/[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 91 by nigel, Sat Feb 24 21:41:34 2007 UTC revision 435 by ph10, Sat Sep 5 10:20:44 2009 UTC
# Line 7  Line 7 
7  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
8    
9                         Written by Philip Hazel                         Written by Philip Hazel
10             Copyright (c) 1997-2006 University of Cambridge             Copyright (c) 1997-2009 University of Cambridge
11    
12  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
13  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 51  functions whose names all begin with "_p Line 51  functions whose names all begin with "_p
51  #define DEBUG  #define DEBUG
52  #endif  #endif
53    
54    /* We do not support both EBCDIC and UTF-8 at the same time. The "configure"
55    script prevents both being selected, but not everybody uses "configure". */
56    
57    #if defined EBCDIC && defined SUPPORT_UTF8
58    #error The use of both EBCDIC and SUPPORT_UTF8 is not supported.
59    #endif
60    
61    /* If SUPPORT_UCP is defined, SUPPORT_UTF8 must also be defined. The
62    "configure" script ensures this, but not everybody uses "configure". */
63    
64    #if defined SUPPORT_UCP && !defined SUPPORT_UTF8
65    #define SUPPORT_UTF8 1
66    #endif
67    
68  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef  /* Use a macro for debugging printing, 'cause that eliminates the use of #ifdef
69  inline, and there are *still* stupid compilers about that don't like indented  inline, and there are *still* stupid compilers about that don't like indented
70  pre-processor statements, or at least there were when I first wrote this. After  pre-processor statements, or at least there were when I first wrote this. After
71  all, it had only been about 10 years then... */  all, it had only been about 10 years then...
72    
73    It turns out that the Mac Debugging.h header also defines the macro DPRINTF, so
74    be absolutely sure we get our version. */
75    
76    #undef DPRINTF
77  #ifdef DEBUG  #ifdef DEBUG
78  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
79  #else  #else
80  #define DPRINTF(p) /*nothing*/  #define DPRINTF(p) /* Nothing */
81  #endif  #endif
82    
83    
 /* Get the definitions provided by running "configure" */  
   
 #include "config.h"  
   
84  /* Standard C headers plus the external interface definition. The only time  /* Standard C headers plus the external interface definition. The only time
85  setjmp and stdarg are used is when NO_RECURSE is set. */  setjmp and stdarg are used is when NO_RECURSE is set. */
86    
# Line 79  setjmp and stdarg are used is when NO_RE Line 93  setjmp and stdarg are used is when NO_RE
93  #include <stdlib.h>  #include <stdlib.h>
94  #include <string.h>  #include <string.h>
95    
96  #ifndef PCRE_SPY  /* When compiling a DLL for Windows, the exported symbols have to be declared
97  #define PCRE_DEFINITION       /* Win32 __declspec(export) trigger for .dll */  using some MS magic. I found some useful information on this web page:
98    http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
99    information there, using __declspec(dllexport) without "extern" we have a
100    definition; with "extern" we have a declaration. The settings here override the
101    setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
102    which is all that is needed for applications (they just import the symbols). We
103    use:
104    
105      PCRE_EXP_DECL       for declarations
106      PCRE_EXP_DEFN       for definitions of exported functions
107      PCRE_EXP_DATA_DEFN  for definitions of exported variables
108    
109    The reason for the two DEFN macros is that in non-Windows environments, one
110    does not want to have "extern" before variable definitions because it leads to
111    compiler warnings. So we distinguish between functions and variables. In
112    Windows, the two should always be the same.
113    
114    The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
115    which is an application, but needs to import this file in order to "peek" at
116    internals, can #include pcre.h first to get an application's-eye view.
117    
118    In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
119    special-purpose environments) might want to stick other stuff in front of
120    exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
121    PCRE_EXP_DATA_DEFN only if they are not already set. */
122    
123    #ifndef PCRE_EXP_DECL
124    #  ifdef _WIN32
125    #    ifndef PCRE_STATIC
126    #      define PCRE_EXP_DECL       extern __declspec(dllexport)
127    #      define PCRE_EXP_DEFN       __declspec(dllexport)
128    #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
129    #    else
130    #      define PCRE_EXP_DECL       extern
131    #      define PCRE_EXP_DEFN
132    #      define PCRE_EXP_DATA_DEFN
133    #    endif
134    #  else
135    #    ifdef __cplusplus
136    #      define PCRE_EXP_DECL       extern "C"
137    #    else
138    #      define PCRE_EXP_DECL       extern
139    #    endif
140    #    ifndef PCRE_EXP_DEFN
141    #      define PCRE_EXP_DEFN       PCRE_EXP_DECL
142    #    endif
143    #    ifndef PCRE_EXP_DATA_DEFN
144    #      define PCRE_EXP_DATA_DEFN
145    #    endif
146    #  endif
147    #endif
148    
149    /* When compiling with the MSVC compiler, it is sometimes necessary to include
150    a "calling convention" before exported function names. (This is secondhand
151    information; I know nothing about MSVC myself). For example, something like
152    
153      void __cdecl function(....)
154    
155    might be needed. In order so make this easy, all the exported functions have
156    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
157    set, we ensure here that it has no effect. */
158    
159    #ifndef PCRE_CALL_CONVENTION
160    #define PCRE_CALL_CONVENTION
161  #endif  #endif
162    
163  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We  /* We need to have types that specify unsigned 16-bit and 32-bit integers. We
# Line 91  preprocessor time in standard C environm Line 168  preprocessor time in standard C environm
168    
169  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
170    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
171      typedef short pcre_int16;
172  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
173    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
174      typedef int pcre_int16;
175  #else  #else
176    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
177  #endif  #endif
178    
179  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
180    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
181      typedef int pcre_int32;
182  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
183    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
184      typedef long int pcre_int32;
185  #else  #else
186    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
187  #endif  #endif
# Line 114  Unix, where it is defined in sys/types, Line 195  Unix, where it is defined in sys/types,
195    
196  typedef unsigned char uschar;  typedef unsigned char uschar;
197    
198  /* PCRE is able to support 3 different kinds of newline (CR, LF, CRLF). The  /* This is an unsigned int value that no character can ever have. UTF-8
199  following macro is used to package up testing for newlines. NLBLOCK is defined  characters only go up to 0x7fffffff (though Unicode doesn't go beyond
200  in the various modules to indicate in which datablock the parameters exist. */  0x0010ffff). */
201    
202    #define NOTACHAR 0xffffffff
203    
204    /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
205    "any" and "anycrlf" at present). The following macros are used to package up
206    testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
207    modules to indicate in which datablock the parameters exist, and what the
208    start/end of string field names are. */
209    
210    #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
211    #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
212    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
213    
214    /* This macro checks for a newline at the given position */
215    
216  #define IS_NEWLINE(p) \  #define IS_NEWLINE(p) \
217    ((p)[0] == NLBLOCK->nl[0] && \    ((NLBLOCK->nltype != NLTYPE_FIXED)? \
218    (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]))      ((p) < NLBLOCK->PSEND && \
219         _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
220           utf8)) \
221        : \
222        ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
223         (p)[0] == NLBLOCK->nl[0] && \
224         (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
225        ) \
226      )
227    
228    /* This macro checks for a newline immediately preceding the given position */
229    
230    #define WAS_NEWLINE(p) \
231      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
232        ((p) > NLBLOCK->PSSTART && \
233         _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
234           &(NLBLOCK->nllen), utf8)) \
235        : \
236        ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
237         (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
238         (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
239        ) \
240      )
241    
242  /* When PCRE is compiled as a C++ library, the subject pointer can be replaced  /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
243  with a custom type. This makes it possible, for example, to allow pcre_exec()  with a custom type. This makes it possible, for example, to allow pcre_exec()
# Line 139  must begin with PCRE_. */ Line 256  must begin with PCRE_. */
256  #define USPTR const unsigned char *  #define USPTR const unsigned char *
257  #endif  #endif
258    
259    
260    
261  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
262  values. */  values. */
263    
# Line 150  need to have their names changed. PCRE m Line 269  need to have their names changed. PCRE m
269  option on the command line. */  option on the command line. */
270    
271  #ifdef VPCOMPAT  #ifdef VPCOMPAT
272    #define strlen(s)        _strlen(s)
273  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
274    #define memcmp(s,c,n)    _memcmp(s,c,n)
275  #define memcpy(d,s,n)    _memcpy(d,s,n)  #define memcpy(d,s,n)    _memcpy(d,s,n)
276  #define memmove(d,s,n)   _memmove(d,s,n)  #define memmove(d,s,n)   _memmove(d,s,n)
277  #define memset(s,c,n)    _memset(s,c,n)  #define memset(s,c,n)    _memset(s,c,n)
# Line 159  option on the command line. */ Line 280  option on the command line. */
280  /* 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(),
281  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
282  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
283  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. */  
284    
285  #if ! HAVE_MEMMOVE  #ifndef HAVE_MEMMOVE
286  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
287  #if HAVE_BCOPY  #ifdef HAVE_BCOPY
288  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
289  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
290  static void *  static void *
291  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)  pcre_memmove(void *d, const void *s, size_t n)
292  {  {
293  size_t i;  size_t i;
294  dest += n;  unsigned char *dest = (unsigned char *)d;
295  src += n;  const unsigned char *src = (const unsigned char *)s;
296  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  if (dest > src)
297  return dest;    {
298      dest += n;
299      src += n;
300      for (i = 0; i < n; ++i) *(--dest) = *(--src);
301      return (void *)dest;
302      }
303    else
304      {
305      for (i = 0; i < n; ++i) *dest++ = *src++;
306      return (void *)(dest - n);
307      }
308  }  }
309  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
310  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 261  capturing parenthesis numbers in back re Line 390  capturing parenthesis numbers in back re
390    
391  /* When UTF-8 encoding is being used, a character is no longer just a single  /* When UTF-8 encoding is being used, a character is no longer just a single
392  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
393  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
394    never be called in byte mode. To make sure it can never even appear when UTF-8
395    support is omitted, we don't even define it. */
396    
397  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
398  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 269  byte-mode, and more complicated ones for Line 400  byte-mode, and more complicated ones for
400  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
401  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
402  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
403  #define BACKCHAR(eptr)  /* #define BACKCHAR(eptr) */
404    
405  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
406    
# Line 278  we know we are in UTF-8 mode. */ Line 409  we know we are in UTF-8 mode. */
409    
410  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
411    c = *eptr; \    c = *eptr; \
412    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
413      { \      { \
414      int gcii; \      int gcii; \
415      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 296  pointer. */ Line 427  pointer. */
427    
428  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
429    c = *eptr; \    c = *eptr; \
430    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
431      { \      { \
432      int gcii; \      int gcii; \
433      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 314  know we are in UTF-8 mode. */ Line 445  know we are in UTF-8 mode. */
445    
446  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
447    c = *eptr++; \    c = *eptr++; \
448    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
449      { \      { \
450      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
451      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 330  know we are in UTF-8 mode. */ Line 461  know we are in UTF-8 mode. */
461    
462  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
463    c = *eptr++; \    c = *eptr++; \
464    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
465      { \      { \
466      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
467      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 347  if there are extra bytes. This is called Line 478  if there are extra bytes. This is called
478    
479  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
480    c = *eptr; \    c = *eptr; \
481    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
482        { \
483        int gcii; \
484        int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
485        int gcss = 6*gcaa; \
486        c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
487        for (gcii = 1; gcii <= gcaa; gcii++) \
488          { \
489          gcss -= 6; \
490          c |= (eptr[gcii] & 0x3f) << gcss; \
491          } \
492        len += gcaa; \
493        }
494    
495    /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
496    pointer, incrementing length if there are extra bytes. This is called when we
497    know we are in UTF-8 mode. */
498    
499    #define GETCHARLENTEST(c, eptr, len) \
500      c = *eptr; \
501      if (utf8 && c >= 0xc0) \
502      { \      { \
503      int gcii; \      int gcii; \
504      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 362  if there are extra bytes. This is called Line 513  if there are extra bytes. This is called
513      }      }
514    
515  /* If the pointer is not at the start of a character, move it back until  /* If the pointer is not at the start of a character, move it back until
516  it is. Called only in UTF-8 mode. */  it is. This is called only in UTF-8 mode - we don't put a test within the macro
517    because almost all calls are already within a block of UTF-8 only code. */
518    
519  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
520    
521  #endif  #endif
522    
# Line 381  Standard C system should have one. */ Line 533  Standard C system should have one. */
533    
534  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
535    
536  /* Private options flags start at the most significant end of the four bytes.  /* Private flags containing information about the compiled regex. They used to
537  The public options defined in pcre.h start at the least significant end. Make  live at the top end of the options word, but that got almost full, so now they
538  sure they don't overlap! The bits are getting a bit scarce now -- when we run  are in a 16-bit flags word. From release 8.00, PCRE_NOPARTIAL is unused, as
539  out, there is a dummy word in the structure that could be used for the private  the restrictions on partial matching have been lifted. It remains for backwards
540  bits. */  compatibility. */
541    
542  #define PCRE_NOPARTIAL     0x80000000  /* can't use partial with this regex */  #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
543  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
544  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
545  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
546  #define PCRE_JCHANGED      0x08000000  /* j option changes within regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
547    #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
548    
549  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
550    
# Line 400  bits. */ Line 553  bits. */
553  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
554  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
555    
556  #define PUBLIC_OPTIONS \  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
557                               PCRE_NEWLINE_ANYCRLF)
558    
559    #define PUBLIC_COMPILE_OPTIONS \
560    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
561     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
562     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \     PCRE_NO_AUTO_CAPTURE|PCRE_NO_UTF8_CHECK|PCRE_AUTO_CALLOUT|PCRE_FIRSTLINE| \
563     PCRE_DUPNAMES|PCRE_NEWLINE_CR|PCRE_NEWLINE_LF)     PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
564       PCRE_JAVASCRIPT_COMPAT)
565    
566  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
567    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
568     PCRE_PARTIAL|PCRE_NEWLINE_CR|PCRE_NEWLINE_LF)     PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF| \
569       PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
570    
571  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
572    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
573     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_CR| \     PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART| \
574     PCRE_NEWLINE_LF)     PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
575    
576  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
577    
# Line 438  variable-length repeat, or a anything ot Line 596  variable-length repeat, or a anything ot
596  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
597  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
598    
599  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
600    environments where these macros are defined elsewhere. Unfortunately, there
601    is no way to do the same for the typedef. */
602    
603  typedef int BOOL;  typedef int BOOL;
604    
605    #ifndef FALSE
606  #define FALSE   0  #define FALSE   0
607  #define TRUE    1  #define TRUE    1
608    #endif
609    
610    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
611    character constants like '*' because the compiler would emit their EBCDIC code,
612    which is different from their ASCII/UTF-8 code. Instead we define macros for
613    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
614    is enabled. When UTF-8 support is not enabled, the definitions use character
615    literals. Both character and string versions of each character are needed, and
616    there are some longer strings as well.
617    
618    This means that, on EBCDIC platforms, the PCRE library can handle either
619    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
620    would need different lookups depending on whether PCRE_UTF8 was set or not.
621    This would make it impossible to use characters in switch/case statements,
622    which would reduce performance. For a theoretical use (which nobody has asked
623    for) in a minority area (EBCDIC platforms), this is not sensible. Any
624    application that did need both could compile two versions of the library, using
625    macros to give the functions distinct names. */
626    
627    #ifndef SUPPORT_UTF8
628    
629  /* Escape items that are just an encoding of a particular data value. Note that  /* UTF-8 support is not enabled; use the platform-dependent character literals
630  ESC_n is defined as yet another macro, which is set in config.h to either \n  so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
631  (the default) or \r (which some people want). */  
632    #define CHAR_HT                     '\t'
633    #define CHAR_VT                     '\v'
634    #define CHAR_FF                     '\f'
635    #define CHAR_CR                     '\r'
636    #define CHAR_NL                     '\n'
637    #define CHAR_BS                     '\b'
638    #define CHAR_BEL                    '\a'
639    #ifdef EBCDIC
640    #define CHAR_ESC                    '\047'
641    #define CHAR_DEL                    '\007'
642    #else
643    #define CHAR_ESC                    '\033'
644    #define CHAR_DEL                    '\177'
645    #endif
646    
647    #define CHAR_SPACE                  ' '
648    #define CHAR_EXCLAMATION_MARK       '!'
649    #define CHAR_QUOTATION_MARK         '"'
650    #define CHAR_NUMBER_SIGN            '#'
651    #define CHAR_DOLLAR_SIGN            '$'
652    #define CHAR_PERCENT_SIGN           '%'
653    #define CHAR_AMPERSAND              '&'
654    #define CHAR_APOSTROPHE             '\''
655    #define CHAR_LEFT_PARENTHESIS       '('
656    #define CHAR_RIGHT_PARENTHESIS      ')'
657    #define CHAR_ASTERISK               '*'
658    #define CHAR_PLUS                   '+'
659    #define CHAR_COMMA                  ','
660    #define CHAR_MINUS                  '-'
661    #define CHAR_DOT                    '.'
662    #define CHAR_SLASH                  '/'
663    #define CHAR_0                      '0'
664    #define CHAR_1                      '1'
665    #define CHAR_2                      '2'
666    #define CHAR_3                      '3'
667    #define CHAR_4                      '4'
668    #define CHAR_5                      '5'
669    #define CHAR_6                      '6'
670    #define CHAR_7                      '7'
671    #define CHAR_8                      '8'
672    #define CHAR_9                      '9'
673    #define CHAR_COLON                  ':'
674    #define CHAR_SEMICOLON              ';'
675    #define CHAR_LESS_THAN_SIGN         '<'
676    #define CHAR_EQUALS_SIGN            '='
677    #define CHAR_GREATER_THAN_SIGN      '>'
678    #define CHAR_QUESTION_MARK          '?'
679    #define CHAR_COMMERCIAL_AT          '@'
680    #define CHAR_A                      'A'
681    #define CHAR_B                      'B'
682    #define CHAR_C                      'C'
683    #define CHAR_D                      'D'
684    #define CHAR_E                      'E'
685    #define CHAR_F                      'F'
686    #define CHAR_G                      'G'
687    #define CHAR_H                      'H'
688    #define CHAR_I                      'I'
689    #define CHAR_J                      'J'
690    #define CHAR_K                      'K'
691    #define CHAR_L                      'L'
692    #define CHAR_M                      'M'
693    #define CHAR_N                      'N'
694    #define CHAR_O                      'O'
695    #define CHAR_P                      'P'
696    #define CHAR_Q                      'Q'
697    #define CHAR_R                      'R'
698    #define CHAR_S                      'S'
699    #define CHAR_T                      'T'
700    #define CHAR_U                      'U'
701    #define CHAR_V                      'V'
702    #define CHAR_W                      'W'
703    #define CHAR_X                      'X'
704    #define CHAR_Y                      'Y'
705    #define CHAR_Z                      'Z'
706    #define CHAR_LEFT_SQUARE_BRACKET    '['
707    #define CHAR_BACKSLASH              '\\'
708    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
709    #define CHAR_CIRCUMFLEX_ACCENT      '^'
710    #define CHAR_UNDERSCORE             '_'
711    #define CHAR_GRAVE_ACCENT           '`'
712    #define CHAR_a                      'a'
713    #define CHAR_b                      'b'
714    #define CHAR_c                      'c'
715    #define CHAR_d                      'd'
716    #define CHAR_e                      'e'
717    #define CHAR_f                      'f'
718    #define CHAR_g                      'g'
719    #define CHAR_h                      'h'
720    #define CHAR_i                      'i'
721    #define CHAR_j                      'j'
722    #define CHAR_k                      'k'
723    #define CHAR_l                      'l'
724    #define CHAR_m                      'm'
725    #define CHAR_n                      'n'
726    #define CHAR_o                      'o'
727    #define CHAR_p                      'p'
728    #define CHAR_q                      'q'
729    #define CHAR_r                      'r'
730    #define CHAR_s                      's'
731    #define CHAR_t                      't'
732    #define CHAR_u                      'u'
733    #define CHAR_v                      'v'
734    #define CHAR_w                      'w'
735    #define CHAR_x                      'x'
736    #define CHAR_y                      'y'
737    #define CHAR_z                      'z'
738    #define CHAR_LEFT_CURLY_BRACKET     '{'
739    #define CHAR_VERTICAL_LINE          '|'
740    #define CHAR_RIGHT_CURLY_BRACKET    '}'
741    #define CHAR_TILDE                  '~'
742    
743    #define STR_HT                      "\t"
744    #define STR_VT                      "\v"
745    #define STR_FF                      "\f"
746    #define STR_CR                      "\r"
747    #define STR_NL                      "\n"
748    #define STR_BS                      "\b"
749    #define STR_BEL                     "\a"
750    #ifdef EBCDIC
751    #define STR_ESC                     "\047"
752    #define STR_DEL                     "\007"
753    #else
754    #define STR_ESC                     "\033"
755    #define STR_DEL                     "\177"
756    #endif
757    
758    #define STR_SPACE                   " "
759    #define STR_EXCLAMATION_MARK        "!"
760    #define STR_QUOTATION_MARK          "\""
761    #define STR_NUMBER_SIGN             "#"
762    #define STR_DOLLAR_SIGN             "$"
763    #define STR_PERCENT_SIGN            "%"
764    #define STR_AMPERSAND               "&"
765    #define STR_APOSTROPHE              "'"
766    #define STR_LEFT_PARENTHESIS        "("
767    #define STR_RIGHT_PARENTHESIS       ")"
768    #define STR_ASTERISK                "*"
769    #define STR_PLUS                    "+"
770    #define STR_COMMA                   ","
771    #define STR_MINUS                   "-"
772    #define STR_DOT                     "."
773    #define STR_SLASH                   "/"
774    #define STR_0                       "0"
775    #define STR_1                       "1"
776    #define STR_2                       "2"
777    #define STR_3                       "3"
778    #define STR_4                       "4"
779    #define STR_5                       "5"
780    #define STR_6                       "6"
781    #define STR_7                       "7"
782    #define STR_8                       "8"
783    #define STR_9                       "9"
784    #define STR_COLON                   ":"
785    #define STR_SEMICOLON               ";"
786    #define STR_LESS_THAN_SIGN          "<"
787    #define STR_EQUALS_SIGN             "="
788    #define STR_GREATER_THAN_SIGN       ">"
789    #define STR_QUESTION_MARK           "?"
790    #define STR_COMMERCIAL_AT           "@"
791    #define STR_A                       "A"
792    #define STR_B                       "B"
793    #define STR_C                       "C"
794    #define STR_D                       "D"
795    #define STR_E                       "E"
796    #define STR_F                       "F"
797    #define STR_G                       "G"
798    #define STR_H                       "H"
799    #define STR_I                       "I"
800    #define STR_J                       "J"
801    #define STR_K                       "K"
802    #define STR_L                       "L"
803    #define STR_M                       "M"
804    #define STR_N                       "N"
805    #define STR_O                       "O"
806    #define STR_P                       "P"
807    #define STR_Q                       "Q"
808    #define STR_R                       "R"
809    #define STR_S                       "S"
810    #define STR_T                       "T"
811    #define STR_U                       "U"
812    #define STR_V                       "V"
813    #define STR_W                       "W"
814    #define STR_X                       "X"
815    #define STR_Y                       "Y"
816    #define STR_Z                       "Z"
817    #define STR_LEFT_SQUARE_BRACKET     "["
818    #define STR_BACKSLASH               "\\"
819    #define STR_RIGHT_SQUARE_BRACKET    "]"
820    #define STR_CIRCUMFLEX_ACCENT       "^"
821    #define STR_UNDERSCORE              "_"
822    #define STR_GRAVE_ACCENT            "`"
823    #define STR_a                       "a"
824    #define STR_b                       "b"
825    #define STR_c                       "c"
826    #define STR_d                       "d"
827    #define STR_e                       "e"
828    #define STR_f                       "f"
829    #define STR_g                       "g"
830    #define STR_h                       "h"
831    #define STR_i                       "i"
832    #define STR_j                       "j"
833    #define STR_k                       "k"
834    #define STR_l                       "l"
835    #define STR_m                       "m"
836    #define STR_n                       "n"
837    #define STR_o                       "o"
838    #define STR_p                       "p"
839    #define STR_q                       "q"
840    #define STR_r                       "r"
841    #define STR_s                       "s"
842    #define STR_t                       "t"
843    #define STR_u                       "u"
844    #define STR_v                       "v"
845    #define STR_w                       "w"
846    #define STR_x                       "x"
847    #define STR_y                       "y"
848    #define STR_z                       "z"
849    #define STR_LEFT_CURLY_BRACKET      "{"
850    #define STR_VERTICAL_LINE           "|"
851    #define STR_RIGHT_CURLY_BRACKET     "}"
852    #define STR_TILDE                   "~"
853    
854    #define STRING_ACCEPT0              "ACCEPT\0"
855    #define STRING_COMMIT0              "COMMIT\0"
856    #define STRING_F0                   "F\0"
857    #define STRING_FAIL0                "FAIL\0"
858    #define STRING_PRUNE0               "PRUNE\0"
859    #define STRING_SKIP0                "SKIP\0"
860    #define STRING_THEN                 "THEN"
861    
862    #define STRING_alpha0               "alpha\0"
863    #define STRING_lower0               "lower\0"
864    #define STRING_upper0               "upper\0"
865    #define STRING_alnum0               "alnum\0"
866    #define STRING_ascii0               "ascii\0"
867    #define STRING_blank0               "blank\0"
868    #define STRING_cntrl0               "cntrl\0"
869    #define STRING_digit0               "digit\0"
870    #define STRING_graph0               "graph\0"
871    #define STRING_print0               "print\0"
872    #define STRING_punct0               "punct\0"
873    #define STRING_space0               "space\0"
874    #define STRING_word0                "word\0"
875    #define STRING_xdigit               "xdigit"
876    
877    #define STRING_DEFINE               "DEFINE"
878    
879    #define STRING_CR_RIGHTPAR          "CR)"
880    #define STRING_LF_RIGHTPAR          "LF)"
881    #define STRING_CRLF_RIGHTPAR        "CRLF)"
882    #define STRING_ANY_RIGHTPAR         "ANY)"
883    #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"
884    #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
885    #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
886    #define STRING_UTF8_RIGHTPAR        "UTF8)"
887    
888    #else  /* SUPPORT_UTF8 */
889    
890    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
891    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
892    only. */
893    
894    #define CHAR_HT                     '\011'
895    #define CHAR_VT                     '\013'
896    #define CHAR_FF                     '\014'
897    #define CHAR_CR                     '\015'
898    #define CHAR_NL                     '\012'
899    #define CHAR_BS                     '\010'
900    #define CHAR_BEL                    '\007'
901    #define CHAR_ESC                    '\033'
902    #define CHAR_DEL                    '\177'
903    
904    #define CHAR_SPACE                  '\040'
905    #define CHAR_EXCLAMATION_MARK       '\041'
906    #define CHAR_QUOTATION_MARK         '\042'
907    #define CHAR_NUMBER_SIGN            '\043'
908    #define CHAR_DOLLAR_SIGN            '\044'
909    #define CHAR_PERCENT_SIGN           '\045'
910    #define CHAR_AMPERSAND              '\046'
911    #define CHAR_APOSTROPHE             '\047'
912    #define CHAR_LEFT_PARENTHESIS       '\050'
913    #define CHAR_RIGHT_PARENTHESIS      '\051'
914    #define CHAR_ASTERISK               '\052'
915    #define CHAR_PLUS                   '\053'
916    #define CHAR_COMMA                  '\054'
917    #define CHAR_MINUS                  '\055'
918    #define CHAR_DOT                    '\056'
919    #define CHAR_SLASH                  '\057'
920    #define CHAR_0                      '\060'
921    #define CHAR_1                      '\061'
922    #define CHAR_2                      '\062'
923    #define CHAR_3                      '\063'
924    #define CHAR_4                      '\064'
925    #define CHAR_5                      '\065'
926    #define CHAR_6                      '\066'
927    #define CHAR_7                      '\067'
928    #define CHAR_8                      '\070'
929    #define CHAR_9                      '\071'
930    #define CHAR_COLON                  '\072'
931    #define CHAR_SEMICOLON              '\073'
932    #define CHAR_LESS_THAN_SIGN         '\074'
933    #define CHAR_EQUALS_SIGN            '\075'
934    #define CHAR_GREATER_THAN_SIGN      '\076'
935    #define CHAR_QUESTION_MARK          '\077'
936    #define CHAR_COMMERCIAL_AT          '\100'
937    #define CHAR_A                      '\101'
938    #define CHAR_B                      '\102'
939    #define CHAR_C                      '\103'
940    #define CHAR_D                      '\104'
941    #define CHAR_E                      '\105'
942    #define CHAR_F                      '\106'
943    #define CHAR_G                      '\107'
944    #define CHAR_H                      '\110'
945    #define CHAR_I                      '\111'
946    #define CHAR_J                      '\112'
947    #define CHAR_K                      '\113'
948    #define CHAR_L                      '\114'
949    #define CHAR_M                      '\115'
950    #define CHAR_N                      '\116'
951    #define CHAR_O                      '\117'
952    #define CHAR_P                      '\120'
953    #define CHAR_Q                      '\121'
954    #define CHAR_R                      '\122'
955    #define CHAR_S                      '\123'
956    #define CHAR_T                      '\124'
957    #define CHAR_U                      '\125'
958    #define CHAR_V                      '\126'
959    #define CHAR_W                      '\127'
960    #define CHAR_X                      '\130'
961    #define CHAR_Y                      '\131'
962    #define CHAR_Z                      '\132'
963    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
964    #define CHAR_BACKSLASH              '\134'
965    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
966    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
967    #define CHAR_UNDERSCORE             '\137'
968    #define CHAR_GRAVE_ACCENT           '\140'
969    #define CHAR_a                      '\141'
970    #define CHAR_b                      '\142'
971    #define CHAR_c                      '\143'
972    #define CHAR_d                      '\144'
973    #define CHAR_e                      '\145'
974    #define CHAR_f                      '\146'
975    #define CHAR_g                      '\147'
976    #define CHAR_h                      '\150'
977    #define CHAR_i                      '\151'
978    #define CHAR_j                      '\152'
979    #define CHAR_k                      '\153'
980    #define CHAR_l                      '\154'
981    #define CHAR_m                      '\155'
982    #define CHAR_n                      '\156'
983    #define CHAR_o                      '\157'
984    #define CHAR_p                      '\160'
985    #define CHAR_q                      '\161'
986    #define CHAR_r                      '\162'
987    #define CHAR_s                      '\163'
988    #define CHAR_t                      '\164'
989    #define CHAR_u                      '\165'
990    #define CHAR_v                      '\166'
991    #define CHAR_w                      '\167'
992    #define CHAR_x                      '\170'
993    #define CHAR_y                      '\171'
994    #define CHAR_z                      '\172'
995    #define CHAR_LEFT_CURLY_BRACKET     '\173'
996    #define CHAR_VERTICAL_LINE          '\174'
997    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
998    #define CHAR_TILDE                  '\176'
999    
1000    #define STR_HT                      "\011"
1001    #define STR_VT                      "\013"
1002    #define STR_FF                      "\014"
1003    #define STR_CR                      "\015"
1004    #define STR_NL                      "\012"
1005    #define STR_BS                      "\010"
1006    #define STR_BEL                     "\007"
1007    #define STR_ESC                     "\033"
1008    #define STR_DEL                     "\177"
1009    
1010    #define STR_SPACE                   "\040"
1011    #define STR_EXCLAMATION_MARK        "\041"
1012    #define STR_QUOTATION_MARK          "\042"
1013    #define STR_NUMBER_SIGN             "\043"
1014    #define STR_DOLLAR_SIGN             "\044"
1015    #define STR_PERCENT_SIGN            "\045"
1016    #define STR_AMPERSAND               "\046"
1017    #define STR_APOSTROPHE              "\047"
1018    #define STR_LEFT_PARENTHESIS        "\050"
1019    #define STR_RIGHT_PARENTHESIS       "\051"
1020    #define STR_ASTERISK                "\052"
1021    #define STR_PLUS                    "\053"
1022    #define STR_COMMA                   "\054"
1023    #define STR_MINUS                   "\055"
1024    #define STR_DOT                     "\056"
1025    #define STR_SLASH                   "\057"
1026    #define STR_0                       "\060"
1027    #define STR_1                       "\061"
1028    #define STR_2                       "\062"
1029    #define STR_3                       "\063"
1030    #define STR_4                       "\064"
1031    #define STR_5                       "\065"
1032    #define STR_6                       "\066"
1033    #define STR_7                       "\067"
1034    #define STR_8                       "\070"
1035    #define STR_9                       "\071"
1036    #define STR_COLON                   "\072"
1037    #define STR_SEMICOLON               "\073"
1038    #define STR_LESS_THAN_SIGN          "\074"
1039    #define STR_EQUALS_SIGN             "\075"
1040    #define STR_GREATER_THAN_SIGN       "\076"
1041    #define STR_QUESTION_MARK           "\077"
1042    #define STR_COMMERCIAL_AT           "\100"
1043    #define STR_A                       "\101"
1044    #define STR_B                       "\102"
1045    #define STR_C                       "\103"
1046    #define STR_D                       "\104"
1047    #define STR_E                       "\105"
1048    #define STR_F                       "\106"
1049    #define STR_G                       "\107"
1050    #define STR_H                       "\110"
1051    #define STR_I                       "\111"
1052    #define STR_J                       "\112"
1053    #define STR_K                       "\113"
1054    #define STR_L                       "\114"
1055    #define STR_M                       "\115"
1056    #define STR_N                       "\116"
1057    #define STR_O                       "\117"
1058    #define STR_P                       "\120"
1059    #define STR_Q                       "\121"
1060    #define STR_R                       "\122"
1061    #define STR_S                       "\123"
1062    #define STR_T                       "\124"
1063    #define STR_U                       "\125"
1064    #define STR_V                       "\126"
1065    #define STR_W                       "\127"
1066    #define STR_X                       "\130"
1067    #define STR_Y                       "\131"
1068    #define STR_Z                       "\132"
1069    #define STR_LEFT_SQUARE_BRACKET     "\133"
1070    #define STR_BACKSLASH               "\134"
1071    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1072    #define STR_CIRCUMFLEX_ACCENT       "\136"
1073    #define STR_UNDERSCORE              "\137"
1074    #define STR_GRAVE_ACCENT            "\140"
1075    #define STR_a                       "\141"
1076    #define STR_b                       "\142"
1077    #define STR_c                       "\143"
1078    #define STR_d                       "\144"
1079    #define STR_e                       "\145"
1080    #define STR_f                       "\146"
1081    #define STR_g                       "\147"
1082    #define STR_h                       "\150"
1083    #define STR_i                       "\151"
1084    #define STR_j                       "\152"
1085    #define STR_k                       "\153"
1086    #define STR_l                       "\154"
1087    #define STR_m                       "\155"
1088    #define STR_n                       "\156"
1089    #define STR_o                       "\157"
1090    #define STR_p                       "\160"
1091    #define STR_q                       "\161"
1092    #define STR_r                       "\162"
1093    #define STR_s                       "\163"
1094    #define STR_t                       "\164"
1095    #define STR_u                       "\165"
1096    #define STR_v                       "\166"
1097    #define STR_w                       "\167"
1098    #define STR_x                       "\170"
1099    #define STR_y                       "\171"
1100    #define STR_z                       "\172"
1101    #define STR_LEFT_CURLY_BRACKET      "\173"
1102    #define STR_VERTICAL_LINE           "\174"
1103    #define STR_RIGHT_CURLY_BRACKET     "\175"
1104    #define STR_TILDE                   "\176"
1105    
1106    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1107    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1108    #define STRING_F0                   STR_F "\0"
1109    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1110    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1111    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1112    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1113    
1114    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1115    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1116    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1117    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1118    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1119    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1120    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1121    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1122    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1123    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1124    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1125    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1126    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1127    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1128    
1129    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1130    
1131    #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS
1132    #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS
1133    #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1134    #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1135    #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1136    #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
1137    #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
1138    #define STRING_UTF8_RIGHTPAR        STR_U STR_T STR_F STR_8 STR_RIGHT_PARENTHESIS
1139    
1140    #endif  /* SUPPORT_UTF8 */
1141    
1142    /* Escape items that are just an encoding of a particular data value. */
1143    
1144  #ifndef ESC_e  #ifndef ESC_e
1145  #define ESC_e 27  #define ESC_e CHAR_ESC
1146  #endif  #endif
1147    
1148  #ifndef ESC_f  #ifndef ESC_f
1149  #define ESC_f '\f'  #define ESC_f CHAR_FF
1150  #endif  #endif
1151    
1152  #ifndef ESC_n  #ifndef ESC_n
1153  #define ESC_n NEWLINE  #define ESC_n CHAR_NL
1154  #endif  #endif
1155    
1156  #ifndef ESC_r  #ifndef ESC_r
1157  #define ESC_r '\r'  #define ESC_r CHAR_CR
1158  #endif  #endif
1159    
1160  /* 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
1161  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1162    
1163  #ifndef ESC_tee  #ifndef ESC_tee
1164  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1165  #endif  #endif
1166    
1167  /* Codes for different types of Unicode property */  /* Codes for different types of Unicode property */
# Line 496  contain UTF-8 characters with values gre Line 1188  contain UTF-8 characters with values gre
1188  value such as \n. They must have non-zero values, as check_escape() returns  value such as \n. They must have non-zero values, as check_escape() returns
1189  their negation. Also, they must appear in the same order as in the opcode  their negation. Also, they must appear in the same order as in the opcode
1190  definitions below, up to ESC_z. There's a dummy for OP_ANY because it  definitions below, up to ESC_z. There's a dummy for OP_ANY because it
1191  corresponds to "." rather than an escape sequence. The final one must be  corresponds to "." rather than an escape sequence, and another for OP_ALLANY
1192  ESC_REF as subsequent values are used for \1, \2, \3, etc. There is are two  (which is used for [^] in JavaScript compatibility mode).
 tests in the code for an escape greater than ESC_b and less than ESC_Z to  
 detect the types that may be repeated. These are the types that consume  
 characters. If any new escapes are put in between that don't consume a  
 character, that code will have to change. */  
   
 enum { ESC_A = 1, ESC_G, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W,  
        ESC_w, ESC_dum1, ESC_C, ESC_P, ESC_p, ESC_X, ESC_Z, ESC_z, ESC_E,  
        ESC_Q, ESC_REF };  
1193    
1194  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  The final escape must be ESC_REF as subsequent values are used for
1195  that extract substrings. Starting from 1 (i.e. after OP_END), the values up to  backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1196    greater than ESC_b and less than ESC_Z to detect the types that may be
1197    repeated. These are the types that consume characters. If any new escapes are
1198    put in between that don't consume a character, that code will have to change.
1199    */
1200    
1201    enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
1202           ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1203           ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
1204           ESC_REF };
1205    
1206    
1207    /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
1208  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.
1209  Note that whenever this list is updated, the two macro definitions that follow  
1210  must also be updated to match. */  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1211    that follow must also be updated to match. There is also a table called
1212    "coptable" in pcre_dfa_exec.c that must be updated. */
1213    
1214  enum {  enum {
1215    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 520  enum { Line 1218  enum {
1218    
1219    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
1220    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
1221    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
1222    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
1223    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
1224    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
1225    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
1226    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
1227    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
1228    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
1229    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
1230    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
1231    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
1232    OP_PROP,           /* 14 \p (Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1233    OP_EXTUNI,         /* 15 \X (extended Unicode sequence */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1234    OP_EODN,           /* 16 End of data or \n at end of data: \Z. */    OP_PROP,           /* 16 \p (Unicode property) */
1235    OP_EOD,            /* 17 End of data: \z */    OP_ANYNL,          /* 17 \R (any newline sequence) */
1236      OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1237    OP_OPT,            /* 18 Set runtime options */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1238    OP_CIRC,           /* 19 Start of line - varies with multiline switch */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1239    OP_DOLL,           /* 20 End of line - varies with multiline switch */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1240    OP_CHAR,           /* 21 Match one character, casefully */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1241    OP_CHARNC,         /* 22 Match one character, caselessly */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1242    OP_NOT,            /* 23 Match one character, not the following one */    OP_EOD,            /* 24 End of data: \z */
1243    
1244    OP_STAR,           /* 24 The maximizing and minimizing versions of */    OP_OPT,            /* 25 Set runtime options */
1245    OP_MINSTAR,        /* 25 all these opcodes must come in pairs, with */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
1246    OP_PLUS,           /* 26 the minimizing one second. */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
1247    OP_MINPLUS,        /* 27 This first set applies to single characters */    OP_CHAR,           /* 28 Match one character, casefully */
1248    OP_QUERY,          /* 28 */    OP_CHARNC,         /* 29 Match one character, caselessly */
1249    OP_MINQUERY,       /* 29 */    OP_NOT,            /* 30 Match one character, not the following one */
1250    OP_UPTO,           /* 30 From 0 to n matches */  
1251    OP_MINUPTO,        /* 31 */    OP_STAR,           /* 31 The maximizing and minimizing versions of */
1252    OP_EXACT,          /* 32 Exactly n matches */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
1253      OP_PLUS,           /* 33 the minimizing one second. */
1254    OP_NOTSTAR,        /* 33 The maximizing and minimizing versions of */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
1255    OP_NOTMINSTAR,     /* 34 all these opcodes must come in pairs, with */    OP_QUERY,          /* 35 */
1256    OP_NOTPLUS,        /* 35 the minimizing one second. */    OP_MINQUERY,       /* 36 */
1257    OP_NOTMINPLUS,     /* 36 This set applies to "not" single characters */  
1258    OP_NOTQUERY,       /* 37 */    OP_UPTO,           /* 37 From 0 to n matches */
1259    OP_NOTMINQUERY,    /* 38 */    OP_MINUPTO,        /* 38 */
1260    OP_NOTUPTO,        /* 39 From 0 to n matches */    OP_EXACT,          /* 39 Exactly n matches */
1261    OP_NOTMINUPTO,     /* 40 */  
1262    OP_NOTEXACT,       /* 41 Exactly n matches */    OP_POSSTAR,        /* 40 Possessified star */
1263      OP_POSPLUS,        /* 41 Possessified plus */
1264    OP_TYPESTAR,       /* 42 The maximizing and minimizing versions of */    OP_POSQUERY,       /* 42 Posesssified query */
1265    OP_TYPEMINSTAR,    /* 43 all these opcodes must come in pairs, with */    OP_POSUPTO,        /* 43 Possessified upto */
1266    OP_TYPEPLUS,       /* 44 the minimizing one second. These codes must */  
1267    OP_TYPEMINPLUS,    /* 45 be in exactly the same order as those above. */    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
1268    OP_TYPEQUERY,      /* 46 This set applies to character types such as \d */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
1269    OP_TYPEMINQUERY,   /* 47 */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
1270    OP_TYPEUPTO,       /* 48 From 0 to n matches */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
1271    OP_TYPEMINUPTO,    /* 49 */    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
1272    OP_TYPEEXACT,      /* 50 Exactly n matches */    OP_NOTMINQUERY,    /* 49 */
1273    
1274    OP_CRSTAR,         /* 51 The maximizing and minimizing versions of */    OP_NOTUPTO,        /* 50 From 0 to n matches */
1275    OP_CRMINSTAR,      /* 52 all these opcodes must come in pairs, with */    OP_NOTMINUPTO,     /* 51 */
1276    OP_CRPLUS,         /* 53 the minimizing one second. These codes must */    OP_NOTEXACT,       /* 52 Exactly n matches */
1277    OP_CRMINPLUS,      /* 54 be in exactly the same order as those above. */  
1278    OP_CRQUERY,        /* 55 These are for character classes and back refs */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
1279    OP_CRMINQUERY,     /* 56 */    OP_NOTPOSPLUS,     /* 54 */
1280    OP_CRRANGE,        /* 57 These are different to the three sets above. */    OP_NOTPOSQUERY,    /* 55 */
1281    OP_CRMINRANGE,     /* 58 */    OP_NOTPOSUPTO,     /* 56 */
1282    
1283      OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
1284      OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
1285      OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
1286      OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
1287      OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
1288      OP_TYPEMINQUERY,   /* 62 */
1289    
1290      OP_TYPEUPTO,       /* 63 From 0 to n matches */
1291      OP_TYPEMINUPTO,    /* 64 */
1292      OP_TYPEEXACT,      /* 65 Exactly n matches */
1293    
1294      OP_TYPEPOSSTAR,    /* 66 Possessified versions */
1295      OP_TYPEPOSPLUS,    /* 67 */
1296      OP_TYPEPOSQUERY,   /* 68 */
1297      OP_TYPEPOSUPTO,    /* 69 */
1298    
1299      OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
1300      OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
1301      OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
1302      OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
1303      OP_CRQUERY,        /* 74 These are for character classes and back refs */
1304      OP_CRMINQUERY,     /* 75 */
1305      OP_CRRANGE,        /* 76 These are different to the three sets above. */
1306      OP_CRMINRANGE,     /* 77 */
1307    
1308    OP_CLASS,          /* 59 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
1309    OP_NCLASS,         /* 60 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
1310                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
1311                             character > 255 is encountered. */                             character > 255 is encountered. */
1312    
1313    OP_XCLASS,         /* 61 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
1314                             class. This does both positive and negative. */                             class. This does both positive and negative. */
1315    
1316    OP_REF,            /* 62 Match a back reference */    OP_REF,            /* 81 Match a back reference */
1317    OP_RECURSE,        /* 63 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
1318    OP_CALLOUT,        /* 64 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
1319    
1320    OP_ALT,            /* 65 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
1321    OP_KET,            /* 66 End of group that doesn't have an unbounded repeat */    OP_KET,            /* 85 End of group that doesn't have an unbounded repeat */
1322    OP_KETRMAX,        /* 67 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
1323    OP_KETRMIN,        /* 68 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
1324    
1325    /* The assertions must come before ONCE and COND */    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
1326    
1327    OP_ASSERT,         /* 69 Positive lookahead */    OP_ASSERT,         /* 88 Positive lookahead */
1328    OP_ASSERT_NOT,     /* 70 Negative lookahead */    OP_ASSERT_NOT,     /* 89 Negative lookahead */
1329    OP_ASSERTBACK,     /* 71 Positive lookbehind */    OP_ASSERTBACK,     /* 90 Positive lookbehind */
1330    OP_ASSERTBACK_NOT, /* 72 Negative lookbehind */    OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
1331    OP_REVERSE,        /* 73 Move pointer back - used in lookbehind assertions */    OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
   
   /* ONCE and COND must come after the assertions, with ONCE first, as there's  
   a test for >= ONCE for a subpattern that isn't an assertion. */  
   
   OP_ONCE,           /* 74 Once matched, don't back up into the subpattern */  
   OP_COND,           /* 75 Conditional group */  
   OP_CREF,           /* 76 Used to hold an extraction string number (cond ref) */  
   
   OP_BRAZERO,        /* 77 These two must remain together and in this */  
   OP_BRAMINZERO,     /* 78 order. */  
   
   OP_BRANUMBER,      /* 79 Used for extracting brackets whose number is greater  
                            than can fit into an opcode. */  
   
   OP_BRA             /* 80 This and greater values are used for brackets that  
                            extract substrings up to EXTRACT_BASIC_MAX. After  
                            that, use is made of OP_BRANUMBER. */  
 };  
1332    
1333  /* WARNING WARNING WARNING: There is an implicit assumption in pcre.c and    /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
1334  study.c that all opcodes are less than 128 in value. This makes handling UTF-8    as there's a test for >= ONCE for a subpattern that isn't an assertion. */
 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. */  
1335    
1336  #define EXTRACT_BASIC_MAX  100    OP_ONCE,           /* 93 Atomic group */
1337      OP_BRA,            /* 94 Start of non-capturing bracket */
1338      OP_CBRA,           /* 95 Start of capturing bracket */
1339      OP_COND,           /* 96 Conditional group */
1340    
1341      /* These three must follow the previous three, in the same order. There's a
1342      check for >= SBRA to distinguish the two sets. */
1343    
1344      OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
1345      OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1346      OP_SCOND,          /* 99 Conditional group, check empty */
1347    
1348      OP_CREF,           /* 100 Used to hold a capture number as condition */
1349      OP_RREF,           /* 101 Used to hold a recursion number as condition */
1350      OP_DEF,            /* 102 The DEFINE condition */
1351    
1352      OP_BRAZERO,        /* 103 These two must remain together and in this */
1353      OP_BRAMINZERO,     /* 104 order. */
1354    
1355      /* These are backtracking control verbs */
1356    
1357      OP_PRUNE,          /* 105 */
1358      OP_SKIP,           /* 106 */
1359      OP_THEN,           /* 107 */
1360      OP_COMMIT,         /* 108 */
1361    
1362      /* These are forced failure and success verbs */
1363    
1364      OP_FAIL,           /* 109 */
1365      OP_ACCEPT,         /* 110 */
1366    
1367      /* This is used to skip a subpattern with a {0} quantifier */
1368    
1369      OP_SKIPZERO        /* 111 */
1370    };
1371    
1372    
1373  /* 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
1374  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
1375    
1376  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1377    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1378    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1379    "notprop", "prop", "extuni",                                    \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1380    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
1381    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
1382    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1383      "*+","++", "?+", "{",                                           \
1384    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1385      "*+","++", "?+", "{",                                           \
1386    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1387      "*+","++", "?+", "{",                                           \
1388    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1389    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
1390    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1391    "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\    "AssertB", "AssertB not", "Reverse",                            \
1392    "Brazero", "Braminzero", "Branumber", "Bra"    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1393      "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
1394      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
1395      "Skip zero"
1396    
1397    
1398  /* 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 668  in UTF-8 mode. The code that uses this t Line 1406  in UTF-8 mode. The code that uses this t
1406    
1407  #define OP_LENGTHS \  #define OP_LENGTHS \
1408    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1409    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                     */ \
1410    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1411      1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1412    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
1413      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1414    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
1415    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1416    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 678  in UTF-8 mode. The code that uses this t Line 1418  in UTF-8 mode. The code that uses this t
1418    /* Positive single-char repeats                            ** These are  */ \    /* Positive single-char repeats                            ** These are  */ \
1419    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \
1420    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \
1421      2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
1422    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1423    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1424    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
1425      2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \
1426    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1427    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1428    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
1429      2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \
1430    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
1431    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
1432    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \
# Line 702  in UTF-8 mode. The code that uses this t Line 1445  in UTF-8 mode. The code that uses this t
1445    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1446    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
1447    1+LINK_SIZE,                   /* Reverse                                */ \    1+LINK_SIZE,                   /* Reverse                                */ \
1448    1+LINK_SIZE,                   /* Once                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1449      1+LINK_SIZE,                   /* BRA                                    */ \
1450      3+LINK_SIZE,                   /* CBRA                                   */ \
1451    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1452      1+LINK_SIZE,                   /* SBRA                                   */ \
1453      3+LINK_SIZE,                   /* SCBRA                                  */ \
1454      1+LINK_SIZE,                   /* SCOND                                  */ \
1455    3,                             /* CREF                                   */ \    3,                             /* CREF                                   */ \
1456      3,                             /* RREF                                   */ \
1457      1,                             /* DEF                                    */ \
1458    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1459    3,                             /* BRANUMBER                              */ \    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
1460    1+LINK_SIZE                    /* BRA                                    */ \    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
1461    
1462    
1463  /* A magic value for OP_CREF to indicate the "in recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
1464    
1465  #define CREF_RECURSE  0xffff  #define RREF_ANY  0xffff
1466    
1467  /* Error code numbers. They are given names so that they can more easily be  /* Error code numbers. They are given names so that they can more easily be
1468  tracked. */  tracked. */
# Line 722  enum { ERR0, ERR1, ERR2, ERR3, ERR4, Line 1472  enum { ERR0, ERR1, ERR2, ERR3, ERR4,
1472         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
1473         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1474         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1475         ERR50, ERR51 };         ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1476           ERR60, ERR61, ERR62, ERR63, ERR64 };
1477    
1478  /* 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
1479  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 744  NOTE NOTE NOTE: Line 1495  NOTE NOTE NOTE:
1495  typedef struct real_pcre {  typedef struct real_pcre {
1496    pcre_uint32 magic_number;    pcre_uint32 magic_number;
1497    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1498    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
1499    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
1500      pcre_uint16 dummy1;             /* For future use */
1501    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
1502    pcre_uint16 top_backref;    pcre_uint16 top_backref;
1503    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 777  typedef struct compile_data { Line 1528  typedef struct compile_data {
1528    const uschar *fcc;            /* Points to case-flipping table */    const uschar *fcc;            /* Points to case-flipping table */
1529    const uschar *cbits;          /* Points to character type table */    const uschar *cbits;          /* Points to character type table */
1530    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
1531      const uschar *start_workspace;/* The start of working space */
1532    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1533    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1534      const uschar *end_pattern;    /* The end of the pattern */
1535      uschar *hwm;                  /* High watermark of workspace */
1536    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1537    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
1538    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
1539      int  bracount;                /* Count of capturing parens as we compile */
1540      int  final_bracount;          /* Saved value after first pass */
1541    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
1542    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
1543      int  external_options;        /* External (initial) options */
1544      int  external_flags;          /* External flag bits to be set */
1545    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1546    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
1547    int  nllen;                   /* 1 or 2 for newline string length */    int  nltype;                  /* Newline type */
1548    uschar nl[4];                 /* Newline string */    int  nllen;                   /* Newline string length */
1549      uschar nl[4];                 /* Newline string when fixed length */
1550  } compile_data;  } compile_data;
1551    
1552  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
# Line 805  typedef struct recursion_info { Line 1564  typedef struct recursion_info {
1564    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1565    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
1566    const uschar *after_call;     /* "Return value": points after the call in the expr */    const uschar *after_call;     /* "Return value": points after the call in the expr */
1567    USPTR save_start;             /* Old value of md->start_match */    USPTR save_start;             /* Old value of mstart */
1568    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1569    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1570  } recursion_info;  } recursion_info;
1571    
1572  /* When compiling in a mode that doesn't use recursive calls to match(),  /* Structure for building a chain of data for holding the values of the subject
1573  a structure is used to remember local variables on the heap. It is defined in  pointer at the start of each subpattern, so as to detect when an empty string
1574  pcre_exec.c, close to the match() function, so that it is easy to keep it in  has been matched by a subpattern - to break infinite loops. */
1575  step with any changes of local variable. However, the pointer to the current  
1576  frame must be saved in some "static" place over a longjmp(). We declare the  typedef struct eptrblock {
1577  structure here so that we can put a pointer in the match_data structure. NOTE:    struct eptrblock *epb_prev;
1578  This isn't used for a "normal" compilation of pcre. */    USPTR epb_saved_eptr;
1579    } eptrblock;
1580    
 struct heapframe;  
1581    
1582  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1583  doing traditional NFA matching, so that they are thread-safe. */  doing traditional NFA matching, so that they are thread-safe. */
# Line 830  typedef struct match_data { Line 1589  typedef struct match_data {
1589    int   *offset_vector;         /* Offset vector */    int   *offset_vector;         /* Offset vector */
1590    int    offset_end;            /* One past the end */    int    offset_end;            /* One past the end */
1591    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1592    int    nllen;                 /* 1 or 2 for newline string length */    int    nltype;                /* Newline type */
1593    uschar nl[4];                 /* Newline string */    int    nllen;                 /* Newline string length */
1594      uschar nl[4];                 /* Newline string when fixed */
1595    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1596    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
1597    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
1598    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1599    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1600    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1601      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1602    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1603    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
   BOOL   partial;               /* PARTIAL flag */  
1604    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1605      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1606    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1607    USPTR  start_subject;         /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1608    USPTR  end_subject;           /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1609    USPTR  start_match;           /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
1610    USPTR  end_match_ptr;         /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1611      USPTR  start_used_ptr;        /* Earliest consulted character */
1612      int    partial;               /* PARTIAL options */
1613    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1614    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1615    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
1616      eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
1617      int    eptrn;                 /* Next free eptrblock */
1618    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1619    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
   struct heapframe *thisframe;  /* Used only when compiling for no recursion */  
1620  } match_data;  } match_data;
1621    
1622  /* A similar structure is used for the same purpose by the DFA matching  /* A similar structure is used for the same purpose by the DFA matching
# Line 862  typedef struct dfa_match_data { Line 1626  typedef struct dfa_match_data {
1626    const uschar *start_code;     /* Start of the compiled pattern */    const uschar *start_code;     /* Start of the compiled pattern */
1627    const uschar *start_subject;  /* Start of the subject string */    const uschar *start_subject;  /* Start of the subject string */
1628    const uschar *end_subject;    /* End of subject string */    const uschar *end_subject;    /* End of subject string */
1629      const uschar *start_used_ptr; /* Earliest consulted character */
1630    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1631    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1632    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1633    int    nllen;                 /* 1 or 2 for newline string length */    int    nltype;                /* Newline type */
1634    uschar nl[4];                 /* Newline string */    int    nllen;                 /* Newline string length */
1635      uschar nl[4];                 /* Newline string when fixed */
1636    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1637  } dfa_match_data;  } dfa_match_data;
1638    
# Line 876  typedef struct dfa_match_data { Line 1642  typedef struct dfa_match_data {
1642  #define ctype_letter  0x02  #define ctype_letter  0x02
1643  #define ctype_digit   0x04  #define ctype_digit   0x04
1644  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1645  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1646  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1647    
1648  /* 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 904  total length. */ Line 1670  total length. */
1670  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1671    
1672  /* Layout of the UCP type table that translates property names into types and  /* Layout of the UCP type table that translates property names into types and
1673  codes. */  codes. Each entry used to point directly to a name, but to reduce the number of
1674    relocations in shared libraries, it now has an offset into a single string
1675    instead. */
1676    
1677  typedef struct {  typedef struct {
1678    const char *name;    pcre_uint16 name_offset;
1679    pcre_uint16 type;    pcre_uint16 type;
1680    pcre_uint16 value;    pcre_uint16 value;
1681  } ucp_type_table;  } ucp_type_table;
# Line 925  extern const uschar _pcre_utf8_table4[]; Line 1693  extern const uschar _pcre_utf8_table4[];
1693    
1694  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1695    
1696    extern const char   _pcre_utt_names[];
1697  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1698  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1699    
# Line 937  extern const uschar _pcre_OP_lengths[]; Line 1706  extern const uschar _pcre_OP_lengths[];
1706  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
1707  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1708    
1709  extern int         _pcre_ord2utf8(int, uschar *);  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,
1710  extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *,                        int *, BOOL);
1711                       const pcre_study_data *, pcre_study_data *);  extern int          _pcre_ord2utf8(int, uschar *);
1712  extern int         _pcre_ucp_findprop(const unsigned int, int *, int *);  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1713  extern int         _pcre_ucp_othercase(const int);                        const pcre_study_data *, pcre_study_data *);
1714  extern int         _pcre_valid_utf8(const uschar *, int);  extern int          _pcre_valid_utf8(const uschar *, int);
1715  extern BOOL        _pcre_xclass(int, const uschar *);  extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1716                          int *, BOOL);
1717    extern BOOL         _pcre_xclass(int, const uschar *);
1718    
1719    
1720    /* Unicode character database (UCD) */
1721    
1722    typedef struct {
1723      uschar script;
1724      uschar chartype;
1725      pcre_int32 other_case;
1726    } ucd_record;
1727    
1728    extern const ucd_record  _pcre_ucd_records[];
1729    extern const uschar      _pcre_ucd_stage1[];
1730    extern const pcre_uint16 _pcre_ucd_stage2[];
1731    extern const int         _pcre_ucp_gentype[];
1732    
1733    
1734    /* UCD access macros */
1735    
1736    #define UCD_BLOCK_SIZE 128
1737    #define GET_UCD(ch) (_pcre_ucd_records + \
1738            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1739            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1740    
1741    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1742    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1743    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1744    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1745    
1746  #endif  #endif
1747    

Legend:
Removed from v.91  
changed lines
  Added in v.435

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12