/[pcre]/code/trunk/pcre_internal.h
ViewVC logotype

Diff of /code/trunk/pcre_internal.h

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

revision 77 by nigel, Sat Feb 24 21:40:45 2007 UTC revision 929 by zherczeg, Fri Feb 24 11:07:47 2012 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) 1997-2005 University of Cambridge             Copyright (c) 1997-2012 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
47    #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
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    /* If SUPPORT_UCP is defined, SUPPORT_UTF must also be defined. The
61    "configure" script ensures this, but not everybody uses "configure". */
62    
63    #if defined SUPPORT_UCP && !(defined SUPPORT_UTF)
64    #define SUPPORT_UTF 1
65    #endif
66    
67    /* We define SUPPORT_UTF if SUPPORT_UTF8 is enabled for compatibility
68    reasons with existing code. */
69    
70    #if defined SUPPORT_UTF8 && !(defined SUPPORT_UTF)
71    #define SUPPORT_UTF 1
72    #endif
73    
74    /* Fixme: SUPPORT_UTF8 should be eventually disappear from the code.
75    Until then we define it if SUPPORT_UTF is defined. */
76    
77    #if defined SUPPORT_UTF && !(defined SUPPORT_UTF8)
78    #define SUPPORT_UTF8 1
79    #endif
80    
81    /* We do not support both EBCDIC and UTF-8/16 at the same time. The "configure"
82    script prevents both being selected, but not everybody uses "configure". */
83    
84    #if defined EBCDIC && defined SUPPORT_UTF
85    #error The use of both EBCDIC and SUPPORT_UTF8/16 is not supported.
86    #endif
87    
88  /* 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
89  inline, and there are *still* stupid compilers about that don't like indented  inline, and there are *still* stupid compilers about that don't like indented
90  pre-processor statements, or at least there were when I first wrote this. After  pre-processor statements, or at least there were when I first wrote this. After
91  all, it had only been about 10 years then... */  all, it had only been about 10 years then...
92    
93  #ifdef DEBUG  It turns out that the Mac Debugging.h header also defines the macro DPRINTF, so
94    be absolutely sure we get our version. */
95    
96    #undef DPRINTF
97    #ifdef PCRE_DEBUG
98  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
99  #else  #else
100  #define DPRINTF(p) /*nothing*/  #define DPRINTF(p) /* Nothing */
101  #endif  #endif
102    
103    
 /* Get the definitions provided by running "configure" */  
   
 #include "config.h"  
   
104  /* Standard C headers plus the external interface definition. The only time  /* Standard C headers plus the external interface definition. The only time
105  setjmp and stdarg are used is when NO_RECURSE is set. */  setjmp and stdarg are used is when NO_RECURSE is set. */
106    
107  #include <ctype.h>  #include <ctype.h>
108  #include <limits.h>  #include <limits.h>
 #include <setjmp.h>  
 #include <stdarg.h>  
109  #include <stddef.h>  #include <stddef.h>
110  #include <stdio.h>  #include <stdio.h>
111  #include <stdlib.h>  #include <stdlib.h>
112  #include <string.h>  #include <string.h>
113    
114  #ifndef PCRE_SPY  /* When compiling a DLL for Windows, the exported symbols have to be declared
115  #define PCRE_DEFINITION       /* Win32 __declspec(export) trigger for .dll */  using some MS magic. I found some useful information on this web page:
116    http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
117    information there, using __declspec(dllexport) without "extern" we have a
118    definition; with "extern" we have a declaration. The settings here override the
119    setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
120    which is all that is needed for applications (they just import the symbols). We
121    use:
122    
123      PCRE_EXP_DECL       for declarations
124      PCRE_EXP_DEFN       for definitions of exported functions
125      PCRE_EXP_DATA_DEFN  for definitions of exported variables
126    
127    The reason for the two DEFN macros is that in non-Windows environments, one
128    does not want to have "extern" before variable definitions because it leads to
129    compiler warnings. So we distinguish between functions and variables. In
130    Windows, the two should always be the same.
131    
132    The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
133    which is an application, but needs to import this file in order to "peek" at
134    internals, can #include pcre.h first to get an application's-eye view.
135    
136    In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
137    special-purpose environments) might want to stick other stuff in front of
138    exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
139    PCRE_EXP_DATA_DEFN only if they are not already set. */
140    
141    #ifndef PCRE_EXP_DECL
142    #  ifdef _WIN32
143    #    ifndef PCRE_STATIC
144    #      define PCRE_EXP_DECL       extern __declspec(dllexport)
145    #      define PCRE_EXP_DEFN       __declspec(dllexport)
146    #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
147    #    else
148    #      define PCRE_EXP_DECL       extern
149    #      define PCRE_EXP_DEFN
150    #      define PCRE_EXP_DATA_DEFN
151    #    endif
152    #  else
153    #    ifdef __cplusplus
154    #      define PCRE_EXP_DECL       extern "C"
155    #    else
156    #      define PCRE_EXP_DECL       extern
157    #    endif
158    #    ifndef PCRE_EXP_DEFN
159    #      define PCRE_EXP_DEFN       PCRE_EXP_DECL
160    #    endif
161    #    ifndef PCRE_EXP_DATA_DEFN
162    #      define PCRE_EXP_DATA_DEFN
163    #    endif
164    #  endif
165    #endif
166    
167    /* When compiling with the MSVC compiler, it is sometimes necessary to include
168    a "calling convention" before exported function names. (This is secondhand
169    information; I know nothing about MSVC myself). For example, something like
170    
171      void __cdecl function(....)
172    
173    might be needed. In order so make this easy, all the exported functions have
174    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
175    set, we ensure here that it has no effect. */
176    
177    #ifndef PCRE_CALL_CONVENTION
178    #define PCRE_CALL_CONVENTION
179  #endif  #endif
180    
181  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 8, 16 and 32-bit integers. We
182  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
183  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
184  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
185  preprocessor time in standard C environments. */  preprocessor time in standard C environments. */
186    
187    typedef unsigned char pcre_uint8;
188    
189  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
190    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
191      typedef short pcre_int16;
192  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
193    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
194      typedef int pcre_int16;
195  #else  #else
196    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
197  #endif  #endif
198    
199  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
200    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
201      typedef int pcre_int32;
202  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
203    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
204      typedef long int pcre_int32;
205  #else  #else
206    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
207  #endif  #endif
208    
209    /* When checking for integer overflow in pcre_compile(), we need to handle
210    large integers. If a 64-bit integer type is available, we can use that.
211    Otherwise we have to cast to double, which of course requires floating point
212    arithmetic. Handle this by defining a macro for the appropriate type. If
213    stdint.h is available, include it; it may define INT64_MAX. Systems that do not
214    have stdint.h (e.g. Solaris) may have inttypes.h. The macro int64_t may be set
215    by "configure". */
216    
217    #if HAVE_STDINT_H
218    #include <stdint.h>
219    #elif HAVE_INTTYPES_H
220    #include <inttypes.h>
221    #endif
222    
223    #if defined INT64_MAX || defined int64_t
224    #define INT64_OR_DOUBLE int64_t
225    #else
226    #define INT64_OR_DOUBLE double
227    #endif
228    
229  /* All character handling must be done as unsigned characters. Otherwise there  /* All character handling must be done as unsigned characters. Otherwise there
230  are problems with top-bit-set characters and functions such as isspace().  are problems with top-bit-set characters and functions such as isspace().
231  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 *,
232  should make things easier for callers. We define a short type for unsigned char  because that should make things easier for callers. This character type is
233  to save lots of typing. I tried "uchar", but it causes problems on Digital  called pcre_uchar.
234  Unix, where it is defined in sys/types, so use "uschar" instead. */  
235    The IN_UCHARS macro multiply its argument with the byte size of the current
236    pcre_uchar type. Useful for memcpy and such operations, whose require the
237    byte size of their input/output buffers.
238    
239    The MAX_255 macro checks whether its pcre_uchar input is less than 256.
240    
241    The TABLE_GET macro is designed for accessing elements of tables whose contain
242    exactly 256 items. When the character is able to contain more than 256
243    items, some check is needed before accessing these tables.
244    */
245    
246  typedef unsigned char uschar;  #ifdef COMPILE_PCRE8
247    
248  /* Include the public PCRE header */  typedef unsigned char pcre_uchar;
249    #define IN_UCHARS(x) (x)
250    #define MAX_255(c) 1
251    #define TABLE_GET(c, table, default) ((table)[c])
252    
253  #include "pcre.h"  #else
254    
255  /* Include the (copy of) the public ucp header, changing the external name into  #ifdef COMPILE_PCRE16
256  a private one. This does no harm, even if we aren't compiling UCP support. */  #if USHRT_MAX != 65535
257    /* This is a warning message. Change PCRE_UCHAR16 to a 16 bit data type in
258    pcre.h(.in) and disable (comment out) this message. */
259    #error Warning: PCRE_UCHAR16 is not a 16 bit data type.
260    #endif
261    
262    typedef pcre_uint16 pcre_uchar;
263    #define IN_UCHARS(x) ((x) << 1)
264    #define MAX_255(c) ((c) <= 255u)
265    #define TABLE_GET(c, table, default) (MAX_255(c)? ((table)[c]):(default))
266    
267    #else
268    #error Unsupported compiling mode
269    #endif /* COMPILE_PCRE16 */
270    
271  #define ucp_findchar _pcre_ucp_findchar  #endif /* COMPILE_PCRE8 */
272    
273    /* This is an unsigned int value that no character can ever have. UTF-8
274    characters only go up to 0x7fffffff (though Unicode doesn't go beyond
275    0x0010ffff). */
276    
277    #define NOTACHAR 0xffffffff
278    
279    /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
280    "any" and "anycrlf" at present). The following macros are used to package up
281    testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
282    modules to indicate in which datablock the parameters exist, and what the
283    start/end of string field names are. */
284    
285    #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
286    #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
287    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
288    
289    /* This macro checks for a newline at the given position */
290    
291    #define IS_NEWLINE(p) \
292      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
293        ((p) < NLBLOCK->PSEND && \
294         PRIV(is_newline)((p), NLBLOCK->nltype, NLBLOCK->PSEND, \
295           &(NLBLOCK->nllen), utf)) \
296        : \
297        ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
298         (p)[0] == NLBLOCK->nl[0] && \
299         (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
300        ) \
301      )
302    
303    /* This macro checks for a newline immediately preceding the given position */
304    
305    #define WAS_NEWLINE(p) \
306      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
307        ((p) > NLBLOCK->PSSTART && \
308         PRIV(was_newline)((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
309           &(NLBLOCK->nllen), utf)) \
310        : \
311        ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
312         (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
313         (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
314        ) \
315      )
316    
317    /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
318    with a custom type. This makes it possible, for example, to allow pcre_exec()
319    to process subject strings that are discontinuous by using a smart pointer
320    class. It must always be possible to inspect all of the subject string in
321    pcre_exec() because of the way it backtracks. Two macros are required in the
322    normal case, for sign-unspecified and unsigned char pointers. The former is
323    used for the external interface and appears in pcre.h, which is why its name
324    must begin with PCRE_. */
325    
326    #ifdef CUSTOM_SUBJECT_PTR
327    #define PCRE_PUCHAR CUSTOM_SUBJECT_PTR
328    #else
329    #define PCRE_PUCHAR const pcre_uchar *
330    #endif
331    
332    /* Include the public PCRE header and the definitions of UCP character property
333    values. */
334    
335    #include "pcre.h"
336  #include "ucp.h"  #include "ucp.h"
337    
338  /* When compiling for use with the Virtual Pascal compiler, these functions  /* When compiling for use with the Virtual Pascal compiler, these functions
# Line 127  need to have their names changed. PCRE m Line 340  need to have their names changed. PCRE m
340  option on the command line. */  option on the command line. */
341    
342  #ifdef VPCOMPAT  #ifdef VPCOMPAT
343    #define strlen(s)        _strlen(s)
344  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
345    #define memcmp(s,c,n)    _memcmp(s,c,n)
346  #define memcpy(d,s,n)    _memcpy(d,s,n)  #define memcpy(d,s,n)    _memcpy(d,s,n)
347  #define memmove(d,s,n)   _memmove(d,s,n)  #define memmove(d,s,n)   _memmove(d,s,n)
348  #define memset(s,c,n)    _memset(s,c,n)  #define memset(s,c,n)    _memset(s,c,n)
# Line 136  option on the command line. */ Line 351  option on the command line. */
351  /* 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(),
352  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
353  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
354  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. */  
355    
356  #if ! HAVE_MEMMOVE  #ifndef HAVE_MEMMOVE
357  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
358  #if HAVE_BCOPY  #ifdef HAVE_BCOPY
359  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
360  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
361  void *  static void *
362  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)  pcre_memmove(void *d, const void *s, size_t n)
363  {  {
364  int i;  size_t i;
365  dest += n;  unsigned char *dest = (unsigned char *)d;
366  src += n;  const unsigned char *src = (const unsigned char *)s;
367  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  if (dest > src)
368      {
369      dest += n;
370      src += n;
371      for (i = 0; i < n; ++i) *(--dest) = *(--src);
372      return (void *)dest;
373      }
374    else
375      {
376      for (i = 0; i < n; ++i) *dest++ = *src++;
377      return (void *)(dest - n);
378      }
379  }  }
380  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
381  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 172  The macros are controlled by the value o Line 396  The macros are controlled by the value o
396  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
397  is automated on Unix systems via the "configure" command. */  is automated on Unix systems via the "configure" command. */
398    
399    #ifdef COMPILE_PCRE8
400    
401  #if LINK_SIZE == 2  #if LINK_SIZE == 2
402    
403  #define PUT(a,n,d)   \  #define PUT(a,n,d)   \
# Line 208  is automated on Unix systems via the "co Line 434  is automated on Unix systems via the "co
434  #define GET(a,n) \  #define GET(a,n) \
435    (((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])
436    
437  #define MAX_PATTERN_SIZE (1 << 30)   /* Keep it positive */  /* Keep it positive */
438    #define MAX_PATTERN_SIZE (1 << 30)
439    
440    #else
441    #error LINK_SIZE must be either 2, 3, or 4
442    #endif
443    
444    #else /* COMPILE_PCRE8 */
445    
446    #ifdef COMPILE_PCRE16
447    
448    #if LINK_SIZE == 2
449    
450    #undef LINK_SIZE
451    #define LINK_SIZE 1
452    
453    #define PUT(a,n,d)   \
454      (a[n] = (d))
455    
456    #define GET(a,n) \
457      (a[n])
458    
459    #define MAX_PATTERN_SIZE (1 << 16)
460    
461    #elif LINK_SIZE == 3 || LINK_SIZE == 4
462    
463    #undef LINK_SIZE
464    #define LINK_SIZE 2
465    
466    #define PUT(a,n,d)   \
467      (a[n] = (d) >> 16), \
468      (a[(n)+1] = (d) & 65535)
469    
470    #define GET(a,n) \
471      (((a)[n] << 16) | (a)[(n)+1])
472    
473    /* Keep it positive */
474    #define MAX_PATTERN_SIZE (1 << 30)
475    
476  #else  #else
477  #error LINK_SIZE must be either 2, 3, or 4  #error LINK_SIZE must be either 2, 3, or 4
478  #endif  #endif
479    
480    #else
481    #error Unsupported compiling mode
482    #endif /* COMPILE_PCRE16 */
483    
484    #endif /* COMPILE_PCRE8 */
485    
486  /* Convenience macro defined in terms of the others */  /* Convenience macro defined in terms of the others */
487    
# Line 225  is automated on Unix systems via the "co Line 492  is automated on Unix systems via the "co
492  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
493  capturing parenthesis numbers in back references. */  capturing parenthesis numbers in back references. */
494    
495    #ifdef COMPILE_PCRE8
496    
497    #define IMM2_SIZE 2
498    
499  #define PUT2(a,n,d)   \  #define PUT2(a,n,d)   \
500    a[n] = (d) >> 8; \    a[n] = (d) >> 8; \
501    a[(n)+1] = (d) & 255    a[(n)+1] = (d) & 255
# Line 232  capturing parenthesis numbers in back re Line 503  capturing parenthesis numbers in back re
503  #define GET2(a,n) \  #define GET2(a,n) \
504    (((a)[n] << 8) | (a)[(n)+1])    (((a)[n] << 8) | (a)[(n)+1])
505    
506  #define PUT2INC(a,n,d)  PUT2(a,n,d), a += 2  #else /* COMPILE_PCRE8 */
507    
508    #ifdef COMPILE_PCRE16
509    
510    #define IMM2_SIZE 1
511    
512    #define PUT2(a,n,d)   \
513       a[n] = d
514    
515    #define GET2(a,n) \
516       a[n]
517    
518    #else
519    #error Unsupported compiling mode
520    #endif /* COMPILE_PCRE16 */
521    
522    #endif /* COMPILE_PCRE8 */
523    
524  /* When UTF-8 encoding is being used, a character is no longer just a single  #define PUT2INC(a,n,d)  PUT2(a,n,d), a += IMM2_SIZE
 byte. The macros for character handling generate simple sequences when used in  
 byte-mode, and more complicated ones for UTF-8 characters. */  
525    
526  #ifndef SUPPORT_UTF8  /* When UTF encoding is being used, a character is no longer just a single
527    character. The macros for character handling generate simple sequences when
528    used in character-mode, and more complicated ones for UTF characters.
529    GETCHARLENTEST and other macros are not used when UTF is not supported,
530    so they are not defined. To make sure they can never even appear when
531    UTF support is omitted, we don't even define them. */
532    
533    #ifndef SUPPORT_UTF
534    
535    /* #define MAX_VALUE_FOR_SINGLE_CHAR */
536    /* #define HAS_EXTRALEN(c) */
537    /* #define GET_EXTRALEN(c) */
538    /* #define NOT_FIRSTCHAR(c) */
539  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
540  #define GETCHARTEST(c, eptr) c = *eptr;  #define GETCHARTEST(c, eptr) c = *eptr;
541  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
542  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
543  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
544  #define BACKCHAR(eptr)  /* #define GETCHARLENTEST(c, eptr, len) */
545    /* #define BACKCHAR(eptr) */
546    /* #define FORWARDCHAR(eptr) */
547    /* #define ACROSSCHAR(condition, eptr, action) */
548    
549  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF */
550    
551    #ifdef COMPILE_PCRE8
552    
553    /* These macros were originally written in the form of loops that used data
554    from the tables whose names start with PRIV(utf8_table). They were rewritten by
555    a user so as not to use loops, because in some environments this gives a
556    significant performance advantage, and it seems never to do any harm. */
557    
558    /* Tells the biggest code point which can be encoded as a single character. */
559    
560    #define MAX_VALUE_FOR_SINGLE_CHAR 127
561    
562    /* Tests whether the code point needs extra characters to decode. */
563    
564    #define HAS_EXTRALEN(c) ((c) >= 0xc0)
565    
566    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
567    Otherwise it has an undefined behaviour. */
568    
569    #define GET_EXTRALEN(c) (PRIV(utf8_table4)[(c) & 0x3f])
570    
571    /* Returns TRUE, if the given character is not the first character
572    of a UTF sequence. */
573    
574    #define NOT_FIRSTCHAR(c) (((c) & 0xc0) == 0x80)
575    
576    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
577    advancing the pointer. */
578    
579    #define GETUTF8(c, eptr) \
580        { \
581        if ((c & 0x20) == 0) \
582          c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
583        else if ((c & 0x10) == 0) \
584          c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
585        else if ((c & 0x08) == 0) \
586          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
587          ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
588        else if ((c & 0x04) == 0) \
589          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
590              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
591              (eptr[4] & 0x3f); \
592        else \
593          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
594              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
595              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
596        }
597    
598  /* 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
599  we know we are in UTF-8 mode. */  we know we are in UTF-8 mode. */
600    
601  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
602    c = *eptr; \    c = *eptr; \
603    if ((c & 0xc0) == 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; \  
       } \  
     }  
604    
605  /* 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
606  pointer. */  pointer. */
607    
608  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
609    c = *eptr; \    c = *eptr; \
610    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf && c >= 0xc0) GETUTF8(c, eptr);
611    
612    /* Base macro to pick up the remaining bytes of a UTF-8 character, advancing
613    the pointer. */
614    
615    #define GETUTF8INC(c, eptr) \
616      { \      { \
617      int gcii; \      if ((c & 0x20) == 0) \
618      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \        c = ((c & 0x1f) << 6) | (*eptr++ & 0x3f); \
619      int gcss = 6*gcaa; \      else if ((c & 0x10) == 0) \
620      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        { \
621      for (gcii = 1; gcii <= gcaa; gcii++) \        c = ((c & 0x0f) << 12) | ((*eptr & 0x3f) << 6) | (eptr[1] & 0x3f); \
622          eptr += 2; \
623          } \
624        else if ((c & 0x08) == 0) \
625        { \        { \
626        gcss -= 6; \        c = ((c & 0x07) << 18) | ((*eptr & 0x3f) << 12) | \
627        c |= (eptr[gcii] & 0x3f) << gcss; \            ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
628          eptr += 3; \
629          } \
630        else if ((c & 0x04) == 0) \
631          { \
632          c = ((c & 0x03) << 24) | ((*eptr & 0x3f) << 18) | \
633              ((eptr[1] & 0x3f) << 12) | ((eptr[2] & 0x3f) << 6) | \
634              (eptr[3] & 0x3f); \
635          eptr += 4; \
636          } \
637        else \
638          { \
639          c = ((c & 0x01) << 30) | ((*eptr & 0x3f) << 24) | \
640              ((eptr[1] & 0x3f) << 18) | ((eptr[2] & 0x3f) << 12) | \
641              ((eptr[3] & 0x3f) << 6) | (eptr[4] & 0x3f); \
642          eptr += 5; \
643        } \        } \
644      }      }
645    
# Line 290  know we are in UTF-8 mode. */ Line 648  know we are in UTF-8 mode. */
648    
649  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
650    c = *eptr++; \    c = *eptr++; \
651    if ((c & 0xc0) == 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; \  
       } \  
     }  
652    
653  /* 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.
654    This is called when we don't know if we are in UTF-8 mode. */
655    
656  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
657    c = *eptr++; \    c = *eptr++; \
658    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf && c >= 0xc0) GETUTF8INC(c, eptr);
659    
660    /* Base macro to pick up the remaining bytes of a UTF-8 character, not
661    advancing the pointer, incrementing the length. */
662    
663    #define GETUTF8LEN(c, eptr, len) \
664      { \      { \
665      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      if ((c & 0x20) == 0) \
666      int gcss = 6*gcaa; \        { \
667      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \        c = ((c & 0x1f) << 6) | (eptr[1] & 0x3f); \
668      while (gcaa-- > 0) \        len++; \
669          } \
670        else if ((c & 0x10)  == 0) \
671        { \        { \
672        gcss -= 6; \        c = ((c & 0x0f) << 12) | ((eptr[1] & 0x3f) << 6) | (eptr[2] & 0x3f); \
673        c |= (*eptr++ & 0x3f) << gcss; \        len += 2; \
674          } \
675        else if ((c & 0x08)  == 0) \
676          {\
677          c = ((c & 0x07) << 18) | ((eptr[1] & 0x3f) << 12) | \
678              ((eptr[2] & 0x3f) << 6) | (eptr[3] & 0x3f); \
679          len += 3; \
680          } \
681        else if ((c & 0x04)  == 0) \
682          { \
683          c = ((c & 0x03) << 24) | ((eptr[1] & 0x3f) << 18) | \
684              ((eptr[2] & 0x3f) << 12) | ((eptr[3] & 0x3f) << 6) | \
685              (eptr[4] & 0x3f); \
686          len += 4; \
687          } \
688        else \
689          {\
690          c = ((c & 0x01) << 30) | ((eptr[1] & 0x3f) << 24) | \
691              ((eptr[2] & 0x3f) << 18) | ((eptr[3] & 0x3f) << 12) | \
692              ((eptr[4] & 0x3f) << 6) | (eptr[5] & 0x3f); \
693          len += 5; \
694        } \        } \
695      }      }
696    
# Line 323  if there are extra bytes. This is called Line 699  if there are extra bytes. This is called
699    
700  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
701    c = *eptr; \    c = *eptr; \
702    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) GETUTF8LEN(c, eptr, len);
703      { \  
704      int gcii; \  /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
705      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \  pointer, incrementing length if there are extra bytes. This is called when we
706      int gcss = 6*gcaa; \  do not know if we are in UTF-8 mode. */
707      c = (c & _pcre_utf8_table3[gcaa]) << gcss; \  
708      for (gcii = 1; gcii <= gcaa; gcii++) \  #define GETCHARLENTEST(c, eptr, len) \
709        { \    c = *eptr; \
710        gcss -= 6; \    if (utf && c >= 0xc0) GETUTF8LEN(c, eptr, len);
       c |= (eptr[gcii] & 0x3f) << gcss; \  
       } \  
     len += gcaa; \  
     }  
711    
712  /* 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
713  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
714    because almost all calls are already within a block of UTF-8 only code. */
715    
716  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
717    
718    /* Same as above, just in the other direction. */
719    #define FORWARDCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr++
720    
721    /* Same as above, but it allows a fully customizable form. */
722    #define ACROSSCHAR(condition, eptr, action) \
723      while((condition) && ((eptr) & 0xc0) == 0x80) action
724    
725    #else /* COMPILE_PCRE8 */
726    
727    #ifdef COMPILE_PCRE16
728    
729    /* Tells the biggest code point which can be encoded as a single character. */
730    
731    #define MAX_VALUE_FOR_SINGLE_CHAR 65535
732    
733    /* Tests whether the code point needs extra characters to decode. */
734    
735    #define HAS_EXTRALEN(c) (((c) & 0xfc00) == 0xd800)
736    
737    /* Returns with the additional number of characters if IS_MULTICHAR(c) is TRUE.
738    Otherwise it has an undefined behaviour. */
739    
740    #define GET_EXTRALEN(c) 1
741    
742    /* Returns TRUE, if the given character is not the first character
743    of a UTF sequence. */
744    
745    #define NOT_FIRSTCHAR(c) (((c) & 0xfc00) == 0xdc00)
746    
747    /* Base macro to pick up the low surrogate of a UTF-16 character, not
748    advancing the pointer. */
749    
750    #define GETUTF16(c, eptr) \
751       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; }
752    
753    /* Get the next UTF-16 character, not advancing the pointer. This is called when
754    we know we are in UTF-16 mode. */
755    
756    #define GETCHAR(c, eptr) \
757      c = *eptr; \
758      if ((c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
759    
760    /* Get the next UTF-16 character, testing for UTF-16 mode, and not advancing the
761    pointer. */
762    
763    #define GETCHARTEST(c, eptr) \
764      c = *eptr; \
765      if (utf && (c & 0xfc00) == 0xd800) GETUTF16(c, eptr);
766    
767    /* Base macro to pick up the low surrogate of a UTF-16 character, advancing
768    the pointer. */
769    
770    #define GETUTF16INC(c, eptr) \
771       { c = (((c & 0x3ff) << 10) | (*eptr++ & 0x3ff)) + 0x10000; }
772    
773    /* Get the next UTF-16 character, advancing the pointer. This is called when we
774    know we are in UTF-16 mode. */
775    
776    #define GETCHARINC(c, eptr) \
777      c = *eptr++; \
778      if ((c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
779    
780    /* Get the next character, testing for UTF-16 mode, and advancing the pointer.
781    This is called when we don't know if we are in UTF-16 mode. */
782    
783    #define GETCHARINCTEST(c, eptr) \
784      c = *eptr++; \
785      if (utf && (c & 0xfc00) == 0xd800) GETUTF16INC(c, eptr);
786    
787    /* Base macro to pick up the low surrogate of a UTF-16 character, not
788    advancing the pointer, incrementing the length. */
789    
790    #define GETUTF16LEN(c, eptr, len) \
791       { c = (((c & 0x3ff) << 10) | (eptr[1] & 0x3ff)) + 0x10000; len++; }
792    
793    /* Get the next UTF-16 character, not advancing the pointer, incrementing
794    length if there is a low surrogate. This is called when we know we are in
795    UTF-16 mode. */
796    
797    #define GETCHARLEN(c, eptr, len) \
798      c = *eptr; \
799      if ((c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
800    
801    /* Get the next UTF-816character, testing for UTF-16 mode, not advancing the
802    pointer, incrementing length if there is a low surrogate. This is called when
803    we do not know if we are in UTF-16 mode. */
804    
805    #define GETCHARLENTEST(c, eptr, len) \
806      c = *eptr; \
807      if (utf && (c & 0xfc00) == 0xd800) GETUTF16LEN(c, eptr, len);
808    
809    /* If the pointer is not at the start of a character, move it back until
810    it is. This is called only in UTF-16 mode - we don't put a test within the
811    macro because almost all calls are already within a block of UTF-16 only
812    code. */
813    
814    #define BACKCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr--
815    
816    /* Same as above, just in the other direction. */
817    #define FORWARDCHAR(eptr) if ((*eptr & 0xfc00) == 0xdc00) eptr++
818    
819    /* Same as above, but it allows a fully customizable form. */
820    #define ACROSSCHAR(condition, eptr, action) \
821      if ((condition) && ((eptr) & 0xfc00) == 0xdc00) action
822    
823  #endif  #endif
824    
825    #endif /* COMPILE_PCRE8 */
826    
827    #endif  /* SUPPORT_UTF */
828    
829    
830  /* In case there is no definition of offsetof() provided - though any proper  /* In case there is no definition of offsetof() provided - though any proper
831  Standard C system should have one. */  Standard C system should have one. */
# Line 353  Standard C system should have one. */ Line 835  Standard C system should have one. */
835  #endif  #endif
836    
837    
838  /* These are the public options that can change during matching. */  /* Private flags containing information about the compiled regex. They used to
839    live at the top end of the options word, but that got almost full, so now they
840  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
841    the restrictions on partial matching have been lifted. It remains for backwards
842    compatibility. */
843    
844  /* Private options flags start at the most significant end of the four bytes,  #ifdef COMPILE_PCRE8
845  but skip the top bit so we can use ints for convenience without getting tangled  #define PCRE_MODE          0x0001  /* compiled in 8 bit mode */
846  with negative values. The public options defined in pcre.h start at the least  #endif
847  significant end. Make sure they don't overlap! */  #ifdef COMPILE_PCRE16
848    #define PCRE_MODE          0x0002  /* compiled in 16 bit mode */
849  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #endif
850  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_FIRSTSET      0x0010  /* first_char is set */
851  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_FCH_CASELESS  0x0020  /* caseless first char */
852  #define PCRE_ICHANGED      0x08000000  /* i option changes within regex */  #define PCRE_REQCHSET      0x0040  /* req_byte is set */
853  #define PCRE_NOPARTIAL     0x04000000  /* can't use partial with this regex */  #define PCRE_RCH_CASELESS  0x0080  /* caseless requested char */
854    #define PCRE_STARTLINE     0x0100  /* start after \n for multiline */
855    #define PCRE_NOPARTIAL     0x0200  /* can't use partial with this regex */
856    #define PCRE_JCHANGED      0x0400  /* j option used in regex */
857    #define PCRE_HASCRORLF     0x0800  /* explicit \r or \n in pattern */
858    #define PCRE_HASTHEN       0x1000  /* pattern contains (*THEN) */
859    
860  /* Options for the "extra" block produced by pcre_study(). */  /* Flags for the "extra" block produced by pcre_study(). */
861    
862  #define PCRE_STUDY_MAPPED   0x01     /* a map of starting chars exists */  #define PCRE_STUDY_MAPPED  0x0001  /* a map of starting chars exists */
863    #define PCRE_STUDY_MINLEN  0x0002  /* a minimum length field exists */
864    
865  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
866  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
867    
868  #define PUBLIC_OPTIONS \  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
869                               PCRE_NEWLINE_ANYCRLF)
870    
871    #define PUBLIC_COMPILE_OPTIONS \
872    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
873     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
874     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| \
875       PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
876       PCRE_JAVASCRIPT_COMPAT|PCRE_UCP|PCRE_NO_START_OPTIMIZE)
877    
878  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
879    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
880     PCRE_PARTIAL)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS| \
881       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
882    
883  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
884    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NOTEMPTY_ATSTART| \
885     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART)     PCRE_NO_UTF8_CHECK|PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST| \
886       PCRE_DFA_RESTART|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
887       PCRE_NO_START_OPTIMIZE)
888    
889    #define PUBLIC_STUDY_OPTIONS \
890       (PCRE_STUDY_JIT_COMPILE|PCRE_STUDY_JIT_PARTIAL_SOFT_COMPILE| \
891        PCRE_STUDY_JIT_PARTIAL_HARD_COMPILE)
892    
893  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  /* Magic number to provide a small check against being handed junk. */
   
 /* Magic number to provide a small check against being handed junk. Also used  
 to detect whether a pattern was compiled on a host of different endianness. */  
894    
895  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */  #define MAGIC_NUMBER  0x50435245UL   /* 'PCRE' */
896    
897    /* This variable is used to detect a loaded regular expression
898    in different endianness. */
899    
900    #define REVERSED_MAGIC_NUMBER  0x45524350UL   /* 'ERCP' */
901    
902  /* Negative values for the firstchar and reqchar variables */  /* Negative values for the firstchar and reqchar variables */
903    
904  #define REQ_UNSET (-2)  #define REQ_UNSET (-2)
# Line 405  req_byte match. */ Line 909  req_byte match. */
909    
910  #define REQ_BYTE_MAX 1000  #define REQ_BYTE_MAX 1000
911    
912  /* Flags added to firstbyte or reqbyte; a "non-literal" item is either a  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
913  variable-length repeat, or a anything other than literal characters. */  environments where these macros are defined elsewhere. Unfortunately, there
914    is no way to do the same for the typedef. */
 #define REQ_CASELESS 0x0100    /* indicates caselessness */  
 #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  
   
 /* Miscellaneous definitions */  
915    
916  typedef int BOOL;  typedef int BOOL;
917    
918    #ifndef FALSE
919  #define FALSE   0  #define FALSE   0
920  #define TRUE    1  #define TRUE    1
921    #endif
922    
923  /* Escape items that are just an encoding of a particular data value. Note that  /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
924  ESC_n is defined as yet another macro, which is set in config.h to either \n  character constants like '*' because the compiler would emit their EBCDIC code,
925  (the default) or \r (which some people want). */  which is different from their ASCII/UTF-8 code. Instead we define macros for
926    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
927    is enabled. When UTF-8 support is not enabled, the definitions use character
928    literals. Both character and string versions of each character are needed, and
929    there are some longer strings as well.
930    
931    This means that, on EBCDIC platforms, the PCRE library can handle either
932    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
933    would need different lookups depending on whether PCRE_UTF8 was set or not.
934    This would make it impossible to use characters in switch/case statements,
935    which would reduce performance. For a theoretical use (which nobody has asked
936    for) in a minority area (EBCDIC platforms), this is not sensible. Any
937    application that did need both could compile two versions of the library, using
938    macros to give the functions distinct names. */
939    
940    #ifndef SUPPORT_UTF
941    
942    /* UTF-8 support is not enabled; use the platform-dependent character literals
943    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
944    
945    #define CHAR_HT                     '\t'
946    #define CHAR_VT                     '\v'
947    #define CHAR_FF                     '\f'
948    #define CHAR_CR                     '\r'
949    #define CHAR_NL                     '\n'
950    #define CHAR_BS                     '\b'
951    #define CHAR_BEL                    '\a'
952    #ifdef EBCDIC
953    #define CHAR_ESC                    '\047'
954    #define CHAR_DEL                    '\007'
955    #else
956    #define CHAR_ESC                    '\033'
957    #define CHAR_DEL                    '\177'
958    #endif
959    
960    #define CHAR_SPACE                  ' '
961    #define CHAR_EXCLAMATION_MARK       '!'
962    #define CHAR_QUOTATION_MARK         '"'
963    #define CHAR_NUMBER_SIGN            '#'
964    #define CHAR_DOLLAR_SIGN            '$'
965    #define CHAR_PERCENT_SIGN           '%'
966    #define CHAR_AMPERSAND              '&'
967    #define CHAR_APOSTROPHE             '\''
968    #define CHAR_LEFT_PARENTHESIS       '('
969    #define CHAR_RIGHT_PARENTHESIS      ')'
970    #define CHAR_ASTERISK               '*'
971    #define CHAR_PLUS                   '+'
972    #define CHAR_COMMA                  ','
973    #define CHAR_MINUS                  '-'
974    #define CHAR_DOT                    '.'
975    #define CHAR_SLASH                  '/'
976    #define CHAR_0                      '0'
977    #define CHAR_1                      '1'
978    #define CHAR_2                      '2'
979    #define CHAR_3                      '3'
980    #define CHAR_4                      '4'
981    #define CHAR_5                      '5'
982    #define CHAR_6                      '6'
983    #define CHAR_7                      '7'
984    #define CHAR_8                      '8'
985    #define CHAR_9                      '9'
986    #define CHAR_COLON                  ':'
987    #define CHAR_SEMICOLON              ';'
988    #define CHAR_LESS_THAN_SIGN         '<'
989    #define CHAR_EQUALS_SIGN            '='
990    #define CHAR_GREATER_THAN_SIGN      '>'
991    #define CHAR_QUESTION_MARK          '?'
992    #define CHAR_COMMERCIAL_AT          '@'
993    #define CHAR_A                      'A'
994    #define CHAR_B                      'B'
995    #define CHAR_C                      'C'
996    #define CHAR_D                      'D'
997    #define CHAR_E                      'E'
998    #define CHAR_F                      'F'
999    #define CHAR_G                      'G'
1000    #define CHAR_H                      'H'
1001    #define CHAR_I                      'I'
1002    #define CHAR_J                      'J'
1003    #define CHAR_K                      'K'
1004    #define CHAR_L                      'L'
1005    #define CHAR_M                      'M'
1006    #define CHAR_N                      'N'
1007    #define CHAR_O                      'O'
1008    #define CHAR_P                      'P'
1009    #define CHAR_Q                      'Q'
1010    #define CHAR_R                      'R'
1011    #define CHAR_S                      'S'
1012    #define CHAR_T                      'T'
1013    #define CHAR_U                      'U'
1014    #define CHAR_V                      'V'
1015    #define CHAR_W                      'W'
1016    #define CHAR_X                      'X'
1017    #define CHAR_Y                      'Y'
1018    #define CHAR_Z                      'Z'
1019    #define CHAR_LEFT_SQUARE_BRACKET    '['
1020    #define CHAR_BACKSLASH              '\\'
1021    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
1022    #define CHAR_CIRCUMFLEX_ACCENT      '^'
1023    #define CHAR_UNDERSCORE             '_'
1024    #define CHAR_GRAVE_ACCENT           '`'
1025    #define CHAR_a                      'a'
1026    #define CHAR_b                      'b'
1027    #define CHAR_c                      'c'
1028    #define CHAR_d                      'd'
1029    #define CHAR_e                      'e'
1030    #define CHAR_f                      'f'
1031    #define CHAR_g                      'g'
1032    #define CHAR_h                      'h'
1033    #define CHAR_i                      'i'
1034    #define CHAR_j                      'j'
1035    #define CHAR_k                      'k'
1036    #define CHAR_l                      'l'
1037    #define CHAR_m                      'm'
1038    #define CHAR_n                      'n'
1039    #define CHAR_o                      'o'
1040    #define CHAR_p                      'p'
1041    #define CHAR_q                      'q'
1042    #define CHAR_r                      'r'
1043    #define CHAR_s                      's'
1044    #define CHAR_t                      't'
1045    #define CHAR_u                      'u'
1046    #define CHAR_v                      'v'
1047    #define CHAR_w                      'w'
1048    #define CHAR_x                      'x'
1049    #define CHAR_y                      'y'
1050    #define CHAR_z                      'z'
1051    #define CHAR_LEFT_CURLY_BRACKET     '{'
1052    #define CHAR_VERTICAL_LINE          '|'
1053    #define CHAR_RIGHT_CURLY_BRACKET    '}'
1054    #define CHAR_TILDE                  '~'
1055    
1056    #define STR_HT                      "\t"
1057    #define STR_VT                      "\v"
1058    #define STR_FF                      "\f"
1059    #define STR_CR                      "\r"
1060    #define STR_NL                      "\n"
1061    #define STR_BS                      "\b"
1062    #define STR_BEL                     "\a"
1063    #ifdef EBCDIC
1064    #define STR_ESC                     "\047"
1065    #define STR_DEL                     "\007"
1066    #else
1067    #define STR_ESC                     "\033"
1068    #define STR_DEL                     "\177"
1069    #endif
1070    
1071    #define STR_SPACE                   " "
1072    #define STR_EXCLAMATION_MARK        "!"
1073    #define STR_QUOTATION_MARK          "\""
1074    #define STR_NUMBER_SIGN             "#"
1075    #define STR_DOLLAR_SIGN             "$"
1076    #define STR_PERCENT_SIGN            "%"
1077    #define STR_AMPERSAND               "&"
1078    #define STR_APOSTROPHE              "'"
1079    #define STR_LEFT_PARENTHESIS        "("
1080    #define STR_RIGHT_PARENTHESIS       ")"
1081    #define STR_ASTERISK                "*"
1082    #define STR_PLUS                    "+"
1083    #define STR_COMMA                   ","
1084    #define STR_MINUS                   "-"
1085    #define STR_DOT                     "."
1086    #define STR_SLASH                   "/"
1087    #define STR_0                       "0"
1088    #define STR_1                       "1"
1089    #define STR_2                       "2"
1090    #define STR_3                       "3"
1091    #define STR_4                       "4"
1092    #define STR_5                       "5"
1093    #define STR_6                       "6"
1094    #define STR_7                       "7"
1095    #define STR_8                       "8"
1096    #define STR_9                       "9"
1097    #define STR_COLON                   ":"
1098    #define STR_SEMICOLON               ";"
1099    #define STR_LESS_THAN_SIGN          "<"
1100    #define STR_EQUALS_SIGN             "="
1101    #define STR_GREATER_THAN_SIGN       ">"
1102    #define STR_QUESTION_MARK           "?"
1103    #define STR_COMMERCIAL_AT           "@"
1104    #define STR_A                       "A"
1105    #define STR_B                       "B"
1106    #define STR_C                       "C"
1107    #define STR_D                       "D"
1108    #define STR_E                       "E"
1109    #define STR_F                       "F"
1110    #define STR_G                       "G"
1111    #define STR_H                       "H"
1112    #define STR_I                       "I"
1113    #define STR_J                       "J"
1114    #define STR_K                       "K"
1115    #define STR_L                       "L"
1116    #define STR_M                       "M"
1117    #define STR_N                       "N"
1118    #define STR_O                       "O"
1119    #define STR_P                       "P"
1120    #define STR_Q                       "Q"
1121    #define STR_R                       "R"
1122    #define STR_S                       "S"
1123    #define STR_T                       "T"
1124    #define STR_U                       "U"
1125    #define STR_V                       "V"
1126    #define STR_W                       "W"
1127    #define STR_X                       "X"
1128    #define STR_Y                       "Y"
1129    #define STR_Z                       "Z"
1130    #define STR_LEFT_SQUARE_BRACKET     "["
1131    #define STR_BACKSLASH               "\\"
1132    #define STR_RIGHT_SQUARE_BRACKET    "]"
1133    #define STR_CIRCUMFLEX_ACCENT       "^"
1134    #define STR_UNDERSCORE              "_"
1135    #define STR_GRAVE_ACCENT            "`"
1136    #define STR_a                       "a"
1137    #define STR_b                       "b"
1138    #define STR_c                       "c"
1139    #define STR_d                       "d"
1140    #define STR_e                       "e"
1141    #define STR_f                       "f"
1142    #define STR_g                       "g"
1143    #define STR_h                       "h"
1144    #define STR_i                       "i"
1145    #define STR_j                       "j"
1146    #define STR_k                       "k"
1147    #define STR_l                       "l"
1148    #define STR_m                       "m"
1149    #define STR_n                       "n"
1150    #define STR_o                       "o"
1151    #define STR_p                       "p"
1152    #define STR_q                       "q"
1153    #define STR_r                       "r"
1154    #define STR_s                       "s"
1155    #define STR_t                       "t"
1156    #define STR_u                       "u"
1157    #define STR_v                       "v"
1158    #define STR_w                       "w"
1159    #define STR_x                       "x"
1160    #define STR_y                       "y"
1161    #define STR_z                       "z"
1162    #define STR_LEFT_CURLY_BRACKET      "{"
1163    #define STR_VERTICAL_LINE           "|"
1164    #define STR_RIGHT_CURLY_BRACKET     "}"
1165    #define STR_TILDE                   "~"
1166    
1167    #define STRING_ACCEPT0              "ACCEPT\0"
1168    #define STRING_COMMIT0              "COMMIT\0"
1169    #define STRING_F0                   "F\0"
1170    #define STRING_FAIL0                "FAIL\0"
1171    #define STRING_MARK0                "MARK\0"
1172    #define STRING_PRUNE0               "PRUNE\0"
1173    #define STRING_SKIP0                "SKIP\0"
1174    #define STRING_THEN                 "THEN"
1175    
1176    #define STRING_alpha0               "alpha\0"
1177    #define STRING_lower0               "lower\0"
1178    #define STRING_upper0               "upper\0"
1179    #define STRING_alnum0               "alnum\0"
1180    #define STRING_ascii0               "ascii\0"
1181    #define STRING_blank0               "blank\0"
1182    #define STRING_cntrl0               "cntrl\0"
1183    #define STRING_digit0               "digit\0"
1184    #define STRING_graph0               "graph\0"
1185    #define STRING_print0               "print\0"
1186    #define STRING_punct0               "punct\0"
1187    #define STRING_space0               "space\0"
1188    #define STRING_word0                "word\0"
1189    #define STRING_xdigit               "xdigit"
1190    
1191    #define STRING_DEFINE               "DEFINE"
1192    
1193    #define STRING_CR_RIGHTPAR             "CR)"
1194    #define STRING_LF_RIGHTPAR             "LF)"
1195    #define STRING_CRLF_RIGHTPAR           "CRLF)"
1196    #define STRING_ANY_RIGHTPAR            "ANY)"
1197    #define STRING_ANYCRLF_RIGHTPAR        "ANYCRLF)"
1198    #define STRING_BSR_ANYCRLF_RIGHTPAR    "BSR_ANYCRLF)"
1199    #define STRING_BSR_UNICODE_RIGHTPAR    "BSR_UNICODE)"
1200    #ifdef COMPILE_PCRE8
1201    #define STRING_UTF_RIGHTPAR            "UTF8)"
1202    #endif
1203    #ifdef COMPILE_PCRE16
1204    #define STRING_UTF_RIGHTPAR            "UTF16)"
1205    #endif
1206    #define STRING_UCP_RIGHTPAR            "UCP)"
1207    #define STRING_NO_START_OPT_RIGHTPAR   "NO_START_OPT)"
1208    
1209    #else  /* SUPPORT_UTF */
1210    
1211    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
1212    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
1213    only. */
1214    
1215    #define CHAR_HT                     '\011'
1216    #define CHAR_VT                     '\013'
1217    #define CHAR_FF                     '\014'
1218    #define CHAR_CR                     '\015'
1219    #define CHAR_NL                     '\012'
1220    #define CHAR_BS                     '\010'
1221    #define CHAR_BEL                    '\007'
1222    #define CHAR_ESC                    '\033'
1223    #define CHAR_DEL                    '\177'
1224    
1225    #define CHAR_SPACE                  '\040'
1226    #define CHAR_EXCLAMATION_MARK       '\041'
1227    #define CHAR_QUOTATION_MARK         '\042'
1228    #define CHAR_NUMBER_SIGN            '\043'
1229    #define CHAR_DOLLAR_SIGN            '\044'
1230    #define CHAR_PERCENT_SIGN           '\045'
1231    #define CHAR_AMPERSAND              '\046'
1232    #define CHAR_APOSTROPHE             '\047'
1233    #define CHAR_LEFT_PARENTHESIS       '\050'
1234    #define CHAR_RIGHT_PARENTHESIS      '\051'
1235    #define CHAR_ASTERISK               '\052'
1236    #define CHAR_PLUS                   '\053'
1237    #define CHAR_COMMA                  '\054'
1238    #define CHAR_MINUS                  '\055'
1239    #define CHAR_DOT                    '\056'
1240    #define CHAR_SLASH                  '\057'
1241    #define CHAR_0                      '\060'
1242    #define CHAR_1                      '\061'
1243    #define CHAR_2                      '\062'
1244    #define CHAR_3                      '\063'
1245    #define CHAR_4                      '\064'
1246    #define CHAR_5                      '\065'
1247    #define CHAR_6                      '\066'
1248    #define CHAR_7                      '\067'
1249    #define CHAR_8                      '\070'
1250    #define CHAR_9                      '\071'
1251    #define CHAR_COLON                  '\072'
1252    #define CHAR_SEMICOLON              '\073'
1253    #define CHAR_LESS_THAN_SIGN         '\074'
1254    #define CHAR_EQUALS_SIGN            '\075'
1255    #define CHAR_GREATER_THAN_SIGN      '\076'
1256    #define CHAR_QUESTION_MARK          '\077'
1257    #define CHAR_COMMERCIAL_AT          '\100'
1258    #define CHAR_A                      '\101'
1259    #define CHAR_B                      '\102'
1260    #define CHAR_C                      '\103'
1261    #define CHAR_D                      '\104'
1262    #define CHAR_E                      '\105'
1263    #define CHAR_F                      '\106'
1264    #define CHAR_G                      '\107'
1265    #define CHAR_H                      '\110'
1266    #define CHAR_I                      '\111'
1267    #define CHAR_J                      '\112'
1268    #define CHAR_K                      '\113'
1269    #define CHAR_L                      '\114'
1270    #define CHAR_M                      '\115'
1271    #define CHAR_N                      '\116'
1272    #define CHAR_O                      '\117'
1273    #define CHAR_P                      '\120'
1274    #define CHAR_Q                      '\121'
1275    #define CHAR_R                      '\122'
1276    #define CHAR_S                      '\123'
1277    #define CHAR_T                      '\124'
1278    #define CHAR_U                      '\125'
1279    #define CHAR_V                      '\126'
1280    #define CHAR_W                      '\127'
1281    #define CHAR_X                      '\130'
1282    #define CHAR_Y                      '\131'
1283    #define CHAR_Z                      '\132'
1284    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
1285    #define CHAR_BACKSLASH              '\134'
1286    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
1287    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
1288    #define CHAR_UNDERSCORE             '\137'
1289    #define CHAR_GRAVE_ACCENT           '\140'
1290    #define CHAR_a                      '\141'
1291    #define CHAR_b                      '\142'
1292    #define CHAR_c                      '\143'
1293    #define CHAR_d                      '\144'
1294    #define CHAR_e                      '\145'
1295    #define CHAR_f                      '\146'
1296    #define CHAR_g                      '\147'
1297    #define CHAR_h                      '\150'
1298    #define CHAR_i                      '\151'
1299    #define CHAR_j                      '\152'
1300    #define CHAR_k                      '\153'
1301    #define CHAR_l                      '\154'
1302    #define CHAR_m                      '\155'
1303    #define CHAR_n                      '\156'
1304    #define CHAR_o                      '\157'
1305    #define CHAR_p                      '\160'
1306    #define CHAR_q                      '\161'
1307    #define CHAR_r                      '\162'
1308    #define CHAR_s                      '\163'
1309    #define CHAR_t                      '\164'
1310    #define CHAR_u                      '\165'
1311    #define CHAR_v                      '\166'
1312    #define CHAR_w                      '\167'
1313    #define CHAR_x                      '\170'
1314    #define CHAR_y                      '\171'
1315    #define CHAR_z                      '\172'
1316    #define CHAR_LEFT_CURLY_BRACKET     '\173'
1317    #define CHAR_VERTICAL_LINE          '\174'
1318    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
1319    #define CHAR_TILDE                  '\176'
1320    
1321    #define STR_HT                      "\011"
1322    #define STR_VT                      "\013"
1323    #define STR_FF                      "\014"
1324    #define STR_CR                      "\015"
1325    #define STR_NL                      "\012"
1326    #define STR_BS                      "\010"
1327    #define STR_BEL                     "\007"
1328    #define STR_ESC                     "\033"
1329    #define STR_DEL                     "\177"
1330    
1331    #define STR_SPACE                   "\040"
1332    #define STR_EXCLAMATION_MARK        "\041"
1333    #define STR_QUOTATION_MARK          "\042"
1334    #define STR_NUMBER_SIGN             "\043"
1335    #define STR_DOLLAR_SIGN             "\044"
1336    #define STR_PERCENT_SIGN            "\045"
1337    #define STR_AMPERSAND               "\046"
1338    #define STR_APOSTROPHE              "\047"
1339    #define STR_LEFT_PARENTHESIS        "\050"
1340    #define STR_RIGHT_PARENTHESIS       "\051"
1341    #define STR_ASTERISK                "\052"
1342    #define STR_PLUS                    "\053"
1343    #define STR_COMMA                   "\054"
1344    #define STR_MINUS                   "\055"
1345    #define STR_DOT                     "\056"
1346    #define STR_SLASH                   "\057"
1347    #define STR_0                       "\060"
1348    #define STR_1                       "\061"
1349    #define STR_2                       "\062"
1350    #define STR_3                       "\063"
1351    #define STR_4                       "\064"
1352    #define STR_5                       "\065"
1353    #define STR_6                       "\066"
1354    #define STR_7                       "\067"
1355    #define STR_8                       "\070"
1356    #define STR_9                       "\071"
1357    #define STR_COLON                   "\072"
1358    #define STR_SEMICOLON               "\073"
1359    #define STR_LESS_THAN_SIGN          "\074"
1360    #define STR_EQUALS_SIGN             "\075"
1361    #define STR_GREATER_THAN_SIGN       "\076"
1362    #define STR_QUESTION_MARK           "\077"
1363    #define STR_COMMERCIAL_AT           "\100"
1364    #define STR_A                       "\101"
1365    #define STR_B                       "\102"
1366    #define STR_C                       "\103"
1367    #define STR_D                       "\104"
1368    #define STR_E                       "\105"
1369    #define STR_F                       "\106"
1370    #define STR_G                       "\107"
1371    #define STR_H                       "\110"
1372    #define STR_I                       "\111"
1373    #define STR_J                       "\112"
1374    #define STR_K                       "\113"
1375    #define STR_L                       "\114"
1376    #define STR_M                       "\115"
1377    #define STR_N                       "\116"
1378    #define STR_O                       "\117"
1379    #define STR_P                       "\120"
1380    #define STR_Q                       "\121"
1381    #define STR_R                       "\122"
1382    #define STR_S                       "\123"
1383    #define STR_T                       "\124"
1384    #define STR_U                       "\125"
1385    #define STR_V                       "\126"
1386    #define STR_W                       "\127"
1387    #define STR_X                       "\130"
1388    #define STR_Y                       "\131"
1389    #define STR_Z                       "\132"
1390    #define STR_LEFT_SQUARE_BRACKET     "\133"
1391    #define STR_BACKSLASH               "\134"
1392    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1393    #define STR_CIRCUMFLEX_ACCENT       "\136"
1394    #define STR_UNDERSCORE              "\137"
1395    #define STR_GRAVE_ACCENT            "\140"
1396    #define STR_a                       "\141"
1397    #define STR_b                       "\142"
1398    #define STR_c                       "\143"
1399    #define STR_d                       "\144"
1400    #define STR_e                       "\145"
1401    #define STR_f                       "\146"
1402    #define STR_g                       "\147"
1403    #define STR_h                       "\150"
1404    #define STR_i                       "\151"
1405    #define STR_j                       "\152"
1406    #define STR_k                       "\153"
1407    #define STR_l                       "\154"
1408    #define STR_m                       "\155"
1409    #define STR_n                       "\156"
1410    #define STR_o                       "\157"
1411    #define STR_p                       "\160"
1412    #define STR_q                       "\161"
1413    #define STR_r                       "\162"
1414    #define STR_s                       "\163"
1415    #define STR_t                       "\164"
1416    #define STR_u                       "\165"
1417    #define STR_v                       "\166"
1418    #define STR_w                       "\167"
1419    #define STR_x                       "\170"
1420    #define STR_y                       "\171"
1421    #define STR_z                       "\172"
1422    #define STR_LEFT_CURLY_BRACKET      "\173"
1423    #define STR_VERTICAL_LINE           "\174"
1424    #define STR_RIGHT_CURLY_BRACKET     "\175"
1425    #define STR_TILDE                   "\176"
1426    
1427    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1428    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1429    #define STRING_F0                   STR_F "\0"
1430    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1431    #define STRING_MARK0                STR_M STR_A STR_R STR_K "\0"
1432    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1433    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1434    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1435    
1436    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1437    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1438    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1439    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1440    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1441    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1442    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1443    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1444    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1445    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1446    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1447    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1448    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1449    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1450    
1451    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1452    
1453    #define STRING_CR_RIGHTPAR             STR_C STR_R STR_RIGHT_PARENTHESIS
1454    #define STRING_LF_RIGHTPAR             STR_L STR_F STR_RIGHT_PARENTHESIS
1455    #define STRING_CRLF_RIGHTPAR           STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1456    #define STRING_ANY_RIGHTPAR            STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1457    #define STRING_ANYCRLF_RIGHTPAR        STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1458    #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
1459    #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
1460    #ifdef COMPILE_PCRE8
1461    #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1462    #endif
1463    #ifdef COMPILE_PCRE16
1464    #define STRING_UTF_RIGHTPAR            STR_U STR_T STR_F STR_1 STR_6 STR_RIGHT_PARENTHESIS
1465    #endif
1466    #define STRING_UCP_RIGHTPAR            STR_U STR_C STR_P STR_RIGHT_PARENTHESIS
1467    #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
1468    
1469    #endif  /* SUPPORT_UTF */
1470    
1471    /* Escape items that are just an encoding of a particular data value. */
1472    
1473  #ifndef ESC_e  #ifndef ESC_e
1474  #define ESC_e 27  #define ESC_e CHAR_ESC
1475  #endif  #endif
1476    
1477  #ifndef ESC_f  #ifndef ESC_f
1478  #define ESC_f '\f'  #define ESC_f CHAR_FF
1479  #endif  #endif
1480    
1481  #ifndef ESC_n  #ifndef ESC_n
1482  #define ESC_n NEWLINE  #define ESC_n CHAR_NL
1483  #endif  #endif
1484    
1485  #ifndef ESC_r  #ifndef ESC_r
1486  #define ESC_r '\r'  #define ESC_r CHAR_CR
1487  #endif  #endif
1488    
1489  /* 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
1490  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1491    
1492  #ifndef ESC_tee  #ifndef ESC_tee
1493  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1494  #endif  #endif
1495    
1496  /* These are escaped items that aren't just an encoding of a particular data  /* Codes for different types of Unicode property */
1497  value such as \n. They must have non-zero values, as check_escape() returns  
1498  their negation. Also, they must appear in the same order as in the opcode  #define PT_ANY        0    /* Any property - matches all chars */
1499  definitions below, up to ESC_z. There's a dummy for OP_ANY because it  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1500  corresponds to "." rather than an escape sequence. The final one must be  #define PT_GC         2    /* Specified general characteristic (e.g. L) */
1501  ESC_REF as subsequent values are used for \1, \2, \3, etc. There is are two  #define PT_PC         3    /* Specified particular characteristic (e.g. Lu) */
1502  tests in the code for an escape greater than ESC_b and less than ESC_Z to  #define PT_SC         4    /* Script (e.g. Han) */
1503  detect the types that may be repeated. These are the types that consume  #define PT_ALNUM      5    /* Alphanumeric - the union of L and N */
1504  characters. If any new escapes are put in between that don't consume a  #define PT_SPACE      6    /* Perl space - Z plus 9,10,12,13 */
1505  character, that code will have to change. */  #define PT_PXSPACE    7    /* POSIX space - Z plus 9,10,11,12,13 */
1506    #define PT_WORD       8    /* Word - L plus N plus underscore */
 enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,  
        ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_X, ESC_Z, ESC_z, ESC_E,  
        ESC_Q, ESC_REF };  
1507    
1508  /* 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
1509  contain UTF-8 characters with values greater than 255. */  contain characters with values greater than 255. */
1510    
1511  #define XCL_NOT    0x01    /* Flag: this is a negative class */  #define XCL_NOT    0x01    /* Flag: this is a negative class */
1512  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */  #define XCL_MAP    0x02    /* Flag: a 32-byte map is present */
# Line 469  contain UTF-8 characters with values gre Line 1514  contain UTF-8 characters with values gre
1514  #define XCL_END       0    /* Marks end of individual items */  #define XCL_END       0    /* Marks end of individual items */
1515  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */
1516  #define XCL_RANGE     2    /* A range (two multibyte chars) follows */  #define XCL_RANGE     2    /* A range (two multibyte chars) follows */
1517  #define XCL_PROP      3    /* Unicode property (one property code) follows */  #define XCL_PROP      3    /* Unicode property (2-byte property code follows) */
1518  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */
1519    
1520    /* These are escaped items that aren't just an encoding of a particular data
1521    value such as \n. They must have non-zero values, as check_escape() returns
1522    their negation. Also, they must appear in the same order as in the opcode
1523    definitions below, up to ESC_z. There's a dummy for OP_ALLANY because it
1524    corresponds to "." in DOTALL mode rather than an escape sequence. It is also
1525    used for [^] in JavaScript compatibility mode, and for \C in non-utf mode. In
1526    non-DOTALL mode, "." behaves like \N.
1527    
1528    The special values ESC_DU, ESC_du, etc. are used instead of ESC_D, ESC_d, etc.
1529    when PCRE_UCP is set, when replacement of \d etc by \p sequences is required.
1530    They must be contiguous, and remain in order so that the replacements can be
1531    looked up from a table.
1532    
1533    The final escape must be ESC_REF as subsequent values are used for
1534    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1535    greater than ESC_b and less than ESC_Z to detect the types that may be
1536    repeated. These are the types that consume characters. If any new escapes are
1537    put in between that don't consume a character, that code will have to change.
1538    */
1539    
1540  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
1541  that extract substrings. Starting from 1 (i.e. after OP_END), the values up to         ESC_W, ESC_w, ESC_N, ESC_dum, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1542           ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z,
1543           ESC_E, ESC_Q, ESC_g, ESC_k,
1544           ESC_DU, ESC_du, ESC_SU, ESC_su, ESC_WU, ESC_wu,
1545           ESC_REF };
1546    
1547    /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
1548  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.
1549  Note that whenever this list is updated, the two macro definitions that follow  
1550  must also be updated to match. */  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1551    that follow must also be updated to match. There are also tables called
1552    "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1553    
1554  enum {  enum {
1555    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 486  enum { Line 1558  enum {
1558    
1559    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
1560    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
1561    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
1562    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
1563    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
1564    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
1565    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
1566    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
1567    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
1568    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
1569    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
1570    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */  
1571    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ANY,            /* 12 Match any character except newline */
1572    OP_PROP,           /* 14 \p (Unicode property) */    OP_ALLANY,         /* 13 Match any character */
1573    OP_EXTUNI,         /* 15 \X (extended Unicode sequence */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1574    OP_EODN,           /* 16 End of data or \n at end of data: \Z. */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1575    OP_EOD,            /* 17 End of data: \z */    OP_PROP,           /* 16 \p (Unicode property) */
1576      OP_ANYNL,          /* 17 \R (any newline sequence) */
1577    OP_OPT,            /* 18 Set runtime options */    OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1578    OP_CIRC,           /* 19 Start of line - varies with multiline switch */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1579    OP_DOLL,           /* 20 End of line - varies with multiline switch */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1580    OP_CHAR,           /* 21 Match one character, casefully */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1581    OP_CHARNC,         /* 22 Match one character, caselessly */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1582    OP_NOT,            /* 23 Match anything but the following char */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1583      OP_EOD,            /* 24 End of data: \z */
1584    OP_STAR,           /* 24 The maximizing and minimizing versions of */  
1585    OP_MINSTAR,        /* 25 all these opcodes must come in pairs, with */    OP_CIRC,           /* 25 Start of line - not multiline */
1586    OP_PLUS,           /* 26 the minimizing one second. */    OP_CIRCM,          /* 26 Start of line - multiline */
1587    OP_MINPLUS,        /* 27 This first set applies to single characters */    OP_DOLL,           /* 27 End of line - not multiline */
1588    OP_QUERY,          /* 28 */    OP_DOLLM,          /* 28 End of line - multiline */
1589    OP_MINQUERY,       /* 29 */    OP_CHAR,           /* 29 Match one character, casefully */
1590    OP_UPTO,           /* 30 From 0 to n matches */    OP_CHARI,          /* 30 Match one character, caselessly */
1591    OP_MINUPTO,        /* 31 */    OP_NOT,            /* 31 Match one character, not the given one, casefully */
1592    OP_EXACT,          /* 32 Exactly n matches */    OP_NOTI,           /* 32 Match one character, not the given one, caselessly */
1593    
1594    OP_NOTSTAR,        /* 33 The maximizing and minimizing versions of */    /* The following sets of 13 opcodes must always be kept in step because
1595    OP_NOTMINSTAR,     /* 34 all these opcodes must come in pairs, with */    the offset from the first one is used to generate the others. */
1596    OP_NOTPLUS,        /* 35 the minimizing one second. */  
1597    OP_NOTMINPLUS,     /* 36 This set applies to "not" single characters */    /**** Single characters, caseful, must precede the caseless ones ****/
1598    OP_NOTQUERY,       /* 37 */  
1599    OP_NOTMINQUERY,    /* 38 */    OP_STAR,           /* 33 The maximizing and minimizing versions of */
1600    OP_NOTUPTO,        /* 39 From 0 to n matches */    OP_MINSTAR,        /* 34 these six opcodes must come in pairs, with */
1601    OP_NOTMINUPTO,     /* 40 */    OP_PLUS,           /* 35 the minimizing one second. */
1602    OP_NOTEXACT,       /* 41 Exactly n matches */    OP_MINPLUS,        /* 36 */
1603      OP_QUERY,          /* 37 */
1604    OP_TYPESTAR,       /* 42 The maximizing and minimizing versions of */    OP_MINQUERY,       /* 38 */
1605    OP_TYPEMINSTAR,    /* 43 all these opcodes must come in pairs, with */  
1606    OP_TYPEPLUS,       /* 44 the minimizing one second. These codes must */    OP_UPTO,           /* 39 From 0 to n matches of one character, caseful*/
1607    OP_TYPEMINPLUS,    /* 45 be in exactly the same order as those above. */    OP_MINUPTO,        /* 40 */
1608    OP_TYPEQUERY,      /* 46 This set applies to character types such as \d */    OP_EXACT,          /* 41 Exactly n matches */
1609    OP_TYPEMINQUERY,   /* 47 */  
1610    OP_TYPEUPTO,       /* 48 From 0 to n matches */    OP_POSSTAR,        /* 42 Possessified star, caseful */
1611    OP_TYPEMINUPTO,    /* 49 */    OP_POSPLUS,        /* 43 Possessified plus, caseful */
1612    OP_TYPEEXACT,      /* 50 Exactly n matches */    OP_POSQUERY,       /* 44 Posesssified query, caseful */
1613      OP_POSUPTO,        /* 45 Possessified upto, caseful */
1614    OP_CRSTAR,         /* 51 The maximizing and minimizing versions of */  
1615    OP_CRMINSTAR,      /* 52 all these opcodes must come in pairs, with */    /**** Single characters, caseless, must follow the caseful ones */
1616    OP_CRPLUS,         /* 53 the minimizing one second. These codes must */  
1617    OP_CRMINPLUS,      /* 54 be in exactly the same order as those above. */    OP_STARI,          /* 46 */
1618    OP_CRQUERY,        /* 55 These are for character classes and back refs */    OP_MINSTARI,       /* 47 */
1619    OP_CRMINQUERY,     /* 56 */    OP_PLUSI,          /* 48 */
1620    OP_CRRANGE,        /* 57 These are different to the three sets above. */    OP_MINPLUSI,       /* 49 */
1621    OP_CRMINRANGE,     /* 58 */    OP_QUERYI,         /* 50 */
1622      OP_MINQUERYI,      /* 51 */
1623    OP_CLASS,          /* 59 Match a character class, chars < 256 only */  
1624    OP_NCLASS,         /* 60 Same, but the bitmap was created from a negative    OP_UPTOI,          /* 52 From 0 to n matches of one character, caseless */
1625                             class - the difference is relevant only when a UTF-8    OP_MINUPTOI,       /* 53 */
1626                             character > 255 is encountered. */    OP_EXACTI,         /* 54 */
1627    
1628    OP_XCLASS,         /* 61 Extended class for handling UTF-8 chars within the    OP_POSSTARI,       /* 55 Possessified star, caseless */
1629                             class. This does both positive and negative. */    OP_POSPLUSI,       /* 56 Possessified plus, caseless */
1630      OP_POSQUERYI,      /* 57 Posesssified query, caseless */
1631    OP_REF,            /* 62 Match a back reference */    OP_POSUPTOI,       /* 58 Possessified upto, caseless */
1632    OP_RECURSE,        /* 63 Match a numbered subpattern (possibly recursive) */  
1633    OP_CALLOUT,        /* 64 Call out to external function if provided */    /**** The negated ones must follow the non-negated ones, and match them ****/
1634      /**** Negated single character, caseful; must precede the caseless ones ****/
1635    OP_ALT,            /* 65 Start of alternation */  
1636    OP_KET,            /* 66 End of group that doesn't have an unbounded repeat */    OP_NOTSTAR,        /* 59 The maximizing and minimizing versions of */
1637    OP_KETRMAX,        /* 67 These two must remain together and in this */    OP_NOTMINSTAR,     /* 60 these six opcodes must come in pairs, with */
1638    OP_KETRMIN,        /* 68 order. They are for groups the repeat for ever. */    OP_NOTPLUS,        /* 61 the minimizing one second. They must be in */
1639      OP_NOTMINPLUS,     /* 62 exactly the same order as those above. */
1640    /* The assertions must come before ONCE and COND */    OP_NOTQUERY,       /* 63 */
1641      OP_NOTMINQUERY,    /* 64 */
1642    OP_ASSERT,         /* 69 Positive lookahead */  
1643    OP_ASSERT_NOT,     /* 70 Negative lookahead */    OP_NOTUPTO,        /* 65 From 0 to n matches, caseful */
1644    OP_ASSERTBACK,     /* 71 Positive lookbehind */    OP_NOTMINUPTO,     /* 66 */
1645    OP_ASSERTBACK_NOT, /* 72 Negative lookbehind */    OP_NOTEXACT,       /* 67 Exactly n matches */
1646    OP_REVERSE,        /* 73 Move pointer back - used in lookbehind assertions */  
1647      OP_NOTPOSSTAR,     /* 68 Possessified versions, caseful */
1648    /* ONCE and COND must come after the assertions, with ONCE first, as there's    OP_NOTPOSPLUS,     /* 69 */
1649    a test for >= ONCE for a subpattern that isn't an assertion. */    OP_NOTPOSQUERY,    /* 70 */
1650      OP_NOTPOSUPTO,     /* 71 */
1651    OP_ONCE,           /* 74 Once matched, don't back up into the subpattern */  
1652    OP_COND,           /* 75 Conditional group */    /**** Negated single character, caseless; must follow the caseful ones ****/
1653    OP_CREF,           /* 76 Used to hold an extraction string number (cond ref) */  
1654      OP_NOTSTARI,       /* 72 */
1655    OP_BRAZERO,        /* 77 These two must remain together and in this */    OP_NOTMINSTARI,    /* 73 */
1656    OP_BRAMINZERO,     /* 78 order. */    OP_NOTPLUSI,       /* 74 */
1657      OP_NOTMINPLUSI,    /* 75 */
1658    OP_BRANUMBER,      /* 79 Used for extracting brackets whose number is greater    OP_NOTQUERYI,      /* 76 */
1659                             than can fit into an opcode. */    OP_NOTMINQUERYI,   /* 77 */
1660    
1661    OP_BRA             /* 80 This and greater values are used for brackets that    OP_NOTUPTOI,       /* 78 From 0 to n matches, caseless */
1662                             extract substrings up to EXTRACT_BASIC_MAX. After    OP_NOTMINUPTOI,    /* 79 */
1663                             that, use is made of OP_BRANUMBER. */    OP_NOTEXACTI,      /* 80 Exactly n matches */
1664  };  
1665      OP_NOTPOSSTARI,    /* 81 Possessified versions, caseless */
1666      OP_NOTPOSPLUSI,    /* 82 */
1667      OP_NOTPOSQUERYI,   /* 83 */
1668      OP_NOTPOSUPTOI,    /* 84 */
1669    
1670      /**** Character types ****/
1671    
1672      OP_TYPESTAR,       /* 85 The maximizing and minimizing versions of */
1673      OP_TYPEMINSTAR,    /* 86 these six opcodes must come in pairs, with */
1674      OP_TYPEPLUS,       /* 87 the minimizing one second. These codes must */
1675      OP_TYPEMINPLUS,    /* 88 be in exactly the same order as those above. */
1676      OP_TYPEQUERY,      /* 89 */
1677      OP_TYPEMINQUERY,   /* 90 */
1678    
1679      OP_TYPEUPTO,       /* 91 From 0 to n matches */
1680      OP_TYPEMINUPTO,    /* 92 */
1681      OP_TYPEEXACT,      /* 93 Exactly n matches */
1682    
1683      OP_TYPEPOSSTAR,    /* 94 Possessified versions */
1684      OP_TYPEPOSPLUS,    /* 95 */
1685      OP_TYPEPOSQUERY,   /* 96 */
1686      OP_TYPEPOSUPTO,    /* 97 */
1687    
1688      /* These are used for character classes and back references; only the
1689      first six are the same as the sets above. */
1690    
1691      OP_CRSTAR,         /* 98 The maximizing and minimizing versions of */
1692      OP_CRMINSTAR,      /* 99 all these opcodes must come in pairs, with */
1693      OP_CRPLUS,         /* 100 the minimizing one second. These codes must */
1694      OP_CRMINPLUS,      /* 101 be in exactly the same order as those above. */
1695      OP_CRQUERY,        /* 102 */
1696      OP_CRMINQUERY,     /* 103 */
1697    
1698      OP_CRRANGE,        /* 104 These are different to the three sets above. */
1699      OP_CRMINRANGE,     /* 105 */
1700    
1701      /* End of quantifier opcodes */
1702    
1703      OP_CLASS,          /* 106 Match a character class, chars < 256 only */
1704      OP_NCLASS,         /* 107 Same, but the bitmap was created from a negative
1705                                  class - the difference is relevant only when a
1706                                  character > 255 is encountered. */
1707      OP_XCLASS,         /* 108 Extended class for handling > 255 chars within the
1708                                  class. This does both positive and negative. */
1709      OP_REF,            /* 109 Match a back reference, casefully */
1710      OP_REFI,           /* 110 Match a back reference, caselessly */
1711      OP_RECURSE,        /* 111 Match a numbered subpattern (possibly recursive) */
1712      OP_CALLOUT,        /* 112 Call out to external function if provided */
1713    
1714      OP_ALT,            /* 113 Start of alternation */
1715      OP_KET,            /* 114 End of group that doesn't have an unbounded repeat */
1716      OP_KETRMAX,        /* 115 These two must remain together and in this */
1717      OP_KETRMIN,        /* 116 order. They are for groups the repeat for ever. */
1718      OP_KETRPOS,        /* 117 Possessive unlimited repeat. */
1719    
1720      /* The assertions must come before BRA, CBRA, ONCE, and COND, and the four
1721      asserts must remain in order. */
1722    
1723      OP_REVERSE,        /* 118 Move pointer back - used in lookbehind assertions */
1724      OP_ASSERT,         /* 119 Positive lookahead */
1725      OP_ASSERT_NOT,     /* 120 Negative lookahead */
1726      OP_ASSERTBACK,     /* 121 Positive lookbehind */
1727      OP_ASSERTBACK_NOT, /* 122 Negative lookbehind */
1728    
1729      /* ONCE, ONCE_NC, BRA, BRAPOS, CBRA, CBRAPOS, and COND must come immediately
1730      after the assertions, with ONCE first, as there's a test for >= ONCE for a
1731      subpattern that isn't an assertion. The POS versions must immediately follow
1732      the non-POS versions in each case. */
1733    
1734      OP_ONCE,           /* 123 Atomic group, contains captures */
1735      OP_ONCE_NC,        /* 124 Atomic group containing no captures */
1736      OP_BRA,            /* 125 Start of non-capturing bracket */
1737      OP_BRAPOS,         /* 126 Ditto, with unlimited, possessive repeat */
1738      OP_CBRA,           /* 127 Start of capturing bracket */
1739      OP_CBRAPOS,        /* 128 Ditto, with unlimited, possessive repeat */
1740      OP_COND,           /* 129 Conditional group */
1741    
1742      /* These five must follow the previous five, in the same order. There's a
1743      check for >= SBRA to distinguish the two sets. */
1744    
1745      OP_SBRA,           /* 130 Start of non-capturing bracket, check empty  */
1746      OP_SBRAPOS,        /* 131 Ditto, with unlimited, possessive repeat */
1747      OP_SCBRA,          /* 132 Start of capturing bracket, check empty */
1748      OP_SCBRAPOS,       /* 133 Ditto, with unlimited, possessive repeat */
1749      OP_SCOND,          /* 134 Conditional group, check empty */
1750    
1751      /* The next two pairs must (respectively) be kept together. */
1752    
1753      OP_CREF,           /* 135 Used to hold a capture number as condition */
1754      OP_NCREF,          /* 136 Same, but generated by a name reference*/
1755      OP_RREF,           /* 137 Used to hold a recursion number as condition */
1756      OP_NRREF,          /* 138 Same, but generated by a name reference*/
1757      OP_DEF,            /* 139 The DEFINE condition */
1758    
1759      OP_BRAZERO,        /* 140 These two must remain together and in this */
1760      OP_BRAMINZERO,     /* 141 order. */
1761      OP_BRAPOSZERO,     /* 142 */
1762    
1763      /* These are backtracking control verbs */
1764    
1765      OP_MARK,           /* 143 always has an argument */
1766      OP_PRUNE,          /* 144 */
1767      OP_PRUNE_ARG,      /* 145 same, but with argument */
1768      OP_SKIP,           /* 146 */
1769      OP_SKIP_ARG,       /* 147 same, but with argument */
1770      OP_THEN,           /* 148 */
1771      OP_THEN_ARG,       /* 149 same, but with argument */
1772      OP_COMMIT,         /* 150 */
1773    
1774      /* These are forced failure and success verbs */
1775    
1776      OP_FAIL,           /* 151 */
1777      OP_ACCEPT,         /* 152 */
1778      OP_ASSERT_ACCEPT,  /* 153 Used inside assertions */
1779      OP_CLOSE,          /* 154 Used before OP_ACCEPT to close open captures */
1780    
1781      /* This is used to skip a subpattern with a {0} quantifier */
1782    
1783      OP_SKIPZERO,       /* 155 */
1784    
1785      /* This is not an opcode, but is used to check that tables indexed by opcode
1786      are the correct length, in order to catch updating errors - there have been
1787      some in the past. */
1788    
1789  /* WARNING WARNING WARNING: There is an implicit assumption in pcre.c and    OP_TABLE_LENGTH
1790  study.c that all opcodes are less than 128 in value. This makes handling UTF-8  };
 character sequences easier. */  
   
 /* The highest extraction number before we have to start using additional  
 bytes. (Originally PCRE didn't have support for extraction counts highter than  
 this number.) The value is limited by the number of opcodes left after OP_BRA,  
 i.e. 255 - OP_BRA. We actually set it a bit lower to leave room for additional  
 opcodes. */  
1791    
1792  #define EXTRACT_BASIC_MAX  100  /* *** NOTE NOTE NOTE *** Whenever the list above is updated, the two macro
1793    definitions that follow must also be updated to match. There are also tables
1794    called "coptable" and "poptable" in pcre_dfa_exec.c that must be updated. */
1795    
1796    
1797  /* 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
1798  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
1799    only in pcre_printint.c, which fills out the full names in many cases (and in
1800    some cases doesn't actually use these names at all). */
1801    
1802  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1803    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1804    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1805    "notprop", "prop", "extuni",                                    \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1806    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
1807    "Opt", "^", "$", "char", "charnc", "not",                       \    "^", "^", "$", "$", "char", "chari", "not", "noti",             \
1808    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??",                                \
1809    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "{", "{", "{",                                                  \
1810      "*+","++", "?+", "{",                                           \
1811      "*", "*?", "+", "+?", "?", "??",                                \
1812      "{", "{", "{",                                                  \
1813      "*+","++", "?+", "{",                                           \
1814      "*", "*?", "+", "+?", "?", "??",                                \
1815      "{", "{", "{",                                                  \
1816      "*+","++", "?+", "{",                                           \
1817      "*", "*?", "+", "+?", "?", "??",                                \
1818      "{", "{", "{",                                                  \
1819      "*+","++", "?+", "{",                                           \
1820    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1821      "*+","++", "?+", "{",                                           \
1822    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1823    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Refi",                     \
1824    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Recurse", "Callout",                                           \
1825    "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\    "Alt", "Ket", "KetRmax", "KetRmin", "KetRpos",                  \
1826    "Brazero", "Braminzero", "Branumber", "Bra"    "Reverse", "Assert", "Assert not", "AssertB", "AssertB not",    \
1827      "Once", "Once_NC",                                              \
1828      "Bra", "BraPos", "CBra", "CBraPos",                             \
1829      "Cond",                                                         \
1830      "SBra", "SBraPos", "SCBra", "SCBraPos",                         \
1831      "SCond",                                                        \
1832      "Cond ref", "Cond nref", "Cond rec", "Cond nrec", "Cond def",   \
1833      "Brazero", "Braminzero", "Braposzero",                          \
1834      "*MARK", "*PRUNE", "*PRUNE", "*SKIP", "*SKIP",                  \
1835      "*THEN", "*THEN", "*COMMIT", "*FAIL",                           \
1836      "*ACCEPT", "*ASSERT_ACCEPT",                                    \
1837      "Close", "Skip zero"
1838    
1839    
1840  /* 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 634  in UTF-8 mode. The code that uses this t Line 1848  in UTF-8 mode. The code that uses this t
1848    
1849  #define OP_LENGTHS \  #define OP_LENGTHS \
1850    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1851    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                     */ \
1852    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1853    2, 2, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1854    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    3, 3,                          /* \P, \p                                 */ \
1855      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1856      1,                             /* \X                                     */ \
1857      1, 1, 1, 1, 1, 1,              /* \Z, \z, ^, ^M, $, $M                   */ \
1858    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1859    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Chari  - the minimum length            */ \
1860    2,                             /* not                                    */ \    2,                             /* not                                    */ \
1861    /* Positive single-char repeats                            ** These are  */ \    2,                             /* noti                                   */ \
1862    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    /* Positive single-char repeats                             ** These are */ \
1863    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??       ** minima in */ \
1864      2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto, minupto             ** mode      */ \
1865      2+IMM2_SIZE,                   /* exact                                  */ \
1866      2, 2, 2, 2+IMM2_SIZE,          /* *+, ++, ?+, upto+                      */ \
1867      2, 2, 2, 2, 2, 2,              /* *I, *?I, +I, +?I, ?I, ??I ** UTF-8     */ \
1868      2+IMM2_SIZE, 2+IMM2_SIZE,      /* upto I, minupto I                      */ \
1869      2+IMM2_SIZE,                   /* exact I                                */ \
1870      2, 2, 2, 2+IMM2_SIZE,          /* *+I, ++I, ?+I, upto+I                  */ \
1871    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1872    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1873    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto, minupto                      */ \
1874      2+IMM2_SIZE,                   /* NOT exact                              */ \
1875      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *, +, ?, upto           */ \
1876      2, 2, 2, 2, 2, 2,              /* NOT *I, *?I, +I, +?I, ?I, ??I          */ \
1877      2+IMM2_SIZE, 2+IMM2_SIZE,      /* NOT upto I, minupto I                  */ \
1878      2+IMM2_SIZE,                   /* NOT exact I                            */ \
1879      2, 2, 2, 2+IMM2_SIZE,          /* Possessive NOT *I, +I, ?I, upto I      */ \
1880    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1881    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1882    4, 4, 4,                       /* Type upto, minupto, exact              */ \    2+IMM2_SIZE, 2+IMM2_SIZE,      /* Type upto, minupto                     */ \
1883      2+IMM2_SIZE,                   /* Type exact                             */ \
1884      2, 2, 2, 2+IMM2_SIZE,          /* Possessive *+, ++, ?+, upto+           */ \
1885    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
1886    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
1887    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    1+2*IMM2_SIZE, 1+2*IMM2_SIZE,  /* CRRANGE, CRMINRANGE                    */ \
1888   33,                             /* CLASS                                  */ \    1+(32/sizeof(pcre_uchar)),     /* CLASS                                  */ \
1889   33,                             /* NCLASS                                 */ \    1+(32/sizeof(pcre_uchar)),     /* NCLASS                                 */ \
1890    0,                             /* XCLASS - variable length               */ \    0,                             /* XCLASS - variable length               */ \
1891    3,                             /* REF                                    */ \    1+IMM2_SIZE,                   /* REF                                    */ \
1892      1+IMM2_SIZE,                   /* REFI                                   */ \
1893    1+LINK_SIZE,                   /* RECURSE                                */ \    1+LINK_SIZE,                   /* RECURSE                                */ \
1894    2+2*LINK_SIZE,                 /* CALLOUT                                */ \    2+2*LINK_SIZE,                 /* CALLOUT                                */ \
1895    1+LINK_SIZE,                   /* Alt                                    */ \    1+LINK_SIZE,                   /* Alt                                    */ \
1896    1+LINK_SIZE,                   /* Ket                                    */ \    1+LINK_SIZE,                   /* Ket                                    */ \
1897    1+LINK_SIZE,                   /* KetRmax                                */ \    1+LINK_SIZE,                   /* KetRmax                                */ \
1898    1+LINK_SIZE,                   /* KetRmin                                */ \    1+LINK_SIZE,                   /* KetRmin                                */ \
1899      1+LINK_SIZE,                   /* KetRpos                                */ \
1900      1+LINK_SIZE,                   /* Reverse                                */ \
1901    1+LINK_SIZE,                   /* Assert                                 */ \    1+LINK_SIZE,                   /* Assert                                 */ \
1902    1+LINK_SIZE,                   /* Assert not                             */ \    1+LINK_SIZE,                   /* Assert not                             */ \
1903    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1904    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
1905    1+LINK_SIZE,                   /* Reverse                                */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1906    1+LINK_SIZE,                   /* Once                                   */ \    1+LINK_SIZE,                   /* ONCE_NC                                */ \
1907      1+LINK_SIZE,                   /* BRA                                    */ \
1908      1+LINK_SIZE,                   /* BRAPOS                                 */ \
1909      1+LINK_SIZE+IMM2_SIZE,         /* CBRA                                   */ \
1910      1+LINK_SIZE+IMM2_SIZE,         /* CBRAPOS                                */ \
1911    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1912    3,                             /* CREF                                   */ \    1+LINK_SIZE,                   /* SBRA                                   */ \
1913    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1+LINK_SIZE,                   /* SBRAPOS                                */ \
1914    3,                             /* BRANUMBER                              */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRA                                  */ \
1915    1+LINK_SIZE                    /* BRA                                    */ \    1+LINK_SIZE+IMM2_SIZE,         /* SCBRAPOS                               */ \
1916      1+LINK_SIZE,                   /* SCOND                                  */ \
1917      1+IMM2_SIZE, 1+IMM2_SIZE,      /* CREF, NCREF                            */ \
1918  /* A magic value for OP_CREF to indicate the "in recursion" condition. */    1+IMM2_SIZE, 1+IMM2_SIZE,      /* RREF, NRREF                            */ \
1919      1,                             /* DEF                                    */ \
1920  #define CREF_RECURSE  0xffff    1, 1, 1,                       /* BRAZERO, BRAMINZERO, BRAPOSZERO        */ \
1921      3, 1, 3,                       /* MARK, PRUNE, PRUNE_ARG                 */ \
1922  /* Error code numbers. They are given names so that they can more easily be    1, 3,                          /* SKIP, SKIP_ARG                         */ \
1923  tracked. */    1, 3,                          /* THEN, THEN_ARG                         */ \
1924      1, 1, 1, 1,                    /* COMMIT, FAIL, ACCEPT, ASSERT_ACCEPT    */ \
1925      1+IMM2_SIZE, 1                 /* CLOSE, SKIPZERO                        */
1926    
1927    /* A magic value for OP_RREF and OP_NRREF to indicate the "any recursion"
1928    condition. */
1929    
1930    #define RREF_ANY  0xffff
1931    
1932    /* Compile time error code numbers. They are given names so that they can more
1933    easily be tracked. When a new number is added, the table called eint in
1934    pcreposix.c must be updated. */
1935    
1936  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,  ERR5,  ERR6,  ERR7,  ERR8,  ERR9,
1937         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
1938         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
1939         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1940         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 };         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1941           ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1942           ERR60, ERR61, ERR62, ERR63, ERR64, ERR65, ERR66, ERR67, ERR68, ERR69,
1943           ERR70, ERR71, ERR72, ERR73, ERR74, ERRCOUNT };
1944    
1945    /* JIT compiling modes. The function list is indexed by them. */
1946    enum { JIT_COMPILE, JIT_PARTIAL_SOFT_COMPILE, JIT_PARTIAL_HARD_COMPILE,
1947           JIT_NUMBER_OF_COMPILE_MODES };
1948    
1949  /* 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
1950  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 703  Because people can now save and re-use c Line 1960  Because people can now save and re-use c
1960  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
1961  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
1962  fields are present. Currently PCRE always sets the dummy fields to zero.  fields are present. Currently PCRE always sets the dummy fields to zero.
1963  NOTE NOTE NOTE:  NOTE NOTE NOTE
1964  */  */
1965    
1966  typedef struct real_pcre {  #ifdef COMPILE_PCRE8
1967    #define REAL_PCRE real_pcre
1968    #else
1969    #define REAL_PCRE real_pcre16
1970    #endif
1971    
1972    typedef struct REAL_PCRE {
1973    pcre_uint32 magic_number;    pcre_uint32 magic_number;
1974    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1975    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
1976    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
1977      pcre_uint16 dummy1;             /* For future use */
1978    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
1979    pcre_uint16 top_backref;    pcre_uint16 top_backref;
1980    pcre_uint16 first_byte;    pcre_uint16 first_char;         /* Starting character */
1981    pcre_uint16 req_byte;    pcre_uint16 req_char;           /* This character must be seen */
1982    pcre_uint16 name_table_offset;  /* Offset to name table that follows */    pcre_uint16 name_table_offset;  /* Offset to name table that follows */
1983    pcre_uint16 name_entry_size;    /* Size of any name items */    pcre_uint16 name_entry_size;    /* Size of any name items */
1984    pcre_uint16 name_count;         /* Number of name items */    pcre_uint16 name_count;         /* Number of name items */
1985    pcre_uint16 ref_count;          /* Reference count */    pcre_uint16 ref_count;          /* Reference count */
1986    
1987    const unsigned char *tables;    /* Pointer to tables or NULL for std */    const pcre_uint8 *tables;       /* Pointer to tables or NULL for std */
1988    const unsigned char *nullpad;   /* NULL padding */    const pcre_uint8 *nullpad;      /* NULL padding */
1989  } real_pcre;  } REAL_PCRE;
1990    
1991  /* 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
1992  remark (see NOTE above) about extending this structure applies. */  remark (see NOTE above) about extending this structure applies. */
1993    
1994  typedef struct pcre_study_data {  typedef struct pcre_study_data {
1995    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1996    pcre_uint32 options;    pcre_uint32 flags;              /* Private flags */
1997    uschar start_bits[32];    pcre_uint8 start_bits[32];      /* Starting char bits */
1998      pcre_uint32 minlength;          /* Minimum subject length */
1999  } pcre_study_data;  } pcre_study_data;
2000    
2001    /* Structure for building a chain of open capturing subpatterns during
2002    compiling, so that instructions to close them can be compiled when (*ACCEPT) is
2003    encountered. This is also used to identify subpatterns that contain recursive
2004    back references to themselves, so that they can be made atomic. */
2005    
2006    typedef struct open_capitem {
2007      struct open_capitem *next;    /* Chain link */
2008      pcre_uint16 number;           /* Capture number */
2009      pcre_uint16 flag;             /* Set TRUE if recursive back ref */
2010    } open_capitem;
2011    
2012  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
2013  doing the compiling, so that they are thread-safe. */  doing the compiling, so that they are thread-safe. */
2014    
2015  typedef struct compile_data {  typedef struct compile_data {
2016    const uschar *lcc;            /* Points to lower casing table */    const pcre_uint8 *lcc;            /* Points to lower casing table */
2017    const uschar *fcc;            /* Points to case-flipping table */    const pcre_uint8 *fcc;            /* Points to case-flipping table */
2018    const uschar *cbits;          /* Points to character type table */    const pcre_uint8 *cbits;          /* Points to character type table */
2019    const uschar *ctypes;         /* Points to table of type maps */    const pcre_uint8 *ctypes;         /* Points to table of type maps */
2020    const uschar *start_code;     /* The start of the compiled code */    const pcre_uchar *start_workspace;/* The start of working space */
2021    const uschar *start_pattern;  /* The start of the pattern */    const pcre_uchar *start_code;     /* The start of the compiled code */
2022    uschar *name_table;           /* The name/number table */    const pcre_uchar *start_pattern;  /* The start of the pattern */
2023    int  names_found;             /* Number of entries so far */    const pcre_uchar *end_pattern;    /* The end of the pattern */
2024    int  name_entry_size;         /* Size of each entry */    open_capitem *open_caps;          /* Chain of open capture items */
2025    int  top_backref;             /* Maximum back reference */    pcre_uchar *hwm;                  /* High watermark of workspace */
2026    unsigned int backref_map;     /* Bitmap of low back refs */    pcre_uchar *name_table;           /* The name/number table */
2027    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  names_found;                 /* Number of entries so far */
2028    BOOL nopartial;               /* Set TRUE if partial won't work */    int  name_entry_size;             /* Size of each entry */
2029      int  workspace_size;              /* Size of workspace */
2030      int  bracount;                    /* Count of capturing parens as we compile */
2031      int  final_bracount;              /* Saved value after first pass */
2032      int  top_backref;                 /* Maximum back reference */
2033      unsigned int backref_map;         /* Bitmap of low back refs */
2034      int  assert_depth;                /* Depth of nested assertions */
2035      int  external_options;            /* External (initial) options */
2036      int  external_flags;              /* External flag bits to be set */
2037      int  req_varyopt;                 /* "After variable item" flag for reqbyte */
2038      BOOL had_accept;                  /* (*ACCEPT) encountered */
2039      BOOL check_lookbehind;            /* Lookbehinds need later checking */
2040      int  nltype;                      /* Newline type */
2041      int  nllen;                       /* Newline string length */
2042      pcre_uchar nl[4];                 /* Newline string when fixed length */
2043  } compile_data;  } compile_data;
2044    
2045  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
2046  branches, for testing for left recursion. */  branches, for testing for left recursion while compiling. */
2047    
2048  typedef struct branch_chain {  typedef struct branch_chain {
2049    struct branch_chain *outer;    struct branch_chain *outer;
2050    uschar *current;    pcre_uchar *current_branch;
2051  } branch_chain;  } branch_chain;
2052    
2053  /* Structure for items in a linked list that represents an explicit recursive  /* Structure for items in a linked list that represents an explicit recursive
2054  call within the pattern. */  call within the pattern; used by pcre_exec(). */
2055    
2056  typedef struct recursion_info {  typedef struct recursion_info {
2057    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
2058    int group_num;                /* Number of group that was called */    int group_num;                  /* Number of group that was called */
2059    const uschar *after_call;     /* "Return value": points after the call in the expr */    int *offset_save;               /* Pointer to start of saved offsets */
2060    const uschar *save_start;     /* Old value of md->start_match */    int saved_max;                  /* Number of saved offsets */
2061    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 */  
2062  } recursion_info;  } recursion_info;
2063    
2064  /* When compiling in a mode that doesn't use recursive calls to match(),  /* A similar structure for pcre_dfa_exec(). */
2065  a structure is used to remember local variables on the heap. It is defined in  
2066  pcre.c, close to the match() function, so that it is easy to keep it in step  typedef struct dfa_recursion_info {
2067  with any changes of local variable. However, the pointer to the current frame    struct dfa_recursion_info *prevrec;
2068  must be saved in some "static" place over a longjmp(). We declare the    int group_num;
2069  structure here so that we can put a pointer in the match_data structure.    PCRE_PUCHAR subject_position;
2070  NOTE: This isn't used for a "normal" compilation of pcre. */  } dfa_recursion_info;
2071    
2072    /* Structure for building a chain of data for holding the values of the subject
2073    pointer at the start of each subpattern, so as to detect when an empty string
2074    has been matched by a subpattern - to break infinite loops; used by
2075    pcre_exec(). */
2076    
2077    typedef struct eptrblock {
2078      struct eptrblock *epb_prev;
2079      PCRE_PUCHAR epb_saved_eptr;
2080    } eptrblock;
2081    
 struct heapframe;  
2082    
2083  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
2084  doing traditional NFA matching, so that they are thread-safe. */  doing traditional NFA matching, so that they are thread-safe. */
2085    
2086  typedef struct match_data {  typedef struct match_data {
2087    unsigned long int match_call_count; /* As it says */    unsigned long int match_call_count;      /* As it says */
2088    unsigned long int match_limit;/* As it says */    unsigned long int match_limit;           /* As it says */
2089    int   *offset_vector;         /* Offset vector */    unsigned long int match_limit_recursion; /* As it says */
2090    int    offset_end;            /* One past the end */    int   *offset_vector;           /* Offset vector */
2091    int    offset_max;            /* The maximum usable for return data */    int    offset_end;              /* One past the end */
2092    const uschar *lcc;            /* Points to lower casing table */    int    offset_max;              /* The maximum usable for return data */
2093    const uschar *ctypes;         /* Points to table of type maps */    int    nltype;                  /* Newline type */
2094    BOOL   offset_overflow;       /* Set if too many extractions */    int    nllen;                   /* Newline string length */
2095    BOOL   notbol;                /* NOTBOL flag */    int    name_count;              /* Number of names in name table */
2096    BOOL   noteol;                /* NOTEOL flag */    int    name_entry_size;         /* Size of entry in names table */
2097    BOOL   utf8;                  /* UTF8 flag */    pcre_uchar *name_table;         /* Table of names */
2098    BOOL   endonly;               /* Dollar not before final \n */    pcre_uchar nl[4];               /* Newline string when fixed */
2099    BOOL   notempty;              /* Empty string match not wanted */    const  pcre_uint8 *lcc;         /* Points to lower casing table */
2100    BOOL   partial;               /* PARTIAL flag */    const  pcre_uint8 *fcc;         /* Points to case-flipping table */
2101    BOOL   hitend;                /* Hit the end of the subject at some point */    const  pcre_uint8 *ctypes;      /* Points to table of type maps */
2102    const uschar *start_code;     /* For use when recursing */    BOOL   offset_overflow;         /* Set if too many extractions */
2103    const uschar *start_subject;  /* Start of the subject string */    BOOL   notbol;                  /* NOTBOL flag */
2104    const uschar *end_subject;    /* End of the subject string */    BOOL   noteol;                  /* NOTEOL flag */
2105    const uschar *start_match;    /* Start of this match attempt */    BOOL   utf;                     /* UTF-8 / UTF-16 flag */
2106    const uschar *end_match_ptr;  /* Subject position at end match */    BOOL   jscript_compat;          /* JAVASCRIPT_COMPAT flag */
2107    int    end_offset_top;        /* Highwater mark at end of match */    BOOL   use_ucp;                 /* PCRE_UCP flag */
2108    int    capture_last;          /* Most recent capture number */    BOOL   endonly;                 /* Dollar not before final \n */
2109    int    start_offset;          /* The start offset value */    BOOL   notempty;                /* Empty string match not wanted */
2110    recursion_info *recursive;    /* Linked list of recursion data */    BOOL   notempty_atstart;        /* Empty string match at start not wanted */
2111    void  *callout_data;          /* To pass back to callouts */    BOOL   hitend;                  /* Hit the end of the subject at some point */
2112    struct heapframe *thisframe;  /* Used only when compiling for no recursion */    BOOL   bsr_anycrlf;             /* \R is just any CRLF, not full Unicode */
2113      BOOL   hasthen;                 /* Pattern contains (*THEN) */
2114      BOOL   ignore_skip_arg;         /* For re-run when SKIP name not found */
2115      const  pcre_uchar *start_code;  /* For use when recursing */
2116      PCRE_PUCHAR start_subject;      /* Start of the subject string */
2117      PCRE_PUCHAR end_subject;        /* End of the subject string */
2118      PCRE_PUCHAR start_match_ptr;    /* Start of matched string */
2119      PCRE_PUCHAR end_match_ptr;      /* Subject position at end match */
2120      PCRE_PUCHAR start_used_ptr;     /* Earliest consulted character */
2121      int    partial;                 /* PARTIAL options */
2122      int    end_offset_top;          /* Highwater mark at end of match */
2123      int    capture_last;            /* Most recent capture number */
2124      int    start_offset;            /* The start offset value */
2125      int    match_function_type;     /* Set for certain special calls of MATCH() */
2126      eptrblock *eptrchain;           /* Chain of eptrblocks for tail recursions */
2127      int    eptrn;                   /* Next free eptrblock */
2128      recursion_info *recursive;      /* Linked list of recursion data */
2129      void  *callout_data;            /* To pass back to callouts */
2130      const  pcre_uchar *mark;        /* Mark pointer to pass back on success */
2131      const  pcre_uchar *nomatch_mark;/* Mark pointer to pass back on failure */
2132      const  pcre_uchar *once_target; /* Where to back up to for atomic groups */
2133  } match_data;  } match_data;
2134    
2135  /* 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
2136  functions. */  functions. */
2137    
2138  typedef struct dfa_match_data {  typedef struct dfa_match_data {
2139    const uschar *start_code;     /* Start of the compiled pattern */    const pcre_uchar *start_code;     /* Start of the compiled pattern */
2140    const uschar *start_subject;  /* Start of the subject string */    const pcre_uchar *start_subject ; /* Start of the subject string */
2141    const uschar *end_subject;    /* End of subject string */    const pcre_uchar *end_subject;    /* End of subject string */
2142    const uschar *tables;         /* Character tables */    const pcre_uchar *start_used_ptr; /* Earliest consulted character */
2143    int   moptions;               /* Match options */    const pcre_uint8 *tables;         /* Character tables */
2144    int   poptions;               /* Pattern options */    int   start_offset;               /* The start offset value */
2145    void  *callout_data;          /* To pass back to callouts */    int   moptions;                   /* Match options */
2146      int   poptions;                   /* Pattern options */
2147      int   nltype;                     /* Newline type */
2148      int   nllen;                      /* Newline string length */
2149      pcre_uchar nl[4];                 /* Newline string when fixed */
2150      void *callout_data;               /* To pass back to callouts */
2151      dfa_recursion_info *recursive;    /* Linked list of recursion data */
2152  } dfa_match_data;  } dfa_match_data;
2153    
2154  /* Bit definitions for entries in the pcre_ctypes table. */  /* Bit definitions for entries in the pcre_ctypes table. */
# Line 834  typedef struct dfa_match_data { Line 2157  typedef struct dfa_match_data {
2157  #define ctype_letter  0x02  #define ctype_letter  0x02
2158  #define ctype_digit   0x04  #define ctype_digit   0x04
2159  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
2160  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
2161  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
2162    
2163  /* 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 861  total length. */ Line 2184  total length. */
2184  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
2185  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
2186    
2187  /* Layout of the UCP type table that translates property names into codes for  /* Internal function and data prefixes. */
2188  ucp_findchar(). */  
2189    #ifdef COMPILE_PCRE8
2190    #ifndef PUBL
2191    #define PUBL(name) pcre_##name
2192    #endif
2193    #ifndef PRIV
2194    #define PRIV(name) _pcre_##name
2195    #endif
2196    #else /* COMPILE_PCRE8 */
2197    #ifdef COMPILE_PCRE16
2198    #ifndef PUBL
2199    #define PUBL(name) pcre16_##name
2200    #endif
2201    #ifndef PRIV
2202    #define PRIV(name) _pcre16_##name
2203    #endif
2204    #else
2205    #error Unsupported compiling mode
2206    #endif /* COMPILE_PCRE16 */
2207    #endif /* COMPILE_PCRE8 */
2208    
2209    /* Layout of the UCP type table that translates property names into types and
2210    codes. Each entry used to point directly to a name, but to reduce the number of
2211    relocations in shared libraries, it now has an offset into a single string
2212    instead. */
2213    
2214  typedef struct {  typedef struct {
2215    const char *name;    pcre_uint16 name_offset;
2216    int value;    pcre_uint16 type;
2217      pcre_uint16 value;
2218  } ucp_type_table;  } ucp_type_table;
2219    
2220    
# Line 875  of the exported public functions. They h Line 2223  of the exported public functions. They h
2223  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
2224  pcre_tables.c module. */  pcre_tables.c module. */
2225    
2226  extern const int    _pcre_utf8_table1[];  #ifdef COMPILE_PCRE8
2227  extern const int    _pcre_utf8_table2[];  
2228  extern const int    _pcre_utf8_table3[];  extern const int            PRIV(utf8_table1)[];
2229  extern const uschar _pcre_utf8_table4[];  extern const int            PRIV(utf8_table1_size);
2230    extern const int            PRIV(utf8_table2)[];
2231    extern const int            PRIV(utf8_table3)[];
2232    extern const pcre_uint8     PRIV(utf8_table4)[];
2233    
2234  extern const int    _pcre_utf8_table1_size;  #endif /* COMPILE_PCRE8 */
2235    
2236  extern const ucp_type_table _pcre_utt[];  extern const char           PRIV(utt_names)[];
2237  extern const int _pcre_utt_size;  extern const ucp_type_table PRIV(utt)[];
2238    extern const int            PRIV(utt_size);
2239    
2240  extern const uschar _pcre_default_tables[];  extern const pcre_uint8     PRIV(default_tables)[];
2241    
2242  extern const uschar _pcre_OP_lengths[];  extern const pcre_uint8     PRIV(OP_lengths)[];
2243    
2244    
2245  /* Internal shared functions. These are functions that are used by more than  /* Internal shared functions. These are functions that are used by more than
2246  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
2247  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
2248    
2249  extern int         _pcre_ord2utf8(int, uschar *);  /* String comparison functions. */
2250  extern void        _pcre_printint(pcre *, FILE *);  #ifdef COMPILE_PCRE8
2251  extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *,  
2252                       const pcre_study_data *, pcre_study_data *);  #define STRCMP_UC_UC(str1, str2) \
2253  extern int         _pcre_ucp_findchar(const int, int *, int *);    strcmp((char *)(str1), (char *)(str2))
2254  extern int         _pcre_valid_utf8(const uschar *, int);  #define STRCMP_UC_C8(str1, str2) \
2255  extern BOOL        _pcre_xclass(int, const uschar *);    strcmp((char *)(str1), (str2))
2256    #define STRNCMP_UC_UC(str1, str2, num) \
2257      strncmp((char *)(str1), (char *)(str2), (num))
2258    #define STRNCMP_UC_C8(str1, str2, num) \
2259      strncmp((char *)(str1), (str2), (num))
2260    #define STRLEN_UC(str) strlen((const char *)str)
2261    
2262    #else
2263    
2264    extern int               PRIV(strcmp_uc_uc)(const pcre_uchar *,
2265                               const pcre_uchar *);
2266    extern int               PRIV(strcmp_uc_c8)(const pcre_uchar *,
2267                               const char *);
2268    extern int               PRIV(strncmp_uc_uc)(const pcre_uchar *,
2269                               const pcre_uchar *, unsigned int num);
2270    extern int               PRIV(strncmp_uc_c8)(const pcre_uchar *,
2271                               const char *, unsigned int num);
2272    extern unsigned int      PRIV(strlen_uc)(const pcre_uchar *str);
2273    
2274    #define STRCMP_UC_UC(str1, str2) \
2275      PRIV(strcmp_uc_uc)((str1), (str2))
2276    #define STRCMP_UC_C8(str1, str2) \
2277      PRIV(strcmp_uc_c8)((str1), (str2))
2278    #define STRNCMP_UC_UC(str1, str2, num) \
2279      PRIV(strncmp_uc_uc)((str1), (str2), (num))
2280    #define STRNCMP_UC_C8(str1, str2, num) \
2281      PRIV(strncmp_uc_c8)((str1), (str2), (num))
2282    #define STRLEN_UC(str) PRIV(strlen_uc)(str)
2283    
2284    #endif /* COMPILE_PCRE8 */
2285    
2286    extern const pcre_uchar *PRIV(find_bracket)(const pcre_uchar *, BOOL, int);
2287    extern BOOL              PRIV(is_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2288                               int *, BOOL);
2289    extern int               PRIV(ord2utf)(pcre_uint32, pcre_uchar *);
2290    extern int               PRIV(valid_utf)(PCRE_PUCHAR, int, int *);
2291    extern BOOL              PRIV(was_newline)(PCRE_PUCHAR, int, PCRE_PUCHAR,
2292                               int *, BOOL);
2293    extern BOOL              PRIV(xclass)(int, const pcre_uchar *, BOOL);
2294    
2295    #ifdef SUPPORT_JIT
2296    extern void              PRIV(jit_compile)(const REAL_PCRE *, PUBL(extra) *, int);
2297    extern int               PRIV(jit_exec)(const REAL_PCRE *, void *,
2298                               const pcre_uchar *, int, int, int, int, int *, int,
2299                               pcre_uchar **);
2300    extern void              PRIV(jit_free)(void *);
2301    extern int               PRIV(jit_get_size)(void *);
2302    extern const char*       PRIV(jit_get_target)(void);
2303    #endif
2304    
2305    /* Unicode character database (UCD) */
2306    
2307    typedef struct {
2308      pcre_uint8 script;
2309      pcre_uint8 chartype;
2310      pcre_int32 other_case;
2311    } ucd_record;
2312    
2313    extern const ucd_record  PRIV(ucd_records)[];
2314    extern const pcre_uint8  PRIV(ucd_stage1)[];
2315    extern const pcre_uint16 PRIV(ucd_stage2)[];
2316    extern const int         PRIV(ucp_gentype)[];
2317    #ifdef SUPPORT_JIT
2318    extern const int         PRIV(ucp_typerange)[];
2319    #endif
2320    
2321    #ifdef SUPPORT_UCP
2322    /* UCD access macros */
2323    
2324    #define UCD_BLOCK_SIZE 128
2325    #define GET_UCD(ch) (PRIV(ucd_records) + \
2326            PRIV(ucd_stage2)[PRIV(ucd_stage1)[(ch) / UCD_BLOCK_SIZE] * \
2327            UCD_BLOCK_SIZE + (ch) % UCD_BLOCK_SIZE])
2328    
2329    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
2330    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
2331    #define UCD_CATEGORY(ch)  PRIV(ucp_gentype)[UCD_CHARTYPE(ch)]
2332    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
2333    
2334    #endif /* SUPPORT_UCP */
2335    
2336    #endif
2337    
2338  /* End of pcre_internal.h */  /* End of pcre_internal.h */

Legend:
Removed from v.77  
changed lines
  Added in v.929

  ViewVC Help
Powered by ViewVC 1.1.5