need later checking */
1660    int  nltype;                  /* Newline type */    int  nltype;                  /* Newline type */
1661    int  nllen;                   /* Newline string length */    int  nllen;                   /* Newline string length */
1662    uschar nl[4];                 /* Newline string when fixed length */    uschar nl[4];                 /* Newline string when fixed length */
# Line 874  branches, for testing for left recursion Line 1667  branches, for testing for left recursion
1667    
1668  typedef struct branch_chain {  typedef struct branch_chain {
1669    struct branch_chain *outer;    struct branch_chain *outer;
1670    uschar *current;    uschar *current_branch;
1671  } branch_chain;  } branch_chain;
1672    
1673  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
# Line 884  typedef struct recursion_info { Line 1677  typedef struct recursion_info {
1677    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1678    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
1679    const uschar *after_call;     /* "Return value": points after the call in the expr */    const uschar *after_call;     /* "Return value": points after the call in the expr */
   USPTR save_start;             /* Old value of md->start_match */  
1680    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1681    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1682      int save_offset_top;          /* Current value of offset_top */
1683  } recursion_info;  } recursion_info;
1684    
 /* 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;  
   
1685  /* 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
1686  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
1687  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops. */
# Line 921  typedef struct match_data { Line 1704  typedef struct match_data {
1704    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1705    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
1706    int    nllen;                 /* Newline string length */    int    nllen;                 /* Newline string length */
1707      int    name_count;            /* Number of names in name table */
1708      int    name_entry_size;       /* Size of entry in names table */
1709      uschar *name_table;           /* Table of names */
1710    uschar nl[4];                 /* Newline string when fixed */    uschar nl[4];                 /* Newline string when fixed */
1711    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1712    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
# Line 928  typedef struct match_data { Line 1714  typedef struct match_data {
1714    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1715    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1716    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1717      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1718      BOOL   use_ucp;               /* PCRE_UCP flag */
1719    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1720    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1721    BOOL   partial;               /* PARTIAL flag */    BOOL   notempty_atstart;      /* Empty string match at start not wanted */
1722    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1723      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1724    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1725    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1726    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1727    USPTR  start_match;           /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
1728    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1729      USPTR  start_used_ptr;        /* Earliest consulted character */
1730      int    partial;               /* PARTIAL options */
1731    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1732    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1733    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
# Line 944  typedef struct match_data { Line 1735  typedef struct match_data {
1735    int    eptrn;                 /* Next free eptrblock */    int    eptrn;                 /* Next free eptrblock */
1736    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1737    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1738    struct heapframe *thisframe;  /* Used only when compiling for no recursion */    const uschar *mark;           /* Mark pointer to pass back */
1739  } match_data;  } match_data;
1740    
1741  /* 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 954  typedef struct dfa_match_data { Line 1745  typedef struct dfa_match_data {
1745    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;     /* Start of the compiled pattern */
1746    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
1747    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;    /* End of subject string */
1748      const uschar *start_used_ptr; /* Earliest consulted character */
1749    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1750      int   start_offset;           /* The start offset value */
1751    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1752    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1753    int    nltype;                /* Newline type */    int    nltype;                /* Newline type */
# Line 969  typedef struct dfa_match_data { Line 1762  typedef struct dfa_match_data {
1762  #define ctype_letter  0x02  #define ctype_letter  0x02
1763  #define ctype_digit   0x04  #define ctype_digit   0x04
1764  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1765  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1766  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1767    
1768  /* 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 997  total length. */ Line 1790  total length. */
1790  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1791    
1792  /* 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
1793  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1794    relocations in shared libraries, it now has an offset into a single string
1795    instead. */
1796    
1797  typedef struct {  typedef struct {
1798    const char *name;    pcre_uint16 name_offset;
1799    pcre_uint16 type;    pcre_uint16 type;
1800    pcre_uint16 value;    pcre_uint16 value;
1801  } ucp_type_table;  } ucp_type_table;
# Line 1018  extern const uschar _pcre_utf8_table4[]; Line 1813  extern const uschar _pcre_utf8_table4[];
1813    
1814  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1815    
1816    extern const char   _pcre_utt_names[];
1817  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1818  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1819    
# Line 1030  extern const uschar _pcre_OP_lengths[]; Line 1826  extern const uschar _pcre_OP_lengths[];
1826  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
1827  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1828    
1829  extern BOOL         _pcre_is_newline(const uschar *, const uschar *, int *,  extern const uschar *_pcre_find_bracket(const uschar *, BOOL, int);
1830                        BOOL);  extern BOOL          _pcre_is_newline(USPTR, int, USPTR, int *, BOOL);
1831  extern int          _pcre_ord2utf8(int, uschar *);  extern int           _pcre_ord2utf8(int, uschar *);
1832  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  extern real_pcre    *_pcre_try_flipped(const real_pcre *, real_pcre *,
1833                        const pcre_study_data *, pcre_study_data *);                         const pcre_study_data *, pcre_study_data *);
1834  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  extern int           _pcre_valid_utf8(USPTR, int);
1835  extern unsigned int _pcre_ucp_othercase(const unsigned int);  extern BOOL          _pcre_was_newline(USPTR, int, USPTR, int *, BOOL);
1836  extern int          _pcre_valid_utf8(const uschar *, int);  extern BOOL          _pcre_xclass(int, const uschar *);
1837  extern BOOL         _pcre_was_newline(const uschar *, const uschar *, int *,  
1838                        BOOL);  
1839  extern BOOL         _pcre_xclass(int, const uschar *);  /* Unicode character database (UCD) */
1840    
1841    typedef struct {
1842      uschar script;
1843      uschar chartype;
1844      pcre_int32 other_case;
1845    } ucd_record;
1846    
1847    extern const ucd_record  _pcre_ucd_records[];
1848    extern const uschar      _pcre_ucd_stage1[];
1849    extern const pcre_uint16 _pcre_ucd_stage2[];
1850    extern const int         _pcre_ucp_gentype[];
1851    
1852    
1853    /* UCD access macros */
1854    
1855    #define UCD_BLOCK_SIZE 128
1856    #define GET_UCD(ch) (_pcre_ucd_records + \
1857            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1858            UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
1859    
1860    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1861    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1862    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1863    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1864    
1865  #endif  #endif
1866    

Legend:
Removed from v.108  
changed lines
  Added in v.593

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12