/[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 85 by nigel, Sat Feb 24 21:41:13 2007 UTC revision 392 by ph10, Tue Mar 17 21:30:30 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) 1997-2005 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  /* 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
62  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
63  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
64  all, it had only been about 10 years then... */  all, it had only been about 10 years then...
65    
66    It turns out that the Mac Debugging.h header also defines the macro DPRINTF, so
67    be absolutely sure we get our version. */
68    
69    #undef DPRINTF
70  #ifdef DEBUG  #ifdef DEBUG
71  #define DPRINTF(p) printf p  #define DPRINTF(p) printf p
72  #else  #else
73  #define DPRINTF(p) /*nothing*/  #define DPRINTF(p) /* Nothing */
74  #endif  #endif
75    
76    
 /* Get the definitions provided by running "configure" */  
   
 #include "config.h"  
   
77  /* Standard C headers plus the external interface definition. The only time  /* Standard C headers plus the external interface definition. The only time
78  setjmp and stdarg are used is when NO_RECURSE is set. */  setjmp and stdarg are used is when NO_RECURSE is set. */
79    
# Line 79  setjmp and stdarg are used is when NO_RE Line 86  setjmp and stdarg are used is when NO_RE
86  #include <stdlib.h>  #include <stdlib.h>
87  #include <string.h>  #include <string.h>
88    
89  #ifndef PCRE_SPY  /* When compiling a DLL for Windows, the exported symbols have to be declared
90  #define PCRE_DEFINITION       /* Win32 __declspec(export) trigger for .dll */  using some MS magic. I found some useful information on this web page:
91    http://msdn2.microsoft.com/en-us/library/y4h7bcy6(VS.80).aspx. According to the
92    information there, using __declspec(dllexport) without "extern" we have a
93    definition; with "extern" we have a declaration. The settings here override the
94    setting in pcre.h (which is included below); it defines only PCRE_EXP_DECL,
95    which is all that is needed for applications (they just import the symbols). We
96    use:
97    
98      PCRE_EXP_DECL       for declarations
99      PCRE_EXP_DEFN       for definitions of exported functions
100      PCRE_EXP_DATA_DEFN  for definitions of exported variables
101    
102    The reason for the two DEFN macros is that in non-Windows environments, one
103    does not want to have "extern" before variable definitions because it leads to
104    compiler warnings. So we distinguish between functions and variables. In
105    Windows, the two should always be the same.
106    
107    The reason for wrapping this in #ifndef PCRE_EXP_DECL is so that pcretest,
108    which is an application, but needs to import this file in order to "peek" at
109    internals, can #include pcre.h first to get an application's-eye view.
110    
111    In principle, people compiling for non-Windows, non-Unix-like (i.e. uncommon,
112    special-purpose environments) might want to stick other stuff in front of
113    exported symbols. That's why, in the non-Windows case, we set PCRE_EXP_DEFN and
114    PCRE_EXP_DATA_DEFN only if they are not already set. */
115    
116    #ifndef PCRE_EXP_DECL
117    #  ifdef _WIN32
118    #    ifndef PCRE_STATIC
119    #      define PCRE_EXP_DECL       extern __declspec(dllexport)
120    #      define PCRE_EXP_DEFN       __declspec(dllexport)
121    #      define PCRE_EXP_DATA_DEFN  __declspec(dllexport)
122    #    else
123    #      define PCRE_EXP_DECL       extern
124    #      define PCRE_EXP_DEFN
125    #      define PCRE_EXP_DATA_DEFN
126    #    endif
127    #  else
128    #    ifdef __cplusplus
129    #      define PCRE_EXP_DECL       extern "C"
130    #    else
131    #      define PCRE_EXP_DECL       extern
132    #    endif
133    #    ifndef PCRE_EXP_DEFN
134    #      define PCRE_EXP_DEFN       PCRE_EXP_DECL
135    #    endif
136    #    ifndef PCRE_EXP_DATA_DEFN
137    #      define PCRE_EXP_DATA_DEFN
138    #    endif
139    #  endif
140    #endif
141    
142    /* When compiling with the MSVC compiler, it is sometimes necessary to include
143    a "calling convention" before exported function names. (This is secondhand
144    information; I know nothing about MSVC myself). For example, something like
145    
146      void __cdecl function(....)
147    
148    might be needed. In order so make this easy, all the exported functions have
149    PCRE_CALL_CONVENTION just before their names. It is rarely needed; if not
150    set, we ensure here that it has no effect. */
151    
152    #ifndef PCRE_CALL_CONVENTION
153    #define PCRE_CALL_CONVENTION
154  #endif  #endif
155    
156  /* 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 161  preprocessor time in standard C environm
161    
162  #if USHRT_MAX == 65535  #if USHRT_MAX == 65535
163    typedef unsigned short pcre_uint16;    typedef unsigned short pcre_uint16;
164      typedef short pcre_int16;
165  #elif UINT_MAX == 65535  #elif UINT_MAX == 65535
166    typedef unsigned int pcre_uint16;    typedef unsigned int pcre_uint16;
167      typedef int pcre_int16;
168  #else  #else
169    #error Cannot determine a type for 16-bit unsigned integers    #error Cannot determine a type for 16-bit unsigned integers
170  #endif  #endif
171    
172  #if UINT_MAX == 4294967295  #if UINT_MAX == 4294967295
173    typedef unsigned int pcre_uint32;    typedef unsigned int pcre_uint32;
174      typedef int pcre_int32;
175  #elif ULONG_MAX == 4294967295  #elif ULONG_MAX == 4294967295
176    typedef unsigned long int pcre_uint32;    typedef unsigned long int pcre_uint32;
177      typedef long int pcre_int32;
178  #else  #else
179    #error Cannot determine a type for 32-bit unsigned integers    #error Cannot determine a type for 32-bit unsigned integers
180  #endif  #endif
# Line 114  Unix, where it is defined in sys/types, Line 188  Unix, where it is defined in sys/types,
188    
189  typedef unsigned char uschar;  typedef unsigned char uschar;
190    
191    /* This is an unsigned int value that no character can ever have. UTF-8
192    characters only go up to 0x7fffffff (though Unicode doesn't go beyond
193    0x0010ffff). */
194    
195    #define NOTACHAR 0xffffffff
196    
197    /* PCRE is able to support several different kinds of newline (CR, LF, CRLF,
198    "any" and "anycrlf" at present). The following macros are used to package up
199    testing for newlines. NLBLOCK, PSSTART, and PSEND are defined in the various
200    modules to indicate in which datablock the parameters exist, and what the
201    start/end of string field names are. */
202    
203    #define NLTYPE_FIXED    0     /* Newline is a fixed length string */
204    #define NLTYPE_ANY      1     /* Newline is any Unicode line ending */
205    #define NLTYPE_ANYCRLF  2     /* Newline is CR, LF, or CRLF */
206    
207    /* This macro checks for a newline at the given position */
208    
209    #define IS_NEWLINE(p) \
210      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
211        ((p) < NLBLOCK->PSEND && \
212         _pcre_is_newline((p), NLBLOCK->nltype, NLBLOCK->PSEND, &(NLBLOCK->nllen),\
213           utf8)) \
214        : \
215        ((p) <= NLBLOCK->PSEND - NLBLOCK->nllen && \
216         (p)[0] == NLBLOCK->nl[0] && \
217         (NLBLOCK->nllen == 1 || (p)[1] == NLBLOCK->nl[1]) \
218        ) \
219      )
220    
221    /* This macro checks for a newline immediately preceding the given position */
222    
223    #define WAS_NEWLINE(p) \
224      ((NLBLOCK->nltype != NLTYPE_FIXED)? \
225        ((p) > NLBLOCK->PSSTART && \
226         _pcre_was_newline((p), NLBLOCK->nltype, NLBLOCK->PSSTART, \
227           &(NLBLOCK->nllen), utf8)) \
228        : \
229        ((p) >= NLBLOCK->PSSTART + NLBLOCK->nllen && \
230         (p)[-NLBLOCK->nllen] == NLBLOCK->nl[0] && \
231         (NLBLOCK->nllen == 1 || (p)[-NLBLOCK->nllen+1] == NLBLOCK->nl[1]) \
232        ) \
233      )
234    
235    /* When PCRE is compiled as a C++ library, the subject pointer can be replaced
236    with a custom type. This makes it possible, for example, to allow pcre_exec()
237    to process subject strings that are discontinuous by using a smart pointer
238    class. It must always be possible to inspect all of the subject string in
239    pcre_exec() because of the way it backtracks. Two macros are required in the
240    normal case, for sign-unspecified and unsigned char pointers. The former is
241    used for the external interface and appears in pcre.h, which is why its name
242    must begin with PCRE_. */
243    
244    #ifdef CUSTOM_SUBJECT_PTR
245    #define PCRE_SPTR CUSTOM_SUBJECT_PTR
246    #define USPTR CUSTOM_SUBJECT_PTR
247    #else
248    #define PCRE_SPTR const char *
249    #define USPTR const unsigned char *
250    #endif
251    
252    
253    
254  /* Include the public PCRE header and the definitions of UCP character property  /* Include the public PCRE header and the definitions of UCP character property
255  values. */  values. */
256    
# Line 125  need to have their names changed. PCRE m Line 262  need to have their names changed. PCRE m
262  option on the command line. */  option on the command line. */
263    
264  #ifdef VPCOMPAT  #ifdef VPCOMPAT
265    #define strlen(s)        _strlen(s)
266  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)  #define strncmp(s1,s2,m) _strncmp(s1,s2,m)
267    #define memcmp(s,c,n)    _memcmp(s,c,n)
268  #define memcpy(d,s,n)    _memcpy(d,s,n)  #define memcpy(d,s,n)    _memcpy(d,s,n)
269  #define memmove(d,s,n)   _memmove(d,s,n)  #define memmove(d,s,n)   _memmove(d,s,n)
270  #define memset(s,c,n)    _memset(s,c,n)  #define memset(s,c,n)    _memset(s,c,n)
# Line 134  option on the command line. */ Line 273  option on the command line. */
273  /* 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(),
274  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
275  is set. Otherwise, include an emulating function for those systems that have  is set. Otherwise, include an emulating function for those systems that have
276  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. */  
277    
278  #if ! HAVE_MEMMOVE  #ifndef HAVE_MEMMOVE
279  #undef  memmove        /* some systems may have a macro */  #undef  memmove        /* some systems may have a macro */
280  #if HAVE_BCOPY  #ifdef HAVE_BCOPY
281  #define memmove(a, b, c) bcopy(b, a, c)  #define memmove(a, b, c) bcopy(b, a, c)
282  #else  /* HAVE_BCOPY */  #else  /* HAVE_BCOPY */
283  void *  static void *
284  pcre_memmove(unsigned char *dest, const unsigned char *src, size_t n)  pcre_memmove(void *d, const void *s, size_t n)
285  {  {
286  int i;  size_t i;
287  dest += n;  unsigned char *dest = (unsigned char *)d;
288  src += n;  const unsigned char *src = (const unsigned char *)s;
289  for (i = 0; i < n; ++i) *(--dest) =  *(--src);  if (dest > src)
290      {
291      dest += n;
292      src += n;
293      for (i = 0; i < n; ++i) *(--dest) = *(--src);
294      return (void *)dest;
295      }
296    else
297      {
298      for (i = 0; i < n; ++i) *dest++ = *src++;
299      return (void *)(dest - n);
300      }
301  }  }
302  #define memmove(a, b, c) pcre_memmove(a, b, c)  #define memmove(a, b, c) pcre_memmove(a, b, c)
303  #endif   /* not HAVE_BCOPY */  #endif   /* not HAVE_BCOPY */
# Line 235  capturing parenthesis numbers in back re Line 383  capturing parenthesis numbers in back re
383    
384  /* 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
385  byte. The macros for character handling generate simple sequences when used in  byte. The macros for character handling generate simple sequences when used in
386  byte-mode, and more complicated ones for UTF-8 characters. */  byte-mode, and more complicated ones for UTF-8 characters. BACKCHAR should
387    never be called in byte mode. To make sure it can never even appear when UTF-8
388    support is omitted, we don't even define it. */
389    
390  #ifndef SUPPORT_UTF8  #ifndef SUPPORT_UTF8
391  #define GETCHAR(c, eptr) c = *eptr;  #define GETCHAR(c, eptr) c = *eptr;
# Line 243  byte-mode, and more complicated ones for Line 393  byte-mode, and more complicated ones for
393  #define GETCHARINC(c, eptr) c = *eptr++;  #define GETCHARINC(c, eptr) c = *eptr++;
394  #define GETCHARINCTEST(c, eptr) c = *eptr++;  #define GETCHARINCTEST(c, eptr) c = *eptr++;
395  #define GETCHARLEN(c, eptr, len) c = *eptr;  #define GETCHARLEN(c, eptr, len) c = *eptr;
396  #define BACKCHAR(eptr)  /* #define BACKCHAR(eptr) */
397    
398  #else   /* SUPPORT_UTF8 */  #else   /* SUPPORT_UTF8 */
399    
# Line 252  we know we are in UTF-8 mode. */ Line 402  we know we are in UTF-8 mode. */
402    
403  #define GETCHAR(c, eptr) \  #define GETCHAR(c, eptr) \
404    c = *eptr; \    c = *eptr; \
405    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
406      { \      { \
407      int gcii; \      int gcii; \
408      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 270  pointer. */ Line 420  pointer. */
420    
421  #define GETCHARTEST(c, eptr) \  #define GETCHARTEST(c, eptr) \
422    c = *eptr; \    c = *eptr; \
423    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
424      { \      { \
425      int gcii; \      int gcii; \
426      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 288  know we are in UTF-8 mode. */ Line 438  know we are in UTF-8 mode. */
438    
439  #define GETCHARINC(c, eptr) \  #define GETCHARINC(c, eptr) \
440    c = *eptr++; \    c = *eptr++; \
441    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
442      { \      { \
443      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
444      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 304  know we are in UTF-8 mode. */ Line 454  know we are in UTF-8 mode. */
454    
455  #define GETCHARINCTEST(c, eptr) \  #define GETCHARINCTEST(c, eptr) \
456    c = *eptr++; \    c = *eptr++; \
457    if (utf8 && (c & 0xc0) == 0xc0) \    if (utf8 && c >= 0xc0) \
458      { \      { \
459      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
460      int gcss = 6*gcaa; \      int gcss = 6*gcaa; \
# Line 321  if there are extra bytes. This is called Line 471  if there are extra bytes. This is called
471    
472  #define GETCHARLEN(c, eptr, len) \  #define GETCHARLEN(c, eptr, len) \
473    c = *eptr; \    c = *eptr; \
474    if ((c & 0xc0) == 0xc0) \    if (c >= 0xc0) \
475        { \
476        int gcii; \
477        int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
478        int gcss = 6*gcaa; \
479        c = (c & _pcre_utf8_table3[gcaa]) << gcss; \
480        for (gcii = 1; gcii <= gcaa; gcii++) \
481          { \
482          gcss -= 6; \
483          c |= (eptr[gcii] & 0x3f) << gcss; \
484          } \
485        len += gcaa; \
486        }
487    
488    /* Get the next UTF-8 character, testing for UTF-8 mode, not advancing the
489    pointer, incrementing length if there are extra bytes. This is called when we
490    know we are in UTF-8 mode. */
491    
492    #define GETCHARLENTEST(c, eptr, len) \
493      c = *eptr; \
494      if (utf8 && c >= 0xc0) \
495      { \      { \
496      int gcii; \      int gcii; \
497      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \      int gcaa = _pcre_utf8_table4[c & 0x3f];  /* Number of additional bytes */ \
# Line 336  if there are extra bytes. This is called Line 506  if there are extra bytes. This is called
506      }      }
507    
508  /* 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
509  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
510    because almost all calls are already within a block of UTF-8 only code. */
511    
512  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--;  #define BACKCHAR(eptr) while((*eptr & 0xc0) == 0x80) eptr--
513    
514  #endif  #endif
515    
# Line 355  Standard C system should have one. */ Line 526  Standard C system should have one. */
526    
527  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)  #define PCRE_IMS (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL)
528    
529  /* Private options flags start at the most significant end of the four bytes,  /* Private flags containing information about the compiled regex. They used to
530  but skip the top bit so we can use ints for convenience without getting tangled  live at the top end of the options word, but that got almost full, so now they
531  with negative values. The public options defined in pcre.h start at the least  are in a 16-bit flags word. */
532  significant end. Make sure they don't overlap! */  
533    #define PCRE_NOPARTIAL     0x0001  /* can't use partial with this regex */
534  #define PCRE_FIRSTSET      0x40000000  /* first_byte is set */  #define PCRE_FIRSTSET      0x0002  /* first_byte is set */
535  #define PCRE_REQCHSET      0x20000000  /* req_byte is set */  #define PCRE_REQCHSET      0x0004  /* req_byte is set */
536  #define PCRE_STARTLINE     0x10000000  /* start after \n for multiline */  #define PCRE_STARTLINE     0x0008  /* start after \n for multiline */
537  #define PCRE_ICHANGED      0x08000000  /* i option changes within regex */  #define PCRE_JCHANGED      0x0010  /* j option used in regex */
538  #define PCRE_NOPARTIAL     0x04000000  /* can't use partial with this regex */  #define PCRE_HASCRORLF     0x0020  /* explicit \r or \n in pattern */
539    
540  /* Options for the "extra" block produced by pcre_study(). */  /* Options for the "extra" block produced by pcre_study(). */
541    
# Line 373  significant end. Make sure they don't ov Line 544  significant end. Make sure they don't ov
544  /* Masks for identifying the public options that are permitted at compile  /* Masks for identifying the public options that are permitted at compile
545  time, run time, or study time, respectively. */  time, run time, or study time, respectively. */
546    
547  #define PUBLIC_OPTIONS \  #define PCRE_NEWLINE_BITS (PCRE_NEWLINE_CR|PCRE_NEWLINE_LF|PCRE_NEWLINE_ANY| \
548                               PCRE_NEWLINE_ANYCRLF)
549    
550    #define PUBLIC_COMPILE_OPTIONS \
551    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \    (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
552     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \     PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY|PCRE_UTF8| \
553     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| \
554       PCRE_DUPNAMES|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
555       PCRE_JAVASCRIPT_COMPAT)
556    
557  #define PUBLIC_EXEC_OPTIONS \  #define PUBLIC_EXEC_OPTIONS \
558    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
559     PCRE_PARTIAL)     PCRE_PARTIAL|PCRE_NEWLINE_BITS|PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE| \
560       PCRE_NO_START_OPTIMIZE)
561    
562  #define PUBLIC_DFA_EXEC_OPTIONS \  #define PUBLIC_DFA_EXEC_OPTIONS \
563    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \    (PCRE_ANCHORED|PCRE_NOTBOL|PCRE_NOTEOL|PCRE_NOTEMPTY|PCRE_NO_UTF8_CHECK| \
564     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART)     PCRE_PARTIAL|PCRE_DFA_SHORTEST|PCRE_DFA_RESTART|PCRE_NEWLINE_BITS| \
565       PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE|PCRE_NO_START_OPTIMIZE)
566    
567  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */  #define PUBLIC_STUDY_OPTIONS 0   /* None defined */
568    
# Line 409  variable-length repeat, or a anything ot Line 587  variable-length repeat, or a anything ot
587  #define REQ_CASELESS 0x0100    /* indicates caselessness */  #define REQ_CASELESS 0x0100    /* indicates caselessness */
588  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */  #define REQ_VARY     0x0200    /* reqbyte followed non-literal item */
589    
590  /* Miscellaneous definitions */  /* Miscellaneous definitions. The #ifndef is to pacify compiler warnings in
591    environments where these macros are defined elsewhere. Unfortunately, there
592    is no way to do the same for the typedef. */
593    
594  typedef int BOOL;  typedef int BOOL;
595    
596    #ifndef FALSE
597  #define FALSE   0  #define FALSE   0
598  #define TRUE    1  #define TRUE    1
599    #endif
600    
601    /* If PCRE is to support UTF-8 on EBCDIC platforms, we cannot use normal
602    character constants like '*' because the compiler would emit their EBCDIC code,
603    which is different from their ASCII/UTF-8 code. Instead we define macros for
604    the characters so that they always use the ASCII/UTF-8 code when UTF-8 support
605    is enabled. When UTF-8 support is not enabled, the definitions use character
606    literals. Both character and string versions of each character are needed, and
607    there are some longer strings as well.
608    
609    This means that, on EBCDIC platforms, the PCRE library can handle either
610    EBCDIC, or UTF-8, but not both. To support both in the same compiled library
611    would need different lookups depending on whether PCRE_UTF8 was set or not.
612    This would make it impossible to use characters in switch/case statements,
613    which would reduce performance. For a theoretical use (which nobody has asked
614    for) in a minority area (EBCDIC platforms), this is not sensible. Any
615    application that did need both could compile two versions of the library, using
616    macros to give the functions distinct names. */
617    
618    #ifndef SUPPORT_UTF8
619    
620    /* UTF-8 support is not enabled; use the platform-dependent character literals
621    so that PCRE works on both ASCII and EBCDIC platforms, in non-UTF-mode only. */
622    
623    #define CHAR_HT                     '\t'
624    #define CHAR_VT                     '\v'
625    #define CHAR_FF                     '\f'
626    #define CHAR_CR                     '\r'
627    #define CHAR_NL                     '\n'
628    #define CHAR_BS                     '\b'
629    #define CHAR_BEL                    '\a'
630    #ifdef EBCDIC
631    #define CHAR_ESC                    '\047'
632    #define CHAR_DEL                    '\007'
633    #else
634    #define CHAR_ESC                    '\033'
635    #define CHAR_DEL                    '\177'
636    #endif
637    
638    #define CHAR_SPACE                  ' '
639    #define CHAR_EXCLAMATION_MARK       '!'
640    #define CHAR_QUOTATION_MARK         '"'
641    #define CHAR_NUMBER_SIGN            '#'
642    #define CHAR_DOLLAR_SIGN            '$'
643    #define CHAR_PERCENT_SIGN           '%'
644    #define CHAR_AMPERSAND              '&'
645    #define CHAR_APOSTROPHE             '\''
646    #define CHAR_LEFT_PARENTHESIS       '('
647    #define CHAR_RIGHT_PARENTHESIS      ')'
648    #define CHAR_ASTERISK               '*'
649    #define CHAR_PLUS                   '+'
650    #define CHAR_COMMA                  ','
651    #define CHAR_MINUS                  '-'
652    #define CHAR_DOT                    '.'
653    #define CHAR_SLASH                  '/'
654    #define CHAR_0                      '0'
655    #define CHAR_1                      '1'
656    #define CHAR_2                      '2'
657    #define CHAR_3                      '3'
658    #define CHAR_4                      '4'
659    #define CHAR_5                      '5'
660    #define CHAR_6                      '6'
661    #define CHAR_7                      '7'
662    #define CHAR_8                      '8'
663    #define CHAR_9                      '9'
664    #define CHAR_COLON                  ':'
665    #define CHAR_SEMICOLON              ';'
666    #define CHAR_LESS_THAN_SIGN         '<'
667    #define CHAR_EQUALS_SIGN            '='
668    #define CHAR_GREATER_THAN_SIGN      '>'
669    #define CHAR_QUESTION_MARK          '?'
670    #define CHAR_COMMERCIAL_AT          '@'
671    #define CHAR_A                      'A'
672    #define CHAR_B                      'B'
673    #define CHAR_C                      'C'
674    #define CHAR_D                      'D'
675    #define CHAR_E                      'E'
676    #define CHAR_F                      'F'
677    #define CHAR_G                      'G'
678    #define CHAR_H                      'H'
679    #define CHAR_I                      'I'
680    #define CHAR_J                      'J'
681    #define CHAR_K                      'K'
682    #define CHAR_L                      'L'
683    #define CHAR_M                      'M'
684    #define CHAR_N                      'N'
685    #define CHAR_O                      'O'
686    #define CHAR_P                      'P'
687    #define CHAR_Q                      'Q'
688    #define CHAR_R                      'R'
689    #define CHAR_S                      'S'
690    #define CHAR_T                      'T'
691    #define CHAR_U                      'U'
692    #define CHAR_V                      'V'
693    #define CHAR_W                      'W'
694    #define CHAR_X                      'X'
695    #define CHAR_Y                      'Y'
696    #define CHAR_Z                      'Z'
697    #define CHAR_LEFT_SQUARE_BRACKET    '['
698    #define CHAR_BACKSLASH              '\\'
699    #define CHAR_RIGHT_SQUARE_BRACKET   ']'
700    #define CHAR_CIRCUMFLEX_ACCENT      '^'
701    #define CHAR_UNDERSCORE             '_'
702    #define CHAR_GRAVE_ACCENT           '`'
703    #define CHAR_a                      'a'
704    #define CHAR_b                      'b'
705    #define CHAR_c                      'c'
706    #define CHAR_d                      'd'
707    #define CHAR_e                      'e'
708    #define CHAR_f                      'f'
709    #define CHAR_g                      'g'
710    #define CHAR_h                      'h'
711    #define CHAR_i                      'i'
712    #define CHAR_j                      'j'
713    #define CHAR_k                      'k'
714    #define CHAR_l                      'l'
715    #define CHAR_m                      'm'
716    #define CHAR_n                      'n'
717    #define CHAR_o                      'o'
718    #define CHAR_p                      'p'
719    #define CHAR_q                      'q'
720    #define CHAR_r                      'r'
721    #define CHAR_s                      's'
722    #define CHAR_t                      't'
723    #define CHAR_u                      'u'
724    #define CHAR_v                      'v'
725    #define CHAR_w                      'w'
726    #define CHAR_x                      'x'
727    #define CHAR_y                      'y'
728    #define CHAR_z                      'z'
729    #define CHAR_LEFT_CURLY_BRACKET     '{'
730    #define CHAR_VERTICAL_LINE          '|'
731    #define CHAR_RIGHT_CURLY_BRACKET    '}'
732    #define CHAR_TILDE                  '~'
733    
734    #define STR_HT                      "\t"
735    #define STR_VT                      "\v"
736    #define STR_FF                      "\f"
737    #define STR_CR                      "\r"
738    #define STR_NL                      "\n"
739    #define STR_BS                      "\b"
740    #define STR_BEL                     "\a"
741    #ifdef EBCDIC
742    #define STR_ESC                     "\047"
743    #define STR_DEL                     "\007"
744    #else
745    #define STR_ESC                     "\033"
746    #define STR_DEL                     "\177"
747    #endif
748    
749    #define STR_SPACE                   " "
750    #define STR_EXCLAMATION_MARK        "!"
751    #define STR_QUOTATION_MARK          "\""
752    #define STR_NUMBER_SIGN             "#"
753    #define STR_DOLLAR_SIGN             "$"
754    #define STR_PERCENT_SIGN            "%"
755    #define STR_AMPERSAND               "&"
756    #define STR_APOSTROPHE              "'"
757    #define STR_LEFT_PARENTHESIS        "("
758    #define STR_RIGHT_PARENTHESIS       ")"
759    #define STR_ASTERISK                "*"
760    #define STR_PLUS                    "+"
761    #define STR_COMMA                   ","
762    #define STR_MINUS                   "-"
763    #define STR_DOT                     "."
764    #define STR_SLASH                   "/"
765    #define STR_0                       "0"
766    #define STR_1                       "1"
767    #define STR_2                       "2"
768    #define STR_3                       "3"
769    #define STR_4                       "4"
770    #define STR_5                       "5"
771    #define STR_6                       "6"
772    #define STR_7                       "7"
773    #define STR_8                       "8"
774    #define STR_9                       "9"
775    #define STR_COLON                   ":"
776    #define STR_SEMICOLON               ";"
777    #define STR_LESS_THAN_SIGN          "<"
778    #define STR_EQUALS_SIGN             "="
779    #define STR_GREATER_THAN_SIGN       ">"
780    #define STR_QUESTION_MARK           "?"
781    #define STR_COMMERCIAL_AT           "@"
782    #define STR_A                       "A"
783    #define STR_B                       "B"
784    #define STR_C                       "C"
785    #define STR_D                       "D"
786    #define STR_E                       "E"
787    #define STR_F                       "F"
788    #define STR_G                       "G"
789    #define STR_H                       "H"
790    #define STR_I                       "I"
791    #define STR_J                       "J"
792    #define STR_K                       "K"
793    #define STR_L                       "L"
794    #define STR_M                       "M"
795    #define STR_N                       "N"
796    #define STR_O                       "O"
797    #define STR_P                       "P"
798    #define STR_Q                       "Q"
799    #define STR_R                       "R"
800    #define STR_S                       "S"
801    #define STR_T                       "T"
802    #define STR_U                       "U"
803    #define STR_V                       "V"
804    #define STR_W                       "W"
805    #define STR_X                       "X"
806    #define STR_Y                       "Y"
807    #define STR_Z                       "Z"
808    #define STR_LEFT_SQUARE_BRACKET     "["
809    #define STR_BACKSLASH               "\\"
810    #define STR_RIGHT_SQUARE_BRACKET    "]"
811    #define STR_CIRCUMFLEX_ACCENT       "^"
812    #define STR_UNDERSCORE              "_"
813    #define STR_GRAVE_ACCENT            "`"
814    #define STR_a                       "a"
815    #define STR_b                       "b"
816    #define STR_c                       "c"
817    #define STR_d                       "d"
818    #define STR_e                       "e"
819    #define STR_f                       "f"
820    #define STR_g                       "g"
821    #define STR_h                       "h"
822    #define STR_i                       "i"
823    #define STR_j                       "j"
824    #define STR_k                       "k"
825    #define STR_l                       "l"
826    #define STR_m                       "m"
827    #define STR_n                       "n"
828    #define STR_o                       "o"
829    #define STR_p                       "p"
830    #define STR_q                       "q"
831    #define STR_r                       "r"
832    #define STR_s                       "s"
833    #define STR_t                       "t"
834    #define STR_u                       "u"
835    #define STR_v                       "v"
836    #define STR_w                       "w"
837    #define STR_x                       "x"
838    #define STR_y                       "y"
839    #define STR_z                       "z"
840    #define STR_LEFT_CURLY_BRACKET      "{"
841    #define STR_VERTICAL_LINE           "|"
842    #define STR_RIGHT_CURLY_BRACKET     "}"
843    #define STR_TILDE                   "~"
844    
845    #define STRING_ACCEPT0              "ACCEPT\0"
846    #define STRING_COMMIT0              "COMMIT\0"
847    #define STRING_F0                   "F\0"
848    #define STRING_FAIL0                "FAIL\0"
849    #define STRING_PRUNE0               "PRUNE\0"
850    #define STRING_SKIP0                "SKIP\0"
851    #define STRING_THEN                 "THEN"
852    
853    #define STRING_alpha0               "alpha\0"
854    #define STRING_lower0               "lower\0"
855    #define STRING_upper0               "upper\0"
856    #define STRING_alnum0               "alnum\0"
857    #define STRING_ascii0               "ascii\0"
858    #define STRING_blank0               "blank\0"
859    #define STRING_cntrl0               "cntrl\0"
860    #define STRING_digit0               "digit\0"
861    #define STRING_graph0               "graph\0"
862    #define STRING_print0               "print\0"
863    #define STRING_punct0               "punct\0"
864    #define STRING_space0               "space\0"
865    #define STRING_word0                "word\0"
866    #define STRING_xdigit               "xdigit"
867    
868    #define STRING_DEFINE               "DEFINE"
869    
870    #define STRING_CR_RIGHTPAR          "CR)"
871    #define STRING_LF_RIGHTPAR          "LF)"
872    #define STRING_CRLF_RIGHTPAR        "CRLF)"
873    #define STRING_ANY_RIGHTPAR         "ANY)"
874    #define STRING_ANYCRLF_RIGHTPAR     "ANYCRLF)"
875    #define STRING_BSR_ANYCRLF_RIGHTPAR "BSR_ANYCRLF)"
876    #define STRING_BSR_UNICODE_RIGHTPAR "BSR_UNICODE)"
877    
878    #else  /* SUPPORT_UTF8 */
879    
880    /* UTF-8 support is enabled; always use UTF-8 (=ASCII) character codes. This
881    works in both modes non-EBCDIC platforms, and on EBCDIC platforms in UTF-8 mode
882    only. */
883    
884    #define CHAR_HT                     '\011'
885    #define CHAR_VT                     '\013'
886    #define CHAR_FF                     '\014'
887    #define CHAR_CR                     '\015'
888    #define CHAR_NL                     '\012'
889    #define CHAR_BS                     '\010'
890    #define CHAR_BEL                    '\007'
891    #define CHAR_ESC                    '\033'
892    #define CHAR_DEL                    '\177'
893    
894    #define CHAR_SPACE                  '\040'
895    #define CHAR_EXCLAMATION_MARK       '\041'
896    #define CHAR_QUOTATION_MARK         '\042'
897    #define CHAR_NUMBER_SIGN            '\043'
898    #define CHAR_DOLLAR_SIGN            '\044'
899    #define CHAR_PERCENT_SIGN           '\045'
900    #define CHAR_AMPERSAND              '\046'
901    #define CHAR_APOSTROPHE             '\047'
902    #define CHAR_LEFT_PARENTHESIS       '\050'
903    #define CHAR_RIGHT_PARENTHESIS      '\051'
904    #define CHAR_ASTERISK               '\052'
905    #define CHAR_PLUS                   '\053'
906    #define CHAR_COMMA                  '\054'
907    #define CHAR_MINUS                  '\055'
908    #define CHAR_DOT                    '\056'
909    #define CHAR_SLASH                  '\057'
910    #define CHAR_0                      '\060'
911    #define CHAR_1                      '\061'
912    #define CHAR_2                      '\062'
913    #define CHAR_3                      '\063'
914    #define CHAR_4                      '\064'
915    #define CHAR_5                      '\065'
916    #define CHAR_6                      '\066'
917    #define CHAR_7                      '\067'
918    #define CHAR_8                      '\070'
919    #define CHAR_9                      '\071'
920    #define CHAR_COLON                  '\072'
921    #define CHAR_SEMICOLON              '\073'
922    #define CHAR_LESS_THAN_SIGN         '\074'
923    #define CHAR_EQUALS_SIGN            '\075'
924    #define CHAR_GREATER_THAN_SIGN      '\076'
925    #define CHAR_QUESTION_MARK          '\077'
926    #define CHAR_COMMERCIAL_AT          '\100'
927    #define CHAR_A                      '\101'
928    #define CHAR_B                      '\102'
929    #define CHAR_C                      '\103'
930    #define CHAR_D                      '\104'
931    #define CHAR_E                      '\105'
932    #define CHAR_F                      '\106'
933    #define CHAR_G                      '\107'
934    #define CHAR_H                      '\110'
935    #define CHAR_I                      '\111'
936    #define CHAR_J                      '\112'
937    #define CHAR_K                      '\113'
938    #define CHAR_L                      '\114'
939    #define CHAR_M                      '\115'
940    #define CHAR_N                      '\116'
941    #define CHAR_O                      '\117'
942    #define CHAR_P                      '\120'
943    #define CHAR_Q                      '\121'
944    #define CHAR_R                      '\122'
945    #define CHAR_S                      '\123'
946    #define CHAR_T                      '\124'
947    #define CHAR_U                      '\125'
948    #define CHAR_V                      '\126'
949    #define CHAR_W                      '\127'
950    #define CHAR_X                      '\130'
951    #define CHAR_Y                      '\131'
952    #define CHAR_Z                      '\132'
953    #define CHAR_LEFT_SQUARE_BRACKET    '\133'
954    #define CHAR_BACKSLASH              '\134'
955    #define CHAR_RIGHT_SQUARE_BRACKET   '\135'
956    #define CHAR_CIRCUMFLEX_ACCENT      '\136'
957    #define CHAR_UNDERSCORE             '\137'
958    #define CHAR_GRAVE_ACCENT           '\140'
959    #define CHAR_a                      '\141'
960    #define CHAR_b                      '\142'
961    #define CHAR_c                      '\143'
962    #define CHAR_d                      '\144'
963    #define CHAR_e                      '\145'
964    #define CHAR_f                      '\146'
965    #define CHAR_g                      '\147'
966    #define CHAR_h                      '\150'
967    #define CHAR_i                      '\151'
968    #define CHAR_j                      '\152'
969    #define CHAR_k                      '\153'
970    #define CHAR_l                      '\154'
971    #define CHAR_m                      '\155'
972    #define CHAR_n                      '\156'
973    #define CHAR_o                      '\157'
974    #define CHAR_p                      '\160'
975    #define CHAR_q                      '\161'
976    #define CHAR_r                      '\162'
977    #define CHAR_s                      '\163'
978    #define CHAR_t                      '\164'
979    #define CHAR_u                      '\165'
980    #define CHAR_v                      '\166'
981    #define CHAR_w                      '\167'
982    #define CHAR_x                      '\170'
983    #define CHAR_y                      '\171'
984    #define CHAR_z                      '\172'
985    #define CHAR_LEFT_CURLY_BRACKET     '\173'
986    #define CHAR_VERTICAL_LINE          '\174'
987    #define CHAR_RIGHT_CURLY_BRACKET    '\175'
988    #define CHAR_TILDE                  '\176'
989    
990    #define STR_HT                      "\011"
991    #define STR_VT                      "\013"
992    #define STR_FF                      "\014"
993    #define STR_CR                      "\015"
994    #define STR_NL                      "\012"
995    #define STR_BS                      "\010"
996    #define STR_BEL                     "\007"
997    #define STR_ESC                     "\033"
998    #define STR_DEL                     "\177"
999    
1000    #define STR_SPACE                   "\040"
1001    #define STR_EXCLAMATION_MARK        "\041"
1002    #define STR_QUOTATION_MARK          "\042"
1003    #define STR_NUMBER_SIGN             "\043"
1004    #define STR_DOLLAR_SIGN             "\044"
1005    #define STR_PERCENT_SIGN            "\045"
1006    #define STR_AMPERSAND               "\046"
1007    #define STR_APOSTROPHE              "\047"
1008    #define STR_LEFT_PARENTHESIS        "\050"
1009    #define STR_RIGHT_PARENTHESIS       "\051"
1010    #define STR_ASTERISK                "\052"
1011    #define STR_PLUS                    "\053"
1012    #define STR_COMMA                   "\054"
1013    #define STR_MINUS                   "\055"
1014    #define STR_DOT                     "\056"
1015    #define STR_SLASH                   "\057"
1016    #define STR_0                       "\060"
1017    #define STR_1                       "\061"
1018    #define STR_2                       "\062"
1019    #define STR_3                       "\063"
1020    #define STR_4                       "\064"
1021    #define STR_5                       "\065"
1022    #define STR_6                       "\066"
1023    #define STR_7                       "\067"
1024    #define STR_8                       "\070"
1025    #define STR_9                       "\071"
1026    #define STR_COLON                   "\072"
1027    #define STR_SEMICOLON               "\073"
1028    #define STR_LESS_THAN_SIGN          "\074"
1029    #define STR_EQUALS_SIGN             "\075"
1030    #define STR_GREATER_THAN_SIGN       "\076"
1031    #define STR_QUESTION_MARK           "\077"
1032    #define STR_COMMERCIAL_AT           "\100"
1033    #define STR_A                       "\101"
1034    #define STR_B                       "\102"
1035    #define STR_C                       "\103"
1036    #define STR_D                       "\104"
1037    #define STR_E                       "\105"
1038    #define STR_F                       "\106"
1039    #define STR_G                       "\107"
1040    #define STR_H                       "\110"
1041    #define STR_I                       "\111"
1042    #define STR_J                       "\112"
1043    #define STR_K                       "\113"
1044    #define STR_L                       "\114"
1045    #define STR_M                       "\115"
1046    #define STR_N                       "\116"
1047    #define STR_O                       "\117"
1048    #define STR_P                       "\120"
1049    #define STR_Q                       "\121"
1050    #define STR_R                       "\122"
1051    #define STR_S                       "\123"
1052    #define STR_T                       "\124"
1053    #define STR_U                       "\125"
1054    #define STR_V                       "\126"
1055    #define STR_W                       "\127"
1056    #define STR_X                       "\130"
1057    #define STR_Y                       "\131"
1058    #define STR_Z                       "\132"
1059    #define STR_LEFT_SQUARE_BRACKET     "\133"
1060    #define STR_BACKSLASH               "\134"
1061    #define STR_RIGHT_SQUARE_BRACKET    "\135"
1062    #define STR_CIRCUMFLEX_ACCENT       "\136"
1063    #define STR_UNDERSCORE              "\137"
1064    #define STR_GRAVE_ACCENT            "\140"
1065    #define STR_a                       "\141"
1066    #define STR_b                       "\142"
1067    #define STR_c                       "\143"
1068    #define STR_d                       "\144"
1069    #define STR_e                       "\145"
1070    #define STR_f                       "\146"
1071    #define STR_g                       "\147"
1072    #define STR_h                       "\150"
1073    #define STR_i                       "\151"
1074    #define STR_j                       "\152"
1075    #define STR_k                       "\153"
1076    #define STR_l                       "\154"
1077    #define STR_m                       "\155"
1078    #define STR_n                       "\156"
1079    #define STR_o                       "\157"
1080    #define STR_p                       "\160"
1081    #define STR_q                       "\161"
1082    #define STR_r                       "\162"
1083    #define STR_s                       "\163"
1084    #define STR_t                       "\164"
1085    #define STR_u                       "\165"
1086    #define STR_v                       "\166"
1087    #define STR_w                       "\167"
1088    #define STR_x                       "\170"
1089    #define STR_y                       "\171"
1090    #define STR_z                       "\172"
1091    #define STR_LEFT_CURLY_BRACKET      "\173"
1092    #define STR_VERTICAL_LINE           "\174"
1093    #define STR_RIGHT_CURLY_BRACKET     "\175"
1094    #define STR_TILDE                   "\176"
1095    
1096    #define STRING_ACCEPT0              STR_A STR_C STR_C STR_E STR_P STR_T "\0"
1097    #define STRING_COMMIT0              STR_C STR_O STR_M STR_M STR_I STR_T "\0"
1098    #define STRING_F0                   STR_F "\0"
1099    #define STRING_FAIL0                STR_F STR_A STR_I STR_L "\0"
1100    #define STRING_PRUNE0               STR_P STR_R STR_U STR_N STR_E "\0"
1101    #define STRING_SKIP0                STR_S STR_K STR_I STR_P "\0"
1102    #define STRING_THEN                 STR_T STR_H STR_E STR_N
1103    
1104    #define STRING_alpha0               STR_a STR_l STR_p STR_h STR_a "\0"
1105    #define STRING_lower0               STR_l STR_o STR_w STR_e STR_r "\0"
1106    #define STRING_upper0               STR_u STR_p STR_p STR_e STR_r "\0"
1107    #define STRING_alnum0               STR_a STR_l STR_n STR_u STR_m "\0"
1108    #define STRING_ascii0               STR_a STR_s STR_c STR_i STR_i "\0"
1109    #define STRING_blank0               STR_b STR_l STR_a STR_n STR_k "\0"
1110    #define STRING_cntrl0               STR_c STR_n STR_t STR_r STR_l "\0"
1111    #define STRING_digit0               STR_d STR_i STR_g STR_i STR_t "\0"
1112    #define STRING_graph0               STR_g STR_r STR_a STR_p STR_h "\0"
1113    #define STRING_print0               STR_p STR_r STR_i STR_n STR_t "\0"
1114    #define STRING_punct0               STR_p STR_u STR_n STR_c STR_t "\0"
1115    #define STRING_space0               STR_s STR_p STR_a STR_c STR_e "\0"
1116    #define STRING_word0                STR_w STR_o STR_r STR_d       "\0"
1117    #define STRING_xdigit               STR_x STR_d STR_i STR_g STR_i STR_t
1118    
1119    #define STRING_DEFINE               STR_D STR_E STR_F STR_I STR_N STR_E
1120    
1121    #define STRING_CR_RIGHTPAR          STR_C STR_R STR_RIGHT_PARENTHESIS
1122    #define STRING_LF_RIGHTPAR          STR_L STR_F STR_RIGHT_PARENTHESIS
1123    #define STRING_CRLF_RIGHTPAR        STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1124    #define STRING_ANY_RIGHTPAR         STR_A STR_N STR_Y STR_RIGHT_PARENTHESIS
1125    #define STRING_ANYCRLF_RIGHTPAR     STR_A STR_N STR_Y STR_C STR_R STR_L STR_F STR_RIGHT_PARENTHESIS
1126    #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
1127    #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
1128    
1129  /* Escape items that are just an encoding of a particular data value. Note that  #endif  /* SUPPORT_UTF8 */
1130  ESC_n is defined as yet another macro, which is set in config.h to either \n  
1131  (the default) or \r (which some people want). */  /* Escape items that are just an encoding of a particular data value. */
1132    
1133  #ifndef ESC_e  #ifndef ESC_e
1134  #define ESC_e 27  #define ESC_e CHAR_ESC
1135  #endif  #endif
1136    
1137  #ifndef ESC_f  #ifndef ESC_f
1138  #define ESC_f '\f'  #define ESC_f CHAR_FF
1139  #endif  #endif
1140    
1141  #ifndef ESC_n  #ifndef ESC_n
1142  #define ESC_n NEWLINE  #define ESC_n CHAR_NL
1143  #endif  #endif
1144    
1145  #ifndef ESC_r  #ifndef ESC_r
1146  #define ESC_r '\r'  #define ESC_r CHAR_CR
1147  #endif  #endif
1148    
1149  /* 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
1150  (presumably because of all the others like size_t). */  (presumably because of all the others like size_t). */
1151    
1152  #ifndef ESC_tee  #ifndef ESC_tee
1153  #define ESC_tee '\t'  #define ESC_tee CHAR_HT
1154  #endif  #endif
1155    
1156  /* These are escaped items that aren't just an encoding of a particular data  /* Codes for different types of Unicode property */
1157  value such as \n. They must have non-zero values, as check_escape() returns  
1158  their negation. Also, they must appear in the same order as in the opcode  #define PT_ANY        0    /* Any property - matches all chars */
1159  definitions below, up to ESC_z. There's a dummy for OP_ANY because it  #define PT_LAMP       1    /* L& - the union of Lu, Ll, Lt */
1160  corresponds to "." rather than an escape sequence. The final one must be  #define PT_GC         2    /* General characteristic (e.g. L) */
1161  ESC_REF as subsequent values are used for \1, \2, \3, etc. There is are two  #define PT_PC         3    /* Particular characteristic (e.g. Lu) */
1162  tests in the code for an escape greater than ESC_b and less than ESC_Z to  #define PT_SC         4    /* Script (e.g. Han) */
 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 };  
1163    
1164  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that  /* Flag bits and data types for the extended class (OP_XCLASS) for classes that
1165  contain UTF-8 characters with values greater than 255. */  contain UTF-8 characters with values greater than 255. */
# Line 467  contain UTF-8 characters with values gre Line 1170  contain UTF-8 characters with values gre
1170  #define XCL_END       0    /* Marks end of individual items */  #define XCL_END       0    /* Marks end of individual items */
1171  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */  #define XCL_SINGLE    1    /* Single item (one multibyte char) follows */
1172  #define XCL_RANGE     2    /* A range (two multibyte chars) follows */  #define XCL_RANGE     2    /* A range (two multibyte chars) follows */
1173  #define XCL_PROP      3    /* Unicode property (one property code) follows */  #define XCL_PROP      3    /* Unicode property (2-byte property code follows) */
1174  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */  #define XCL_NOTPROP   4    /* Unicode inverted property (ditto) */
1175    
1176    /* These are escaped items that aren't just an encoding of a particular data
1177    value such as \n. They must have non-zero values, as check_escape() returns
1178    their negation. Also, they must appear in the same order as in the opcode
1179    definitions below, up to ESC_z. There's a dummy for OP_ANY because it
1180    corresponds to "." rather than an escape sequence, and another for OP_ALLANY
1181    (which is used for [^] in JavaScript compatibility mode).
1182    
1183    The final escape must be ESC_REF as subsequent values are used for
1184    backreferences (\1, \2, \3, etc). There are two tests in the code for an escape
1185    greater than ESC_b and less than ESC_Z to detect the types that may be
1186    repeated. These are the types that consume characters. If any new escapes are
1187    put in between that don't consume a character, that code will have to change.
1188    */
1189    
1190    enum { ESC_A = 1, ESC_G, ESC_K, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s,
1191           ESC_W, ESC_w, ESC_dum1, ESC_dum2, ESC_C, ESC_P, ESC_p, ESC_R, ESC_H,
1192           ESC_h, ESC_V, ESC_v, ESC_X, ESC_Z, ESC_z, ESC_E, ESC_Q, ESC_g, ESC_k,
1193           ESC_REF };
1194    
1195    
1196  /* Opcode table: OP_BRA must be last, as all values >= it are used for brackets  /* Opcode table: Starting from 1 (i.e. after OP_END), the values up to
 that extract substrings. Starting from 1 (i.e. after OP_END), the values up to  
1197  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.
1198  Note that whenever this list is updated, the two macro definitions that follow  
1199  must also be updated to match. */  *** NOTE NOTE NOTE *** Whenever this list is updated, the two macro definitions
1200    that follow must also be updated to match. There is also a table called
1201    "coptable" in pcre_dfa_exec.c that must be updated. */
1202    
1203  enum {  enum {
1204    OP_END,            /* 0 End of pattern */    OP_END,            /* 0 End of pattern */
# Line 484  enum { Line 1207  enum {
1207    
1208    OP_SOD,            /* 1 Start of data: \A */    OP_SOD,            /* 1 Start of data: \A */
1209    OP_SOM,            /* 2 Start of match (subject + offset): \G */    OP_SOM,            /* 2 Start of match (subject + offset): \G */
1210    OP_NOT_WORD_BOUNDARY,  /*  3 \B */    OP_SET_SOM,        /* 3 Set start of match (\K) */
1211    OP_WORD_BOUNDARY,      /*  4 \b */    OP_NOT_WORD_BOUNDARY,  /*  4 \B */
1212    OP_NOT_DIGIT,          /*  5 \D */    OP_WORD_BOUNDARY,      /*  5 \b */
1213    OP_DIGIT,              /*  6 \d */    OP_NOT_DIGIT,          /*  6 \D */
1214    OP_NOT_WHITESPACE,     /*  7 \S */    OP_DIGIT,              /*  7 \d */
1215    OP_WHITESPACE,         /*  8 \s */    OP_NOT_WHITESPACE,     /*  8 \S */
1216    OP_NOT_WORDCHAR,       /*  9 \W */    OP_WHITESPACE,         /*  9 \s */
1217    OP_WORDCHAR,           /* 10 \w */    OP_NOT_WORDCHAR,       /* 10 \W */
1218    OP_ANY,            /* 11 Match any character */    OP_WORDCHAR,           /* 11 \w */
1219    OP_ANYBYTE,        /* 12 Match any byte (\C); different to OP_ANY for UTF-8 */    OP_ANY,            /* 12 Match any character (subject to DOTALL) */
1220    OP_NOTPROP,        /* 13 \P (not Unicode property) */    OP_ALLANY,         /* 13 Match any character (not subject to DOTALL) */
1221    OP_PROP,           /* 14 \p (Unicode property) */    OP_ANYBYTE,        /* 14 Match any byte (\C); different to OP_ANY for UTF-8 */
1222    OP_EXTUNI,         /* 15 \X (extended Unicode sequence */    OP_NOTPROP,        /* 15 \P (not Unicode property) */
1223    OP_EODN,           /* 16 End of data or \n at end of data: \Z. */    OP_PROP,           /* 16 \p (Unicode property) */
1224    OP_EOD,            /* 17 End of data: \z */    OP_ANYNL,          /* 17 \R (any newline sequence) */
1225      OP_NOT_HSPACE,     /* 18 \H (not horizontal whitespace) */
1226    OP_OPT,            /* 18 Set runtime options */    OP_HSPACE,         /* 19 \h (horizontal whitespace) */
1227    OP_CIRC,           /* 19 Start of line - varies with multiline switch */    OP_NOT_VSPACE,     /* 20 \V (not vertical whitespace) */
1228    OP_DOLL,           /* 20 End of line - varies with multiline switch */    OP_VSPACE,         /* 21 \v (vertical whitespace) */
1229    OP_CHAR,           /* 21 Match one character, casefully */    OP_EXTUNI,         /* 22 \X (extended Unicode sequence */
1230    OP_CHARNC,         /* 22 Match one character, caselessly */    OP_EODN,           /* 23 End of data or \n at end of data: \Z. */
1231    OP_NOT,            /* 23 Match anything but the following char */    OP_EOD,            /* 24 End of data: \z */
1232    
1233    OP_STAR,           /* 24 The maximizing and minimizing versions of */    OP_OPT,            /* 25 Set runtime options */
1234    OP_MINSTAR,        /* 25 all these opcodes must come in pairs, with */    OP_CIRC,           /* 26 Start of line - varies with multiline switch */
1235    OP_PLUS,           /* 26 the minimizing one second. */    OP_DOLL,           /* 27 End of line - varies with multiline switch */
1236    OP_MINPLUS,        /* 27 This first set applies to single characters */    OP_CHAR,           /* 28 Match one character, casefully */
1237    OP_QUERY,          /* 28 */    OP_CHARNC,         /* 29 Match one character, caselessly */
1238    OP_MINQUERY,       /* 29 */    OP_NOT,            /* 30 Match one character, not the following one */
1239    OP_UPTO,           /* 30 From 0 to n matches */  
1240    OP_MINUPTO,        /* 31 */    OP_STAR,           /* 31 The maximizing and minimizing versions of */
1241    OP_EXACT,          /* 32 Exactly n matches */    OP_MINSTAR,        /* 32 these six opcodes must come in pairs, with */
1242      OP_PLUS,           /* 33 the minimizing one second. */
1243    OP_NOTSTAR,        /* 33 The maximizing and minimizing versions of */    OP_MINPLUS,        /* 34 This first set applies to single characters.*/
1244    OP_NOTMINSTAR,     /* 34 all these opcodes must come in pairs, with */    OP_QUERY,          /* 35 */
1245    OP_NOTPLUS,        /* 35 the minimizing one second. */    OP_MINQUERY,       /* 36 */
1246    OP_NOTMINPLUS,     /* 36 This set applies to "not" single characters */  
1247    OP_NOTQUERY,       /* 37 */    OP_UPTO,           /* 37 From 0 to n matches */
1248    OP_NOTMINQUERY,    /* 38 */    OP_MINUPTO,        /* 38 */
1249    OP_NOTUPTO,        /* 39 From 0 to n matches */    OP_EXACT,          /* 39 Exactly n matches */
1250    OP_NOTMINUPTO,     /* 40 */  
1251    OP_NOTEXACT,       /* 41 Exactly n matches */    OP_POSSTAR,        /* 40 Possessified star */
1252      OP_POSPLUS,        /* 41 Possessified plus */
1253    OP_TYPESTAR,       /* 42 The maximizing and minimizing versions of */    OP_POSQUERY,       /* 42 Posesssified query */
1254    OP_TYPEMINSTAR,    /* 43 all these opcodes must come in pairs, with */    OP_POSUPTO,        /* 43 Possessified upto */
1255    OP_TYPEPLUS,       /* 44 the minimizing one second. These codes must */  
1256    OP_TYPEMINPLUS,    /* 45 be in exactly the same order as those above. */    OP_NOTSTAR,        /* 44 The maximizing and minimizing versions of */
1257    OP_TYPEQUERY,      /* 46 This set applies to character types such as \d */    OP_NOTMINSTAR,     /* 45 these six opcodes must come in pairs, with */
1258    OP_TYPEMINQUERY,   /* 47 */    OP_NOTPLUS,        /* 46 the minimizing one second. They must be in */
1259    OP_TYPEUPTO,       /* 48 From 0 to n matches */    OP_NOTMINPLUS,     /* 47 exactly the same order as those above. */
1260    OP_TYPEMINUPTO,    /* 49 */    OP_NOTQUERY,       /* 48 This set applies to "not" single characters. */
1261    OP_TYPEEXACT,      /* 50 Exactly n matches */    OP_NOTMINQUERY,    /* 49 */
1262    
1263    OP_CRSTAR,         /* 51 The maximizing and minimizing versions of */    OP_NOTUPTO,        /* 50 From 0 to n matches */
1264    OP_CRMINSTAR,      /* 52 all these opcodes must come in pairs, with */    OP_NOTMINUPTO,     /* 51 */
1265    OP_CRPLUS,         /* 53 the minimizing one second. These codes must */    OP_NOTEXACT,       /* 52 Exactly n matches */
1266    OP_CRMINPLUS,      /* 54 be in exactly the same order as those above. */  
1267    OP_CRQUERY,        /* 55 These are for character classes and back refs */    OP_NOTPOSSTAR,     /* 53 Possessified versions */
1268    OP_CRMINQUERY,     /* 56 */    OP_NOTPOSPLUS,     /* 54 */
1269    OP_CRRANGE,        /* 57 These are different to the three sets above. */    OP_NOTPOSQUERY,    /* 55 */
1270    OP_CRMINRANGE,     /* 58 */    OP_NOTPOSUPTO,     /* 56 */
1271    
1272      OP_TYPESTAR,       /* 57 The maximizing and minimizing versions of */
1273      OP_TYPEMINSTAR,    /* 58 these six opcodes must come in pairs, with */
1274      OP_TYPEPLUS,       /* 59 the minimizing one second. These codes must */
1275      OP_TYPEMINPLUS,    /* 60 be in exactly the same order as those above. */
1276      OP_TYPEQUERY,      /* 61 This set applies to character types such as \d */
1277      OP_TYPEMINQUERY,   /* 62 */
1278    
1279      OP_TYPEUPTO,       /* 63 From 0 to n matches */
1280      OP_TYPEMINUPTO,    /* 64 */
1281      OP_TYPEEXACT,      /* 65 Exactly n matches */
1282    
1283      OP_TYPEPOSSTAR,    /* 66 Possessified versions */
1284      OP_TYPEPOSPLUS,    /* 67 */
1285      OP_TYPEPOSQUERY,   /* 68 */
1286      OP_TYPEPOSUPTO,    /* 69 */
1287    
1288      OP_CRSTAR,         /* 70 The maximizing and minimizing versions of */
1289      OP_CRMINSTAR,      /* 71 all these opcodes must come in pairs, with */
1290      OP_CRPLUS,         /* 72 the minimizing one second. These codes must */
1291      OP_CRMINPLUS,      /* 73 be in exactly the same order as those above. */
1292      OP_CRQUERY,        /* 74 These are for character classes and back refs */
1293      OP_CRMINQUERY,     /* 75 */
1294      OP_CRRANGE,        /* 76 These are different to the three sets above. */
1295      OP_CRMINRANGE,     /* 77 */
1296    
1297    OP_CLASS,          /* 59 Match a character class, chars < 256 only */    OP_CLASS,          /* 78 Match a character class, chars < 256 only */
1298    OP_NCLASS,         /* 60 Same, but the bitmap was created from a negative    OP_NCLASS,         /* 79 Same, but the bitmap was created from a negative
1299                             class - the difference is relevant only when a UTF-8                             class - the difference is relevant only when a UTF-8
1300                             character > 255 is encountered. */                             character > 255 is encountered. */
1301    
1302    OP_XCLASS,         /* 61 Extended class for handling UTF-8 chars within the    OP_XCLASS,         /* 80 Extended class for handling UTF-8 chars within the
1303                             class. This does both positive and negative. */                             class. This does both positive and negative. */
1304    
1305    OP_REF,            /* 62 Match a back reference */    OP_REF,            /* 81 Match a back reference */
1306    OP_RECURSE,        /* 63 Match a numbered subpattern (possibly recursive) */    OP_RECURSE,        /* 82 Match a numbered subpattern (possibly recursive) */
1307    OP_CALLOUT,        /* 64 Call out to external function if provided */    OP_CALLOUT,        /* 83 Call out to external function if provided */
1308    
1309    OP_ALT,            /* 65 Start of alternation */    OP_ALT,            /* 84 Start of alternation */
1310    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 */
1311    OP_KETRMAX,        /* 67 These two must remain together and in this */    OP_KETRMAX,        /* 86 These two must remain together and in this */
1312    OP_KETRMIN,        /* 68 order. They are for groups the repeat for ever. */    OP_KETRMIN,        /* 87 order. They are for groups the repeat for ever. */
   
   /* The assertions must come before ONCE and COND */  
   
   OP_ASSERT,         /* 69 Positive lookahead */  
   OP_ASSERT_NOT,     /* 70 Negative lookahead */  
   OP_ASSERTBACK,     /* 71 Positive lookbehind */  
   OP_ASSERTBACK_NOT, /* 72 Negative lookbehind */  
   OP_REVERSE,        /* 73 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. */  
 };  
1313    
1314  /* WARNING WARNING WARNING: There is an implicit assumption in pcre.c and    /* The assertions must come before BRA, CBRA, ONCE, and COND.*/
 study.c that all opcodes are less than 128 in value. This makes handling UTF-8  
 character sequences easier. */  
   
 /* The highest extraction number before we have to start using additional  
 bytes. (Originally PCRE didn't have support for extraction counts highter than  
 this number.) The value is limited by the number of opcodes left after OP_BRA,  
 i.e. 255 - OP_BRA. We actually set it a bit lower to leave room for additional  
 opcodes. */  
1315    
1316  #define EXTRACT_BASIC_MAX  100    OP_ASSERT,         /* 88 Positive lookahead */
1317      OP_ASSERT_NOT,     /* 89 Negative lookahead */
1318      OP_ASSERTBACK,     /* 90 Positive lookbehind */
1319      OP_ASSERTBACK_NOT, /* 91 Negative lookbehind */
1320      OP_REVERSE,        /* 92 Move pointer back - used in lookbehind assertions */
1321    
1322      /* ONCE, BRA, CBRA, and COND must come after the assertions, with ONCE first,
1323      as there's a test for >= ONCE for a subpattern that isn't an assertion. */
1324    
1325      OP_ONCE,           /* 93 Atomic group */
1326      OP_BRA,            /* 94 Start of non-capturing bracket */
1327      OP_CBRA,           /* 95 Start of capturing bracket */
1328      OP_COND,           /* 96 Conditional group */
1329    
1330      /* These three must follow the previous three, in the same order. There's a
1331      check for >= SBRA to distinguish the two sets. */
1332    
1333      OP_SBRA,           /* 97 Start of non-capturing bracket, check empty  */
1334      OP_SCBRA,          /* 98 Start of capturing bracket, check empty */
1335      OP_SCOND,          /* 99 Conditional group, check empty */
1336    
1337      OP_CREF,           /* 100 Used to hold a capture number as condition */
1338      OP_RREF,           /* 101 Used to hold a recursion number as condition */
1339      OP_DEF,            /* 102 The DEFINE condition */
1340    
1341      OP_BRAZERO,        /* 103 These two must remain together and in this */
1342      OP_BRAMINZERO,     /* 104 order. */
1343    
1344      /* These are backtracking control verbs */
1345    
1346      OP_PRUNE,          /* 105 */
1347      OP_SKIP,           /* 106 */
1348      OP_THEN,           /* 107 */
1349      OP_COMMIT,         /* 108 */
1350    
1351      /* These are forced failure and success verbs */
1352    
1353      OP_FAIL,           /* 109 */
1354      OP_ACCEPT,         /* 110 */
1355    
1356      /* This is used to skip a subpattern with a {0} quantifier */
1357    
1358      OP_SKIPZERO        /* 111 */
1359    };
1360    
1361    
1362  /* 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
1363  for debugging. The macro is referenced only in pcre_printint.c. */  for debugging. The macro is referenced only in pcre_printint.c. */
1364    
1365  #define OP_NAME_LIST \  #define OP_NAME_LIST \
1366    "End", "\\A", "\\G", "\\B", "\\b", "\\D", "\\d",                \    "End", "\\A", "\\G", "\\K", "\\B", "\\b", "\\D", "\\d",         \
1367    "\\S", "\\s", "\\W", "\\w", "Any", "Anybyte",                   \    "\\S", "\\s", "\\W", "\\w", "Any", "AllAny", "Anybyte",         \
1368    "notprop", "prop", "extuni",                                    \    "notprop", "prop", "\\R", "\\H", "\\h", "\\V", "\\v",           \
1369    "\\Z", "\\z",                                                   \    "extuni",  "\\Z", "\\z",                                        \
1370    "Opt", "^", "$", "char", "charnc", "not",                       \    "Opt", "^", "$", "char", "charnc", "not",                       \
1371    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1372      "*+","++", "?+", "{",                                           \
1373    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1374      "*+","++", "?+", "{",                                           \
1375    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \    "*", "*?", "+", "+?", "?", "??", "{", "{", "{",                 \
1376      "*+","++", "?+", "{",                                           \
1377    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \    "*", "*?", "+", "+?", "?", "??", "{", "{",                      \
1378    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \    "class", "nclass", "xclass", "Ref", "Recurse", "Callout",       \
1379    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \    "Alt", "Ket", "KetRmax", "KetRmin", "Assert", "Assert not",     \
1380    "AssertB", "AssertB not", "Reverse", "Once", "Cond", "Cond ref",\    "AssertB", "AssertB not", "Reverse",                            \
1381    "Brazero", "Braminzero", "Branumber", "Bra"    "Once", "Bra", "CBra", "Cond", "SBra", "SCBra", "SCond",        \
1382      "Cond ref", "Cond rec", "Cond def", "Brazero", "Braminzero",    \
1383      "*PRUNE", "*SKIP", "*THEN", "*COMMIT", "*FAIL", "*ACCEPT",      \
1384      "Skip zero"
1385    
1386    
1387  /* 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 632  in UTF-8 mode. The code that uses this t Line 1395  in UTF-8 mode. The code that uses this t
1395    
1396  #define OP_LENGTHS \  #define OP_LENGTHS \
1397    1,                             /* End                                    */ \    1,                             /* End                                    */ \
1398    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                     */ \
1399    1, 1,                          /* Any, Anybyte                           */ \    1, 1, 1, 1, 1, 1,              /* \D, \d, \S, \s, \W, \w                 */ \
1400    2, 2, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \    1, 1, 1,                       /* Any, AllAny, Anybyte                   */ \
1401      3, 3, 1,                       /* NOTPROP, PROP, EXTUNI                  */ \
1402      1, 1, 1, 1, 1,                 /* \R, \H, \h, \V, \v                     */ \
1403    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \    1, 1, 2, 1, 1,                 /* \Z, \z, Opt, ^, $                      */ \
1404    2,                             /* Char  - the minimum length             */ \    2,                             /* Char  - the minimum length             */ \
1405    2,                             /* Charnc  - the minimum length           */ \    2,                             /* Charnc  - the minimum length           */ \
# Line 642  in UTF-8 mode. The code that uses this t Line 1407  in UTF-8 mode. The code that uses this t
1407    /* Positive single-char repeats                            ** These are  */ \    /* Positive single-char repeats                            ** These are  */ \
1408    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \    2, 2, 2, 2, 2, 2,              /* *, *?, +, +?, ?, ??      ** minima in  */ \
1409    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \    4, 4, 4,                       /* upto, minupto, exact     ** UTF-8 mode */ \
1410      2, 2, 2, 4,                    /* *+, ++, ?+, upto+                      */ \
1411    /* Negative single-char repeats - only for chars < 256                   */ \    /* Negative single-char repeats - only for chars < 256                   */ \
1412    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \    2, 2, 2, 2, 2, 2,              /* NOT *, *?, +, +?, ?, ??                */ \
1413    4, 4, 4,                       /* NOT upto, minupto, exact               */ \    4, 4, 4,                       /* NOT upto, minupto, exact               */ \
1414      2, 2, 2, 4,                    /* Possessive *, +, ?, upto               */ \
1415    /* Positive type repeats                                                 */ \    /* Positive type repeats                                                 */ \
1416    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \    2, 2, 2, 2, 2, 2,              /* Type *, *?, +, +?, ?, ??               */ \
1417    4, 4, 4,                       /* Type upto, minupto, exact              */ \    4, 4, 4,                       /* Type upto, minupto, exact              */ \
1418      2, 2, 2, 4,                    /* Possessive *+, ++, ?+, upto+           */ \
1419    /* Character class & ref repeats                                         */ \    /* Character class & ref repeats                                         */ \
1420    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \    1, 1, 1, 1, 1, 1,              /* *, *?, +, +?, ?, ??                    */ \
1421    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \    5, 5,                          /* CRRANGE, CRMINRANGE                    */ \
# Line 666  in UTF-8 mode. The code that uses this t Line 1434  in UTF-8 mode. The code that uses this t
1434    1+LINK_SIZE,                   /* Assert behind                          */ \    1+LINK_SIZE,                   /* Assert behind                          */ \
1435    1+LINK_SIZE,                   /* Assert behind not                      */ \    1+LINK_SIZE,                   /* Assert behind not                      */ \
1436    1+LINK_SIZE,                   /* Reverse                                */ \    1+LINK_SIZE,                   /* Reverse                                */ \
1437    1+LINK_SIZE,                   /* Once                                   */ \    1+LINK_SIZE,                   /* ONCE                                   */ \
1438      1+LINK_SIZE,                   /* BRA                                    */ \
1439      3+LINK_SIZE,                   /* CBRA                                   */ \
1440    1+LINK_SIZE,                   /* COND                                   */ \    1+LINK_SIZE,                   /* COND                                   */ \
1441      1+LINK_SIZE,                   /* SBRA                                   */ \
1442      3+LINK_SIZE,                   /* SCBRA                                  */ \
1443      1+LINK_SIZE,                   /* SCOND                                  */ \
1444    3,                             /* CREF                                   */ \    3,                             /* CREF                                   */ \
1445      3,                             /* RREF                                   */ \
1446      1,                             /* DEF                                    */ \
1447    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \    1, 1,                          /* BRAZERO, BRAMINZERO                    */ \
1448    3,                             /* BRANUMBER                              */ \    1, 1, 1, 1,                    /* PRUNE, SKIP, THEN, COMMIT,             */ \
1449    1+LINK_SIZE                    /* BRA                                    */ \    1, 1, 1                        /* FAIL, ACCEPT, SKIPZERO                 */
1450    
1451    
1452  /* A magic value for OP_CREF to indicate the "in recursion" condition. */  /* A magic value for OP_RREF to indicate the "any recursion" condition. */
1453    
1454  #define CREF_RECURSE  0xffff  #define RREF_ANY  0xffff
1455    
1456  /* 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
1457  tracked. */  tracked. */
# Line 685  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4, Line 1460  enum { ERR0,  ERR1,  ERR2,  ERR3,  ERR4,
1460         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,         ERR10, ERR11, ERR12, ERR13, ERR14, ERR15, ERR16, ERR17, ERR18, ERR19,
1461         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,         ERR20, ERR21, ERR22, ERR23, ERR24, ERR25, ERR26, ERR27, ERR28, ERR29,
1462         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,         ERR30, ERR31, ERR32, ERR33, ERR34, ERR35, ERR36, ERR37, ERR38, ERR39,
1463         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47 };         ERR40, ERR41, ERR42, ERR43, ERR44, ERR45, ERR46, ERR47, ERR48, ERR49,
1464           ERR50, ERR51, ERR52, ERR53, ERR54, ERR55, ERR56, ERR57, ERR58, ERR59,
1465           ERR60, ERR61, ERR62, ERR63, ERR64 };
1466    
1467  /* 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
1468  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 707  NOTE NOTE NOTE: Line 1484  NOTE NOTE NOTE:
1484  typedef struct real_pcre {  typedef struct real_pcre {
1485    pcre_uint32 magic_number;    pcre_uint32 magic_number;
1486    pcre_uint32 size;               /* Total that was malloced */    pcre_uint32 size;               /* Total that was malloced */
1487    pcre_uint32 options;    pcre_uint32 options;            /* Public options */
1488    pcre_uint32 dummy1;             /* For future use, maybe */    pcre_uint16 flags;              /* Private flags */
1489      pcre_uint16 dummy1;             /* For future use */
1490    pcre_uint16 top_bracket;    pcre_uint16 top_bracket;
1491    pcre_uint16 top_backref;    pcre_uint16 top_backref;
1492    pcre_uint16 first_byte;    pcre_uint16 first_byte;
# Line 740  typedef struct compile_data { Line 1517  typedef struct compile_data {
1517    const uschar *fcc;            /* Points to case-flipping table */    const uschar *fcc;            /* Points to case-flipping table */
1518    const uschar *cbits;          /* Points to character type table */    const uschar *cbits;          /* Points to character type table */
1519    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
1520      const uschar *start_workspace;/* The start of working space */
1521    const uschar *start_code;     /* The start of the compiled code */    const uschar *start_code;     /* The start of the compiled code */
1522    const uschar *start_pattern;  /* The start of the pattern */    const uschar *start_pattern;  /* The start of the pattern */
1523      const uschar *end_pattern;    /* The end of the pattern */
1524      uschar *hwm;                  /* High watermark of workspace */
1525    uschar *name_table;           /* The name/number table */    uschar *name_table;           /* The name/number table */
1526    int  names_found;             /* Number of entries so far */    int  names_found;             /* Number of entries so far */
1527    int  name_entry_size;         /* Size of each entry */    int  name_entry_size;         /* Size of each entry */
1528      int  bracount;                /* Count of capturing parens as we compile */
1529      int  final_bracount;          /* Saved value after first pass */
1530    int  top_backref;             /* Maximum back reference */    int  top_backref;             /* Maximum back reference */
1531    unsigned int backref_map;     /* Bitmap of low back refs */    unsigned int backref_map;     /* Bitmap of low back refs */
1532      int  external_options;        /* External (initial) options */
1533      int  external_flags;          /* External flag bits to be set */
1534    int  req_varyopt;             /* "After variable item" flag for reqbyte */    int  req_varyopt;             /* "After variable item" flag for reqbyte */
1535    BOOL nopartial;               /* Set TRUE if partial won't work */    BOOL had_accept;              /* (*ACCEPT) encountered */
1536      int  nltype;                  /* Newline type */
1537      int  nllen;                   /* Newline string length */
1538      uschar nl[4];                 /* Newline string when fixed length */
1539  } compile_data;  } compile_data;
1540    
1541  /* Structure for maintaining a chain of pointers to the currently incomplete  /* Structure for maintaining a chain of pointers to the currently incomplete
# Line 766  typedef struct recursion_info { Line 1553  typedef struct recursion_info {
1553    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */    struct recursion_info *prevrec; /* Previous recursion record (or NULL) */
1554    int group_num;                /* Number of group that was called */    int group_num;                /* Number of group that was called */
1555    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 */
1556    const uschar *save_start;     /* Old value of md->start_match */    USPTR save_start;             /* Old value of mstart */
1557    int *offset_save;             /* Pointer to start of saved offsets */    int *offset_save;             /* Pointer to start of saved offsets */
1558    int saved_max;                /* Number of saved offsets */    int saved_max;                /* Number of saved offsets */
1559  } recursion_info;  } recursion_info;
1560    
1561  /* 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
1562  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
1563  pcre.c, close to the match() function, so that it is easy to keep it in step  has been matched by a subpattern - to break infinite loops. */
1564  with any changes of local variable. However, the pointer to the current frame  
1565  must be saved in some "static" place over a longjmp(). We declare the  typedef struct eptrblock {
1566  structure here so that we can put a pointer in the match_data structure.    struct eptrblock *epb_prev;
1567  NOTE: This isn't used for a "normal" compilation of pcre. */    USPTR epb_saved_eptr;
1568    } eptrblock;
1569    
 struct heapframe;  
1570    
1571  /* Structure for passing "static" information around between the functions  /* Structure for passing "static" information around between the functions
1572  doing traditional NFA matching, so that they are thread-safe. */  doing traditional NFA matching, so that they are thread-safe. */
1573    
1574  typedef struct match_data {  typedef struct match_data {
1575    unsigned long int match_call_count; /* As it says */    unsigned long int match_call_count;      /* As it says */
1576    unsigned long int match_limit;/* As it says */    unsigned long int match_limit;           /* As it says */
1577      unsigned long int match_limit_recursion; /* As it says */
1578    int   *offset_vector;         /* Offset vector */    int   *offset_vector;         /* Offset vector */
1579    int    offset_end;            /* One past the end */    int    offset_end;            /* One past the end */
1580    int    offset_max;            /* The maximum usable for return data */    int    offset_max;            /* The maximum usable for return data */
1581      int    nltype;                /* Newline type */
1582      int    nllen;                 /* Newline string length */
1583      uschar nl[4];                 /* Newline string when fixed */
1584    const uschar *lcc;            /* Points to lower casing table */    const uschar *lcc;            /* Points to lower casing table */
1585    const uschar *ctypes;         /* Points to table of type maps */    const uschar *ctypes;         /* Points to table of type maps */
1586    BOOL   offset_overflow;       /* Set if too many extractions */    BOOL   offset_overflow;       /* Set if too many extractions */
1587    BOOL   notbol;                /* NOTBOL flag */    BOOL   notbol;                /* NOTBOL flag */
1588    BOOL   noteol;                /* NOTEOL flag */    BOOL   noteol;                /* NOTEOL flag */
1589    BOOL   utf8;                  /* UTF8 flag */    BOOL   utf8;                  /* UTF8 flag */
1590      BOOL   jscript_compat;        /* JAVASCRIPT_COMPAT flag */
1591    BOOL   endonly;               /* Dollar not before final \n */    BOOL   endonly;               /* Dollar not before final \n */
1592    BOOL   notempty;              /* Empty string match not wanted */    BOOL   notempty;              /* Empty string match not wanted */
1593    BOOL   partial;               /* PARTIAL flag */    BOOL   partial;               /* PARTIAL flag */
1594    BOOL   hitend;                /* Hit the end of the subject at some point */    BOOL   hitend;                /* Hit the end of the subject at some point */
1595      BOOL   bsr_anycrlf;           /* \R is just any CRLF, not full Unicode */
1596    const uschar *start_code;     /* For use when recursing */    const uschar *start_code;     /* For use when recursing */
1597    const uschar *start_subject;  /* Start of the subject string */    USPTR  start_subject;         /* Start of the subject string */
1598    const uschar *end_subject;    /* End of the subject string */    USPTR  end_subject;           /* End of the subject string */
1599    const uschar *start_match;    /* Start of this match attempt */    USPTR  start_match_ptr;       /* Start of matched string */
1600    const uschar *end_match_ptr;  /* Subject position at end match */    USPTR  end_match_ptr;         /* Subject position at end match */
1601    int    end_offset_top;        /* Highwater mark at end of match */    int    end_offset_top;        /* Highwater mark at end of match */
1602    int    capture_last;          /* Most recent capture number */    int    capture_last;          /* Most recent capture number */
1603    int    start_offset;          /* The start offset value */    int    start_offset;          /* The start offset value */
1604      eptrblock *eptrchain;         /* Chain of eptrblocks for tail recursions */
1605      int    eptrn;                 /* Next free eptrblock */
1606    recursion_info *recursive;    /* Linked list of recursion data */    recursion_info *recursive;    /* Linked list of recursion data */
1607    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 */  
1608  } match_data;  } match_data;
1609    
1610  /* 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 823  typedef struct dfa_match_data { Line 1617  typedef struct dfa_match_data {
1617    const uschar *tables;         /* Character tables */    const uschar *tables;         /* Character tables */
1618    int   moptions;               /* Match options */    int   moptions;               /* Match options */
1619    int   poptions;               /* Pattern options */    int   poptions;               /* Pattern options */
1620      int    nltype;                /* Newline type */
1621      int    nllen;                 /* Newline string length */
1622      uschar nl[4];                 /* Newline string when fixed */
1623    void  *callout_data;          /* To pass back to callouts */    void  *callout_data;          /* To pass back to callouts */
1624  } dfa_match_data;  } dfa_match_data;
1625    
# Line 832  typedef struct dfa_match_data { Line 1629  typedef struct dfa_match_data {
1629  #define ctype_letter  0x02  #define ctype_letter  0x02
1630  #define ctype_digit   0x04  #define ctype_digit   0x04
1631  #define ctype_xdigit  0x08  #define ctype_xdigit  0x08
1632  #define ctype_word    0x10   /* alphameric or '_' */  #define ctype_word    0x10   /* alphanumeric or '_' */
1633  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */  #define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */
1634    
1635  /* 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 859  total length. */ Line 1656  total length. */
1656  #define ctypes_offset (cbits_offset + cbit_length)  #define ctypes_offset (cbits_offset + cbit_length)
1657  #define tables_length (ctypes_offset + 256)  #define tables_length (ctypes_offset + 256)
1658    
1659  /* Layout of the UCP type table that translates property names into codes for  /* Layout of the UCP type table that translates property names into types and
1660  pcre_ucp_findchar(). */  codes. Each entry used to point directly to a name, but to reduce the number of
1661    relocations in shared libraries, it now has an offset into a single string
1662    instead. */
1663    
1664  typedef struct {  typedef struct {
1665    const char *name;    pcre_uint16 name_offset;
1666    int value;    pcre_uint16 type;
1667      pcre_uint16 value;
1668  } ucp_type_table;  } ucp_type_table;
1669    
1670    
# Line 880  extern const uschar _pcre_utf8_table4[]; Line 1680  extern const uschar _pcre_utf8_table4[];
1680    
1681  extern const int    _pcre_utf8_table1_size;  extern const int    _pcre_utf8_table1_size;
1682    
1683    extern const char   _pcre_utt_names[];
1684  extern const ucp_type_table _pcre_utt[];  extern const ucp_type_table _pcre_utt[];
1685  extern const int _pcre_utt_size;  extern const int _pcre_utt_size;
1686    
# Line 892  extern const uschar _pcre_OP_lengths[]; Line 1693  extern const uschar _pcre_OP_lengths[];
1693  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
1694  sense, but are not part of the PCRE public API. */  sense, but are not part of the PCRE public API. */
1695    
1696  extern int         _pcre_ord2utf8(int, uschar *);  extern BOOL         _pcre_is_newline(const uschar *, int, const uschar *,
1697  extern real_pcre * _pcre_try_flipped(const real_pcre *, real_pcre *,                        int *, BOOL);
1698                       const pcre_study_data *, pcre_study_data *);  extern int          _pcre_ord2utf8(int, uschar *);
1699  extern int         _pcre_ucp_findchar(const int, int *, int *);  extern real_pcre   *_pcre_try_flipped(const real_pcre *, real_pcre *,
1700  extern int         _pcre_valid_utf8(const uschar *, int);                        const pcre_study_data *, pcre_study_data *);
1701  extern BOOL        _pcre_xclass(int, const uschar *);  extern int          _pcre_valid_utf8(const uschar *, int);
1702    extern BOOL         _pcre_was_newline(const uschar *, int, const uschar *,
1703                          int *, BOOL);
1704    extern BOOL         _pcre_xclass(int, const uschar *);
1705    
1706    
1707    /* Unicode character database (UCD) */
1708    
1709    typedef struct {
1710      uschar script;
1711      uschar chartype;
1712      pcre_int32 other_case;
1713    } ucd_record;
1714    
1715    extern const ucd_record  _pcre_ucd_records[];
1716    extern const uschar      _pcre_ucd_stage1[];
1717    extern const pcre_uint16 _pcre_ucd_stage2[];
1718    extern const int         _pcre_ucp_gentype[];
1719    
1720    
1721    /* UCD access macros */
1722    
1723    #define UCD_BLOCK_SIZE 128
1724    #define GET_UCD(ch) (_pcre_ucd_records + \
1725            _pcre_ucd_stage2[_pcre_ucd_stage1[(ch) / UCD_BLOCK_SIZE] * \
1726            UCD_BLOCK_SIZE + ch % UCD_BLOCK_SIZE])
1727    
1728    #define UCD_CHARTYPE(ch)  GET_UCD(ch)->chartype
1729    #define UCD_SCRIPT(ch)    GET_UCD(ch)->script
1730    #define UCD_CATEGORY(ch)  _pcre_ucp_gentype[UCD_CHARTYPE(ch)]
1731    #define UCD_OTHERCASE(ch) (ch + GET_UCD(ch)->other_case)
1732    
1733  #endif  #endif
1734    

Legend:
Removed from v.85  
changed lines
  Added in v.392

  ViewVC Help
Powered by ViewVC 1.1.5