/[pcre]/code/branches/pcre16/pcre_internal.h
ViewVC logotype

Diff of /code/branches/pcre16/pcre_internal.h

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

code/trunk/pcre_internal.h revision 101 by ph10, Tue Mar 6 15:19:44 2007 UTC code/branches/pcre16/pcre_internal.h revision 769 by zherczeg, Sun Nov 27 17:40:33 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) 1997-2011 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 40  POSSIBILITY OF SUCH DAMAGE. Line 40  POSSIBILITY OF SUCH DAMAGE.
40    
41  /* This header contains definitions that are shared between the different  /* This header contains definitions that are shared between the different
42  modules, but which are not relevant to the exported API. This includes some  modules, but which are not relevant to the exported API. This includes some
43  functions whose names all begin with "_pcre_". */  functions whose names all begin with "_pcre_" or "_pcre16_" depending on
44    the PRIV macro. */
45    
46  #ifndef PCRE_INTERNAL_H  #ifndef PCRE_INTERNAL_H
47  #define PCRE_INTERNAL_H  #define PCRE_INTERNAL_H
48    
49  /* Define DEBUG to get debugging output on stdout. */  /* Define PCRE_DEBUG to get debugging output on stdout. */
50    
51  #if 0  #if 0
52  #define DEBUG  #define PCRE_DEBUG
53    #endif
54    
55    /* PCRE is compiled as an 8 bit library if it is not requested otherwise. */
56    #ifndef COMPILE_PCRE16
57    #define COMPILE_PCRE8
58    #endif
59    
60    /* We do not support both EBCDIC and UTF-8/16 at the same time. The "configure"
61    script prevents both being selected, but not everybody uses "configure". */
62    
63    #if defined EBCDIC && (defined SUPPORT_UTF8 || defined SUPPORT_UTF16)
64    #error The use of both EBCDIC and SUPPORT_UTF8/16 is not supported.
65    #endif
66    
67    /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
68    "configure" script ensures this, but not everybody uses "configure". */
69    
70    #if defined SUPPORT_UCP && !defined SUPPORT_UTF8
71    #define SUPPORT_UTF8 1
72  #endif  #endif
73    
74  /* 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 80  It turns out that the Mac Debugging.h he
80  be absolutely sure we get our version. */  be absolutely sure we get our version. */
81    
82  #undef DPRINTF  #undef DPRINTF
83  #ifdef DEBUG  #ifdef PCRE_DEBUG
84  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
85  #else  #else
86  #define DPRINTF(p) /* Nothing */  #define DPRINTF(p) /* Nothing */
87  #endif  #endif
88    
89    
 /* Get the definitions provided by running "configure" */  
   
 #include "config.h"  
   
90  /* Standard C headers plus the external interface definition. The only time  /* Standard C headers plus the external interface definition. The only time
91  setjmp and stdarg are used is when NO_RECURSE is set. */  setjmp and stdarg are used is when NO_RECURSE is set. */
92    
93  #include <ctype.h>  #include <ctype.h>
94  #include <limits.h>  #include <limits.h>
 #include <setjmp.h>  
 #include <stdarg.h>  
95  #include <stddef.h>  #include <stddef.h>
96  #include <stdio.h>  #include <stdio.h>
97  #include <stdlib.h>  #include <stdlib.h>
98  #include <string.h>  #include <string.h>
99    
100  #ifndef PCRE_SPY  /* When compiling a DLL for Windows, the exported symbols have to be declared
101  #define PCRE_DEFINITION       /* Win32 __declspec(export) trigger for .dll */  using some MS magic. I found some useful information on this web page:
102    http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
103    information there, using __declspec(dllexport) without "extern" we have a
104    definition; with "extern" we have a declaration. The settings here override the
105    setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
106    which is all that is needed for applications (they just import the symbols). We
107    use:
108    
109      PCRE_EXP_DECL       for declarations
110      PCRE_EXP_DEFN       for definitions of exported functions
111      PCRE_EXP_DATA_DEFN  for definitions of exported variables
112    
113    The reason for the two DEFN macros is that in non-Windows environments, one
114    does not want to have "extern" before variable definitions because it leads to
115    compiler warnings. So we distinguish between functions and variables. In
116    Windows, the two should always be the same.
117    
118    The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
119    which is an application, but needs to import this file in order to "peek" at
120    internals, can #include pcre.h first to get an application's-eye view.
121    
122    In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
123    special-purpose environments) might want to stick other stuff in front of
124    exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
125    PCRE_EXP_DATA_DEFN only if they are not already set. */
126    
127    #ifndef PCRE_EXP_DECL
128    #  ifdef _WIN32
129    #    ifndef PCRE_STATIC
130    #      define PCRE_EXP_DECL       extern __declspec(dllexport)
131    #      define PCRE_EXP_DEFN       __declspec(dllexport)
132    #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
133    #    else
134    #      define PCRE_EXP_DECL       extern
135    #      define PCRE_EXP_DEFN
136    #      define PCRE_EXP_DATA_DEFN
137    #    endif
138    #  else
139    #    ifdef __cplusplus
140    #      define PCRE_EXP_DECL       extern "C"
141    #    else
142    #      define PCRE_EXP_DECL       extern
143    #    endif
144    #    ifndef PCRE_EXP_DEFN
145    #      define PCRE_EXP_DEFN       PCRE_EXP_DECL
146    #    endif
147    #    ifndef PCRE_EXP_DATA_DEFN
148    #      define PCRE_EXP_DATA_DEFN
149    #    endif
150    #  endif
151  #endif  #endif
152    
153  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* When compiling with the MSVC compiler, it is sometimes necessary to include
154    a "calling convention" before exported function names. (This is secondhand
155    information; I know nothing about MSVC myself). For example, something like
156    
157      void __cdecl function(....)
158    
159    might be needed. In order so make this easy, all the exported functions have
160    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
161    set, we ensure here that it has no effect. */
162    
163    #ifndef PCRE_CALL_CONVENTION
164    #define PCRE_CALL_CONVENTION
165    #endif
166    
167    /* We need to have types that specify unsigned 8, 16 and 32-bit integers. We
168  cannot determine these outside the compilation (e.g. by running a program as  cannot determine these outside the compilation (e.g. by running a program as
169  part of "configure") because PCRE is often cross-compiled for use on other  part of "configure") because PCRE is often cross-compiled for use on other
170  systems. Instead we make use of the maximum sizes that are available at  systems. Instead we make use of the maximum sizes that are available at
171  preprocessor time in standard C environments. */  preprocessor time in standard C environments. */
172    
173    typedef unsigned char pcre_uint8;
174    
175  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
176    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
177      typedef short pcre_int16;
178  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
179    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
180      typedef int pcre_int16;
181  #else  #else
182    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
183  #endif  #endif
184    
185  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
186    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
187      typedef int pcre_int32;
188  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
189    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
190      typedef long int pcre_int32;
191  #else  #else
192    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
193  #endif  #endif
194    
195    /* When checking for integer overflow in pcre_compile(), we need to handle
196    large integers. If a 64-bit integer type is available, we can use that.
197    Otherwise we have to cast to double, which of course requires floating point
198    arithmetic. Handle this by defining a macro for the appropriate type. If
199    stdint.h is available, include it; it may define INT64_MAX. Systems that do not
200    have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
201    by "configure". */
202    
203    #if HAVE_STDINT_H
204    #include <stdint.h>
205    #elif HAVE_INTTYPES_H
206    #include <inttypes.h>
207    #endif
208    
209    #if defined INT64_MAX || defined int64_t
210    #define INT64_OR_DOUBLE int64_t
211    #else
212    #define INT64_OR_DOUBLE double
213    #endif
214    
215  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
216  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
217  However, we leave the interface to the outside world as char *, because that  However, we leave the interface to the outside world as char * or short *,
218  should make things easier for callers. We define a short type for unsigned char  because that should make things easier for callers. We define a short type
219  to save lots of typing. I tried "uchar", but it causes problems on Digital  for the current character representation (either 8 or 16 bit) to save lots
220  Unix, where it is defined in sys/types, so use "uschar" instead. */  of typing. I tried "uchar", but it causes problems on Digital Unix, where
221    it is defined in sys/types, so use "uschar" instead. */
222    
223  typedef unsigned char uschar;  #ifdef COMPILE_PCRE8
224    
225    typedef unsigned char pcre_uchar;
226    #define IN_UCHARS(x) (x)
227    
228    #else
229    
230    #ifdef COMPILE_PCRE16
231    #if USHRT_MAX != 65535
232    /* This is a warning message. Change PCRE_SCHAR16 to a 16 bit data type in
233    pcre.h(.in) and disable (comment out) this message. */
234    #error Warning: PCRE_SCHAR16 is not a 16 bit data type.
235    #endif
236    typedef pcre_uint16 pcre_uchar;
237    #define IN_UCHARS(x) ((x) << 1)
238    
239    #else
240    #error Unsupported compiling mode
241    #endif /* COMPILE_PCRE16 */
242    
243    #endif /* COMPILE_PCRE8 */
244    
245  /* This is an unsigned int value that no character can ever have. UTF-8  /* This is an unsigned int value that no character can ever have. UTF-8
246  characters only go up to 0x7fffffff (though Unicode doesn't go beyond  characters only go up to 0x7fffffff (though Unicode doesn't go beyond
# Line 125  characters only go up to 0x7fffffff (tho Line 249  characters only go up to 0x7fffffff (tho
249  #define NOTACHAR 0xffffffff  #define NOTACHAR 0xffffffff
250    
251  /* 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,
252  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
253  newlines. NLBLOCK, PSSTART, and PSEND are defined in the various modules to  testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
254  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
255  string field names are. */  start/end of string field names are. */
256    
257  #define NLTYPE_FIXED   0     /* Newline is a fixed length string */  #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
258  #define NLTYPE_ANY     1     /* Newline is any Unicode line ending */  #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
259    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
260    
261  /* This macro checks for a newline at the given position */  /* This macro checks for a newline at the given position */
262    
263  #define IS_NEWLINE(p) \  #define IS_NEWLINE(p) \
264    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
265      ((p) < NLBLOCK->PSEND && \      ((p) < NLBLOCK->PSEND && \
266       _pcre_is_newline((p), NLBLOCK->PSEND, &(NLBLOCK->nllen), utf8) \       PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
267      ) \         utf8)) \
268      : \      : \
269      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \      ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
270       (p)[0] == NLBLOCK->nl[0] && \       (p)[0] == NLBLOCK->nl[0] && \
# Line 152  string field names are. */ Line 277  string field names are. */
277  #define WAS_NEWLINE(p) \  #define WAS_NEWLINE(p) \
278    ((NLBLOCK->nltype != NLTYPE_FIXED)? \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
279      ((p) > NLBLOCK->PSSTART && \      ((p) > NLBLOCK->PSSTART && \
280       _pcre_was_newline((p), NLBLOCK->PSSTART, &(NLBLOCK->nllen), utf8) \       PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
281      ) \         &(NLBLOCK->nllen), utf8)) \
282      : \      : \
283      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \      ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
284       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \       (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
# Line 171  used for the external interface and appe Line 296  used for the external interface and appe
296  must begin with PCRE_. */  must begin with PCRE_. */
297    
298  #ifdef CUSTOM_SUBJECT_PTR  #ifdef CUSTOM_SUBJECT_PTR
299  #define PCRE_SPTR CUSTOM_SUBJECT_PTR  #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
 #define USPTR CUSTOM_SUBJECT_PTR  
300  #else  #else
301  #define PCRE_SPTR const char *  #define PCRE_PUCHAR const pcre_uchar *
 #define USPTR const unsigned char *  
302  #endif  #endif
303    
304  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
# Line 200  option on the command line. */ Line 323  option on the command line. */
323  /* 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(),
324  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
325  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
326  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. */  
327    
328  #if ! HAVE_MEMMOVE  #ifndef HAVE_MEMMOVE
329  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
330  #if HAVE_BCOPY  #ifdef HAVE_BCOPY
331  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
332  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
333  static void *  static void *
334  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)  pcre_memmove(void *d, const void *s, size_t n)
335  {  {
336  size_t i;  size_t i;
337  dest += n;  unsigned char *dest = (unsigned char *)d;
338  src += n;  const unsigned char *src = (const unsigned char *)s;
339  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  if (dest > src)
340  return dest;    {
341      dest += n;
342      src += n;
343      for (i = 0; i < n; ++i) *(--dest) = *(--src);
344      return (void *)dest;
345      }
346    else
347      {
348      for (i = 0; i < n; ++i) *dest++ = *src++;
349      return (void *)(dest - n);
350      }
351  }  }
352  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
353  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 237  The macros are controlled by the value o Line 368  The macros are controlled by the value o
368  the config.h file, but can be overridden by using -D on the command line. This  the config.h file, but can be overridden by using -D on the command line. This
369  is automated on Unix systems via the "configure" command. */  is automated on Unix systems via the "configure" command. */
370    
371    #ifdef COMPILE_PCRE8
372    
373  #if LINK_SIZE == 2  #if LINK_SIZE == 2
374    
375  #define PUT(a,n,d)   \  #define PUT(a,n,d)   \
# Line 273  is automated on Unix systems via the "co Line 406  is automated on Unix systems via the "co
406  #define GET(a,n) \  #define GET(a,n) \
407    (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])    (((a)[n] << 24) | ((a)[(n)+1] << 16) | ((a)[(n)+2] << 8) | (a)[(n)+3])
408    
409  #define MAX_PATTERN_SIZE (1 << 30)   /* Keep it positive */  /* Keep it positive */
410    #define MAX_PATTERN_SIZE (1 << 30)
411    
412    #else
413    #error LINK_SIZE must be either 2, 3, or 4
414    #endif
415    
416    #else /* COMPILE_PCRE8 */
417    
418    #ifdef COMPILE_PCRE16
419    
420    #if LINK_SIZE == 2
421    
422    #undef LINK_SIZE
423    #define LINK_SIZE 1
424    
425    #define PUT(a,n,d)   \
426      (a[n] = (d))
427    
428    #define GET(a,n) \
429      (a[n])
430    
431    #define MAX_PATTERN_SIZE (1 << 16)
432    
433    #elif LINK_SIZE == 3 || LINK_SIZE == 4
434    
435    #undef LINK_SIZE
436    #define LINK_SIZE 2
437    
438    #define PUT(a,n,d)   \
439      (a[n] = (d) >> 16), \
440      (a[(n)+1] = (d) & 65536)
441    
442    #define GET(a,n) \
443      (((a)[n] << 16) | (a)[(n)+1])
444    
445    /* Keep it positive */
446    #define MAX_PATTERN_SIZE (1 << 30)
447    
448  #else  #else
449  #error LINK_SIZE must be either 2, 3, or 4  #error LINK_SIZE must be either 2, 3, or 4
450  #endif  #endif
451    
452    #else
453    #error Unsupported compiling mode
454    #endif /* COMPILE_PCRE16 */
455    
456    #endif /* COMPILE_PCRE8 */
457    
458  /* Convenience macro defined in terms of the others */  /* Convenience macro defined in terms of the others */
459    
# Line 290  is automated on Unix systems via the "co Line 464  is automated on Unix systems via the "co
464  offsets changes. There are used for repeat counts and for other things such as  offsets changes. There are used for repeat counts and for other things such as
465  capturing parenthesis numbers in back references. */  capturing parenthesis numbers in back references. */
466    
467    #ifdef COMPILE_PCRE8
468    
469    #define IMM2_SIZE 2
470    
471  #define PUT2(a,n,d)   \  #define PUT2(a,n,d)   \
472    a[n] = (d) >> 8; \    a[n] = (d) >> 8; \
473    a[(n)+1] = (d) & 255    a[(n)+1] = (d) & 255
# Line 297  capturing parenthesis numbers in back re Line 475  capturing parenthesis numbers in back re
475  #define GET2(a,n) \  #define GET2(a,n) \
476    (((a)[n] << 8) | (a)[(n)+1])    (((a)[n] << 8) | (a)[(n)+1])
477    
478  #define PUT2INC(a,n,d)  PUT2(a,n,d), a += 2  #else /* COMPILE_PCRE8 */
479    
480    #ifdef COMPILE_PCRE16
481    
482    #define IMM2_SIZE 1
483    
484    #define PUT2(a,n,d)   \
485       a[n] = d
486    
487    #define GET2(a,n) \
488       a[n]
489    
490    #else
491    #error Unsupported compiling mode
492    #endif /* COMPILE_PCRE16 */
493    
494    #endif /* COMPILE_PCRE8 */
495    
496    #define PUT2INC(a,n,d)  PUT2(a,n,d), a += IMM2_SIZE
497    
498  /* 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
499  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
500  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. GETCHARLENTEST is
501    not used when UTF-8 is not supported, so it is not defined, and BACKCHAR should
502    never be called in byte mode. To make sure they can never even appear when
503    UTF-8 support is omitted, we don't even define them. */
504    
505  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
506  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 310  byte-mode, and more complicated ones for Line 508  byte-mode, and more complicated ones for
508  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
509  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
510  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
511  #define BACKCHAR(eptr)  /* #define GETCHARLENTEST(c, eptr, len) */
512    /* #define BACKCHAR(eptr) */
513    
514  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
515    
516    /* These macros were originally written in the form of loops that used data
517    from the tables whose names start with PRIV(utf8_table). They were rewritten by
518    a user so as not to use loops, because in some environments this gives a
519    significant performance advantage, and it seems never to do any harm. */
520    
521    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
522    advancing the pointer. */
523    
524    #define GETUTF8(c, eptr) \
525        { \
526        if ((c & 0x20) == 0) \
527          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
528        else if ((c & 0x10) == 0) \
529          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
530        else if ((c & 0x08) == 0) \
531          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
532          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
533        else if ((c & 0x04) == 0) \
534          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
535              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
536              (eptr[4] & 0x3f); \
537        else \
538          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
539              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
540              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
541        }
542    
543  /* 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
544  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
545    
546  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
547    c = *eptr; \    c = *eptr; \
548    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; \  
       } \  
     }  
549    
550  /* 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
551  pointer. */  pointer. */
552    
553  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
554    c = *eptr; \    c = *eptr; \
555    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8(c, eptr);
556    
557    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
558    the pointer. */
559    
560    #define GETUTF8INC(c, eptr) \
561      { \      { \
562      int gcii; \      if ((c & 0x20) == 0) \
563      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
564      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
565      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        { \
566      for (gcii = 1; gcii <= gcaa; gcii++) \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
567          eptr += 2; \
568          } \
569        else if ((c & 0x08) == 0) \
570          { \
571          c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
572              ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
573          eptr += 3; \
574          } \
575        else if ((c & 0x04) == 0) \
576          { \
577          c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
578              ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
579              (eptr[3] & 0x3f); \
580          eptr += 4; \
581          } \
582        else \
583        { \        { \
584        gcss -= 6; \        c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
585        c |= (eptr[gcii] & 0x3f) << gcss; \            ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
586              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
587          eptr += 5; \
588        } \        } \
589      }      }
590    
# Line 355  know we are in UTF-8 mode. */ Line 593  know we are in UTF-8 mode. */
593    
594  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
595    c = *eptr++; \    c = *eptr++; \
596    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; \  
       } \  
     }  
597    
598  /* 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.
599    This is called when we don't know if we are in UTF-8 mode. */
600    
601  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
602    c = *eptr++; \    c = *eptr++; \
603    if (utf8 && c >= 0xc0) \    if (utf8 && c >= 0xc0) GETUTF8INC(c, eptr);
604    
605    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
606    advancing the pointer, incrementing the length. */
607    
608    #define GETUTF8LEN(c, eptr, len) \
609      { \      { \
610      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
611      int gcss = 6*gcaa; \        { \
612      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
613      while (gcaa-- > 0) \        len++; \
614          } \
615        else if ((c & 0x10)  == 0) \
616          { \
617          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
618          len += 2; \
619          } \
620        else if ((c & 0x08)  == 0) \
621          {\
622          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
623              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
624          len += 3; \
625          } \
626        else if ((c & 0x04)  == 0) \
627        { \        { \
628        gcss -= 6; \        c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
629        c |= (*eptr++ & 0x3f) << gcss; \            ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
630              (eptr[4] & 0x3f); \
631          len += 4; \
632          } \
633        else \
634          {\
635          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
636              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
637              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
638          len += 5; \
639        } \        } \
640      }      }
641    
# Line 388  if there are extra bytes. This is called Line 644  if there are extra bytes. This is called
644    
645  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
646    c = *eptr; \    c = *eptr; \
647    if (c >= 0xc0) \    if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
648      { \  
649      int gcii; \  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
650      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  pointer, incrementing length if there are extra bytes. This is called when we
651      int gcss = 6*gcaa; \  do not know if we are in UTF-8 mode. */
652      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
653      for (gcii = 1; gcii <= gcaa; gcii++) \  #define GETCHARLENTEST(c, eptr, len) \
654        { \    c = *eptr; \
655        gcss -= 6; \    if (utf8 && c >= 0xc0) GETUTF8LEN(c, eptr, len);
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
656    
657  /* 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
658  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
659    because almost all calls are already within a block of UTF-8 only code. */
660    
661  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
662    
663  #endif  #endif  /* SUPPORT_UTF8 */
664    
665    
666  /* 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 418  Standard C system should have one. */ Line 671  Standard C system should have one. */
671  #endif  #endif
672    
673    
674  /* These are the public options that can change during matching. */  /* Private flags containing information about the compiled regex. They used to
675    live at the top end of the options word, but that got almost full, so now they
676    are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
677    the restrictions on partial matching have been lifted. It remains for backwards
678    compatibility. */
679    
680    #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
681    #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
682    #define PCRE_REQCHSET      0x0004  /* req_byte is set */
683    #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
684    #define PCRE_JCHANGED      0x0010  /* j option used in regex */
685    #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
686    #define PCRE_HASTHEN       0x0040  /* pattern contains (*THEN) */
687    
688  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  /* Flags for the "extra" block produced by pcre_study(). */
689    
690  /* Private options flags start at the most significant end of the four bytes.  #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
691  The public options defined in pcre.h start at the least significant end. Make  #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
 sure they don't overlap! The bits are getting a bit scarce now -- when we run  
 out, there is a dummy word in the structure that could be used for the private  
 bits. */  
   
 #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  
 #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  
 #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  
 #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  
 #define PCRE_JCHANGED      0x08000000  /* j option changes within regex */  
   
 /* Options for the "extra" block produced by pcre_study(). */  
   
 #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  
692    
693  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
694  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
695    
696  #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| \
697                               PCRE_NEWLINE_ANYCRLF)
698    
699  #define PUBLIC_OPTIONS \  #define PUBLIC_COMPILE_OPTIONS \
700    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
701     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
702     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| \
703     PCRE_DUPNAMES|PCRE_NEWLINE_BITS)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
704       PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
705    
706  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
707    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
708     PCRE_PARTIAL|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
709       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
710    
711  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
712    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
713     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
714       PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
715       PCRE_NO_START_OPTIMIZE)
716    
717  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS \
718       PCRE_STUDY_JIT_COMPILE
719    
720  /* Magic number to provide a small check against being handed junk. Also used  /* Magic number to provide a small check against being handed junk. Also used
721  to detect whether a pattern was compiled on a host of different endianness. */  to detect whether a pattern was compiled on a host of different endianness. */
# Line 480  variable-length repeat, or a anything ot Line 738  variable-length repeat, or a anything ot
738  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
739  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
740    
741  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
742    environments where these macros are defined elsewhere. Unfortunately, there
743    is no way to do the same for the typedef. */
744    
745  typedef int BOOL;  typedef int BOOL;
746    
747    #ifndef FALSE
748  #define FALSE   0  #define FALSE   0
749  #define TRUE    1  #define TRUE    1
750    #endif
751    
752    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
753    character constants like '*' because the compiler would emit their EBCDIC code,
754    which is different from their ASCII/UTF-8 code. Instead we define macros for
755    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
756    is enabled. When UTF-8 support is not enabled, the definitions use character
757    literals. Both character and string versions of each character are needed, and
758    there are some longer strings as well.
759    
760    This means that, on EBCDIC platforms, the PCRE library can handle either
761    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
762    would need different lookups depending on whether PCRE_UTF8 was set or not.
763    This would make it impossible to use characters in switch/case statements,
764    which would reduce performance. For a theoretical use (which nobody has asked
765    for) in a minority area (EBCDIC platforms), this is not sensible. Any
766    application that did need both could compile two versions of the library, using
767    macros to give the functions distinct names. */
768    
769    #ifndef SUPPORT_UTF8
770    
771    /* UTF-8 support is not enabled; use the platform-dependent character literals
772    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
773    
774    #define CHAR_HT                     '\t'
775    #define CHAR_VT                     '\v'
776    #define CHAR_FF                     '\f'
777    #define CHAR_CR                     '\r'
778    #define CHAR_NL                     '\n'
779    #define CHAR_BS                     '\b'
780    #define CHAR_BEL                    '\a'
781    #ifdef EBCDIC
782    #define CHAR_ESC                    '\047'
783    #define CHAR_DEL                    '\007'
784    #else
785    #define CHAR_ESC                    '\033'
786    #define CHAR_DEL                    '\177'
787    #endif
788    
789    #define CHAR_SPACE                  ' '
790    #define CHAR_EXCLAMATION_MARK       '!'
791    #define CHAR_QUOTATION_MARK         '"'
792    #define CHAR_NUMBER_SIGN            '#'
793    #define CHAR_DOLLAR_SIGN            '$'
794    #define CHAR_PERCENT_SIGN           '%'
795    #define CHAR_AMPERSAND              '&'
796    #define CHAR_APOSTROPHE             '\''
797    #define CHAR_LEFT_PARENTHESIS       '('
798    #define CHAR_RIGHT_PARENTHESIS      ')'
799    #define CHAR_ASTERISK               '*'
800    #define CHAR_PLUS                   '+'
801    #define CHAR_COMMA                  ','
802    #define CHAR_MINUS                  '-'
803    #define CHAR_DOT                    '.'
804    #define CHAR_SLASH                  '/'
805    #define CHAR_0                      '0'
806    #define CHAR_1                      '1'
807    #define CHAR_2                      '2'
808    #define CHAR_3                      '3'
809    #define CHAR_4                      '4'
810    #define CHAR_5                      '5'
811    #define CHAR_6                      '6'
812    #define CHAR_7                      '7'
813    #define CHAR_8                      '8'
814    #define CHAR_9                      '9'
815    #define CHAR_COLON                  ':'
816    #define CHAR_SEMICOLON              ';'
817    #define CHAR_LESS_THAN_SIGN         '<'
818    #define CHAR_EQUALS_SIGN            '='
819    #define CHAR_GREATER_THAN_SIGN      '>'
820    #define CHAR_QUESTION_MARK          '?'
821    #define CHAR_COMMERCIAL_AT          '@'
822    #define CHAR_A                      'A'
823    #define CHAR_B                      'B'
824    #define CHAR_C                      'C'
825    #define CHAR_D                      'D'
826    #define CHAR_E                      'E'
827    #define CHAR_F                      'F'
828    #define CHAR_G                      'G'
829    #define CHAR_H                      'H'
830    #define CHAR_I                      'I'
831    #define CHAR_J                      'J'
832    #define CHAR_K                      'K'
833    #define CHAR_L                      'L'
834    #define CHAR_M                      'M'
835    #define CHAR_N                      'N'
836    #define CHAR_O                      'O'
837    #define CHAR_P                      'P'
838    #define CHAR_Q                      'Q'
839    #define CHAR_R                      'R'
840    #define CHAR_S                      'S'
841    #define CHAR_T                      'T'
842    #define CHAR_U                      'U'
843    #define CHAR_V                      'V'
844    #define CHAR_W                      'W'
845    #define CHAR_X                      'X'
846    #define CHAR_Y                      'Y'
847    #define CHAR_Z                      'Z'
848    #define CHAR_LEFT_SQUARE_BRACKET    '['
849    #define CHAR_BACKSLASH              '\\'
850    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
851    #define CHAR_CIRCUMFLEX_ACCENT      '^'
852    #define CHAR_UNDERSCORE             '_'
853    #define CHAR_GRAVE_ACCENT           '`'
854    #define CHAR_a                      'a'
855    #define CHAR_b                      'b'
856    #define CHAR_c                      'c'
857    #define CHAR_d                      'd'
858    #define CHAR_e                      'e'
859    #define CHAR_f                      'f'
860    #define CHAR_g                      'g'
861    #define CHAR_h                      'h'
862    #define CHAR_i                      'i'
863    #define CHAR_j                      'j'
864    #define CHAR_k                      'k'
865    #define CHAR_l                      'l'
866    #define CHAR_m                      'm'
867    #define CHAR_n                      'n'
868    #define CHAR_o                      'o'
869    #define CHAR_p                      'p'
870    #define CHAR_q                      'q'
871    #define CHAR_r                      'r'
872    #define CHAR_s                      's'
873    #define CHAR_t                      't'
874    #define CHAR_u                      'u'
875    #define CHAR_v                      'v'
876    #define CHAR_w                      'w'
877    #define CHAR_x                      'x'
878    #define CHAR_y                      'y'
879    #define CHAR_z                      'z'
880    #define CHAR_LEFT_CURLY_BRACKET     '{'
881    #define CHAR_VERTICAL_LINE          '|'
882    #define CHAR_RIGHT_CURLY_BRACKET    '}'
883    #define CHAR_TILDE                  '~'
884    
885    #define STR_HT                      "\t"
886    #define STR_VT                      "\v"
887    #define STR_FF                      "\f"
888    #define STR_CR                      "\r"
889    #define STR_NL                      "\n"
890    #define STR_BS                      "\b"
891    #define STR_BEL                     "\a"
892    #ifdef EBCDIC
893    #define STR_ESC                     "\047"
894    #define STR_DEL                     "\007"
895    #else
896    #define STR_ESC                     "\033"
897    #define STR_DEL                     "\177"
898    #endif
899    
900    #define STR_SPACE                   " "
901    #define STR_EXCLAMATION_MARK        "!"
902    #define STR_QUOTATION_MARK          "\""
903    #define STR_NUMBER_SIGN             "#"
904    #define STR_DOLLAR_SIGN             "$"
905    #define STR_PERCENT_SIGN            "%"
906    #define STR_AMPERSAND               "&"
907    #define STR_APOSTROPHE              "'"
908    #define STR_LEFT_PARENTHESIS        "("
909    #define STR_RIGHT_PARENTHESIS       ")"
910    #define STR_ASTERISK                "*"
911    #define STR_PLUS                    "+"
912    #define STR_COMMA                   ","
913    #define STR_MINUS                   "-"
914    #define STR_DOT                     "."
915    #define STR_SLASH                   "/"
916    #define STR_0                       "0"
917    #define STR_1                       "1"
918    #define STR_2                       "2"
919    #define STR_3                       "3"
920    #define STR_4                       "4"
921    #define STR_5                       "5"
922    #define STR_6                       "6"
923    #define STR_7                       "7"
924    #define STR_8                       "8"
925    #define STR_9                       "9"
926    #define STR_COLON                   ":"
927    #define STR_SEMICOLON               ";"
928    #define STR_LESS_THAN_SIGN          "<"
929    #define STR_EQUALS_SIGN             "="
930    #define STR_GREATER_THAN_SIGN       ">"
931    #define STR_QUESTION_MARK           "?"
932    #define STR_COMMERCIAL_AT           "@"
933    #define STR_A                       "A"
934    #define STR_B                       "B"
935    #define STR_C                       "C"
936    #define STR_D                       "D"
937    #define STR_E                       "E"
938    #define STR_F                       "F"
939    #define STR_G                       "G"
940    #define STR_H                       "H"
941    #define STR_I                       "I"
942    #define STR_J                       "J"
943    #define STR_K                       "K"
944    #define STR_L                       "L"
945    #define STR_M                       "M"
946    #define STR_N                       "N"
947    #define STR_O                       "O"
948    #define STR_P                       "P"
949    #define STR_Q                       "Q"
950    #define STR_R                       "R"
951    #define STR_S                       "S"
952    #define STR_T                       "T"
953    #define STR_U                       "U"
954    #define STR_V                       "V"
955    #define STR_W                       "W"
956    #define STR_X                       "X"
957    #define STR_Y                       "Y"
958    #define STR_Z                       "Z"
959    #define STR_LEFT_SQUARE_BRACKET     "["
960    #define STR_BACKSLASH               "\\"
961    #define STR_RIGHT_SQUARE_BRACKET    "]"
962    #define STR_CIRCUMFLEX_ACCENT       "^"
963    #define STR_UNDERSCORE              "_"
964    #define STR_GRAVE_ACCENT            "`"
965    #define STR_a                       "a"
966    #define STR_b                       "b"
967    #define STR_c                       "c"
968    #define STR_d                       "d"
969    #define STR_e                       "e"
970    #define STR_f                       "f"
971    #define STR_g                       "g"
972    #define STR_h                       "h"
973    #define STR_i                       "i"
974    #define STR_j                       "j"
975    #define STR_k                       "k"
976    #define STR_l                       "l"
977    #define STR_m                       "m"
978    #define STR_n                       "n"
979    #define STR_o                       "o"
980    #define STR_p                       "p"
981    #define STR_q                       "q"
982    #define STR_r                       "r"
983    #define STR_s                       "s"
984    #define STR_t                       "t"
985    #define STR_u                       "u"
986    #define STR_v                       "v"
987    #define STR_w                       "w"
988    #define STR_x                       "x"
989    #define STR_y                       "y"
990    #define STR_z                       "z"
991    #define STR_LEFT_CURLY_BRACKET      "{"
992    #define STR_VERTICAL_LINE           "|"
993    #define STR_RIGHT_CURLY_BRACKET     "}"
994    #define STR_TILDE                   "~"
995    
996    #define STRING_ACCEPT0              "ACCEPT\0"
997    #define STRING_COMMIT0              "COMMIT\0"
998    #define STRING_F0                   "F\0"
999    #define STRING_FAIL0                "FAIL\0"
1000    #define STRING_MARK0                "MARK\0"
1001    #define STRING_PRUNE0               "PRUNE\0"
1002    #define STRING_SKIP0                "SKIP\0"
1003    #define STRING_THEN                 "THEN"
1004    
1005    #define STRING_alpha0               "alpha\0"
1006    #define STRING_lower0               "lower\0"
1007    #define STRING_upper0               "upper\0"
1008    #define STRING_alnum0               "alnum\0"
1009    #define STRING_ascii0               "ascii\0"
1010    #define STRING_blank0               "blank\0"
1011    #define STRING_cntrl0               "cntrl\0"
1012    #define STRING_digit0               "digit\0"
1013    #define STRING_graph0               "graph\0"
1014    #define STRING_print0               "print\0"
1015    #define STRING_punct0               "punct\0"
1016    #define STRING_space0               "space\0"
1017    #define STRING_word0                "word\0"
1018    #define STRING_xdigit               "xdigit"
1019    
1020    #define STRING_DEFINE               "DEFINE"
1021    
1022    #define STRING_CR_RIGHTPAR             "CR)"
1023    #define STRING_LF_RIGHTPAR             "LF)"
1024    #define STRING_CRLF_RIGHTPAR           "CRLF)"
1025    #define STRING_ANY_RIGHTPAR            "ANY)"
1026    #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
1027    #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
1028    #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
1029    #define STRING_UTF8_RIGHTPAR           "UTF8)"
1030    #define STRING_UCP_RIGHTPAR            "UCP)"
1031    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
1032    
1033    #else  /* SUPPORT_UTF8 */
1034    
1035    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
1036    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
1037    only. */
1038    
1039    #define CHAR_HT                     '\011'
1040    #define CHAR_VT                     '\013'
1041    #define CHAR_FF                     '\014'
1042    #define CHAR_CR                     '\015'
1043    #define CHAR_NL                     '\012'
1044    #define CHAR_BS                     '\010'
1045    #define CHAR_BEL                    '\007'
1046    #define CHAR_ESC                    '\033'
1047    #define CHAR_DEL                    '\177'
1048    
1049    #define CHAR_SPACE                  '\040'
1050    #define CHAR_EXCLAMATION_MARK       '\041'
1051    #define CHAR_QUOTATION_MARK         '\042'
1052    #define CHAR_NUMBER_SIGN            '\043'
1053    #define CHAR_DOLLAR_SIGN            '\044'
1054    #define CHAR_PERCENT_SIGN           '\045'
1055    #define CHAR_AMPERSAND              '\046'
1056    #define CHAR_APOSTROPHE             '\047'
1057    #define CHAR_LEFT_PARENTHESIS       '\050'
1058    #define CHAR_RIGHT_PARENTHESIS      '\051'
1059    #define CHAR_ASTERISK               '\052'
1060    #define CHAR_PLUS                   '\053'
1061    #define CHAR_COMMA                  '\054'
1062    #define CHAR_MINUS                  '\055'
1063    #define CHAR_DOT                    '\056'
1064    #define CHAR_SLASH                  '\057'
1065    #define CHAR_0                      '\060'
1066    #define CHAR_1                      '\061'
1067    #define CHAR_2                      '\062'
1068    #define CHAR_3                      '\063'
1069    #define CHAR_4                      '\064'
1070    #define CHAR_5                      '\065'
1071    #define CHAR_6                      '\066'
1072    #define CHAR_7                      '\067'
1073    #define CHAR_8                      '\070'
1074    #define CHAR_9                      '\071'
1075    #define CHAR_COLON                  '\072'
1076    #define CHAR_SEMICOLON              '\073'
1077    #define CHAR_LESS_THAN_SIGN         '\074'
1078    #define CHAR_EQUALS_SIGN            '\075'
1079    #define CHAR_GREATER_THAN_SIGN      '\076'
1080    #define CHAR_QUESTION_MARK          '\077'
1081    #define CHAR_COMMERCIAL_AT          '\100'
1082    #define CHAR_A                      '\101'
1083    #define CHAR_B                      '\102'
1084    #define CHAR_C                      '\103'
1085    #define CHAR_D                      '\104'
1086    #define CHAR_E                      '\105'
1087    #define CHAR_F                      '\106'
1088    #define CHAR_G                      '\107'
1089    #define CHAR_H                      '\110'
1090    #define CHAR_I                      '\111'
1091    #define CHAR_J                      '\112'
1092    #define CHAR_K                      '\113'
1093    #define CHAR_L                      '\114'
1094    #define CHAR_M                      '\115'
1095    #define CHAR_N                      '\116'
1096    #define CHAR_O                      '\117'
1097    #define CHAR_P                      '\120'
1098    #define CHAR_Q                      '\121'
1099    #define CHAR_R                      '\122'
1100    #define CHAR_S                      '\123'
1101    #define CHAR_T                      '\124'
1102    #define CHAR_U                      '\125'
1103    #define CHAR_V                      '\126'
1104    #define CHAR_W                      '\127'
1105    #define CHAR_X                      '\130'
1106    #define CHAR_Y                      '\131'
1107    #define CHAR_Z                      '\132'
1108    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
1109    #define CHAR_BACKSLASH              '\134'
1110    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
1111    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
1112    #define CHAR_UNDERSCORE             '\137'
1113    #define CHAR_GRAVE_ACCENT           '\140'
1114    #define CHAR_a                      '\141'
1115    #define CHAR_b                      '\142'
1116    #define CHAR_c                      '\143'
1117    #define CHAR_d                      '\144'
1118    #define CHAR_e                      '\145'
1119    #define CHAR_f                      '\146'
1120    #define CHAR_g                      '\147'
1121    #define CHAR_h                      '\150'
1122    #define CHAR_i                      '\151'
1123    #define CHAR_j                      '\152'
1124    #define CHAR_k                      '\153'
1125    #define CHAR_l                      '\154'
1126    #define CHAR_m                      '\155'
1127    #define CHAR_n                      '\156'
1128    #define CHAR_o                      '\157'
1129    #define CHAR_p                      '\160'
1130    #define CHAR_q                      '\161'
1131    #define CHAR_r                      '\162'
1132    #define CHAR_s                      '\163'
1133    #define CHAR_t                      '\164'
1134    #define CHAR_u                      '\165'
1135    #define CHAR_v                      '\166'
1136    #define CHAR_w                      '\167'
1137    #define CHAR_x                      '\170'
1138    #define CHAR_y                      '\171'
1139    #define CHAR_z                      '\172'
1140    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1141    #define CHAR_VERTICAL_LINE          '\174'
1142    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1143    #define CHAR_TILDE                  '\176'
1144    
1145    #define STR_HT                      "\011"
1146    #define STR_VT                      "\013"
1147    #define STR_FF                      "\014"
1148    #define STR_CR                      "\015"
1149    #define STR_NL                      "\012"
1150    #define STR_BS                      "\010"
1151    #define STR_BEL                     "\007"
1152    #define STR_ESC                     "\033"
1153    #define STR_DEL                     "\177"
1154    
1155    #define STR_SPACE                   "\040"
1156    #define STR_EXCLAMATION_MARK        "\041"
1157    #define STR_QUOTATION_MARK          "\042"
1158    #define STR_NUMBER_SIGN             "\043"
1159    #define STR_DOLLAR_SIGN             "\044"
1160    #define STR_PERCENT_SIGN            "\045"
1161    #define STR_AMPERSAND               "\046"
1162    #define STR_APOSTROPHE              "\047"
1163    #define STR_LEFT_PARENTHESIS        "\050"
1164    #define STR_RIGHT_PARENTHESIS       "\051"
1165    #define STR_ASTERISK                "\052"
1166    #define STR_PLUS                    "\053"
1167    #define STR_COMMA                   "\054"
1168    #define STR_MINUS                   "\055"
1169    #define STR_DOT                     "\056"
1170    #define STR_SLASH                   "\057"
1171    #define STR_0                       "\060"
1172    #define STR_1                       "\061"
1173    #define STR_2                       "\062"
1174    #define STR_3                       "\063"
1175    #define STR_4                       "\064"
1176    #define STR_5                       "\065"
1177    #define STR_6                       "\066"
1178    #define STR_7                       "\067"
1179    #define STR_8                       "\070"
1180    #define STR_9                       "\071"
1181    #define STR_COLON                   "\072"
1182    #define STR_SEMICOLON               "\073"
1183    #define STR_LESS_THAN_SIGN          "\074"
1184    #define STR_EQUALS_SIGN             "\075"
1185    #define STR_GREATER_THAN_SIGN       "\076"
1186    #define STR_QUESTION_MARK           "\077"
1187    #define STR_COMMERCIAL_AT           "\100"
1188    #define STR_A                       "\101"
1189    #define STR_B                       "\102"
1190    #define STR_C                       "\103"
1191    #define STR_D                       "\104"
1192    #define STR_E                       "\105"
1193    #define STR_F                       "\106"
1194    #define STR_G                       "\107"
1195    #define STR_H                       "\110"
1196    #define STR_I                       "\111"
1197    #define STR_J                       "\112"
1198    #define STR_K                       "\113"
1199    #define STR_L                       "\114"
1200    #define STR_M                       "\115"
1201    #define STR_N                       "\116"
1202    #define STR_O                       "\117"
1203    #define STR_P                       "\120"
1204    #define STR_Q                       "\121"
1205    #define STR_R                       "\122"
1206    #define STR_S                       "\123"
1207    #define STR_T                       "\124"
1208    #define STR_U                       "\125"
1209    #define STR_V                       "\126"
1210    #define STR_W                       "\127"
1211    #define STR_X                       "\130"
1212    #define STR_Y                       "\131"
1213    #define STR_Z                       "\132"
1214    #define STR_LEFT_SQUARE_BRACKET     "\133"
1215    #define STR_BACKSLASH               "\134"
1216    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1217    #define STR_CIRCUMFLEX_ACCENT       "\136"
1218    #define STR_UNDERSCORE              "\137"
1219    #define STR_GRAVE_ACCENT            "\140"
1220    #define STR_a                       "\141"
1221    #define STR_b                       "\142"
1222    #define STR_c                       "\143"
1223    #define STR_d                       "\144"
1224    #define STR_e                       "\145"
1225    #define STR_f                       "\146"
1226    #define STR_g                       "\147"
1227    #define STR_h                       "\150"
1228    #define STR_i                       "\151"
1229    #define STR_j                       "\152"
1230    #define STR_k                       "\153"
1231    #define STR_l                       "\154"
1232    #define STR_m                       "\155"
1233    #define STR_n                       "\156"
1234    #define STR_o                       "\157"
1235    #define STR_p                       "\160"
1236    #define STR_q                       "\161"
1237    #define STR_r                       "\162"
1238    #define STR_s                       "\163"
1239    #define STR_t                       "\164"
1240    #define STR_u                       "\165"
1241    #define STR_v                       "\166"
1242    #define STR_w                       "\167"
1243    #define STR_x                       "\170"
1244    #define STR_y                       "\171"
1245    #define STR_z                       "\172"
1246    #define STR_LEFT_CURLY_BRACKET      "\173"
1247    #define STR_VERTICAL_LINE           "\174"
1248    #define STR_RIGHT_CURLY_BRACKET     "\175"
1249    #define STR_TILDE                   "\176"
1250    
1251    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1252    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1253    #define STRING_F0                   STR_F "\0"
1254    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1255    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1256    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1257    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1258    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1259    
1260    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1261    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1262    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1263    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1264    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1265    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1266    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1267    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1268    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1269    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1270    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1271    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1272    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1273    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1274    
1275    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1276    
1277    #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1278    #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1279    #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1280    #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1281    #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1282    #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
1283    #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
1284    #define STRING_UTF8_RIGHTPAR           STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1285    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1286    #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
1287    
1288    #endif  /* SUPPORT_UTF8 */
1289    
1290  /* Escape items that are just an encoding of a particular data value. */  /* Escape items that are just an encoding of a particular data value. */
1291    
1292  #ifndef ESC_e  #ifndef ESC_e
1293  #define ESC_e 27  #define ESC_e CHAR_ESC
1294  #endif  #endif
1295    
1296  #ifndef ESC_f  #ifndef ESC_f
1297  #define ESC_f '\f'  #define ESC_f CHAR_FF
1298  #endif  #endif
1299    
1300  #ifndef ESC_n  #ifndef ESC_n
1301  #define ESC_n '\n'  #define ESC_n CHAR_NL
1302  #endif  #endif
1303    
1304  #ifndef ESC_r  #ifndef ESC_r
1305  #define ESC_r '\r'  #define ESC_r CHAR_CR
1306  #endif  #endif
1307    
1308  /* 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
1309  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1310    
1311  #ifndef ESC_tee  #ifndef ESC_tee
1312  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1313  #endif  #endif
1314    
1315  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
1316    
1317  #define PT_ANY        0    /* Any property - matches all chars */  #define PT_ANY        0    /* Any property - matches all chars */
1318  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1319  #define PT_GC         2    /* General characteristic (e.g. L) */  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1320  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1321  #define PT_SC         4    /* Script (e.g. Han) */  #define PT_SC         4    /* Script (e.g. Han) */
1322    #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1323    #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1324    #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1325    #define PT_WORD       8    /* Word - L plus N plus underscore */
1326    
1327  /* 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
1328  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 535  contain UTF-8 characters with values gre Line 1339  contain UTF-8 characters with values gre
1339  /* 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
1340  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
1341  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
1342  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
1343  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
1344  ESC_REF as subsequent values are used for backreferences (\1, \2, \3, etc).  used for [^] in JavaScript compatibility mode, and for \C in non-utf8 mode. In
1345  There are two tests in the code for an escape greater than ESC_b and less than  non-DOTALL mode, "." behaves like \N.
1346  ESC_Z to detect the types that may be repeated. These are the types that  
1347  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.
1348  character, that code will have to change. */  when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1349    They must be contiguous, and remain in order so that the replacements can be
1350  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.
1351         ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_R, ESC_X, ESC_Z, ESC_z,  
1352         ESC_E, ESC_Q, ESC_k, ESC_REF };  The final escape must be ESC_REF as subsequent values are used for
1353    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1354    greater than ESC_b and less than ESC_Z to detect the types that may be
1355    repeated. These are the types that consume characters. If any new escapes are
1356    put in between that don't consume a character, that code will have to change.
1357    */
1358    
1359    enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
1360           ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1361           ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1362           ESC_E, ESC_Q, ESC_g, ESC_k,
1363           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1364           ESC_REF };
1365    
1366  /* 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  
1367  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.
1368    
 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.  
   
1369  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1370  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
1371  "coptable" in pcre_dfa_exec.c that must be updated. */  "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1372    
1373  enum {  enum {
1374    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 567  enum { Line 1377  enum {
1377    
1378    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
1379    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
1380    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
1381    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
1382    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
1383    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
1384    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
1385    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
1386    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
1387    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
1388    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
1389    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */  
1390    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ANY,            /* 12 Match any character except newline */
1391    OP_PROP,           /* 14 \p (Unicode property) */    OP_ALLANY,         /* 13 Match any character */
1392    OP_ANYNL,          /* 15 \R (any newline sequence) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1393    OP_EXTUNI,         /* 16 \X (extended Unicode sequence */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1394    OP_EODN,           /* 17 End of data or \n at end of data: \Z. */    OP_PROP,           /* 16 \p (Unicode property) */
1395    OP_EOD,            /* 18 End of data: \z */    OP_ANYNL,          /* 17 \R (any newline sequence) */
1396      OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1397    OP_OPT,            /* 19 Set runtime options */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1398    OP_CIRC,           /* 20 Start of line - varies with multiline switch */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1399    OP_DOLL,           /* 21 End of line - varies with multiline switch */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1400    OP_CHAR,           /* 22 Match one character, casefully */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1401    OP_CHARNC,         /* 23 Match one character, caselessly */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1402    OP_NOT,            /* 24 Match one character, not the following one */    OP_EOD,            /* 24 End of data: \z */
1403    
1404    OP_STAR,           /* 25 The maximizing and minimizing versions of */    OP_CIRC,           /* 25 Start of line - not multiline */
1405    OP_MINSTAR,        /* 26 these six opcodes must come in pairs, with */    OP_CIRCM,          /* 26 Start of line - multiline */
1406    OP_PLUS,           /* 27 the minimizing one second. */    OP_DOLL,           /* 27 End of line - not multiline */
1407    OP_MINPLUS,        /* 28 This first set applies to single characters.*/    OP_DOLLM,          /* 28 End of line - multiline */
1408    OP_QUERY,          /* 29 */    OP_CHAR,           /* 29 Match one character, casefully */
1409    OP_MINQUERY,       /* 30 */    OP_CHARI,          /* 30 Match one character, caselessly */
1410      OP_NOT,            /* 31 Match one character, not the given one, casefully */
1411    OP_UPTO,           /* 31 From 0 to n matches */    OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1412    OP_MINUPTO,        /* 32 */  
1413    OP_EXACT,          /* 33 Exactly n matches */    /* The following sets of 13 opcodes must always be kept in step because
1414      the offset from the first one is used to generate the others. */
1415    OP_POSSTAR,        /* 34 Possessified star */  
1416    OP_POSPLUS,        /* 35 Possessified plus */    /**** Single characters, caseful, must precede the caseless ones ****/
1417    OP_POSQUERY,       /* 36 Posesssified query */  
1418    OP_POSUPTO,        /* 37 Possessified upto */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1419      OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1420    OP_NOTSTAR,        /* 38 The maximizing and minimizing versions of */    OP_PLUS,           /* 35 the minimizing one second. */
1421    OP_NOTMINSTAR,     /* 39 these six opcodes must come in pairs, with */    OP_MINPLUS,        /* 36 */
1422    OP_NOTPLUS,        /* 40 the minimizing one second. They must be in */    OP_QUERY,          /* 37 */
1423    OP_NOTMINPLUS,     /* 41 exactly the same order as those above. */    OP_MINQUERY,       /* 38 */
1424    OP_NOTQUERY,       /* 42 This set applies to "not" single characters. */  
1425    OP_NOTMINQUERY,    /* 43 */    OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1426      OP_MINUPTO,        /* 40 */
1427    OP_NOTUPTO,        /* 44 From 0 to n matches */    OP_EXACT,          /* 41 Exactly n matches */
1428    OP_NOTMINUPTO,     /* 45 */  
1429    OP_NOTEXACT,       /* 46 Exactly n matches */    OP_POSSTAR,        /* 42 Possessified star, caseful */
1430      OP_POSPLUS,        /* 43 Possessified plus, caseful */
1431    OP_NOTPOSSTAR,     /* 47 Possessified versions */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
1432    OP_NOTPOSPLUS,     /* 48 */    OP_POSUPTO,        /* 45 Possessified upto, caseful */
1433    OP_NOTPOSQUERY,    /* 49 */  
1434    OP_NOTPOSUPTO,     /* 50 */    /**** Single characters, caseless, must follow the caseful ones */
1435    
1436    OP_TYPESTAR,       /* 51 The maximizing and minimizing versions of */    OP_STARI,          /* 46 */
1437    OP_TYPEMINSTAR,    /* 52 these six opcodes must come in pairs, with */    OP_MINSTARI,       /* 47 */
1438    OP_TYPEPLUS,       /* 53 the minimizing one second. These codes must */    OP_PLUSI,          /* 48 */
1439    OP_TYPEMINPLUS,    /* 54 be in exactly the same order as those above. */    OP_MINPLUSI,       /* 49 */
1440    OP_TYPEQUERY,      /* 55 This set applies to character types such as \d */    OP_QUERYI,         /* 50 */
1441    OP_TYPEMINQUERY,   /* 56 */    OP_MINQUERYI,      /* 51 */
1442    
1443    OP_TYPEUPTO,       /* 57 From 0 to n matches */    OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1444    OP_TYPEMINUPTO,    /* 58 */    OP_MINUPTOI,       /* 53 */
1445    OP_TYPEEXACT,      /* 59 Exactly n matches */    OP_EXACTI,         /* 54 */
1446    
1447    OP_TYPEPOSSTAR,    /* 60 Possessified versions */    OP_POSSTARI,       /* 55 Possessified star, caseless */
1448    OP_TYPEPOSPLUS,    /* 61 */    OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1449    OP_TYPEPOSQUERY,   /* 62 */    OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1450    OP_TYPEPOSUPTO,    /* 63 */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1451    
1452    OP_CRSTAR,         /* 64 The maximizing and minimizing versions of */    /**** The negated ones must follow the non-negated ones, and match them ****/
1453    OP_CRMINSTAR,      /* 65 all these opcodes must come in pairs, with */    /**** Negated single character, caseful; must precede the caseless ones ****/
1454    OP_CRPLUS,         /* 66 the minimizing one second. These codes must */  
1455    OP_CRMINPLUS,      /* 67 be in exactly the same order as those above. */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1456    OP_CRQUERY,        /* 68 These are for character classes and back refs */    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1457    OP_CRMINQUERY,     /* 69 */    OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1458    OP_CRRANGE,        /* 70 These are different to the three sets above. */    OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1459    OP_CRMINRANGE,     /* 71 */    OP_NOTQUERY,       /* 63 */
1460      OP_NOTMINQUERY,    /* 64 */
1461    OP_CLASS,          /* 72 Match a character class, chars < 256 only */  
1462    OP_NCLASS,         /* 73 Same, but the bitmap was created from a negative    OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1463                             class - the difference is relevant only when a UTF-8    OP_NOTMINUPTO,     /* 66 */
1464                             character > 255 is encountered. */    OP_NOTEXACT,       /* 67 Exactly n matches */
1465    
1466    OP_XCLASS,         /* 74 Extended class for handling UTF-8 chars within the    OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1467                             class. This does both positive and negative. */    OP_NOTPOSPLUS,     /* 69 */
1468      OP_NOTPOSQUERY,    /* 70 */
1469    OP_REF,            /* 75 Match a back reference */    OP_NOTPOSUPTO,     /* 71 */
1470    OP_RECURSE,        /* 76 Match a numbered subpattern (possibly recursive) */  
1471    OP_CALLOUT,        /* 77 Call out to external function if provided */    /**** Negated single character, caseless; must follow the caseful ones ****/
1472    
1473    OP_ALT,            /* 78 Start of alternation */    OP_NOTSTARI,       /* 72 */
1474    OP_KET,            /* 79 End of group that doesn't have an unbounded repeat */    OP_NOTMINSTARI,    /* 73 */
1475    OP_KETRMAX,        /* 80 These two must remain together and in this */    OP_NOTPLUSI,       /* 74 */
1476    OP_KETRMIN,        /* 81 order. They are for groups the repeat for ever. */    OP_NOTMINPLUSI,    /* 75 */
1477      OP_NOTQUERYI,      /* 76 */
1478    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/    OP_NOTMINQUERYI,   /* 77 */
1479    
1480    OP_ASSERT,         /* 82 Positive lookahead */    OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1481    OP_ASSERT_NOT,     /* 83 Negative lookahead */    OP_NOTMINUPTOI,    /* 79 */
1482    OP_ASSERTBACK,     /* 84 Positive lookbehind */    OP_NOTEXACTI,      /* 80 Exactly n matches */
1483    OP_ASSERTBACK_NOT, /* 85 Negative lookbehind */  
1484    OP_REVERSE,        /* 86 Move pointer back - used in lookbehind assertions */    OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1485      OP_NOTPOSPLUSI,    /* 82 */
1486    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,    OP_NOTPOSQUERYI,   /* 83 */
1487    as there's a test for >= ONCE for a subpattern that isn't an assertion. */    OP_NOTPOSUPTOI,    /* 84 */
1488    
1489    OP_ONCE,           /* 87 Atomic group */    /**** Character types ****/
1490    OP_BRA,            /* 88 Start of non-capturing bracket */  
1491    OP_CBRA,           /* 89 Start of capturing bracket */    OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1492    OP_COND,           /* 90 Conditional group */    OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1493      OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1494      OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
1495      OP_TYPEQUERY,      /* 89 */
1496      OP_TYPEMINQUERY,   /* 90 */
1497    
1498      OP_TYPEUPTO,       /* 91 From 0 to n matches */
1499      OP_TYPEMINUPTO,    /* 92 */
1500      OP_TYPEEXACT,      /* 93 Exactly n matches */
1501    
1502      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1503      OP_TYPEPOSPLUS,    /* 95 */
1504      OP_TYPEPOSQUERY,   /* 96 */
1505      OP_TYPEPOSUPTO,    /* 97 */
1506    
1507      /* These are used for character classes and back references; only the
1508      first six are the same as the sets above. */
1509    
1510      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1511      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1512      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1513      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1514      OP_CRQUERY,        /* 102 */
1515      OP_CRMINQUERY,     /* 103 */
1516    
1517      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1518      OP_CRMINRANGE,     /* 105 */
1519    
1520      /* End of quantifier opcodes */
1521    
1522      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1523      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1524                                  class - the difference is relevant only when a
1525                                  UTF-8 character > 255 is encountered. */
1526      OP_XCLASS,         /* 108 Extended class for handling UTF-8 chars within the
1527                                  class. This does both positive and negative. */
1528      OP_REF,            /* 109 Match a back reference, casefully */
1529      OP_REFI,           /* 110 Match a back reference, caselessly */
1530      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1531      OP_CALLOUT,        /* 112 Call out to external function if provided */
1532    
1533      OP_ALT,            /* 113 Start of alternation */
1534      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1535      OP_KETRMAX,        /* 115 These two must remain together and in this */
1536      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1537      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1538    
1539      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1540      asserts must remain in order. */
1541    
1542      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1543      OP_ASSERT,         /* 119 Positive lookahead */
1544      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1545      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1546      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1547    
1548      /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
1549      after the assertions, with ONCE first, as there's a test for >= ONCE for a
1550      subpattern that isn't an assertion. The POS versions must immediately follow
1551      the non-POS versions in each case. */
1552    
1553      OP_ONCE,           /* 123 Atomic group, contains captures */
1554      OP_ONCE_NC,        /* 124 Atomic group containing no captures */
1555      OP_BRA,            /* 125 Start of non-capturing bracket */
1556      OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
1557      OP_CBRA,           /* 127 Start of capturing bracket */
1558      OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
1559      OP_COND,           /* 129 Conditional group */
1560    
1561    /* These three must follow the previous three, in the same order. There's a    /* These five must follow the previous five, in the same order. There's a
1562    check for >= SBRA to distinguish the two sets. */    check for >= SBRA to distinguish the two sets. */
1563    
1564    OP_SBRA,           /* 91 Start of non-capturing bracket, check empty  */    OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
1565    OP_SCBRA,          /* 92 Start of capturing bracket, check empty */    OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
1566    OP_SCOND,          /* 93 Conditional group, check empty */    OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
1567      OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
1568    OP_CREF,           /* 94 Used to hold a capture number as condition */    OP_SCOND,          /* 134 Conditional group, check empty */
1569    OP_RREF,           /* 95 Used to hold a recursion number as condition */  
1570    OP_DEF,            /* 96 The DEFINE condition */    /* The next two pairs must (respectively) be kept together. */
1571    
1572      OP_CREF,           /* 135 Used to hold a capture number as condition */
1573      OP_NCREF,          /* 136 Same, but generated by a name reference*/
1574      OP_RREF,           /* 137 Used to hold a recursion number as condition */
1575      OP_NRREF,          /* 138 Same, but generated by a name reference*/
1576      OP_DEF,            /* 139 The DEFINE condition */
1577    
1578      OP_BRAZERO,        /* 140 These two must remain together and in this */
1579      OP_BRAMINZERO,     /* 141 order. */
1580      OP_BRAPOSZERO,     /* 142 */
1581    
1582      /* These are backtracking control verbs */
1583    
1584      OP_MARK,           /* 143 always has an argument */
1585      OP_PRUNE,          /* 144 */
1586      OP_PRUNE_ARG,      /* 145 same, but with argument */
1587      OP_SKIP,           /* 146 */
1588      OP_SKIP_ARG,       /* 147 same, but with argument */
1589      OP_THEN,           /* 148 */
1590      OP_THEN_ARG,       /* 149 same, but with argument */
1591      OP_COMMIT,         /* 150 */
1592    
1593      /* These are forced failure and success verbs */
1594    
1595      OP_FAIL,           /* 151 */
1596      OP_ACCEPT,         /* 152 */
1597      OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
1598      OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
1599    
1600      /* This is used to skip a subpattern with a {0} quantifier */
1601    
1602      OP_SKIPZERO,       /* 155 */
1603    
1604      /* This is not an opcode, but is used to check that tables indexed by opcode
1605      are the correct length, in order to catch updating errors - there have been
1606      some in the past. */
1607    
1608    OP_BRAZERO,        /* 97 These two must remain together and in this */    OP_TABLE_LENGTH
   OP_BRAMINZERO      /* 98 order. */  
1609  };  };
1610    
1611    /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1612    definitions that follow must also be updated to match. There are also tables
1613    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1614    
1615    
1616  /* 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
1617  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging, and some of them are only partial names. The macro is referenced
1618    only in pcre_printint.c, which fills out the full names in many cases (and in
1619    some cases doesn't actually use these names at all). */
1620    
1621  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1622    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1623    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1624    "notprop", "prop", "anynl", "extuni",                           \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1625    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
1626    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1627    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1628      "{", "{", "{",                                                  \
1629    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1630    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1631      "{", "{", "{",                                                  \
1632      "*+","++", "?+", "{",                                           \
1633      "*", "*?", "+", "+?", "?", "??",                                \
1634      "{", "{", "{",                                                  \
1635      "*+","++", "?+", "{",                                           \
1636      "*", "*?", "+", "+?", "?", "??",                                \
1637      "{", "{", "{",                                                  \
1638    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1639    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1640    "*+","++", "?+", "{",                                           \    "*+","++", "?+", "{",                                           \
1641    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1642    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1643    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1644    "AssertB", "AssertB not", "Reverse",                            \    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1645    "Once", "Bra 0", "Bra", "Cond", "SBra 0", "SBra", "SCond",      \    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1646    "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero"    "Once", "Once_NC",                                              \
1647      "Bra", "BraPos", "CBra", "CBraPos",                             \
1648      "Cond",                                                         \
1649      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1650      "SCond",                                                        \
1651      "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1652      "Brazero", "Braminzero", "Braposzero",                          \
1653      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1654      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1655      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1656      "Close", "Skip zero"
1657    
1658    
1659  /* 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 731  in UTF-8 mode. The code that uses this t Line 1667  in UTF-8 mode. The code that uses this t
1667    
1668  #define OP_LENGTHS \  #define OP_LENGTHS \
1669    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1670    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                     */ \
1671    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1672    3, 3, 1, 1,                    /* NOTPROP, PROP, EXTUNI, ANYNL           */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1673    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    3, 3,                          /* \P, \p                                 */ \
1674      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1675      1,                             /* \X                                     */ \
1676      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1677    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1678    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1679    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1680    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1681    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1682    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1683    2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto, minupto             ** mode      */ \
1684      2+IMM2_SIZE,                   /* exact                                  */ \
1685      2, 2, 2, 2+IMM2_SIZE,          /* *+, ++, ?+, upto+                      */ \
1686      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1687      2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto I, minupto I                      */ \
1688      2+IMM2_SIZE,                   /* exact I                                */ \
1689      2, 2, 2, 2+IMM2_SIZE,          /* *+I, ++I, ?+I, upto+I                  */ \
1690    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1691    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1692    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto, minupto                      */ \
1693    2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \    2+IMM2_SIZE,                   /* NOT exact                              */ \
1694      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *, +, ?, upto           */ \
1695      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1696      2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto I, minupto I                  */ \
1697      2+IMM2_SIZE,                   /* NOT exact I                            */ \
1698      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *I, +I, ?I, upto I      */ \
1699    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1700    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1701    4, 4, 4,                       /* Type upto, minupto, exact              */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* Type upto, minupto                     */ \
1702    2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \    2+IMM2_SIZE,                   /* Type exact                             */ \
1703      2, 2, 2, 2+IMM2_SIZE,          /* Possessive *+, ++, ?+, upto+           */ \
1704    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
1705    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
1706    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
1707   33,                             /* CLASS                                  */ \   33,                             /* CLASS                                  */ \
1708   33,                             /* NCLASS                                 */ \   33,                             /* NCLASS                                 */ \
1709    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1710    3,                             /* REF                                    */ \    1+IMM2_SIZE,                   /* REF                                    */ \
1711      1+IMM2_SIZE,                   /* REFI                                   */ \
1712    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1713    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1714    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
1715    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
1716    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
1717    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
1718      1+LINK_SIZE,                   /* KetRpos                                */ \
1719      1+LINK_SIZE,                   /* Reverse                                */ \
1720    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
1721    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
1722    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1723    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
   1+LINK_SIZE,                   /* Reverse                                */ \  
1724    1+LINK_SIZE,                   /* ONCE                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1725      1+LINK_SIZE,                   /* ONCE_NC                                */ \
1726    1+LINK_SIZE,                   /* BRA                                    */ \    1+LINK_SIZE,                   /* BRA                                    */ \
1727    3+LINK_SIZE,                   /* CBRA                                   */ \    1+LINK_SIZE,                   /* BRAPOS                                 */ \
1728      1+LINK_SIZE+IMM2_SIZE,         /* CBRA                                   */ \
1729      1+LINK_SIZE+IMM2_SIZE,         /* CBRAPOS                                */ \
1730    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1731    1+LINK_SIZE,                   /* SBRA                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1732    3+LINK_SIZE,                   /* SCBRA                                  */ \    1+LINK_SIZE,                   /* SBRAPOS                                */ \
1733      1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
1734      1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
1735    1+LINK_SIZE,                   /* SCOND                                  */ \    1+LINK_SIZE,                   /* SCOND                                  */ \
1736    3,                             /* CREF                                   */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \
1737    3,                             /* RREF                                   */ \    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \
1738    1,                             /* DEF                                    */ \    1,                             /* DEF                                    */ \
1739    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
1740      3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1741      1, 3,                          /* SKIP, SKIP_ARG                         */ \
1742      1, 3,                          /* THEN, THEN_ARG                         */ \
1743      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
1744      1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
1745    
1746  /* 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"
1747    condition. */
1748    
1749  #define RREF_ANY  0xffff  #define RREF_ANY  0xffff
1750    
1751  /* 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
1752  tracked. */  easily be tracked. When a new number is added, the table called eint in
1753    pcreposix.c must be updated. */
1754    
1755  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1756         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
1757         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
1758         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1759         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1760         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1761           ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
1762           ERR70, ERRCOUNT };
1763    
1764  /* 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
1765  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 809  Because people can now save and re-use c Line 1775  Because people can now save and re-use c
1775  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
1776  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
1777  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1778  NOTE NOTE NOTE:  NOTE NOTE NOTE
1779  */  */
1780    
1781  typedef struct real_pcre {  typedef struct real_pcre {
1782    pcre_uint32 magic_number;    pcre_uint32 magic_number;
1783    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1784    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
1785    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
1786      pcre_uint16 dummy1;             /* For future use */
1787    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
1788    pcre_uint16 top_backref;    pcre_uint16 top_backref;
1789    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 827  typedef struct real_pcre { Line 1793  typedef struct real_pcre {
1793    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
1794    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
1795    
1796    const unsigned char *tables;    /* Pointer to tables or NULL for std */    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
1797    const unsigned char *nullpad;   /* NULL padding */    const pcre_uint8 *nullpad;      /* NULL padding */
1798  } real_pcre;  } real_pcre;
1799    
1800  /* The format of the block used to store data from pcre_study(). The same  /* The format of the block used to store data from pcre_study(). The same
# Line 836  remark (see NOTE above) about extending Line 1802  remark (see NOTE above) about extending
1802    
1803  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1804    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1805    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1806    uschar start_bits[32];    pcre_uint8 start_bits[32];      /* Starting char bits */
1807      pcre_uint32 minlength;          /* Minimum subject length */
1808  } pcre_study_data;  } pcre_study_data;
1809    
1810    /* Structure for building a chain of open capturing subpatterns during
1811    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
1812    encountered. This is also used to identify subpatterns that contain recursive
1813    back references to themselves, so that they can be made atomic. */
1814    
1815    typedef struct open_capitem {
1816      struct open_capitem *next;    /* Chain link */
1817      pcre_uint16 number;           /* Capture number */
1818      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
1819    } open_capitem;
1820    
1821  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1822  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
1823    
1824  typedef struct compile_data {  typedef struct compile_data {
1825    const uschar *lcc;            /* Points to lower casing table */    const pcre_uint8 *lcc;            /* Points to lower casing table */
1826    const uschar *fcc;            /* Points to case-flipping table */    const pcre_uint8 *fcc;            /* Points to case-flipping table */
1827    const uschar *cbits;          /* Points to character type table */    const pcre_uint8 *cbits;          /* Points to character type table */
1828    const uschar *ctypes;         /* Points to table of type maps */    const pcre_uint8 *ctypes;         /* Points to table of type maps */
1829    const uschar *start_workspace;/* The start of working space */    const pcre_uchar *start_workspace;/* The start of working space */
1830    const uschar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_code;     /* The start of the compiled code */
1831    const uschar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_pattern;  /* The start of the pattern */
1832    const uschar *end_pattern;    /* The end of the pattern */    const pcre_uchar *end_pattern;    /* The end of the pattern */
1833    uschar *hwm;                  /* High watermark of workspace */    open_capitem *open_caps;          /* Chain of open capture items */
1834    uschar *name_table;           /* The name/number table */    pcre_uchar *hwm;                  /* High watermark of workspace */
1835    int  names_found;             /* Number of entries so far */    pcre_uchar *name_table;           /* The name/number table */
1836    int  name_entry_size;         /* Size of each entry */    int  names_found;                 /* Number of entries so far */
1837    int  bracount;                /* Count of capturing parens */    int  name_entry_size;             /* Size of each entry */
1838    int  top_backref;             /* Maximum back reference */    int  bracount;                    /* Count of capturing parens as we compile */
1839    unsigned int backref_map;     /* Bitmap of low back refs */    int  final_bracount;              /* Saved value after first pass */
1840    int  external_options;        /* External (initial) options */    int  top_backref;                 /* Maximum back reference */
1841    int  req_varyopt;             /* "After variable item" flag for reqbyte */    unsigned int backref_map;         /* Bitmap of low back refs */
1842    BOOL nopartial;               /* Set TRUE if partial won't work */    int  assert_depth;                /* Depth of nested assertions */
1843    int  nltype;                  /* Newline type */    int  external_options;            /* External (initial) options */
1844    int  nllen;                   /* Newline string length */    int  external_flags;              /* External flag bits to be set */
1845    uschar nl[4];                 /* Newline string when fixed length */    int  req_varyopt;                 /* "After variable item" flag for reqbyte */
1846      BOOL had_accept;                  /* (*ACCEPT) encountered */
1847      BOOL check_lookbehind;            /* Lookbehinds need later checking */
1848      int  nltype;                      /* Newline type */
1849      int  nllen;                       /* Newline string length */
1850      pcre_uchar nl[4];                 /* Newline string when fixed length */
1851  } compile_data;  } compile_data;
1852    
1853  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
1854  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
1855    
1856  typedef struct branch_chain {  typedef struct branch_chain {
1857    struct branch_chain *outer;    struct branch_chain *outer;
1858    uschar *current;    pcre_uchar *current_branch;
1859  } branch_chain;  } branch_chain;
1860    
1861  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
1862  call within the pattern. */  call within the pattern; used by pcre_exec(). */
1863    
1864  typedef struct recursion_info {  typedef struct recursion_info {
1865    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1866    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
1867    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
1868    USPTR save_start;             /* Old value of md->start_match */    int saved_max;                  /* Number of saved offsets */
1869    int *offset_save;             /* Pointer to start of saved offsets */    PCRE_PUCHAR subject_position;   /* Position at start of recursion */
   int saved_max;                /* Number of saved offsets */  
1870  } recursion_info;  } recursion_info;
1871    
1872  /* When compiling in a mode that doesn't use recursive calls to match(),  /* A similar structure for pcre_dfa_exec(). */
 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. */  
1873    
1874  struct heapframe;  typedef struct dfa_recursion_info {
1875      struct dfa_recursion_info *prevrec;
1876      int group_num;
1877      PCRE_PUCHAR subject_position;
1878    } dfa_recursion_info;
1879    
1880  /* 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
1881  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
1882  has been matched by a subpattern - to break infinite loops. */  has been matched by a subpattern - to break infinite loops; used by
1883    pcre_exec(). */
1884    
1885  typedef struct eptrblock {  typedef struct eptrblock {
1886    struct eptrblock *epb_prev;    struct eptrblock *epb_prev;
1887    USPTR epb_saved_eptr;    PCRE_PUCHAR epb_saved_eptr;
1888  } eptrblock;  } eptrblock;
1889    
1890    
# Line 914  typedef struct match_data { Line 1895  typedef struct match_data {
1895    unsigned long int match_call_count;      /* As it says */    unsigned long int match_call_count;      /* As it says */
1896    unsigned long int match_limit;           /* As it says */    unsigned long int match_limit;           /* As it says */
1897    unsigned long int match_limit_recursion; /* As it says */    unsigned long int match_limit_recursion; /* As it says */
1898    int   *offset_vector;         /* Offset vector */    int   *offset_vector;           /* Offset vector */
1899    int    offset_end;            /* One past the end */    int    offset_end;              /* One past the end */
1900    int    offset_max;            /* The maximum usable for return data */    int    offset_max;              /* The maximum usable for return data */
1901    int    nltype;                /* Newline type */    int    nltype;                  /* Newline type */
1902    int    nllen;                 /* Newline string length */    int    nllen;                   /* Newline string length */
1903    uschar nl[4];                 /* Newline string when fixed */    int    name_count;              /* Number of names in name table */
1904    const uschar *lcc;            /* Points to lower casing table */    int    name_entry_size;         /* Size of entry in names table */
1905    const uschar *ctypes;         /* Points to table of type maps */    pcre_uchar *name_table;         /* Table of names */
1906    BOOL   offset_overflow;       /* Set if too many extractions */    pcre_uchar nl[4];               /* Newline string when fixed */
1907    BOOL   notbol;                /* NOTBOL flag */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
1908    BOOL   noteol;                /* NOTEOL flag */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
1909    BOOL   utf8;                  /* UTF8 flag */    BOOL   offset_overflow;         /* Set if too many extractions */
1910    BOOL   endonly;               /* Dollar not before final \n */    BOOL   notbol;                  /* NOTBOL flag */
1911    BOOL   notempty;              /* Empty string match not wanted */    BOOL   noteol;                  /* NOTEOL flag */
1912    BOOL   partial;               /* PARTIAL flag */    BOOL   utf8;                    /* UTF8 flag */
1913    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
1914    const uschar *start_code;     /* For use when recursing */    BOOL   use_ucp;                 /* PCRE_UCP flag */
1915    USPTR  start_subject;         /* Start of the subject string */    BOOL   endonly;                 /* Dollar not before final \n */
1916    USPTR  end_subject;           /* End of the subject string */    BOOL   notempty;                /* Empty string match not wanted */
1917    USPTR  start_match;           /* Start of this match attempt */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
1918    USPTR  end_match_ptr;         /* Subject position at end match */    BOOL   hitend;                  /* Hit the end of the subject at some point */
1919    int    end_offset_top;        /* Highwater mark at end of match */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
1920    int    capture_last;          /* Most recent capture number */    BOOL   hasthen;                 /* Pattern contains (*THEN) */
1921    int    start_offset;          /* The start offset value */    const  pcre_uchar *start_code;  /* For use when recursing */
1922    eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */    PCRE_PUCHAR start_subject;      /* Start of the subject string */
1923    int    eptrn;                 /* Next free eptrblock */    PCRE_PUCHAR end_subject;        /* End of the subject string */
1924    recursion_info *recursive;    /* Linked list of recursion data */    PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
1925    void  *callout_data;          /* To pass back to callouts */    PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
1926    struct heapframe *thisframe;  /* Used only when compiling for no recursion */    PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
1927      int    partial;                 /* PARTIAL options */
1928      int    end_offset_top;          /* Highwater mark at end of match */
1929      int    capture_last;            /* Most recent capture number */
1930      int    start_offset;            /* The start offset value */
1931      int    match_function_type;     /* Set for certain special calls of MATCH() */
1932      eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
1933      int    eptrn;                   /* Next free eptrblock */
1934      recursion_info *recursive;      /* Linked list of recursion data */
1935      void  *callout_data;            /* To pass back to callouts */
1936      const  pcre_uchar *mark;        /* Mark pointer to pass back */
1937      const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
1938  } match_data;  } match_data;
1939    
1940  /* 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
1941  functions. */  functions. */
1942    
1943  typedef struct dfa_match_data {  typedef struct dfa_match_data {
1944    const uschar *start_code;     /* Start of the compiled pattern */    const pcre_uchar *start_code;     /* Start of the compiled pattern */
1945    const uschar *start_subject;  /* Start of the subject string */    const pcre_uchar *start_subject ; /* Start of the subject string */
1946    const uschar *end_subject;    /* End of subject string */    const pcre_uchar *end_subject;    /* End of subject string */
1947    const uschar *tables;         /* Character tables */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
1948    int   moptions;               /* Match options */    const pcre_uint8 *tables;         /* Character tables */
1949    int   poptions;               /* Pattern options */    int   start_offset;               /* The start offset value */
1950    int    nltype;                /* Newline type */    int   moptions;                   /* Match options */
1951    int    nllen;                 /* Newline string length */    int   poptions;                   /* Pattern options */
1952    uschar nl[4];                 /* Newline string when fixed */    int   nltype;                     /* Newline type */
1953    void  *callout_data;          /* To pass back to callouts */    int   nllen;                      /* Newline string length */
1954      pcre_uchar nl[4];                 /* Newline string when fixed */
1955      void *callout_data;               /* To pass back to callouts */
1956      dfa_recursion_info *recursive;    /* Linked list of recursion data */
1957  } dfa_match_data;  } dfa_match_data;
1958    
1959  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 967  typedef struct dfa_match_data { Line 1962  typedef struct dfa_match_data {
1962  #define ctype_letter  0x02  #define ctype_letter  0x02
1963  #define ctype_digit   0x04  #define ctype_digit   0x04
1964  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1965  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1966  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1967    
1968  /* 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 994  total length. */ Line 1989  total length. */
1989  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
1990  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1991    
1992    /* Internal function prefix */
1993    
1994    #ifdef COMPILE_PCRE8
1995    #define PRIV(name) _pcre_##name
1996    #else
1997    #ifdef COMPILE_PCRE16
1998    #define PRIV(name) _pcre16_##name
1999    #else
2000    #error Unsupported compiling mode
2001    #endif /* COMPILE_PCRE16 */
2002    #endif /* COMPILE_PCRE8 */
2003    
2004  /* 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
2005  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
2006    relocations in shared libraries, it now has an offset into a single string
2007    instead. */
2008    
2009  typedef struct {  typedef struct {
2010    const char *name;    pcre_uint16 name_offset;
2011    pcre_uint16 type;    pcre_uint16 type;
2012    pcre_uint16 value;    pcre_uint16 value;
2013  } ucp_type_table;  } ucp_type_table;
# Line 1009  of the exported public functions. They h Line 2018  of the exported public functions. They h
2018  but are not part of the PCRE public API. The data for these tables is in the  but are not part of the PCRE public API. The data for these tables is in the
2019  pcre_tables.c module. */  pcre_tables.c module. */
2020    
2021  extern const int    _pcre_utf8_table1[];  extern const int            PRIV(utf8_table1)[];
2022  extern const int    _pcre_utf8_table2[];  extern const int            PRIV(utf8_table2)[];
2023  extern const int    _pcre_utf8_table3[];  extern const int            PRIV(utf8_table3)[];
2024  extern const uschar _pcre_utf8_table4[];  extern const pcre_uint8     PRIV(utf8_table4)[];
2025    
2026  extern const int    _pcre_utf8_table1_size;  extern const int            PRIV(utf8_table1_size);
2027    
2028  extern const ucp_type_table _pcre_utt[];  extern const char           PRIV(utt_names)[];
2029  extern const int _pcre_utt_size;  extern const ucp_type_table PRIV(utt)[];
2030    extern const int            PRIV(utt_size);
2031    
2032  extern const uschar _pcre_default_tables[];  extern const pcre_uint8     PRIV(default_tables)[];
2033    
2034  extern const uschar _pcre_OP_lengths[];  extern const pcre_uint8     PRIV(OP_lengths)[];
2035    
2036    
2037  /* Internal shared functions. These are functions that are used by more than  /* Internal shared functions. These are functions that are used by more than
2038  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
2039  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
2040    
2041  extern BOOL         _pcre_is_newline(const uschar *, const uschar *, int *,  /* String comparison functions. */
2042                        BOOL);  #ifdef COMPILE_PCRE8
2043  extern int          _pcre_ord2utf8(int, uschar *);  
2044  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,  #define STRCMP_UC_UC(str1, str2) \
2045                        const pcre_study_data *, pcre_study_data *);    strcmp((char *)(str1), (char *)(str2))
2046  extern int          _pcre_ucp_findprop(const unsigned int, int *, int *);  #define STRCMP_UC_C8(str1, str2) \
2047  extern unsigned int _pcre_ucp_othercase(const unsigned int);    strcmp((char *)(str1), (str2))
2048  extern int          _pcre_valid_utf8(const uschar *, int);  #define STRNCMP_UC_UC(str1, str2, num) \
2049  extern BOOL         _pcre_was_newline(const uschar *, const uschar *, int *,    strncmp((char *)(str1), (char *)(str2), (num))
2050                        BOOL);  #define STRNCMP_UC_C8(str1, str2, num) \
2051  extern BOOL         _pcre_xclass(int, const uschar *);    strncmp((char *)(str1), (str2), (num))
2052    #define STRLEN_UC(str) strlen((const char *)str)
2053    
2054    #else
2055    
2056    extern int               PRIV(strcmp_uc_uc)(const pcre_uchar *,
2057                               const pcre_uchar *);
2058    extern int               PRIV(strcmp_uc_c8)(const pcre_uchar *,
2059                               const char *);
2060    extern int               PRIV(strncmp_uc_uc)(const pcre_uchar *,
2061                               const pcre_uchar *, unsigned int num);
2062    extern int               PRIV(strncmp_uc_c8)(const pcre_uchar *,
2063                               const char *, unsigned int num);
2064    extern unsigned int      PRIV(strlen_uc)(const pcre_uchar *str);
2065    
2066    #define STRCMP_UC_UC(str1, str2) \
2067      PRIV(strcmp_uc_uc)((str1), (str2))
2068    #define STRCMP_UC_C8(str1, str2) \
2069      PRIV(strcmp_uc_c8)((str1), (str2))
2070    #define STRNCMP_UC_UC(str1, str2, num) \
2071      PRIV(strncmp_uc_uc)((str1), (str2), (num))
2072    #define STRNCMP_UC_C8(str1, str2, num) \
2073      PRIV(strncmp_uc_c8)((str1), (str2), (num))
2074    #define STRLEN_UC(str) PRIV(strlen_uc)(str)
2075    
2076    #endif /* COMPILE_PCRE8 */
2077    
2078    extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2079    extern BOOL              PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2080                               int *, BOOL);
2081    extern int               PRIV(ord2utf8)(int, pcre_uint8 *);
2082    extern real_pcre        *PRIV(try_flipped)(const real_pcre *, real_pcre *,
2083                               const pcre_study_data *, pcre_study_data *);
2084    #ifndef COMPILE_PCRE16
2085    extern int               PRIV(valid_utf8)(PCRE_PUCHAR, int, int *);
2086    #else
2087    extern int               PRIV(valid_utf16)(PCRE_PUCHAR, int, int *);
2088    #endif
2089    extern BOOL              PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2090                               int *, BOOL);
2091    extern BOOL              PRIV(xclass)(int, const pcre_uchar *);
2092    
2093    #ifdef SUPPORT_JIT
2094    extern void              PRIV(jit_compile)(const real_pcre *, pcre_extra *);
2095    extern int               PRIV(jit_exec)(const real_pcre *, void *,
2096                               const pcre_uchar *, int, int, int, int, int *, int);
2097    extern void              PRIV(jit_free)(void *);
2098    #endif
2099    
2100    /* Unicode character database (UCD) */
2101    
2102    typedef struct {
2103      pcre_uint8 script;
2104      pcre_uint8 chartype;
2105      pcre_int32 other_case;
2106    } ucd_record;
2107    
2108    extern const ucd_record  PRIV(ucd_records)[];
2109    extern const pcre_uint8  PRIV(ucd_stage1)[];
2110    extern const pcre_uint16 PRIV(ucd_stage2)[];
2111    extern const int         PRIV(ucp_gentype)[];
2112    #ifdef SUPPORT_JIT
2113    extern const int         PRIV(ucp_typerange)[];
2114    #endif
2115    
2116    /* UCD access macros */
2117    
2118    #define UCD_BLOCK_SIZE 128
2119    #define GET_UCD(ch) (PRIV(ucd_records) + \
2120            PRIV(ucd_stage2)[PRIV(ucd_stage1)[(ch) / UCD_BLOCK_SIZE] * \
2121            UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
2122    
2123    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
2124    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
2125    #define UCD_CATEGORY(ch)  PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2126    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
2127    
2128  #endif  #endif
2129    

Legend:
Removed from v.101  
changed lines
  Added in v.769

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12